International Microbiology

CONTENTS International Microbiology (2015) 18:1-70 ISSN (print): 1139-6709. e-ISSN: 1618-1095 www.im.microbios.org

Volume 18, Number 1, March 2015

RESEARCH REVIEW

Kristiansen JE, Dastidar SG, Palchoudhuri S, Roy DS, Das S, Hendricks O, Christensen JB Phenothiazines as a solution for multidrug resistant tuberculosis: From the origin to present

1

RESEARCH ARTICLES

Montes-Borrego M, Lopes JRS, Jiménez-Díaz RM, Landa BB Combined use of a new SNP-based assay and multilocus SSR markers to assess genetic diversity of Xylella fastidiosa subsp. pauca infecting citrus and coffee plants Guirao-Abad JP, González-Párraga P, Argüelles JC Strong correlation between the antifungal effect of amphotericin B and its inhibitory action on germ-tube formation in a Candida albicans URA+ strain Velasco R, Ordóñez JA, Cambero MI, Cabeza MC Use of E-beam radiation to eliminate Listeria monocytogenes from surface mould cheese

13

25

33

Polo D, García-Fernández I, Fernández-Ibáñez P, Romalde JL Solar water disinfection (SODIS): Impact on hepatitis A virus and on a human Norovirus surrogate under natural solar conditions

41

Orús P, Gomez-Perez L, Leranoz S, Berlanga M Increasing antibiotic resistance in preservative-tolerant bacterial strains isolated from cosmetic products

51

Huang HY, Tang YJ, King VAE, Chou JW, Tsen JH Properties of Lactobacillus reuteri chitosan-calcium-alginate encapsulation under simulated gastrointestinal conditions

61

Journal Citations Reports 5-year Impact Factor of International Microbiology is 2,10.

The journal is covered in several leading abstracting and indexing databases, including the following ones: Agricultural & Environmental Bio­­technology Abstracts; ASFA/Aquatic Sciences & Fisheries Abstracts; BIOSIS; CAB Abstracts; Chemical Abstracts; SCOPUS; Current Contents/Agriculture, Biology & Environmental Sciences; EBSCO; EMBASE/Elsevier Bibliographic Databases; Food Science & Technology Abstracts; ICYT/CINDOC; IBECS/ BNCS; ISI Alerting Services; MEDLINE/Index Medicus; Latindex; MedBioWorld; PubMed; SciELO-Spain; Science Citation Index Expanded; SciSearch.

A1

Front cover legends Upper left. Electron micrograph showing morphology of bacteriophages that infect Rhizobium etli. They were obtained from rhizosphere soil of bean plants from agricultural lands in Mexico using an enrichment method. Micrograph by Víctor González, Evolutive Genomics, Center of Genomic Sciences, UNAM, Cuernavaca, Morelos, México. (Magnification, 200,000×) Upper right. Darkfield micrograph of the cyanobacterium Nostoc sp., isolated from a freshwater pond. Note the differentiated cells known as heterocysts that fix atmospheric nitrogen. Photo by Rubén Duro, Center for Microbiological Research (CIM), Barcelona. (Magnification, 1000×) Center. Symptomatic branch (left) with shortened internodes and smaller leaves showing tip discoloration or necrosis in a coffee tree affected by coffee leaf scorch (CLS), a serious disease caused by plant pathogen bacterium Xylella fastidiosa. Healthy plant (right). Photographs by JRS Lopes, Dept. Entomology and Acarology, ESALQ/University of Sao Paulo, Piracicaba, SP, Brazil. Snapshot Nikon Camera, CoolPix 3700.

[See article by Montes-Borrego et al., pp 13-24, this issue]

Lower right. Photonic micrograph of spores of the fungus Alternaria. This fungus is a major aeroallergen in many parts of the world. Sensitivity to Alternaria has been increasingly recognized as a risk factor for the development and persistence of asthma. It is most common as an outdoor mold, as it thrives on various types of plants–including the black rot commonly seen on tomato fruit. by Rubén Duro, CIM. (Magnification, 1000×)

Lower left. Darkfield microsgraph of the unicellular cilliated Paramecium sp. Paramecia are widespread in freshwater, brackish and marine environments and are often very abundant in stagnant basins and ponds. Some species of Paramecium form mutualistic relationships with other organisms. Paramecium bursaria and P. chlorelligerum harbor endosymbiotic green algae, from which they obtain nutrients and protection from predators. Photo by Rubén Duro, CIM. (Magnification, 1000×)

Back cover: Pioneers in Microbiology Pedro Gutiérrez Igaravídez (1871–1935), Puerto Rico Pedro Gutiérrez Igaravídez (1871–1935), born in San Juan, was a Puerto Rican physician who left his successful private medical practice to collaborate with Bailey K. Ashford (1873–1934) to fight endemic tropical diseases in Puerto Rico, mainly uncinariasis. (Ashford had gone to Puerto Rico in 1898 to treat wounded American soldiers in the Spanish American War and after its end he remained in Puerto Rico to fight the disease that affected mostly the agricultural laborers or jíbaros.)

There is some controversy information about where he studied medicine, either Barcelona or Seville, but it seems sure that he went to Spain to study and that he worked for Jaime Ferran at the Laboratorio Microbiológico Municipal of Barcelona from 1895 to 1897. Before establishing as a private physician in his country, in Bayamón, in 1899, he did postgraduate studies in the USA, in Philadelphia. That very year, Ashford had discovered that the anemia that many rural Puerto Ricans suffered was not due to malnutrition but to a parasite—Necator americanus, a hookworm—that entered the body mainly through the feet. In 1904, following Ashford’s suggestion, the Government created the first Puerto Rico Anemia Commission to study the disease and provide treatment and prevention. Gutiérrez Igaravídez, who had also studied anemia, joined the Commission as one of its directors,

along with Ashford and Walter W. King. A second Commission followed in 1905-1906 with a main station in Aibonito and substations in each district of the island. In 1908, the Commission was converted into the Anemia Dispensary Service, with forty-two dispensaries in opeartion in the island, and Gutiérrez Igarávidez was appointed its director. This project made it possible to reduce by 90% the number of death caused by uncinariasis. As Puerto Ricans suffered also from tropical diseases other than anemia, the Anemia Dispensary Service was reorganized and resulted into the creation of the Transmittable and Tropical Disease Service. However, in the country there was not any institution devoted to researching tropical diseases, and in 1912 the Institute of Tropical Medicine was founded. Research on diseases such as malaria, yellow fever, tuberculosis, syphilis, typhoid fever, and bronchopneumonia was carried there.

Gutiérrez Igaravídez, Ashford and José Jenaro Soler were the members of the first provisional Board of the Institute. Gutiérrez Igaravidez was commissioned to visit several European schools of tropical medicine in Paris, London and Liverpool. In 1913, the Institute was renamed as Institute of Tropical Medicine and Hygiene, and was the precursor of the School of Medicine of the University of Puerto Rico. In 1918, the Government provided funds to build new labo­ ra­tories and the various members of the Institute specialized in the study of different diseases. Gutiérrez Igaravídez continued the study of uncinariasis, on which he produced several works including Uncinariasis in Porto Rico: A Medical and Economic Problem, which went beyond the medical aspects of the disease. He died in Puerto Rico in 1935.

Front cover and back cover design by MBerlanga & RGuerrero

A2

RESEARCH REVIEW International Microbiology (2015) 18:1-12 doi:10.2436/20.1501.01.229. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Phenothiazines as a solution for multidrug resistant tuberculosis: From the origin to present Jette E. Kristiansen,1¶ Sujata G. Dastidar,2¶ Shauroseni Palchoudhuri,2 Debalina Sinha Roy,2 Sukhen Das,3 Oliver Hendricks,4 Jørn B. Christensen5* Memphys Centre for Biomembrane Physics, Department of Physics and Chemistry, University of Southern Denmark, Odense, Denmark. 2Department of Microbiology, Herbicure Healthcare Bio-Herbal Research Foundation, Saraldighi (E), Boral, Kolkata, India. 3Department of Physics, Jadavpur University, Kolkata, India.

4 King Christian X Hospital for Rheumatic Diseases, University of Southern Denmark, Grasten, Denmark.

5 Department of Chemistry, University of Copenhagen, Copenhagen, Denmark 1

Received 16 January 2015 · Accepted 6 March 2015 Summary. Historically, multiplicity of actions in synthetic compounds is a rule rather than exception. The science of non-antibiotics evolved in this background. From the antimalarial and antitrypanosomial dye methylene blue, chemically similar compounds, the phenothiazines, were developed. The phenothiazines were first recognised for their antipsychotic properties, but soon after their antimicrobial functions came to be known and then such compounds were designated as non-antibiotics. The emergence of highly drug-resistant bacteria had initiated an urgent need to search for novel affordable compounds. Several phenothiazines awakened the interest among scientists to determine their antimycobacterial activity.

Chlorpromazine, trifluoperazine, methdilazine and thioridazine were found to have distinct antitubercular action. Thioridazine took the lead as researchers repeatedly claimed its potentiality. Although thioridazine is known for its central nervous system and cardiotoxic side-effects, extensive and repeated in vitro and in vivo studies by several research groups revealed that a very small dose of thioridazine is required to kill tubercle bacilli inside macrophages in the lungs, where the bacteria try to remain and multiply silently. Such a small dose is devoid of its adverse side-effects. Recent studies have shown that the (–) thioridazine is a more active antimicrobial agent and devoid of the toxic side effects normally encountered. This review describes the possibilities of bringing down thioridazine and its (–) form to be combined with other antitubercular drugs to treat infections by drug-resistant strains of Mycobacterium tuberculosis and try to eradicate this deadly disease. [Int Microbiol 2015; 18(1):1-12] Keywords: Mycobacterium tuberculosis · phenotiazines · thioridazine · tuberculosis

Introduction The origin of phenothiazine dates back to 1883, when German chemist August Heinrich Bernthsen was carrying out analyses to determine the chemical constituents of two dyes, Lauth’s Corresponding author: J.B. Christensen Department of Chemistry, University of Copenhagen Copenhagen, Denmark E-mail: jbc@chem.ku.dk

*

¶

These authors contributed equally to the work.

violet and methylene blue. The dye methylene blue, whose chemical nucleus is phenothiazine, had been produced even earlier by another German, Heinrich Caro [11]. The dye industry originated in 1856 with William Perkin’s preparation of the first synthetic dye aniline purple. The increasing demand of organic chemicals such as dyes promoted the develop­ment of synthetic organic chemistry. It was soon realised that knowing structure-activity relations of organic compounds was essential for the synthesis and manufacture of new dyes. Bernthsen synthesised the phenothiazine named as methylene blue on this background [11].

2

Int. Microbiol. Vol. 18, 2015

One of the first major applications of phenothiazines beyond their use in dye industry was made in 1934 by entomologists in the US Department of Agriculture. They prepared dozens of synthetic organic sulphur compounds, among which, the most toxic compound for mosquito larvae was a phenothiazine, thiodiphenylamine [12]. Another important veterinary use of a synthetic phenothiazine was its application as an anthelmintic due to its effectiveness against swine ascaris [31]. The third action of phenothiazine to be known was its antimalarial activity reported in the extensive studies by Paul Guttmann and Paul Ehrlich [29] on methylene blue, which basically is a phenothiazine. In this connection, note that, although Ehrlich was not involved directly with the diagnosis or therapy of tuberculosis, he was instrumental in the successful staining of tubercle bacilli.

Robert Koch started his investigation staining and identifying the infective organism of tuberculosis from the grey tubercles in the lungs of animals that had died from tuberculosis. On March 24, 1882, Koch reported, in a lecture in Berlin, his success in staining and cultivating tubercle bacilli from infected lung tissues [34]. Ehrlich, who was overwhelmed with the discovery, obtained a pure culture from Koch and experimented with his various stains. He used a shorter staining time and applied nitric acid and alcohol to decolourise the surrounding tissues. By accident, he learned the benefit of heating the slide where the staining was carried out. In May 1882, Ehrlich published a detailed description of his staining technique [23]. Ehrlich’s improvement of the staining technique was soon acknowledged by Koch [50]. However, the first demonstration of a bacterium from a human infection had been made in 1875. Later Ziehl and Nielsen improved Ehrlich’s method and stained with carbol fuchsin so that the bacilli became red and very prominently visible [59]. The work

by Guttman and Ehrlich was confirmed much later by Fourneau et al. [25], who synthesised many compounds by modifying the main molecular structure of the phenothiazine methylene blue.

Syntheses and characterization of phenothiazines Halpern and Ducrot [30] synthesized many phenothiazine­ amine derivatives among which fenethazine was found to have a long lasting antihistaminic property. Charpentier et al.

[14] described the synthesis of promethazine, which surpassed the action of fenethazine in its antihistaminic action; this was later used to treat motion sickness.

KRISTIANSEN ET AL.

Subsequently, many more phenothiazines were synthesised to determine other effects. Bovet et al. [11] declared that their new phenothiazine compound diethazine had effects on sympathetic and parasympathetic nerve functions in dogs and rabbits. Macht and Hoffmaster [39] were the first to report on the use of conditioned reflex tests in animals to identify an effect of synthetic antihistamines on the central nervous system. Winter and Flataker [58] described the central antihistaminic effect of phenothiazines in animals with the help of the method of Macht and Mora [40]. They believed that the antihistaminic action of phenothiazines achieved their depressant effect by central action. The effect of phenothiazines could be nullified by caffeine to a large extent. Hence, they thought that cerebral cortex was the probable site of phenothiazine’s central action. Their observations awoke the interest to synthesise many more phenothiazine compounds.

In 1949, Henri Laborit, a French navy neurosurgeon, reported on the use of antihistamines to combat circulatory shock during and after surgery. Laborit started using the phenothiazine compound promethazine as potentiator for general anaesthesia and also to facilitate artificial hibernation in surgery. Based on Laborit’s work, the French chemical company Rhône-Poulenc decided to synthesize many more such chemicals, which led to the synthesis of more than 4000 different phenothiazines. In 1951, Laborit received samples including chlorpromazine labelled as 4560RP from Rhône-Poulenc for his clinical and pharmacological studies to evaluate them as potentiators of anaesthesia in his surgical practice. Although he found these compounds useful for artificial hibernation, Laborit emphasised on the undeniable complexity of chlorpromazine. Historically, the major impetus to move chlorpromazine into psychiatry did not come directly from RhônePoulenc. In fact, the studies carried out by Laborit and his co-workers were more responsible for the first trials of chlorpromazine by French psychiatrists, including Sigwald and Bouthier [51], and Delay and Deniker [18], each team working independently in 1952. Thus, within a year chlorpromazine became world’s best choice for psychiatry.

Although the credit of having been the first miracle drug for psychiatry is given to chlorpromazine, in 1899 the Italian physician Pietro Bodoni [9] had demonstrated that methylene blue had sedative effect in a variety of psychotic conditions in his patients. Bodoni further suggested that methylene blue could be given to psychiatric patients on a regular basis [9].

However, when chlorpromazine came on the market, only a very few papers referred to the fact that chlorpromazine was a successor of methylene blue and also that all synthesised phenothiazines and related compounds were antimicrobial and

THE JOURNEY OF PHENOTHIAZINES

neuroleptic, the so-called “narcobiotics.” Rhône-Poulenc was busy synthesizing various phenothiazines and distributing them to be tested in order to determine which of their compounds should be the best neuroleptic for patients. Several researchers found that one particular compound labelled as 3277 RP had antitubercular properties in vitro [37]. Such an antitubercular activity of chlorpromazine was obvious among different clinicians engaged in the treatment of psychoses and severe neuroses [27]. However, this unique function of chlorpromazine was not considered for further development to call it and label it as an antimycobacterial drug due to the serious side-effects produced by its prolonged administration. Meanwhile isoniazid had come on the market in 1952 and its success for treating patients suffering from Mycobacterium tu­ berculosis, coupled with information about other effective drugs, including streptomycin and rifampicin, lessened the chances of chlorpromazine being considered as an antimycobacterial agent [16]. Nevertheless, studies from different parts of the world reported on the antimycobacterial properties of several phenothiazines, including chlorpromazine.

Reports of antimycobacterial pro­perties in phenothiazines Phenothiazines as a class of easily available compounds have been repeatedly reported to have a moderate to powerful effect against many clinical strains of Gram-positive and Gramnegative bacteria both in vitro and in vivo [17]. Soon after the results of astounding success in treating neuroleptic patients with chlorpromazine became known, reports on its action against tuberculosis started appearing in scientific journals.

One such important work was by Popper and Lorian [46] in 1959, where the in vitro minimum inhibitory concentration (MIC) of chlorpromazine against M. tuberculosis was found to be as low as 25 mg/ml (Table 1). Subsequently, in 1961 Bourdon [10] again proved that chlorpromazine had anti-tubercular action in vitro. Those studies, however, did not create interests in the development of chlorpromazine as a possible antitubercular agent because of the terrible side-effects observed in patients receiving chlorpromazine in routine therapy.

While studying the action of a few phenothiazines on different types of bacteria, Molnar et al. [42] described that the growth of M. tuberculosis, M. bovis and M. butyricum was inhibited by chlorpromazine practically at an identical concentration. The MICs for M. tuberculosis were 10 mg/ml for chlorpromazine and levomepromazine, 20 mg/ml for diethazine and promethazine, whilst chlorpromazine sulfoxyde was

Int. Microbiol. Vol. 18, 2015

3

ineffective even at 100 mg/ml. Both chlorpromazine and promethazine exerted a measurable bactericidal activity on M. tuberculosis at 50 mg/ml; total destruction of the organism and partial loss of acid fastness of the cells could be observed at 300 mg/ml of chlorpromazine. In 1986, Kristiansen and Vergmann [35] made an intensive research on the antibacterial effect of various phenothiazine and thioxanthene derivatives on different species of mycobacteria in vitro (Table 1). They observed that the MICs of levomepromazine-maleate and chlorpromazine against mycobacterial strains were 25 mg/ml and 12.5 mg/ml, respectively. However, all stereo-isomers of clopenthixol were demonstrated to be twice as potent as chlorpromazine, but equally potent as chlorprothixen. Again, following the same procedure, the stereo-isomeric compounds of flupenthixol were shown to be more efficient than those of clopenthixol and chlorprothixene.

All these observations suggested that the stereo-isomeric analogue of thioxanthene derivatives should have significant antibacterial activity against the slow-growing mycobacteria.

This in vitro study also revealed that these compounds were effective with other resistant mycobacterial strains, including M. avium and M. intracellulare, within the same concentration range as mentioned earlier. Subsequently, chlorpromazine was tested by Crowle et al. [16] for its ability to inhibit the replication of M. tuberculosis and M. avium in cultured normal human macrophages, as determined by counts of viable bacteria 0, 4, and 7 days after bacterial infection of the macrophages. Chlorpromazine was able to inhibit the intracellular bacteria at concentrations ranging from 0.23 mg/ml to 3.6 mg/ml, and was more effective intracellularly than extracellularly (Table 1).

According to Ratnakar and Murthy [47], trifluoperazine, a known calmodulin antagonist, completely inhibits the growth of mycobacteria. In a synthetic medium containing 0.2% Tween 80, the MIC of this drug ranged from 5 to 8 mg/ml for the human pathogenic strain M. tuberculosis H37RV, and M. tuberculosis was resistant to isoniazid. When added to a growing culture of M. tuberculosis H37RV on the 10th day (mid exponential phase), trifluoperazine at the level of 50 mg/ml further arrested growth of this organism. Those authors [47] also observed that trifluoperazine, at a concentration of 50 µg/ml, when added to the cells along with the labeled precursors, inhibited the incorporation of 14C acetate into lipids (63%) and uptake of 14C glycine (74%) and 3H thymidine (52%) by whole cells of M. tuberculosis H37RV, after 6 h of exposure.

However, after 48 h, the inhibition was 87%, 97% and 74% respectively, in comparison with the labelled compounds as mentioned above. They further reported that when the drug

4

Int. Microbiol. Vol. 18, 2015

KRISTIANSEN ET AL.

Table 1. Distribution pattern of minimum inhibitory concentration (MIC) of different phenothiazines on clinical isolates of Mycobacterium tuberculosis Phenothiazine Chlorpromazine

MIC (µg/ml) 25

Method used ADTa

Reference [46]

Chlorpromazine

12

ADT

[10]

Chlorpromazine

10

ADT

[42]

Levomepromazine

10

Promethazine

20

Diethazine

20

Chlorpromazine sulfoxide

>100 ADT

[35]

Macb

[16]

Chlorpromazine

12.5

Levomepromazine

25

Trans (E)-flupenthixol

6.25

Cis (Z)-flupenthixol

6.25

Trans (E)-clopenthixol

12.5

Cis (Z)-clopenthixol

6.25

Trans (E)-chlorprothixen

6.25

Cis (Z)-chlorprothixen

12.5

Chlorpromazine

0.9

Trifluoperazine

5–8

ADT

[47]

Methildiazine

5–12.5

ADT

[13]

Thioridazine

8-32

Bactecc

[4]

Chlorpromazine

4–32

Bactecc

[4]

Trifluoperazine

8–32

ADT

[26]

Chlorpromazine*

10

Bactecc

[44]

Chlorpromazine**

20–30

Thioridazine*

15

Thioridazine**

20–30

Thioridazine

4

ADT

[54]

Trifluoperazine

2.5–7.5

ADT

[3]

ADT

[52]

SILA 421

2–16

Thioridazine

2–16

Thioridazine S -enantiomer

4–16

Thioridazine R -enantiomer

4–16

Chlorpromazine

1–16

Promazine

16 to >32

ADT, Agar Diffusion Test.

Mac, macrophage. c Bactec, generation of 14CO2 in Bactec 460 System, 12B vials. *MIC observed against sensitive strains of M. tuberculosis. **MIC observed against multidrug resistant strains of M. tuberculosis. a b

was added to cells taking up and metabolizing the labelled precursors at a later point 3 h for 14C acetate and 3H thymidine and 12 h for 14C glycine, trifluoperazine inhibited the uptake of all the precursors up to 24 h. It was suggested that this phe-

nothiazine would have multiple sites of action and acted probably by affecting the synthesis of lipids, proteins and DNA. Dastidar, Chakrabarty and their collaborators had worked on the detection and determination of antimicrobial action of

THE JOURNEY OF PHENOTHIAZINES

various categories of pharmacological agents since 1976 [17].

During their search they were able to detect moderate to powerful activity against grampositive and gramnegative bacteria in several antihistaminic phenothiazines. One such potent compound was methdilazine. In an intensive study, Chakrabarty et al. [13] employed 14 different reference strains of the genus Mycobacterium and screened them for their in vitro action against methdilazine along with two known antitubercular agents, streptomycin and rifampicin.

The regular medium for determining in vitro activity was Kirchner’s liquid medium, while the results were subsequently confirmed with the help of Lowenstein–Jensen medium. The MIC of methdilazine varied from 5 mg/ml to 12.5 mg/ml with respect to six test strains of M. tuberculosis; but most of the other strains of Mycobacterium were inhibited at 12.5–15.0 mg/ml of the drug, only except M. gordonae, whose MIC was as low as 5.0 mg/ml. Both streptomycin and rifampicin were highly effective on these strains, with MIC values ranging from 1.0 mg/ml to 2.0 mg/ml in respect of all the tested mycobacterial strains.

In 1996, Amaral and Kristiansen [4] observed that both chlorpromazine and thioridazine, another antipsychotic phenothiazine, could inhibit the respiration of clinical isolates of multiple drug-resistant M. tuberculosis. All these isolates were resistant to isoniazid and at least to another or even three of the other antitubercular drugs listed as first-line drugs, including streptomycin, rifampicin, ethambutol, and pyrazinamide. The authors emphasised that thioridazine, being much less toxic compared to chlorpromazine, might have a greater potential for its usage in freshly diagnosed patients of tuberculosis before determining their antibiotic sensitivity profile (Table 1).

Gadre and co-workers [26], while determining the in vitro action of trifluoperazine on clinical isolates of drug sensitive and resistant M. tuberculosis, observed that the MIC of trifluo­perazine ranged from 4 mg/ml to 8 mg/ml among most of the test strains. However, the levels of MIC of trifluoperazine were between 4 mg/ml to16 mg/ml in strains that were resistant to one or two common antitubercular agents. Gadre et al. [26] further suggested that trifluoperazine, being a known calmodulin antagonist, might inhibit the growth of tubercle bacilli, which have been shown to contain calmodulin-like protein inside their cells (Table 1).

Ordway and her collaborators [44] became interested in determining the intracellular killing capacity of thioridazine versus chlorpromazine against sensitive as well as drug-resistant M. tuberculosis with the help of BACTEC 460-TB

Int. Microbiol. Vol. 18, 2015

5

method (Table 1). They observed that both chlorpromazine and thioridazine killed intracellular antibiotic sensitive and resistant M. tuberculosis at concentrations, in the medium, well below those present in the plasma of patients treated with these agents for psychosis. Such concentrations in vitro were not toxic to macrophages, nor did they have any effect on the in vitro cellular immune processes. The authors further mentioned that as the phenothiazines are known to be concentrated by macrophages that phagocytose and have in situ activity against mycobacteria, these agents could be considered for use as adjuvants for the management of freshly diagnosed tuberculosis in patients from populations with a high prevalence of multidrug-resistant tuberculosis.

Furthermore, chlorpromazine, thioridazine, and promethazine were shown to enhance the activity of rifampicin and streptomycin when used in combinations at concentrations that are minimally effective when employed separately against clinical strains of M. tuberculosis resistant to two or more antibiotics (poly-drug-resistant M. tuberculosis). The phenothiazines had no effect on the activity of isoniazid against test strains of poly-drug-resistant M. tu­ber­culosis.

Since the toxic side effects due to systemic administration of thioridazine in patients were much lower than those recorded after chronic administration of chlorpromazine, Ordway et al. [44] postulated that thioridazine could be considered as a suitable anti-tubercular drug and could be given to recently diagnosed patients suffering from pulmonary tuberculosis.

Van Ingen et al. [54] carried out a detailed in vitro experiment to find out the effect of thioridazine on 25 strains of various species of Mycobacterium including 8 strains of M. tuber­ culosis. Susceptibility to thioridazine was tested at concentrations ranging from 1 to 128 µg/ml in Middlebrook 7H10 medium, where sufficient growth of rapid growers was noted after 4 days and that of slow growers was between 8 and 11 days.

The MIC of thioridazine with respect to all the 8 strains of M. tuberculosis (including extremely-drug-resistant, multidrug-resistant and susceptible mycobacteria) was 4 mg/ml (Table 1), whereas the MIC of thioridazine ranged from 16 mg/ml to 32 mg/ml among most of the other mycobacteria.

The level of activity of thioridazine against M. abs­cessus was maximum (64 mg/ml). This particular organism is known to cause severe disease in humans, but whether the concentration effect of thioridazine is achieved within macro­phages for infection due to such a bacterium is yet to be ascertained [54]. However, with respect to M. kansasii type 1 and M. xenopi the MIC of thioridazine was 2 µg/ml and 8 µg/ml respectively.

6

Int. Microbiol. Vol. 18, 2015

Properties of selected phenothiazines A tentative epidemiological cut-off of thioridazine was obtained by Angeby et al. [8] in 51 strains of sensitive M. tuber­ culosis and 67 strains of multidrug-resistant/extremely-drug resistant M. tuberculosis with the help of MIC determination using Middlebrook 7HI0 medium. Finally, a cut off value of 16 mg/l was proposed by the authors [8]. Although such a concentration was clinically not achievable in serum, thioridazine was found to become concentrated intracellularly and an amount of only 0.1 mg/l was able to kill M. tuberculo­ sis cells residing inside cells [8]. The MIC value >16 mg/l was found in 6% of multidrug and extremely-drug-resistant strains for which the authors stated that resistance mechanism against thioridazine had already been present in the drug-resistant clinical strains of M. tuberculosis.

The calmodulin antagonist trifluoperazine was studied in detail by Advani et al. [3] for its activity towards M. tubercu­ losis, since such compounds are known to have multiple sites of action, including lipid synthesis, DNA processing, protein synthesis and respiration. Sensitive wild type, as well as multidrug-resistant strains of M. tuberculosis, were treated with trifluoperazine under different growth conditions of stress, including low pH, starvation, presence of nitric oxide and inactivated THP-1 infectious model. Perturbation in growth kinetics of tubercle bacilli at different concentrations of trifluoperazine was checked to determine the MIC of trifluoperazine per active as also dominant bacterial cells; they observed that trifluoperazine was able to significantly reduce the actively replicating as well as non-replicating cells of M. tuberculosis, thereby producing inhibition in growth of multidrug-resistant tuberculosis bacilli (Table 1). Since trifluoperazine showed enhanced action against intracellular bacilli, the authors postulated that this phenothiazine could also get accumulated in macrophages. Furthermore, the concentration required to produce such a phenomenon was non-toxic to macro­phages [3].

Organo-silicon compounds are known efflux pump inhibitors and have anti-tubercular activity. One such compound, SILA 421, had the same pathway as the other efflux pump inhibitors of M. tuberculosis. Furthermore SILA 421 was found to modify the mdr-1 efflux pumps of M. tuberculosis and could enhance the killing action of M. tuberculosis by macrophage. Simons et al. carried out a comparative study to find out the efficacy of SILA 421 versus several known antitubercular phenothiazines using 21 clinical isolates of sensitive and resistant M. tuberculosis strains. SILA 421 was found

KRISTIANSEN ET AL.

to be equally as in vitro as the other well known efflux pump inhibitor thioridazine [52] (Table 1).

Methdilazine was assessed in a thoroughly designed in vivo experiment using Swiss albino mice. In that test, > 9 × 109 viable cells of M. tuberculosis H37Rv 102 were given as the challenge to two groups of mice, one of which received methdilazine daily. After six weeks, all the animals in both groups were sacrificed and bacterial load was determined individually from liver, spleen, intestine and lung of each animal. The results showed that the protective capacity of methdilazine was statistically significant. Further studies on the action of methdilazine were carried out by Dutta et al. in 2009 [20].

They administered methdilazine singly and in combination with either streptomycin or isoniazid to separate groups of mice infected with M. tuberculosis H37Rv 102. Maintaining proper control groups throughout the entire period of investigation, those authors found that, although the combination of methdilazine and streptomycin did not result in a significant synergism, methdilazine was highly significantly synergistic with isoniazid [20].

Apart from methdilazine, the only other phenothiazine that has been studied in detail in animal systems is thioridazine. In 2007, Martins and her colleagues [41] tried to evaluate the effectiveness of thioridazine at different dose levels in BALB/c mice that had been infected intraperitoneally with a high dose of M. tuberculosis ATCC H37Rv, thirty days prior to initiation of the treatment. A group of mice receiving no drug remained as the control. The dosages of thioridazine, calculated on the basis of human application, ranged from 0.05 to 0.5 mg/day. The group of animals that received 0.5 mg of thioridazine every day revealed a reduction of the colony forming units counts (CFU) in the lungs within 30 days. The therapeutic schedule was continued and after 300 days the number of mycobacteria in the lungs was found to be distinctly low. With the help of the same mouse model, van Sooligen et al. [55] determined the action of thioridazine in separate groups of animals infected with susceptible and multidrug-resistant M. tuberculosis. After a two-month period of oral administration of 32 and 70 mg/kg of thioridazine to two separate groups of animals, the CFU in the lungs were reduced significantly. Moreover, when thioridazine was added to a regimen containing rifampicin, isoniazid and pyrazinamide for infection with susceptible bacilli, a statistically significant synergistic effect (P < 0.01) was achieved. In a thoroughly designed pharmacokinetic study Dutta et al. [54] tried to determine the tuberculocidal activity of thioridazine in guinea pigs. The animals were aerosol infected with M. tuber­

THE JOURNEY OF PHENOTHIAZINES

culosis and single drug treatment with thioridazine was initiated 4 weeks later. The human equivalent dose of thioridazine was determined to be 5 mg/kg, which was given for 5 days per week. The initial bacterial burden in the lungs reduced after treatment with thioridazine; however, such a reduction was less than what was observed in animals treated with isoniazid.

The tolerance limit of thioridazine in guinea pigs was found to be 40 mg/kg [21].

For assessing the activity of thioridazine singly and in combination with standard anti-tubercular drugs, single dose and steady state pharmacokinetic study [22] were performed in BALB/c mice to establish human equivalent doses of thioridazine. In order to determine the bactericidal activity of thioridazine against M. tuberculosis, three separate experiments were carried out including a dose ranging study of thioridazine monotherapy and effectiveness of human equivalent doses of thioridazine with or without isoniazid/or rifampicin.

Therapeutic efficacy was assessed by the deviation in mycobacterial load in the lungs of test animals. The human-equiva­ lent dose of thioridazine was found to be 25 mg/kg, which was very well tolerated by test mice [22]. Furthermore thioridazine was found to accumulate at higher concentrations in lung tissue compared to their amount in the serum. The authors also observed that thioridazine was not only synergistic with isoniazid, but prevented emergence of isoniazid resistant mutants in lung tissues of mice [22].

Multidrug resistance in Mycobacterium tuberculosis Tuberculosis is a highly potent communicable disease worldwide. The causative organism, M. tuberculosis, is airborne and is frequently transmitted from person to person particularly among people of lower income group suffering from malnutrition and immunodeficiency. According to the World Health Organisation (WHO), in 2012 an estimated 8.6 million new cases was reported and 1.3 million died from the disease; of those, 320,000 were HIV positive. The distribution pattern of cases worldwide showed 29% in South-East Asia, 27% in Africa, 19% in Western Pacific regions. Note that India and China accounted for 26% and 12% of total cases respectively.

The treatment regime recommended by the WHO is known as Directly Observed Therapy, Short-course (DOTS), in which age-old drugs such as isoniazid, rifampicin, ethambutol, and pyrazinamide are administered simultaneously for the first two months after presumptive diagnosis. Such a schedule is

Int. Microbiol. Vol. 18, 2015

7

followed by a prolonged therapy with isoniazid and rifampicin for subsequent 4â&#x20AC;&#x201C;7 months depending on the severity of the infection. But this does not guarantee a complete cure from the disease state. Moreover, more virulent forms, designated as multidrug-resistant and extremely drug-resistant tuberculosis and their frequent association with HIV has further provoked the crisis of the infection.

Mycobacterium tuberculosis resides within the pulmonary cavities where the supply of oxygen, pH and nutrition are quantitatively very low. The bacterium causes the host to have an active immune response during the infection, which results in the release of highly reactive oxygen and nitrogen species that may at first sight seem toxic to the mycobacterium bacilli.

But the organism has developed resistance mechanisms to counteract that crisis. This resistance accounts for the success of M. tuberculosis as an intracellular pathogen. Most antibiotics that are used for the treatment of tuberculosis are only active against growing mycobacteria, but not against the dormant pathogens. The correlation between antibiotic activity and bacterial growth state in streptomycin-dependent M. tu­ berculosis has been reported. The main reasons for such a resistance are the change in bacterial metabolic pathway or a difference in physiological state that is described as phenotypic resistance. Due to this phenotypic resistance, the dormant bacilli effectively escape the immune system to the current line drugs, although genetic resistance in some organisms contributes to the resistance to tuberculosis chemotherapy.

The term multidrug-resistant is applied to those isolates of M. tuberculosis that are resistant to isoniazid and rifampicin, while extremely-drug-resistant isolates are often resistant to isoniazid, rifampicin, streptomycin, any flouroquinolone and any of the injectible anti-tuberculosis drugs like amikacin/kanamycin/capreomycin. The involvement of multidrug-resistant or extremely-drug-resistant M. tuberculosis strain in an infection is often due to misuse or overuse of the scheduled drugs or the failure of prolonged therapy that is required to be continued for months for the complete cure of the infection.

Multidrug-resistant/extremely-drug-resistant strains may also arise if the treatment schedule is allowed for more than 12 months as in the case of a new severe infection of tuberculosis. Since many drugs are prescribed and given simultaneously for such cases, sufficient pressure is applied to the causative agent to select multidrug-resistant mutants. Such a situation could not occur by a single mode of antibacterial action, as different drugs have different target sites. A single-point mutation for resistance to any type of agents may influence the occurrence of resistance to the same class of compounds.

8

Int. Microbiol. Vol. 18, 2015

However, anti-tubercular drugs are very varied as far as their mode of action is concerned. Therefore, structurally similar drugs may have the same mutated target. Thus, there is an urgent need to use less targeted therapeutics in terms of microbial anatomy. Multi-site target attack is rather less common in actual practice. But, usage of such drugs may result in damaging the internal microbiota of individuals, particularly elderly patients who might easily be infected by opportunistic microorganisms [59]. Thus there is a crucial necessity to explore novel therapeutics with improved activities such as enhanced function against multidrug-resistant and extremelydrug-resistant tuberculosis, rapid mycobactericidal mechanism, and ability to fast penetration of host cells followed by bactericidal effect in the intracellular environment.

According to DiMasi et al. [19], the cost of a new drug discovery may exceed 800 million USD for a novel chemical entity and hence there is insufficient economic incentive for a pharmaceutical industry to develop novel drugs against infectious diseases endemic in developing and underdeveloped countries. Under such circumstances, the advancement of existing chemical compounds with well-documented antimicrobial potentiality should be explored in order to obtain affordable drugs to the great multitude of patients worldwide suffering from virulent multidrug-resistant infections including tuberculosis.

During the past sixty years phenothiazines have turned out to be a very important class of compounds owing to their remarkable biological and pharmacological properties. Several studies revealed that they were the first antipsychotics and are still the best choice for psychic patients. They are not only antibacterial or antifungal, but have many other properties, such as antiviral, anticancer, antimalarial, antifilarial, trypanocidal, antihistaminic, analgesic, anti-inflammatory activities. Various studies have further revealed that some phenothiazines are capable of inducing drug resistances in bacteria [43,45]. Phenothiazines are capable of exerting activities on biological systems via the interaction of multicyclic ring system like π to π interaction or interaction in DNA, or via the lipophilic character leading to penetration through biological membranes to reach the site of action [45]. The detailed account of numerous studies presented in this review indicates that several phenothiazines act both in vitro and in vivo against sensitive as well as multidrug-resistant (MDR) tubercle bacilli. Such active phenothiazines include chlorpromazine, methdilazine, trifluoperazine and thioridazine. Many researchers have tested chlorpromazine for its potent antibacterial and antimalarial properties. This compound, however,

KRISTIANSEN ET AL.

could not be considered for therapy of bacterial infections due to its frequent serious side-effects [5]. But in the early 1980s, when multidrug-resistant tuberculosis surfaced in several developed countries, the search for new anti-tuberculosis compounds became a social necessity. At this time there was a resurgence of the anti-tubercular activity of chlorpromazine.

This activity was ensured at such high concentrations that were far beyond those clinically achievable in serum, although it was known that phenothiazines had a great affinity for lung tissue which is <100 mg/g wet tissue [28]. Kristiansen and Vergmann [35] had reported that all of their test mycobacteria were inhibited at <25 mg/ml of chlorpromazine. Hetero­ genous populations of both actively living and latent tuberculosis lesions are known to be present in individuals infected with tuberculosis.

Several studies [3,13] indicate the action of both trifluoperazine and methdilazine that could kill both the above populations which implied that the target pathways are common to both the multiplying and dormant forms of the organism.

Multidrug therapy replaced monotherapy in tuberculosis infections since the early 1960s as this was considered to be the main cause for the development of drug resistance. Various approaches should be used in testing the antimicrobial susceptibility against M. tuberculosis so that further experiments with drug combinations can be made in vivo. The synergistic action of phenothiazines with a number of anti-tubercular agents has already been reported [20,22]. Both trifluoperazine and methdilazine tend to cause accumulation and retention of antimycobacterial drugs in macrophages. Thus there should be approaches for developing more effective and less harmful derivatives of such drugs that can be used as an alternative to trigger the function of the compound.

Although chlorpromazine was found to kill M. tuberculosis in newly phagocytosed human macrophages [4,44] , due to the side-effects on the central nervous system, the use of the compound as an anti-tubercular drug remained restricted. Chlorpromazine was largely replaced by the neuroleptic drug thioridazine, which is less toxic and is much more effective in in vitro systems against all forms of antibiotic resistant strains of M. tuberculosis and also it has fewer side effects than chlorpromazine [4]. Thus the interest in thioridazine began; this compound promotes the killing of multidrug-resistant M. tuber­ culosis by macrophages, at a much lower concentration used in the therapy of psychosis, in the experimental medium [44].

In the murine model, thioridazine has been found to be effective against both antibiotic susceptible and multidrugresistant tuberculosis infection [41,55]. It has been effective

THE JOURNEY OF PHENOTHIAZINES

also against ten out of twelve cases of extremely-drug-resistant tuberculosis patients [6]. Therefore, the therapy of extremely-drug-resistant tuberculosis may be conducted on a compassionate basis with thioridazine plus standard therapeutic regimen.

Chlorpromazine has been found to affect the bacterial efflux pump; therefore, it reduces the resistance to antibiotics in bacterial strains [56]. The result was found to be similar in the case of phenothiazines. They could inhibit Nor A efflux pump of Staphylococcus aureus and the QAC efflux pump of plasmid-carrying multidrug-resistant Staphylococcus aureus [33].

The Acr AB efflux pump of E.coli was found to be inhibited by both phenothiazines, chlorpromazine and thioridazine [57]. These phenothiazines could also inhibit the main efflux pump of M. smegmatis and the efflux pump of M. avium [48].

Since it affects the survival genes of M. tuberculosis, the effect of thioridazine may presumably be universal. Phenothiazines act by inhibiting the activity of calcium-dependent ATPase, which leads to the acidification of phagolysosome and its subsequent activation of the hydrolases, resulting in the inhibition of replication of the bacterium [7]. Thioridazine inhibits the EmrE-encoded efflux pump in M. tuberculosis [49]. Further, thioridazine treated cell cause the expression of Rv3065 gene, which encodes the multidrug transport integral membrane protein EmrE, and that of another putative efflux pump gene Rv1634. When Rv3065 homologue is deleted in M. smegmatis, it has been seen that there is an increased susceptibility to a number of drugs including fluoroquinolones, ethidium bromide, and acriflavin [49].

During the macrophage infection by M. tuberculosis, the action of several efflux pumps and their regulators induce the tolerance of thioridazine inside the mycobacterial cell [2]. The accumulation of mutations in isoniazid targets is caused by the overexpression of the efflux pumps and the treatment with thioridazine, chlorpromazine and verapamil can reduce the resistance of M. tuberculosis to isoniazid [38]. Efflux pump inhibitors enhance the killing of intracellular M. tuberculosis by nonkilling macrophages. Thioridazine is able to inhibit the expression of efflux pumps in M. tuberculosis in such a manner that the anti-mycobacterial drugs fail to arrive at their intended targets in the cell. Thioridazine can also inhibit the activity of existing efflux pumps that are responsible for the multidrug-resistant phenotype of M. tuberculosis. Thus, for preventing the emergence of multidrug-resistance tuberculosis thioridazine should be used for developing new therapeutic strategies [38].

Kristiansen and her co-workers probed on various properties of phenothiazines including inhibiting potassium efflux,

Int. Microbiol. Vol. 18, 2015

9

antimicrobial potentialities and reversal of drug resistances among pathogenic bacteria [36]. They reported that antimicrobial action of phenothiazines and central nervous system activity of the same compounds are not related [36]. In their studies they further observed that phenothiazines containing an exocyclic double bond, termed as thioxanthenes, had a more potent antimicrobial activity [36]. However, such thioxanthenes acted more intensely on the central nervous system than the actual phenothiazines [24,36]. Therefore, there is a possibility to separate central nervous system and antimicrobial functions by synthesising isomeric forms of the samecompound with an additional halogen moiety, which is known to potentiate antimicrobial action [36].

Classical phenothiazine compounds having different stereoisomeric forms (+/–) have distinct differences in producing antimicrobial action [32,36,53]. In 2013, Christensen et al.

[15] made a detailed comparative analysis on the in vitro and in vivo efficacies of the (+) and (–) enantiomers of thioridazine along with its racemic form. Their study yielded a significant antibacterial action in all the forms of thioridazine, indicating of further the levorotatory (–) form to be superior in terms of both its in vitro and in vivo efficacies [15]. Since the racemic form of thioridazine containing both (+) and (–) is known to possess very potent anti-mycobacterial action [36] the (–) thioridazine with no central nervous system activity should lead to a breakthrough in the treatment of tuberculosis, particularly synergistic action of thioridazine with isoniazid [22] should also be further investigated with this form in vitro and in vivo. Thus all the recent studies presented here have established the addition of thioridazine to the existing combined regimen for the treatment of multidrug-resistant and extremely-drug-resistant M. tuberculosis.

One hundred years ago Paul Ehrlich had been treating patients suffering from trypanosomiasis and malaria with the help of methylene blue, the first ever compound of phenothiazines. Since the late 1940s the entire class of phenothiazines has evolved not only as neuroleptics but also as antimicrobials. The potentiality of thioridazine as a highly suitable candidate for multidrug-resistant/extremely-drug-resistant tuberculosis has been repeatedly proved in Argentina, where patients dying from infection by extremely drug-resistant tuberculosis were cured with a combination of linezolid, moxifloxacin and thioridazine [1]. In this way this review offers a new horizon for successful therapy of patients suffering from multidrugresistant/extremely-drug-resistant tuberculosis with thioridazine, preferably its levorotatory form, as a drug additional to the list of conventional therapeutic regimen.

10

Int. Microbiol. Vol. 18, 2015

Conclusion Tuberculosis is one of the most serious human infections throughout the world. One of the foremost reasons for its prevalence both in developed and developing countries is the lack of and inability to afford proper healthcare for such a serious disease. Overcrowding and malnutrition of poor communities in large cities is a major cause of the high incidence of tuberculosis particularly due to the easy transmission. The accelerated speed of the spread of tuberculosis during the last few decades and the higher incidence of infection by multidrug-resistant and extremely-drug-resistant M. tuberculosis have led to an alarming situation in all the continents. Although the antimycobacterial property of chlorpromazine became known soon after its widespread use as a potent antipsychotic agent, chlorpromazine failed to receive attention of the clinicians not only due to the toxic side effects but also due to the discovery of many pharmacologically active antitubercular compounds. In course of time, attention to chlorpromazine was replaced by the cogener thioridazine which continuously received attention of researchers for its antimycobacterial potentialities. Thioridazine is not only active against tubercle bacilli but also against a series of grampositive and gramnegative bacteria. The understanding of the concept of non-antibiotics is primarily centered on recognition of antimicrobial properties in commercially available pharmacologically diverse compounds. Following this pathway chlorpromazine, trifluoperazine, methdilazine and thioridazine have evolved as potent antimycobacterial drugs. Although many phenothiazines have been retracted from the market for their cardiotoxicity, the studies presented above clearly prove that the requirement of the amount of thioridazine for treating tuberculosis in BALB/c mice is rather small due to its unique property of being concentrated inside macrophages that host tubercle bacilli. The report by Dutta et al. [22] on thioridazine on the well validated murine model used for preclinical screening of antitubercular drug claimed that the human equivalent dose of thioridazine was well tolerated by test animals and that thioridazine was found to accumulate at high concentrations in lung tissues compared to serum levels.

Moreover, a very recent study has revealed that QT prolongation effect was centered on racemate and (+) thioridazine while such an action was much less in (–) thioridazine. Thus for the treatment of the multidrug-resistant/extremely-drugresistant tuberculosis a new avenue will open up by synthesising and initiating clinical trials with (–) thioridazine to facilitate an orchestrated response against this deadly pathogen.

KRISTIANSEN ET AL.

Competing interests. The authors Oliver Hendricks, Jørn B. Christensen and Jette E. Kristiansen own shares in NOA-SIC. The company holds Patent US 8,623,864, EP 168 940 5A1 & WO 2005 046694 A1 concerning “Thioridazine and Derivatives Thereof for Reversing Anti-Microbial DrugResistance”.

References 1. Abbate E, Vescovo M, Natiello M, Cufre M, Garcia A, Gonzalez Montaner P, Ambroggi M, Ritacco V, van Soolingen D (2012) Successful alternative treatment of extensively drug resistant tuberculosis in Argentina with a combination of linezolid, moxifloxacin, and thioridazine. J Anti­ microb Chemother 67:473-477 2. Adams KN, Takaki K, Connolly LE, Weidenhoft H, Winglee K, Humbert O, Edelstein PH, Cosma CL, Ramakrishnan L (2011) Drug tolerance in replicating mycobacteria mediated by a macrophage-induced efflux mechanism. Cell 145:39-53 3. Advani MJ, Siddique I, Sharma P, Reddy H (2012) Activity of trifluope­ razine against replicating, non-replicating and drug resistant M. tuber­ culosis. PLoS ONE 7:1-9 4. Amaral L , Kristiansen JE, Abebe LS, Millett W (1996) Inhibition of the respiration of multidrug resistant clinical isolates of Mycobacterium tuber­culosis by thioridazine: potential use for initial therapy of freshly diagnosed tuberculosis. J Antimicrob Chemother 38:1049-1053 5. Amaral L,Viveiros M, Kristiansen JE (2001) Phenothiazines: potential alternatives for the management of antibiotic resistant infections of tuber­culosis and malaria in developing countries. Trop Med Int Health 6:1016-1022 6. Amaral L, Boeree MJ, Gillespie SH, Udwadia ZF, van Soolingen D (2010) Thioridazine cures extensively drug resistant tuberculosis (XDRTB) and the need for global trial is now! Int J Antimicrob Agents 35:524526 7. Amaral L, Martins M, Viveiros M (2007) Enhanced killing of intracellular multidrug-resistant Mycobacterium tuberculosis by compounds that affect the activity of efflux pumps. J Antimicrob Chemother 59:12371246 8. Angeby K, Juréen P, Chryssanthou E, Schön T (2012) Tentative susceptibility testing breakpoint for the neuroleptic drug thioridazine, a treatment option for multi- and extensively drug resistant tuberculosis. Int J Mycobacteriol 1:177-179 9. Bodoni P (1899) Dell’azione sedative del blue di metilene in varie forme di psicosi. Clinica Medica Italiana 24:217-222 10. Bourdon JL (1961) Contribution to the study of the antibiotic properties of chlorpromazine or 4560 RP. Ann Inst Pasteur 101:876-386 11. Bovet D, Fournel J, Charpentier P (1947) Pharmacodynamic properties of diethylaminoethylphenothiazine hydrochloride Therapie (Paris) 2:115-119 12. Campbell FL, Sullivan WN, Smith LE, Haller HL (1934) Insecticidal tests of synthetic organic compounds—chiefly tests of sulphur compounds against culicine mosquito larvae. J Econ Entomol 27:1176-1186 13. Chakrabarty AN, Bhattacharya CP, Dastidar SG (1993) Antimycobacterial activity of methdilazine(Md), an antimicrobic phenothiazine. APMIS 101:449-454 14. Charpentier P, Guillot P, Jacob R, Gaudecho J, Buisson P (1952) Recherches sur les di-methylamino-propyl-N-phenothiazine. C.R. Acad Sc 235: 59-60 15. Christensen JB, Hendricks O, Chaki S, Mukherjee S, Das A, Pal TK, Dastidar SG Kristiansen JE (2013) A comparative analysis of in vitro and in vivo efficacies of the enantiomers of thioridazine and its racemate.

PLoS ONE 8:1-6

THE JOURNEY OF PHENOTHIAZINES

16. Crowle JA, Douvas GS, May MH (1992) Chlorpromazine: A drug potentially useful for treating mycobacterial infections. Chemotherapy 38:410-419 17. Dastidar SG, Kristiansen JE, Molnar J, Amaral L (2013) Role of phenothiazines and structurally similar compounds of plant origin in the fight against infections by drug resistant bacteria. Antibiotics 2:58-71 18. Delay J, Deniker P (1952) Réactions biologiques observées au cours du traitement par le chlorhydrate de diméthyl-amino-propyl-N-chlorophénothiazine (4560 R.P.). C.R. Congrés Médicine Alién Neurology France 50:514-518 19. DiMasi JA, Hansen RW, Grabowski HG (2003) The price of innovation: new estimates of drug development costs. J Health Econo 22:151-185 20. Dutta NK, Mazumdar K, Dasgupta A, Dastidar SG (2009) Activity of the phenothiazine methdilazine alone or in combination with isoniazid or streptomycin against Mycobacterium tuberculosis in mice. J Medical Microbiol 58:1667-1678 21. Dutta NK, Alsultan A, Peloquin CA, Karakousis PC (2013) Preliminary pharmacokinetic study of repeated doses of rifampin and rifapentine in guinea pigs. Antimicrob Agents Chemother 57: 1535-1537 22. Dutta NK, Pinn ML, KaraKousis PC (2014) Reduce emergence of isoniazid resistance with concurrent use of thioridazine against acute murine tuberculosis. Antimicrob Agents Chemother 58:4048-4053 23. Ehrlich P (1956) Modifikation der von Koch angegebenen Methode der Färbung von Tuberkelbazillen. Dtsch. Med. Wschr. In: Himmelweit F, Marquardt M, Dale H (eds) The collected papers of Paul Ehrlich, Vol 1.

Histology, Biochemistry and Pathology, Pergamon Press, London, pp 311-313 24. Feinberg AP, Snyder SH (1975) Phenothiazine drugs: structure activity relationships explained by a confirmation that mimics dopamine. Proc Nat Acad Sc USA 72:1899-1903 25. Fourneau E,Trefouel J, Trefouel M, Bovet D, Benoit G (1931) Contribution à la chimiothérapie du paludisme essais sur les calfates. Ann Inst Pasteur 46: 514-541 26. Gadre DV, Talwar V (1999) In vitro susceptibility testing of Mycobacte­ rium tuberculosis strains to trifluoperazine. J Chemother 11:203-206 27. Geiger H, Finkelstein BA (1954) Largactil in the treatment of tuberculosis. Schweiz Med Wschr 84:1063-1064 28. Gruenke LD, Craig JC, Klein FD, Nguyen TL, Hitzemann BA, Holaday JW, Loh HH, Braff L, Fischer A, Glick ID, et al. (1985) Determination of chlorpromazine and its major metabolites by GC/MS: application to biological fluids. Biomed Mass Spectrom 12:707-713 29. Guttmann P, Ehrlich P (1891) Über die Wirkung des Methylenblau bei Malaria. Berliner Klinische Wschr 28:953-956 30. Halpern BN, Ducroit R (1946) Recherches expérimentales sur une nouvelle série chimique des corps doués de propriétés antihistaminiques puissantes; les dérivés de la thiodiphénylamine (T.D.A). C.R. �������������� Soc Biology 140:361-363 31. Harwood PD (1946) Research on phenothiazine as an anthelmintic. Science Model 62:32-42 32. Hendricks O, Christensen JB, Kristiansen JE (2004) Antibakterielle Eigenschaften der Phenothiazine. Eine Behandlungsoption für die Zukunft?

Chemother J 13:203-205 33. Kaatz GW, Moudgal W, Seo SM, Kristiansen JE (2003) Phenothiazines and thioxanthenes inhibit multidrug efflux pump activity in Staphylococ­ cus aureus. Antimicrob Agents Ch 47:719-726 34. Koch R (1882) Die Ätiologie der Tuberkulose. Berliner Klinische Wschr 19:221-230 35. Kristiansen JE, Vergmann B (1986) The antibacterial effect of selected phenothiazines and thioxanthenes on slow growing mycobacteria. Acta Path Micro IM B 94:393-398

Int. Microbiol. Vol. 18, 2015

11

36. Kristiansen JE, Hendricks O, Delvin T, Butterworth TS, Aagaard L, Christiansen JB, Flores VC, Keyzer H (2007) Reversal of resistance in microorganisms by help of non-antibiotics. J Antimicrob Chemother 59:1271-1279 37. Levaditi C, Chaigneau-Erhard H, Henry-Eveno J (1951) L’anti­ hista­ minique 3277 RP (Phenergan) agit-il curativement dans la tuberculosis experimentale de la souris? C.R. Soc Biol 145:1454-1456 38. Machado D, Couto I, Perdigao J, Rodrigues L, Portugal I, Baptista P, Veigas B, Amaral L, Viveiros M (2012) Contribution of efflux to the emergence of isoniazid and multidrug resistance in Mycobacterium tu­ berculosis. PloS ONE 7:e34538.

39. Macht DI, Hoffmaster T (1948) Influence of benadryl and pyribehzas­ sine on the neuromuscular system of rats. Federal Proceedings 7:242 40. Macht DI, Mora CF (1921) Effect of opium alkaloids on the behaviour of rats in the circular maze. J. Pharmacol Exp Therap 16:219-223 41. Martins M, Viveiros M, Kristiansen JE, Molnar J, Amaral L (2007) The curative activity of thioridazine on mice infected with Mycobacterium tuberculosis. In Vivo 21:771-776 42. Molnar J, Beladi I, Foldes I (1977) Studies on anti-tuberculotic action of some phenothiazine derivatives in vitro. Zentral Bakteriol (Orig A) 239:521-526 43. Mukherjee S, Chaki S, Barman S, Das S, Koley H, Dastidar SG (2012) Effective elimination of drug resistance genes in pathogenic Pseudomo­ nas aeruginosa by antipsychotic agent thioridazine. Current Res Bacteriol 5:36-41 44. Ordway D, Viveiros M, Leandro C, Bettencourt R, Almeida J, Martins M, Kristiansen JE, Molnar J, Amaral L (2003) Clinical concentrations of thioridazine kill intracellular multidrug-resistant Mycobacterium tuber­ culosis. Antimicrob Agents Chemother 47:917-922.

45. Pluta K, Morak-Milodawska B, Jelen M (2011) Recent progess in biological activities of synthesized phenothiazines. Europ J Med Chem 46:3179-3189 46. ������������������������������������������������������������������� Popper M, Lorian V (1959) L’action de la chlorpromazine sur les microbes et le complexe bacteriostatique in vitro. Presse Med 67:212213 47. Ratnakar P, Murthy PS (1993) Trifluoperazine inhibits the incorporation of labelled precursors into lipids, proteins and DNA of Mycobacterium tuberculosis H37Rv. FEMS Microb Lett 110:291-294 48. Rodrigues L, Sampaio D, Couto I, Machado D, Kern WV, Amaral L, Viveiros M (2009) The role of efflux pumps in macrolide resistance in Mycobacterium avium complex. Int J Antimicrob Agents 34:529533 49. Rodrigues L, Wagner D, Viveiros M, Sampaio D, Couto I, Vavra M, Kern WV, Amaral L (2008) Thioridazine and chlorpromazine inhibition of ethidium bromide efflux in Mycobacterium avium and Mycobacterium smegmatis. J Antimicrob Chemother 61:1076-1082 50. Sakula A (1979) Robert Koch: Founder of the science of bacteriology and discoverer of the tubercle bacillus: A study of his life and work. Brit J Dis Chest 73:389-394 51. Sigwald J, Bouthier D (1953) Le chlorhydrate de chloro-3(dimethyl­ amino–3-propyl)-10-phenothiazine en practique neuro-psychiatrique courante. Ann Medicine 54:150-182 52. Simons SO, Kristiansen JE, Hajos G, et al. (2013) Activity of the efflux pump inhibitor SILA 421 against drug-resistant tuberculosis. Int J Antimicrob Agents 41:488-489 53. Svendsen CN, Froimowitz M, Hrbek C, Campbell A, Kula N, Baldessarini RJ, Cohen BM, Babb S, Teicher MH, Bird ED (1988) Receptor affinity, neurochemistry and behavioral characteristics of the enantiomers of thioridazine: Evidence for different stereoselectives at D1 and D2 receptors in rat brain. Neuropharmacology 27:1117-1124.

12

Int. Microbiol. Vol. 18, 2015

54. van Ingen J, van der Laan T, Amaral L, Dekhuijzen R, Boeree MJ, van Soolingen D (2009) In vitro activity of thioridazine against mycobacteria. Internat J Antimicrob Agents 34:190-191 55. van Soolingen D, Hernandez-Pando R, Orozco H, Aguilar D, MagisEscurra C, Amaral L, van Ingen J, Boeree MJ (2010) The antipsychotic thioridazine shows promising theraupetic activity in a mouse model of multidrug-resistant tuberculosis. PLoS ONE 5:e12640 56. Viveiros M, Leandro C, Amaral L (2003) Mycobacterial efflux pumps and chemotherapeutic implications. Int J Antimicrob Agents 22:274-278

KRISTIANSEN ET AL.

57. Viveiros M, Dupont M, Rodrigues L, Couto I, Davin-Regli A, Martins M, Pages JM, Amaral L (2007) Antibiotic stress, genetic response and altered permeability of E. coli. PLoS ONE 2:365-372 58. Winter C, Flataker L (1951) The effect of antihistaminic drugs upon the performance of trained rats. J Pharmacol Exp Ther 101:56-162 59. Wainwright M, Amaral L, Kristiansen JE (2012) The evolution of antimycobacterial agents from non-antibiotics. Open J Pharmacol 2:1-11 60. Zeihl F (1882) Zur F채rbung des Tuberkelbacillus. Dtsch Med Wshr 8: 451-453

RESEARCH ARTICLE International Microbiology (2015) 18:13-24 doi:10.2436/20.1501.01.230. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Combined use of a new SNP-based assay and multilocus SSR markers to assess genetic diversity of Xylella fastidiosa subsp. pauca infecting citrus and coffee plants Miguel Montes-Borrego,1 Joao R. S. Lopes,2 Rafael M. Jiménez-Díaz,1,3 Blanca B. Landa1* Institute for Sustainable Agriculture (IAS), CSIC, and Agrifood Campus of International Excellence, Córdoba, Spain.

2 Department of Entomology and Acarology, São Paulo University, Piracicaba, São Paulo, Brazil.

3 IAS-CSIC and College of Agriculture and Forestry, University of Córdoba, Córdoba, Spain

1

Received 17 November 2014 · Accepted 2 February 2015 Summary. Two haplotypes of Xylella fastidiosa subsp. pauca (Xfp) that correlated with their host of origin were identified in a collection of 90 isolates infecting citrus and coffee plants in Brazil, based on a single-nucleotide polymorphism in the gyrB sequence. A new single-nucleotide primer extension (SNuPE) protocol was designed for rapid identification of Xfp according to the host source. The protocol proved to be robust for the prediction of the Xfp host source in blind tests using DNA from cultures of the bacterium, infected plants, and insect vectors allowed to feed on Xfp-infected citrus plants. AMOVA and STRUCTURE analyses of microsatellite data separated most Xfp populations on the basis of their host source, indicating that they were genetically distinct. The combined use of the SNaPshot protocol and three previously developed multilocus SSR markers showed that two haplotypes and distinct isolates of Xfp infect citrus and coffee in Brazil and that multiple, genetically different isolates can be present in a single orchard or infect a single tree. This combined approach will be very useful in studies of the epidemiology of Xfp-induced diseases, host specificity of bacterial genotypes, the occurrence of Xfp host jumping, vector feeding habits, etc., in economically important cultivated plants or weed host reservoirs of Xfp in Brazil and elsewhere [Int Microbiol 2015; 18(1):13-24] Keywords: Citrus variegated chlorosis · coffee leaf scorch · vector transmission· xylem-limited bacteria · haplotype characterization · host-plant association

Introduction Xylella fastidiosa is a Gram-negative, xylem-inhabiting bacterium with very slow in vitro growth. It is non-specifically transmitted by several xylem-fluid feeder insect species of sharpshooter leafhoppers (Hemiptera: Cicadellidae: Cicadel* Corresponding author: B.B. Landa Instituto de Agricultura Sostenible-CSIC Alameda del Obispo, s/n 14080 Córdoba, Spain Tel. + 34-957499279. Fax + 34-957499252 E-mail: blanca.landa@ias.csic.es

linae) and spittlebugs or froghoppers (Hemiptera: Cercopidea) [3,6,15,19,34]. Xylella fastidiosa causes enormous yield losses as the etiological agent of Pierce’s disease (PD) of grapevine, Vitis vinifera; phony peach disease in peach, Prunus persica; and citrus variegated chlorosis (CVC) in Citrus spp. It is also the cause of a number of so-called leaf scorch diseases in Prunus spp. (including almond leaf scorch in Prunus amygdalus and plum leaf scald in Prunus domestica), Acer spp., Carya illinoinensis, Coffea arabica, Hedera helix, Morus rubra, Nerium oleander, Olea europaea, Platanus occidentalis, Quercus spp., and Ulmus americana [1,6,15,19].

14

Int. Microbiol. Vol. 18, 2015

Several pathogenic variants of the bacterium have been described; these are often host specific and have been given the category of subspecies [37]: Xylella fastidiosa subsp. fastidiosa, X. fastidiosa subsp. multiplex, X. fastidiosa subsp. pauca, and X. fastidiosa subsp. sandyi. Of these, X. fastidiosa subsp. pauca (Xfp) has caused significant economic losses to Brazilian agriculture since it was first reported as the etiologic agent of CVC in 1987 [3,6], and currently poses a serious potential threat to the citrus industry worldwide. This subspecies also infects coffee, another economically important crop in Brazil, causing coffee leaf scorch (CLS), a disease that was first documented in Brazil in 1995 [4]. Xilella fastidiosa is a quarantine bacterium in the European Union (EU) [11]. A strain of X. fastidiosa was recently associated with olive quick decline syndrome (OQDS), which is devastating olive trees in the Apulian province of Lecce, Salento peninsula. OQDS is a destructive disorder that developed suddenly, a few years ago, in the olive groves of a restricted area close to the city of Gallipoli, in southeast Italy. From there, the disease expanded to a wider area, currently estimated at ~10,000 ha [11,24,36]. This OQDS infestation is the first detection and establishment of this quarantine bacterium in Europe. Preliminary phylogenetic analysis using the gene encoding the β-subunit polypeptide of the DNA gyrase (gyrB) indicates that isolates from olive group are close to the branch comprising X. fastidiosa isolates that belong to the subspecies pauca [5]. Furthermore, multilocus sequence typing indicates that isolates of X. fastidiosa from olive in Apulia represent a novel strain within the subspecies pauca [23].

Xfp isolates from citrus and coffee are generally reciprocally host specific [2,25], although artificial inoculation assays have shown that isolates from citrus can sometimes infect coffee plants [21,31]. Additionally, molecular studies using a limited number of Xfp isolates have shown that they are genetically distinct [1,28,33,38]. The main objective of this study was to develop a simple molecular protocol to Xfp strains from coffee and citrus that could be used in epidemiological studies. Following the detection of a single nucleotide polymorphism in the gyrB gene that differentiates between citrus and coffee strains of Xfp, we developed a simple, reliable, and fast single nucleotide primer extension (SNuPE), or SNaPshot, protocol that can be used to differentiate between these strains (haplotypes) in samples of total DNA extracted from bacteria, infected plants, and insect vectors. Additionally, we analyzed the genetic structure of a large collection of Xfp strains from infected citrus and coffee plants grown in different regions in Brazil by using multilocus simple sequence repeat markers (SSR) (microsatellite).

MONTES-BORREGO ET AL.

Materials and methods Xylella fastidiosa isolates. Xylella fastidiosa (Xfp) is a difficult-togrow (i.e., fastidious, exigent) Gram-negative xylem-limited gammaproteobacterium in the family Xanthomonadaceae. It is rod-shaped with distinctive rippled cell walls, non-flagellate, does not form spores and measures 0.1–0.5 × 1–5 μm. Ninety Xfp isolates from Brazil were analyzed in this study (Table 1), including 28 from citrus plants showing CVC and 62 from coffee plants showing CLS symptoms. All isolates were obtained using standard isolation and triple-cloning protocols [1,2]. Some of the isolates were obtained from different trees within the same orchard in different years. The orchards were located in 19 counties in four states of Brazil (Table 1). Of the 90 isolates, 27 were used in a previous study (Table 1) [2,28], as it was the reference strain 9a5c (first X. fastidiosa isolate sequenced).

Molecular characterization of Xylella fastidiosa isolates.

All bacterial isolates were confirmed as X. fastidiosa based on their in vitro fastidious growth and PCR assays using the primer pairs RST 33/RST 31 [27] and S-S-X.fas-0067-a-S-19/S-S-X.fas-0838-a-A-21 [35], which are universal for all X. fastidiosa subspecies, as well as primer pair CVC-1/272-2-int, which is specific for Xfp causing CVC and CLS [30] (Table 2).

An in silico analysis was first performed using the X. fastidiosa multilocus database [http://pubmlst.org/xfastidiosa/] hosted on the PubMLST-Public databases for molecular typing and microbial genome diversity. The X. fastidiosa database includes data from seven loci (leuA, petC, malF, cysG, holC, nuoL, gltT). We also used information contained in Nunney et al. [28], which included some of the isolates listed in Table 1. The seven loci as well as other sequences (16S rRNA and gyrB) harbored in the GenBank database were used to test for the presence of candidate SNPs that could serve as markers in the differentiation of citrus and coffee isolates by a simple approach.

Sequencing and phylogenetic analysis. The gyrB gene was selected as the candidate gene for SNP validation. Sequences were amplified using the primer pair FXYgyr499/FXYgyr907 as described by Rodrigues et al. [35] (Table 2), purified with a gel extraction kit (Geneclean turbo; QBIOgene, Illkirch, France), quantified as described for genomic DNA, and used for direct partial sequencing, carried out at the University of Córdoba sequencing facilities. Sequencing was done with the primer FXYgyr499 with a terminator cycle sequencing ready reaction kit (BigDye; Perkin-Elmer Applied Biosystems, Madrid, Spain) according to the manufacturer’s instructions and using a DNA multicapillary sequencer (model 3100 genetic analyzer; Applied Biosystems). The sequences were deposited in the Genbank database under accession numbers DQ223435–DQ223506. The gyrB sequences from this study were edited and aligned with all GenBank published gyrB sequences of X. fastidiosa from different hosts and geographic origins.

Phylogenetic analysis was performed using Bionumerics 6.6 software (Applied Maths, Sint-Martens-Latem, Belgium).

SNuPE protocol development and validation. Based on the identification of a single nucleotide polymorphism in the gyrB sequence, a SNuPE assay was developed that differentiates between Xfp genotypes according to the host source, i.e., citrus (G genotype) and coffee (A genotype). The SNuPE protocol was designed with reference to the single-base extension (SBE) protocol of the ABI SNaPshot multiplex kit (Applied Biosystems). SNaPshot is a commercially available kit for the multiplex detection of SNPs that relies on the extension of a primer annealed immediately adjacent to the SNP of interest.

Detection is possible via the use of fluorescently labeled dideoxynucleotides, each of which emits a different wavelength such that each base is identified by a specific color. The fluorescently labeled extension products can be visualized by electrophoresis using a capillary automated sequencer [12].

A NEW SNP-BASED ASSAY FOR XYLELLA FASTIDIOSA SUBSP. PAUCA

Int. Microbiol. Vol. 18, 2015

15

Table 1. Xylella fastidiosa subsp. pauca isolates from Brazil used in the study: host source, isolation information, code number, and haplotype as defined by the SNuPE protocol developed in this study Host plant

Statea

County

Treeb

Prefix or strain

Isolate numberc

Haploype

MG

Ervália

A B C D ----

COFCOFCOFCOF-

E1, E2, E15, E17, E18 E3, E19 E4, E5, E8, E11, E22 E1N, E2N, E3N E10N, E13N, E15N

A A A A A

2003 2003 2003 2005 2005

MG

São Gotardo

E -------------------

COFCOFCOFCOFCOFCOFCOF-

E10*, E13*, E14*, E21* E22N, E23N, E25N, E26N, E24N*, E27N* J68*, J69*, J70, J71, J72, J73, J74, J75, J76, J78*, J77, J79, J80, J81, J82, J83

G A G G A G A

2003 2005 2005 2006 2006 2006 2006

MG

Lavras

-------

COFCOF-

J48*, J49* J50, J51

G A

2005 2005

MG

Varginha

----

COF-

J33

A

2002

SP

Cravinhos

----

COF-

J4

A

2002

SP

Garça

----

COF-

J29

A

2002

SP

Matao

A

1999

SP

Muritinga do Sul

----

COF-

J32

A

2002

SP

Neves Paulista

----

COF-

J17

A

2002

DF

Planaltina

----

COF-

J56, J57, J58, J59, J60, J61, J62, J63

A

2005

SP

Bebedouro

----

CIT-

J1, J2, J5, J6, J7, J8

G

2003

G

1997

G

2005

Year of isolationd

Coffea arabica

6756

Citrus sinensis 6570 SP

Gaviao Peixoto

----

CIT-

J45, J46, J47

SP

Macaubal

G

1993

SP

São Carlos

----

9a5c CIT-

J11, J12, J15, J16, J18, J19, J20

G

2003

SP

Taquaritinga

----

CIT-

J41, J42, J43

G

2005

SP

Ubirajara

----

CIT-

J44

G

2005

BA

Itapirucu

----

CIT-

J65

G

2005

MG

Comendador Gomes

----

CIT-

J35 J52, J53

G

2002 2005

MG

Frutal

----

CIT-

J54, J55

G

2005

BA = Bahia, MG = Minas Gerais, SP = Sao Paulo, DF = Distrito Federal. bSame letters identified single trees from which different isolates were obtained. ---- = Indicate that each strain was isolated from a different tree.cUnderlined isolates were used in a previous study [2]. (*) Indicates the coffee isolates that showed the G haplotype. a

Two different primers were designed in the reverse and two in the forward direction, so that they exactly adjoined to stop just 5′ of the SNP found in the gyrB sequence. The criteria for primer design were the avoidance of 3′ self-priming as well as 3′ dimerization and a size of 20–30 nucleotides.

The gyrB PCR products were purified by incubating 15 µl of each one at 37ºC for 1 h in a PCR tube containing 10 µl������������������������ �������������������������� of ExoSAP-IT (U.S. Biochemical, Cleveland, OH). An additional incubation of 15 min at 75ºC resulted in deactivation of the enzyme. Primer extension minisequencing reactions were performed according to the protocol of the SNaPshot multi-

plex kit in a total volume of 10 µl: 3 µl of the PCR purified product was mixed with 2.5 µl of SNaPshot multiplex ready reaction mix, 3.5 µl������� ��������� of ultrapure water, and 1 µl of the SNP-gyrB primer diluted to 0.6 pmol/µl. This 10-µl mixture was placed in a thermal cycler under the following cycle conditions: 25 cycles of 96ºC for 10 s, 50ºC for 5 s, and 60ºC for 30 s.

Samples were analysed with a DNA 3130 genetic analyzer (Applied Biosystems) at the University of Córdoba sequencing facilities according to the SNaPshot reaction mix protocol. The results were read using the GeneScan software (Applied Biosystems).

16

Int. Microbiol. Vol. 18, 2015

MONTES-BORREGO ET AL.

Table 2. List of PCR primers used to amplify the different Xylella fastidiosa loci Primer name

Sequence (5′→3′)

272-2-int

GCCGCTTCGGAGAGCATTCCT

[30]

CVC-1

AGATGAAAACAATCATGCAAA

[30]

S-S-X.fas-0067-a-S-19

CGGCAGCACATTGGTAGTA

[35]

S-S-X.fas-0838-a-A-21

CGATACTGAGTGCCAATTTGC

[35]

FXYgyr499

CAGTTAGGGGTGTCAGCG

[35]

FXYgyr907

CTCAATGTAATTACCCAAGGT

[35]

RST31

GCGTTAATTTTCGAAGTGATTCGATTGC

[27]

RST33

CACCATTCGTATCCCGGTG

[27]

CSSR-6 Fw

CGCACTGTCATCCATTTAATC

[22]

CSSR-6 Rv

GCTGCTTCATCTAGACGTG

[22]

CSSR-20 Fw

GGTATCGCCTTTGGTTCTGG

[22]

CSSR-20 Rv

GACAACCGACATCCTCATGG

[22]

OSSR-17 Fw

AGTACAGCGAACAGGCATTG

[22]

OSSR-17 Rv

AGCAACCAGGACGGGAAC

[22]

The reproducibility of the primer selected for the SNuPE protocol was validated by means of two approaches: (1) the SNaPshot protocol was first applied to a collection of eight Xfp isolates belonging either to the A or to the G haplotype, as determined by sequencing of the gyrB before primer design and the SNaPshot protocol; (2) a blind test was performed to determine the utility of the technique for genotyping Xfp isolates or determining the host of origin. For that purpose we used a collection of DNA from 20 isolates (J33, J35, J63, J65, and J68–J83; Table 1) obtained by J. S. Lopes and provided to B. B. Landa, but with information regarding the host of origin withheld (Table 1).

Implementation of the SNuPE protocol in epidemiological studies. The utility of the SNuPE protocol for epidemiological studies was validated using leaf and insect vector samples. Leaves were sampled from four CVC-affected plants from a citrus orchard at São Carlos (São Paulo, SP) and ten CLS-affected plants from a coffee orchard at São Gotardo (Minas Gerais, MG) that also were used to isolate bacterial strains.

The leaves were placed inside labeled, dark plastic bags under moist conditions, transported on ice to the laboratory (2–4 h until delivery time), and kept in a cold room (10°C) for 1 day before DNA extraction. Total DNA was extracted from sampled leaves using cetyltrimethylammonium bromide buffer as described previously [35]. Additionally, DNA was extracted from adults of the sharpshooter vector Bucephalogonia xanthophis Berg (Hemiptera: Cicadellidae: Cicadellinae) that were fed for 48 h on citrus plants infected with Xfp belonging to the G genotype. The extraction protocol was that described by Ciapina et al. [8]. The gyrB gene from leaf and insect DNA was amplified using one or two rounds of the same PCR protocol [35], in which 1 µl of the first PCR run was used as template in the second run. The aim of this approach was to counteract the low concentration of the bacterium in some of the sampled leaves and insects. When a positive PCR amplification was achieved, the SNuPE protocol was performed as described above.

Reference

Microsatellite data analyses. Multilocus SSR PCR assays using three primer pairs (CSSR-6, CSSR-20, and OSSR-17) were used to assess the genetic diversity within Xfp strains from citrus and coffee. Primers were selected among the 34 SSR primer sets designed by Lin et al. [22] that maximize the number of haplotypes that could be differentiated according to the host of origin. PCR was performed as described by Lin et al. [22], with the forward primer labeled with HEX fluorescent dye. PCR amplification was first tested by agarose gel electrophoresis, after which allele sizes were determined with a LIZ500 size standard and a DNA 3130XL genetic analyzer (Applied Biosystems) at the University of Córdoba sequencing facilities and by using Genemapper version 3.7 (Applied Biosystems). All SSR fragments were scored as one putative locus with two alleles; one allele indicated the presence of a fragment and the other its absence. The datasets were compiled as a matrix of isolates and SSR fragments. Only reproducible bands were scored. Approximately half of the assayed isolates were tested twice, with identical results. The resulting matrix of isolates and SSR fragments was analyzed in two ways.

First, a distance dendrogram was calculated with the simple matching coefficient and the Ward algorithm using Bionumerics 6.6 software. Second, an approach similar to that recently described by Almeida et al. [2] was used, in which the hypothesis tested was whether citrus and coffee Xfp isolates were genetically clustered in different groups based on the host plant.

GenAlEx Version 6.5 [29] was used to calculate the average number of alleles (Na), average number of effective alleles (Ne), and haploid genetic diversity (H) at each locus and across all loci for the different groups of strains according to their host of origin. Analyses of molecular variance (AMOVAs), as implemented in GenAlEx 6.5, were used to determine the covariance between groups by grouping populations by host plant. The significance of the AMOVA and of the results of the population pairwise Fst comparisons was tested with 1000 permutations. The number of genetic clusters in the microsatellite dataset was determined using the software package STRUCTURE 2.2 [32]. The posterior likelihood of the samples

A NEW SNP-BASED ASSAY FOR XYLELLA FASTIDIOSA SUBSP. PAUCA

being divided into between one and six genetic clusters (k) was tested by resampling the dataset ten times (burn-in, 10,000 steps; run, 100,000 steps) as described before [2].

Results and Discussion Selection of candidate locus and phylogenetic analysis. In silico analysis of the seven loci (leuA, petC, malF, cysG, holC, nuoL, gltT) included in the public X. fastidiosa MLST database, the study from Nunney et al. [28], and the 16S rRNA sequences indicated that they were less appropriate than the gyrB gene in the differentiation between citrus and coffee strains of Xfp based on the potential presence of a single SNP (data not shown). Indeed, some of the loci (leuA, petC, nuoL, gltT) showed no variation between a subset of the coffee or citrus isolates from Brazil while others contained variable positions within Brazilian coffee (cysG, holC, malF, leuA, petC, nuoL, gltT, from 8 to 13 nucleotides) or citrus (e.g., petC, holC, nuoL, from 1 to 10 nucleotides) isolates. However, the analysis of 90 gyrB sequences from all citrus and coffee isolates of Xfp used in the present study showed that the sequences of the isolates were identical, with the exception of a SNP (a purine transition A→G) at position 831 with respect to the gyrB sequence (AE003849, locus tag XF0005) of strain 9a5c.

Phylogenetic analysis of all X. fastidiosa isolates yielded a topology similar to those reported in other studies [5,28,35], with all Xfp isolates differentiated in two groups (data not shown). Within these two groups, 100% of the citrus isolates showed the same G haplotype, and most (50 out of 62) of the coffee isolates showed the A haplotype, whereas only a small proportion (12 out of 62) of the isolates from coffee shared the G haplotype of the citrus isolates. Those coffee isolates with the G haplotype were isolated from two coffee orchards in São Gotardo, MG, and Lavras, MG (Table 1).

These data indicated that two different haplotypes would infect citrus and coffee plants in Brazil, that they could co-infect a single coffee orchard, and that host jumping between haplotypes in certain regions would be possible. This result is consistent with reports that Xfp isolates causing CLS in coffee do not colonize citrus plants, whereas those pathogenic to citrus and causing CVC can infect and multiply in coffee plants [2,21,31]. Chen et al. [7] were able to differentiate X. fastidiosa subsp. multiplex and X. fastidiosa subsp. fastidiosa based on a SNP in the 16S rRNA gene and a multiplex PCR assay, which showed the simultaneous presence of the two X. fastidiosa subspecies in the same infected almond orchard. However, to the best of our knowledge, mixed infections by different genotypes

Int. Microbiol. Vol. 18, 2015

17

belonging to the same X. fastidiosa subspecies have not been previously demonstrated. The potential for mixed genotype infection ocurrence could affect current epidemiological studies and, therefore, X. fastidiosa disease management strategies.

Previous attempts to analyze the genetic relatedness of Xfp strains through sequencing of the 16S rRNA gene and 16S-23S intergenic space region [13,26] were hampered by the limited variability, which did not consistently cluster with the host [35].

Similar limited variability was found in the sequence analysis of some of the seven loci referred to above. In contrast, using gyrB, a gene that evolves much faster than rRNAs, we were able to differentiate between Xfp isolates according to the host of origin, i.e., coffee or citrus. Other studies have reported the utility of SNP molecular markers in the analysis of X. fastidiosa genetic diversity in isolates from coffee and citrus plants. However, those studies found a larger number of SNPs (10–24 and 2–12 for coffee citrus isolates, respectively), which, unlike the gyrB SNP, must be used together to distinguish between strains according to the host of origin of the isolates [38].

SNuPE protocol development and validation.

Out of the four primers designed, SNP-gyr-25 (5′-GGACTGATGCCTACCAAGAAACAAT-3′) yielded the most reproducible results in the SNuPE protocol, whereas migrations by the other three primers were not reproducible among different runs for the same isolate and the efficiency of the amplification was lower (data not shown). The results obtained with the SNuPE protocol were in complete agreement with those derived from the sequencing of the gyrB gene. The results for a given isolate were also identical when the SNuPE reactions were performed with the same PCR products in different runs, or with PCR products resulting from independent amplifications (data not shown). An example of an electropherogram of two SNuPE reactions for bacterial isolates of haplotypes A and G is shown in Fig. 1. For each genotype, a single peak was observed for each of the isolates: green (adenine) for isolates belonging to the A haplotype (e.g., isolate E2 from coffee, A haplotype), and blue (guanine) for isolates belonging to the G haplotype (e.g., isolate 6750 from citrus, G haplotype).

The SNuPE technique was originally developed to detect the Phe508 mutation in the human CFTR gene [20]. However, it has been recently applied to bacteria, in particular, to identify bacterial species and variants relevant to animal or human health as well as in the identification and differentiation of various probiotic bacteria at the species or sub-species level [9,12,16,17]. However, ours is the first application of the technique to differentiate among strains or haplotypes of plant pathogenic bacteria.

Int. Microbiol. Vol. 18, 2015

MONTES-BORREGO ET AL.

Int Microbiol

18

Fig. 1. SNaPshot analysis migration profile of Xylella fastidiosa subsp. pauca (Xfp) isolates E2 and J70 from coffee (A genotype; green peak, adenine), and isolates 6750 and J35N from citrus and J68 from coffee (G genotype; blue peak, guanine). The migrations correspond to the primer length. Note the slightly different migration for the adenine to guanine incorporation after the SNaPshot reaction.

To further demonstrate the usefulness of the newly developed SNuPE technique, the robustness of the protocol was tested in blind tests using total DNA extracted from bacterial isolates or from infected plants. All bacterial isolates were assigned correctly to their respective host of origin by the SNuPE protocol; e.g., isolate J35 (citrus, G haplotype) and isolate J70 (coffee, A haplotype) (Fig. 1). As determined from the sequence analysis of gyrB, a small proportion (13.3%) of the bacterial isolates from coffee in the orchard located in S達o Gotardo, MG, showed the G haplotype (e.g., isolate J68). In addition, amplification of the gyrB gene showed that all symptomatic coffee and citrus plants were infected by Xfp. The SNuPE protocol indicated that all citrus-infected plants showed the G haplotype, as expected (Fig. 2E), whereas the analysis of coffee plants indicated that seven were infected by bacteria of the A haplotype (Fig. 2C,D), and three were infected jointly by bacteria of the two haplotypes (Fig. 2A,B).

These plants came from the CLS-affected orchard in S達o Gotardo, MG, from which isolates of Xfp belonging to haplotypes A and G were isolated. This is the first reported demonstration that two Xfp haplotypes can co-infect trees in a single orchard, and more importantly a single coffee plant within an

orchard. This observation may explain the results reported by Almeida et al. [2] and Nunney et al. [28], in which CVC and CLS were shown to be caused by genetically distinct, frequently recombining groups of X. fastidiosa. Recombination is thought to occur widely in this bacterium, allowing both its adaptation to new host plants and speciation. In addition, horizontal gene transfer of pathogenicity factors may drive the emergence of new diseases caused by X. fastidiosa [2].

A nested-PCR assay for the gyrB gene [35] using total DNA extracted from insects previously fed for 2 days on a citrus plant infected with the G haplotype of a Xfp isolate yielded positive results. A subsequent assay using the SNuPE protocol confirmed the identity of the isolate as the G haplotype (Fig. 2F,G). No PCR amplification occurred when DNA extracted from healthy laboratory-reared insects allowed to feed on non-infected citrus plants was used as the template (data not shown). Thus, the SNuPE technique also has applications in genotyping for host association of X. fastidiosa using total DNA extracted directly from plant and insect tissues.

The need for a nested-PCR protocol for the analysis of insect DNA samples in our study was probably due to the low number of X. fastidiosa cells in the insect vector, as previously

Int. Microbiol. Vol. 18, 2015

19

Int Microbiol

A NEW SNP-BASED ASSAY FOR XYLELLA FASTIDIOSA SUBSP. PAUCA

Fig. 2. SNaPshot analysis migration profile of plant samples. DNA samples extracted from four different coffee (a, b, c, and d) or citrus (e) representative plants showing CLS or CVC symptoms, respectively, and from insect vectors allowed to feed on Xfp-infected citrus tree (f). Note the presence of both genotypes (A genotype; green peak, adenine; and G genotype; blue peak, guanine) in representative coffee plants 1 and 2.

observed in culture and by the quantification of colony-forming units [14]. Nevertheless, the combined use of the nestedPCR and the SNuPE assays in a sequence would be very useful for epidemiological studies of CVC and CLS diseases, as well as for finding new potential vectors in Xfp-susceptible crops. Also, the identification of new SNPs in gyrB genes and other loci will facilitate the implementation of the SNuPE protocol in a multiplex approach to analyze and differentiate Xfp variants, such as the one associated with OQDS in Italy, as well as the remaining X. fastidiosa subspecies. Indeed, the combination of only four SNPs was shown to be sufficient to differentiate all X. fastidiosa subspecies, including the one associated with OQDS in Italy (M. Montes-Borrego, M. Saponari, B.B. Landa, unpublished results).

Characteristics of SSR loci, genotypes and genetic diversity. SSR amplification was successful with the three primer pairs, which amplified products ranging from 224 to 367 bp, as scored with the multicapillary sequencer

and GenScan software. A double dendrogram (one for the isolates and another for the SSR loci) was generated with the three sets of SSR loci to visually summarize the variability within the collection of Xfp isolates (Fig. 3). Overall, 41 SSR products were scored with the three primer pairs and the 90 Xfp isolates analyzed (Fig. 3). All loci were polymorphic, with the number of alleles ranging from a minimum of 11 (OSSR17) to a maximum of 16 (CSSR20). Among the 56 genotypes (i.e., combined SSR profiles) that were identified, 34 were from coffee and 22 from citrus; 37 were unique (i.e., a single isolate). Although some of the loci were common to coffee and citrus isolates (i.e., 2/11, 5/14 and 7/16 for OSSR17, CSSR16, and CSSR20 respectively), no common combined SSR profiles were detected among coffee and citrus isolates.

The frequencies of some alleles varied considerably with respect to the host source and geographic origin (states of Brazil and counties within each state) of the bacterial isolates (Fig. 3). Among the products amplified by the SSR primer pairs CSSR6, CSSR20, and OSSR17, those of sizes 253, 290, and

Int. Microbiol. Vol. 18, 2015

MONTES-BORREGO ET AL.

Int Microbiol

20

A NEW SNP-BASED ASSAY FOR XYLELLA FASTIDIOSA SUBSP. PAUCA

Int. Microbiol. Vol. 18, 2015

21

Table 3. Descriptive statistics and genetic diversity of Xylella fastidiosa subsp. pauca isolates across three microsatellite loci, as determined in Brazilian populations of the bacterium collected in coffee and citrus plants showing leaf scorch and variegated chlorosis symptoms, respectively Population

Locus

Number of isolates

Number of alleles

Number of effective alleles

Haploid genetic diversity

Frequency of occurrence of alleles

Coffee

CSSR6

62

10

1.99

0.498

0.016–0.694

CSSR20

62

13

10.01

0.900

0.016–0.161

OSSR17

62

8

5.29

0.811

0.016–0.226

CSSR6

28

9.00

5.44

0.816

0.036–0.286

CSSR20

28

10.00

5.85

0.829

0.036–0.321

OSSR17

28

5.00

1.86

0.462

0.036–0.714

CSSR6

90

14.00

3.84

0.739

0.011–0.478

CSSR20

90

16.00

11.16

0.910

0.011–0.178

OSSR17

90

11.00

6.72

0.851

0.011–0.233

Citrus

Total

234 bp were the most frequent (frequencies of 0.478, 0.178, and 0.233, respectively) and the most effective in discriminating Xfp groups (Fig. 3; dendrogram for SSR loci). Locusbased haploid diversity ranged from 0.498 to 0.900 for coffee isolates and from 0.462 to 0.829 for citrus isolates. The overall haploid genetic diversity was very similar for coffee (0.736) and citrus isolates (0.702) (Table 3). Almeida et al. [2] also analyzed the genetic diversity among citrus and coffee strains from Brazil using microsatellite analysis, but they did not provide any information concerning the number of alleles, genotypes, or haploid genetic diversity of each population.

We also wondered whether Xfp isolates collected from symptomatic citrus and coffee plants in different Brazilian states could be grouped genetically based on the host plant.

The AMOVA results showed a statistically significant difference between the two population groups (Fst = ΦPT = 0.664; P = 0.001), supporting the hypothesis that the genetic structures of the groups of isolates were driven by the plants from which they were obtained. Indeed, most of the variation observed occurred among populations (66%). In contrast, Almeida et al. [2] did not find significant genetic differences associated with the host source of 46 of X. fastidiosa isolates. Most of the variation occurred among individuals within populations, although populations within groups showed statistical differences when grouped by state, indicative of genetic differ-

ences between coffee and citrus isolates from a same region.

The discrepancies between our results and those of Almeida et al. [2] might be due to differences in the sizes of the datasets (90 vs. 46 isolates, respectively) and in the SSR loci selected for our study. Thus, we used three SSR loci showing the highest number of different alleles among those described by Lin et al.

[22], whereas Almeida et al. [2] used six SRR loci from those described by Coletta-Filho et al. [9], which in general showed lower haploid genetic diversity than was the case in our study (0–0.85 for citrus isolates and 0–0.75 for coffee isolates).

Genetic relationship and structure. Two main clusters (I and II) and six subclusters (Ia, Ib, Ic, and Id, and IIa, IIb, and IIc) were identified in the cluster analysis of the 90 bacterial isolates that were defined based on a 35.5% and 60.0% similarity coefficient, respectively. This clustering of the isolates correlated well with their host of origin and bacterial haplotype, with a few exceptions (Fig. 3). For instance, all isolates in cluster I were isolated from coffee plants and had the A haplotype, with the exception of one isolate from citrus (J47) that was of the G haplotype (subcluster Ia). Within cluster II, subclusters IIb and IIc were formed only by citrus strains, whereas subcluster IIa was the most diverse and included all coffee isolates that were of the G haplotype and some citrus isolates (Fig. 3).

← Fig. 3. Dendrogram based on similarity data (simple matching coefficient) from short sequence repeats (SSR) fragment analysis with the Ward algorithm using Bionumerics 6.6 software for the 90 Xfp isolates used in the study. BA = Bahia, MG = Minas Gerais, SP = Sao Paulo, DF = Distrito Federal. Cophenetic correlation values are indicated in each node. *Cluster and subcluster groups were defined based on a cluster cutoff value of 35.5% and 60.0%, respectively.

Int. Microbiol. Vol. 18, 2015

MONTES-BORREGO ET AL.

Int Microbiol

22

Our results also showed that coffee trees in a single orchard can be infected by different isolates and that a single coffee plant can be infected by different isolates, as determined by different SSR profiles. For instance, isolates E15, E17, and E18 (subcluster Ib) and isolates E1 and E2 (subcluster Ib) were obtained from the same coffee tree from Ervália, MG, in 2003 (Fig. 3, Table 1). This was also the case for isolates E11 and E22 and for isolates E4, E5, and E8 (obtained from the same orchard as those above, but from a different tree), which grouped in subcluster Ic but differed in their SSR profiles, as well as for isolates E10, E13, E14, and E21 (sampled in São

Fig. 4. Individual assignments of 'X. fasti­ diosa subsp. pauca' isolates to four genetic clusters obtained by STRUCTURE analysis and indicated in different colors. Disagreement with respect to the assignment of isolates with the clustering obtained in Fig. 3 is indicated by an asterisk (*).

Gotardo, MG, in 2003), which varied in their SSR profiles and were grouped in subcluster IIa (Fig. 3). Note that, although coffee orchards from Ervália and São Gotardo, MG were sampled in different years (Table 1), isolates with the same SSR genotype were recovered from the different coffee plants sampled.

Note also that some of the SSR profiles were present in coffee or citrus plants in different states and counties, whereas isolates of other SSR profiles were unique to a single orchard or county (Fig. 3).

To further analyze the possible genetic structure of Xfp isolates, a second clustering approach was used to infer the

A NEW SNP-BASED ASSAY FOR XYLELLA FASTIDIOSA SUBSP. PAUCA

most likely group of origin of each isolate. An analysis using the STRUCTURE software showed that the probability was the highest with four clusters (k), which was confirmed by the larger ∆k value with four populations (data not shown). The membership of each isolate obtained from the STRUCTURE analysis, estimated as (q), or the ancestry coefficient, is shown in Fig. 4. The q values vary between 0 and 1.0; with 1.0 indicating full membership in a population. Individual isolates can be assigned to multiple clusters (with values of q summing to 1.0), indicating that they are admixed. If isolates with q ≥ 0.90 are defined as those of a single lineage, 18 of the 90 Xfp isolates can be considered as admixed lineages [18]. The STRUCTURE analysis, without prior information regarding geography or host cultivar, placed all isolates into four major clusters (Fig. 4). The clustering of the isolates was consistent with their lineage assignment, as determined by the cluster analysis (Fig. 3). The exception was a few isolates in cluster IIb, which according to STRUCTURE were grouped with all isolates in cluster IIc. Almeida et al. [2], using the STRUCTURE package, grouped their coffee and citrus X. fastidiosa dataset into three genetic clusters, one of which comprised both citrus and coffee isolates, as was the case in the present study for isolates within cluster IIa (Figs. 3 and 4).

The combined use of microsatellite- and SNuPE protocolbased assays in the present study yielded new insights into the genetic relationships among Xfp isolates collected from symptomatic citrus and coffee plants in Brazil. Our fingdings support previous results showing that CVC and CLS diseases are caused by genetically distinct, frequently recombining groups of Xfp isolates [2]. In addition, we were able to show that some of the isolates obtained from coffee plants shared the G haplotype and clustered with citrus isolates. This result is consistent with previous observations that CVC-associated isolates can multiply to some extent in coffee plants, while CLSassociated isolates do not colonize citrus plants [2,31].

In conclusion, although the use of a single molecular marker for the specific detection of X. fastidiosa strains originating from one or another host can have limitations, due to the frequent mutation and recombination events that characterize populations of this bacterium [2,28], the SNuPE protocol developed in this study is a rapid and robust assay that can be used in a first screening to differentiate among Xfp haplotypes recovered from citrus or coffee. The combined use of the SNuPE protocol with microsatellite analysis can be very useful for studies on the epidemiology of the diseases caused

by Xfp, the host specificity of bacterial genotypes, the occurrence of host jumping, insect vector feeding habits, etc., in economically important cultivated host plants and weed host

Int. Microbiol. Vol. 18, 2015

23

reservoirs of the bacterium in Brazil. Also, the combined use of microsatellite- and SNuPE protocol-based assays has applications in studies on the new variant of Xfp associated with the recent epidemic outbreak of OQDS, devastating olive trees in Italy [5,24,36].

Acknowledgements. We acknowledge financial support from the EU grant ICA4-CT-2001-10005 and an ‘Intramural Project’ to B. B. Landa from the Spanish National Research Council (CSIC), as well as CNPq for a scholarship to J. R. S. Lopes in Brazil. We thank E. Zambolin, L. Zambolin, C. M. Oliveira, and S. Lopes for collecting coffee and citrus plants with CLS and CVC symptoms.

Competing interests. None declared.

References 1. Almeida RPP, Purcell AH (2003) Biological traits of Xylella fastidiosa strains from grapes and almonds. Appl Environ Microbiol 69:7447-7452 2. Almeida RPP, Nascimento FE, Chau J, Prado SS, Tsai CW, Lopes SA, Lopes JRS (2008) Genetic structure and biology of Xylella fastidiosa strains causing disease in citrus and coffee in Brazil. Appl Environ Microbiol 74:3690-3701 3 Almeida RPP, Pereira EF, Purcell AH, Lopes JRS (2001) Multiplication and movement of a citrus strain of Xylella fastidiosa within sweet orange. Plant Dis 85:382-386 4. Beretta MJG, Harakava R, Chagas CM (1996) First report of Xylella fastidiosa in coffee. Plant Dis 80:821 5. Cariddi C, Saponari M, Boscia D, De Stradis A, Loconsole G, Nigro F, Porcelli F, Potere O, Martelli GP (2014) Isolation of a Xylella fastidiosa strain infecting olive and oleander in Apulia, Italy. J Plant Pathol 96:425-429 6. Chang CJ, Garnier M, Zreik L, Rossetti V, Bové JM (1993) Culture and serological detection of the xylem-limited bacterium causing citrus variegated chlorosis and its identification as a strain of Xylella fastidiosa.

Curr Microbiol 27:137-142 7. Chen J, Groves R, Civerolo EL, Viveros M, Freeman M, Zheng Y (2005) Two Xylella fastidiosa genotypes associated with almond leaf scorch disease on the same location in California. Phytopathology 95:708-714 8. Ciapina LP, Carareto Alves LM, Lemos EGM (2004) A nested-PCR assay for detection of Xylella fastidiosa in citrus plants and sharpshooter leafhoppers. J Appl Microbiol 96:546-551 9. Coletta-Filho HD, Takita MA, Souza AA, Aguilar-Vildoso CI, Machado MA (2001) Differentiation of strains of Xylella fastidiosa by a variable number of tandem repeat analysis. Appl Environ Microbiol 67:4091-4095 10. Dalmasso A, Civera T, Bottero MT (2009) Multiplex primer-extension assay for identification of six pathogenic vibrios. Int J Food Microbiol 129:21-25 11. European Food Safety Authority (2013) Statement of EFSA on host plants, entry and spread pathways and risk reduction options for Xylella fastidiosa Wells et al. EFSA J 2013:11 3468 12. Ferri L, Perrin E, Campana S, Tabacchioni S, Taccetti G, Cocchi P, Ravenni N, Dalmastric C, Chiarini L, Bevivino A, Manno G, Mentasti M, Fani R (2010) Application of multiplex single nucleotide primer extension (mSNuPE) to the identification of bacteria: the Burkholderia cepacia complex case. J Microbiol Methods 80:251-256

24

Int. Microbiol. Vol. 18, 2015

13. Hendson M, Purcell AH, Chen D, Smart C, Guilhabert M, Kirkpatrick B (2001) Genetic diversity of Pierce’s disease strains and other pathotypes of Xylella fastidiosa. Appl Environ Microbiol 67:895-903 14. Hill BL, Purcell AH (1995) Acquisition and retention of Xylella fastidiosa by an efficient vector, Graphocephala atropunctata. Phytopathology 85:209-212.

15. Hopkins DL (1989) Xylella fastidiosa: xylem-limited bacterial pathogen of plants. Annu Rev Phytopathol 27:271-290 16. Huang CH, Chang MT, Huang MC, Lee FL (2011a) Application of the SNaPshot minisequencing assay to species identification in the Lactobacillus casei group. Mol Cell Probes 25:153-157 17. Huang CH, Chang MT, Huang MC, Lee FL (2011b) Rapid identification of Lactobacillus plantarum group using the SNaPshot minisequencing assay. Syst Appl Microbiol 34:586-589 18. Islam Md-S, Glynn JM, Bai Y, Duan Y-P, Coletta-Filho HD, Kuruba G, Civerolo EL, Lin H (2012) Multilocus microsatellite analysis of ‘Candidatus Liberibacter asiaticus’ associated with citrus Huanglongbing worldwide. BMC Microbiol 12:39, doi:10.1186/1471-2180-12-39 19. Janse JD, Obradovic A (2010) Xylella fastidiosa: Its biology, diagnosis, control and risks. J Plant Pathol 92:S1.35-S1.48 20. Kuppuswamy MN, Hoffmann JW, Kasper CK, Spitzer SG, Groce SL, Bajaj SP (1991) Single nucleotide primer extension to detect genetic diseases: experimental application to hemophilia B (factor IX) and cystic fibrosis genes. Proc Natl Acad Sci USA 88:1143-1147 21. Li W-B, Pria WD Jr, Teixeira DC, Miranda VS, Ayres AJ, Franco CF, Costa MG, He C-X, Costa PI, Hartung JS (2001) Coffee leaf scorch caused by a strain of Xylella fastidiosa from citrus. Plant Dis 85:501-505 22. Lin H, Civerolo EL, Hu R, Barros S, Francis M, Walker MA (2005) Multilocus simple sequence repeat markers for differentiating strains and evaluating genetic diversity of Xylella fastidiosa. Appl Environ Microbiol 71:4888-4892 23. Loconsole G, Almeida R, Boscia D, Martelli GP, Saponari M (2014) Multilocus sequence typing reveals the genetic distinctiveness of the Xylella fastidiosa strain CoDiro. Proc Internat Symp European outbreak of Xylella fastidiosa in olive. Gallipoli, Locorotondo, Italy 21-24 October: p 55.

24. Loconsole G, Potere O, Boscia D, Altamura G, Djelouah K, Elbeaino T, Frasheri D, Lorusso D, Palmisano F, Pollastro P, Silletti MR, Trisciuzzi N, Valentini F, Savino V, Saponari M (2014) Detection of Xylella fastidiosa in olive trees by molecular and serological methods. J Plant Pathol 96:1-8 25. Lopes SA, Marcussi S, Torres SCZ, Souza V, Fagan C, França SC, Fernandes NG, Lopes JRS (2003) Weeds as alternative hosts of the citrus, coffee, and plum strains of Xylella fastidiosa in Brazil. Plant Dis 87:544-549

MONTES-BORREGO ET AL.

26. Metha A, Rosato YB (2001) Phylogenetic relationships of Xylella fastidiosa strains from different hosts, based on 16S rDNA and 16S-23S intergenic spacer sequences. Int J Syst Evol Microbiol 51:311-318 27. Minsavage GV, Thompson CM, Hopkins DL, Leite RMVBC, Stal RE (1994) Development of a polymerase chain reaction protocol for detection of Xylella fastidiosa in plant tissue. Phytopathology 84:456-461 28. Nunney L, Yuan X and Bromley RE (2012) Detecting genetic introgression: high levels of intersubspecific recombination found in Xylella fastidiosa in Brazil. Appl Environ Microbiol 78:4702-4714 29. Peakall R, Smouse PE (2012) GenAlEx 6.5: genetic analysis in Excel.

Population genetic software for teaching and research–An update.

Bioinformatics 28:2537-2539 30. Pooler MR, Hartung JS (1995) Specific PCR detection and identification of Xylella fastidiosa strains causing citrus variegated chlorosis. Curr Microbiol 31:134-137 31. Prado S, Lopes JRS, Demetrio C, Borgatto A, Almeida RPP (2008). Host colonization differences between citrus and coffee isolates of Xylella fastidiosa in reciprocal inoculation. Sci Agricola 65:251-258 32. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155:945-959 33. Qin X, Miranda VS, Machado MA, Lemos EGM, Hartung JS (2001) Na evaluation of the genetic diversity of Xylella fastidiosa isolated from diseased citrus and coffee in São Paulo, Brazil. Phytopathology 91:599605 34. Redak RA, Purcell AH, Lopes JRS, Blua MJ, Mizell III RF, Andersenm PC (2004) The biology of xylem fluid-feeding insect vectors of Xylella fastidiosa and their relation to disease epidemiology. Annu Rev Entomol 49:243-270 35. Rodrigues JLM, Silva-Stenico ME, Gomes JE, Lopes JRS, Tsai SM (2003) Detection and diversity assessment of Xylella fastidiosa in field-collected plant and insect samples by using 16S rRNA and gyrB sequences. Appl Environ Microbiol 69:4249-4255 36. Saponari M, Boscia D, Nigro F, Martelli GP (2013) Identification of DNA sequences related Xylella fastidiosa in oleander, almond and olive trees exhibiting leaf scorch symptoms in Apulia (Southern Italy). J Plant Pathol 95:659-668 37. Schaad NW, Postnikova E, Lacy G, Fatmi M, Chang CJ (2004) Xylella fastidiosa subspecies: X. fastidiosa subsp. piercei subsp. nov., X. fastidiosa subsp. multiplex subsp. nov., X. fastidiosa subsp. pauca subsp. nov. Syst Appl Microbiol 27:290-300 38. Wickert E, Machado MA, Lemos EG (2007) Evaluation of the genetic diversity of Xylella fastidiosa strains from citrus and coffee hosts by single-nucleotide polymorphism markers. Phytopathology 97:1543-1549

RESEARCH ARTICLE International Microbiology (2015) 18:25-31 doi:10.2436/20.1501.01.231. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Strong correlation between the antifungal effect of amphotericin B and its inhibitory action on germ-tube formation in a Candida albicans URA+ strain José P. Guirao-Abad, Pilar González-Párraga, Juan-Carlos Argüelles* Microbiology Area, School of Biology, University of Murcia, and IMIB-Arrixaca, Murcia, Spain Received 22 December 2014 · Accepted 30 March 2015

Summary. The hypothetical capacity of amphotericin B to suppress the formation of germ-tubes, which is the first step of yeast-to-hypha conversion in Candida albicans, has been investigated in the wild-type strain CEY.1 (CAI.4-URA+). Exponential cells exposed to concentrations of amphotericin B below or around the MIC90, exhibited a weak reduction in the percentage of human serum-induced germ-tube formation at 37ºC compared with a non-exposed control. However, the dimorphic transition was drastically suppressed after addition of potentially lethal doses of amphotericin B, which also caused severe cell killing. In contrast, an identical experimental approach carried out with the fungistatic compound 5-fluorocytosine had no significant effect on the level of the germ-tube formation. Together, these results strongly point to a close correlation between the fungicidal action of amphotericin B and its ability to impair morphogenetic conversion in C. albicans. [Int Microbiol 2015; 18(1):25-31] Keywords: Candida albicans · amphotericin B · 5-fluorocytosine · germ-tube · cell killing

Introduction Antifungal therapy is less developed than the antibacterial equivalent. However, the striking increase in morbidity and mortality caused by invasive mycosis, which mainly affects the human immunocompromised population, demands new effort in this pharmaceutical field [22,24,25]. Together with the development of safer, more potent and less toxic compounds, another line of work involves improving the currently Corresponding author: J.C. Argüelles Area de Microbiología, Facultad de Biología Campus de Espinardo, Universidad de Murcia 30071 Murcia, Spain Tel. + 34-868887131. Fax +34-868883963 E-mail: arguelle@um.es *

available arsenal of fungicidal antibiotics [1]. A good example of this pursuit is the polyene macrolide amphotericin B (AmB). Since its discovery in the 1950s [20] AmB has been the predominant compound utilized in clinical practice for the treatment of systemic infections caused by species of Candida, Aspergillus, Cryptococcus and many other pathogenic fungi. However, this extensive use has provoked some toxic side-effects (nephrotoxicity and hepatic damage, mainly), which have successfully been surmounted by new liposomal formulations, permitting its current extensive utilization against invasive candidiasis [13,19]. Interestingly, clinical resistance to AmB remains extremely rare despite 50 years of use as a monotherapy [28].

Many clinical trials have taken as a target the polimorphic opportunistic yeast Candida albicans, which is still the most

26

Int. Microbiol. Vol. 18, 2015

prevalent fungal pathogen in humans, where it is responsible for mucosal and disseminated infections [7,19,22,25]. Furthermore, the incidence of C. albicans infections in neonates and patients subjected to extensive surgery therapy, as well as nosocomical bloodstream candidemia, has increased noticeably in the last two decades [7,19,22,24]. Several studies on C. albi­cans suggest that AmB may exert a complex mode of action, interacting with different cellular targets. Although the specific binding to the membrane ergosterol triggers perturbations in the selective membrane permeability that eventually lead to cell killing, the formation of pores and membrane instability has been dissociated from the fungicidal effect [21].

Other evidences gathered in C. albicans and Cryptococcus neoformans [29] involve the generation of oxidative stress through the release of reactive oxygen species, would reinforce the fungicidal effect of AmB-induced cell damage [4,6,9,16].

Candida albicans is a polimorphic organism able to grow as a unicellular budding yeast or as mycelial forms (hypha and pseudohypha). The morphological transition from yeast to hypha has been considered as a contributory factor of virulence [10,15]. It has been proposed that exposure to AmB causes a dose-dependent severe reduction in the germ-tube development, which is the first step of hypha formation [8,26].

In fact, the success obtained with early therapies that used AmB to treat oral candidiasis seemed to involve a significant reduction in the percentage of germ-tube formation, which is the first step in the yeast-to-hypha conversion of this opportunistic pathogen [8]. Oral candidiasis remains among the main manifestations of mycosis caused by C. albicans, including a great proportion of HIV-infected patients [18].

However, such observations are far from conclusive [7,24,25].

We have examined this suggestion in a prototypic wild-type C. albicans URA3+ strain, by testing putative effects on dimorphism and cell viability triggered by the polyene AmB and the fungistatic antibiotic 5-fluorocytosine (5-FC), a pyrimidine analogue that interferes with DNA synthesis by reducing the available nucleotide pool [12]. Our rationale was to try to dissociate morphological inhibition from killing cell processes. According to our data, the blockage of germ-tube development seems to be a direct consequence of the previous drastic fungicidal action of AmB.

Material and methods Yeast strains and culture conditions. Because an ura3 auxotrophy might have side effects on the physiology and virulence of C. albicans, a CEY.1 (CAI4-URA3+) strain was used throughout this study. A detailed description of the constructions and procedures followed to obtain this strain is

GUIRAO-ABAD ET AL.

reported elsewhere [23]. This same strain has been used in previous research on the antifungal action of AmB [11]. Yeast cell cultures were grown at 28ºC. They were shaken in a medium consisting of 2% peptone, 1% yeast extract and 2% glucose (YPD). Strains were maintained by periodic subculturing in solid YPD. Human-serum induction of germ tube formation. For germ-tube induction, cultures were either directly supplemented with 10% human blood serum or, alternatively, samples were harvested at different stages of growth, quickly washed with water and resuspended at a density of 1–5 ×106 cells/ml in YPD prewarmed to 37ºC together with 10% human serum. Before addition, serum was sterilized by filtration (0.45 mm). The appearance of germ tubes was monitored as indicated by phase contrast light microscope with a haemocytometer. When required, clumped cells were dispersed prior to microscopic examination by mild sonication (10–15 s). At least 250 cells were counted each time and the percentage of dimorphism was represented as the ratio of germ tube-forming cells to the total number of cells [3,23].

AmB and 5-FC treatments. AmB was obtained from Sigma (80% purity) and prepared in DMSO (100%). It was then maintained at room temperature for autosterilisation, since it is not possible to filter the compound. Because the solutions are all very light-sensitive, they were manipulated in dark conditions. 5-FC was also from Sigma, and was dissolved in MiliQ water.

Cultures were grown in YPD until the exponential phase (OD600 = 0.3) and then divided into several identical aliquots, which were treated with different concentrations of AmB or 5-FC and incubated at 28ºC with shaking. The samples were harvested at the indicated times and viability was determined after appropriate dilution of the samples with sterile water by plating in triplicate on solid YPD. Between 30 and 300 colonies were counted per plate. Survival was normalized to control non-treated samples (100% viability).

MICs determination. The MIC90 for AmB and 5-FC against CEY.1 (URA3+) cells was determined according to a normalized protocol for yeasts (EUCAST with minor modifications). Briefly, a microtiter plate (96 wells) (Brand 781660, Wertheim, Germany) was filled with 100 µl of different concentrations of AmB (0.0125; 0.025; 0.05; 0.1; 0.2; 0.4; 0.8; 1.6; 3.2; 6.4 and 12.8 mg/ml) in RPMI 1640 medium plus 2% glucose buffered at pH 7.0 with MOPS 0.164 M. One hundred ml of a yeast suspension in sterile saline solution (105 cells/ml) was added to each well. The plates were incubated at 35°C for 24 h and 48 h and read at 550 nm on a microtiter reading instrument (Asys Jupiter). In accordance with the EUCAST protocol, the strain Candida parapsilosis ATCC 22019 has been included as quality control. The MIC was defined as the lowest concentration which inhibited 90% of the cell growth.

Results Effects of AmB treatments on germ-tube formation and cell killing in the C. albicans CEY.1 (URA) strain. In a set of preliminary experiments, we have determined the MIC90 values for CEY.1 (CAI.4-URA3+) cells against AmB and 5-FC. They were 0.12 and 0.25 mg/l, res­ pectively. The quality control required by the EUCAST procedure was included to ensure the validity of the calculations.

These concentrations are within the range previously reported for other C. albicans genetic backgrounds of clinical or laboratory origin [8].

Int. Microbiol. Vol. 18, 2015

Int Microbiol

GERM-TUBE FORMATION IN C. ALBICANS

Fig. 1. Effect of rising concentrations of amphotericin B (AmB) on the time-course of germ-tube formation induced by human serum (A), the percentage of survival (B) and the microscopic cell morphologies (C and D) in the wild type strain CEY.1 (CAI-4 URA+) of C. albicans. The cultures were incubated at 28ºC in YPD until early exponential phase (OD = 0.3), harvested, washed and resuspended at a cellular density of 2 × 106 cells/ml in YPD supplemented with 10% human serum and immediately transferred at 37ºC. Identical samples were supplemented at time zero with the following doses of AmB (mg/l): 0 (control, ♦), 0.05 (■); 0.1 (▲); 0.5 (●) and 1.0 (ÿ). Error bars are omitted for the sake of clarity, but the standard deviation was always lower than 12%. (A and B). Seruminduced samples were photographed after 5 h of treatment with 0.05 mg/l (C) or 0.5 mg/l (D) AmB.

27

Int. Microbiol. Vol. 18, 2015

GUIRAO-ABAD ET AL.

Int Microbiol

28

Fig. 2. Reversible action of AmB on human serum-induced germ-tube formation (A) and cell viability (B) in C. albicans CEY.1 strain. An exponential culture growing on YPD at 28ºC was divided into identical aliquots, which were quickly supplemented with 10% human serum and transferred to 37ºC. Then, they were subjected to the following treatments: addition of 0.5 mg/l AmB for 30 min (■) or 60 min (▲) followed by rapid washing out AmB and resuspension in YPD plus human serum (time zero). Alternatively, control samples were kept without AmB (♦) or with that toxic concentration of AmB throughout the assay (●). More details in Figure 1.

The addition of 10% human serum together with a temperature up-shift from 28 to 37ºC has been demonstrated as an efficient procedure for inducing filamentation in C. albicans [3,11,23]. This was also the case for the wild-type strain CEY.1 (CAI.4-URA+). Exponentially-growing yeast cells (blastoconidia) in a glucose-rich medium (YPD) are fully competent to enter the dimorphic program when they are transferred to the same medium prewarmed at 37ºC, supplemented with human serum and further incubated at 37ºC (Fig. 1A). It has been postulated that sub-lethal concentrations of AmB induce a complete suppression of the capacity to issue germ-tubes [8,26]. In the present study, the effect of several

doses of AmB on the degree of human serum-induced initial filamentation in CEY.1 growing cells was studied. As shown in Figure 1A, the addition of AmB doses below or about the MIC90 value had a negligible effect on the level of germ-tube formation (about 10% reduction at 0.1 mg/l AmB; Fig. 1A). It was necessary to increase the AmB concentration to potentially lethal levels (0.5 or 1.0 mg/l) in order to bring about a significant decay of mycelial outgrowth in comparison with a control sample maintained at 37ºC (Fig. 1A).

In equivalent cell aliquots withdrawn from the same cultures, the effect of those selected AmB doses on the degree of cell killing was analyzed. The time-course percentage of via-

Int. Microbiol. Vol. 18, 2015

29

Int Microbiol

GERM-TUBE FORMATION IN C. ALBICANS

Fig. 3. Level of dimorphic conversion (A) and cell growth (B) in CEY.1 cells induced by several concentrations of the fungistatic antibiotic 5-fluorocytosine (5-FC). A growing exponential culture was processed as indicated in the legend for Figure 2. Identical aliquots were treated with the following concentrations of 5-FC (mg/l): 0 (♦), 0.25 (●), 1.0 (■) and 2.5 (▲). A sample containing 0.25 mg/l of AmB was introduced as control of the positive antifungal effect (ÿ). Error bars are omitted for the sake of clarity, but the standard deviation was always lower than 12%.

ble CEY.1 blastoconidia underwent a minor reduction upon the addition of tolerated concentrations of AmB (ca. 20% at 0.05 mg/l and 40% at 0.1 mg/l) (Fig. 1B), whereas fungicidal doses of the polyene (4 × MIC90 and 8 × MIC90) caused a more drastic and progressive loss of viable cells (Fig. 1B). Optical examination of samples treated for 1 h with 0.05 mg/l AmB showed normal serum-induced germ-tube emergence (Fig. 1C), whereas a 10-fold increase in the AmB concentration (0.5 mg/l) completely prevented the dimorphic transition (Fig. 1D). Hence, these data strongly suggest that the suppressive action on germ-tubes emergence in C. albicans after AmB supply might likely be an off-side indirect action de-

rived from its intrinsic fungicidal power rather than a direct target. Interestingly, AmB triggers a conspicuous individual reduction of cell size in yeast cultures [11].

In order to confirm this putative relationship between the antifungal power and suppression of hypha formation exerted

by AmB, the following experiment was carried out. Growing CEY.1 blastoconidia were adjusted to a similar cell density (2x106 cells/ml) and preincubated with a toxic concentration of AmB (0.5 mg/l) for 30 or 60 min. Then, the antifungal was removed by washing the cells, which were quickly resuspended in human serum (10%) at 37ºC at an identical cell density.

The degree of germ-tube development and the percentage of

30

Int. Microbiol. Vol. 18, 2015

cell survival were determined in parallel. According to the results presented in Figure 2, pretreatment with 0.5 mg/l AmB caused a time-dependent loss of viable cells (30 or 60 min) (Fig. 2B). Afterthat, however, the surviving fraction was still able to resume active grow to a certain extent as well as to enter the dimorphic program by producing conspicuous germtubes (Fig. 2; results not shown). This dimorphic conversion markedly increased as the incubation period was prolonged (Fig. 2A). As expected, a sample subjected to the permanent presence of AmB lost its capacity to form germ-tubes, the reduction in cell viability being drastic and the subsequent level of survival extremely low (Fig. 2A and B). In turn, in a control culture treated with human serum at 37ºC, the degree of myceliation was higher than 80% after 2 h (Fig. 2A). In this set of experiments, a control sample was maintained at 37ºC without serum addition, and the level of germ-tube formation was about 10–20% (results not shown).

Analysis of the action of 5-FC. The suppressive action triggered by fungicidal exposure of AmB on the yeast-tohypha conversion capacity in C. albicans was also reinforced by studies carried out with the fungistatic compound 5-fluorocytosine (5-FC). In exponential CEY-1 cells, the addition of concentrations around the MIC90 (0.25 mg/l) had no relevant effect on the percentage of filamentation or cell viability (Fig. 3A and 3B). A moderate increase (4 × MIC90) in the dose of 5-FC only caused a partial and sustained decrease in the level of survival and had an irrelevant effect on dimorphic conversion, which was manifested by the rather similar ability to produce germ-tubes (Fig. 3B). The effect of 5-FC was not dose-dependent, since an additional rise (10 × MIC90) only caused a weak additional decrease in the level of survival and the percentage of germ-tube formation (Fig. 3). In this assay, an aliquot exposed to a lower concentration of AmB (0.25 mg/l) was introduced as a positive control of the fungicidal action. As expected, the polyene provoked a severe degree of cell killing and, simultaneously, a pronounced inhibition in the ability to issue germ tubes (Fig. 3).

Discussion The morphological conversion from yeast cells (blastoconidia) to mycelial structures (hypha and/or pseudohypha) in C. albicans has been considered a factor of virulence, although the matter is open to dispute [10]. The genetic evidence is not entirely conclusive: whereas some mutants unable to filament are avirulent [15], homozygous null mutants deficient in the

GUIRAO-ABAD ET AL.

MAP kinase HOG1 and in the transcriptional repressor factor TUP1 show a phenotype of hyperfilamentation but lack infectivity in a murine model for systemic candidiasis [2,5]. Furthermore, in other dimorphic fungi with great incidence in pathogenesis, e.g. Histoplasma capsulatum or Blastomyces dermatitidis, yeast-like cells instead of mycelia are the predominant invasive morphology in clinical mycoses [10].

These fungi are very susceptible to both AmB and azoles [13,14]. In turn, Candida glabrata, which displays high clinical incidence due to its inherent resistance towards usual antifungals, shows a typical growth as oval yeast cells [17].

In this context, a strong and inhibitory action of the polyene AmB on the ability to produce germ-tubes in C. albicans has been proposed [8,26], although the mechanism for this suggested targeted inhibition remains unknown. However, the evidence provided in this study seems to dismiss this proposal and strongly points to the fungicidal action triggered by AmB as the main cause of the reduced formation of germ tubes in C. albicans. Thus, exponential blastoconida of the wild type CEY.1 strain (URA+) did not show significant loss of cell growth upon addition of sublethal concentrations of AmB (Fig. 1B) and were able to enter the dimorphic program, issuing conspicuous germ tubes (Fig. 1A and 1C), although with a certain delay compared to the untreated samples (Fig. 1A).

Only the treatment with toxic doses of AmB impeded the formation of evident mycelial structures (Fig. 1D), probably due to the halt in cell growth (Fig. 1B).

On the other hand, these harmful effects on cell growth and hypha formation are reversible, conditioned to the presence of the antifungal in the culture broth. When a 4 × MIC90 concentration of AmB was applied for a short time (30 or 60 min) and then quickly removed from the medium, the cells suffered an initial decay in cell survival with no production of germ-tubes as long as the antifungal was present (Fig. 2). But, afterwards, they were able to resume active growth (Fig. 2B) and showed a capacity to undergo a noticeable dimorphic transition, which had remained unaffected during the transitory exposure to AmB (Fig. 2A). Similar experiments performed with the fungistatic 5-FC provided equivalent results and confirmed the validity of this proposal (Fig. 3).

The relationship between fungicidal action and hypha formation might be relevant for clinical therapeutical purposes, particularly in the light of the ability of C. albicans to develop biofilms as a main pathogenesis mechanism [27]. Laboratory studies have proven that C. albicans biofilms have high intrinsic resistance to several antifungals, including polyenes, 5-fluorocytosine (5-FC) and azoles [27], although they still remain susceptible to the application of echinocandins and

GERM-TUBE FORMATION IN C. ALBICANS

some lipid preparations of AmB, in the range of doses commonly used in clinical assays. Because fungal biofilms are heterogeneous communities of blastospores, pseudohyphal and hyphal cells embedded in an extracellular complex matrix, experiments on the susceptibility of each population to AmB should be performed to clarify this matter. In conclusion, we exclude yeast-to-hypha conversion as the main target of AmB. Rather, early AmB-induced germ-tube inhibition seems to be a side effect derived from the drastic lytic action of the polyene on cell integrity.

Acknowledgements. We thank Dr. O. Zaragoza (Instituto de Salud Carlos III, Madrid) for the critical reading of the manuscript, useful suggestions and warm support. J.P.G-A received a partial fellowship from VitalGaia, S.L. The experimental work was supported by grant PI12/01797 (Ministerio de Economía y Competitividad, ISCIII, Spain). We are also indebted to the financial contract provided by Cespa, Servicios Auxiliares de Murcia, S.A. Competing interests. None declared.

References 1. Akins RA (2005) An update on antifungal targets and mechanisms of resistance in Candida albicans. Med Mycol 43:285-318 2. Alonso-Monge R, Navarro-García F, Molero G, Diez-Orejas R, Gustin M, Pla J, Sánchez M, Nombela C (1999) Role of the mitogen-activated protein kinase Hog1p in morphogenesis and virulence of Candida albicans. J Bacteriol 181:3058-3068 3. Argüelles JC, Rodriguez T, Alvarez-Peral FJ (1999) Trehalose ������������������ hydrolysis is not required for human serum-induced dimorphic transition in Candida albicans: evidence from a tps1/tps1 mutant deficient in trehalose synthesis. Res Microbiol 150:521-529 4. Belenky P, Camacho D, Collins JJ (2013) Fungicidal drugs induce a common oxidative-damage cellular death pathway. Cell Reports 3:350-358 5. Braun BR, Johnson AD (1997) Control of filament formation in Candida albicans by the transcriptional repressor TUP1. Science 277:105-109 6. Cao Y, Zhu Z, Chen X, Yao X, Zhao L, Wang H, Yan L, Wu H, Chai, Y, Jiang Y (2013) Effect of Amphotericin B on the metabolic profiles of Candida albicans. J Proteomics 12:2921-2932 7. Cisterna R, Ezpeleta G, Telleria O, Guinea J, Regueiro B, Rodriguez J, Esperalba J (2010). A nationwide sentinel surveillance of bloodstream Candida infections in 40 tertiary care hospitals in Spain. J Clin Microbiol 48:4200-4206 8. Ellepola AN, Samaranayake LB (1998) The effect of limited exposure to antifungal agents on the germ tube formation of oral Candida albicans. J Oral Pathol Med 27:213-219 9. González-Párraga P, Sánchez-Fresneda R, Zaragoza O, Argüelles JC (2011) Amphotericin B induces trehalose synthesis and simultaneously activates an antioxidant enzymatic response in Candida albicans. Biochim Biophys Acta 1810:777-783 10. Gow NAR, Brown AJP, Odds FC (2002) Fungal morphogenesis and host invasion. Curr Opin Microbiol 5:366-371 11. Guirao-Abad JP, Sánchez-Fresneda R, Valentin E, Martínez-Esparza M, Argüelles JC (2013) Analysis of Validamycin A as a potential antifungal compound against Candida albicans. Int Microbiol 16:217-225

Int. Microbiol. Vol. 18, 2015

31

12. Hope WW, Tabernero L, Denning DW, Anderson MJ (2004). Molecular mechanisms of primary resistance to flucytosine in Candida albicans.

Antimicrob Agents Chemother 48:4377-4386 13. ����������������������������������������������������������������������� Klepser M (2011). The value of amphotericin B in the treatment of invasive fungal infections. J Crit Care 26:225.e1-10 14. Li RK, Ciblak MA, Nordoff N, Pasarell L, Warnock DW, McGinnis MR (2000). In vitro activities of voriconazole, itraconazole, and amphotericin B against Blastomyces dermatitidis, Coccidioides immitis, and Histoplasma capsulatum. Antimicrob Agents Chemother 44:1734-1736 15. Lo HJ, Köhler JR, DiDomenico B, Loebenberg D, Cacciapuoti A, Fink GR (1997) Nonfilamentous Candida albicans mutants are avirulent. Cell 90:939-949 16. Mesa-Arango A, Trevijano-Contador N, Román E, Sánchez-Fresneda R, Casas C, Herrero E, Argüelles JC, Plá J, Cuenca-Estrella M, Zaragoza O (2014) ��������������������������������������������������������������� The production of oxygen reactive species is a universal mechanism of amphotericin B against pathogenic yeasts and contributes to the fungicidal effect of this drug. Antimicrob Agents Chemother 58:66276638 17. Miramón P, Kasper L, Hube B (2013) Thriving within the host: Candida spp. interactions with phagocytic cells. Med Microbiol Immunol 202:183-195 18. Nishimura S, Shimbo T, Asayama N, Akiyama J, Ohmagari N, Yazaki H, Oka S, Uemura N (2013) Factors associated with esophageal candidiasis and its endoscopic severity in the era of antiretroviral therapy. Plos One 8:e58217 19. Ortega M, Marco F, Soriano A, Almela M, Martínez JA, López J, Pitart C, Mensa J. (2011). Candida species bloodstream infection: epidemiology and outcome in a single institution from 1991 to 2008. J Hosp Infect 77:157-161 20. Oura M, Sternberg ET, Wright ET (1955) A new antifungal antibiotic, amphotericin B. Antibiot Annu 3:566-573 21. Palacios DS, Anderson, TM, Burker MD (2007) A post-PKS oxidation of the amphotericin B skeleton predicted to be critical for the channel formation is not required for potent antifungal activity. J Am Chem Soc 129:13804-13805 22. Patterson TF (2005) Advances and challenges in management of invasive mycoses. Lancet 366:1013-1025 23. Pedreño Y, González-Párraga P, Martínez-Esparza M, Sentandreu R, Valentín E, Argüelles JC (2007) Disruption of the Candida albicans ATC1 gene encoding a cell-linked acid trehalase decreases hypha formation and infectivity without affecting resistance to oxidative stress. Microbiology 153:1372-1381 24. Pfaller MA, Diekema DJ (2007) Epidemiology of invasive candidiasis: a persistent public health problem. Clin Microbiol Rev 20:133-163 25. Pfaller MA, Diekema DJ (2010) Epidemiology of invasive mycoses in North America. Critical Rev Microbiol 36:1-53 26. Vale-Silva LA, Buchita V, Valentova E (2007) Effect of subinhibitory concentrations of some established and experimental antifungal compounds on the germ tube formation in Candida albicans. Folia Microbiol 52:39-43 27. Vediyappan G, Rossignol T, d’Enfert C (2010) Interaction of Candida albicans biofilms with antifungals: transcriptional response and binding of antifungals to beta-glucans. Antimicrob Agents Chemother 54:2096-2111 28. Vincent BM, Lancaster AK, Scherz-Souval R, Whitesell L, Lindquist S (2013) Fitness trade-offs restrict the evolution of resistance to Amphotericin B. Plos Biology 11:e1001692 29. Wang Y, Casadevall A (1994) Growth of Cryptococcus neoformans in presence of L-dopa decreases its susceptibility to amphotericin B. Antimicrob Agents Chemother 38:2648-2650

RESEARCH ARTICLE International Microbiology (2015) 18:33-40 doi:10.2436/20.1501.01.232. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Use of E-beam radiation to eliminate Listeria monocytogenes from surface mould cheese Raquel Velasco, Juan A. Ordóñez, M. Isabel Cambero, M. Concepción Cabeza* Department of Nutrition, Food Science and Food Technology. School of Veterinary, Complutense University of Madrid, Madrid, Spain Received 20 January 2015 · Accepted 29 March 2015

Summary. Camembert and Brie soft cheese varieties were subjected to E-beam irradiation as a sanitation treatment. The effects of treatments on microbiota and selected physicochemical properties were also studied. The absorbed doses required to meet the food safety objective (FSO) according to EU and USDA criteria for Listeria monocytogenes were 1.27 and 2.59 kGy, respectively. The bacterial load, mainly lactic acid bacteria, was reduced by the treatment but injured cells were recovered during storage at 14°C. The radiation treatment gave rise to negligible changes in the pH and water activity at doses required to achieve microbial safety. [Int Microbiol 2015; 18(1):33-40] Keywords: Listeria monocytogenes · food safety objective (FSO) · soft mould-ripened cheeses · E-beam radiation

Introduction Camembert and Brie cheeses are included in the soft cheeses types. They are manufactured by a similar technology and the main difference between them is the diameter size, i.e., 10–11 cm and 22–36 cm for Camembert and Brie, respectively [13]. The main feature of both cheeses is that, after the lactic acid fermentation, a white crust is formed due to the growth of Penicillium camemberti in the surface. Due to the mould activity, the ripening is very fast at room temperature and to decelerate this phenomenon the cheese is commonly stored under refrig-

Corresponding author: M.C. Cabeza Depto. de Nutrición, Bromatología y Tecnología de los Alimentos Facultad de Veterinaria Universidad Complutense de Madrid Av. Puerta de Hierro, s/n 28040 Madrid, Spain Tel. +34-913944091. Fax +34-913943743 E-mail: ccabezab@ucm.es *

eration. Since Listeria monocytogenes is a psychrotrophic organism, it may grow if it is present.

Listeria monocytogenes is the causative agent of a disease that may be acquired by food ingestion. Several L. monocytogenes outbreaks have been reported due to soft cheeses [11,20, 30,53]. This bacterium can reach the product mainly from environmental contamination or, more rarely, bovine mastitis [7,15,45]. It has been isolated from cheeses made from raw, low heat-treated and pasteurized milk [11,31] even more frequently than in those made from unpasteurized [20], which is probably due to a post-treatment contamination [36].

The ubiquity of L. monocytogenes in nature and its recognized presence in food-processing environments [49] explain the difficulty in producing either minimally processed foods or ones handled after processing that are free of the pathogen.

A relatively small number of cases of foodborne diseases are caused by L. monocytogenes in comparison to the most common pathogen outbreaks such as those caused by Salmonella spp. or Campylobacter jejuni [12, 21]. Despite its low incidence, the infection is considered of great impor-

34

Int. Microbiol. Vol. 18, 2015

tance due to the severity of the illness and high mortality rates (e.g., 12.70% in 2011 vs. 0.12% for Salmonella) [21].

Recently, the effectiveness of accelerated electrons (Ebeam) treatment to eliminate pathogens was demonstrated [8,10,28,39] in a variety of ready-to-eat (RTE) foods. Moreover, a higher radioresistance of L. innocua vs. L. monocytogenes has been consistently observed [8,10,28,42] permitting the former species to be used as a surrogate of L. monocytogenes.

Previously, we have used E-beam radiation [52] to minimize the incidence of late blowing of cheese but an important reduction in lactic acid bacteria (LAB) was also produced, although after damage reparation an almost normal growth was recorded.

Taking these findings into account, it was postulated that E-beam technology could be highly appropriate for sanitizing RTE cheeses. This paper describes a study to optimize Ebeam treatment of the above mentioned cheese varieties with the final goal of reducing the number of L. monocytogenes to a safe level, i.e., in compliance with the criteria of EU and USDA regulations for this bacterium.

Materials and methods Organisms. One strain of Listeria innocua (NCTC 11288) and another of L. monocytogenes (Scott A, CIP 103575, serotype 4b) were used. The strains were maintained by freezing (−40°C) in trypticase soy broth (TSB; Difco, Detroit, MI, USA) with 10% glycerol added as a cryogenic agent. Fresh cultures were prepared by removing a piece of frozen culture from vials and inoculating it into 9 ml of TSB, then incubating it at 32°C for 24 h. The culture was then centrifuged at 4ºC (3000 × g for 30 minutes) and the pellet was suspended in a beaker with 50 ml sterile saline solution, which yielded a bacterial load of approximately 108 cells/ml. In experiments, large numbers of cells were used to accurately determine the bacterium inactivation kinetics.

Sample preparation and irradiation treatment. Individually wrapped portions (30 g) of Brie and Camembert cheeses made from pasteurized milk were acquired at a local supermarket. Only samples used to study the effect of E-beam on the survival of both L. monocytogenes and L. innocua were deliberately contaminated. To do this, 0.5 ml of the bacterial suspension was divided in five aliquots of 0.1 ml, which were injected with disposable syringes in five separated zones of the cheese portion after removing wrappings and rinds aseptically. Then, samples were vacuum-packaged in 20 × 20 cm laminated film bags of low gas permeability (oxygen transmission rate of 35 cm3/24 h m2 bar and 150 cm3/24 h m2 bar to carbon dioxide) until vacuum reached 20 kPa. Once the samples were ready, they were treated in the irradiation plant (Ionisos Iberica, SA, Tarancón, Spain) and irradiated under an electron beam radiation source, which operates at 10 MeV. The radiation doses employed were between 0.2 and 2 kGy for kinetics studies and from 1 to 3 kGy for texture and sensory analyses. The dose absorbed by samples was checked by determining the absorbance of cellulose triacetate dosimeters [1] simultaneously irradiated with samples. Experiments were made in triplicate and performed at room temperature (18–20°C). The product temperature increased less than 2ºC during treatment. Following the treatment, samples

VELASCO ET AL.

were transferred to the laboratory where samples intended for determining the death kinetics and the selected parameters corresponding to the day zero were immediately processed. The remaining samples were stored at 4 and 14°C (the latter as an example of temperature abuse) to study the microbiota and behaviour of selected characteristics over time.

Physicochemical analysis. Moisture, ash, protein, fat, pH and aw were measured as previously described [52]. The pH was determined in the rind and core of cheese samples, using a Crison Digit-501 pH meter (Crison Instruments LTD, Barcelona, Spain).

Microbial analyses. To count survivors, the samples, after weighing, were homogenized with 20 ml of sterile saline solution in a Stomacher bag.

Total viable counts (TVC) were determined by the pour-plate method using Plate Count Agar (PCA; Difco) as culture medium. Lactobacilli counts were performed in double layer acidified (pH 5.5) MRS agar (Conda-Pronadisa, Madrid, Spain) as recommended by Henri-Dubernet et al. [26]. Incubation was carried out at 32°C for 48 h.

To determine inactivation parameters of both L. monocytogenes and L. innocua, Palcam agar base (Oxoid, Basingstoke, UK) with egg yolk emulsion (Oxoid) and selective supplement (polymyxin B, acriflavine HCl and ceftazimide, Oxoid) was used. Plates were incubated at 37°C for 48 h. Despite the selective agents present in the Palcam agar, this medium has been found previously to be perfectly suitable for the determination of listeria death kinetics [9]. Colonies were enumerated with a Digital S Colony counter (J. P. Selecta, Barcelona, Spain).

Survival curves were constructed by plotting log CFU/g against irradiation dose. Decimal reduction doses (D-values) were calculated from the linear regression equation of survival curves. The TVC and LAB growth curves were constructed according to the Baranyi and Roberts model [4] using the Excel add-in fitting curves Dmfit [http://www.ifr.ac.uk/safety/dmfit/].

Risk assessment. For risk assessment the USDA recommends a “zero tolerance” policy for L. monocytogenes in RTE products, equivalent to a Food Safety Objective (FSO) of 4 CFU/100g (log10 = −1.39). In the EU, European Commission Regulation (EC) No. 1441/2007 [18] divides the RTE foods into two categories according to the ability of L. monocytogenes to grow in them. So, for the material with an aw above 0.92, which permits the growth of L. monocytogenes, the “zero tolerance” criterion would be applied.

Products must not contain L. monocytogenes in 25 g at the time they leave the production plant. When the product is on the market, the limit is 100 CFU/g (log10 = 2) throughout the shelf life [18].

To establish the Process Criteria (PC), i.e., the irradiation dose required to reach the FSO, the initial contamination and the listeria growth throughout the shelf-life must be taken into account. The current milk pasteurization conditions are sufficient to achieve a decrease of 5-6 decimal reductions in L. monocytogenes [16]. In cheeses made from pasteurized milk, the original listeria level therefore depends upon the potential contamination during the preparation of portions.

The post-milking environment has been identified as the main source of contamination with L. monocytogenes. This was detected in milk from 90% of herds although at low concentrations, accounting for around 2.25 CFU/ml in bulk tank milk [5], but pasteurization causes a reduction of 5.2 log units of L. monocytogenes according to the data reported by Mackey and Bratchell [37], resulting in a listeria concentration of 1.42 × 10–5 cells/ml. To manufacture each piece of Camembert (250 g) and Brie cheese (2,250 g), 2.2 and 22.5 litres of milk are required, respectively [45]. During whey draining, 10% of bacterial cells should be eliminated [45]. So, 0.03 and 0.29 cells should be retained in each Camembert and Brie curd, respectively (ca. 10–4 cell/g). Ryser and Marth [44] analysed the changes in L. monocytogenes numbers during the manufacture and ripening of Camembert cheeses made from raw milk artificially contaminated with 2.6–2.9 log CFU/ml. After 65 days of ripening,

E-BEAM RADIATION ON L. MONOCYTOGENES

counts reached 6.30-7.51 log CFU/g. A similar increase was obtained in the listeriosis risk analysis carried out by Sanaa et al. [45] for other white mould ripened cheeses. From an initial load of 0.8 and 0.3 cells of L. monocytogenes per litre of raw milk, the former authors calculated, at the time of consumption, 3 and 5 cells/g in Camembert and Brie varieties, respectively. Therefore, if the estimated increase is about 4 log units, from an initial concentration in the curd of 1.12 × 10–4 and 1.29 × 10–4 cell/g (i.e., 0.03 and 0.29 cells/curd) the Camembert and Brie cheeses should contain a final load of 1.12 and 1.28 CFU/g (log10 = 0.05–0.11). These values are lower compared to the contamination that could occur post-processing or during size reduction adopted by other authors for products such as sausages [29] and cooked ham [8], assumed to be around 10 cells/g (log10 = 1). Following the same criterion for the preparation of cheese portions, L. monocytogenes contamination during this operation could reach, in the worst case, the amount of 10 cells/g. On the other hand, from the results of several authors compiled by FDA [22], a growth rate of 0.071 log CFU/day for L. monocytogenes in Camembert and Brie cheeses stored at 4°C has been calculated. Assuming for cheeses, a shelf-life of 60 days at 4°C, a final bacterial load of 5.26 log units (0.071× 60 + 1) would be achieved. From these data, it can be estimated that 3.26 and 6.65 decimal reductions would be required (performance criteria) to reach the FSO for EU and USDA statements, respectively.

Texture analysis. Cheeses were sampled on day 0 and after 15 (two weeks) and 42 days (six weeks) of storage. The effect of E-beam radiation on the texture of cheeses and how it changes during storage were examined through a puncture test, as described by Herrero et al. [27], and a texture profile analysis (TPA) as previously reported [52]. Hardness, springiness, adhesiveness, cohesiveness, gumminess and chewiness parameters of samples were calculated.

Sensory analysis. A triangular test, performed as described by Velasco et al. [52], was carried out at different days during storage at both temperatures to study the behaviour of appearance, flavour and odour of samples Ebeam treated. In addition, panellists were asked to justify their answer with brief descriptions.

Statistical analysis. Excel (Microsoft, Redmond, WA, USA) was used to calculate the coefficients of determination (R2) of survival curves and to conduct the F test to compare them. For statistical analysis of the physicochemical tests results, a one-way ANOVA and Duncan’s test for multiplerange test were performed using Statgraphics Centurion XVI for Windows (Statistical Graphics Corporation, Rockville, MD, USA).

Results and Discussion Physicochemical characteristics. The chemical com­ po­sition (% on wet matter) of Camembert and Brie cheeses was: moisture, 45.08 ± 1.09 and 47.13 ± 0.85; fat, 31.90 ± 0.14 and 36.77 ± 3.38; protein 24.52 ± 1.56 and 18.98 ± 0.65; and ash 2.88 ± 0.11 and 2.66 ± 0.21, respectively. Moisture values were slightly lower than those reported in the literature (45–47% vs.

48–52%) while the fat (32–37% vs. 24–27%) and protein (19– 24% vs. 19–21%) contents were higher [13,23], but, consequently, both compounds were similar in dry matter terms.

No significant (P > 0.05) dose effect was observed in the aw values immediately after irradiation treatment in both

Int. Microbiol. Vol. 18, 2015

35

cheeses (Table 1), as previously recorded in cheese slices [52]. The initial values were 0.960 and 0.964 for Camembert and Brie, respectively (Table 1), similar to values (0.967 and 0.965) found by Marcos et al. [38] for both cheeses and slightly lower than the value (0.97) reported by Guinee and Fox [25]. At the end of storage, the higher the dose the higher the aw of Brie samples, which reached values of 0.983 and 0.974 in samples treated at 3 kGy and stored for 41 days at 4°C and 19 days at 14°C, respectively with some significant differences (P < 0.05). However, this trend was not observed in the Camembert cheese because there was a significant increase (P < 0.05) in this parameter during storage, independently of the dose applied. After 41 days storage at 4°C and 19 days at 14°C, the average values were 0.978 and 0.973, respectively. The radiolysis of water present in both the cheese and the package environment resulted in the release of hydroxyl radicals, whose concentration diminished over time to return to form water molecules [46] and could contribute to an increase in water activity at the end of storage. Nevertheless, although in some instances significant differences were found, the range of values, from 0.983 and 0.957 (Table 1), was not broad enough to considerably inhibit either the growth of L. monocytogenes or the dominant microbiota (LAB), since the minimum aw for growth of the former pathogen has been established at 0.92 [16] and LAB may grow perfectly well at aw below 0.96 [50]. Indeed, the drop in aw is the basis of the lactic fermentation that occurs in traditional dry fermented sausages by LAB [43].

In the core of both cheeses, no relation between storage time and pH values was observed (data not shown). No dose effect (P > 0.05) was found in the pH at any of the times tested, consistent with the findings reported by various authors for irradiated cheeses, e.g., Konteles et al. [34]. Figure 1 shows the effect of E-beam treatment on the pH values in rind of Camembert and Brie cheeses and their changes during storage at 4 and 14°C. At the beginning of storage, the rinds had an average pH of 7.21 (Camembert) and 6.90 for (Brie), whereas inside the pH was significantly (P < 0.05) lower, with values of ca. 5.95 and 6.29. These values are in agreement with those reported by Spinnler and Gripon [48]. The increasing pH gradient from the centre to the surface is characteristic of white mould-ripened cheeses since Penicillium camemberti, present on the cheese surface, uses lactate as a carbon source for cell growth [35], which would result in a deacidification of the surface of the cheese. Geotrichum candidum might be implicated in the ripening of Camemberttype cheese [48], also helping to raise the pH either by the metabolism of amino acids as carbon and nitrogen sources

36

Int. Microbiol. Vol. 18, 2015

VELASCO ET AL.

Table 1. Effect of E-beam treatment and storage conditions on aw (mean values ± SD) of Camembert and Brie cheeses Storage Cheese Camembert

(ºC—days) 4

14

Brie

4

14

Dose 0 kGy

1 kGy

2 kGy

3 kGy

0

0.960 ± 0.002 b

0.957 ± 0.002 c

0.964 ± 0.005 b

0.962 ± 0.006 b

26

0.957 ± 0.002 β,b

0.966 ± 0.002 α,b

0.967 ± 0.002 α,b

0.969 ± 0.003 α,b

41

0.977 ± 0.003 a

0.979 ± 0.001 a

0.976 ± 0.003 a

0.981 ± 0.005 a

0

0.960 ± 0.002 b

0.957 ± 0.002 b

0.964 ± 0.005

0.962 ± 0.006

19

0.975 ± 0.003 α,a

0.972 ± 0.003 α,β,a

0.969 ± 0.002 β

0.972 ± 0.003 α,β

0

0.964 ± 0.009

0.958 ± 0.001 f

0.974 ± 0.007

0.965 ± 0.011 e

27

0.966 ± 0.007

0.969 ± 0.001 e

0.969 ± 0.009

0.970 ± 0.001 e

41

0.966 ± 0.003 γ

0.976 ± 0.004 β,d

0.978 ± 0.003 α,β

0.983 ± 0.001 α,d

0

0.964 ± 0.009

0.958 ± 0.001

0.974 ± 0.007

0.965 ± 0.011

19

0.957 ± 0.001 β

0.959 ± 0.002 β

0.971 ± 0.002 α

0.974 ± 0.004 α

: values in the same row with a different letter are significantly different (P < 0.05). and d, e, f: values in the same column with a different letter are significantly different in Camembert and Brie cheese, respectively (P < 0.05).

α, β, γ a, b, c

yielding ammonia or lactate utilization to maintain the cell structure during the stationary phase [3]. Furthermore, Debaryomyces hansenii, a yeast not included in the starters but possibly present because of their ubiquitous nature, could contribute to the surface alkalinisation [35]. During storage, the pH of the rinds became significantly more acid (P < 0.05) in all the samples until reaching a value of around 6.0 for both Camembert and Brie cheeses after 41 days of storage at 4°C and 5.67 and 5.84, respectively, after 19 days of storage at 14°C. This phenomenon could be explained by both the facultative anaerobic nature of LAB and the progressive death of the moulds caused by the low level of oxygen in the package, supported by the progressive disappearance of mycelium.

Microbial results. Lactococcus [13] and lactobacilli [26] are the dominant organisms in the paste of Camembert and Brie cheeses. Then, the fate of LAB could be assimilated to that of the TVC. The response of LAB microbiota (PCA counts) and lactobacilli (MRS counts) to the action of E-beam treatment is shown in Fig. 2. Initially, the TVC accounted was about 8.0–8.5 log CFU/g in both cheeses, which is normal for most cheese varieties [24]. The 1 kGy treatment resulted in a

decrease of about 2 (Camembert) and 3 (Brie) log units. When 2 kGy were applied, the reductions were around 5 and 4 log units, respectively. At 3 kGy a “tail” was observed in both cheeses, which may be explained by the presence of a heterogeneous indigenous microbiota consisting of the most radioresistant organisms, which would be responsible for the tail.

These organisms would belong to bacterial groups other than LAB because at 3 kGy, the lactobacilli (Fig. 2) had practically disappeared since the values came from counts of 3–4 colonies developed on the agar plates. Although lactobacilli counts of 3 kGy samples were approximated, they were fitted to a straight line, which confirms the well-known first order kinetics of bacteria inactivation by ionizing radiation. Since the straight portions of survivor curves were parallel, global Dvalues of 0.42 and 0.51 kGy could be roughly estimated for Camembert and Brie cheeses, respectively. These results are close to the range recorded by other authors. For example, a D-value of 0.66 kGy has been reported for mesophilic aerobic count in cheese slices [52] and 0.39 kGy could be estimated for TVC in Cottage cheese [32]. The differences may be explained by the fact that cheese always presents a mixed microbiota and the radioresistance of bacteria may be variable [52]. However, in some cheese varieties, much higher D-values have been re-

Int. Microbiol. Vol. 18, 2015

37

Int Microbiol

E-BEAM RADIATION ON L. MONOCYTOGENES

Fig. 1. Changes in the rind pH during storage at 4 and 14째C of Camembert (striped bars) and Brie (solid bars) cheeses treated by E-beam irradiation.

biota recovery was observed when 1 kGy was applied in such a way that in a couple of days the number of TVC superposed that of control (non-treated) samples, growing at a doubling time (g value) of 12 h. The same pattern was detected at 2-kGy doses but the growth rate was slower (g of about 24 h) achieving the 108 CFU/g level after 12 days of storage. The bacteria subjected to 3 kGy were probably not able to completely repair the damage produced by the E-beam and the highest level of

Int Microbiol

ported, such as 1.8 kGy in Feta cheese [34], probably due to factors related to bacterial radioresistance, which depend on both the inherent resistance of the organisms and the conditions of the matrix food [2,6].

The changes in TVC (total LAB) in Camembert cheese during storage at 4 and 14째C after E-beam treatment are shown in Fig. 3. As expected, no changes occurred in control samples throughout storage. In samples stored at 14째C, a quick micro-

Fig. 2. Survival curves of LAB (circles) and lactobacilli (triangles) for Camembert (A) and Brie (B) rindless cheeses.

Int. Microbiol. Vol. 18, 2015

VELASCO ET AL.

Int Microbiol

38

Fig. 3. TVC of control (diamonds) and irradiated Camembert cheese at 1 (squares), 2 (triangles) and 3 kGy (circles) stored at 4°C (full symbols) and 14°C (empty symbols).

them did not exceed 106 CFU/g after 33 days of storage. At 4°C, the TVC was maintained (at doses of 1 and 2 kGy) at the levels achieved just after the treatment application, but at 3 kGy a decrease was observed afterwards and no colonies were detected after 20 days of storage. TVC changes in the Brie cheese showed a similar trend than in the Camembert (data not shown). Briefly, the revival of microbiota was observed estimating a lag phase of about 10 days at 14°C when doses of 2 and 3 kGy were used and at 4°C even a slight growth occurred when 1 kGy was applied, giving a g value of about 500 h. In the lactobacilli counts, the same pattern was observed in both Camembert and Brie cheeses (data not shown). From all the former data taken together, it would seem that, although the LAB are the most important bacteria in both cheeses, the species present in the Brie cheese used in this work were more

sensitive than those of the Camembert variety. In a previous study on the use of E-beam radiation to diminish the late blowing of cheese, a recovery of the total mesophilic aerobic counts was reported, although the profile observed was slightly different [52].

Food safety aspects. According to the EU regulation [18], both cheeses are included in the category of RTE foods able to support the growth of L. monocytogenes since both the pH and aw are higher than 4.4 and 0.92, respectively. It is, therefore, a risk for the consumer if this bacterium is present.

The response of both L. innocua and L. monocytogenes to the E-beam treatment resulted in a first-order inactivation kinetics, as repeatedly reported [e.g. 39,40]. From the regression equations shown in Table 2, D-values 0.39 kGy for L. innocua and

Table 2. Irradiation decimal reduction values (D-values) for Listeria monocytogenes and Listeria innocua in Camembert and Brie cheeses Cheese Camembert

Brie

Survival equations

R2

D-value

L. innocua NCTC 11288

Log CFU/g = 7.42 – 2.80·Dose

0.96

0.36 kGy

L. monocytogenes Scott A

Log CFU/g = 9.26 – 2.95·Dose

0.995

0.34 kGy

L. innocua NCTC 11288

Log CFU/g = 7.29 – 2.55·Dose

0.94

0.39 kGy

L. monocytogenes Scott A

Log CFU/g = 9.26 – 2.92·Dose

0.999

0.34 kGy

Microorganism

E-BEAM RADIATION ON L. MONOCYTOGENES

0.34 kGy for L. monocytogenes can be calculated in Brie cheese and 0.36 and 0.34 kGy, respectively, in Camembert cheese. D-values of a similar order have been reported for L. monocytogenes in processed cheese slices treated with gamma radiation [47]. Higher D-values have been reported at refrigeration temperatures [32,33,50]: 0.84–0.93 kGy in sliced and pizza cheeses (10°C), 1.33 kGy in Feta cheese (0–2°C), and 1.38 kGy in soft whey cheese Anthotyros (4°C), respectively.

To estimate the dose required to meet the FSO (process criteria), the most unfavourable case was taken into account.

Thus, the highest D-value, matching L. innocua in Brie cheese, was chosen (0.39 kGy). According to the previously calculated performance criterion, 3.26 and 6.65 decimal reductions would be required depending on the destination country. Therefore, the process criteria would be 1.27 and 2.59 kGy, respectively. These doses are much lower (the former) and of the similar level (the latter) than that considered as acceptable (i.e., 2.5 kGy) by the Scientific Committee on Food of EU and allowed in France for the treatment of Camem­bert cheese made from raw milk [19]. However, Bougle and Stahl [7] have detected viable listeria but they were not able to grow at 12°C in Camembert cheese treated at 2.6 kGy. Likewise, a slower growth rate of L. monocytogenes and S. aureus has been observed in vacuum packaged cooked ham after treatment at 2 and 3 kGy [9].

Texture and sensory aspects. Immediately after Ebeam treatment, no significant differences (P > 0.05) in some textural attributes (adhesiveness, springiness and chewiness) were found in both cheeses. However, significant differences (P < 0.05) in other parameters (hardness, cohesiveness, gumminess and breaking force) were detected only in Camembert cheese while values for the Brie samples remained fairly constant. Nevertheless, no significant differences (P > 0.05) between untreated and treated samples after two weeks of storage were detected. A similar behaviour was observed in sensorial parameters. These results are in agreement with those by other authors, e.g., no significant differences were found in the sensory attributes of Camembert samples treated with doses up to 2.5 kGy [14]. Similarly, it has been reported that the texture of Feta and smear-ripened cheeses is not affected by irradiation treatments [17,50].

From the results obtained, it can be concluded that E-beam treatments at 1.27 and 2.59 kGy allowed the control of L. mono­cytogenes growth, achieving the FSO in soft mouldripened cheeses (Camembert and Brie varieties) according to EU and USDA criteria, respectively. Although the lactic acid microbiota was also reduced, its recovery occurred afterwards

Int. Microbiol. Vol. 18, 2015

39

in samples stored at 14°C (a favourable temperature for LAB growth) until reaching values close to the initial numbers.

Acknowledgements. This work has been supported by the Projects AGL2010-19158 and CARNISENUSA (CSD0007-00016) and the Group 920276 of the Complutense University of Madrid. R. Velasco was the beneficiary of a grant financed by the former CARNISENUSA project.

Competing interests. None declared.

References 1. American Society for Testing and Materials (2000) E1650-97e1 standard practice for use of cellulose acetate dosimetry systems. 12:2. ASTM, West Conshohocken, PA, USA 2. Augustin JC (1996) Resistance of Listeria monocytogenes to physical exposure. Pathol Biol 44:790-807 3. Aziza M, Adour L, Amran, A (2008) Assimilation of peptides and amino acids and dissimilation of lactate during submerged pure cultures of Penicillium camembertii and Geotrichum candidum. J Microbiol Biotechn 18:124-127 4. Baranyi J, Roberts, TA (1994) A dynamic approach to predicting bacterial growth in food. Int J Food Microbiol 23:277-294 5. Bemrah N, Sanaa M, Cassin MH, Griffiths MW, Cerf O (1998) Quantitative risk assessment of human listeriosis from consumption of soft cheese made from raw milk. Prev Vet Med 37:129-145 6. Black JL, Jaczynski J (2008) Effect of water activity on the inactivation kinetics of Escherichia coli O157:H7 by electron beam in ground beef, chicken breast meat, and trout fillets. Int J Food Sci Tech 43:579–586 7. Bougle DL, Stahl V (1994) Survival of Listeria monocytogenes after irradiation treatment of Camembert cheeses made from raw milk. J Food Protect 57: 811-813 8. Cabeza MC, Cambero I, de la Hoz L, Ordóñez JA (2007) Optimization of E-beam irradiation treatment to eliminate Listeria monocytogenes from ready-to-eat (RTE) cooked ham. Innov Food Sci Emerg 8:299-305 9. Cabeza MC, Cambero MI, Núñez M, Medina M, de la Hoz L, Ordóñez J (2010) Lack of growth of Listeria monocytogenes and Staphylococcus aureus in temperature abuse of E-beam treated ready-to-eat (RTE) cooked ham. Food Microbiol 27:777-782 10. Cambero MI, Cabeza MC, Escudero R, Manzano S, García-Márquez I, Velasco R, Ordóñez JA (2012) Sanitation of selected ready-to-eat (RTE) intermediate moisture foods (IMF) of animal origin by E-beam irradiation. Foodborne Pathog Dis 9:594-599 11. Centers for Disease Control and Prevention (2013) Multistate outbreak of listeriosis linked to Crave Brothers Farmstead cheeses. Available at: http://www.cdc.gov/listeria/outbreaks/cheese-07-13/index.html.

12. Centers for Disease Control and Prevention (2013) Summary of notifiable diseases—United States 2011. Morb Mortal Wkly Rep 60:1-117 13. Chapman HR, Sharpe ME (1990) Microbiology of cheese. In Robinson RK (ed) Dairy microbiology, vol. 2 The microbiology of milk products.

2nd ed. Elsevier, Essex, UK, pp 203-289 14. Chincholle R (1991) Action of the ionization treatment on the soft cheese made from unpasteurized milk. C R Acad Agric Fr 77:29-35 15. Codex Alimentarius Commission (2009) Guidelines on the application of general principles of food hygiene to the control of Listeria monocytogenes in Ready-To-Eat foods. CAC/GL 61-2007, Rev. 1. Available at: http:// www.codexalimentarius.org/download/standards/10740/CXG_061e. pdf. Accessed 9 January 2015

40

Int. Microbiol. Vol. 18, 2015

16. Doyle ME, Mazzota AS, Wang T, Wiseman DW, Scott VN (2001) Heat resistance of Listeria monocytogenes. J Food Protect 64:410-429 17. Ennahar S, Kuntz F, Strasser A, Bergaentzle M, Hasselmann C, Stahl V (1994) Elimination of Listeria monocytogenes in soft and red smear cheeses by irradiation with low-energy electrons. Int J Food Sci Tech 29:395-403 18.

European Commission (2007) Commission Regulation (EC) No 1441/2007 of 5 December 2007 amending Regulation No 2073/2005 on microbiological criteria for foodstuffs. Off J Eur Union L322:12-29 19. European Food Safety Authority (2011) Statement summarizing the conclusions and recommendations from the opinions on the safety of irradiation of food adopted by the BIOHAZ and CEF Panels. EFSA J 9: 2107 20. European Food Safety Authority (2011) The European Union summary report on trends and sources of zoonoses, zoonotic agents and foodborne outbreaks in 2009. EFSA J 9:2090 21. European Food Safety Authority (2013) The European Union summary report on trends and sources of zoonoses, zoonotic agents and foodborne outbreaks in 2011. EFSA J 11:3129 22. US Food and Drug Administration, US Food Safety and Inspection Service (2003) Quantitative assessment of relative risk to public health from foodborne Listeria monocytogenes among selected categories of ready-to-eat foods. Appendix 8: Growth of Listeria monocytogenes in foods. Available at: http://www.fda.gov/downloads/Food/ /FoodScienceResearch/UCM197321.pdf 23. Food Standards Agency (2002) McCance and Widdowson’s the composition of foods, 6th ed. Royal Society of Chemistry, Cambridge, UK 24. Fox PF, McSweeney PLH, Cogan TM, Guinee TP (2004) Cheese: Chemistry, physics and microbiology, Vol. 2 Major Cheese Groups. 3rd ed. Elsevier Applied Science, London, UK 25. Guinee TP, Fox PF (2004) Salt in cheese: Physical, chemical and biological aspects. In Fox PF, McSweeney PLH, Cogan TM, Guinee TP (eds) Cheese: Chemistry, physics and microbiology, Vol. 2 Major Cheese Groups, 3rd ed. Elsevier Applied Science, London, UK, pp 208-259 26. Henri-Dubernet S, Desmasures N, Guéguen M (2008) Diversity and dynamics of lactobacilli populations during ripening of RDO Camembert cheese. Can J Microbiol 54:218-228 27. Herrero AM, Cambero MI, Ordóñez JA, de la Hoz L, Carmona P (2009) Plasma powder as cold-set binding agent for meat system: Rheological and Raman spectroscopy study. Food Chem 113:493-499 28. Hoz L, Cambero MI, Cabeza MC, Herrero AM, Ordóñez JA (2008) Elimination of Listeria monocytogenes from vacuum-packed dry-cured ham by E-Beam radiation. J Food Protect 71:2001-2006 29. International Commission on Microbiological Specifications for Foods (2002) Listeria monocytogenes in cooked sausage (Frankfurters). In Tompkin RB, Garam L, Roberts TA, Buchanan RL, van Schothorst M, Dahms S, Cole MB (eds), Microorganisms in foods: Microbiological testing in food safety management, vol. 7. Kluwer/Plenum, New York, USA, pp 285-312 30. Jackson KA, Biggerstaff M, Tobin-D’Angelo M, Sweat D, Klos R, Nosari J, Garrison O, Boothe E, et al. (2011) Multistate outbreak of Listeria monocytogenes associated with Mexican-style cheese made from pasteurized milk among pregnant, Hispanic women. J Food Protect 76:949-953 31. Johnson EA, Nelson JH, Johnson M (1990) Microbiological safety of cheese made from heat-treated milk, Part I. Executive summary, introduction and history. J Food Protect 53:441-452 32. Jones TH, Jelen P (1988) Low dose-irradiation of Camembert, Cottage cheese and Cottage whey. Milchwissenschaft 43:233-235 33. Kim HJ, Ham JS, Lee JW, Kim K, Ha SD, Jo C (2010) Effects of gamma and electron beam irradiation on the survival of pathogens inoculated into sliced and pizza cheeses. Radiat Phys Chem 79:731-734

VELASCO ET AL.

34. Konteles S, Sinanoglou VJ, Batrinou A, Sflomos K (2009) Effects of γ-irradiation on Listeria monocytogenes population, colour, texture and sensory properties of Feta cheese during cold storage. Food Microbiol 26:157-165 35. Lessard MH, Bélanger G, St-Gelais D, Labrie S (2012) The composition of Camembert cheese ripening cultures modulates both mycelial growth and appearance. Appl Environ Microb 78:1813-1819 36. Lianou A, Sofos JN (2007) A review of the incidence and transmission of Listeria monocytogenes in ready-to-eat products in retail and food service environments. J Food Protect 70:2172-2198 37. Mackey BM, Bratchell N (1989) The heat resistance of Listeria monocytogenes: A review. Lett Appl Microbiol 9:89-94 35. Marcos A, Esteban MA, Alcalá M (1990) Determination of water activity in Brie and Camembert cheese varieties by four different methods. Food Chem 38:189-199 39. Medina M, Cabeza MC, Bravo D, Cambero I, Montiel R, Ordóñez JA, Nuñez M, Hoz L (2009) A comparison between E-beam irradiation and high pressure treatment for cold-smoked salmon sanitation: microbiological aspects. Food Microbiol 26:224–227 40. Miller RB (2005) Electronic irradiation of foods: An introduction to the technology. Springer Science, New York, USA 41. Moseley BEB (1989) Ionizing radiation: Action and repair. In Gould GA (ed.), Mechanisms of action of food preservation procedures. Elsevier, Essex, UK, pp 43-70 42. Niemira BA, Fan X, Sokorai KJB, Sommers CH (2003) Ionizing radiation sensitivity of Listeria monocytogenes ATCC 49594 and Listeria innocua ATCC 51742 inoculated on endive (Cichorium endiva). J Food Protect 66:993-998 43. Ordóñez JA, Hierro EM, Bruna JM, de la Hoz L (1999) Changes in the components of dry-fermented sausages during ripening. Crit Rev Food Sci 39:329-367 44. Ryser ET, Marth EH (1987) Fate of Listeria monocytogenes during manufacture and ripening of Camembert cheese. J Food Protect 50:372378 45. Sanaa M, Coroller L, Cerf O (2004) Risk assessment of listeriosis linked to the consumption of two Soft cheeses made from raw milk: Camembert of Normandy and Brie of Meaux. Risk Anal 24:389-399 46. Seisa D, Osthoff G, Hugo C, Hugo A, Bothma C, Van der Merwe J (2004) The effect of low-dose gamma irradiation and temperature on the microbiological and chemical changes during ripening of Cheddar cheese. Radiat Phys Chem 69:419-431 47. Sommers CH, Boyd G (2005) Elimination of Listeria monocytogenes from Ready-to-Eat turkey and cheese tortilla wraps using ionizing radiation. J Food Protect 68:164-167 48. Spinnler HE, Gripon JC (2004) Surface mould-ripened cheeses. In Fox PF, McSweeney PLH, Cogan TM, Guinee TP (eds) Cheese: Chemistry, physics and microbiology, Vol. 2 Major Cheese Groups, 3rd ed. Elsevier Applied Science, London, UK, pp 208-259 49. Taormina PJ, Beuchat LR (2001) Survival and heat resistance of Listeria monocytogenes after exposure to alkali and chlorine. Appl Environ Microb 67:2555-2563 50. Troller JA, Stinson JV (1981) Moisture requirements for growth and metabolite production by Lactic Acid Bacteria. Appl Environ Microb 42:682-687 51. Tsiotsias A, Savvaidis I, Vassila A, Kontominas M, Kotzekidou P (2002) Control of Listeria monocytogenes by low-dose irradiation in combination with refrigeration in the soft whey cheese ‘Anthotyros’.

Food Microbiol 19:117-126 52. Velasco R., Ordóñez JA, Cabeza MC, Hoz L, Cambero MI (2011) Use of the E-beam radiation to diminish the late blowing of cheese. Int Dairy J 21:493-500 53. Warriner K, Namvar A (2009) What is the hysteria with Listeria? Trends Food Sci Tech 20:245-254

RESEARCH ARTICLE International Microbiology (2015) 18:41-49 doi:10.2436/20.1501.01.233. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Solar water disinfection (SODIS): Impact on hepatitis A virus and on a human Norovirus surrogate under natural solar conditions David Polo,1 Irene García-Fernández,2 Pilar Fernández-Ibáñez,2 Jesús L. Romalde1* Department of Microbiology and Parasitology, CIBUS-School of Biology, University of Santiago de Compostela, Santiago de Compostela, Spain. 2Solar Platform of Almería, CIEMAT, Tabernas, Spain

1

Received 10 March 2015 · Accepted 31 March 2015

Summary. This study evaluates the effectiveness of solar water disinfection (SODIS) in the reduction and inactivation of hepatitis A virus (HAV) and of the human Norovirus surrogate, murine Norovirus (MNV-1), under natural solar conditions.

Experiments were performed in 330 ml polyethylene terephthalate (PET) bottles containing HAV or MNV-1 contaminated waters (103 PFU/ml) that were exposed to natural sunlight for 2 to 8 h. Parallel experiments under controlled temperature and/ or in darkness conditions were also included. Samples were concentrated by electropositive charged filters and analysed by RT-real time PCR (RT-qPCR) and infectivity assays. Temperature reached in bottles throughout the exposure period ranged from 22 to 40ºC. After 8 h of solar exposure (cumulative UV dose of ~828 kJ/m2 and UV irradiance of ~20 kJ/l), the results showed significant (P < 0.05) reductions from 4.0 (±0.56) ×104 to 3.15 (±0.69) × 103 RNA copies/100 ml (92.1%, 1.1 log) for HAV and from 5.91 (±0.59) × 104 to 9.24 (±3.91) × 103 RNA copies/100 ml (84.4%, 0.81 log) for MNV-1. SODIS conditions induced a loss of infectivity between 33.4% and 83.4% after 4 to 8 h in HAV trials, and between 33.4% and 66.7% after 6 h to 8 h in MNV-1 trials. The results obtained indicated a greater importance of sunlight radiation over the temperature as the main factor for viral reduction. [Int Microbiol 2015; 18(1):41-49] Keywords: Solar water disinfection (SODIS) · water disinfection · hepatitis A virus (HAV) · murine Norovirus (MNV-1)

Introduction Water scarcity and the lack of access to sanitation in developing countries continue to be global health challenges. Despite progress towards the Millennium Development Goals, more * Corresponding author: J.L. Romalde Departamento de Microbiología y Parasitología Facultad de Biología Universidad de Santiago de Compostela 15782 Santiago de Compostela, Spain Tel. +34-881816908. Fax +34-881816966 E-mail: jesus.romalde@usc.es

than 768 million people (~11% of the global population) remain without access to safe drinking water sources [42]. Consumption of untreated or improperly treated water is one of the most common routes for enteric disease outbreaks and is a priority issue to solve in order to reduce morbidity and mortality in the developing world [7,18].

Household water treatment and storage (HWTS) have demonstrated to be among the most effective ways to reduce the incidence of waterborne diseases in regions without access to adequately treated drinking water. It constitutes a low cost, easy to use and sustainable water treatment, complying with basic criteria for acceptance in these developing zones

42

Int. Microbiol. Vol. 18, 2015

[9,18,32]. Solar disinfection (SODIS) is a HWTS method that has been gaining popularity over the last 30 years [1,32,41].

The SODIS technique consists of exposing small-volumes (up to 3 l) of contaminated water with low turbidity (<30 NTU) in transparent containers (usually polyethylene-terephthalate [PET] bottles) to direct sunlight for at least 6 h (or 2 consecutive days if there is more than 50% of cloud cover) during the maximum intensity of radiation [32]. Biocidal effects of SODIS are attributed to optical (UVA) and solar mildheating mechanisms [8,26]. The disinfection efficacy of SODIS depends principally on the solar irradiance, water temperature, turbidity, dissolved oxygen and resistance of the type of microorganism [31].

Solar UV radiation consists of UV-C (λ = 100–280 nm), UV-B (λ = 280–320 nm) and UV-A (λ = 320–400). However, only UV-A, and a small part of UV-B reach the earth surface [20,40]. UV-B may directly damage nucleic acids through formation of pyrimidine dimers. UV-A photons, the main component responsible for the disinfecting action of SODIS, are not sufficiently energetic to modify directly nucleic acids like UV-B and UV-C. However, it causes indirect damage to structural components and DNA of cells through photosensitizers and may generate reactive oxygen species (ROS) in water including singlet oxygen (1O2), superoxide (O2–), hydrogen peroxide (H2O2), and hydroxyl radicals (•OH), which can form single strand breaks, nucleic base modifications as well as induce oxidations in proteins and membrane lipids [15,26,36].

Previous studies reported reductions in the incidence of diarrhoea using the SODIS method [10]. Recently, new research proved that SODIS can significantly reduce rates of childhood dysentery and infantile diarrhoea by 45% [13]. Beyond its health benefits through reduction of waterborne diseases, additional studies have also demonstrated that SODIS has positive health impact for Kenyan children under 5 years as they showed significant increase of weight and height [13].

SODIS efficacy has been already demonstrated for a wide range of microorganisms including bacteria, fungi and protozoan parasites [8,19,28,37]. Studies evaluating SODIS against human viruses under real field conditions are scarce. Some research has been done using enteric viruses and viral indicators such as bacteriophages under simulated sunlight and laboratory conditions or real sunlight in natural conditions [2,22, 23,29,45]. From a public health perspective, only a few human enteric viruses have been shown epidemiologically to be waterborne transmitted and widely detected in the environment. Hepatitis A virus (HAV) (fam. Picornaviridae) and Norovirus (NoV) genogroups I and II (fam. Caliciviridae) are

POLO ET AL.

among the leading aetiological viral pathogens transmitted by water and food [6,34,43,44]. HAV is the main cause of acute hepatitis worldwide and the WHO regard it as reference patho­gen for drinking water risk analysis [17]. NoV is the most important foodborne infectious agent of gastroenteritis outbreaks worldwide [3,34]. Human NoV are non-culturable under laboratory conditions, however, murine norovirus (MNV) has been demonstrated to be an useful human NoV surrogate due to its similarity in genetic and environmental stability properties [4]. In addition, viral pathogens, especially non-enveloped viruses such as hepatitis A virus and noroviruses, besides having a high environmental stability and a low minimal infective dose, are resistant to commonly disinfection processes and persistent in water supply systems [6,43].

The main objective of this study was to evaluate and compare the effectiveness of SODIS method for the disinfection of HAV and MNV-1, two high resistant pathogens, in distilled water under natural solar conditions.

Materials and methods Cell culture and viral stocks. HAV HM-175/18f was obtained from the ATCC as a cell culture-adapted cytopathic clone of strain HM-175. MNV-1, a culturable calicivirus genetically similar to human NoV [4,47] was kindly provided by Dr Herbert W. Virgin IV (University of Washington, USA). A mutant non-virulent infective strain of Mengovirus (vMC0), kindly provided

by Dr. Albert Bosch (University of Barcelona) was employed as RNA extraction control as it was previously described [11]. Stocks of each viral strain were generated by inoculation onto confluent monolayers of appropriate cell lines (FRhK-4, RAW 267.4 and HeLa for HAV HM-175, MNV-1 and vMC0, respectively). The stocks were purified previously to their use performing 5 series of freezing/thawing to complete release of the viral particles from the cells and centrifuged at 2000 ×g for about 20 minutes at 37°C to remove cellular debris and titrated by plaque assay [12]. (Plaque forming units: PFU.) Final concentration of each viral stock was 1 × 105 PFU/ml for mengovirus vMC0, and 1 × 106 PFU/ml for HAV and MNV-1.

Solar experiments. All assays were carried out at Plataforma Solar de Almería, Tabernas dessert (Spain) (37.09º N, 2.36º W). SODIS experiments were performed contaminating volumes of 330-ml distilled water contained in PET bottles at initial HAV or MNV-1 concentration of 103 PFU/ml and then exposing the bottles (from here SODIS-bottles) on concrete surface directly to the action of natural solar radiation during 2, 4, 6 and 8 h. The distilled water used had a conductivity of <10µS/cm, Cl– = 0.7–0.8 mg/l, NO3– = 0.5 mg/l and dissolved organic carbon <0.5 mg/l. Water temperature was measured during the experiments every two hours with a thermometer (model HI 98509-1, Hanna Instruments, Eibar, Spain).

In order to discriminate the effect of solar radiation and temperature, parallel assays under controlled temperature and in darkness conditions were also included. Bottles under controlled temperature (from here Bath-bottles) were maintained at 25 ± 1°C in a cooled water bath. The cooled water bath consisted of a plastic container filled with water up to half of the bottles, which were placed in a horizontal position as the rest of the bottles not cooled. The water was renewed every so often to maintain the temperature.

SODIS OF HAV AND MURINE NOROVIRUS

Bottles in darkness (from here Dark-bottles) were wrapped in opaque aluminium foil and exposed to solar radiation. Control-bottles were wrapped in opaque aluminium foil and maintained at 25 ± 1°C in a cooled water bath.

Each treatment was performed in triplicate and in perfectly clear sunny days of July 2012 at Plataforma Solar de Almería (South of Spain). A solar energy unit, QUV, is a term commonly used to compare results under different conditions [16]. UV radiation was measured in continuum with a global UV-A radiometer (295–385 nm, Model CUV3, Kipp & Zonen, Netherlands) on a horizontal platform, with a typical sensibility of 264 mV/W m2. The radio­ meter provides data in terms of incident W/m2, which is defined as the solar radiant energy rate incident on a surface per unit area. UV dose (kJ/m2) is dependent on UV intensity and time, and is given by Equation 1: dose = I × ∆t where I is the average irradiation intensity, W/m2, and Dt is the experimental time, in seconds. Moreover, the inactivation kinetics can be plotted as function of cumulative energy per unit of volume (QUV, kJ/l) received by the bottles, and calculated by Equation 2:

where Quv,n, Quv, n–1 , are the UV energy accumulated per unit volume (kJ/l) at times n and n–1, respectively, UVG,n is the average incident irradiation on the irradiated area, Δtn is the experimental time of sample, Ar is the illuminated area of the solar bottle (m2), and Vt is the total volume (l) of treated water.

Virus concentration and RNA extraction. Virus recovery from water samples was carried out following the principles outlined in the recently developed standard method for virus detection in foodstuffs, included bottled water (ISO/TS 15216-1:2013) with minor modifications. The concentration of viral particles from each sample was performed by filtration using electro-positive charged filters (Virocap filters, Scientific Methods, USA).

After the adsorption to filters, the viral particles were eluted by an alkaline solution of pH = 9.5 (Beef extract 1.5%, 0.25 mol/l glycine, Tween 80 0.1%) in a final volume of 5 ml. Additional 2 ml of the alkaline solution were added to the empty bottle, shaked during 10 min and added to the previous eluate.

Then, pH was adjusted to 7.5 with 0.1 mol/l HCl and viral particles were concentrated by PEG 8000 (8%) with a vigorous stirring for 2 h. After centrifugation at 10,000 ×g for 1 h the pellet was resuspended in 1 ml of PBS. The viral RNA from each sample was extracted using a commercial kit (NucleoSpin RNA virus, Macherey-Nagel, Germany). This method is based on the guanidine thiocyanate disruption and the adsorption of RNA to silica columns. Known amounts of mengovirus clone (vMC0) (10 µl of mengovirus stock) were previously spiked to each sample as an independent nucleic acid extraction efficiency control [11]. To determine the extraction efficiency, cycle threshold (Ct) value for the Mengovirus-positive amplification control and the Ct value of each sample for the Mengovirus were compared and classified as valid (> 5%) or invalid (<5%). Following the ISO technical specifications, samples with a < 5% extraction efficiency were re-extracted again.

Reverse transcriptase-real time PCR (RT-qPCR). RT-qPCR method was carried out according to the CEN/ISO standard method. RT-qPCR was performed on an Mx3005p QPCR System (Stratagene, USA) thermocycler, using TaqMan probes and Platinum Quantitative RT-PCR Thermoscript One-step System kit (Invitrogen, Saint Aubin, France) (25 µl final volume) with 5 µl of extracted RNA. Primer set and probe used were: 0.9 μmol/l of reverse primer HAV240 (5′-GGAGAGCCCTGGAAGAAAG-3′),

Int. Microbiol. Vol. 18, 2015

43

0.5 μmol l–1 of forward primer HAV68 (5′-TCACCGCCGTTTGCCTAG-3′) and 0.45 μmol l–1 of probe HAV150 (6-FAM-CCTGAACCTGCAGGAATTAA-MGB) for HAV [11]. For MNV-1, 0.2 μmol/l of reverse primer RvORF1/ORF2 (5′-GCGCTGCGCCATCACTC-3′), 0.2 μmol/l of forward primer Fw-ORF1/ORF2 (5′-CACGCCACCGATCTGTTCTG-3′) and 0.2 μmol/l of probe MGB-ORF1/ORF2 (6-FAM-CGCTTTGGAACAATGMGB) [5].

Amplification conditions for HAV were: reverse transcription at 55ºC for 1 h, denaturation at 95ºC for 5 min, followed by 45 cycles of amplification with a denaturation at 95ºC for 15 s, annealing at 60ºC for 1 min, and extension at 65ºC for 1 min. Amplification conditions for MNV-1 were as previously described [5] with minor modifications. Briefly, after a RT step at 45ºC for 1 h, PCR amplification was carried out with a initial denaturation at 95ºC for 5 min, and 50 cycles of amplification with a denaturation at 95ºC for 15 s and annealing-extension step at 60ºC for 1 min. Primers/probe set and amplification conditions for Mengovirus are those specified in the standard method ISO/TS 15216-1:2013.

The presence of RT-PCR inhibitors and the determination of the RT-qPCR efficiency were tested by means of the external controls (EC) included for each reaction. Briefly, 2.5 µl of EC, containing 103 genome copies of appropriated virus (HAV or MNV-1), were mixed with 2.5 µl of each sample extracted RNA and the Ct values of these reactions were compared with the Ct value obtained for the EC in RNA-free sterile water. Then the efficiency was classified as valid (> 25%) or invalid (< 25%). Following the ISO technical specifications, samples with < 25% RT-qPCR efficiency were tested again.

Negative controls containing no nucleic acid as well as positive controls were also introduced in each run. A sample displaying a Ct ≤ 41, with no evidence of amplification in the negative controls, was considered as positive. Quantification was estimated by standard curves constructed with serial dilutions of HAV or MNV-1 RNA, plotting the number of genome copies against the Ct. This quantification was not corrected with the extraction or RT-qPCR efficiencies, following the recommendations of the ISO standard method.

Infectivity assays. The infectivity of HAV and MNV-1 remaining in water samples at the end of each experimental period was evaluated in confluent FRhK-4 cells for HAV and RAW 267.4 for MNV-1 in 48-well cell culture plates. Each sample was tested in duplicate using 100 µl of viral concentrate per well. Once inoculated, plates were incubated 1 h at 37°C with slow agitation to promote virus attachment and internalization. After this period, the cells were washed with PBS (pH 7.4) to avoid the toxicity of beef extract components [25], and then maintenance medium was added to each well.

Maintenance medium consisted on DMEM supplemented with 2% foetal bovine serum, 1X non-essential amino acids, 2 mmol/l l-glutamine and 100 UI-100 UI mg/ml penicillin-streptomicin (Lonza-BioWhittaker, Belgium).

The plates were then incubated at 37°C and 5% of CO2 and microscopically examined daily for cytopathic effect (CPE) during 21 and 5 days for HAV and MNV-1, respectively. Negative samples were subjected to a blind passage to avoid false negative results. Appropriate negative and positive controls were included. Negative controls consist on FRhK-4 or RAW 267.4 cells inoculated with sterile water filtered and subjected to the same conditions than the samples. Positive controls consist on the appropriate cell line inoculated with HAV or MNV-1 stock solutions.

Statistical analysis. One-way ANOVA analysis was performed to compare the differences in the percentage of viral elimination obtained between viruses and exposure conditions. Moreover, post-hoc tests were employed to determine the statistical significance of the viral reduction between each exposure conditions using the Tukey’s and Dunnett’s tests. Significance level was established at P < 0.05. All statistical analyses were performed using the SPSS v20.0.0 software statistical package (IBM Corp., Madrid, Spain).

Int. Microbiol. Vol. 18, 2015

Results The maximum local noon UV irradiances recorded for HAV and MNV-1 experiments were 35.4 W/m2 and 38.2 W/m2, respectively (Fig. 1). The accumulated UV dose and QUV at the end of the exposure period were 820.692 kJ/m2 and 19.9 kJ/l for HAV, and 834.876 kJ/m2 and 20.24 kJ/l for MNV-1 (Fig. 2).

The average water temperature profiles reached in bottles along the exposure time for HAV and MNV-1 experiments are shown in Figure 1. Maximum water temperatures recorded within the PET bottles in HAV trials were 38.6°C in SODISbottles, 37.4°C in Dark-bottles, 29.1ºC in Bath-bottles and 27.3ºC in Control-bottles. In MNV-1 trials, the maximum water temperatures reached were: 40.7°C in SODIS-bottles, 39.7°C in Dark-bottles, 30.9°C in Bath-bottles and 27.8°C in Control-bottles.

All samples yield valid extraction and RT-qPCR efficiency values. Extraction values ranged from 20 to 100% for HAV

POLO ET AL.

and from 26 to 100% for MNV-1. RT-qPCR efficiencies ranged from 52 to 100% for HAV and from 43 to 100% for MNV-1. The average viral quantification at initial time (0 h of solar exposure) was 4.0 (± 0.56) ×104 and 5.91(± 0.59) × 104 RNA copies/100 ml (RNAc/100 ml) for HAV and MNV-1, respectively. After 8h of solar exposure, average quantification values in SODIS bottles were 3.15 (± 0.69) × 103 and 9.24(±3.91) × 103 RNAc 100 ml–1 for HAV and MNV-1, respectively. These values represent an average decrease of 92.1% (1.1 log) for HAV (Table 1) and 84.4% (0.81 log) for MNV-1 (Table 2). The average decreases for Bath-, Dark- and Control-bottles were 85.1% (0.83 log), 36.7% (0.20 log) and 17.4% (0.08 log) for HAV; and 61.6% (0.42 log), 37.4% (0.20 log) and 10.3% (0.05 log) for MNV-1 (Tables 1 and 2; Fig. 2).

Statistical analyses did not show significant differences between HAV and MNV-1 removal rates (P > 0.05). However, statistical differences were observed with Dunnet’s and Tukey tests between treatments, both in HAV and MNV-1 trials (Table 1 and 2). Significant differences with regard to the control were

Int Microbiol

44

Fig. 1. Average solar UV irradiance (295–385 nm) along the solar exposure period and profiles of the mean water temperatures recorded in bottles for each exposure condition performed with HAV (A) and MNV-1 (B).

SODIS OF HAV AND MURINE NOROVIRUS

Int. Microbiol. Vol. 18, 2015

Int Microbiol

Fig. 2. HAV (A) and MNV-1 (B) RNA removal expressed as % RNA copies along the exposure time (2, 4, 6 and 8 h). Solar UV dose (W/m2) and cumulative UV radiation (QUV, kJ/l) are also shown in secondary axis.

observed for SODIS (P < 0.01) and constant temperature conditions (Bath-bottles) (P < 0.05), but no for darkness conditions (Dark-bottles) in HAV trials. For MNV-1, all conditions showed significant differences, P < 0.001 for SODIS, P < 0.01 for constant temperature conditions and P < 0.05 for darkness conditions.

With regard to infectivity assays, water samples from control-bottles and Dark-bottles maintained their infectivity capacity along the study period for both viruses. Water samples from Bath-bottles showed a one third decrease (33.4%) in their infectivity capacity after 6 and 8 h of exposure in HAV

Table 1. Quantification data of HAV expressed as viral RNA copies/100 ml for each solar exposure condition along the study period (t) HAV

RNA copies/100 ml¶

t

SODIS

0h

4.00 (± 0.56) × 10

2h

Bath

Dark

Control

4.00 (± 0.56) × 10

4 A,a

4.00 (± 0.56) × 10

4.00 (± 0.56) × 104 A,a

1.27 (± 0.41) × 104 B,a

1.66 (± 1.74) × 104 AB,a

3.70 (± 3.77) × 104 A,a

3.64 (± 0.21) × 104 A,a

4h

3.27 (± 0.75) × 103 C,b

1.28 (± 3.10) × 104 AB,ab

3.18 (± 1.92) × 104 A,ab

4.00 (± 1.21) × 104 A,a

6h

2.65 (± 0.31) × 103 C,a

7.30 (± 1.78) × 103 BC,a

2.58 (± 6.23) × 104 A,a

3.97 (± 1.40) × 104 A,a

8h

3.15 (± 0.69) × 103 C,b

5.97 (± 0.11) × 103 C,b

2.53 (± 3.79) × 104 A,a

3.30 (± 0.43) × 104 A,a

%r*

92.1 (1.1 Log)

85.1 (0.83 Log)

36.7 (0.20 Log)

17.4 (0.08 Log)

4 A,a

4 A,a

All data correspond to the geometric mean of three replicates. Standard deviation is shown in parentheses. *Percentage (and log units) of total viral removal. Statistical differences among results of the different sampling times within the same treatment are indicated by capital letters. Statistical differences among results of the different treatments within the same sampling time are indicated by small letters. Results with the same letter did not show significant differences (P > 0.05). ¶

45

46

Int. Microbiol. Vol. 18, 2015

POLO ET AL.

Table 2. Quantification data of MNV-1 expressed as viral RNA copies/100 ml for each solar exposure condition along the study period (t) MNV-1

RNA copies/100 ml¶

t

SODIS

Bath

Dark

0h

5.91 (± 0.59) × 104 A,a

5.91 (± 0.59) × 10

5.91 (± 0.59) × 10

5.91 (± 0.59) × 104 A,a

2h

2.52 (± 3.20) × 104 AB,a

4.52 (± 0.72) × 104 AB,a

4.73 (± 0.60) × 104 AB,a

5.58 (± 0.71) × 104 A,a

4h

1.68 (± 3.01) × 104 AB,b

4.06 (± 0.40) × 104 B,ab

4.21 (± 0.43) × 104 B,ab

5.91 (± 0.21) × 104 A,a

6h

1.88 (± 2.07) × 104 AB,c

3.36 (± 0.50) × 104 BC,bc

4.85 (± 0.51) × 104 B,ab

5.85 (± 0.51) × 104 A,a

8h

9.24 (± 3.91) × 103 B,d

2.27 (± 3.92) × 104 C,c

3.70 (± 0.33) × 104 B,b

5.30 (± 0.41) × 104 A,a

% r*

84.4 (0.81 Log)

61.6 (0.42 Log)

37.4 (0.20 Log)

10.3 (0.05 Log)

4 A,a

Control 4 A,a

All data are the geometric mean of three replicates. Standard deviation is shown in paretheses. *Percentage (and log units) of total viral removal. Statistical differences among results of the different periods within the same treatment are indicated by capital letters. Statistical differences among results of the different treatments within the same sampling time are indicated by small letters. Results with the same letter did not show significant differences (P > 0.05). ¶

trials and after 8 h in MNV-1 trials. Bottles exposed to SODIS conditions showed a decrease in the infectivity capacity of their water samples after 4 h (infectivity loss of 33.4%), 6 h (66.7%) and 8 h (83.4%) in HAV trials, and after 6 h (33.4%) and 8 h (66.7%) in MNV-1 trials (Table 3). In addition, a delay in the appearance of the CPE from 2–3 days to 6 days in water samples from Bath-bottles (after 8 h of exposure) and SODISbottles (after 4, 6 and 8 h of exposure) was observed for MNV-1 (data not shown).

Discussion The aim of this study was to obtain a preliminary picture of the water disinfection for HAV and MNV-1 by SODIS. To our knowledge, this is the first study that evaluates and compares

by RT-qPCR the efficacy of this method to reduce and inactivate HAV and MNV-1 under natural solar conditions. In an attempt to establish the baseline of the impact of these condi-

tions on viral RNA and infectivity, the study was carried out using distilled water as an approach with as fewer variables as possible (i.e., organic matter that could interfere in the process). On the basis of the results obtained, future studies with natural waters (ground, river or tap water), which are likely to be used in developing countries, could be designed in order to optimize SODIS method for viral elimination.

The procedures employed involved the inclusion of reliable controls of RNA extraction and amplification steps. Results showed that extraction and RT-qPCR efficiencies did not showed important variations between HAV and MNV-1, making data consistent and suitable for quantification. SODIS is recommended to be practiced in regions with > 500 W × m–2 of global sunlight irradiance during 35 h [14]. Here, global sunlight irradiance values between 800 and 1000 W × m–2 were recorded during 3–4 h (data not shown). A strong synergistic effect has been observed between optical and thermal inactivation processes at water temperatures above 45ºC [33,45]. However, the maximum temperatures reached in SO-

Table 3. Infectivity assays carried out with HAV and MNV-1 HAV

MNV-1

Solar exposed for

Solar exposed for

Treatment

0h

2h

4h

6h

8h

0h

2h

4h

6h

8h

Control

100

100

100

100

100

100

100

100

100

100

Dark

100

100

100

100

100

100

100

100

Bath

100

100

66.6

66.6

100

100

83.3

66.6

SODIS

100

66.6

33.3

16.6

100

100

66.6

33.3

Results are expressed as bottles that showed infectivity/total bottles × 100.

SODIS OF HAV AND MURINE NOROVIRUS

DIS bottles were between 38 and 40ºC and at least 4 h were necessary to reach the temperatures between 35–40ºC from the beginning of the exposure. This is an important fact since, as it was also previously suggested [8], it may be a cause of the disparity between simulated and natural sunlight results.

Results showed a significant reduction in RNA levels after 8 h under SODIS conditions (1.1 log units for HAV and 0.81 log units for MNV-1) although final RNA counts remained relatively high (~103 RNAc/100 ml). In addition, infectivity assays reflected a decrease in the infectivity capacity of water samples from bottles in SODIS conditions 4 h after the beginning of the exposure (Table 3). Harding and Schwab [21] have reported 0.4 log and 1.4 log reduction in infectious MNV by plaque assay after a 2.5 and 6 hour of SODIS, respectively.

They have also reported better reductions in MS2 than MNV, suggesting that MNV would be highly resistant to damage by SODIS. Here, RNA removal rate and infectivity assays seem to indicate a higher resistance of MNV-1 than HAV, but without statistical differences.

The SODIS principle relies on the action of the solar UV radiation and the water temperature. The comparison between exposure conditions suggests a greater importance of sunlight radiation over the temperature as a principal factor of viral reduction. Previous studies have reported that high temperatures have a major effect on viral capsids proteins but limited effect on the viral genome [5,24,35]. In addition, a synergistic effect between heating and UV inactivation over 45ºC which leads to improved disinfection has been reported [33,45]. Although these temperatures never were reached in this study, better results were achieved with the combination of radiation and heat. Wegelin et al. [45] have reported similar findings with other viruses at comparable temperatures (<40°C).

Viral inactivation by heat relies basically in the loss of ability to bind with its host cell, by structural changes in the viral capsid proteins that disrupt the specific structures needed to recognize and bind the host cells [35,46]. Nevertheless, in this study, water samples from bottles maintained in darkness retained its infectivity capacity. Although high temperatures clearly denature capsids, the natural mode of indirect transmission of enteric viruses like HAV and NoV confers high stability in harsh environments outside the host’s body, including food and water at physiological temperatures as recorded here.

On the other hand, UV radiation seems to be crucial in this study for viral reduction. The mechanisms involved in viral inactivation may be either by direct UV damage on viral components, by indirect damage by reactive intermediates, such as ROS, or both [39]. The main components of non-enveloped

Int. Microbiol. Vol. 18, 2015

47

viruses (proteins and nucleic acids) do not absorb light at wavelengths > 320 nm, so direct damage mechanisms are conducted by UV-B light portion. The ROS mechanism, on the other hand, is initiated by UV-A light [27]. ROS mechanism was the major destroying viral capsid, as •OH have a high reactivity and the oxidative action alters membrane permeability reacting and oxidising capsid proteins, therefore diffusion of viral components to the medium occur, ending in viral inactivation.

It was reported that, although certain modifications in viral proteins can occur, the exposure to UV irradiation and ROS damages by 1O2 seems to be a more strong-genome damage component, since it transforms RNA itself by dimers or RNA−RNA and RNA–protein cross-links [46]. This suggests that RT-qPCR amplification loss by UV genome damages could be an appropriate proxy for SODIS evaluations in certain conditions, like low or middle temperatures, when damages to viral capsids are minimal. However, note that other genome damages outside the amplification regions can be underestimated. In this sense, long range RT-qPCR could be a useful solution [48].

From a viral perspective, subtle differences in viral genome and capsid composition affect disinfection kinetics and mechanisms between closely related viruses [38]. Differences in nucleic acid type (single- or double-stranded DNA or RNA), genome length and structure (longer genomes offer more targets for attack) and composition (% of adjacent pyrimidines or content of guanines, the most easily oxidized bases) could account for the variability in direct and indirect sunlight damages and inactivation rates [27,30,38]. HAV HM175 and MNV-1 have similar genome lengths (7478 and 7382 bases, respectively) and with a similar % of pyrimidine bases (51 and 48%, respectively) but with different base composition (HAV 32.9% U; 16.1% C; 29.3% A; 21.8% G and MNV-1 22.2% U; 28.9 C; 21.1% A; 27.8 G). How these differences in genomic composition could affect to the disinfection rates remains unclear and future research is needed in this sense.

With regard to the inactivation kinetic, for HAV and in lesser extent for MNV-1, a clear stabilization of the inactivating effect after 4 h is observed. This effect could not be related in a direct way to solar exposure since values of solar exposure at 6 h are still quite high (Fig. 1). Factors such as viral aggregation could affect the inactivation kinetics through subpopulations of non-easily-inactivated viruses. In this sense future experiments are needed with longer exposure times.

In summary, our results indicate that, under appropriate conditions, SODIS may be an effective and acceptable intervention against certain waterborne viruses including HAV

48

Int. Microbiol. Vol. 18, 2015

and NoV. Viral disinfection rates are relatively lower than bacterial disinfection rates reported in other studies, and the requireder exposure periods are longer. Although the drinking water requirements for a total protection against viral illness are not completely fulfilled, our results point out that SODIS could contribute to reduce the risk of viral infection, supporting its use as an emergency intervention for vulnerable communities. Viral inactivation of non-enveloped viruses is scarce and this study provides new comparative data on HAV and MNV-1 RNA damage and infecti­vity after sunlight exposure.

Further research is required to determine and validate the efficacy and limits of SODIS to eliminate HAV and MNV-1 under different conditions, including in natural waters.

Acknowledgements. This work was supported in part by Grants of the Ministerio de Ciencia e Innovación (Spain) for financing stays at the PSA (by DP and JLR). Authors thank the European Commission funds under SFERA program (Solar Facilities for the European Research Area, EC Grant agreement no. 228296).

Competing interests. None declared.

References 1. Acra A, Jurdi M, Mu’Allem H, Karahagopian Y, Raffoul Z (1989) Sunlight as disinfectant. Lancet 1:280 2. Agulló-Barceló M, Polo-López MI, Lucena F, Jofre J, Fernández-Ibáñez P (2013) Solar advanced oxidation processes as disinfection tertiary treatments for real wastewater: implications for water reclamation. Appl Catal B: Environ 136-137:341-350 3. Atmar RL, Estes MK (2006) The epidemiologic and clinical importance of norovirus infection. Gastroenterol Clin North Am 35:275-290 4. Bae J, Schwab KJ (2008) Evaluation of Murine Norovirus, feline Calicivirus, poliovirus, and MS2 as surrogates for human norovirus in a model of viral persistence in surface water and groundwater. Appl Environ Microbiol 74:477-484.

5. Baert L, Wobus CE, Van Coillie E, Thackray LB, Debevere J, Uyttendaele M (2008) Detection of murine norovirus 1 by using plaque assay, transfection assay, and real-time reverse transcription-PCR before and after heat exposure. Appl Environ Microbiol 74:543-546 6. Bosch A (1998) Human enteric viruses in the water environment: A minireview. Int Microbiol 1:191-196 7. Boschi-Pinto C, Velebit L, Shibuya K (2008) Estimating child mortality due to diarrhoea in developing countries. Bull World Health Organ 86:710-717 8. Boyle M, Sichel C, Fernández-Ibáñez P, Arias-Quiroz GB, Iriarte-Puña M, Ubomba-Jaswa E, McGuigan KG (2008) Bactericidal effect of solar water disinfection under real sunlight conditions. Appl Environ Microbiol 74:2997-3001 9. Clasen T, Cairncross S, Haller L, Bartram J, Walker D (2007) Cost-effectiveness of water quality interventions for preventing diarrhoeal disease in developing countries. J Water Health 5:599-608 10. Conroy RM, Elmore-Meegan M, Joyce T, McGuigan KG, Barnes J (1999) Solar disinfection of water reduces diarrhoeal disease: An update.

Arch Dis Child 81:337-338

POLO ET AL.

11. Costafreda MI, Bosch A, Pintó RM (2006) Development, evaluation, and standardization of a real-time TaqMan reverse transcription-PCR assay for quantification of hepatitis A virus in clinical and shellfish samples.

Appl Environ Microbiol 72:3846-3855 12. Cromeans T, Sobsey MD, Fields, HA (1987) Development of a plaque assay for a cytopathic, rapidly replicating isolate of hepatitis A. J Med Virol 2:45-56 13. du Preez M, Conroy RM, Ligondo S, Hennessy J, Elmore-Meegan M, Soita A, McGuigan KG (2011) Solar disinfection of drinking water (SODIS) in the prevention of dysentery in Kenyan children aged under 5 years. Environ Sci Technol 45:9315-9323 14. EAWAG (2008) Solar water disinfection: the method. http://www.sodis. ch/Text2002/T-TheMethod.html 15. Eisenstark A (1987) Mutagenic and lethal effects of near-ultraviolet radiation (290-400nm) on bacteria and phage. Environ Mol Mutagen 10:317-337 16. Fernández-Ibáñez P, Blanco J, Sichel C, Malato S (2009) Water disinfection by solar photocatalysis using compound parabolic collectors. Catal Today 101:345-352 17. Fewtrell L, Bartram J (2001) Water quality: guidelines, standards and health. Assessment of risk and risk management for water-related infectious disease. IWA Publishing, London 18. Fewtrell L, Kaufmann R, Kay D, Enanoria W, Haller L, Colford J (2005) Water, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis. Lancet Infect Dis 5:42-52 19. Gómez-Couso H, Fontán-Saínz M, Sichel C, Fernández-Ibáñez P, AresMazás E (2009) Efficacy of the solar water disinfection method in turbid waters experimentally contaminated with Cryptosporidium parvum oocysts under real field conditions. Trop Med Int Health 14:620-627 20. Goodsell DS (2001) The molecular perspective: ultraviolet light and pyrimidine dimmers. Oncologist 6:298-299 21. Harding AS, Schwab KJ (2012) Using limes and synthetic psoralens to enhance solar disinfection of water (SODIS): a laboratory evaluation with norovirus, Escherichia coli, and MS2. Am J Trop Med Hyg 86:566-572 22. Heaselgrave W, Kilvington S (2012) The efficacy of simulated solar disinfection (SODIS) against coxsackievirus, poliovirus and hepatitis A virus. J Water Health 10:531-538 23. Heaselgrave W, Patel N, Kehoe SC, Kilvington S, McGuigan KG (2006) Solar disinfection of poliovirus and Acanthamoeba polyphaga cysts in water – a laboratory study using simulated sunlight. Lett Appl Microbiol 43:125-130 24. Hewitt J, Rivera-Aban M, Greening GE (2009) Evaluation of murine norovirus as a surrogate for human norovirus and hepatitis A virus in heat inactivation studies. J Appl Microbiol 107:65-71 25. Hurst CJ, Goyke T (1983) Reduction of interfering cytotoxicity associated with wastewater sludge concentrates assayed for indigenous enteric viruses. Appl Environ Microbiol 46:133-139 26. Jagger J (1981) Near-UV radiation effects on microorganisms. Photochem Photobiol 34:761-768 27. Kawanishi S, Hiraku Y, Oikawa S (2001) Mechanism of guanine-specific DNA damage by oxidative stress and its role in carcinogenesis and aging. Mutat Res 488:65-76 28. Lonnen J, Kilvington S, Kehoe SC, Al-Touati F, McGuigan KG (2005) Solar and photocatalytic disinfection of protozoan, fungal and bacterial microbes in drinking water. Water Res 39:877-883 29. Love DC, Silverman A, Nelson, KL (2010) Human virus and bacteriophage inactivation in clear water by simulated sunlight compared to bacteriophage inactivation at a Southern California beach. Environ Sci Technol 44:6965-6970

SODIS OF HAV AND MURINE NOROVIRUS

30. Lytle CD, Sagripanti JL (2005) Predicted inactivation of viruses of relevance to biodefense by solar radiation. J Virol 79:14244-14252 31. Malato S, Fernández-Ibáñez P, Maldonado MI, Blancom J, Gernjak W (2009) Decontamination and disinfection of water by solar photocatalysis: recent overview and trends. Catal Today 147:1-59 32. McGuigan KG, Conroy RM, Mosler HJ, du Preez M, Ubomba-Jaswa E, Fernandez-Ibañez P (2012) Solar water disinfection (SODIS): A review from bench-top to roof-top. J Hazard Mater 235-236:29-46 33. McGuigan KG, Joyce TM, Conroy RM, Gillespie JB, Elmore-Meegan M (1998) Solar disinfection of drinking water contained in transparent plastic bottles: characterizing the bacterial inactivation process. J Appl Microbiol 84:1138-1148 34. Moore MD, Goulter RM, Jaykus LA (2015) Human norovirus as a food pathogen: challenges and developments. Annu Rev Food Sci Technol 6:411-433 35. Nuanualsuwan S, Cliver DO (2003) Capsid functions of inactivated human picornaviruses and feline caliciviruses. Appl Environ Microbiol 69:350-357 36. Reed RH, Mani SK, Meyer V (2000) Solar photo-oxidative disinfection of drinking water: preliminary field observations. Lett Appl Microbiol 30:432-436 37. Sichel C, Tello J, de Cara M, Fernandez-Ibañez P (2007) Effect of UV solar intensity and dose on the photocatalytic disinfection of bacteria and fungi. Catal Today 129:152-160 38. Sigstam T, Gannon G, Cascella M, Pecson BM, Wigginton KR, Kohn T (2013) Subtle differences in virus composition affect disinfection kinetics and mechanisms. Appl Environ Microbiol 79:3455-3467 39. Silverman AI, Peterson BM, Boehm AB, McNeill K, Nelson KL (2013) Sunlight inactivation of human viruses and bacteriophages in coastal waters containing natural photosensitizers. Environ Sci Technol 47:1870-1878

Int. Microbiol. Vol. 18, 2015

49

40. Sinha RP, Hader DP (2002) UV-induced DNA damage and repair: a review. Photochem Photobiol Sci 1:225-236 41. Sommer B, Mariño A, Solarte Y, Salas ML, Dierolf C, Valiente C, Mora D, Rechsteiner R, Setter P, Wirojanagud W, Ajarmeh H, Al-Hassan A, Wegelin M (1997) SODIS - an emerging water treatment process. J Water SRT – Aqua 46:127-137 42. United Nations (UN) (2013) The Millennium Development Goals Report. United Nations, New York 43. Vaughan G, Rossi LMG, Forbi JC, de Paula VS, Purdy MA, Xia G, Khudyakov YE (2014) Hepatitis A virus: host interactions, molecular epidemiology and evolution. Infect Genet Evol 21:227-243 44. Verhoef L, Hewitt J, Barclay L, Ahmed SM, Lake R, Hall AJ, Lopman B, Kroneman A, Vennema H, Vinjé J, Koopmans M (2015) Norovirus genotype profiles associated with foodborne transmission, 1999-2012.

Emerg Infect Dis 21:592-599 45. Wegelin M, Canonica S, Mechsner K, Fleischmann T, Pesaro F, Metzler A (1994) Solar water disinfection: scope of the process and analysis of radiation experiments. J Water SRT – Aqua 43:154-169 46. Wigginton KR, Pecson BM, Sigstam T, Bosshard F, Khon T (2012) Virus inactivation mechanisms: impact of disinfectants on virus function and structural integrity. Environ Sci Technol 46:12069-12078 47. Wobus CE, Thackray LB, Virgin HW (2006) Murine norovirus: a model system to study norovirus biology and pathogenesis. J Virol 80:5104-5112 48. Wolf S, Rivera-Aban M, Greening GE (2009) Long-range reverse transcription as a useful tool to asses the genomic integrity of Norovirus.

Food Environ Virol 1:129-136

RESEARCH ARTICLE International Microbiology (2015) 18:51-59 doi:10.2436/20.1501.01.234. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Increasing antibiotic resistance in preservative-tolerant bacterial strains isolated from cosmetic products Pilar Orús,1 Laura Gomez-Perez,2,3 Sonia Leranoz,4 Mercedes Berlanga2* Beautyge-Revlon S.L., Barcelona, Spain. 2Department of Microbiology and Parasitology, Faculty of Pharmacy, University of Barcelona, Barcelona, Spain. 3School of Biological Sciences, University of East Anglia, Norwich, UK. 4 Supply Quality-Reckitt Benkiser Granollers S.L., Spain 1

Received 17 December 2014 · Accepted 22 February 2015

Summary. To ensure the microbiological quality, consumer safety and organoleptic properties of cosmetic products, manufacturers need to comply with defined standards using several preservatives and disinfectants. A drawback regarding the use of these preservatives is the possibility of generating cross-insusceptibility to other disinfectants or preservatives, as well as cross resistance to antibiotics. Therefore, the objective of this study was to understand the adaptive mechanisms of Enterobacter gergoviae, Pseudomonas putida and Burkholderia cepacia that are involved in recurrent contamination in cosmetic products containing preservatives. Diminished susceptibility to formaldehyde-donors was detected in isolates but not to other preservatives commonly used in the cosmetics industry, although increasing resistance to different antibiotics (β-lactams, quinolones, rifampicin, and tetracycline) was demonstrated in these strains when compared with the wild-type strain. The outer membrane protein modifications and efflux mechanism activities responsible for the resistance trait were evaluated. The development of antibiotic-resistant microorganisms due to the selective pressure from preservatives included in cosmetic products could be a risk for the emergence and spread of bacterial resistance in the environment. Nevertheless, the large contribution of disinfection and preservation cannot be denied in cosmetic products. [Int Microbiol 2015; 18(1):51-59] Keywords: Enterobacter · Pseudomonas · Burkholderia · cosmetic preservatives · antibiotics · cross-resistance

Introduction Microbial contamination of cosmetic products is a matter of great importance to the industry and it is potentially a major cause of both product and economic losses. Water and nutriCorresponding author: M. Berlanga Department of Microbiology and Parasitology Faculty of Pharmacy, University of Barcelona Av. Joan XXIII, s/n 08028 Barcelona, Spain Tel. +34-934024497. Fax +34-934024498 E-mail: mberlanga@ub.edu *

ents present in cosmetics make them susceptible to microbial growth. Most often, microorganisms are the cause of organoleptic alterations, such as offensive odors, and changes in viscosity and color. Moreover, in a few cases, contaminating microorganisms or their activity may cause human health problems, such as skin irritation, allergic contact dermatitis and infection, especially in the eyes, mouth or wounds [10,25,33,35]. To ensure microbiological quality, consumer safety and the organoleptic properties of cosmetic products, manufacturers need to use disinfectants and preservatives.

Therefore, cosmetics need preservatives that are able to reduce the microbial load to acceptable levels during the period

52

Int. Microbiol. Vol. 18, 2015

of validity [17]. Regulations in the EU and in other countries permit specified preservatives, define their maximum concentrations, and provide other controls specifically related to the type of cosmetic product. Preservatives are added to cosmetics in order to inhibit the proliferation of spoiling microorganisms. Microbial quality assurance in cosmetics aims to prevent the transmission of diseases by using these products properly, and ensures their stability and effectiveness [28].

Antibiotics are generally considered to be selectively toxic agents suitable for administration to patients, whereas biocides have been traditionally regarded as antiseptics, disinfectants or preservatives. Additionally, antibiotics are considered to have a specific target site within a bacterial cell, whereas biocides have multiple target sites. Unlike antibiotics, which are selectively toxic, most biocides do not act against a defined target cell [22,23,32,36]. For example, triclosan inhibits the synthesis of agents that bind to enoyl-acyl carrier protein reductase, causing inhibition of fatty acid biosynthesis and disruption of the membrane, thus precipitating the cytoplasmic compounds [13]. Formaldehyde-donors/releasers act against bacterial cells by releasing formaldehyde into the medium, and their biocidal effect is due to the cross-linking of proteins in the cell envelope and elsewhere in the cell, as well as cross-linking of RNA and DNA [34].

Exposure of bacteria to biocides can select for mutants with decreased biocide susceptibility, and these mutants often display a decreased susceptibility to various antibiotics, indicating that biocides can act as drivers of antibiotic resistance under laboratory conditions [7,37]. It has been suggested that cross-resistance to antimicrobial compounds, following exposure and adaptation to a biocide, could occur in a limited number of situations, such as when a biocide and antibiotic compound use the same entry mechanisms, have the same cellular target, can develop the same resistance mechanism and, finally, when biocide tolerance and antibiotic resistance are potentially carried by the same mobile genetic element [7]. Resistance to antimicrobial compounds can emerge following one or more target gene mutations, but it is difficult for bacteria to become resistant to the recommended concentrations used for many biocides, since mutations within a single gene will not usually confer resistance. In contrast, when tolerance to biocides arises, it is mediated by mechanisms that are less well characterized.

Some of the modifications that can occur in a bacterial cell include upregulation of the efflux pump activity or structural alterations in the cell wall, which impact permeability.

Efflux systems can export both antibiotics and many biocides [36].

ORÚS ET AL.

The extensive use of microbicides in a wide range of applications has been questioned with regard to their role in the development of bacterial resistance to antimicrobials. However, for natural isolates, there was no evidence that cross-resistance between cosmetic preservatives and antibiotics could occur [30]. The aim of this study was to determine the mechanisms of tolerance to formaldehyde-donors shown by bacterial strains belonging to species of Enterobacter gergoviae, Pseudomonas putida and Burkholderia cepacia isolated from cosmetics products, and, because of human health concerns, to search for possible cross-resistance to antibiotics.

Materials and methods Bacterial strains. A total of 46 strains belonging to Enterobacter gergo­ viae, 22 to Pseudomonas putida and 44 to Burkholderia cepacia were obtained from contaminated cosmetic formulations (shampoos, lotions, conditioners) preserved with parabens and formaldehyde-donors. Strains were identified by the API system 20NE (Biomeriéux, Marcy l’Etoile, France).

Reference strains from the American Type Culture Collection (ATCC) were also included in the study. Strains were cultured on trypticase soy agar (Difco, Detroit, MI, USA) for E. gergoviae and P. putida, and on plate count agar (Difco) for B. cepacia.

Minimal inhibitory concentration (MIC) in solid medium.

Susceptibility tests to 16 preservatives and 12 antibiotics were performed by serial dilutions in Mueller-Hinton agar (Difco). The inoculum consisted of 104 colony forming units (CFU) per spot and the MIC was defined as the lowest concentration that prevented visible growth after incubation for 18 h at 35ºC. The following preservatives were studied: methylchloroisothiazolinone/methylisothiazolinone and polyaminopropyl biguanide (Thor, Speyer, Germany), DMDM hydantoin (McIntyre, Halifax, Canada), Quaternium-15 (Evonik Degussa Ibérica, Barcelona, Spain), diazolidinyl urea, imidazolidinyl urea and sodium hydroxymethylglycinate (International Specialty Products, West Milford, NJ, USA), methylparaben and propylparaben (Sharon Labs., Ashdod, Israel), methyldibromo glutaronitrile (Shülke, Nordersted, Germany), phenoxyethanol (Seppic, Paris, France), hexamidine isethionate (Laboratoires Sérobiologiques, Pulnoy, France), chlorphenesin (Arnaud, Paris, France), benzalkonium chloride (Comercial Química Massó, Barcelona, Spain), bronopol (Basf, Ludwigshafen, Germany), and triclosan (Ciba Specialty Chemicals, Basel, Switzerland). Antibiotics were purchased from Sigma (Madrid, Spain). Working solutions were prepared daily in suitable sterile diluents. Antibiotics used were: cefotaxime, ceftazidime, ceftriaxone, kanamycin, streptomycin, tetracycline, erythromycin, ciprofloxacin, penicillin, ampicillin, chloramphenicol, and novobiocin.

Outer membrane isolation and SDS-PAGE. Outer membrane proteins were obtained from cells cultivated in Luria Bertani broth (Difco), as described previously [20]. The inner membrane fraction was solubilized from disrupted cells by direct extraction with 2% sodium dodecyl sulfate and the outer membrane was separated by centrifugation (12,000 rpm, 60 min). The proteins obtained were analyzed by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (SDS-PAGE). In order to enhance the resolution of protein bands, 4 M urea was added to the resolving gel.

ANTIBIOTIC CROSS-RESISTANCE IN COSMETIC PRESERVATIVE

Microbial affinity to solvents (MATS) for selected strains.

The MATS assay was carried out as previously described [11]. After overnight culture, cells were harvested by centrifugation (7500 rpm, 10 min), washed twice with PBS (phosphate buffered saline) at pH 7.0, and resuspended in the same solution at a final optical density (OD600) of 0.8. Each bacterial suspension (2.4 ml) was mixed for 60 s at maximum intensity on a vortex-type agitator with 0.4 ml of chloroform, hexadecane, diethyl ether, or hexane (Panreac, Barcelona, Spain). The mixtures were allowed to stand for 60 min to ensure complete separation of the two phases. A 1-ml sample was then carefully removed from the aqueous phase and its optical density was measured at 600 nm. The microbial affinity for each solvent was calculated using the formula: % Affinity = (OD0 − ODf / OD0) × 100 where OD0 is the optical density of the bacterial suspension before mixing with the solvent and ODf the absorbance after mixing and phase separation.

Each measurement was performed in duplicate and the experiment was repeated three times with independent bacterial cultures. Two solvent pairs were assayed: (i) chloroform (an acidic solvent) and hexadecane (an apolar solvent), and (ii) diethyl ether (a strong basic solvent) and hexane (an apolar solvent).

Ciprofloxacin accumulation and efflux assays. Ciprofloxacin accumulation and efflux assays were performed by a previously described fluorimetric method [3]. Briefly, isolates were incubated at 37°C until OD600 = 0.5–0.7. Bacteria were harvested by centrifugation (7000 rpm) at room temperature, washed, and concentrated 10-fold in PBS pH 7.5. Ciprofloxacin was added to a final concentration of 10 µg/ ml. At timed intervals of 30 s, 1.5 min, 3 min, 6 min, and 9 min, samples were centrifuged in a microfuge at 10,000 rpm at 4°C for 15 s. Pellets were washed in 1 ml of chilled PBS at pH 7.5, and suspended in 1 ml 0.1 M glycine–HCl buffer at pH 3.0, and finally incubated at room temperature overnight to allow bacterial lysis. The suspensions were centrifuged at 20°C for 25 min in order to remove bacterial debris.

The concentration of the antibiotic in the supernatants was determined fluorometrically using an SLM Aminco 8100 spectrofluorometer. For the efflux assay, cells were incubated for 3 min with antibiotic before addition of the metabolic inhibitor CCCP (carbonyl cyanide m-chlorophenylhydrazone) at a final concentration of 100 µM, and samples were manipulated in the same way as for quinolone accumulation. The specific extinction and emission wavelengths used to identify ciprofloxacin were 279 and 447, respectively, and they were determined in 0.1 M glycine-HCl at pH 3.0.

Results Susceptibility of natural preservative-tolerant strains isolated from cosmetic products. The MICs of different preservatives commonly used in the cosmetics industry were determined for Enterobacter gergoviae, Pseudomonas putida and Burkholderia cepacia strains. Results are shown in Table 1. Tolerance to preservatives was defined as more than eight-fold the MIC compared to susceptible strains [24]. Regarding E. gergoviae, some strains showed tolerance to formaldehyde-donors, such as diazolidinyl urea, imidazolidinyl urea, Quaternium-15, sodium hydroxymethylglycinate and DMDM hydantoin. No significant changes in

Int. Microbiol. Vol. 18, 2015

53

susceptibility were found when other preservatives were tested, such as parabens, isothiazolinone, bronopol, biguanide, methyldibromo glutaronitrile, chlorphenesin, phenoxyethanol, benzalkonium chloride, triclosan or hexamidine. On the other hand, the strains with decreased susceptibility to formaldehyde-donors showed reduced susceptibility to β-lactam antibiotics (penicillin and cephalosporin), tetracycline and ciprofloxacin when compared to the preservative-susceptible Enterobacter strains. No increased resistance to novobiocin, macrolides (erythromycin), aminoglycosides (kanamycin, streptomycin) or chloramphenicol was found (Table 2). Similar results were observed for P. putida and B. cepacia (Table 1).

Regarding antibiotics, reduced susceptibility was found in nearly all the drugs tested (Table 2).

Surface characterization and permeability. To exert an antibacterial action, antimicrobial (preservatives/antibiotics) must penetrate the cell envelope or accumulate therein at a sufficiently high concentration. Adaptation of the microbial cell envelope may contribute to the mechanism responsible for resistance to antimicrobial agents. As bacterial species differ in their envelope structures and, hence, in their intrinsic resistance to antibiotics [4,6,19,39], we studied outer membrane proteins, physicochemical characteristics of the cell surface and permeability that may influence cell susceptibility to preservatives/antibiotics in three selected tolerant strains (E. gergoviae EU36, P. putida EU34 and B. cepacia EU67).

The initial reaction of a biocide with a microbial cell involves initial binding to the cell surface, although target sites might be found within the cell. The first reaction of any antibacterial agent involves interaction with the outer cell membrane in the case of Gram-negative bacteria and subsequent passage of the agent to the target site. Outer membrane proteins from Enterobacter contain three porins, named OmpF, OmpC and OmpD [19,20]. E. gergoviae EU36 did not express the OmpF porin. Therefore, it seems that formaldehyde-donors must enter into the bacterial cell through this porin, since the cell became tolerant when OmpF was not expressed in the outer membrane (Fig. 1). It has been described that hydrophilic molecules pass through the outer membrane by porins [19,26]. Accordingly, porin-deficient strains did not show either tolerance to hydrophobic preservatives (parabens, chlorphenesin, triclosan), or resistance to hydrophobic and highermolecular weight antibiotics, such as erythromycin. Pseudo­ monas putida EU34 and B. cepacia EU67 were tolerant to preservatives and did not show any changes in porin composition with respect to the susceptible strains (Fig. 1). Therefore,

54

Int. Microbiol. Vol. 18, 2015

ORÚS ET AL.

Table 1. Minimal inhitory concenctratio (MIC) for the three species tested for preservatives. Results are expressed in µg/ml MIC Enterobacter gergoviae Preservative DMDM hydantoin

No. strainsa

MIC90b

Pseudomonas putida No. strains

MIC90

Burkholderia cepacia No. strains

MIC90

S (26)

1500

S (4)

412.5

S (16)

1650

R (20)

6000

R (18)

3850

R (28)

16500

Quaternium-15

S (26)

600

S (4)

300

S (16)

300

R (20)

2500

R (18)

7500

R (28)

2500

Imidazolidinyl urea

S (26)

1500

S (4)

5000

S (16)

300

R (20)

>7500

R (18)

17500

R (28)

6000

Diazolidinyl urea

S (26)

300

S (4)

300

S (16)

600

R (20)

5000

R (18)

2500

R (28)

12500

Sodium hydroxymethylglycinate Methylparaben

S (26)

625

S (4)

625

S (16)

1250

R (20)

1250

R (18)

3500

R (28)

2500

S (46)

2500

S (22)

2500

S (44)

2500

R (0)

–

R (0)

–

R (0)

–

Propylparaben

S (46)

500

S (22)

500

S (44)

500

R (0)

–

R (0)

–

R (0)

–

MCI/MIc

S (46)

15

S (22)

10

S (44)

12.5

R (0)

–

R (0)

–

R (0)

–

Methyldibromo glutaronitrile Phenoxyethanol Hexamidine isethionate Chlorphenesin Benzalkonium chloride Bronopol Polyaminopropyl biguanide Triclosan

S (46)

120

S (22)

100

S (44)

120

R (0)

–

R (0)

–

R (0)

–

S (46)

5000

S (22)

5000

S (44)

2500

R (0)

–

R (0)

–

R (0)

–

S (46)

12.5

S (22)

25

S (44)

300

R (0)

–

R (0)

–

R (0)

–

S (46)

3000

S ()

3000

S (44)

1500

R (0)

–

R (0)

–

R (0)

–

S (46)

125

S (22)

250

S (44)

100

R (0)

–

R (0)

–

R (0)

–

S (46)

15

S (22)

30

S (44)

25

R (0)

–

R (0)

–

R (0)

–

S (46)

25

S (22)

2500

S (44)

180

R (0)

–

R (0)

–

R (0)

–

S (46)

<2.5

S (22)

50

S (44)

5

R (0)

–

R (0)

–

R (0)

–

S: susceptible strains; R: resistant strains. b MIC90: MIC at which 90% of the isolates are inhibited. c MCI/MI: methylchloroisothiazolinone/methylisothiazolinone. a

Pseudomonas and Burkholderia must have another mechanism implicated in tolerance to formaldehyde-donor preservatives, as well as for increasing antibiotic resistance (this pattern of porins was also observed in all the preservative-tolerant strains in this study).

The charge and cell surface hydrophobicity may influence the interaction with antimicrobial agents [2]. Cationic compounds, such as chlorhexidine and benzalkonium chloride, are thought to interact with negative charges in the bacterial cell wall and outer membrane [12]. The Lewis acid/base and

ANTIBIOTIC CROSS-RESISTANCE IN COSMETIC PRESERVATIVE

Int. Microbiol. Vol. 18, 2015

55

Table 2. MIC results of antibiotics for the three species tested. Results are expressed in 袖g/ml MIC Enterobacter gergoviae Antibiotic Cefotaxime Ceftazidime Ceftriaxone Kanamycin Streptomycin Tetracycline Erythromycin Ciprofloxacin Penicillin Ampicillin Chloramphenicol Novobiocin

Strainsa

MIC90b

Pseudomonas putida Strains

MIC90

Burkholderia cepacia Strains

MIC90

S

0.075

S

6

S

5

R

0.30

R

15

R

20

S

0.0125

S

0.5

S

1

R

0.5

R

2.75

R

3.5

S

0.025

S

2.75

S

7.5

R

0.15

R

12.5

R

17.5

S

0.3

S

0.25

S

15

R

0.6

R

0.75

R

75

S

0.5

S

2

S

10

R

0.5

R

6.5

R

100

S

1.25

S

2

S

35

R

4

R

8

R

35

S

10

S

10

S

70

R

10

R

100

R

125

S

0.0025

S

0.0125

S

0.125

R

0.01

R

0.05

R

1

S

20

S

180

S

100

R

50

R

250

R

550

S

6

S

100

S

350

R

12.5

R

100

R

850

S

10

S

75

S

17.5

R

25

R

125

R

125

S

350

S

700

S

3.1

R

350

R

700

R

6.2

S/R: susceptible/resistant to formaldehyde donors (DMDM hydantoin, Quarternium-15, imidazolidinyl urea, diazolidinyl urea, sodium hydroxymethylglycinate), respectively. For E. gergoviae, n (S) = 26, n (R) = 20.

For P. putida, n (S) = 4, n (R ) = 18 and for B. cepacia n (S) = 16, n (R) = 28. b MIC90: MIC at which 90% of the isolates are inhibited. a

hydrophobicity of cell envelopes can be assessed by microbial affinity to solvents (MATS) [8]. Monopolar solvents (i.e., chloroform and diethyl ether) were selected for the estimation of the Lewis acid/base character, and apolar solvents (hexane and hexadecane) were used to estimate the hydrophobicity properties. B. cepacia EU67 showed the most affinity to chloroform (Lewis acid character or electron acceptor). Affinity for chloroform implies an increased number of protonated groups, such as NH3 and/or OH groups, on the bacterial surface [15]. Affinity to diethyl ether (Lewis base or electron donor) was observed in E. gergoviae EU36. Pseudomonas puti足 da EU34 had a moderate affinity to both chloroform (Lewis acid) and diethyl ether (Lewis base). In general, all strains

showed a hydrophilic character due to the low affinity for apolar solvents such as hexadecane and hexane. However, Burkholderia displayed slowly increasing hydrophobicity compared to Enterobacter and Pseudomonas (Table 3).

Figure 2 shows the results of experiments comparing ci足 pro足floxacin accumulation with and without inhibition by CCCP. After 6 min of bacterial exposure to ciprofloxacin, a species-specific steady-state concentration was achieved, confirming previously published data [3]. Ciprofloxacin accumulation in E. gergoviae EU36 was lower than in the susceptible type strain ATTCC 12833 because of deficiency of the OmpF porin. Denergized cells with CCCP had similar steady-state accumulation. A slower entry of ciprofloxacin and a similarly

Int. Microbiol. Vol. 18, 2015

ORÚS ET AL.

Int Microbiol

56

efficient efflux pump compared to the susceptible type strain may be sufficient to explain the resistance observed in E. ger­ goviae EU36. Pseudomonas putida EU34 accumulated slightly less than E. gergoviae EU36, but denergizated cells that accumulated ciprofloxacin had similar values to denergized En­ terobacter cells. In this case, the low entry for Pseudomonas and effective efflux pumps could explain the relative resistance observed, since the susceptible Pseudomonas type strain had a similar behavior to its preservative-tolerant counterpart. Burk­ holderia strains accumulated less ciprofloxacin than Pseudo­ monas, especially B. cepacia EU67, but Burkholderia had a very effective efflux system that was higher than those observed in Pseudomonas and Enterobacter. Burkholderia preservative-tolerant strains had a dramatically low entry, but no porin suppression was observed. Probably, surface characteristics, such as hydrophobicity and the Lewis acid character, may have influenced the uptake of ciprofloxacin.

Fig. 1. SDS-PAGE of outer membrane proteins from Enterobacter gergoviae (A), Pseudomonas putida (B), and Burk­ holderia cepacia (C).

Discussion Preservatives are added to cosmetic formulations in order to inhibit the growth of microorganisms. They are normally used at high concentrations, which are rapidly bactericidal, but, in some circumstances, sublethal concentrations can occur in the products and this exerts selective pressure on the bacteria leading to reduced susceptibility to these preservatives.

It has been proposed that antibiotics/biocides can penetrate the envelopes of Gram-negative bacteria by three routes: (i) the hydrophilic pathway, through water-filled porin channels; (ii) the hydrophobic pathway, through the lipid bilayer; and (iii) self-promoted, which involves the displacement of divalent cations that bridge adjacent lipopolysaccharide (LPS) molecules, thereby disrupting the structure of the outer membrane and exposing areas of the phospholipid bilayer [27].

Table 3. Lewis acid-base and hydrophobicity surface characteristics Microbial affinity to solvents (MATS) Strains

Chloroform

Hexadecane

Diethyl ether

Hexane

Enterobacter gergoviae (EU36)

33.80 ± 0.7

7.17 ± 0.36

50.54 ± 3.11

13.89 ± 0.77

Pseudomonas putida (EU34)

68.77 ± 7.62

14.14 ± 5.6

66.38 ± 2.6

33.24 ± 3.86

Burkholderia cepacia (EU67)

86.60 ± 6.74

29.89 ± 3.59

69.62 ± 6.62

51.24 ± 5.64

Int. Microbiol. Vol. 18, 2015

57

Int Microbiol

ANTIBIOTIC CROSS-RESISTANCE IN COSMETIC PRESERVATIVE

Fig. 2. Steady-state accumulation of ciprofloxacin (cip) after 6 min exposure. The height of each column represents the mean of the results of three independent experiments. The standard deviations are indicated by the bars. Eg, Enterobacter gergoviae; Pp, Pseudomonas putida; Bc, Burkholderia cepacia.

Permeability (reduced entry and/or an overexpressed efflux system) may be considered as a common and basic mechanism of resistance, which is perhaps even more frequent than target modification or production of antibiotic-inactivating enzymes.

Multidrug efflux pumps, especially those belonging to the resistance-nodulation-division family, play a major role in establishing the “intrinsic or acquired” resistance of Gram-negative bacteria to a wide range of toxic compounds, including antibiotics [16,31]. Efflux through RND-family pumps (e.g.

AcrAB, MexAB–OprM) is driven by the proton motive force, an electrochemical gradient in which the movement of hydrogen ions drives the transport of the substrate that can be inhibited by CCCP. AcrAB-TolC multidrug resistance confers resistance to a wide variety of lipophilic and amphiphilic compounds, such as dyes, detergents, and antimicrobial agents (ethidium bromide, crystal violet, sodium dodecyl sulfate, bile acids, tetracycline, chloramphenicol, fluoroquinolones, β-lac­ tams, erythromycin) [31]. The presence of similar systems has been reported for Enterobacteriaceae, and a strain of Entero­ bacter cloacae overexpressing the AcrAB system exhibited a reduction of porin gene expression [29]. A similar mechanism was probably responsible for the natural E. gergoviae isolates

that were tolerant to preservatives and had increasing resistance to several antibiotics used in this study. In previous studies using a series of E. gergoviae derivatives isolated with increasing methylisothiazolinone–chloromethylisothiazolinone (MIT-CMIT) and triclosan concentrations, antibiotic susceptibility has not been altered and a different mechanism of tolerance has been described [30]. Note that, no alteration in porin expression was observed in Pseudomonas nor in Burkholde­ ria; overexpression of the efflux system probably belongs to the RND (resistance-nodulation-cell division) family of transporters [38]. Most of the efflux systems characterized in pseudomonads export both antibiotics and biocides. Resistance to formaldehyde-donors both in pseudomonads and enterobacteria has been described by the action of formaldehyde-dehydrogenases [18,21]; however, no specific assays were carried out to determine this enzymatic activity, since a large difference in the MIC of formaldehyde would be expected.

To date, there have been several reports of cross-resistance between antibiotics and disinfectants used in the food industry and hospital environment [5,7,9,39], but, as far as we know, there has been no evidence that this could be possible for preservatives in cosmetics, except for laboratory selected bacterial strains [30]. In our work, increasing resistance to an-

58

Int. Microbiol. Vol. 18, 2015

tibiotics in natural preservative-tolerant isolates was more than eight-fold the MIC with respect to susceptible strains. In some cases, such as ciprofloxacin, this increasing resistance was insufficient to encourage clinical resistance because the serum concentration of ciprofloxacin was still high. However, surviving cells would have the potential to mutate spontaneously to higher, more clinically relevant levels of resistance to quinolones (such as ciprofloxacin) or to other antibiotics [1,14]. Antibacterial agents and antibiotics share the same resistance problem: resistance will certainly increase as the drug persists, especially at low levels (e.g., residues) for long periods of time. However, this concern is irrelevant with substances that do not leave residues (alcohols, bleaches, peroxides), although it could be possible with preservatives.

It is claimed that all formaldehyde-donors/releasers are microbicidal on account of the formaldehyde released. However, in some cases, the low quantity of released formaldehyde may not be sufficient to create a biocidal action. We have shown that mechanisms other than formaldehyde-dehydrogenase activity could be responsible for a moderate tolerance to these preservatives. Also, we showed that formaldehyde-donors must penetrate into the cell through a hydrophilic pathway (porins) like some antibiotics and they could be substrates for efflux pumps.

As in clinical or veterinary practice, the development of antibiotic resistant strains, due in this case to the selective pressure from preservatives included in consumer products, could be a risk for human health. Nevertheless, the great contribution of disinfection, preservation and acceptance of hygienic measures that have supported advances in public health over the last century cannot be denied. Indeed, if reductions in the number of infections requiring antibiotic treatment can be achieved through the use of biocidal products, then this is likely to decrease rather than increase the incidence of antibiotic resistance. However, in order to preserve the role of biocides in hygiene, it is paramount to prevent the emergence of bacterial resistance and cross-resistance through their appropriate and prudent use.

Competing interests. None declared.

References 1. Berlanga M, Viñas M (2000) Salicylate induction of phenotypic resistance to quinolones in Serratia marcescens. 46:279-282 2. Berlanga M, Ruiz N, Hernandez-Borrell J, Montero T, Viñas M (2000) Role of the outer membrane in the accumulation of quinolones by Ser­ ratia marcescens. Can J Microbiol 46:716-722

ORÚS ET AL.

3. Berlanga M, Montero MT, Hernández-Borrell J, Viñas M (2004) Influence of the cell wall on ciprofloxacin susceptibility in selected wild-type Gram-negative and Gram-positive bacteria. Int J Antimicrob Agents 23:627-630 4. Bruchmann S, Dötsch A, Nouri B, Chaberny IF, Häussler S (2013) Quantitative contributions of target alteration and decreased drug accumulation to Pseudomonas aeruginosa fluoroquinolone resistance. Antimicrob Agents Chemother 57:1361-1368 5. Capita R, Riesco-Peláez F, Alonso-Hernando A, Alonso-Calleja C (2014) Exposure of Escherichia coli ATTCC 12806 to sublethal concentrations of food-grade biocides influences its ability to form biofilm, resistance to antimicrobials, and ultrastructure. Appl Environ Microbiol 80:1268-1280 6. Choi C-W, Park EC, Yun SH, Lee S-Y, Lee YG et al. (2014) Proteomic characterization of the outer membrane vesicle of Pseudomonas putida KT2440. J Proteome Res 13:4298-4309 7. Condell O, Iversen C, Cooney S, Power KA, Walsh C, Burgess C, Fanning S (2012) Efficacy of biocides used in the modern food industry to control Salmonella enterica, and links between biocide tolerance and resistance to clinically relevant antimicrobial compounds. Appl Environ Microbiol doi:10.1128/AEM.075334-11 8. Djeribi R, Boucherit Z, Bouchloukh W, Zouaoui W, Latrache H, Hamadi F, Menaa B (2013). A study of pH effects on the bacterial surface physicochemical properties of Acinetobacter baumannii. Colloids Surface B 102:540-545 9. Furi L, Ciusa ML, Knight D, Di Lorenzo V, Tocci N et al. (2013) Evaluation of reduced susceptibility to quaternary ammonium compounds and bisbiguanides in clinical isolates and laboratory-generated mutants of Staphylococcus aureus. Antimicrob Agents Chemother 57:34883497 10. Giacomel CB, Dartora G, Dienfethaeler HS, Haas SE (2013) Investigation on the use of expired make-up and microbiological contamination of mascaras. Int J Cosmetic Sci doi:10.1111/ics12053 11. Giaouris E, Chapot-Chartier M-P, Briandet R (2009) Surface physicochemical analysis of natural Lactococcus lactis strains reveals the existence of hydrophobic and low charged strains with altered adhesive properties. J Food Microbiol 131:2-9 12. Gilber P, Moore LE (2005) Cationic antiseptics: diversity of action under a common epithet. J Appl Microbiol 99:703-715 13. Gomez-Escalada M, Russell AD, Maillard J-Y, Ochs D (2005) Triclosan–bacteria interactions: single or multiple target sites? Lett Appl Microbiol 41:476-481 14. González-Zorn B, Escudero JA (2012) Ecology of antimicrobial resistance: humans, animals, food and environment. Int Microbiol 15:101-109 15. Hamadi F, Latrache H, Zahir H, El Abed S, Ellouali M, Saad IK (2012) The relation between the surface chemical composition of Escherichia coli and their electron donor/electron acceptor (acid-base) properties.

Res J Microbiol 7:32-40 16. Hinchliffe P, Symmons MF, Hughes C, Koronakis V (2013) Structure and operation of bacterial tripartite pumps. Annu Rev Microbiol 67:221-242 17. Hiom SJ (2013) Preservation of medicines and cosmetics. In: Fraise AP, Maillard J-Y, Sattar SA (eds.) Principles and practice of disinfection, preservation and sterilization, John Wiley, Oxford, pp 388-407 18. Ito K, Takahashi M, Yoshimoto T, Tsuru D (1994) Cloning and highlevel expression of the glutathione-independent formaldehyde dehydrogenase gene from Pseudomonas putida. J Bacteriol 176:2483-2491 19. Jaskulski MR, Medeiros BC, Borges JV, Zalewsky R, Fonseca MEC et al. (2013) Assessment of extended-spectrum β-lactamase, KPC carbapen­ emase and porin resistance mechanisms in clinical samples of Klebsiella pneumonia and Enterobacter spp. Int Antimicrob Agents 42:76-79

ANTIBIOTIC CROSS-RESISTANCE IN COSMETIC PRESERVATIVE

20. Kaneko M, Yamaguchi A, Sawai T (1984) Purification and characterization of two kinds of porins from the Enterobacter cloacae outer membrane. J Bacteriol 158:1179-1181 21. Kaulfers PM, Marquardt A (1991) Demonstration of formaldehyde dehydrogenase activity in formaldehyde-resistant Enterobacteriaceae. FEMS Microbiol Lett 79:335-338 22. Kohanski MA, Dwyer DJ, Collins JJ (2010) How antibiotics kill bacteria: from targets to networks. Nat Rev Microbiol doi:10.1038/nrmicro2333 23. Maillard JY (2002) Bacterial target sites for biocide action. J Appl Microbiol 92:16S-27S 24. McDonell G, Russell AD (1999) Antiseptics and disinfectants: activity, action, and resistance. Clin Microbiol Rev 12:147-179 25. Molina-Cabrillana J, Bolaños-Rivero M, Á��������������������������� ���������������������������� lvarez-León EE, Martín Sánchez AM, Sánchez-Palacios M, ��������������������������������������� Ál������������������������������������� varez D, Sáez-Nieto JA (2006) Intrinsically contaminated alcohol-free mouthwash implicated in a nosocomial outbreak of Burkholderia cepacia colonization and infection. Infect Control Hosp Epidemiol 27:1281-1282 26. Nikaido H (2003) Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593-656 27. Ortega Morente E, Fernández-Fuentes MA, Grande Burgos MJ, Abriouel H, Pérez Pulido R, Gálvez A (2013) Biocide tolerance in bacteria. Int J Food Microbiol doi:10.1016/j.ijfoodmicro.2012.12.028 28. Orús P, Leranoz S (2005) Current trends in cosmetic microbiology. Int Microbiol 8:77-79 29. Pérez A, Canle D, Latasa C, Poza M, Beceiro A, et al., (2007) Cloning, nucleotide sequencing, and analysis of the AcrAB-TolC eflux pump of Enterobacter cloacae and determination of its involvement in antibiotic resistance in a clinical isolate. Antimicrob Agents Chemother 51:3247-3253

Int. Microbiol. Vol. 18, 2015

59

30. Périamé M, Pagès J-M, Davin-Regli (2014) Enterobacter gergoviae membrane modifications are involved in the adaptive response to preservatives used in cosmetic industry. J Appl Microbiol 118:49-61 31. Piddock LJV (2006) Multidrug-resistance efflux pumps—not just for resistance. Nat Rev Microbiol 4:629-636 32. Russell AD, McDonnell G (2000) Concentration: a major factor in studying biocidal action. J Hosp Infect 44:1-3 33. Saeb-Lima M, Solis-Arreola GV, Fernandez-Flores A (2015) Mycobacterial infection after cosmetic procedure with botulinum toxin A. J Clin Diagn Res doi:10.7860/JCDR/2015/11741.5756 34. Sheldon ATJr (2005) Antiseptic “resistance”: real or perceived threat?

Clin Infect Dis 40:1650-1656 35. Warburton KL, Wilkinson M (2015) Contact allergy to methylisothiazolinone: Has there been any change? Experience of a UK centre. Contact Derm doi:10.1111/cod.12403 36. Webber MA, Whitehead RN, Mount M, Loman NJ, Pallen MJ, Piddock LJV (2015) Parallel evolutionary pathways to antibiotic resistance selected by biocide exposure. J Antimicrob Chemother. doi:10.1093/jac/ dkv109 37. Whitehead RN, Overton TW, Kemp CL, Webber MA (2011) Exposure of Salmonella enterica serovar Typhimurium to high level biocide challenge can select multidrug resistant mutants in a single step. PLoS One 6:e22833 38. Zhang L, Li XZ, Poole K (2001) Fluoroquinolone susceptibilities of efflux-mediated multidrug-resistant Pseudomonas aeruginosa, Stenotro­ phomonas maltophilia and Burkholderia cepacia. J Antimicrob Chemother 48:549-552 39. Zhou G, Shi Q-S, Ouyang Y-S, Chen Y-B (2014) Involvement of outer membrane proteins and peroxide-sensor genes in Burkholderia cepacia resistance to isothiazolone. World J Microbiol Biotechnol 30:1251-1260

60

Int. Microbiol. Vol. 18, 2015

ORĂ&#x161;S ET AL.

RESEARCH ARTICLE International Microbiology (2015) 18:61-69 doi:10.2436/20.1501.01.235. ISSN (print): 1139-6709. e-ISSN: 1618-1095

www.im.microbios.org

Properties of Lactobacillus reuteri chitosancalcium-alginate encapsulation under simulated gastrointestinal conditions Hui-Ying Huang,1¶ Yi-Ju Tang,1¶ V. An-Erl King,2 Jen-Wei Chou,3 Jen-Horng Tsen1* Department of Nutrition, China Medical University, Taichung, Taiwan, ROC. 2Department of Food Science and Biotechnology, National Chung-Hsing University, Taichung, Taiwan, ROC. 3Division of Gastroenterology and Hepatology, Department of Internal Medicine, China Medical University Hospital, College of Medicine, China Medical University, Taichung, Taiwan, ROC 1

Received 17 January 2014 · Accepted 30 March 2015

Summary. The protective effects of encapsulation on the survival of Lactobacillus reuteri and the retention of the bacterium’s probiotic properties under simulated gastrointestinal conditions were investigated. Viable counts and the remaining probiotic properties of calcium (Ca)-alginate encapsulated (A group), chitosan-Ca-alginate encapsulated (CA group), and unencapsulated, free L. reuteri (F group) were determined. Encapsulation improved the survival of L. reuteri subjected to simulated gastrointestinal conditions, with the greatest protective effect achieved in the CA group. The degree of cell membrane injury increased with increasing bile salt concentrations at constant pH, but the extent of injury was less in the encapsulated than in the free cells. Adherence rates were, in descending order: CA (0.524%) > A (0.360%) > F (0.275%). Lactobacillus reuteri cells retained their antagonistic activity toward Listeria monocytogenes even after incubation of the lactobacilli under simulated gastrointestinal conditions. Displacement of the pathogen by cells released from either of the encapsulation matrices was higher than that by free cells. The safety of L. reuteri was demonstrated in an in vitro invasion assay. [Int Microbiol 2015; 18(1):61-69] Keywords: Lactobacillus reuteri · Listeria monocytogenes · chitosan–calcium-alginate encapsulation · probiotic properties · simulated gastrointestinal conditions

Introduction Most probiotics belong to strains of lactic acid bacteria (LAB) and their positive effects on human health are well established [40]. These benefits of LAB include: balancing gut microbiota [38], inhibiting infection by pathogens [11], lowering blood cholesterol [20], and reducing the risk of colon cancer [19].

Corresponding author: J.-H.Tsen Department of Nutrition China Medical University 91, Hsueh-Shih Road Taichung, 404 Taiwan, ROC E-mail: jhtsen@mail.cmu.edu.tw *

¶

These authors contributed equally to the work.

For probiotics to exert these and other beneficial effects, they must pass safely through the gastrointestinal tract and then adhere to and colonize the intestinal canal [42]. However, the survival of many microbes exposed to gastric acid and bile salts is poor.

Using encapsulation to immobilize LAB cells can improve their survival under adverse conditions [45]. Encap­ sulation offers many advantages for the encapsulated cells, including the maintenance of stability, activity, and high volumetric productivity, the improvement of process control, protection against damage, and a reduced susceptibility to contamination [25,41]. Several encapsulation techniques to improve the survival of microorganisms in dairy products [35] and artificial gastrointestinal juice [10] have been tested, with

62

Int. Microbiol. Vol. 18, 2015

alginate encapsulation as one of the most successful [18]. The encapsulation of LAB in Ca-alginate beads improved the survival of the cells under harsh conditions [25,31]. Chitosan has also been used as the encapsulation material, especially for drug delivery in the gastrointestinal tract, based on its absorption-enhancing, controlled-release, and bioadhesive properties [8]. Chito­san is obtained from chitin via N-deace­tylation and the immersion of Ca-alginate beads in chitosan solution results in the formation of an outer, protective membrane [21]. Many studies have shown that encapsulation of LAB in a chitosan-alginate complex is effective in reducing the decline of viable cells exposed to simulated gastric and intestinal juice and improves their survival during refrigerated storage [2,22,24].

In addition to tolerating gastric acid and bile salts, probio­ tics must be able to adhere to gut surfaces, colonize the gut together with the resident microbiota, and compete with pathogens [11]. The probiotic traits of LAB with respect to gastric acid and bile tolerance, adhesiveness, and competition with pathogens have been investigated [13,31,49] but these studies often failed to examine the influence of sequential acid and bile salt exposure, as occurs under physiological conditions in the human gastrointestinal tract [25]. Moreover, few studies have considered the remaining probiotic properties of LAB cells after their exposure to simulated gastrointestinal conditions. In a previous work [16], we found that the exposure of LAB strains to low pH followed by high concentrations of bile salts led to a decline in cell survival. In addition, there was a loss of adhesiveness of viable LAB, perhaps due to the cell injury caused by the harsh conditions. LAB strains that survived sequential incubations at pH 4 and 0.1% bile salt had higher adherence rates but a slightly lower rate of pathogen displacement than an unexposed LAB strain. In this study, Lactobacillus reuteri was encapsulated in Ca-alginate and chitosan-Ca-alginate and the protective effects of encapsulation on the survival of cells exposed to sequential acid and bile treatments was evaluated and compared. In addition, we assessed the adhesiveness, safety, and displacement of Listeria monocytogenes by free and encapsulated bacteria subjected to simulated gastrointestinal conditions.

Materials and methods Bacterial strains and culture conditions. Lactobacillus reuteri BCRC 14625 and Listeria monocytogenes BCRC 14847 were purchased from the Bioresource Collection and Research Center of the Food Industry Research and Development Institute at Hsinchu, Taiwan. Lactobacillus reuteri was grown in de Man, Rogosa and Sharpe (MRS) broth or on MRS agar

HUANG ET AL.

medium (Difco). Listeria monocytogenes was cultured with Bacto brain heart infusion (BHI; Difco). Both strains were incubated at 37ºC and propagated under microaerophilic conditions. The stock cultures were preserved at –80ºC in 20% glycerol.

Encapsulation of Lactobacillus reuteri. Bacteria grown at 37ºC for 24 h were centrifuged at 4ºC, 8000 ×g for 10 min. The harvested cell pellets were washed and then resuspended in phosphate-buffered saline (PBS).

The bacterial suspension was mixed with an equal volume of 2% (w/v) sodium alginate (Na-alginate) solution to obtain a cell suspension containing 1.0– 9.9 × 108 colony-forming units (CFU)/ml. The solution was transferred dropwise into 0.1 M CaCl2 through the tube of a peristaltic pump, resulting in the formation of calcium alginate (Ca-alginate) beads containing L. reuteri cells (A group). A suspension of LAB mixed with an equal volume of H2O served as the control free cell group (F group). The chitosan-Ca-alginate beads used to encapsulate L. reuteri cells were prepared as previously described [23].

Briefly, 0.4 g of chitosan (Sigma) was dissolved in 90 ml of distilled water (DW) previously acidified with 0.4 ml of glacial acetic acid. The final concentration of chitosan solution was first adjusted to 0.4% (w/v) and then to pH 5.7–6.0 with 1 N NaOH. The chitosan solution was then filtered and the volume was increased to 100 ml with DW. Ca-alginate beads containing L. reuteri prepared as described above were then immersed in the 0.4% chitosan solution and shaken at 100 rpm for 1 min to produce chitosan-coated-Ca-alginate beads (CA group). The interactions of the three groups with the pathogen List. monocytogenes were assessed in wells containing Caco-2 cells and pre-adhered, FITC-labeled List. monocytogenes. The adherence rate of the pathogen under LAB-free conditions was designated as 100%.

Microscopic observation of bead structure. The microstructures of the beads and the encapsulated LAB were examined using scanning electron microscopy [34]. The Ca-alginate and chitosan-Ca-alginate beads were fixed with 2.5% glutaraldehyde at 4ºC for 16 h, washed with PBS, and dehydrated in an ethanol gradient (30, 50, 70, 80, 85, 90, 95, and 100% ethanol, each for 15 min). The beads were then dried by critical point drying, coated with gold, and observed in a JEOL, JSM-7401F (Japan) scanning electron microscope.

Tolerance of simulated digestive juice. The initial bacterial counts were obtained as follows: for the F group, a previously prepared L.­ reuteri suspension was mixed with an equal volume of aseptic water. For the A and CA groups, the L. reuteri suspension was mixed with an equal volume of Na-alginate solution. The number of cells in 1 ml of each of the three suspensions was counted after 24–48 h of incubation at 37ºC using the standard plate count (SPC) method. The data were expressed as the log CFU/ml. The initial LAB counts were expected to be the same in the free and encapsulated samples. One ml of free L. reuteri cells for the F group and gel beads prepared from 1 ml of an equal volume LAB cells and Na-alginate solution for the A and CA groups (~40 gel beads for each encapsulation group) were added to 10 ml of 0.1% peptone water adjusted to pH 2–3 [11]. The samples were shaken at 150 rpm at 37°C for 3 h. The acid-treated free cells were centrifuged (25°C, 8000 ×g, 10 min) and the pellets then dissolved in 1 ml of aseptic water. Acid-treated beads of the A and CA groups were collected and drained dry. The dried beads and 1 ml of acid-treated free cells were then transferred to 0.1%, 0.5%, or 1% oxgall bile (Sigma) and shaken as described for the acid treatment. After bile treatment, the free cells and gel beads were treated as previously described and then transferred to 10 ml of simulated colon fluid (0.1 M KH2PO4, pH 7.4 ± 0.2), with shaking for 1 h. Under these conditions, the gel beads depolymerized, releasing encapsulated L. reuteri.

Both the free and the released cells were centrifuged, re-suspended in 1 ml of aseptic water, and the number of viable cells was determined.

Int. Microbiol. Vol. 18, 2015

PROTECTIVE ENCAPSULATION OF PROBIOTIC L. REUTERI

63

β-Galactosidase activity of Lactobacillus reuteri after simulated digestive juice treatment. The β-galactosidase activity of L. reuteri after simulated digestive juice treatment was assayed as described previously [32]. Both free and encapsulated LAB cells were treated with acid solutions of pH 2 or pH 3, then with 0.1%, 0.5% or 1% bile salt solutions, each for 3 h, and then with simulated colon fluid for1 h. The free cells and the cells released from the beads were collected by centrifugation, washed, and re-dissolved in 1 ml of aseptic water. The cell suspensions were then mixed with 4 ml of 0.005 M o-nitrophenyl-β-d-galactopyranoside (ONPG; Sigma) and incubated at 37ºC for 10 min to allow color formation.

After the reaction was stopped by the addition of 2 ml of 0.1 M sodium carbonate, the samples were centrifuged at 8000 ×g at 1°C for 10 min to precipitate and remove the cells. The absorbance of the supernatant was measured at 420 nm with a spectrophotometer (Hitachi, U-2000, Japan). The amount of o-nitrophenol (ONP) formed in the sample was calculated by comparison with a standard curve. β-galactosidase activity was expressed as micromoles of ONP formed per milliliter of cell suspension in 10 min.

PBS. To test the displacement effect of L. reuteri, 1 ml of FITC-labeled List. monocytogenes suspension was added to a well with a fixed Caco-2 cell monolayer and incubated for 2 h. One ml of gel-released or free LAB cells treated as described for the adhesion assay was added to a well with pre-adherent List. monocytogenes cells. After 2 h of incubation at 37°C, the incubation medium was discarded and the Caco-2 cell monolayer was washed five times with PBS to remove both non-adherent LAB and pathogen cells. The adherence of FITC-labeled List. monocytogenes cells to the Caco-2 cell monolayer was assessed with a spectrofluorometer (Bio-Tek, FLX-800, USA) by recording the fluorescence intensity at 485 nm excitation and 528 nm emission. The antagonistic activity of L. reuteri was expressed as the adherence rate (%) of the pathogen (non-displaced List. monocytogenes cells) as determined by its fluorescence intensity; that is, the greater the antagonism of L. reuteri, the greater the displacement of List. monocytogenes. The “LABfree” control consisted of FITC-labeled List. monocytogenes adhered to the Caco-2 cells and not exposed to competitive LAB. The adherence rate under these conditions was defined as 100%.

Cell culture. The epithelial-like Caco-2 cell line BCRC 60182, originally isolated from a human colon adenocarcinoma [36], was purchased from Bioresource Collection and Research Center of the Food Industry Research and Development Institute at Hsinchu, Taiwan. The cells were grown in Dulbecco’s modified Eagle’s minimal essential medium (DMEM; Biochrom AG, Germany) supplemented with 10% (v/v) heat-inactivated (56ºC, 30 min) fetal bovine serum (FBS; Biological, Israel) and 1% (v/v) penicillin-streptomycin (stock solution 100 unit/ml; BioSource, USA) and incubated at 37ºC in a CO2 incubator (NUAIRE, NU-5500, USA) with a 5% CO2/95% air atmosphere. Subcultured cells were seeded at a concentration of 3 × 105 cells per well in a six-well tissue culture plate. The culture medium was replaced every other day. Mature monolayers, obtained after 15 days of incubation, were used for the following assays [4].

Invasion assay. The in vitro safety of L. reuteri was evaluated in an invasion assay as previously reported [44]. The Caco-2 cell monolayer was washed twice with PBS and re-incubated for 1.5 h in fresh DMEM medium without penicillin and streptomycin. Gel-released and free LAB cells, both pre-treated with simulated gastrointestinal conditions, were suspended in antibiotic-free DMEM medium. One ml of LAB suspension was loaded into a well containing a Caco-2 cell monolayer and incubated for 2 h at 37ºC in a 5% CO2/95% air. The medium was then aspirated and Caco-2 cells were washed three times with PBS to remove non-adherent LAB. DMEM containing 10 mg tetracycline/ml was added to the well and incubated for another 2 h to kill the adherent LAB. The Caco-2 cells were washed five times with PBS and lysed with 1% Triton X-100. Viable counts of LAB released from the Caco-2 cells were determined. The safety of L. reuteri was expressed as the invasion rate and calculated as follows:

Adhesion assay. At least 1.5 h before the adhesion assay, the DMEM medium was aspirated from the wells and the Caco-2 cell monolayer was fixed with 0.25% glutaraldehyde for 15 min followed by three washes with PBS. Fresh DMEM not containing FBS or antibiotics was added to each well [28]. After exposure to the above-described simulated gastrointestinal conditions (pH 3, 0.1% bile salt, and simulated colon fluid), 1 ml of L. reuteri cells released from the gel beads or the free L. reuteri cell suspension was added to the respective wells and incubated at 37ºC for 2 h in a 5% CO2/95% air atmosphere. The monolayers were then washed five times with PBS to remove non-adherent LAB cells [14] and lysed with 1 ml of a 1% Triton X-100 solution (Sigma) for 5 min to release adherent bacteria. The number of adherent LAB released from the surface of the Caco-2 cells was determined by the SPC method using plates incubated for 24-48 h at 37°C [17]. The adherence rate (%) was calculated as:

Adherence rate (%) =

No. of adherent LAB (CFU/ml) No. of LAB added to each well (CFU/ml)

× 100%

Antagonistic activity of Lactobacillus reuteri against Lis­ teria monocytogenes. A suspension of List. monocytogenes cells prepared from a culture was centrifuged (25ºC, 8000 ×g, 10 min) to pellet the cells, which were then washed three times and resuspended in PBS. The suspension was mixed with an equal volume of 2 mg of fluorescein isothiocyanate (FITC; Sigma)/ml for 20 min to stain the bacterial cells. Excess fluorochrome was removed by centrifugation of the FITC-labeled bacterial suspension, discarding the supernatant and washing the cell pellet three times with

Invasion rate (%) =

No. of LAB released from Caco-2 cells (CFU/ml) No. of LAB added to each well (CFU/ml)

× 100%

Statistical analysis. Statistical differences between samples were determined using Statistical Analysis System software (version 9.2, 2008).

Duncan’s new multiple range test, Dunnett’s test of the GLM, and Student’s t-test were used to determine significance, defined as a P < 0.05.

Results Microscopy of the gel beads and encapsulated cells. Figure 1 shows the scanning electron micrographs of the internal and external structures of the Ca-alginate (Fig. 1A, B), the chitosan-Ca-alginate (Fig. 1C,D) gel beads and the respective encapsulated L. reuteri cells. Rod-shaped L. reuteri spread throughout the gel beads of both encapsulation matrices but the density of bacteria was lower on the surface of the chitosan-Ca-alginate beads than on the surface of the Ca-alginate beads. By more effectively encapsulating LAB cells, chitosanCa-alginate may better protect cells from unfavorable environmental conditions.

Int. Microbiol. Vol. 18, 2015

HUANG ET AL.

Int Microbiol

64

Fig. 1. Scanning electron micrograph of gel beads containing encapsulated Lactobacillus reuteri cells (5000× magnification).

(A) Internal and (B) external structure of a Ca-alginate gel bead. (C) Internal and (D) external structure of a chitosan-Ca-alginate gel bead. Scale bar 1 µm.

Protection of Lactobacillus reuteri by encapsulation against simulated digestive juice. Table 1 shows the viable counts (log CFU/ml) of encapsulated and free L. reuteri after treatment first with an acid solution of pH 2 or 3, then with different concentrations of bile salt for 3 h, and then with simulated colon fluid for 1 h. LAB survival increased with increasing pH and lower bile salt concentrations.

No viable counts were found for groups F and A group at pH

2 in combination with any of the bile salt concentrations. By contrast, under the same treatment conditions, a portion of the cells in the CA group survived. The viable counts of the CA group were the highest among the free and encapsulated groups. These results demonstrated that encapsulation improves the survival of L. reuteri after acid and bile salt treatments, with CA having a significantly better protective effect (P < 0.05).

Table 1. Viable counts (log CFU/ml*) of encapsulated and free Lactobacillus reuteri pre-treated with a pH 2 or 3 solution for 3 h, then with 0.1, 0.5, or 1% bile salt for 3 h, and finally with simulated colonic fluid for 1 h Cell treatment

Initial bacteria count (log CFU/ml)

pH 2 0.1% bile

0.5% bile

pH 3 1% bile

0.1% bile

0.5% bile

1% bile

Free cells (control)

8.83 ± 0.07

NDbx

NDbx

NDbx

3.14 ± 0.02cx

1.22 ± 0.07cy

1.00 ± 0.00cy

Alginate-coated

8.83 ± 0.07

NDbx

NDbx

NDbx

5.70 ± 0.08bx

3.72 ± 0.05by

3.15 ± 0.02by

Chitosan-Ca-alginate-coated

8.83 ± 0.07

2.62 ± 0.07ax

2.07 ± 0.07ay

1.40 ± 0.07az

5.96 ± 0.067ax

3.99 ± 0.09ay

3.37 ± 0.08ay

*One ml of an equal volume of LAB and aseptic water (free LAB) or LAB and Na-alginate (encapsulated LAB). a-c Different letters within the same column differ significantly (P < 0.05; n = 3). x-z Different letters within a row of the same pH values differ significantly (P < 0.05; n = 3).

Int. Microbiol. Vol. 18, 2015

PROTECTIVE ENCAPSULATION OF PROBIOTIC L. REUTERI

65

Table 2. β-Galactosidase activity (μmol/10 min/ml) of encapsulated and free Lactobacillus reuteri pre-treated with a pH 2 or 3 solution for 3 h, then with 0.1, 0.5, or 1% bile salt for 3 h, and finally with simulated colonic fluid for 1 h pH 2

pH 3

Cell treatment

Blank*

0.1% bile

Free cells (control)

0.066x

0.008bx

0.013bx

0.052bx

1781.769ax

194176.592ay

338656.801az

Alginate-coated

0.066x

0.023bx

0.095bx

0.160bx

16.850bx

1906.229by

7585.133bz

Chitosan-Ca-alginate-coated

0.066x

558.554ax

6780.474ay

58144.768az

9.668bx

964.351bx

4530.215bxy

0.5% bile

1% bile

0.1% bile

0.5% bile

1% bile

*Blank: β-galactosidase activity of cells not exposed to acid, bile salt and simulated colonic fluid. a-b Different letters within the same column differ significantly (P < 0.05; n = 3). x-z Different letters within the same row differ significantly (P < 0.05; n = 3).

β-Galactosidase activity of encapsulated Lacto­bacillus reuteri exposed to simulated digestive juice. Table 2 shows the β-galactosidase activity (μmol/10 min/ml) of encapsulated and free L. reuteri after exposure of the cells to simulated gastrointestinal conditions.

As evidenced by the activity values, injury of the L. reuteri cell membrane increased with increasing bile salt concentrations at the same pH value. At pH 3, encapsulated cells of the A and CA groups had significantly lower β-galactosidase activity than did the free cells.

Adhesion of encapsulated Lactobacillus reuteri after simulated digestive juice treatment. The adherence rates of encapsulated and free L. reuteri after treatment with simulated gastrointestinal conditions are shown in Table 3. L. reuteri cells sequentially exposed to pH 3, 0.1% bile salt, and simulated colonic fluid retained their adhesion ability, with the adhesion of the CA group (0.524%) > A group (0.360%) > F group (0.275%). Thus, the encapsulated cells were more adhesive than the free cells.

Antagonistic activity toward Listeria mono­­­­ cytoge­nes. Figure 2 shows the displacement effect of L. reuteri on FITC-labeled List. monocytogenes previously adhered to Caco-2 cells. The corresponding fluorescence microscopy photographs are shown in Fig. 3. After treatment with simulated gastrointestinal conditions, all three groups of

L. reuteri were able to displace List. monocytogenes. The adherence rate (%) of the LAB-free group in the absence of L. reuteri cells was designated as 100%. After the addition of LAB cells of the F, A, and CA groups to pre-adhered List. monocytogenes, adherence of the pathogen was reduced to 93.61%, 69.63%, and 67.02%, respectively. Thus, after sequential acid and bile treatments, the antagonism of List. monocytogenes adhered to Caco-2 cells was significantly greater by the encapsulated cells than by free L. reuteri (P < 0.05). The fluorescence images in Fig. 3 qualitatively confirm these results, as the density of List. monocytogenes luminescence was progressively reduced from the F group to the A group and the CA group (Fig. 3).

Safety of Lactobacillus reuteri after simulated digestive juice treatment. The invasion rates of encapsulated and free L. reuteri after exposure of the cells to simulated gastrointestinal conditions were zero in all cases.

The lack of invasiveness means that, after acid and bile salt treatments, LAB are recognized as safe in vitro.

Discussion To achieve the health benefits of dietary probiotics, the food product should contain at least 107 live organisms per gram or milliliter [33,43]. Encapsulation improves the survival of pro-

Table 3. Adherence rates (%) of encapsulated and free Lactobacillus reuteri cells after their exposure to simulated gastrointestinal conditions Cell treatment

Adherence rate (%)

Free cells

0.275a

Alginate-coated

0.360b

Chitosan-Ca-alginate-coated

0.524c

Means followed by different superscript lowercase letters differ significantly from each other (P < 0.05: n = 3).

a-c

Int. Microbiol. Vol. 18, 2015

HUANG ET AL.

Int Microbiol

66

biotic microorganisms during both food processing and passage through the gastrointestinal tract [35,13]. The protective effects of the Ca-alginate encapsulation of probiotics against the action of simulated gastric acid and bile salt have been reported previously [13,25]. However, bacterial viability varies depending on the encapsulation method, wall material, and probiotic strains. Ca-alginate beads with and without chitosan coating have been used to encapsulate Lactobacillus acidophilus 547, Bifidobacterium bifidum ATCC 1994, and L. casei 01, and the survival advantage conferred by the encapsulation of probiotics when subjected to simulated gastrointestinal conditions have been investigated [22]. The results of those studies were similar to those of the present study of L. reuteri, as they also demonstrated the superiority of chitosan-coated alginate beads in protecting against the action of acid and bile salt.

The acid and bile tolerance properties of LAB have been investigated in several in vitro studies [13,31,49]. In most cases, acid and bile tolerance were examined and reported separately; however, these conditions do not resemble those of the human gastrointestinal tract [2,7], where ingested LAB first pass through the highly acidic environment of the stomach, are then transported along the small intestine, and finally colonize the colon [30,37]. Thus, simulated gastrointestinal conditions in vitro studies should be based on the sequential exposure of LAB to acid and bile [25]. In this study, L. reuteri encapsulated using Ca-alginate and chitosan-Ca-alginate were treated sequentially with acid and bile salt before being released into simulated colonic fluid. In agreement with previous reports [22,25], the results showed that the viability of encapsulated L. reuteri, especially those in the CA group, was

Fig. 2. Displacement by Lactobacillus reuteri of pre-adhered and FITClabeled Listeria monocytogenes from Caco-2 cells. a-c Groups with different letters differ significantly from each other (P < 0.05). LAB-free group: pathogen adhesion in the absence of L. reuteri; F group: non-encapsulated, free L. reuteri cells; A group: Ca-alginate encapsulated L. reuteri cells; CA group: chitosan-Ca-alginate encapsulated L. reuteri cells.

significantly greater than that of the free cells (Table 1). Thus, encapsulation can protect probiotics from the detrimental effects of the harsh conditions of the gastrointestinal tract.

Because the L. reuteri tolerance tests were conducted immediately after the beads had been prepared, most of the LAB cells were still encapsulated inside the Ca-alginate or chitosan-Ca-alginate matrix. During their brief, sequential acid and bile salt treatments, there was little release of the encapsulated LAB cells, as confirmed by the fact that no live L. reuteri cells were detected in the acid or bile salt solutions even under the mildest conditions (data not shown).

Injury to the bacterial cells was monitored based on the action of β-galactosidase, located on the inner membrane of LAB. ONPG (o-nitrophenyl-β-galactoside) is a colorless substrate analog for the detection of β-galactosidase activity.

When the cell membrane is injured, causing permeability to rise, ONPG enters the cell and reacts with β-galactosidase to produce the yellow-colored o-nitrophenol (ONP), which can be quantified by colorimetric assay. In a previous study, the simulated intestinal juice of oxgall increased cellular permeability and therefore β-galactosidase activity [32]. In this study, β-galactosidase activity increased with increasing bile salt concentrations whether at pH 2 or pH 3. Encapsulated L. reuteri and especially in the CA group had the lowest β-galactosidase activity at pH 3.

Since the first step in microbial colonization, and therefore the initiation of health benefits, is adherence to the intestinal surfaces, measurements of adhesiveness are essential in probiotic screening [15]. The adhesiveness of probiotics as well as pathogens reflects the characteristic of the bacterial cell surface [1] and involves receptors on the surface of the host

Int. Microbiol. Vol. 18, 2015

67

Int Microbiol

PROTECTIVE ENCAPSULATION OF PROBIOTIC L. REUTERI

Fig. 3. Fluorescence microscopy photographs showing the displacement of pre-adhered and FITC-labeled Listeria monocytogenes by Lactobacillus reuteri. (A) LAB-free group; (B) free L. reuteri cells (F group); (C) Ca-alginate group (A group); (D) chitosan-Ca-alginate group (CA group).

intestinal cells [29]. In the case of LAB, adherence to human intestinal mucosa requires extracellular glycoprotein or protein structures on the bacterial cell surface [39,46]. Accord足 ing to the adherence rates in Table 3, both released and free L. reuteri subjected to simulated gastrointestinal conditions retained their ability to adhere to Caco-2 cells, with the released encapsulated L. reuteri displaying greater adhesiveness than free cells. This result suggests that encapsulation should protect cell surface factors such as proteins or carbohydrates from the injury caused by acid and bile salt and thus would maintain bacterial adhesiveness.

Inhibition of the adhesion of pathogens to intestinal epithelial cells by LAB has been demonstrated in vivo [3,26] and may be related to competition for specific adhesion receptors on the surface of gut cells [27]. Another study [9] has found that reuterin, produced by L. reuteri, inhibits both Grampositive and Gram-negative bacteria. The inhibition of Heli足 co足bacter pylori by L. reuteri was proposed to involve bacteriostatic activity, the production of inhibitory compounds such as lactate and bactericidal substances, competition for nutrients, immunostimulation of mucosal IgA production,

and/or the adherent capacity of L. reuteri to epithelial cells [12]. In this study, pathogen displacement from the intestinal surface was modeled using Caco-2 cells incubated first with List. monocytogenes and then with LAB previously exposed to simulated gastrointestinal conditions. Pathogen adherence was significantly reduced by cells in both the free and encapsulated groups, with the latter having a more potent effect.

In a previous study, we investigated the adhesion inhibition and displacement abilities of three LAB strains (Lactobacillus acidophilus BCRC 10695, L. paracasei BCRC 14023, Bifidobacterium bifidum BCRC 14615) using the pathogen Clostridium perfringens BCRC 13019 [16]. The adhesion inhibition test measured the ability of LAB cells preadhered to Caco-2 cells to inhibit the adhesion of C. perfringens. The displacement assay examined the displacement effect of LAB cells pretreated or not with acid and bile salt.

Because the experimental design used in the present work sought to mimic the conditions of the gastrointestinal tract, only the displacement assay was performed and the antagonistic ability of encapsulated and free L. reuteri cells after sequential acid and bile salt treatments was compared. The re-

68

Int. Microbiol. Vol. 18, 2015

sults confirmed the protective effect of encapsulation based on the lower adherence rates of List. monocytogenes achieved with A and CA group cells than with F group cells (Fig. 2).

The invasion assay was used to determine the infectivity and pathogenicity of the tested LAB, as previously described [44]. LAB are generally recognized as safe for multiple uses.

The health benefits of L. reuteri based on their colonization of the gastrointestinal tract and immunomodulation have been reported [47] and their daily ingestion is safe, with no negative side effects [6,48]. The safety of L. reuteri BCRC strain 14625 used in this research had been previously demonstrated [5] and was confirmed in this study by the absence of invasion by free or released cells exposed to acid (pH 3) and then bile salt (0.1%).

In summary, this study investigated the protective effects of Ca-alginate and chitosan-Ca-alginate encapsulation of L. reuteri, and the remaining probiotic properties of the released bacteria after their sequential exposure to simulated gastrointestinal conditions. The results showed that, compared to free (non-encapsulated) cells, encapsulation, especially using chitosan-Ca-alginate gel beads, improves the survival of acid- and bile-salt-treated L. reuteri, reduces injury to the cell membrane, and preserves the probiotic properties of adhesiveness and pathogen antagonism. Our method of challenging gel-encapsulated lactic acid bacteria with sequential simulated acid and bile salt treatment is a suitable approach to investigating their potential probiotic properties [16].

Acknowledgements. We thank the National Science Council, Taiwan, ROC, for funding this project (grant number: NSC 96-2313-B-039-003).

Competing interests. None declared.

References 1. Bibiloni R, Perez PF, de Antoni GL (1999) Will a high adhering capacity in a probiotic strain guarantee exclusion of pathogens from intestinal epithelia? Anaerobe 5:519-524 2. Chávarri M, Marañón I, Ares R, Ibáñez FC, Marzo F, Villarán MC (2010) Microencapsulation of a probiotic and prebiotic in alginate-chitosan capsules improves survival in simulated gastro-intestinal conditions. Int J Food Microbiol 142:185-189 3. Coconnier MH, Bernet MF, Kernéis S, Chauvière G, Fourniat J, Servin AL (1993) Inhibition of adhesion of enteroinvasive pathogens to human intestinal Caco-2 cells by Lactobacillus acidophilus strain LB decreases bacterial invasion. FEMS Microbiol Lett 110:299-306 4. Coconnier MH, Klaenhammer TR, Kernéis S, Bernet MF, Servin AL (1992) Protein-mediated adhesion of Lactobacillus acidophilus BG2FO4 on human enterocyte and mucus-secreting cell lines in culture.

Appl Environ Microbiol 58:2034-2039

HUANG ET AL.

5. Collado MC, Grześkowiak Ł, Salminen S (2007) Probiotic strains and their combination inhibit in vitro adhesion of pathogens to pig intestinal mucosa. Curr Microbiol 55:260-265 6. Connolly E, Abrahamsson T, Björkstén B (2005) Safety of d(–)-lactic acid producing bacteria in the human infant. J Pediatr Gastroenterol Nutr 41:489-492 7. Curto AL, Pitino I, Mandalari G, Dainty JR, Faulks RM, Wickham MSJ (2011) Survival of probiotic lactobacilli in the upper gastrointestinal tract using an in vitro gastric model of digestion. Food Microbiol 28:13591366 8. Dodane V, Vilivalam VD (1998) Pharmaceutical applications of chitosan.

Pharm Sci Technolo Today1:246-253 9. El-Ziney M, van den Tempel T, Debevere J, Jakobsen M (1999) Application of reuterin produced by Lactobacillus reuteri 12002 for meat decontamination and preservation. J Food Prot 62:257-261 10. Fávaro-Trindale CS, Grosso CRF (2002) Microencapsulation of L. aci­ do­philus (La-05) and B. lactis (Bb-12) and evaluation of their survival at the pH values of the stomach and in bile. J Microencapsul 19:485-494 11. Fernández MF, Boris S, Barbés C (2003) Probiotic properties of human lactobacilli strains to be used in the gastrointestinal tract. J Appl Microbiol 94:449-455 12. Francavilla R, Lionetti E, Castellaneta SP, Magistà AM, Maurogiovanni G, Bucci N, De Canio A, Indrio F, Cavallo L, Ierardi E, Miniello VL (2008) Inhibition of Helicobacter pylori infection in humans by Lactobacillus reuteri ATCC 55730 and effect on eradication therapy: a pilot study. Helicobacter 13:127-134 13. Goderska K, Zybals M, Czarnecki Z (2003) Characterization of microencapsulated Lactobacillus rhamnosus LR7 strain. Pol J Food Nutr Sci 12:1-24 14. Gopal PK, Prasad J, Smart J, Gill HS (2001) In vitro adherence properties of Lactobacillus rhamnosus DR20 and Bifidobacterium lactis DR10 strains and their antagonistic activity against an enterotoxigenic Escherichia coli. Int J Food Microbiol 67:207-216 15. Havenaar R, Brink NG, Huis In’t Ved JHJ (1992) Selection of strains for probiotics use. In: Fuller R (ed) Probiotics, the scientific basis. Chapman & Hall, London, pp 210-224 16. Huang HY, Hsieh HY, King VAE, Chi LL, Tsen JH (2014) Lactic acid bacteria pre-challenged by simulated gastrointestinal conditions is a suitable approach for study of potential probiotic properties. J Microbiol Meth 107:138-146 17. Huang HY, Huang SY, Chen PY, King VAE, Lin YP, Tsen JH (2007) Basic characteristics of Sporolactobacillus inulinus BCRC 14647 for potential probiotic properties. Curr Microbiol 54:396-404 18. Kailasapathy K (2002) Microencapsulation of probiotic bacteria: technology and potential applications. Curr Issues Intest Microbiol 3:39-48 19. Kazuhiro H, Joseph R (2000) The role of probiotic bacteria in cancer prevention. Microbes Infect 2:681-686 20. Klaver FAM, van der Meer R (1993) The assumed assimilation of cholesterol by lactobacilli and Bifidobacterium bifidum is due to their bile salt deconjugating activity. Appl Environ Microbiol 59:1120-1124 21. Klinkenberg G, Lystad KQ, Levine TDW, Dyrset N (2001) Cell release from alginate immobilized Lactococcus lactis ssp. lactis in chitosan and alginate coated beads. J Dairy Sci 84:1118-1127 22. Krasaekoopt W, Bhandari B, Deeth H (2004) The influence of coating materials on some properties of alginate beads and survivability of microencapsulated probiotic bacteria. Int Dairy J 14:737-743 23. Krasaekoopt W, Bhandari B, Deeth HC (2006) Survival of probiotics encapsulated in chitosan-coated alginate beads in yoghurt from UHTand conventionally treated milk during storage. LWT-Food Sci Technol 39:177-183

PROTECTIVE ENCAPSULATION OF PROBIOTIC L. REUTERI

24. Lee JS, Cha DS, Park HJ (2004) Survival of freeze-dried Lactobacillus bulgaricus KFRI 673 in chitosan-coated calcium alginate microparticles.

J Agric Food Chem 52:7300-7305 25. Lee KY, Heo TR (2000) Survival of Bifidobacterium longum immobilized in calcium alginate beads in simulated gastric juices and bile salt solution.

Appl Environ Microbiol 66:869-873 26. Lee YK, Lim CY, Teng WL, Ouwehand AC, Tuomola EM, Salminen S (2000) Quantitative approach in the study of adhesion of lactic acid bacteria to intestinal cells and their competition with enterobacteria.

Appl Environ Microbiol 66:3692-3697 27. Lee YK, Puong KY (2002) Competition for adhesion between probiotics and human gastrointestinal pathogens in presence of carbohydrate. Br J Nutr 88:S101-S108 28. Lee YK, Puong KY, Ouwehand AC, Salminen S (2003) Displacement of bacterial pathogens from mucus and Caco-2 cell surface by lactobacilli.

J Med Microbiol 52:925-930 29. Li XJ, Yue LY, Guan XF, Qiao SY (2008) The adhesion of putative probiotic lactobacilli to cultured epithelial cells and porcine intestinal mucus. J Appl Microbiol 104:1082-1091 30. Mainville I, Arcand Y, Farnworth ER (2005) A dynamic model that simulates the human upper gastrointestinal tract for the study of probiotics. Int J Food Microbiol 99:287-296 31. Muthukumarasamy P, Allan-Wojtas P, Holley RA (2006) Stability of Lactobacillus reuteri in different types of microcapsules. J Food Sci 71:M20-M24 32. Noh DO, Gilliland SE (1993) Influence of bile on cellular integrity and β-galactosidase activity of Lactobacillus acidophilus. J Dairy Sci 76:1253-1259 33. Ouwehand AC, Salminen SJ (1998) The health effects of cultured milk products with viable and nonviable bacteria. Int J Food Microbiol 8:749-758 34. Ouwerx C, Velings N, Mestdagh MM, Axelos MAV (1998) Physicochemical properties and rheology of alginate gel beads formed with various divalent cations. Polym Gels Networks 6:393-408 35. Picot A, Lacroix C (2004) Encapsulation of bifidobacteria in whey protein-based microcapsules and survival in simulated gastrointestinal conditions and in yoghurt. Int Dairy J 14:505-515 36. Pinto M, Robine-Leon S, Appay MD, Kedinger M, Triadou N, Dussaulx E, Lacroix B, Simon-Assmann P, Haffen K, Fogh J, Zweibaum A (1983) Enterocyte-like differentiation and polarization of the human colon carcinoma cell line Caco-2 in culture. Biol Cell 47:323-333

Int. Microbiol. Vol. 18, 2015

69

37. Pouchart P, Marteau P, Bouhnik Y, Goderel I, Bourlioux P, Rambaud JC (1992) Survival of bifidobacteria ingested via fermented milk during their passage through the human small intestine: an in vivo study using intestinal perfusion. Am J Clin Nutr 55:78-80 38. Restall RA (2004) Bacteria in the gut: friends and foes and how to alter the balance. J Nutr 134 (8 suppl.):2022s-2026s 39. Rojas M, Ascencio F, Conway PL (2002) Purification and characterization of a surface protein from Lactobacillus fermentum 104R that binds to porcine small intestinal mucus and gastric mucin. Appl Environ Microbiol 68:2330-2336 40. Saarela M, Lähteenmäki L, Crittenden R, Salminen S, Mattila-Sandholm T (2002) Gut bacteria and health foods- the European perspective. Int J Food Microbiol 78:99-117 41. Scott CD (1987) Immobilized cells: a review of recent literature. Enzyme Microb Technol 9:66-73 42. Servin AL, Coconnier MH (2003) Adhesion of probiotic strains to the intestinal mucosa and interaction with pathogens. Best Pract Res Clin Gastroenterol 17:741-754 43. Shah NP (2000) Probiotic bacteria: Selective enumeration and survival in dairy foods. J Dairy Sci 83:894-910 44. Tang P, Foubister V, Pucciarelli MG, Finlay BB (1993) Methods to study bacterial invasion. J Microbiol Methods 18:227-240 45. Tsen JH, Lin YP, King VAE (2004) Fermentation of banana media by using κ-carrageenan immobilized Lactobacillus acidophilus. Int J Food Microbiol 91:215-220 46. Tuomola EM, Ouwehand AC, Salminen SJ (2000) Chemical, physical and enzymatic pre-treatments of probiotic lactobacilli alter their adhesion to human intestinal mucus glycoproteins. Int J Food Microbiol 60:75-81 47. Valeur N, Engel P, Carbajal N, Connolly E, Ladefoged K (2004) Colonization and immunomodulation by Lactobacillus reuteri ATCC 55730 in the human gastrointestinal tract. Appl Environ Microbiol 70:1176-1181 48. Weizman Z., Alsheikh A (2006) Safety and tolerance of a probiotic formula in early infancy comparing two probiotic agents: a pilot study. J Am Coll Nutr 25:415-419 49. Zhang Y, Zhang L, Du M, Yi H, Guo C, Tuo Y, Han X, Li J, Zhang L, Yang L, (2011) Antimicrobial activity against Shigella sonnei and probiotic properties of wild lactobacilli from fermented food. Microbiol Res 167:27-31

Instructions for authors Preparation of manuscripts General information Research articles and research reviews should not exceed 12 pages, including tables and figures. The text should be typed in 12-point, Times New Roman font, with one and a half line spacing, left justification, and no line numbering. All pages must be numbered consecutively, starting with the tile page.

The Title page should comprise: title of the manuscript, first name and surname and affiliation (department, university, city, state/province, and country) for all authors. The address, telephone and fax numbers, and e-mail address of the corresponding author should also be included.

The Summary should be informative and completely comprehensible, briefly present the topic, state the scope of the experiments, indicate significant data, and point out major findings and conclusions. It should not exceed 200 words. Standard nomenclature should be used and abbreviations should be avoided or defined. No references should be cited. Immediately following the Summary, up to five Keywords should be provided; these will be used for indexing purposes.

The Introduction should be concise and define the objectives of the work in relation to other work done in the same field. It should not give an exhaustive review of the literature.

Materials and methods should provide sufficient detail to allow the experiments to be reproduced. However, only truly new procedures should be described in detail; previously published procedures should be cited, and important modifications of published procedures should be mentioned briefly. The suppliers of chemicals and equipment should be indicated if this might affect the results. Subheadings may be used. Statistical techniques used must be specified.

Results should be presented with clarity and precision. The results should be written in the past tense when describing findings in the author’s experiments. Previously published findings should be written in the present tense.

Results should be explained, but largely without referring to the literature.

The Discussion should be confined to interpretation of the results (not to recapitulating them), also in light of the pertinent literature on the subject.

When appropriate, the Results and Discussion sections can be combined.

This will be the case in research notes.

Acknowledgements should be presented after the Discussion section.

Personal acknowledgements should only be made with the permission of the person(s) named.

Competing interests should be declared by authors at submission indicating whether they have any financial, personal, or professional interests that could be construed to have influenced their paper.

References should be listed and numbered in alphabetical order. In the text, citations should be indicated by the reference number in square brackets. The list of references should include only works that are cited in the text and that have been published or accepted for publication. Unpublished work in preparation, Ph.D. and Masters theses, etc., should be mentioned in the text only, in parentheses. The author(s) must obtain written permission for the citation of a personal communication or other’s researchers’ unpublished results. References cited in the text should be numbered and placed within square brackets, referring to an alphabetized list at the end of the paper.

References should be in the following style: Published papers Venugopalan VP, Kuehn A, Hausner M, Springael D, Wilderer PA, Wuertz

70

S (2005) Architecture of a nascent Sphingomonas sp. biofilm under varied hydrodynamic conditions. Appl Environ Microbiol 71:2677-2686 Books Miller JH (1972) Experiments in molecular genetics. 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA Book chapters Lo N, Eggleton P (2011) Termite phylogenetics and co-cladogenesis with symbionts. In: Bignell DE, Yves R, Nathan L (eds) Biology of termites: a modern synthesis, 2nd ed. Springer, Heidelberg, Germany, pp.27-50 Please list the first eight authors and then add “et al.” if there are additional authors. Citation of articles that have appeared in electronic journals is allowed if access to them is unlimited and their URL or DOI number to the full-text article is supplied.

Tables and Figures should be restricted to the minimum needed to clarify the text; a total number (F + T) of five is recommended. Neither tables nor figures should be used to present results that can be described with a short statement in the text. They also must not be integrated into the text. Figure legends must be typed double-spaced on a separate page and appended to the text. Photographs should be well contrasted and not exceed the printing area (17.6 × 23.6 cm). Magnification of micrographs should be shown by a bar marker. For color illustrations, the authors will be expected to pay the extra costs of 600.00 € per article. Color figures may be accepted for use on the cover of the issue in which the paper will appear. Tables must be numbered consecutively with Arabic numerals and submitted separately from the text at the end of the paper. Tables may be edited to permit more compact typesetting. The publisher reserves the right to reduce or enlarge figures and tables.

Electronic Supporting Information (SI) such as supplemental figures, tables, videos, micrographs, etc. may be published as additional materials, when details are too voluminous to appear in the printed version.

SI is referred to in the article’s text and is ported on the journal’s website (www.im.microbios.org) at the time of publication.

Abbreviations and units should follow the recommendations of the IUPAC-IUB Commission. Information can be obtained at: http://www.chem.qmw.ac.uk/iupac/.

Common abbreviations such as cDNA, NADH and PCR need not to be defined. Non-standard abbreviation should be defined at first mention in the Summary and again in the main body of the text and used consistently thereafter.SI units should be used throughout.

For Nomenclature of organisms genus and species names must be in italics. Each genus should be written out in full in the title and at first mention in the text. Thereafter, the genus may be abbreviated, provided there is no danger of confusion with other genera discussed in the paper. Bacterial names should follow the instructions to authors of the International Journal of Systematic and Evolutionary Microbiology. Nomenclature of protists should follow the Handbook of Protoctista (Jones and Bartlett, Boston).

Outline of the Editorial Process Peer-Review Process All submitted manuscripts judged potentially suitable for the journal are formally peer reviewed. Manuscripts are evaluated by a minimum of two and a maximum of five external reviewers working in the paper’s specific area. Reviewers submit their reports on the manuscripts along with their recommendation and the journal’s editors will then make a decision based on the reviewers.

Acceptance, article preparation, and proofs Once an article has been accepted for publication, manuscripts are thoroughly revised, formatted, copy-edited, and typeset. PDF proofs are generated so that the authors can approve the final article. Only typesetting errors should be corrected at this stage. Corrections of errors that were present in the original manuscript will be subject to additional charges. Corrected page proofs must be returned by the date requested.