AGROS 2021

ANO MMXXI

E depois do IS A...?

Editorial

Cada vez mais alunos desistem do ensino superior. Segundo um estudo da Dire ção-Geral de Estatística da Educação e Ciência (DGEEC), cerca de 29% dos estu dantes abandonam os respetivos cursos e apenas metade dos alunos terminam o curso nos anos estipulados. No ISA, cerca de 23,03% dos alunos ou abandonam a licen ciatura após o primeiro ano ou mudam de curso, existindo licenciaturas no ISA onde essa percentagem atinge os 37,8%, segundo a DGEEC.

Existem várias causas para este problema, como a falta de aulas práticas, desmo tivação e pressão social, mas também a falta de informação sobre o curso e sobre as áreas de trabalho da respetiva área. Foi neste tema que focámos a AGROS deste ano, para que os nossos estudantes possam ter um “cheirinho” daquela que pode ser a sua vida pós ISA.

Também a falta de informação no ensino secundário é um fator decisivo que influên cia o número de desistências no ensino superior, muitas das licenciaturas nem sequer são faladas ao longo do percurso académico, muitas delas essenciais como são as que temos no ISA que sustentam o nosso país e que durante este tempo de pandemia foram prova do quão fulcrais são os trabalhos na área do setor primário.

Que esta 101.ª edição da AGROS seja um auxílio a todos os alunos que se sintam des motivados ou desinteressados no seu curso, para que compreendam melhor a sua área de estudo e futura área de trabalho.

Assim sendo, fica aqui a centésima pri meira edição da AGROS: “E depois do ISA...? O mundo de trabalho visto por ex-alunos", que esta revista sirva de motivação não só para nossos novos alunos, mas também para todos os que se sentem deslocados no seu trabalho, para os alunos que se sentem desamparados no seu curso e para aqueles que ainda não se decidiram.

Uma boa leitura a todos.

Que esta 101.ª edição da AGROS seja um auxílio a todos os alunos que se sintam desmotivados ou desinteressados no seu curso, para que compreendam melhor a sua área de estudo e futura área de trabalho.

Por João Adriano Tavares Coordenador do Apoio ao Estudante

Alimentação. A urgência da inovação Paulo Marques A Floresta e as alterações climáticas Catarina Tonelo

Qual é o sector? Diga-me um! Filipe Figueira O maneio e bem estar animal nos ovinos Sara Fernandes Restoration, Habitat Creation, and Placemaking: The Mangrove

ALIMENTAÇÃO A urgência da inovação

Ainda é relativa mente comum não ser reconhecida a motivação dos jovens para con tribuir para algo maior do que meros interesses egoístas. Trata-se de uma posição injusta, questionada pela frequência dos exemplos que vamos tendo de iniciativas tomadas por jovens relativas a crises societais, como a urgência do problema das alterações climáticas ou as desi gualdades no acesso a uma alimentação saudável. Pelo contrário, estudos mostram que a preocupação dos jovens com as gerações futuras e com o seu pró prio legado é um fator de motivação pelo menos tão importante quanto para os adultos.

Os jovens, quando podem optar entre alternativas de emprego, estão a dar uma ponderação crescente ao sentido que lhes faz o trabalho que irão realizar

– vai ou não contribuir para um mundo melhor?

No caso do agroalimentar não parece difícil encontrarem problemas / desafios para darem o seu contributo para um mundo melhor. No alimentar, o “melhor para a sua saúde e melhor para o planeta” é cada vez mais uma condição de sucesso comer cial nos mercados mais desenvolvidos. Estas motivações de saúde pessoal / saúde do plane ta que começam a generalizar-se na população, são também motivações dos profissionais do agroali mentar – dão sentido ao nosso trabalho.

Deixo-vos de seguida algumas notas sobre desa fios societais ligados ao agroalimentar e sobre o enorme esforço de inovação indispensável para os ultrapassar.

DESAFIOS SOCIETAIS

O mundo confronta-se com vários desafios da maior relevância. Por exemplo, evitar uma nova pandemia ou evitar uma guerra biológica, escolhendo dois para que podemos ter sensibilidade, nós que temos como base as ciências da vida. Alguns desafios têm mais a ver connosco do que outros. De entre aqueles em que podemos contribuir mais diretamente para as solu ções, destaco:

Acesso a alimentação saudável

Ex: disponibilizar produtos de maior densidade nutricional a preços mais acessíveis (a uma população que ainda deverá crescer cerca de 30%)

Sustentabilidade dos sistemas de saúde Ex: relevância da alimentação saudável na prevenção da doença (o investimento em medidas preventivas é da maior importância para a sustentabilidade dos sistemas de saúde)

Neutralidade nas emissões / captura de gases de efeito de estufa Ex: impactos das escolhas alimentares, das tecnologias de produção (sobretudo primária) e da proximidade produção / consumo

Preservação da biodiversidade Ex: impactos das práticas agrícolas

Escassez no abastecimento de água Ex: utilização excessiva de água (stress hídrico); emissões poluentes

Impactos ambientais das embalagens alimentares

Ex: redução, reutilização e reciclagem / circularização de embalagens

têm impactos enormes na solução dos desafios que enfrentamos.

Os novos produtos que a indústria alimentar lança no mercado e os que promove estão ou não a con tribuir para soluções? E as condições que coloca aos seus fornecedores de matérias primas contribuem ou não para uma maior eficiência no uso da terra e da água, para a preservação da biodiversidade, para menores emissões de gases de efeito de estufa?

INVESTIGAÇÃO E DESENVOLVIMENTO

Uma parte importante da inovação na indústria ali mentar é uma resposta à necessidade de conseguir ganhos sistemáticos de eficiência, de modo a que as empresas garantam a sua competitividade de cus tos, sem a qual lhes é difícil concorrer em mercados abertos – as decisões dos consumidores estão muito dependentes do preço. De notar que este foco nos ganhos de eficiência tem vindo a contribuir para a solução de alguns dos desafios ambientais, como a redução das emissões de gases de efeito de estu fa (dependente da eficiência energética) ou a redu ção dos impactos ambientais das embalagens (com a redução do peso / embalagem).

Os consumidores estão cada vez mais conscien tes destes desafios e o retalho alimentar está cada vez mais decidido a fazer o seu papel, sendo que ambos têm um poder crescente nas suas mãos. Também as empresas alimentares, com o seu papel de charnei ra nas fileiras, entre a produção primária e o reta lho, assumem uma grande responsabilidade ao defi nirem as suas estratégias de negócio, cujas opções

A dimensão dos desafios societais exige, porém, uma abordagem dedicada que considere e leve a mudanças em todos os elos da fileira, sem esquecer os consumidores. Muitas soluções (parciais) são já conhecidas – o nosso contributo passa por ajudar a promovê-las e a implementá-las – mas continuamos a precisar de encontrar cada vez melhores soluções. O investimento em investigação e desenvolvimento é essencial na procura de soluções para problemas muito concretos, frequentemente através do cruza mento de conhecimento científico e tecnológico já existente.

A investigação de cariz aplicado é mais rápida, mais barata e tem menor incerteza nos resultados, ou seja, tem uma maior probabilidade de sucesso e efi ciência. Tem frutos mais fáceis de colher, mas deve

Por Paulo Marques

ser complementada com investigação fundamen tal que tenha potencial para, a prazo, trazer ainda mais caminhos e maiores probabilidades de suces so para os projetos dirigidos a solucionar problemas específicos.

Qualquer sistema de ID, por muitos resultados que produza, terá um custo difícil de justificar se não contribuir para a melhoria das condições de vida das pessoas. É preciso que a investigação fundamental traga conhecimento útil para a investigação aplica da e que esta se articule com projetos de desenvolvi mento de processos e produtos. Sem o envolvimento das empresas a ID acabará por não concretizar o seu potencial – a inovação não acontecerá.

de saúde relativamente tangíveis (comparando, por exemplo, com a redução de emissões de gases de efeito de estufa).

O efeito de determinadas alterações alimenta res na redução do risco relativo de certas doen ças, mesmo quando não seja devidamente valori zado pela maioria dos consumidores, não deixa de ter impactos muito positivos nos sistemas de saúde. Tendo estas alterações alimentares um interes se público, justifica-se a existência de programas de incentivo à inovação que permitam avançar rapida mente no lançamento de produtos mais alinhados com elas.

Aos estudantes do ISA são dadas bases mais do que suficientes para participarem na procura de res postas a estas perguntas.

OPORTUNIDADES PARA TODOS

A multiplicidade e a dimensão dos desafios que enfrentamos têm vindo a traduzir-se no fervilhar da inovação alimentar, nacional e internacional, assim como na centralidade da alimentação enquanto tema de discussão nas redes sociais e na programação de conteúdos pelos órgãos de comunicação.

INOVAÇÃO ALIMENTAR E SAÚDE

Muitas empresas alimentares nacionais elegem como vetor estratégico essencial a diferenciação dos produ tos. Muito mais do que tentarem competir pelo menor custo, difícil sem uma escala internacional, procuram a diferenciação por via da origem das suas matérias primas e do valor que conseguem acrescentar a partir do seu conhecimento tecnológico (e, com menor fre quência, científico). De que valor estamos a falar?

O que os consumidores mais valori zam é sem dúvida o prazer, que tem uma importância determinante nas escolhas alimentares (dentro de um preço razoá vel). Há também que considerar fatores higiénicos, muitas vezes inconscientes, de ordem cultural e religiosa, por exemplo.

Uma tendência generalizada dos mer cados, sobretudo dos mais desenvolvi dos, é a valorização crescente da associa ção da alimentação aos seus impactos na saúde. Os consumidores começam a ser particularmente sensíveis aos impactos das suas escolhas alimentares na sua pró pria saúde e na dos que lhe são próximos (enquanto prescritores e não só), no curto como também no longo prazo. Nas esco lhas alimentares estão em causa ganhos

A inovação que responde a requisitos de saúde tende a ter um peso relativamente grande da inter face da ciência & tecnologia, enquanto fonte de ino vação, comparativamente à interface do mercado (clientes e consumidores). É certo que para que a ino vação tenha condições de sucesso terá sempre de ser centrada nos consumidores, respondendo às suas necessidades explícitas e implícitas. Mas, sabendo da sua preocupação crescente com os impactos da ali mentação na saúde, que oportunidades nos traz a interface da ciência & tecnologia? Que problemas ali mentares e que soluções podem ser incorporadas em novos produtos, com sabor e preço ganhadores?

A urgência da inovação alimentar vai continuar a abrir oportunidades para profissionais com perfis pessoais muito distintos –desde os ratos de laboratório da investigação (pacientes e analíticos), até aos mãos na massa da implementação operacional da inovação (práticos e empreendedores).

O diálogo e a definição de parcerias academia / empresas são essenciais para o alinhamento das estratégias e projetos de ID indispensáveis para uma diferenciação dos produtos assente em conhecimen to científico e tecnológico.

Às empresas exige-se muito mais do que inves timento em ID. As empresas têm de ter capacidade para investir valores bem superiores no desenvolvi mento de conceitos de produto, no estudo dos con sumidores, em comunicação de marca. Para além da grande exigência do marketing de novos produtos alimentares, é frequente que o lançamento de uma inovação também exija à empresa o investimento em novas capacidades industriais.

A dimensão das empresas pode ser determinan te para o sucesso da inovação, atendendo à capaci dade de investimento e de execução necessárias, mas mesmo as empresas de menor dimensão têm encon trado espaços por ocupar, beneficiando da sua criati vidade, entusiasmo ou rapidez de decisão.

A urgência da inovação alimentar vai continuar a abrir oportunidades para profissionais com perfis pessoais muito distintos – desde os ratos de labora tório da investigação (pacientes e analíticos), até aos mãos na massa da implementação operacional da inovação (práticos e empreendedores).

Os profissionais formados no ISA têm pois à sua frente, independentemente do perfil de cada um, oportunidades desafiantes e gratificantes.

Paulo Marques, nascido em 1959, frequentou o Instituto Superior de Agronomia desde 1977 a 1982, primeiramente como estudante, tendo continuado posteriormente nos anos de 1982 e 1987, como monitor, engenheiro convidado e assistente, nas disciplinas de Termodinâmica, Química Geral e Análise, Bioquímica e Química Inorgânica. Obteve o Diploma de Aptidão Pedagógica e Capacidade Científica no ISA em 1987 e o grau de Docteur do Institut National Agronomique Paris-Grignon, em 1991. Hoje em dia admite funções de Assessor da Administração da SUMOL+COMPAL, dirigindo equipas de investigação e desenvolvimento.

A FLORESTA E AS ALTERAÇÕES CLIMÁTICAS

Hoje em dia, um engenheiro do ambiente pode intervir em várias áreas de atividade para além da tradicional área do saneamen to público, em grande parte, devido à ideia da econo mia circular. Este conceito está assente num modelo de desenvolvimento económico que pretende valori zar os subprodutos e promove a utilização de mate riais que possam ser reutilizados, recuperados ou reciclados. O valor deste tipo de modelo é inegável, já que permite uma maior eficiência no uso de recur sos e, sendo obviamente importante a nível finan ceiro. No entanto, não foca os problemas a mon tante na cadeia de produção, nomeadamente, a problemática relacionada com o nível insu ficiente de recursos disponíveis para uma população crescente. Assim surge o conceito de bioeconomia, que visa a substituição de recursos não -renováveis por recursos reno váveis, quando possível.

Na última década, fala

-se cada vez mais do valor da floresta portuguesa e explora-se a produção dos produtos florestais não -lenhosos. A floresta já não é só uma fonte de madei ra como matéria-prima para materiais. É também uma fonte de biomassa para produção de energia, e cada vez mais se valoriza a floresta pelos serviços de ecossistema relacionados com o recreio. Embora esta mudança de perspetiva seja relativamente recente, a verdade é que os produtos florestais não-lenhosos, como a cortiça, os cogumelos, a resina, o mel, sem pre representaram uma fonte de rendimento para o proprietário.

A valorização deste tipo de produtos, como qual quer outro, passa muito por conseguir estimar a produção e o seu valor monetário. A cortiça é um dos maiores produtos florestais portugueses e tem ganho relevância como matéria-prima para outras indústrias que não a indústria rolheira. Devido à sua natureza, importância no mercado mundial, e tam bém graças ao trabalho de investigação do grupo FORCHANGE (Centro de Estudos Florestais), é pos

sível obter boas estimativas da produção de corti ça. O mesmo já não se pode dizer de outros produ tos não-lenhosos, como os cogumelos, que para além de serem difíceis de quantificar, são frequentemente alvo de furtos. Na verdade, vários dos produtos não lenhosos têm esta problemática. A cortiça, por exem plo, tem de secar no campo em pilhas para que possa perder humidade e ser comercializada a peso cons tante. Mas não só durante esse tempo está vulnerável, pois em anos recentes, o furto de cortiça na árvore é uma realidade. Tudo isto apesar da legislação para a atividade económica em volta do sobreiro. No caso dos cogumelos, por exemplo, a falta de aplicação e fiscalização da lei não é o problema, mas sim a sua existência.

A tarefa difícil de quantificar os serviços do ecos sistema é conseguida com a ajuda de modelos mate máticos, mas devido ao contexto das alterações cli máticas, essa tarefa é cada vez mais difícil com os modelos florestais tradicionais. Portanto, existe uma necessidade de conseguir explicar e estimar a produ ção florestal com base em parâmetros fisiológicos da árvore e variáveis climáticas. A resposta a esta neces sidade é o desenvolvimento de modelos florestais com base nos processos do ecossistema, um tipo de modelo que não é puramente matemático e usa fato res que permitem prever o fornecimento de produ tos florestais em situações climáticas incertas ou, sim plesmente situações para as quais não existem dados de produção.

Por fim, vale a pena dizer que não há interesse em ter recursos naturais degradados. E ainda, que a indústria madeireira continua a ser uma grande parte do rendimento proveniente da exploração des ses recursos porque é vital para o desenvolvimento económico e da sociedade. A “inovação ambiental” está em conseguir valorizar todos os produtos e ser viços florestais. Contudo, sem um suporte financeiro e legislativo, a conservação da floresta e valorização dos recursos naturais torna-se muito difícil.

Catarina Tonelo, nascida em 1993, ingressou no ISA em 2013 onde permanece até aos dias de hoje, embora tenha concluído o curso em 2019. Frequentou a Tunassa e sente que foi das melhores experiências que podia ter tido no nosso Instituto, para além do facto do mesmo lhe ter aberto várias portas graças à comunidade unida que apresenta. Neste momento é bolseira de doutoramento em Engenharia Florestal, embora tenha feito todo o seu percurso académico em Engenharia do Ambiente. Realizou vários estágios, tais como a Consultadoria a nível de recursos hídricos dos parques públicos para a Câmara de Lisboa, projetos de conservação de ecossistemas marinhos, na Grécia e o impacto do clima no crescimento de árvores através de isótopos, na Finlândia.

Por Catarina Tonelo

QUAL É O SECTOR? DIGA-ME UM!

Por Filipe Figueira

Não sendo um “guru” da motivação, nem nada que se pareça, venho-vos apresentar alguns factos e tendências que a vós, como futuros intervenientes no setor agrícola, poderão fazer subir a moral e aumentar a confiança com que farão a escolha futura de integrar tão nobre ofício, estando já, deseja velmente, a preparar-se para o efeito naquela que é a melhor escola de Ciências Agrárias do País.

Ao contrário do que é a perceção ou “opinião públi ca”, a agricultura não é um setor atrasado, não é um campo perdido no Passado nem é uma profissão estag nada no tempo. A agricultura tem evoluído muito.

Arrisco dizer - fundamentando em seguida - a agri cultura tem sido a que mais tem evoluído entre qual quer outro setor económico (dos que já cá estão há algumas décadas, claro).

A agricultura evolui mais do que todos os outros setores e, ainda assim, consegue continuar a ser vista como um setor composto por um bando de iletra dos e analfabetos funcionais. O “velhinho de enxada” está bem presente no imagético do comum português. Explicar porquê vai para lá da minha compreensão, bem como do escopo deste artigo...

Ora vejamos, com diferentes escalas temporais, diferentes medidas e, principalmente, numa lista não inclusiva de todos os progressos de que a agricultura foi protagonista nos últimos tempos.

PRODUTIVIDADE E ECONOMIA

A tendência de quantos seres humanos um agricul tor alimenta é inatacável: em 1920 eram 19, em 1970 “ainda” eram 26, em 2013 passaram às 155 e, atualmen te, já superam as 200 pessoas. Em média, cada agri

cultor produz alimentos suficientes para alimen tar 200 pessoas! Reflitam sobre estes dados e sobre as mais de 200 pessoas que vão depender do vosso trabalho para (sobre)viverem.

Isto nota-se, a título de exemplo, num aumento médio anual da produtividade do trabalho agrícola em Portugal no período entre 2000 e 2014 de 2,6%!

Claro que, com a degradação dos preços agríco las face aos do resto da economia, uma parte signi ficativa dos ganhos de produtividade foi eliminada do rendimento real dos agricultores, que apenas aumentou 0,6%, em média, anualmente.

De qualquer das formas, o valor acrescentado gera do pelo complexo agro-alimentar, que inclui a agricul tura e as indústrias agro-alimentares, tem apresenta do um crescimento estrutural superior ao conjunto da economia – no período 2000-2014 o VAB em volu me do setor agroalimentar cresceu a uma taxa média anual de 0,5% enquanto o PIB registou uma variação quase nula – 0,1%.

COMÉRCIO ASSOCIADO

Entre 1980 e 2014, isto é, nuns meros 35 anos, o comér cio mundial de produtos agrícolas duplicou em ter mos reais, ultrapassando os €1.5 biliões (€1.5 mil mil milhões para os mais distraídos!).

As vantagens são inúmeras e, democraticamente, tocam-nos a todos, na forma de alimentos mais bara tos, maior diversidade e escolha (o que permite a exis

tência de certas dietas “alternativas” de pessoas que tanto gostam de criticar os próprios agricultores que os alimentam), quebra de sazonalidade, entre tantas outras.

ÁGUA

Neste ponto, reutilizando uma pesquisa que tinha feito para Portugal, de realçar que, na última déca da, o setor agrícola aumentou em 70% a produtivida de da água (água utilizada por unidade de alimento produzido).

O investimento, sempre curto, na modernização e reabilitação dos nossos sistemas de regadio conduzi ram, nas últimas décadas, a uma redução de 50% no consumo unitário de água por hectare regado.

Em 10 anos, o volume total de água utilizada sofreu um forte decréscimo – de 7.500hm3 para 4.200hm3, muito à custa do enorme crescimento da rega sob

Entre 1980 e 2014, isto é, nuns meros 35 anos, o comércio mundial de produtos agrícolas duplicou em termos reais, ultrapassando os €1,5 biliões (€1,5 mil mil milhões para os mais distraídos!).

pressão, com a sua superior eficiência, que passou de 22% em 1999 para 68% do total de regadio em 2009, a par da correspondente perda de expressão da rega por gravidade que decresceu de 78% para 32% no mesmo período.

AMBIENTE

Globalmente, as emissões de gases com efeito de estu fa por unidade de alimento produzida decresceu 39% para as culturas vegetais entre 1970 e 2007. Para a pro dução pecuária reduziu 44% no mesmo período.

Desde 1950, a dose média de produtos fitofarma cêuticos aplicada por hectare reduziu em 95% - sim, parte dessa redução deve-se a maiores concentrações de substância ativa, mas a melhoria é inquestionável – a quantidade de alimentos produzida por cada qui lograma de substância ativa aplicada aumentou mais de 10% desde a década de 1980. Tudo isto enquanto se procedeu a uma redução média de 40% na toxicidade aguda das substâncias ativas introduzidas no mercado desde os anos 60.

Estamos, de facto, já hoje a evoluir do modelo quí mico-mecânico para um de intensificação sustentável (prestar atenção às aulas do Professor Lima Santos, sff).

CONCLUSÃO

No fundo, pode-se afirmar que, felizmente, a agricul tura já não é o que era. Como alguém dizia: vêm-se “as olheiras dos meus professores que já viveram a mudança e o brilho sonhador dos meus colegas que concebem as ideias do futuro. A chamada agricultura convencional, já pouco tem de convencional...”

A amálgama de apontamentos e indicadores que vos dei aqui não são para memorização, nem para os conhecerem detalhadamente – tanto que nem dei as fontes que incluem o GPP, o Prof. Avillez, as Contas Económicas da Agricultura, os Censos, etc. – são sim para que fique firme a imagem de que os agriculto res são uns verdadeiros progressistas, andamos para a frente e, cada vez mais, fazemos mais e melhor!

Tudo isto enquanto é tolerado um rendimen to estruturalmente mais baixo do que as outras ativi dades, sendo que um agricultor europeu aufere em média 40% do rendimento médio europeu.

Já sabemos que os factos não convencem pessoas já convencidas pelas suas próprias ideias e precon

ceitos, porém que isto sirva para que, tu, futuro Enge nheiro Agrónomo, te sintas motivado a não deixar cair o momento, a contribuíres para esta trajetória impa rável de crescimento pois, por muito que te critiquem por seres o que és, lembra-te: vais ter sempre de lhes pôr comida no prato!

PS: Como o ISA é para quem os tem, também lhes podes, à mesma, apontar alguns destes factos ;)

NA NATUREZA DO SEU NEGÓCIO

Filipe Figueira, nascido em 1997, frequentou o Instituto Superior de Agronomia de 2015 a 2020. Tirou o mestrado em Engenharia Agronómica sem especialização, complementado concomitantemente com uma Pós-Graduação em Finanças Corporativas na NOVA SBE. Hoje em dia é Analista Financeiro na Duck River. Realizou diversos estágios ao longo da licenciatura, em Portugal e fora, tendo também participado nas variadas atividades que o nosso Instituto oferece, tais como o Conselho de Escola, o Conselho Pedagógico, a AEISA, a IAAS, o NagroISA, o Futebol e Futsal e ainda, citando o próprio, “(Jantares do) Rugby”. Hoje em dia é um orgulhoso membro da AlumnISA

O MANEIO E BEM ESTAR ANIMAL NOS OVINOS

Nos últimos anos tem sido notável o aumento da procura de carne e leite, os ovinicultores têm recorrido cada vez mai s ao cruzamento de raças, de forma a obter se as suas melhores qua lidades, conforme as necessidades do produtor. Os principais fatores que influenciam o uso do cruza mento em produção animal são a oportunidade de se utilizar o efeito de raça para a obtenção de cara terísticas de maior importância económica e a com binação de genes de diversas fontes, possibilitando a combinação de caraterísticas que não existem nas populações parentais. No cruzamento de raças, é importante selecionar primeiramente as raças pater nas, que fornecerão os reprodutores, e posterior mente as raças maternas. Sendo que existem raças de rápido ritmo de crescimento, mais indicadas para serem paternas, e outras que apresentam melhor fertilidade, além de boa habilidade materna, usual mente s elecionadas para as raças maternas. A esco lha das raças a utilizar nos cruzamentos depende rá da finalidade que o produtor pretende, produção leiteira ou produção de carne. Em Portugal, as raças paternas mais utilizadas nestes cruzamentos, são a Ile de France e a Merino Precoce.

Mais recentemente, no nosso país, a raça Suffolk tem sido vista como uma boa opção para raça paterna.

Na exploração Kicando, em Benavente, onde fiz a minha dissertação, foi realizado o estudo do cruza mento de fêmeas Inra 401 com machos da raça Suf folk, para produção de carne. A raça Inra 401 apre senta boas caraterísticas maternas, enquanto a raça Suffolk é caraterizada pela sua boa qualidade de carne. O intuito deste cruzamento seria obter um maior número de borregos à nascença que adquiris sem as melhores caraterísticas de um Suffolk.

Como é de esperar em qualquer cruzamento, consegue se borregos com pesos intermédios entre os caraterísticos de cada linhagem. Por este motivo, cada vez mais são feitos estudos de diferentes cruza mentos e também, é cada vez mais importante a sua realização tirando proveito das melhores qualidades de cada linhagem.

Os resultados deste estudo podem ser vistos ao detalhe em: Fernandes SV, Bengala Freire JP, Almei da AM (2020). Avaliar o crescimento de borregos. Ruminantes Out/Nov/Dez 2020: 12 16.

Visto que, existe cada vez mais a necessidade de aumentar a produção de carne e leite, de forma, a responder à sua grande procura. Uma das solu ções propostas passa pela substituição da produção extensiva por intensiva, quando possível. Assim, os animais estão confinados em instalações apropria das, facilitando a sua alimentação de acordo com a fase de crescimento em que se encontram.

Existem outros fatores importantes a ter em con sideração, como, o maneio alimentar, produtivo e reprodutivo. O maneio alimentar, tanto em ovi nos como em qualquer animal, deve suprir as suas necessidades alimentares. Uma alimentação adequa da requer uma estimativa das necessidades nutri cionais e da ingestão voluntária dos animais junta mente com uma estimativa do valor nutritivo dos alimentos fornecidos. Os alimentos mais utilizados

Visto que, existe cada vez mais a necessidade de aumentar a produção de carne e leite, de forma, a responder à sua grande procura. Uma das soluções propostas passa pela substituição da produção extensiva por intensiva, quando possível.

na espécie ovina são: pastagens, alimentos compos tos, fenos, palhas e silagens. O tamanho dos bebe douros e comedouros, assim como a higiene destes e as condições ambientais em que se encontram care cem igualmente de importância, a fim de se garan tir um bem estar animal e as maiores produtividades associadas ao mesmo. Os animais devem ser alimen tados em intervalos adequados às suas necessidades fisiológicas. É necessário que existia também um for necimento diário e suficiente de água fresca a todos os animais e que esta se encontre sempre dispo nível, especialmente durante a lac tação. A alimentação dos animais jovens é de extrema importância em explorações que criam os seus próprios animais de substituição. A este tipo de animais, deve ser for necida uma alimentação adequada gara ntindo que os animais tenham um bom desenvolvimento, uma taxa de fertilização elevada, uma longa vida útil, sejam produtivos e igualmente saudáveis.

Por Sara Fernandes

Figura 1 : Ovelha da raça INRA 401 com cria.

Figura 2: Ovelha da raça Suffolk com cria.

Como se pode observar na Fig. 3., a exploração Kicando, para produção de carne, apresenta um ovil com caraterísticas distintas do maneio tradicional, usualmente encontrado neste género de exploração. Possui um corredor central, onde é distribuído o ali mento fornecido aos animais, com auxílio de um tra tor. O controle da alimentação é realizado através de um sistema de semi contenção, como se pode obser var na fig 3. Desta forma, os animais têm acesso ao alimento de igual forma e é possível uma melhor monitorização. Outro aspeto positivo e importante, é que neste caso, enquanto os animais se alimentam, são facilmente observados, permitindo depreender se encontram se bem.

A Fertiland, para produção leiteira, apresenta um ovil com as características tradicionais, onde cada parque possui um comedouro, fig 4. Por vezes, neste caso, quando o alimento é fornecido nem todos os animais têm acesso de imediato. Ao contrário do que se encontra num ovil como, na fig 3., há um menor controle do que cada animal ingere.

Relativamente ao maneio reprodutivo, a cobri ção natural é o método mais utilizado em Portu gal, sendo o mais simples e prático. É importante, que haja uma sincronização do estro nas ovelhas, de forma a conseguir se uma concentração de par tos, nas épocas desejadas. Por norma, os animais são dividido em grupos de seis fêmeas para um macho,

como se pode ver na fig 5., embora varie de explora ção para exploração.

Na Kicando, as fêmeas gestantes são colocadas num parque até parirem. Após o nascimento dos borregos são colocadas com as crias noutro parque. Mesmo depois dos borregos já se alimentarem de ração e água, normalmente, permanecem junto das mães até irem para abate.

No caso da Fertiland, onde também estagiei, e apesar de também se tratar de uma exploração semi intensiva, muitas das ovelhas pariam no campo, e só depois eram transportadas para um parque onde ficam com os borregos nas suas primeiras 24 horas de vida. Visto que a imunidade das crias é adquiri da com a ingestão do colostro, este período é bastan te importante para a sobrevivência dos borregos. No dia seguinte, a cria é retirada da mãe e colocada nou tro parque, onde serão alimentados com um biberão durante um dia. No terceiro dia de vida, são transfe ridos para outro parque, onde são ensinados a ali mentarem se na máquina de amamentação artificial.

Posteriormente, são transferidos para outro par que onde têm ao seu dispor ração e água.

Embora, sejam exemplos de produção distin tas, cada exploração funciona de maneira diferente e deve adaptar se conforme as suas necessidades e recursos, com vista no seu crescimento e, acima de tudo, proporcionar a todos os animais as melhores condições e cuidados. O maneio produtivo é igual mente importante em qualquer exploração. Desde a vacinação e desparasitação dos animais, à pesa gem dos borregos, de forma a haver um controlo de crescimento, à limpeza das camas, à observação dos animais e ao tratamento dos que possam estar doentes até ao corte e tratamento de unhas. Todas as tarefas associadas ao maneiro produtivo são imprescindíveis.

Cada vez mais, nos dias de hoje se fala no bem estar animal e são defendidos os direitos dos ani mais. Nós zootécnicos, queremos acima de tudo o bem estar do animal e proporcionar lhe as melhores condições. Acredito que os ovinicultores trabalham com o mesmo objetivo, animais saudáveis têm uma melhor produtividade e isso tornará uma exploração mais rentável.

Sara Fernandes, nascida a 1992, frequentou o Instituto Superior de Agronomia de 2011 a 2020. Sente que o ISA lhe abriu várias portas, tanto através do estágio obrigatório na Fertiprado que necessitou de fazer na licenciatura, como através de todas as pessoas que conheceu ao longo do seu percurso académico. Acabou por tirar o mestrado em Engenharia Agronómica, com especialidade em Agropecuária e neste momento encontra-se a trabalhar na Syngenta.

Figura 3: Ovil da exploração Kicando, Benavente

Figura 5 : Divisão dos grupos de animais para cobrições

Figura 4: Ovil da exploração Fertiland, Vaiamonte

RESTORATION, HABITAT CREATION, AND PLACEMAKING: THE MANGROVE MUSEUM MOUNTAIN TO SEA ECOLOGICAL CORRIDOR

Por Michael Blier

INTRODUCTION

Project Context

Located on the east bank of the Pearl River estuary in the Guangdong Province, the booming metrop olis of Shenzhen is an essential link between Hong Kong and Mainland China. Since 1980 it has been one of the fastest-growing cities in the world. Once a small fishing community comprised of a series of waterside villages, the city has mushroomed to a population of over 20 million people stretching more than 15km west to east. Consequently, Shenz hen has become one of the world's most influential financial centers.

Pre-existing Site Morphology

In its former natural state, the area where Shenzhen presently sits had been defined by a contiguous top

ographic and ecological gradient rising from the rich Pearl River to the mountainous region (Antuo Moun tain) along the northern boundary of the city. The continuity of surface from low to high as part of the larger, regional morphology resulted in a series of robust transitional ecologies moving from Mountain to Sea: a multi-layered, overlapping, resilient, and diverse ecological network.

Project Objective

The Mountain to Sea Ecological Corridor Framework Plan is a joint initiative between the National Forestry and Grassland Administration, the Shenzhen Munic ipal People's Government, and the adjacent Man grove Wetland Museum. The Mountain to Sea study

came about as the city has entered a period of ecolog ical retrospection. Shenzhen is a stunning contempo rary City. However, only in recent years has the city begun to think more comprehensively and collective ly about the role of ecology as a primary ingredient of resilient urbanism. A careful review of the expan sion of urban growth over time reveals a wholesale ecological disconnect between the Mountains' nat ural resources and the resource area of the Bay and

Pearl River basin. To rectify this, the City Government has identified seven potential, though latent, corri dors within the urban fabric where reconnection of mountain and sea could be realized. The City Govern ment then designated the central corridor as a case study (one of eight) where basic principles for ecolog ical and cultural reconnection are established, tested, and described within the context of the attached Eco logical Framework Plan.

OBSERVABLE EXISTING CONDITIONS

Parcelization

To address the impacts of rapid urbanization, histor ic planning efforts focused primarily on transporta tion, largely defined, at least originally, by the private automobile and resulted in the creation of a series of thoroughfares and highways (elevated, surface, and depressed) connecting the cities western neigh borhoods to the eastern side. The multiple arteries are more or less parallel to one another and gener ally run parallel to the waterfront to the south. A sec ond roadway network establishes a cadence of north/ south connections through the urban fabric. While the crisscrossing pattern of roadways achieved mean ingful urban connections, they also had the nega tive effect of chopping into pieces what had once been a vibrant and interconnected ecology linking the Mountain to the Sea. Disconnected and isolated, the remaining ecological patches cling to life as they have been left on their own to respond to the harmful aspects of extreme and rapid urbanization.

Loss of Biodiversity/Habitat

The focus for the landscapes across many stages of urbanization had more to do with establishing the horticultural, social, and visual attributes to devel op a dynamic visual image for the city and provide support for various cultural sites and festivals. The

landscape development was less interested in eco logical performativity associated with more diverse, local, and native plant communities. Consequent ly, Shenzhen's constructed landscapes rely heavily on non-native species and ornamentals. Often, these imported species provide little in the way of ecologi cal compensation for the loss of the original environ mental contexts.

The loss of more diverse and layered plant com munities has led to the loss of requisite habitat for the region's many species of animals and birdlife. Addi tionally, Shenzhen is also a world-famous landing spot for regional migratory birds. Urbanization has led to a reduced area for migratory bird landing. Fur ther, associated urban noise and light pollution have negatively impacted local nesting and migratory bird landing. With the loss of large swathes of forested area, a significant increase of the ambient air temper ature at the ground level and the worsening air quali ty, exasperated by congestion, have further negative ly impacted area species habitat.

Flooding and Stormwater Management

The rapid transformation of forests and farmlands into vast expanses of impervious networks of road ways and rooftops has created unprecedented epi sodes of seasonal flooding. With considerably less

available land and vegetation to absorb stormwa ter and with little in the way of an integrated man agement strategy to address such rapid growth, the issue of flooding has become chronic, negatively impacting city infrastructural systems and providing no ecological benefit.

Large Expanses of Centralized Urban Infrastructures

The designated Study Area for the Framework Plan is centrally located within the city's urban fabric. As it happens, within this same centralized location is housed a very active railyard and a 25-acre podium structure under which the city's subway car stor age and repair facility operate. Each entity is highly secure, noisy, and to some extent polluting, yet over which the new Mountain to Sea corridor must pass.

THE APPROACH

Classification of Existing Landscape Typologies: The Biomes

Careful identification and then analysis of the Case Study Area revealed five distinct, adjacent ecolog ical zones operating to their own degree of ecologi cal success – some better than others and a few, not at all. Each Biome was identified and cataloged, each with its particular characteristics, benefits, and vul

nerabilities. Because the various Biomes are relics of a continuous ecological corridor that once exist ed and were fragmented due to urbanization, each Biome lacked the benefit of overlap with the adja cent Biome. While the team believed that each Biome shared some DNA with the adjacent, it was under stood that the critical transitional ecologies that will ultimately serve to strengthen the health of individu al Biomes had, for the most part, been lost.

Classification and Location of Transitional Landscape Typologies: The Ecotones

To enhance overlap and continuity within a land scape of transition from Mountain to Bay, special 'connectors' were established to provide elongat ed ecological experiences comprised of elements from each adjacent Biome with enough overlap such that proper thickening over time of local, adaptive ecologies would ensure a more sturdy and resilient cord between the two Biomes, but in a larger sense, across the entire transect. In this case, the Ecotones are interpreted as vegetated, earthen bridges span ning the highways to foster more beneficial physi cal connections, which enable the passage of small mammals and people alike. The Ecotone also con tributes to the continuity of the tree canopy, fur ther enhancing bird habitat and other benefits of the urban tree system.

The team assessed and assigned each Biome and Ecotone with a particular set of qualifications; mate rials, forest scenarios, plant selections, and the extent to which cultural interplay could integrate within and between each area. In this way, a more seamless and resilient ecological and cultural corridor would emerge, one that allows for more balanced urbanism for both people, ecologies, animals, and, of course, the birds. This adaptable process can be applied to other similar corridors throughout The City. Though each Biome is defined by careful research of local conditions and adjacencies, each new corridor will have its own unique look, feel and experience.

Team Structure

The multidisciplinary team and extended consult ant team is comprised of architects, urbanists, ecol ogists, cultural Historians, transportation engineers, arborists, structural engineers, and civil engineers. The team is global and consisting of practitioners and academics.

Our work evolved along a line of (pre-COVID) scheduled workshops and Design Team-led Semi nars. Various experts were invited to attend a series of presentations over a few days and offer input back to The City Government, the larger design team. This was a beneficial interaction process as it provided real-time feedback from highly qualified specialists from around the world to corroborate and challenge.

For more information about this project or others please contact Adrienne Allard, aallard@landworksstudio.com or visitor our website www.landworksstudio.com

About Landworks Studio, Inc.

Landworks Studio is an award-winning, internationally recognized design firm with over two decades of professional practice located just north of Boston in Salem, Massachusetts. As an interdisciplinary team of designers, artists, and innovators, we deliver a boutique experience, drawing upon our collective expertise in landscape architecture, architecture, forestry, furniture design, urban planning, and creative placemaking. Rooted in investigative practices such as hand-sketching, modeling, material and fabrication testing, Landworks’ production method ensures no element is overlooked and enables our team to generate concepts in a timely and consistent manner throughout the design process. We examine all possibilities and outcomes from diverse perspectives, ensuring our designs are not only innovative but constructible.

At Landworks Studio, we are profoundly interested in the relationship between ecological and cultural sustainability as a mechanism for establishing lasting, meaningful, and vibrant new landscapes. Our projects develop through a careful weave between program use, existing ecological systems, and responsible planting palettes. When approaching a new site, we always consider its current and future use, how it will evolve, and in which ways the landscape can integrate and extend into the urban environment.

A FLORESTA E SETOR FLORESTAL: IMPORTÂNCIA ESTRATÉGICA NUM CONTEXTO DE MUDANÇA

IMPORTÂNCIA ESTRATÉGICA

É, hoje em dia, consensual que a floresta e o setor flo restal assumem uma importância estratégica deci siva no desenvolvimento sustentável em Portugal, destacando:

• A floresta enquanto um espaço multifuncional, de elevado valor económico, quer na sua dimen são comercial, quer nos serviços ambientais que presta.

• O significado no Produto Interno Bruto portu guês, comparativamente a outros países do espa ço comunitário, e se, ainda assim, a expressivida de em percentagem possa ser considerada menos significativa, ela é bastante reforçada pela susten tação das fileiras de base florestal em matérias-pri mas nacionais;

• A Matriz vincadamente exportadora de produ tos de valor acrescentado (papel e carão, rolhas

de cortiça, painéis de madeira, castanha, pinhão, alfarroba etc.);

• A Criação de emprego e de polos de animação económica e social em todos os concelhos do continente.

O sector florestal, no seu conjunto, mostra-se dinâ mico e empreendedor, continuando a serem efetua dos investimentos em diferentes áreas, na perspe tiva da sua modernização e de resposta aos novos desafios.

Neste campo há que salientar as novas perspeti vas que o uso da biomassa para energia em sistemas de cogeração ou em centrais dedicadas represen ta, sendo, igualmente, fundamentais para o cum primento dos compromissos de Portugal ao nível internacional.

A dinâmica proactiva, resiliência e adaptabilida de a novos contextos, designadamente no âmbito da promoção da bioeconomia é demostrado em recen tes notícias sobre a fileira da pasta, papel e cartão, referindo a intensão de investimento na vertente de produção de embalagens, pelo expectável aumento da procura deste produto final, face às recentes medi das de política visando a substituição das embalagens de plástico por bioprodutos.

ECONOMIA FLORESTAL

No corrente milénio, o conjunto do setor florestal tem tido um desempenho macroeconómico global positivo, relevando a componente do investimento industrial, das exportações, balança comercial (Figu ra 1 e Quadro 1) como as que mais contribuíram no mesmo (INE, 2021a).

Exportações do setor florestal em percentagem das exportações nacionais Importações do setor florestal em percentagem das importações nacionais

Balança comercial (milhões €) Taxa de cobertura das exportações sobre as importações (%)

Ano Sector Nacional Sector Nacional florestal florestal

2000 1 061 -18 491 153 60

2001 994 -18 701 149 60

2002 1 078 -16 619 153 63

2003 1 389 -15 181 170 66

2004 1 328 -18 340 164 63

2005 1 304 -20 242 163 61

2006 1 532 -20 654 170 63

2007 1 539 -21 632 163 64

2008 1 484 -25 347 161 61

2009 1 321 -19 682 164 62

2010 1 771 -21 379 175 64

2011 1 954 -16 723 181 72

2012 2 456 -11 161 223 80

2013 2 664 -9 710 230 83

2014 2 657 -10 978 219 81

2015 2 800 -10 711 217 82

2016 2 712 -11 385 207 81

2017 2 682 -14 671 197 79

2018 2 758 -17 589 194 77

2019 2 777 -20 074 193 75

2020 2 531 -14 388 198 79

Quadro 1: Balança comercial nacional e do setor florestal

CONSTRANGIMENTOS E DESAFIOS

Porém, as fileiras florestais apresentam ainda fragilidades estruturais que ameaçam a sua sustentabilidade.

Desde logo, o desinvestimento ao nível produção

silvícola é preocupante, até pela importância estraté gica da atividade primária na sustentação e oferta de matéria-prima (madeira, resina, castanha, pinhões, alfarroba, etc.) às indústrias de transformação.

Os desequilíbrios na produção na floresta são demonstrados na análise das estatísticas desenvolvi das no quadro das contas económicas da silvicultu ra, pelo Instituto Nacional da Estatistica (INE, 2021b & ICNF, 2021). Esta informação indica que a produ ção, em valor (euros), de madeira de coníferas para fins industriais tem vindo a decrescer à média anual de -2%, no corrente milénio. Ainda assim nota-se o crescimento médio anual de 3%, já na última déca da. Em sentido inverso, a madeira de folhosas para fins industriais tem verificado crescimento, à média anual de 3% (Quadro 2).

Quadro 2: Estrutura da Produção Florestal entre 2000 e 2019

A análise da produção em volume (metros cúbi cos) de madeira redonda para fins industriais no pre sente século (2000 a 2020), com base na compilação publicada pela FAO (“Food and Agriculture Organi zation of the United Nations”) no sistema FAOSTAT e pela UNECE (“United Nations Economic Commission for Europe”), mostra também (FAO, 2021 & UNECE, 2021):

A redução da produção de madeira redonda de coníferas à taxa média anual de -1%, sendo a variação total de -21%;

O aumento de madeira redonda de folhosas à taxa média anual de 2%, sendo a variação total de 48%.

Já na presente década (2010 a 2020), a madeira redonda para fins indústrias verificou, no seu conjun to, aumentos em volume, embora os mesmos tenham sido superiores na madeira de folhosas (média anual 2%;variação total 39%), e inferiores na madeira de coníferas (média anual 1%;variação total 14%).

ESTRATÉGIAS E DESENVOLVIMENTO

É assim consensual que a aplicação de medidas que mitiguem a degradação do espaço florestal, promo vam o crescimento da produção de bens silvícolas e ganhos de produtividade através da melhoria da ges tão florestal dos povoamentos existentes, e a insta lar de novo, são essenciais à sustentabilidade no con junto do sector florestal. A valorização económica da produção na floresta será, assim, um fator da maior relevância no desenvolvimento equilibrado e susten tável do setor florestal em Portugal. Os novos desa fios e crises resultantes do contexto de mudanças cli máticas, crescimento demográfico, e outros, impõem também urgência e maior eficiência na gestão e utili zação dos recursos naturais.

Em coerência com esse desígnio, a estratégia nacional para as florestas (RCM n. º 6-B/2015) assume a maximização do valor económico total das florestas enquanto objetivo enquadrador primordial às ações para as florestas nacionais, organizando os seus obje tivos estratégicos em: minimização dos riscos (incên dios, pragas, doenças e invasoras lenhosas) e melho ria da produtividade, designadamente através da gestão florestal sustentável.

Ao nível da União Europeia, a nova estratégia para as florestas 2030, iniciativa emblemática do Pacto Ecológico Europeu, apoiando-se na estratégia de bio diversidade da UE para 2030, pretende abordar as múltiplas funcionalidades da floresta e promover a sustentabilidade, resiliência e biodiversidade das flo restas na Europa.

A Estratégia Florestal preconiza também alcançar a meta de redução das emissões de gases com efei to de estufa (GEE) de, pelo menos, 55% em 2030 e a neutralidade climática em 2050. Nesse quadro as flo restas são assumidas como particularmente relevan tes para o compromisso de aumentar as remoções e potenciar os sumidouros naturais de GEE.

REFERÊNCIAS

FAO, 2021, FAOSTAT, Food and Agriculture Orga nization of the United Nations, URL http://www.fao. org/faostat/en/#home

ICNF, 2021, MERCADOS E PRODUÇÃO FLORES TAIS 2020, Instituto da Conservação da Nature za e das Florestas, URL http://www2.icnf.pt/portal/ florestas/fileiras/econ

INE, 2021a, Estatísticas do comércio internacional, Instituto Nacional de Estatísticas, URL: www.ine.pt INE, 2021b, Contas económicas da silvicultura, Ins tituto Nacional de Estatísticas, URL: www.ine.pt

UNECE, 2021, INForest, United Nations Econo mic Commission for Europe, URL https://forest-data. unece.org/

Graça Louro, doutorada em engenharia florestal e dos recursos naturais, ingressou no Instituto Superior de Agronomia em 1980, onde concluiu os seus estudos em 1984. Tirou o mestrado no Instituto Superior Técnico em Georecursos mas acabou por regressar ao ISA em 2008 onde permaneceu até 2012. Hoje em dia trabalha no Instituto da Conservação da Natureza e das Florestas (ICNF).

MITOFUSINS: DISEASE GATEKEEPERS AND HUBS IN MITOCHONDRIAL QUALITY CONTROL BY E3 LIGASES

By Mafalda Escobar-Henriques and Mariana Joaquim

Mitochondria are dynamic organelles engaged in quality control and aging processes. They constantly undergo fusion, fission, transport, and anchoring events, which empower mitochondria with a very interactive behavior. The membrane remodeling processes needed for fusion require conserved proteins named mitofusins, MFN1 and MFN2 in mammals and Fzo1 in yeast. They are the first determinants deciding on whether communication and content exchange between different mitochondrial populations should occur. Importantly, each cell possesses hundreds of mitochondria, with a different severity of mitochondrial mutations or dysfunctional proteins, which potentially spread damage to the entire network. Therefore, the degree of their merging capacity critically influences cellular fitness. In turn, the mitochondrial network rapidly and dramatically changes in response to metabolic and environmental cues. Notably, cancer or obesity conditions, and stress experienced by neurons and cardiomyocytes, for example, triggers the downregulation of mitofusins and thus fragmentation of mitochondria. This places mitofusins upfront in sensing and transmitting stress. In fact, mitofusins are almost entirely exposed to the cytoplasm, a topology suitable for a critical relay point in information exchange between mitochondria and their cellular environment. Consistent with their topology, mitofusins are either activated or repressed by cytosolic post-translational modifiers, mainly by ubiquitin. Ubiquitin is a ubiquitous small protein orchestrating multiple quality control pathways, which is covalently attached to lysine residues in its substrates, or in ubiquitin itself. Importantly, from a chain of events also mediated by E1 and E2 enzymes, E3 ligases perform the ultimate and determinant step in substrate choice. Here, we review the ubiquitin E3 ligases that modify mitofusins. Two mitochondrial E3 enzymes— March5 and MUL1—one ligase located to the ER—Gp78—and finally three cytosolic enzymes—MGRN1, HUWE1, and Parkin—were shown to ubiquitylate mitofusins, in response to a variety of cellular inputs. The respective outcomes on mitochondrial morphology, on contact sites to the endoplasmic reticulum and on destructive processes, like mitophagy or apoptosis, are presented. Ultimately, understanding the mechanisms by which E3 ligases and mitofusins sense and bi-directionally signal mitochondriacytosolic dysfunctions could pave the way for therapeutic approaches in neurodegenerative, cardiovascular, and obesity-linked diseases.

Keywords: E3 ligases, ubiquitin, mitofusins, MFN1/MFN2, mitochondria, quality control, mitophagy, ER

INTRODUCTION

Mitochondria were considered as static and isolat ed bean-shaped organelles for a long time, being labeled “power house of the cell” given the assump tion that ATP production by oxidative phosphoryla tion (OXPHOS) was their main function (Mcbride and Neuspiel, 2006). However, as soon as researchers start ed to look into it by live imaging, it was quickly percep tible the existence of a high dynamism (Bereiter-Hahn and Voth, 1994; Nunnari et al., 1997), later proved to be associated with many new mitochondrial functions (Westermann, 2010). Mitochondria possess proteins that enable plastic responses, depending on the cellu lar conditions, by fusion, fission, and transport process es (Friedman and Nunnari, 2014). Another hallmark in the field was the awareness of the importance of mito chondrial transport and positioning within the cell and thereby interaction with other cellular compartments surrounding it. Pioneering studies unraveling physi cal tethering between mitochondria and the endoplas mic reticulum (ER; Kornmann et al., 2009) paved the way for subsequent discoveries on several other mito chondrial contact sites (Eisenberg- bord and Schuld iner, 2017; Cohen et al., 2018). These contacts coordi nate a continuous communication of mitochondria with other organelles to support important cellular functions. Finally, the functional impact of mitochon drial interaction with soluble components present in the cytoplasm, like ubiquitin, revealed another layer of the integrative behavior of these organelles (Esco bar-Henriques and Langer, 2014; Bragoszewski et al., 2017). Ubiquitylation is a post-translational modifica tion (PTM) that occurs through the addition of a ubiqui tin moiety to substrates (Yau and Rape, 2016; Kwon and Ciechanover, 2017). Ubiquitin is required for many cel lular pathways, and the discovery of its regulatory func tions is constantly increasing (Rape, 2018). Consistently, ubiquitin targets at mitochondria are associated with several distinct and important cellular processes, main ly with cellular quality control functions (Escobar-Hen riques and Langer, 2014).

Here we present the current knowledge on the E3 ligases modifying mitochondrial proteins, focusing on the mitochondrial fusion factors Mitofusin 1 (MFN1) and Mitofusin 2 (MFN2). Mitofusins appear to be pre ferred targets, constituting a cellular hub in response to metabolic needs of the cell (Chan et al., 2011; Sarraf et al., 2013; Bingol et al., 2014).

Ubiquitylation

Ubiquitylation of proteins is one of the cellular PTMs, which allow diversifying the coding capacity of genes by covalent modifications, mostly enzyme-catalyzed, of nascent or folded proteins. Therefore, PTMs create a bigger pool of protein diversity (Walsh et al., 2005). The most common small PTMs are phosphorylation, acetylation, glycosylation, carboxylation, methylation, nitrosylation, and S-glycation, which are characterized by the addition of the respective chemical moieties to proteins (Walsh et al., 2005). Moreover, ubiquitin and ubiquitin-like modifiers constitute a set of addition al PTMs: ubiquitylation, sumoylation, rubylation, lip idation, ISGylation, and FATylation (Cappadocia and Lima, 2017). Interestingly, ubiquitin itself was shown to be post- translationally modified by phosphorylation and acetylation (Herhaus and Dikic, 2015; Swatek and Komander, 2016).

Ubiquitin is a small highly conserved eukaryotic pro tein and ubiquitylation is the process by which ubiqui tin molecules are added to a substrate (Ciechanover, 2005) (Figure 1). It occurs via an enzymatic cascade involving three elements: an E1 ubiquitin-activating enzyme, an E2 ubiquitin-conjugating enzyme, and an E3 ubiquitin ligase. First, the E1 enzyme activates ubiq uitin and transfers it to the E2 enzyme, in an ATP-de pendent manner. Subsequently, the ubiquitin molecule is transferred from the E2 enzyme to a specific target substrate. This requires substrate recognition by an E3 ligase, which either actively receives ubiquitin from the E2 and then covalently binds it to the substrate (HECT, RBR) or serves as a binding platform between the E2 and the substrate (RING) (Komander and Rape, 2012; Yau and Rape, 2016). E3 ligases are of extreme impor tance in this enzymatic cascade, since they select the specific substrates to be modified (Zheng and Shabek, 2017). Importantly, ubiquitylation is a reversible pro cess, where deubiquitylases are able to remove the ubiq uitin moiety from a substrate, resulting in free ubiqui tin (Mevissen and Komander, 2017; Clague et al., 2019). Ubiquitin can be present in substrates in the form of one ubiquitin moiety (mono-ubiquitylation) or several moi eties (multi-monoubiquitylation). Moreover, poly- ubiq uitin chains of different topologies can also form, via any of the seven internal lysine residues in ubiquitin (Lys6, Lys11, Lys27, Lys29, Lys33, Lys48, and Lys63; Komander and Rape, 2012; Yau and Rape, 2016). Due to their differ ent surfaces, these ubiquitin chains attract diverse effec tors, giving origin to a variety of functions (Kwon and Ciechanover, 2017). For example, Lys48-linked chains

Langer, 2014). Here we present the current knowledge on the E3 ligases modifying mitochondrial proteins, focusing on the mitochondrial fusion factors Mitofusin 1 (MFN1) and Mitofusin 2 (MFN2).

DUB

A.

Thioester intermediate B. Direct transfer E2 to Substrate

S

FIGURE 1 | Ubiquitylation cascade. Ubiquitylation of substrates requires a cascade of events involving three enzymes: an E1 ubiquitin activating enzyme, an E2 ubiquitin conjugating enzyme, and an E3 ubiquitin ligase. First in this cascade, the E1 enzyme activates ubiquitin and transfers it to the E2 enzyme in an ATP dependent manner with which ubiquitin is conjugated. Afterward, the ubiquitin molecule is transferred from the E2 enzyme to the specific target substrate by the E3 ligase enzymes, which either actively receives ubiquitin from E2 and then transfers it to the substrate or serves as a binding platform between the E2 and the substrate. Finally, on the target substrate, mono, mono multi, or polyubiquitylation can occur.

FIGURE 1 | Ubiquitylation cascade. Ubiquitylation of substrates requires a cascade of events involving three enzymes: an E1 ubiquitin-activating enzyme, an E2 ubiquitin-conjugating enzyme, and an E3 ubiquitin ligase. First in this cascade, the E1 enzyme activates ubiquitin and transfers it to the E2 enzyme in an ATP- dependent manner with which ubiquitin is conjugated. Afterward, the ubiquitin molecule is transferred from the E2 enzyme to the specific target substrate by the E3 ligase enzymes, which either actively receives ubiquitin from E2 and then transfers it to the substrate or serves as a binding platform between the E2 and the substrate. Finally, on the target substrate, mono, mono-multi, or polyubiquitylation can occur.

are mostly known to mark proteins for proteasomal deg radation via the ubiquitin-proteasome system (UPS), whereas Lys63-linked chains are mainly associated with regulatory functions (Kwon and Ciechanover, 2017).

Mitochondria and Mitofusins

Mitochondria are double membrane organelles com posed by the outer mitochondrial membrane (OMM) and the inner mitochondrial membrane (IMM), which are separated by the intermembrane space (IMS; Figure 2; Pfanner et al., 2019). The IMM encloses the mitochon drial matrix and forms invaginations called cristae (Frey et al., 2002). The OXPHOS system locates along the cris tae and provides the mitochondrial membrane poten tial, necessary for the production of energy in the form of ATP (Nunnari and Suomalainen, 2012). Besides oxi dative phosphorylation, mitochondria perform sever al other important functions, such as phospholipid syn thesis and assembly of iron-sulfur clusters (Braymer and Lill, 2017; Tatsuta and Langer, 2017; Cardenas et al., 2018). In addition, mitochondria are important for cellu lar responses, such as calcium (Ca2+) buffering (Marchi et al., 2017; Xie et al., 2018), mitophagy (Harper et al., 2018; Pickles et al., 2018), and regulation of programmed cell death (Xie et al., 2018; Sedlackova and Korolchuk, 2019). The functional plasticity of mitochondria is intimate ly linked to its morphology (Friedman and Nunnari, 2014; Tilokani et al., 2018). Fusion and fission events are majorly important for the regulation of mitochondrial morphology, whereas mitochondrial transport is of par ticular importance in cells with high-energy demands, such as neurons (Knott and Bossy-Wetzel, 2008; Rakov

ic et al., 2011). The dynamics between mitochondri al fusion and fission may result in several possible morphological outcomes, from a tubular mitochon drial network, sometimes massively interconnected, to several fragments. This plasticity is fundamental for the maintenance of proper mitochondrial func tion and to assist mitochondria in response to sever al stress situations (Liesa and Shirihai, 2013; Schrepfer and Scorrano, 2016; Chen and Chan, 2017). For exam ple, loss of membrane potential and nutrient excess have been shown to induce mitochondrial fragmen tation (Yu et al., 2006), whereas nutrient starvation was shown to shift the balance toward a tubular mito chondrial network (Tondera et al., 2009; Gomes et al., 2011; Rambold et al., 2011). These membrane remodel ing events are mediated by conserved large dynaminlike GTPase proteins (Praefcke and McMahon, 2004). Drp1 is responsible for fission (Dnm1 in yeast), MFN1/ MFN2 for OMM fusion (Fzo1 in yeast), and Opa1 for IMM fusion (Mgm1 in yeast; Youle and van der Bliek, 2012). They are main targets of PTMs, being either activated or repressed in order to push the morphol ogy toward a fused or a fragmented state (Figure 2; Escobar- Henriques and Langer, 2014; Macvicar and Langer, 2016; Mishra and Chan, 2016).

Mitofusins are OMM proteins, with the GTPase domain locating at the N-terminal, followed by one hydrophobic heptad repeat (HR1), the transmem brane anchor(s) and finally possessing a second pro tein-protein interaction domain, HR2 (Figure 3). First, it was proposed that both N- and C-terminus face the cytosol, connected by two transmembrane domains

and a short loop in the IMS (Rojo et al., 2002). This topol ogy is in agreement with fusion- compatible structural information from both MFN1 and the bacterial homo logue BDLP (Low and Löwe, 2006; Low et al., 2009; Qi et al., 2016; Cao et al., 2017). However, an alternative topolo gy for MFN1 and MFN2 was proposed, with a single span ning-membrane domain, instead of two, therefore plac ing the C-terminus in the IMS (Mattie et al., 2018). Further studies will be necessary to elucidate which topolo gy of mitofusins reflects fusion-dependent or perhaps fusion-independent roles of mitofusins. MFN1 and MFN2 proteins are 62% identical and 77% similar to each other (Zorzano and Pich, 2006). Interestingly, despite being extremely similar, depletion of each mitofusin has differ ent effects on mitochondrial morphology. While deple tion of MFN1 leads to highly fragmented mitochondria, in the shape of small fragments, depletion of MFN2 leads to bigger mitochondrial fragments that aggregate into clusters (Figure 4; Chen et al., 2003). Strikingly, homo zygous knockout of either MFN1 or MFN2 in mice was shown to be lethal, with death occurring in midgestation (Chen et al., 2003). Additionally, MFN2 depleted mice presented placental defects within the giant cell layer, with a reduced number of giant cells and a reduced number of nuclei on the few cells still observed. Howev

er, no placental developmental defects were observed in MFN1 mutants. This suggests that MFN1 and MFN2 may have distinct functions, perhaps independent of their roles in mitochondrial fusion (Loiseau et al., 2007; Guil let et al., 2010; Codron et al., 2016; Beręsewicz et al., 2017; El Fissi et al., 2018; Zhou et al., 2019). For example, a cor relation observed between the levels of MFN2 and oxi dative phosphorylation was suggested to be dependent on coenzyme Q deficiency, independently of the fusion capacity of MFN2 (Pich et al., 2005; Segalés et al., 2013; Mourier et al., 2015).

CELLULAR PROCESSES AFFECTED BY UBIQUITYLATION OF MITOFUSINS

Ubiquitylation of both MFN1 and MFN2 has been report ed and associated with diverse cellular processes. First, responses directly affecting mitochondria themselves were described, either by changing their morphology or by extending mitochondrial contacts to the ER. Sec ond, effects on mitophagy or apoptosis are the most described effects. Nonetheless, links of mitofusin ubiq uitylation with hypoxic and genotoxic stress have also been made.

FIGURE 2 | Mitochondrial roles and post-translational modifications of mitochondrial dynamic proteins. Mitochondria are involved in several cellular processes other than ATP production, such as calcium (Ca2+) buffering, phospholipid biosynthesis, iron-sulfur clusters (Fe-S) assembly from iron (Fe2+), regulation of programmed cell death via cytochrome c (Cytc c) release, and mitophagy. Mitochondrial proteins known for their role in mitochondrial dynamics, such as Drp1, Mfn1/2, and Opa1, are a target for several post-translational modifications, which differently regulate their level and, hence, function. Drp1, on the OMM, can be modified by sumoylation, phosphorylation, acetylation, ubiquitylation, and S-nitrosylation. Mitofusins, also on the OMM, can be a target for phosphorylation, ubiquitylation, or acetylation. Opa1, in the IMM, can suffer processing or be modified through acetylation.

FIGURE 3 | Structure and topology models of mitofusins. (A) Linear structure of mitofusin, with the GTPase domain locating at the N-terminal, one hydrophobic heptad repeat (HR1), the transmembrane anchor(s), and a second hydrophobic heptad repeat (HR2). (B) Crystal structure of MFN1 and MFN2 modeled on BDLP and mini-MFN1, according to the first topology proposed, with two transmembrane domains and both the N- and C-terminus facing the cytosol (Rojo et al., 2002; Low and Löwe, 2006; Low et al., 2009; Qi et al., 2016; Cao et al., 2017). (C) Structural scheme of MFN1 and MFN2 according to the second topology proposed with a single spanning-membrane domain, instead of two, and the C-terminus residing in the IMS and not facing the cytosol (Mattie et al., 2018).

Mitochondrial Morphology

Ubiquitylation of mitofusins is promptly observed in yeast, flies, and mammals (Cohen et al., 2008; Ziviani et al., 2010; Rakovic et al., 2011). Interestingly, it plays a dual role in mitochondrial morphology either by addressing mitofusins for proteasomal turnover or by activating mitofusins and therefore promoting mem brane merging. It was first suggested that the steadystate levels of Fzo1 are regulated (Fritz et al., 2003), and it was later shown that the turnover of this protein is proteasome-dependent in response to mating factor (Neutzner and Youle, 2005). Moreover, it was shown that the AAA protein and ubiquitin-selective chaper one VCP/ p97/Cdc48 is required for proteasomal-de pendent degradation of mitofusins (Tanaka et al., 2010; Kim et al., 2013; Zhang et al., 2017). This results in mito chondrial fragmentation, due to ongoing fission events. Moreover, it is associated with stress responses, mediat ed by several E3 ligases, as outlined later. This induces different outcomes, either regulating mitochondria-ER contact sites or affecting mitophagy and apoptosis.

On the other hand, pro-fusion ubiquitylation of Fzo1 occurs constitutively and is tightly controlled by deu

biquitylases and Cdc48 (Anton et al., 2013; Yue et al., 2014; Simões et al., 2018).

ER-Mitochondria Contacts

Mitochondria are responsible for a number of cellu lar and subcellular processes. Consistently, mitochon dria form a dynamic network with several other organ elles. For example, the junctions formed with the ER are known as the mitochondria-associated ER mem brane (MAM) sites and can cover up to 5% of mitochon dria (Wu et al., 2018). These junctions are of extreme importance for several processes such as lipids biosyn thesis, mitochondrial dynamics, and Ca2+ transfer (Fig ure 5; Szabadkai et al., 2006; Friedman et al., 2011; Row land and Voeltz, 2012; Kojima et al., 2016; Wu et al., 2018; Doghman-Bouguerra and Lalli, 2019).

The majority of enzymes necessary for lipid biosyn thesis are found on the ER membrane. However, synthe sis of phosphatidylethanolamine (PE) and phosphati dylcholine (PC), the two most abundant phospholipids, requires lipid trafficking between mitochondria and ER, due to the localization of the required enzymes (Row land and Voeltz, 2012; Kojima et al., 2016; Tatsuta and

Langer, 2017). PE is produced from phosphatidylserine (PS), which is synthetized on the ER membrane. In turn, the enzyme phosphatidylserine decarboxylase, which is responsible for the majority of PE biosynthesis, locates mostly at the IMM of mitochondria (Tatsuta and Langer, 2017; Friedman et al., 2018). Therefore, PS must be trans ferred to mitochondria. In turn, to produce PC, PE must be transferred back to the ER, again requiring lipid trans fer events. In conclusion, the biosynthesis of both PE and PC demonstrates the importance of ER-mitochon dria contact sites in lipid biosynthesis (Figure 5). Interestingly, the ER was shown to be an active reg ulator of mitochondrial dynamics. ER tubules that con tact with mitochondria were found to correlate with the presence of the mitochondrial fission factor Drp1 (Fried man et al., 2011). In agreement with the idea that ER-mi tochondria contacts might regulate mitochondrial divi sion, they correlated with the presence of constricted mitochondria, prior to Drp1 recruitment (Figure 5; Fried man et al., 2011). In addition, MFN2 was suggested to be present at the ER, in MAM sites, directly acting as a teth er between these two organelles (De Brito and Scorra

no, 2008). However, whether ER-mitochondria juxtapo sition is promoted (De Brito and Scorrano, 2008; Naon et al., 2016; Basso et al., 2018; McLelland et al., 2018) or inhibited (Cosson et al., 2012; Filadi et al., 2015; Wang et al., 2015) by MFN2 is controversially discussed, depend ing perhaps on the cellular growth conditions. Interest ingly, a role in ER-mitochondria contacts in inhibiting mitophagy was recently shown (Basso et al., 2018; McLel land et al., 2018).

One of the most well-characterized processes where ER-mitochondria contacts are indispensable is Ca2+ buff ering (Marchi et al., 2017). Ca2+ is transferred from the ER through the inositol 1, 4,5-triphosphate receptor (I3PR) to the voltage- dependent anion-selective chan nel (VDAC) on the OMM (Figure 5; Rizzuto et al., 1998; Szabadkai et al., 2006). In turn, Ca2+ influx to the mito chondrial matrix occurs via the mitochondrial calcium uniporter (MCU; Kirichok et al., 2004; Baughman et al., 2012). The transfer of Ca2+ to mitochondria is required for several mitochondrial proteins or processes, includ ing the TCA cycle enzymes (Bustos et al., 2017). In addi tion, mitochondrial division has been shown to be affect ed by Ca2+ levels, in a Drp1-dependent manner (Friedman et al., 2011). In fact, some factors required for mitochondri al division and found in MAM sites are regulated by Ca2+ binding, as for exam ple MIRO, a protein mainly required for mitochondrial trafficking (Saotome et al., 2008). Finally, Ca2+ transferring at the ER-mitochondria contacts was also shown to activate apoptosis (Dogh man-Bouguerra and Lalli, 2019). Ca2+ influx to mitochondria is able to open the mitochondrial permeability transi tion pore, leading to the release of cyto chrome c and further apoptosis induc tion (Scorrano et al., 2003).

FIGURE 4 | Mitochondrial morphology upon knockout of Mitofusin 1 or 2 and its disease-associated roles. Although extremely similar in their sequence and structure, each mitofusin ablation leads to strikingly different mitochondrial morphologies. Mitofusin 1 knockout gives origin to a highly fragmented mitochondrial network composed of many small fragments, whereas depletion of Mitofusin 2 leads to a network where mitochondrial fragments are found enlarged and aggregated in clusters, commonly perinuclearly organized. Several diseases have been associated with knockout of Mitofusin 1 or 2. Homozygous knockout of Mitofusin 1 or 2 leads to embryonic defects such as defective giant cell layer and leads, ultimately, to lethality. While Mitofusin 1 only appears to have an effect at the embryonic level, knockout of its homologue protein, Mitofusin 2, has been shown to relate to several other defects. Animal models depleted for Mitofusin 2 display severe cardiac defects such as cardiomyocyte dysfunction, rapid progressive dilated cardiomyopathy, and final heart failure. Moreover, Mitofusin 2 mutations are the primary cause of the incurable neuropathy CharcotMarie Tooth Type 2A for which no disease-underlying functions have been yet identified. Additionally, links with other neurodegenerative diseases such as Alzheimer’s and Parkinson’s diseases have been made, although the molecular mechanisms underlying it are not fully understood. Low levels of Mitofusin 2 were also shown to have a strong positive correlation with diabetes type 2 and obesity.

Ca2+ Ca2+ Ca2+ I3PR

Ca2+ VDAC

Ca2+

stood how, both BAX and BAK form pores on the OMM, enabling the release of pro-apoptotic molecules from the IMS to the cytosol (Wang and Youle, 2009). Once in the cytosol, cytochrome c binds to the apoptotic prote ase activating factor 1 (Apaf-1) (Liu et al., 1996), forming the apoptosome. This complex cleaves and activates the pro-caspase 9, followed by the activation of effector caspases (Xiong et al., 2014).

function, the quality control mechanism activated is mitophagy. Mitophagy is a selective form of macroau tophagy that eliminates damaged mitochondrial pro teins, or portions of damaged mitochondrial network. It occurs via their engulfment by autophagosomes, which subsequently fuse with the lysosome, where degrada tion occurs (Figure 7).

Mitochondrial Fission Phospholipids Synthesis

FIGURE 5 Roles of mitochondria ER contacts. The physical contacts established between mitochondria and ER are responsible for several cellular processes such as mitochondrial fission, calcium buffering, and phospholipid synthesis. During mitochondrial fission, ER tubules are found in contact with mitochondria in the future fission sites. Drp1 is recruited to these sites and, together with ER tubules, promotes constriction and fission of mitochondria. Calcium (Ca2+) is transported from the ER to mitochondria via transporters in each membrane. On the ER membrane, Ca2+ is exported via the inositol 1,4,5 triphosphate receptor (I3PR) to the voltage dependent anion selective channel (VDAC) on the mitochondrial outer membrane. For the synthesis of some phospholipids, the required enzymes are found in the mitochondria and not in the endoplasmic reticulum (ER). This implies the transfer of precursor forms of phospholipids from the ER to mitochondria where they can be modified and then re transferred to the ER. For example, the production of the mitochondrial phospholipid cardiolipin (CL) depends on precursor lipids present at the ER. Phosphatidic acid (PA) is transferred from the ER, across the IMS to the IMM, where it is enzymatically modified to form CL, which is further transported to the OMM. An example of a bidirectional movement of lipids between the ER and mitochondria is the production of phosphatidylethanolamine (PE) and phosphatidylcholine (PC). Phosphatidylserine (PS) is first produced on the ER membrane and then translocated to the OMM. On the OMM, PS is transferred to the IMM where it is enzymatically modified to form PE. Finally, in order to produce PC, the precursor PE must be transferred back to the ER where specific enzymes modify it into PC.

or inhibited (Cosson et al., 2012; Filadi et al., 2015; Wang et al., 2015) by MFN2 is controversially discussed, depending perhaps on the cellular growth conditions. Interestingly, a role in ER mitochondria contacts in inhibiting mitophagy was recently shown (Basso et al., 2018; McLelland et al., 2018).

FIGURE 5 | Roles of mitochondria-ER contacts. The physical contacts established between mitochondria and ER are responsible for several cellular processes such as mitochondrial fission, calcium buffering, and phospholipid synthesis. During mitochondrial fission, ER tubules are found in contact with mitochondria in the future fission sites. Drp1 is recruited to these sites and, together with ER tubules, promotes constriction and fission of mitochondria. Calcium (Ca2+) is transported from the ER to mitochondria via transporters in each membrane. On the ER membrane, Ca2+ is exported via the inositol 1,4,5-triphosphate receptor (I3PR) to the voltage- dependent anion-selective channel (VDAC) on the mitochondrial outer membrane. For the synthesis of some phospholipids, the required enzymes are found in the mitochondria and not in the endoplasmic reticulum (ER). This implies the transfer of precursor forms of phospholipids from the ER to mitochondria where they can be modified and then re-transferred to the ER. For example, the production of the mitochondrial phospholipid cardiolipin (CL) depends on precursor lipids present at the ER. Phosphatidic acid (PA) is transferred from the ER, across the IMS to the IMM, where it is enzymatically modified to form CL, which is further transported to the OMM. An example of a bidirectional movement of lipids between the ER and mitochondria is the production of phosphatidylethanolamine (PE) and phosphatidylcholine (PC). Phosphatidylserine (PS) is first produced on the ER membrane and then translocated to the OMM. On the OMM, PS is transferred to the IMM where it is enzymatically modified to form PE. Finally, in order to produce PC, the precursor PE must be transferred back to the ER where specific enzymes modify it into PC.

One of the most well characterized processes where ER mitochondria contacts are indispensable is Ca 2+ buffering (Marchi et al., 2017). Ca2+ is transferred from the ER through the inositol 1, 4,5 triphosphate receptor (I3PR) to the voltage dependent anion selective channel (VDAC) on the OMM (Figure 5; Rizzuto et al., 1998; Szabadkai et al., 2006). In turn, Ca2+ influx to the mitochondrial matrix occurs via the mitochondrial calcium uniporter (MCU; Kirichok et al., 2004; Baughman et al., 2012). The transfer of Ca2+ to mitochondria is required for several mitochondrial proteins or processes, including the TCA cycle enzymes (Bustos et al., 2017).

Apoptosis

Apoptosis is a highly regulated programmed form of cell death that occurs in response to stress. The apoptotic cascade can be activated via the extrinsic or the intrin sic pathway depending on whether the stress signals are extra or intracellular, respectively (Elmore, 2007; Gallu zzi and Vitale, 2018). Both pathways culminate with the activation of caspases, the final effectors of apoptosis. The extrinsic pathway is initiated with the binding of an extracellular death ligand to a cell-surface death recep tor. In turn, internal death stimuli are, for example, DNA damage, oncogene activation, the absence of certain growth factors/hormones, or viral infection. The apop totic intrinsic pathway is mediated by mitochondria (and therefore also known as mitochondrial pathway). It occurs through the release of pro-apoptotic molecules

In fact, some factors required for mitochondrial division and found in MAM sites are regulated by Ca2+ binding, as for example MIRO, a protein mainly required for mitochondrial trafficking (Saotome et al., 2008). Finally, Ca2+ transferring at the ER mitochondria contacts was also shown to activate apoptosis (Doghman Bouguerra and Lalli, 2019). Ca2+ influx to mitochondria is able to open the mitochondrial permeability transition pore, leading to the release of cytochrome c and further apoptosis induction (Scorrano et al., 2003).

Apoptosis

from the IMS to the cytosol, for example, cytochrome c or SMAC/DIABLO (Figure 6; Xiong et al., 2014; Ugar te-Uribe and García-Sáez, 2017).

In addition, mitochondrial division has been shown to be affected by Ca2+ levels, in a Drp1 dependent manner (Friedman et al., 2011).

Apoptosis is a highly regulated programmed form of cell death that occurs in response to stress. The apoptotic cascade can be activated via the extrinsic or the intrinsic pathway depending on whether the stress signals are extra or intracellular, respectively (Elmore, 2007; Galluzzi and Vitale, 2018). Both pathways culminate with the activation of caspases, the final effectors of apoptosis. The extrinsic pathway is initiated with

The major players in the apoptotic mitochondri al pathway are proteins belonging to the Bcl-2 fami ly, which can be divided into pro-survival, pro-apop totic, or apoptosis initiators (Xiong et al., 2014). Within the pro-apoptotic Bcl-2 proteins, BAX and BAK are the two main regulators (Wei et al., 2001; Ugarte-Uribe and García-Sáez, 2017). BAX is a cytosolic protein that trans locates to mitochondria upon apoptotic stimuli, where it oligomerizes (Antonsson et al., 2001). Simultaneously, BAK, which locates to mitochondria, undergoes confor mational changes and oligomerization upon death stim uli (Griffiths et al., 1999). Although not completely under

Interestingly, BAX and BAK can interact with mito fusins and Drp1, thus placing apoptosis in close relation with mitochondrial dynamics (Karbowski et al., 2002; Brooks et al., 2007). However, the impact of mitochon drial dynamics and morphology on apoptosis is still con troversially discussed (Xie et al., 2018). On one hand, mitochondrial fragmentation was suggested to induce cell death, because fragmented or clustered mitochon dria correlated with increased apoptosis, whereas Drp1 loss-of-function prevented apoptosis (Frank et al., 2001; Huang et al., 2007). BAX was shown to translocate to specific sites on mitochondria during the early stages of apoptosis, which subsequently become mitochon drial fission sites (Karbowski et al., 2002), Consistently, Drp1 was able to permeabilize the OMM by BAX recruit ment to mitochondria (Montessuit et al., 2010). More over, a pro-apoptotic role of Drp1 by stabilizing ER-mito chondria contact sites was recently shown (Prudent et al., 2015; Prudent and Mcbride, 2017). On the other hand, caspase-3 activation and enhanced apoptosis could be observed in Drp1-deficient mice or derived colon can cer cells, attributing an anti-apoptotic role to mitochon drial fragmentation (Wakabayashi et al., 2009; Inoue-ya mauchi and Oda, 2012). Reciprocally, a role of BAK and BAX in the regulation of mitochondrial fusion was pro posed (Brooks et al., 2007; Hoppins et al., 2011). First, a role of BAK in promoting mitochondrial fragmentation during apoptosis was suggested, along with the disasso ciation from MFN2 and association with MFN1 (Brooks et al., 2007). Second, under non-apoptotic conditions, solu ble BAK activated mitochondrial fusion via MFN2 (Hop pins et al., 2011). In conclusion, the reciprocal relation between mitochondrial dynamics and apoptosis is com plex and context-specific.

Mitophagy

Mitochondria are kept in vigilance by a multi-layered quality control system that protects it against all sorts of stress, ensuring maintenance of healthy mitochondria (Harper et al., 2018; Pickles et al., 2018). Upon extreme stress, such as loss of membrane potential, failure of mitochondrial channels, or severe mitochondrial dys

- 34 - - 35 -

Mitophagy requires the presence of specific recep tors linking the autophagosome membrane to the mito chondrial portion destined for degradation, and it can be either dependent or independent on ubiquitin. More over, in most cases, mitophagy is also dependent on the ubiquitin-like modifier Atg8 (yeast)/LC3 (mammals), whose lipidated and active form (LC3-II) integrates in the autophagosome membrane. However, LC3-inde pendent mitophagy has also been reported (Nishida et al., 2009; Saito et al., 2019). In Saccharomyces cerevisiae, mitochondria are targeted via the OMM protein recep tor Atg32, its binding to Atg8, and, consequently, the acti vation of mitophagy (Kanki et al., 2009; Okamoto et al., 2009). In mammals, the homologue of Atg32, Bcl2-L-13, was reported to bind LC3-II on the autophagosome membrane and to be required for mitophagy induction (Otsu et al., 2015). Furthermore, other OMM proteins containing an LC3 interacting (LIR) motif, such as BNIP3, NIX and FUNDC1, were also described to target mito chondria for mitophagic destruction (Novak et al., 2010; Rikka et al., 2011; Liu et al., 2012; Wu et al., 2014).

The mostly described factors mediating ubiqui tin-dependent regulation of mitophagy are the kinase PINK1 and the E3 ligase Parkin (Figure 7). Upon loss of membrane potential, PINK1 accumulates at the OMM and recruits Parkin to the mitochondria. Once at the OMM, Parkin is phosphorylated by PINK1 and there by activated (Shiba-Fukushima et al., 2012). Activat ed Parkin initiates ubiquitylation of several OMM pro teins, including MFN1 and MFN2, which immediately leads to loss of fusion events and to mitochondrial frag mentation, characteristic of mitophagy (Gegg et al., 2010; Poole et al., 2010; Tanaka et al., 2010; Ziviani et al., 2010). Furthermore, the poly-ubiquitin chains on sur face proteins get bound to LC3-II via several adaptors, such as optineurin, NDP52, and p62 (Geisler et al., 2010; Narendra et al., 2010b; Lazarou et al., 2015; Khaminets et al., 2016; Mcwilliams and Muqit, 2017), thus allow ing the association of mitochondria to the autophago somes (Geisler et al., 2010).

In addition to Parkin, the E3 ligase Gp78 also activates mitophagy upon mitochondrial depolarization (Fu et al., 2013). Moreover, other ligases were reported to induce

PA PS PE PC CL PA CL PS PE

Calcium Buffering Drp1

the binding of an extracellular death ligand to a cell surface death receptor. In turn, internal death stimuli are, for example, DNA damage, oncogene activation, the absence of certain growth factors/hormones, or viral infection. The apoptotic intrinsic pathway is mediated by mitochondria (and therefore also known as mitochondrial pathway). It occurs through the release of pro apoptotic molecules from the IMS to the cytosol, for example, cytochrome c or SMAC/DIABLO (Figure 6; Xiong et al., 2014; Ugarte Uribe and García Sáez, 2017).

mitophagy as a response to other stress factors, for example, MARCH5, upon disruption of oxygen homeo stasis (Daskalaki et al., 2018; Ferrucci et al., 2018; Shefa et al., 2019). In fact, deficiency of O2 causes hypoxic stress, but excess of O2 may lead to excessive reactive oxygen species, both with toxic consequences for the cells. Con sequently, eukaryotes have developed complex systems to maintain their oxygen homeostasis. Not surprisingly, hypoxic stress was shown to induce mitophagy, depen dent on the receptor FUNDC1 (Liu et al., 2012; Lampert et al., 2019) and its ubiquitin- dependent regulation by MARCH5 (Chen et al., 2017).

The major players in the apoptotic mitochondrial pathway are proteins belonging to the Bcl 2 family, which can be divided into pro survival, pro apoptotic, or apoptosis initiators (Xiong et al., 2014). Within the pro apoptotic Bcl 2 proteins, BAX and BAK are the two main regulators (Wei et al., 2001; Ugarte Uribe and García Sáez, 2017). BAX is a cytosolic protein that translocates to mitochondria upon apoptotic stimuli, where it oligomerizes (Antonsson et al., 2001). Simultaneously, BAK, which locates to mitochondria, undergoes conformational changes and oligomerization upon death stimuli (Griffiths et al., 1999). Although not completely understood how, both BAX and BAK form pores on the OMM, enabling the release of pro apoptotic molecules from the IMS to the cytosol (Wang and Youle, 2009 ). Once in the cytosol, cytochrome c binds to the apoptotic protease activating factor 1 (Apaf 1) (Liu et al., 1996), forming the apoptosome. This complex cleaves and activates the pro caspase 9, followed by the activation of effector caspases (Xiong et al., 2014).

E3 LIGASES ACTING ON MITOFUSINS

Various E3 ligases, soluble or membrane embedded and either located to the cytoplasm, the OMM, or the ER, have been shown to regulate either one or both mitofusins, as a response to various physiological or stress-induced conditions (Figures 8, 9A,B). The OMM E3 MARCH5 was implicated in the regulation of mitochon drial morphology, apoptosis, and ER-mitochondria con tacts and in responses to toxic stress, via both MFN1 and MFN2. In turn, ubiquitylation of MFN2 by the OMM E3 MUL1 is linked to mitochondrial morphology, mitopha gy, and neurodegeneration (Cilenti et al., 2014; Yun et al., 2014; Tang et al., 2015). The ER-located E3 Gp78 affected

mitochondrial fragmentation was suggested to induce cell death, because fragmented or clustered mitochondria correlated with increased apoptosis, whereas Drp1 loss of function prevented apoptosis (Frank et al., 2001; Huang et al., 2007). BAX was shown to translocate to specific sites on mitochondria during the early stages of apoptosis, which subsequently become mitochondrial fission sites (Karbowski et al., 2002), Consistently, Drp1 was able to permeabilize the OMM by BAX recruitment to mitochondria (Montessuit et al., 2010). Moreover, a pro apoptotic role of Drp1 by stabilizing ER mitochondria contact sites was recently shown (Prudent et al., 2015; Prudent and Mcbride, 2017 ). On the other hand, caspase 3 activation and enhanced apoptosis could be observed in Drp1 deficient mice or derived colon cancer cells, attributing an anti apoptotic role to mitochondrial fragmentation (Wakabayashi et al., 2009 ; Inoue yamauchi and Oda, 2012 ). Reciprocally, a role of BAK and BAX in the regulation of mitochondrial fusion was proposed (Brooks et al., 2007; Hoppins et al., 2011). First, a role of BAK in promoting mitochondrial fragmentation during apoptosis was suggested, along with the disassociation from MFN2 and association with MFN1 ( Brooks et al., 2007). Second, under non apoptotic conditions, soluble BAK activated mitochondrial fusion via MFN2 (Hoppins et al., 2011). In conclusion, the reciprocal relation between mitochondrial dynamics and apoptosis is complex and context specific.

Interestingly, BAX and BAK can interact with mitofusins and Drp1, thus placing apoptosis in close relation with mitochondrial dynamics (Karbowski et al., 2002; Brooks et al., 2007). However, the impact of mitochondrial dynamics and morphology on apoptosis is still controversially discussed (Xie et al., 2018). On one hand,

Death stimuli

BAX

mitophagy and ER-mitochondria contacts, via ubiquityl ation of both mitofusins (Fu et al., 2013; Wang et al., 2015). Interestingly, the cytosolic E3 MGRN1 was proposed to coordinate the balance between mitochondrial fusion and mitophagy, via Gp78 (Mukherjee and Chakrabarti, 2016a,b). Constitutively, MGRN1 promotes a stabilizing ubiquitylation on MFN1, concomitant with a destabiliz ing ubiquitylation on Gp78, thus preventing mitophagy and instead promoting fusion. By contrast, stress pre vented MGRN1-dependent ubiquitylation and turnover of Gp78, consequently leading to MFN1 turnover, mito chondrial fragmentation, and induction of mitopha gy. Ubiquitylation of mitofusins by another cytosolic E3, HUWE1, is linked to both genotoxic stress and mito phagy (Leboucher et al., 2012; Di Rita et al., 2018). Final ly, the cytosolic E3 Parkin was shown to be recruited to mitochondria under stress, thus ubiquitylating mito fusins and promoting mitophagy, but was also suggested to regulate ER-mitochondria contact sites, both in mam mals and in Drosophila (Narendra et al., 2008; Gegg et al., 2010; Poole et al., 2010; Tanaka et al., 2010; Ziviani et al., 2010; Glauser et al., 2011; Rakovic et al., 2011).

and a RING-finger domain at its N-terminus (Nakamura et al., 2006; Yonashiro et al., 2006). MARCH5 is associat ed with the ubiquitylation and degradation of proteins regulating mitochondrial dynamics. It was shown that its overexpression increased mitochondrial tubulation and that its depletion or the presence of a RING-inactive mutant leads to mitochondrial fragmentation (Nakamu ra et al., 2006; Yonashiro et al., 2006). These results sup porting a “pro-tubulation” role of MARCH5, meaning either promoting fusion or via inhibition of mitochon drial fission, the later suggested by Xu et al. (2016). How ever and by contrast, downregulation or RING-inactive mutants of MARCH5 were also shown to induce abnor mal elongation of mitochondria (Karbowski et al., 2007; Park et al., 2010). In fact, depending on the circumstanc es, MARCH5 ubiquitylates or interacts with both fission and fusion components, suggesting a plastic role in the regulation of mitochondrial morphology, as a response to different stimuli.

to mitochondrial aggregation and cellular senescence, which is counteracted by MARCH5. MARCH5 interacts with MFN1 (Park et al., 2010), preferentially binding to acetylated MFN1 on K491 (Park et al., 2014), which is con served in yeast but not in MFN2. Then, MARCH5 assem bles K-48-linked ubiquitin chains on MFN1, addressing it for proteasomal degradation (Park et al., 2014). More over, cellular senescence of MARCH5 depleted cells could be rescued by further knockdown of MFN1, espe cially under Antimycin A induced stress (Park et al., 2014). This suggests an important pro-survival role of MARCH5 upon mitochondrial stress via MFN1, concom itant with increased acetylation of MFN1, rendering it a preferential substrate for MARCH5- dependent degrada tion (Park et al., 2014).

Mitophagy

Mitochondria are kept in vigilance by a multi layered quality control system that protects it against all sorts of stress, ensuring maintenance of healthy mitochondria (Harper et al., 2018; Pickles et al., 2018). Upon extreme stress, such as loss of membrane potential, failure of mitochondrial

MARCH5

The E3 ligase Human membrane-associated RING-CH-V (MARCH5, also named MARCHV or MITOL) is an integral OMM protein with four membrane-spanning segments

Consistent with a plastic role of MARCH5, this ligase was reported to control a fine balance of MFN1 levels and mitochondrial fusion, in order to avoid cellular senes cence (Park et al., 2014). First, MARCH5 downregulation led to intense elongation of mitochondria, a pro-surviv al effect (Park et al., 2010). However, persistent downreg ulation caused aggregation of mitochondria, progressive cellular enlargement and flattening as well as increased senescence (Park et al., 2010). Consistently, the same authors showed that upon mitochondrial stress, caused by Antimycin A (an inhibitor of complex 3 of the respi ratory chain), mitochondria first elongate along with increased levels of MFN1. However, excessive MFN1 leads

A link of MARCH5 with cell death, mediated by ubiq uitylation of MFN1, was observed upon the addition of CGP37157 (CGP), an inhibitor of mitochondrial calci um efflux, thus an enhancer of apoptosis (Choudhary et al., 2014). Induction of cell death in prostate cancer cells with CGP led to ubiquitylation and degradation of MFN1. MFN1 turnover was dependent on MARCH5, sug gesting that it could directly modify MFN1. Moreover, MFN1 depletion in prostate cancer cells increased the cell death response to CGP. Therefore, a pro-apoptot ic role of MARCH5 and the potential therapeutic bene fits of MFN1 inhibition are suggested (Choudhary et al., 2014). Consistently, in induced pluripotent stem cells (iPSCs), a decrease in cell viability and ATP content, as well as mitochondrial fragmentation, was observed with tributyltin (TBT; Yamada et al., 2016), an endocrine dis ruptor that causes neurotoxicity and immunotoxici

FIGURE 6 | Apoptotic intrinsic pathway mediated by mitochondria. The programmed and regulated cell death, apoptosis, can occur via two different pathways— intrinsic or extrinsic—according to the origin of the death stimuli, whether it is intrinsic or extrinsic to the cell. Upon intrinsic death stimuli, such as, for example, DNA damage or oncogene activation, the intrinsic apoptotic pathway is activated, which is mediated by mitochondria. Intrinsic stimuli induce the oligomerization of a pro-apoptotic BcL-2 protein—BAX. These oligomers are able to permeabilize the mitochondrial membrane by pore formation on the OMM. Membrane permeabilization allows the release of pro-apoptotic molecules from the IMS, importantly, cytochrome c. In a complex together with other pro-apoptotic proteins, cytochrome c activates caspases, the effectors of apoptosis.

FIGURE 6 | Apoptotic intrinsic pathway mediated by mitochondria. The programmed and regulated cell death, apoptosis, can occur via two different pathways intrinsic or extrinsic according to the origin of the death stimuli, whether it is intrinsic or extrinsic to the cell. Upon intrinsic death stimuli, such as, for example, DNA damage or oncogene activation, the intrinsic apoptotic pathway is activated, which is mediated by mitochondria. Intrinsic stimuli induce the oligomerization of a pro apoptotic BcL 2 protein BAX. These oligomers are able to permeabilize the mitochondrial membrane by pore formation on the OMM. Membrane permeabilization allows the release of pro apoptotic molecules from the IMS, importantly, cytochrome c. In a complex together with other pro apoptotic proteins, cytochrome c activates caspases, the effectors of apoptosis.

FIGURE 7 | Mitochondrial clearance via the ubiquitin-mediated pathway. Upon mitochondrial depolarization, a cascade of events is initiated, which targets damaged mitochondria, or portions of it, for degradation by the autophagy machinery. In a first step, the kinase PINK1 accumulates at the mitochondrial outer membrane and initiates phosphorylation (P) and recruitment of the E3 ligase Parkin. Activated Parkin ubiquitylates (U) several outer mitochondrial membrane proteins such as Mitofusins 1 and 2. Additional phosphorylation of ubiquitin generates a positive feedback loop increasing Parkin recruitment and further ubiquitylation. The formation of ubiquitin chains on mitochondrial surface proteins promotes its binding to lipidated LC3, an autophagosome receptor, via the mitochondrial receptors (R) such as optineurin, NDP52, or p62. From this point, the mitochondria, or its fragments, meant to be degraded are surrounded by the autophagosome, which finally fuses with the lysosome for final destruction.

BAX recruitment

Membrane pore formation Cytochrome release and caspases activation

Cyt c Caspase

Mitochondrial morphology / Cell cycle

Mitochondrial morphology

ER mitochondria contacts Apoptosis

MARCH5

Mitophagy

ER mitochondria contacts Parkin

Mitochondrial morphology MUL1 Mitophagy MGRN1 MFN1 MFN2 Mitophagy

Mitophagy HUWE1

FIGURE 8 | E3 ligases that modify mitofusins and cellular processes associated. MARCH5, Parkin, and Gp78 regulate both mitofusins, whereas MGRN1 affects MFN1 and HUWE1 and MUL1 affect MFN2.

Gp78

ER mitochondria Mitophagy contacts

Genotoxic stress

instead promoting the formation of MFN2 higher oligo mers, important to maintain ER-mitochondria con tacts (Sugiura et al., 2013). Consistently, MARCH5 knock down caused MFN2 decrease at MAM sites as well as reduced co-localization of mitochondria and ER, which was rescued upon re-expression of MARCH5. More over, a decrease in mitochondrial Ca2+ uptake could be observed, suggesting a functional role of MFN2-mediat ed contact sites, dependent on MARCH5 ubiquitylation of MFN2 (Sugiura et al., 2013).

HUWE1

phosphorylation of MFN2 by JNK leads to recruitment of the E3 ligase HUWE1 to phosphorylated MFN2 and subsequent acceleration of its degradation by the pro teasome. Consequently, degradation of MFN2 leads to enhanced mitochondrial fragmentation and apoptosis (Leboucher et al., 2012).

ty (Kotake, 2012). TBT led to a decrease in MFN1 levels, which depended on MARCH5, presumably via direct ubiquitylation of MFN1 (Yamada et al., 2016).

FIGURE 8 E3 ligases that modify mitofusins and cellular processes associated. MARCH5, Parkin, and Gp78 regulate both mitofusins, whereas MGRN1 affects MFN1 and HUWE1 and MUL1 affect MFN2.

Wang et al., 2015). Interestingly, the cytosolic E3 MGRN1 was proposed to coordinate the balance between mitochondrial fusion and mitophagy, via Gp78 (Mukherjee and Chakrabarti, 2016a,b). Constitutively, MGRN1 promotes a stabilizing ubiquitylation on MFN1, concomitant with a destabilizing ubiquitylation on Gp78, thus preventing mitophagy and instead promoting fusion. By contrast, stress prevented MGRN1 dependent ubiquitylation and turnover of Gp78, consequently leading to MFN1 turnover, mitochondrial fragmentation, and induction of mitophagy. Ubiquitylation of mitofusins by another cytosolic E3, HUWE1, is linked to both genotoxic stress and mitophagy (Leboucher et al., 2012; Di Rita et al., 2018). Finally, the cytosolic E3 Parkin was shown to be recruited to mitochondria under stress, thus ubiquitylating mitofusins and promoting mitophagy, but was also suggested to regulate ER mitochondria contact sites, both in mammals and in Drosophila (Narendra et al., 2008; Gegg et al., 2010; Poole et al., 2010; Tanaka et al., 2010; Ziviani et al., 2010; Glauser et al., 2011; Rakovic et al., 2011).

MARCH5

The apparent contradictory observations of a proand anti- survival role of MARCH5 upon Antimycin A or CGP/TBT treatment, respectively, could be explained by the extension of MFN1 overexpression achieved by each stress. Consistently, the levels of acetylated MFN1 prob ably regulate a fine-tuned balance of fusion activity as well. Indeed, the ubiquitin binding deacetylase HDAC6, mostly cytosolic, was shown to interact with MFN1, mainly under glucose deprivation (Lee et al., 2014). Inter estingly, HDAC6 also bound to MFN2, but its acetylation or interaction with HDAC6 was not altered under glu cose deprivation (Lee et al., 2014). Importantly, acetyla tion of MFN1 at K222 was shown to inhibit its fusogen ic activity. Moreover, HDAC6-dependent deacetylation of MFN1 ameliorated ROS production during glucose starvation, supporting a role of MFN1-mediated fusion to cope with metabolic stress (Lee et al., 2014). There fore, HDAC6 promotes MFN1-dependent hyperfusion, observed both upon glucose starvation in cells and upon fasting in mice (Lee et al., 2014). In conclusion, various stress conditions lead to MARCH5- dependent ubiquityl ation and degradation of MFN1.

In addition to MFN1, MARCH5 and HDAC6 regu late stress- induced MFN2 turnover (Kim et al., 2015). MARCH5 interacts with MFN2 (Nakamura et al., 2006), which occurs between the C-terminal domain of MARCH5 and the HR1 domain of MFN2 (Sugiura et al., 2013). Moreover, MARCH5 was shown to be responsi

The E3 ligase Human membrane associated RING CH V (MARCH5, also named MARCHV or MITOL) is an integral OMM protein with four membrane spanning segments and a RING finger domain at its N terminus (Nakamura et al., 2006; Yonashiro et al., 2006). MARCH5 is associated with the ubiquitylation and degradation of proteins regulating mitochondrial dynamics. It was shown that its overexpression increased mitochondrial tubulation and that its depletion or the presence of a RING inactive mutant leads to mitochondrial fragmentation (Nakamura et al., 2006; Yonashiro et al., 2006). These results supporting a “pro tubulation” role of MARCH5, meaning either

promoting fusion or via inhibition of mitochondrial fission, the later suggested by Xu et al. (2016). However and by contrast, downregulation or RING inactive mutants of MARCH5 were also shown to induce abnormal elongation of mitochondria (Karbowski et al., 2007; Park et al., 2010). In fact, depending on the circumstances, MARCH5 ubiquitylates or interacts with both fission and fusion components, suggesting a plastic role in the regulation of mitochondrial morphology, as a response to different stimuli.

Consistent with a plastic role of MARCH5, this ligase was reported to control a fine balance of MFN1 levels and mitochondrial fusion, in order to avoid cellular senescence ( Park et al., 2014). First, MARCH5 downregulation led to intense elongation of mitochondria, a pro survival effect (Park et al., 2010). However, persistent downregulation caused aggregation of mitochondria, progressive cellular enlargement and flattening as well as increased senescence (Park et al., 2010). Consistently, the same authors showed that upon mitochondrial stress, caused by Antimycin A (an inhibitor of complex 3 of the respiratory chain), mitochondria first elongate along with increased levels of MFN1. However, excessive MFN1 leads to mitochondrial aggregation and cellular senescence, which is counteracted by MARCH5. MARCH5 interacts with MFN1 (Park et al., 2010), preferentially binding to acetylated MFN1 on K491 ( Park et al., 2014), which is conserved in yeast but not in MFN2. Then, MARCH5 assembles K 48 linked ubiquitin chains on MFN1, addressing it for proteasomal degradation ( Park et al., 2014). Moreover, cellular senescence of MARCH5 depleted cells could be rescued by further knockdown of MFN1, especially under Antimycin A induced stress (Park et al., 2014). This suggests an important pro survival role of MARCH5 upon mitochondrial stress via MFN1, concomitant with increased acetylation of

ble for ubiquitylation and degradation of MFN2 under hypoxic stress, provoked by adding deferoxamine (DFO), in cells lacking HDAC6 (Kim et al., 2015). However, HDAC6 bound strongly to MFN2 in the presence of DFO, thus inhibiting MFN2 turnover and preventing mito chondrial fragmentation (Kim et al., 2015). Therefore, HDAC6 can protect mitochondria and activate adaptive mitochondrial fusion under hypoxic stress (via deacetyl ation of MFN2) and under starvation (via deacetyla tion of MFN1). In both cases, deacetylation of mitofusins induced their fusogenic capacity by preventing ubiqui tylation and degradation by MARCH5. The regulation of MFN1 and MFN2 by MARCH5 was also observed in the absence of external stress, being instead either linked to cell cycle, via MFN1, or linked to ER-mitochondria exchanges, via MFN2. Cell cycle transitions are accompa nied by alterations in mitochondrial morphology, which are actively regulated by both fusion and fission proteins (Mitra, 2013; Horbay and Bilyy, 2016). During the G2/M phase, mitochondria fragment; then, after cellular divi sion, they fuse again (Mitra, 2013). Consistent with the fragmentation, it was shown that MARCH5 ubiquitylates MFN1 at G2/M, addressing it for proteasomal degrada tion (Park and Cho, 2012). By contrast, a non-proteolyt ic role of MARCH5 acting on MFN2 was demonstrated, through the addition of K63-linked polyubiquitin chains at K192, located on the GTPase domain and not con served in MFN1 (Sugiura et al., 2013). Despite the sug gestion that MFN2 also locates to the ER, MARCH5 only ubiquitylated MFN2 present at the mitochondria, not ER-associated MFN2. This created a non-proteolytic tag,

HUWE1 is a multifunctional E3 ligase belonging to the HECT- domain E3 ligase family, therefore forming a ubiquitin-thioester intermediate with ubiquitin, before transferring it to the substrate (Bernassola et al., 2008). HUWE1 is composed by an ubiquitin- associated domain, a WWE domain (involved in proteolysis), a BH3 domain (common to the family), and two N-terminal domains. Finally, the HECT domain is found in its C-terminus (Kao et al., 2018). HUWE1 mediates not only K48- and K63linked poly-ubiquitylation (Adhikary et al., 2005; Zhao et al., 2008) but also mono-ubiquitylation (Parsons et al., 2009) and K11-/ K6-linked ubiquitylation (Michel et al., 2017), therefore being suggested to assemble a powerful ubiquitin combination for proteasomal turnover (Meyer and Rape, 2014). This E3 ligase is majorly known for reg ulation of proliferation, differentiation, and apoptosis, being therefore associated with cancer and metastasis (Kao et al., 2018).

MFN2 was recently shown to be modified in a HUWE1- dependent manner with K6-linked polyubiqui tin chains (Michel et al., 2017). Moreover, it was demon strated that HUWE1 ubiquitylates MFN2 in response to genotoxic stress (Leboucher et al., 2012) and upon induction of mitophagy (Di Rita et al., 2018), thereby addressing MFN2 for proteasomal degradation. First, in sarcoma cells, HUWE1 ubiquitylated MFN2 upon acti vation of the c-Jun N-terminal kinase (JNK) pathway, by the addition of Doxorubicin, a genotoxic stressor well known for inducing apoptosis (Leboucher et al., 2012). Doxorubicin led to mitochondrial fragmentation and to MFN2 ubiquitylation and phosphorylation at serine 27, signaling proteasomal-dependent loss of MFN2. The authors identified the JNK as the stress-activated kinase phosphorylating MFN2. In addition, MFN2 bound to the BH3 domain of HUWE1, suggesting a role in apop tosis. Consistently, knockdown of HUWE1 prevented apoptosis, which was restored upon further knock down of MFN2. In summary, this study proposes that

Second, a role of HUWE1 in mitophagy induction was recently identified. This was observed upon the addition of Oligomycin (an inhibitor of complex 5 of OXPHOS) and Antimycin A, by a pathway mediated by AMBRA1 (Di Rita et al., 2018). AMBRA1 is an inducer of autopha gy and a regulator of mitophagy, involved in both PINK1/ Parkin- dependent and independent pathways, through binding to LC3 (Strappazzon et al., 2015). HUWE1 inter acted with AMBRA1, especially under mitophagy condi tions, and translocated to mitochondria. Moreover, both HUWE1-dependent ubiquitylation and proteasomal-de pendent degradation of MFN2 were observed. Impor tantly, HUWE1 depletion impaired AMBRA1-mediated mitophagy (Di Rita et al., 2018).

Gp78

Glycoprotein 78 (Gp78) is an ER membrane-anchored E3 ubiquitin ligase (Nabi and Raz, 1987). Gp78 is inhibit ed by the ubiquitous cytokine autocrine motility factor (AMF) and was first identified as the autocrine motility factor receptor (AMFR), for its role in a signaling cascade regulating cancer cell motility and metastasis (Nabi et al., 1990; Silletti et al., 1991). Gp78 regulates protein qual ity control via the ER-associated degradation machin ery (ERAD; Fang et al., 2001). ERAD is responsible for the degradation of misfolded or functionally denatured pro teins from the ER, occurring via proteasomal degrada tion after their retro-translocation to the cytosol (Mehr tash and Hochstrasser, 2018). This E3 ligase possesses a G2BR domain, for E2-binding, has five N-terminal trans membrane domains, and, in its C-terminus, has the RING-Finger and VIM domain facing the cytosol (Joshi et al., 2017). The E3 ligase activity of Gp78 is associated with cell signaling and motility, metabolism, neurodegenera tion, and cancer/metastasis (Liottaab et al., 1986; Wata nabes et al., 1991; Nabi et al., 1992; Luo et al., 2002).

Regarding mitofusins, a role of Gp78 was observed for mitophagy (Fu et al., 2013) and ER-mitochondria contact sites (Wang et al., 2015). First, overexpression of Gp78 was shown to induce mitochondrial fragmen tation, dependent on its E3 ligase activity (Fu et al., 2013). Moreover, mitochondrial loss was observed, fur ther aggravated upon mitochondrial uncoupling with CCCP, but rescued by Gp78 downregulation, pointing

to its role in mediating mitophagy. Consistently, Gp78 expression was shown to partially recruit LC3 to the ER, co-localizing with Gp78 itself (Fu et al., 2013). More over, recruitment of LC3 to Gp78-positive ER domains was dramatically increased upon CCCP treatment, sup porting a direct role of Gp78 for clearance of damaged mitochondria. Concomitantly, decreased levels of MFN1 and MFN2 were observed, especially in the presence of CCCP, an effect prevented by proteasomal inhibition (Fu et al., 2013). Moreover, Gp78 interacted with both mito fusins (Fu et al., 2013) and its inhibitor, AMF, prevent ed Gp78-induced degradation of both MFN1 and MFN2 (Shankar et al., 2013). However, knockdown of MFN1, but not of MFN2, inhibited induction of mitophagy, suggest ing that albeit degraded, MFN1 is required for mitophagy. In summary, the role of Gp78 in mitophagy appears to be specifically dependent on MFN1.

In addition to mitophagy, Gp78 promoted contact sites between the ER and mitochondria, which were functional in calcium transfer (Wang et al., 2015). How ever, despite Gp78 being mostly localized to smooth-ER (Benlimame et al., 1998), it specifically affected contacts of mitochondria to rough ER, in fibrosarcoma cancer cells (Wang et al., 2015). Moreover, in this case, the selec tive regulation of rough ER by Gp78 depended specif ically on the presence of MFN2. Given that MFN2 lev els increased upon downregulation of Gp78, the role of the ligase in promoting rough ER-mitochondria contacts might occur via ubiquitylation and subsequent degrada tion of MFN2 (Wang et al., 2015). By contrast, although the levels of MFN1 also increased in the absence of Gp78, MFN1 knockdown did not affect rough ER-mitochon dria contacts. Instead, MFN1 knockdown inhibited the contacts between mitochondria and smooth ER, which were not affected by Gp78 knockdown or by the addi tion of its inhibitor AMF. In summary, both MFN1 and MFN2 behaved as inhibitors of ER-mitochondria contact formation, albeit through different mechanisms and at different ER sites (Wang et al., 2015).

MGRN1

FIGURE 9 | Residues, E3 ligases, and processes regulating MFN1 (A) MFN2 (B). Representation of the triggers identified to modify mammalian mitofusins by ubiquitylation, phosphorylation, and acetylation. The enzymes evolved in each case, and the cellular outcome is also depicted. The vertical bar on each side of the structure denotes that the residues modified by ubiquitylation in MFN1 or MFN2 are not known. See text for details.

Residues, E3 ligases, and processes regulating MFN1 (A) MFN2 (B) Representation of the triggers identified to modify mammalian mitofusins by phosphorylation, and acetylation. The enzymes evolved in each case, and the cellular outcome is also depicted. The vertical bar on each side of the that the residues modified by ubiquitylation in MFN1 or MFN2 are not known. See text for details.

The E3 ligase Mahogunin Ring Finger-1 (MGRN1) was first discovered to be the gene mutated in a color-coat mutant mice, mahoganoid (Phan et al., 2002; He et al., 2003; Upadhyay et al., 2016). MGRN1 is a soluble E3 ligase, however, locating to the cytoplasm, plasma mem brane, endosomes, and nucleus (Bagher et al., 2006). Interestingly, a relation between MGRN1 and both Gp78 and MFN1 was suggested, affecting both mitophagy and mitochondrial fusion. First, the levels of Gp78 itself

were negatively regulated by MGRN1, an effect observed in the absence of external stress (Mukherjee and Chakrabarti, 2016a,b). Depletion of MGRN1, or deletion of the RING-finger domain, caused perinuclear cluster ing of mitochondria and increased oxidatively modified proteins, indicative of ROS (Mukherjee and Chakrabarti, 2016a,b). Moreover, MGRN1-dependent ubiquitylation of Gp78 addressed it for proteasomal degradation. Thus, by repressing Gp78, MGRN1 indirectly prevented mitopha gy, MFN1 turnover, and perinuclear clustering. However, depolarization conditions compromised ubiquitylation of Gp78 by MGRN1, thus rescuing the levels of Gp78 and favoring mitophagy of damaged mitochondria (Mukher jee and Chakrabarti, 2016b). Indeed, cells expressing the Ring-Finger mutant MGRN1 reveal higher propensity for mitophagy: they displayed higher levels of LC3-positive mitochondria and presented increased co-localization of mitochondria with p62-positive autophagic vesicles.

Second, MGRN1 also preserved MFN1 stability by directly interacting with it. Moreover, expression of a RING-finger mutant impaired higher oligomerization of MFN1 and mitochondrial fusion (Mukherjee and Chakrabarti, 2016a). Thus, MGRN1 was suggested to actively promote mitochondrial fusion via non-degra dative ubiquitylation of MFN1, consistent with previous observations (Yue et al., 2014). This is reminiscent of the E3 ligase SCFMdm30 that modifies the yeast mito fusin Fzo1 (Fritz et al., 2003; Escobar-henriques et al., 2006; Cohen et al., 2008, 2011; Anton et al., 2011) with a stabilizing ubiquitylation (Anton et al., 2013; Simões et al., 2018). Indeed, a dual and interdependent balance between constitutive/ non-degradative ubiquitylation of MFN1 vs. stress-induced/ degradative ubiquitylation resembles the regulation of Fzo1 in yeast (Anton et al., 2013; Simões et al., 2018).

Parkin

Parkin, an E3 ligase associated with Parkinson’s disease (Kitada et al., 1998), belongs to the RING-between-RING family of E3 ligases (Smit and Sixma, 2014; Walden and Rittinger, 2018): it has an N-terminal Ub-like (UBL) domain, a zinc-binding domain (RING0, unique to Par kin), a RING domain (RING1, a canonical domain), and two linear zinc-binding folds (IBR and RING2; Panick er et al., 2017). Importantly, Parkin recruitment to the mitochondria requires PINK1 (Geisler et al., 2010; Mat suda et al., 2010; Vives-Bauza et al., 2010; Narendra et al., 2010a), namely its kinase activity and localization at the mitochondrial surface (Okatsu et al., 2012; Shi ba-Fukushima et al., 2012). Moreover, crystal structures

rendering it a preferential substrate for MARCH5 and immunotoxicity (Kotake, 2012). TBT led to a decrease in

revealed that Parkin assumes an auto-inhibited confor mation that is activated by undergoing major structural rearrangements, which require PINK1-dependent phos phorylation at serine 65 of its UBL domain. In addition, it requires binding of ubiquitin itself, also phosphory lated by PINK1 at serine 65 (Riley et al., 2013; Spratt et al., 2013; Wauer and Komander, 2013; Caulfield et al., 2014; Kane et al., 2014; Kazlauskaite et al., 2014; Koyano et al., 2014; Swatek and Komander, 2016; Wauer et al., 2016; Kumar et al., 2017; Mcwilliams et al., 2018; Glad kova et al., 2019). Parkin is able to ubiquitylate itself as well as a large variety of both cytosolic and OMM pro teins. Mono-ubiquitylation and K63-, K48-, K11-, and K6-linked poly-ubiquitylation have been reported for this E3 ligase (Cunningham et al., 2015; Seirafi et al., 2015; Mouton- liger et al., 2017). Despite its diverse func tions (Scarffe et al., 2014), the extensive body of litera ture regarding Parkin and its E3 ligase activity is main ly gathered from its role on mitochondrial clearance (Harper et al., 2018; Pickles et al., 2018), whereby mito fusins are ubiquitylated. Nevertheless, Parkin has also been shown to act on mitofusins for the regulation of ER-mitochondria contact sites (Basso et al., 2018; McLelland et al., 2018).

Among the E3 ligases acting on mitofusins, Parkin is definitely the most well studied. Parkin was initial ly shown to translocate from the cytoplasm to depo larized mitochondria in mammalian cells (Narendra et al., 2008). This study led to the development of the hypothesis that Parkin not only is responsible for the ubiquitylation of proteins, leading to their subsequent degradation by the UPS, but also acts on the selective elimination of impaired mitochondria. Strikingly, this dramatically raised the interest of many researchers. Shortly after, several mitochondrial ubiquitylation tar gets of Parkin were identified (Chan et al., 2011). These included dMfn (MARF) in Drosophila (Poole et al., 2010; Ziviani et al., 2010) and MFN1 and MFN2 in murine and mammalian cells (Gegg et al., 2010; Tanaka et al., 2010; Glauser et al., 2011; Rakovic et al., 2011), as well as other OMM proteins like VDAC1, Fis1, or Tom20 (Chan et al., 2011). The ubiquitylation of several mitochondrial pro teins by Parkin precedes mitophagy and is accom plished by the proteasome, independently of autoph agy (Chan et al., 2011; Sarraf et al., 2013). Furthermore, this was shown to be essential for mitophagy, since inhibition of the 26S proteasome fully abrogated Par kin-mediated mitophagy (Tanaka et al., 2010; Chan et al., 2011). In conclusion, upon mitochondrial depolar ization, Parkin mediates ubiquitylation of both mito

fusins, thus addressing them for proteasomal turnover. However, whether mitofusins are actively required for Parkin-mediated mitophagy is still controversial (Nar endra et al., 2008; Chan et al., 2011; Chen and Dorn, 2013), as discussed later.

Interestingly, a direct role of MFN2 in preventing mitophagy dependent on PINK1/Parkin, by keeping mitochondria tethered to the ER, was recently demon strated (Basso et al., 2018; McLelland et al., 2018). First, depolarization impaired the connection between mito chondria and the ER, a process that depended on pro teasomal turnover and was further exacerbated by overexpression of Parkin (McLelland et al., 2018). In fact, turnover of both mitofusins is a very early event after the addition of CCCP (Chan et al., 2011; Sarraf et al., 2013; McLelland et al., 2018). MFN2 was phos pho-ubiquitylated, as expected, and its subsequent degradation contributed to the recruitment of Parkin to mitochondria, thus activating the general ubiquityl ation of other outer membrane proteins and acceler ating mitophagy (McLelland et al., 2018). Consistently, the Parkin-resistant MFN2 HR1 mutants K406R, K416R and K420R failed to induce mitophagy, a specific effect since their capacity to promote mitochondrial fusion was not affected.

Second, in contrast, Parkin-mediated ubiquitylation of mitofusins, also at K416, promoted ER-mitochon dria contact sites (Basso et al., 2018). Indeed, down regulation of Parkin in Drosophila or mouse embryon ic fibroblast (MEF) cells led to decreased ubiquitylation of mitofusins, concomitant with a significant decrease in ER-mitochondria contacts. Moreover, the ubiqui tin dead mutant of dMFN, corresponding to K416R in mammals, was also deficient in establishing ER-mito chondria contacts and in mitochondrial Ca2+ uptake (Basso et al., 2018). These results suggest that ubiqui tylation of dMfn at “lysine 416” regulates physical and functional ER-mitochondria contacts in Drosophila Finally, expressing an ER-mitochondria synthetic linker rescued locomotor deficits associated with Parkinson’s disease. In conclusion, although somehow contradicto ry, both studies suggest an active role of MFN2 in pre venting mitophagy by tethering mitochondria to the ER. It is possible that in the absence of external stress mild MFN2 ubiquitylation by Parkin keeps the two organelles together. Then, upon depolarization, exces sive ubiquitylation occurs instead targeting MFN2 for proteasomal turnover, again reminiscent of findings in yeast (Anton et al., 2013; Simões et al., 2018).

MUL1/MULAN/MAPL/GIDE

Like MARCH5, the E3 ligase MUL1 is an integral outer membrane protein regulating mitochondrial dynamics. MUL1/ MULAN was first identified as an NF- B activa tor with E3 ligase activity (Matsuda et al., 2003). Later, this E3 ligase was identified as a mitochondrial protein, being also known as mitochondria-anchored protein ligase (MAPL) or growth inhibition and death E3 (GIDE) ligase (Li et al., 2008; Neuspiel et al., 2008; Zhang et al., 2008). MUL1 has two transmembrane domains, with its RING-finger domain facing the cytosol, whose overex pression induces fragmentation and perinuclear clus tering of mitochondria (Li et al., 2008; Neuspiel et al., 2008). A clear role of MUL1/MAPL as a SUMO E3 ligase was demonstrated (Braschi et al., 2009). For example, MUL1/MAPL acts in the formation of mitochondria-de rived vesicles addressed to peroxisomes (Neuspiel et al., 2008; Sugiura et al., 2014), in inflammation (Barry et al., 2018), and in innate immunity (Doiron et al., 2017). In addition, other SUMO-dependent roles were shown, in the regulation of mitochondrial fission (Braschi et al., 2009) and activation of apoptosis, by stabilizing ER-mitochondria contact sites via Drp1 SUMOylation (Prudent et al., 2015). Consistently, Zhang et al. (2008) reported MUL1/ GIDE as a pro-apoptotic enzyme. How ever, in what regards the regulation of mitofusins by MUL1, so far only ubiquitin ligase activity was shown (Yun et al., 2014; Tang et al., 2015). MUL1 was suggest ed to compensate for PINK1 and Parkin deficiencies, for example, by ubiquitylating mitofusin in Drosophila, dMfn/MARF (Yun et al., 2014), and MFN2 in mammals (Tang et al., 2015), thus contributing to mitochondri al integrity by promoting mitophagy. In the fly, over expression of dMfn aggravated the lethality and neu rodegeneration phenotypes seen upon PINK1/Parkin deficiency (Yun et al., 2014). Moreover, these deficien cies could be compensated by MUL1-dependent ubiq uitylation and proteasomal degradation of dMfn. Con sistently, loss of MUL1 resulted in an increase in dMfn levels. Importantly, increased levels of dMfn, observed in dopaminergic neurons and muscle of PINK1/Par kin mutant flies, could be rescued upon MUL1 overex pression. Moreover, HeLa cells exposed to cyclohexim ide presented decreased MFN1 and MFN2 levels, which could be stabilized upon MUL1 silencing (Yun et al., 2014). Consistently, in the mouse model for neurode generation, mnd2, where MUL1 accumulated, the lev els of MFN2 decreased and mitophagy was enhanced (Cilenti et al., 2014). Moreover, a role of MUL1 in the degradation of MFN2 was suggested in the context of dopaminergic (DA) neuronal loss (Tang et al., 2015),

- 42 - - 43 -

closely related to Parkinson’s and Alzheimer’s diseases. Indeed, Parkinson’s-linked mutations of VPS35, a ret romer component for endosomal trafficking, correlat ed with increased ubiquitylation and decreased levels of MFN2, dependent on the proteasome (Tang et al., 2015). In addition, these neurons presented mitochon drial fragmentation, along with impaired OXPHOS, pointing to mitochondrial dysfunction. Finally, MUL1 inhibition re-increased MFN2, suggesting that MUL1 directly ubiquitylates MFN2, signaling its degradation by the proteasome (Tang et al., 2015). Together, these results indicate that MUL1 compensates for Parkin son’s phenotypes, caused by the loss of PINK1/Parkin or VSP35, by decreasing MFN2.

MITOFUSINS: PRO- OR ANTI- MITOPHAGIC PROTEINS?

During mitophagy induction, mitofusins are clearly among the first substrates to be ubiquitylated by Par kin (Chan et al., 2011; Sarraf et al., 2013; McLelland et al., 2018). However, it was first reported that the absence of both mitofusins did not affect mitophagy (Narendra et al., 2008; Chan et al., 2011). Nevertheless, mutations or modifications of MFN1 and MFN2, either by ubiq uitylation, acetylation, or phosphorylation, have also been proposed to directly affect mitophagy (Figure 10). As detailed below, whether mitophagy is enhanced or instead repressed by mitofusins and their ubiquitylation is still controversially discussed.

Pro-mitophagic Role

An active and essential role of MFN2 for mitophagy was proposed under diverse mitophagy induction condi tions and different cell lines, tissues, and animal models. First, MFN2- phosphorylation by PINK1 was required for recruiting Parkin to damaged mitochondria (Chen and Dorn, 2013). This study could show that Parkin binds to MFN2 in a PINK1-dependent manner after which PINK1 phosphorylates MFN2. Phosphorylated MFN2 further promotes ubiquitylation of other OMM proteins by Par kin, targeting mitochondria for degradation. This model was supported by the observation that MFN2 depletion from murine cardiomyocytes prevented Parkin trans location to mitochondria upon membrane depolar ization and, consequently, decreased mitophagy levels (Chen and Dorn, 2013). Consistently, depletion of MFN1 and MFN2 caused accumulation of defective mitochon dria, but no increase in mitophagy levels was observed (Song et al., 2015). These results suggest that ablation of

both mitofusins interferes with a step in the mitophag ic process, stopping defective mitochondria from being degraded and leading to their accumulation, point ing, thereby, to the fact that both MFN1 and MFN2 are required for mitochondrial clearance in the murine car diac system. However, it is important to note that other studies have reported that mitophagy still occurs in the absence of MFN1 or MFN2, pointing to the existence of other proteins that serve as Parkin receptors upon mito phagy induction (Narendra et al., 2008; Chan et al., 2011).

Further supporting a pro-mitophagic role, MFN1-knockdown inhibited mitophagy that was caused by overexpression of Gp78. (Fu et al., 2013). Induc tion of mitophagy by Gp78 and MFN1 was Parkin-inde pendent (Fu et al., 2013). Moreover, MFN2 was also not required. However, MFN2 was reported as a central play er in autophagosome-lysosome formation in human cardiomyocytes (Zhao et al., 2012). Deletion of MFN2 in cardiomyocytes led to the extensive accumulation of autophagosomes, in response to ischemia-reperfusion stress, a condition that induces mitophagy (Zhao et al., 2012). Nevertheless, both autophagosome and lysosome formation remained unaltered (Zhao et al., 2012). Instead, autophagosome accumulation was due to marked retar dation of the fusion step between autophagosomes and lysosomes, in the absence of MFN2, a phenotype res cued by re-expression of MFN2 (Zhao et al., 2012). Addi tional evidences for a pro-mitophagic role of MFN2 were provided in murine skeletal muscle, in the context of aging (Sebastián et al., 2016). Aging in mice was accom panied by a decrease in MFN2 levels and, consistently, MFN2 depletion generated aging signatures (Sebastián et al., 2016). In parallel, MFN2 ablation impaired auto phagy and led to excessive mitochondrial dysfunc tion (Sebastián et al., 2016). In conclusion, during aging MFN2 levels decreased, consequently impairing mitoph agy. Also in the context of aging, MFN2 was reported to induce mitophagy in aged human liver cells upon isch emia/ reperfusion (I/R) injury (Chun et al., 2018). This study reported a pro-mitophagic role of MFN2, not pro moted by the canonical Parkin-dependent ubiquitylation of MFN2, but by another type of PTM: deacetylation via sirtuin 1 (SIRT1). Overexpression of either MFN2 or SIRT1 alone failed to rescue I/R injury, mitochondrial dysfunc tion, and cell death. However, their co-expression pro moted autophagy in aged hepatocytes (Chun et al., 2018). Furthermore, the authors showed that deacetylation of MFN2 by SIRT1 occurs at K655 and K662 residues, locat ed in the C-terminus, and directly linked these modifica tion to an increase in autophagy (Chun et al., 2018).

Finally, supporting pro-mitophagic evidences were shown in primary cultured neurons, where I/R injury phenotypes were ameliorated by MFN2 expression and aggravated by its downregulation (Peng et al., 2018). In fact, MFN2 expression led to increased autophagosome formation and fusion with lysosomes (Peng et al., 2018), suggesting once again an important role of MFN2 factor for mitochondrial clearance.

Anti-mitophagic Role

In the context of the neuropathy Charcot-Marie-Tooth Type 2A (CMT2A), which is caused by dominant-neg ative mutations in MFN2, it was suggested that MFN2 behaves as an inhibitor of mitophagy (Rizzo et al., 2016). Indeed, motor neurons derived from iPSCS gen erated from CMT2A fibroblasts had an increased auto phagic flux (Rizzo et al., 2016). This suggests an anti-mi tophagic role of mutant MFN2 in CMT2A patients. Consistently, in the CD4+ T immune system cells, and in the presence of several stimuli such as rapamycin, ion omycin, or starvation, overexpression of MFN2 led to an impairment in autophagy (Ying et al., 2017).

Moreover, an inhibitory effect of MFN2 toward mito phagy was also suggested in the context of its function as an ER-mitochondria tether. ER-mitochondria con tacts are destroyed during mitochondrial clearance process, and, consistently, a reduction in these con tacts leads to an increase in mitochondrial degrada tion (McLelland et al., 2018). Importantly, MFN2 direct ly promoted contacts between mitochondria and ER, therefore preventing mitophagy (McLelland et al., 2018). Moreover, degradation of MFN2 was necessary for mitophagy to occur. In fact, ubiquitylated MFN2 was suggested to be the active form of MFN2 in promoting ER-mitochondria tethering (Basso et al., 2018). Finally, another anti-mitophagic role of MFN2 was proposed, upon compromised PINK1/Parkin, which depended on ubiquitylation of MFN2 by MUL1, followed by proteaso mal degradation (Yun et al., 2014).

DISEASE-ASSOCIATED ROLES OF MFN2

A broad spectrum of disorders has been linked to muta tions or altered levels in MFN2, underlining its physi ological relevance (Figure 4). MFN2 correlation with disease also supports the idea that MFN2, more than MFN1, is involved in several other disease-relevant roles besides mitochondrial fusion.

First, MFN2 mutations cause a rare neurodegener ative disease. In addition, a link of MFN2 to the most

common neuropathies has also been suggested. Impor tantly, so far no cure is possible for these diseases, being only symptomatic treatments available. Second, a role of MFN2 in mitophagy or ER-contact sites pre venting heart failure was proposed. Finally, MFN2 is also linked to diabetes, an aspect of crucial importance in our present society.

Neurodegeneration

The major causal link between MFN2 dysfunction and disease lies with the CMT2A neuropathy (Züchner et al., 2004; Verhoeven et al., 2006). CMT2A is a subtype of an incurable neuropathy, Charcot-Marie-Tooth (CMT), char acterized by progressive distal weakness, muscular atro phy, and sensory abnormalities, affecting 1 in 2,500 peo ple (Tazir et al., 2013, 2014; El-abassi et al., 2014; Stuppia et al., 2015; Stojkovic, 2016; Barbullushi et al., 2019). CMT is one of the most common inherited neurological dis eases, usually inherited as an autosomal dominant trait but sometimes as autosomal recessive and X-linked trait. CMT2A subtype presents an earlier onset, with motor symptoms mainly affecting the lower limbs. More than hundred MFN2 mutations are described as causative of CMT2A, comprising one-fifth of all CMT2A cases (Sto jkovic, 2016; Dohrn et al., 2017). Strikingly, MFN2 muta tions account for approximately 90% of the most severe cases of CMT (Feely et al., 2011).

To date, the disease-underlying functions of MFN2 have not been identified. Recent studies with CMT2A-associated mutations highlighted the impor

tance of carefully addressing membrane potential, apoptosis, ER-mitochondria contacts, and mitopha gy in CMT2A, further suggesting fusion-independent roles of Mitofusin 2 (Saporta et al., 2015; Rizzo et al., 2016; Bernard-Marissal et al., 2018; Larrea et al., 2019). Two very recent studies support a pathogenic role of reduced ER-mitochondria contacts caused by MFN2 mutants associated with CMT2A. First, the presence of the most common CMT2A disease mutant MFN 2R94Q in patient-derived fibroblasts, primary neurons, and in vivo motor neurons of CMT2A mouse model, impaired ER-mitochondria contacts (Bernard-Maris sal et al., 2018). Importantly, ER stress, Ca2+ defective uptake, and alteration in the axonal transport of mito chondria, could also be observed. Second, the extent of ER-mitochondria contacts also diminished with three different CMT2A variants: MFN2R364W, MFN2M376V, and MFN2W740S (Larrea et al., 2019). Moreover, phospholipid synthesis and trafficking were affected in cells express ing these pathogenic mutants, further supporting the functional relevance of ER-mitochondria contact sites for MFN2-related neurodegeneration.

In addition to CMT, links have been made between MFN2 dysregulation and both Parkinson’s and Alzhei mer’s diseases (Han et al., 2011). Indeed, the frontal cor tex of Alzheimer’s disease patients displays a reduction in MFN2 levels, as well as the hippocampal neurons of post-mortem patients (Wang et al., 2009; Manczak et al., 2011), which is recapitulated in Alzheimer’s dis ease models. In fact, production of amyloid ß-peptide

FIGURE 10 | Pro- and anti-mitophagic roles of mitofusins. Mitofusins are reported to induce and inhibit mitophagy through a variety of processes. Pro-mitophagic roles of mitofusins can be promoted by (1) phosphorylation of MFN2 by PINK1, and subsequent recruitment of Parkin to mitochondria, (2) Gp78-mediated ubiquitylation of MFN1, and (3) SIRT1-mediated deacetylation of MFN2. Anti-mitophagic roles have been described for Mitofusin 2 via (1) MULANmediated ubiquitylation, (2) increase of ER-mitochondria contacts directly mediated by MFN2, and (3) mutations in MFN2 associated with Charcot-Marie-Tooth Type 2A (CMT2A) neuropathy.

(aß), the main component of amyloid plaques caus ative of Alzheimer’s disease, was found to be decreased upon knockdown of MFN2 (Leal et al., 2016). Further more, silencing of MFN2 led to an increase in ER-mito chondria contacts, characteristic of Alzheimer’s (Hedsk og et al., 2013). However, contradictory results suggest that overexpression of MFN2 leads instead to reduction of aß-mediated neuronal cell death (Park et al., 2015). In addition to ER-mitochondria contacts, a protective role of mitophagy has also been suggested (Kerr et al., 2017; Fang et al., 2019). Indeed, neurons affected in Alz heimer’s disease presented compromised mitophagy and accumulated defective mitochondria, consistent with the mitochondrial dysfunction characteristic of Alzheimer’s disease. Importantly, mitophagy induction diminished aß as well as tau hyperphosphorylation, two hallmarks of Alzheimer’s disease (Fang et al., 2019). Sup porting this idea, mitophagy induction prevented cogni tive impairment in an Alzheimer’s disease mouse model and reversed memory impairment in both transgen ic tau nematodes and mice (Fang et al., 2019). Regard ing Parkinson’s disease, the link originates from MFN2 being a target of Parkin. First, loss of function mutations in the genes encoding for Parkin or PINK1 is found at the origin of early-onset PD (Kitada et al., 1998; Valente et al., 2004; Corti et al., 2011). In fact, Parkin mutations are pri marily associated with autosomal recessive Parkinson’s disease and are the most known cause for this neurop athy. PD is characterized by progressive loss of dopami nergic neurons within substantia nigra, which attributes to Parkin an extremely important neuroprotective role. Second, the E3 ligase Parkin ubiquitylates MFN2 (Tana ka et al., 2010), and other OM proteins (Sarraf et al., 2013), being mitofusins rapidly degraded, among other Parkin targets (Chan et al., 2011; Sarraf et al., 2013). How ever, causative evidence for a role of MFN2-dependent mitophagy via PINK1/Parkin in Parkinson’s disease is still missing.

Cardiac Dysfunction

A link between cardiac dysfunction and mitofusin impairment has been established over the years (Hall et al., 2014). First, ablation of mitofusins in cardiomy ocytes, Drosophila, and mice caused cardiomyocyte dysfunction, rapid progressive dilated cardiomyop athy, and finally heart failure (Chen et al., 2011; Dorn et al., 2011; Papanicolaou et al., 2012; Song et al., 2017). Furthermore, heart phenotypes present in the fly were rescued upon expression of human mitofusins, supporting evolutionary conserved roles of mitofusins

in the heart (Dorn et al., 2011). Interestingly, MFN2 was found to have an essential role in mice heart, in the metabolic shift from carbohydrates to fatty acids as the substrate preference, which occurs during peri natal period (Gong et al., 2015). An MFN2 mutant that cannot be phosphorylated by PINK1 and, therefore, inhibits Parkin-mediated mitophagy, prevented this mitochondrial metabolic maturation, and the respec tive hearts maintained the fetal metabolic signature (Gong et al., 2015). The authors suggested that mito chondria in fetal cardiomyocyte undergo MFN2-de pendent mitophagy in order to allow their replace ment by mature mitochondria. This places MFN2 as a central player in this essential metabolic shift. In fact, an active role of MFN2 for mitophagy of cardiac mito chondria had previously been suggested (Chen et al., 2011). Consistently, depletion of MFN1 and MFN2 in cardiomyocytes caused accumulation of defective mitochondria (Song et al., 2015, 2017). Moreover, no increase in mitophagy levels was observed, suggesting that ablation of both mitofusins interferes with a step in the mitophagic process required for the degrada tion of defective mitochondria (Song et al., 2015, 2017). In fact, despite the existence of fusion events in car diac mitochondria (Weaver et al., 2014), they are dis crete organelles rather than the continuous networks observed in other cell types, perhaps explaining the particular importance of mitophagy (Dorn, 2018).

Type 2 Diabetes

A critical role of MFN2 for integrative physiology of whole body energy and glucose metabolism has been proposed, in neurons expressing orexigenic neuropep tide agouti-related protein (Agrp) and neurons express ing anorexigenic pro- opiomelanocortin (POMC; Ozcan, 2013; Dietrich et al., 2014; Schneeberger et al., 2014; Tim per and Brüning, 2017). Moreover, MFN2 has been linked with diabetes and obesity, which are, per se, intimately related (Zorzano et al., 2009). MFN2 is found decreased in skeletal muscle from both obese and type 2 diabet ic patients (Bach et al., 2005), in line with increased mitochondrial fission and reduced mitochondrial size, which are characteristic hallmarks of diabetes type 2 (Kelley et al., 2002; Toledo et al., 2006). Interestingly, body weight loss in obese subjects increased the expres sion of MFN2 in skeletal muscle and rescued mitochon drial size and number (Bach et al., 2005; Toledo et al., 2006). However, in adipose tissues of mice subjected to high-fat diet, the levels of MFN2 were increased, sug gesting tissue-specific responses (Boutant et al., 2017).

Moreover, ablation of MFN2 in adipocytes was bene ficial, as it conferred better tolerance to glucose and protected against high-fat induced insulin resistance (Boutant et al., 2017; Mahdaviani et al., 2017). Interesting ly, a concomitant decrease in mitochondria-lipid drop let interaction was observed, which decreased the lip olytic response of adipose tissues (Boutant et al., 2017). Lipid droplets are organelles present in adipose tissues responsible for the storage of lipid molecules. Upon spe cific metabolic conditions, these organelles release the lipid molecules, which can be further used for ATP pro duction (Londos et al., 1999). Interestingly, the exces sive storage of lipids in these organelles was shown to underlie metabolic disorders such as diabetes and obe sity. These results suggest an interesting role of MFN2 in mitochondria-lipid droplets interaction, which had been previously described (Pidoux et al., 2011; Rambold et al., 2015). Consistently, lack of MFN2 in mice and Dro sophila impairs lipid droplet formation and morpholo gy (Sandoval et al., 2014; Mcfie et al., 2016). Interesting ly, enzymes necessary for specific lipids biosynthesis commonly stored in lipid droplets are found enriched at MAM sites, where human MFN2 was also found to be enriched (De Brito and Scorrano, 2008). The nature of this relation is still poorly understood but might depend on a role of MFN2 in ER-mitochondria contacts, shown to be altered in different models of obesity and insulin resistance (Sebastián et al., 2012; Schneeberger et al., 2014; Filadi et al., 2017).

CONCLUDING REMARKS

The importance of mitochondrial morphology for cel lular fitness is increasingly clear. Moreover, the role of mitofusins as a response hub to different metabolic or stress states, by their selective recognition and modifi cation has also been demonstrated. Ubiquitylation by E3 ligases and also phosphorylation or acetylation modu late a myriad of activating or repressing states in mito fusins. Further studies are required to elucidate con troversial results and decipher the pathophysiological impact of mitofusin modifications in not only neurode generation and cardiac diseases but also energy expen diture-related diseases, such as obesity and diabetes.

For more: https://www.frontiersin.org/articles/10.3389/ fphys.2019.00517/full

Mariana Joaquim – Biologia

Nasceu a 1994, frequentou o ISA de 2012 a 2015, onde participou no núcleo de estudantes de Biologia. Fez o projeto final de licenciatura no IMM sobre modelo celular para cardiomiopatia hipertrófica. Fez o mestrado na FCT em Genética molecular e biomedicina e a tese baseou-se no estudo e investigação da síndrome de Angelman. Após a aquisição do grau de mestre, rumou até à Alemanha onde permanece até hoje, em Colónia, no instituto CECAD, a estudar o Role of mitofusin tool in quality control processes and diseases e onde publicou o artigo apresentado nesta edição da AGROS. Atualmente encontra-se a fazer o Doutoramento pela universidade de Colónia e não sabe o que o futuro lhe reserva.

CONCLUSÃO E AGRADECIMENTOS

Este ano decidimos dedicar a nossa AGROS aos estudantes de todo o Instituto Superior de Agronomia. Como tal, envolvemos toda a comunidade estudantil nesta edição especial. Contámos com a colaboração de alumnis de cada uma das áreas do nosso Instituto e até de várias partes do globo, a quem devemos o nosso maior agradecimento, tendo como principal objetivo mostrar aos estudantes desta geração o mundo para além do ISA.

Todos percorremos caminhos diferentes e todos nos sentimos um pouco perdidos nalgum momento do nosso percurso académico. Serve a presente edição para entre gar alguma esperança às almas mais perdidas e para mostrar um pouco o trabalho que é feito hoje em dia por indivíduos que já pisaram as terras desta grande instituição. Cada área tem o seu interesse, cada artigo tem a sua opinião, cada Alumni tem o seu valor e cada estudante tem a sua vida inteira pela frente.

Esta é A edição realizada por (ex-)alunos para alunos. Espero que tenham gostado tanto dela como nós, tal como esperamos que ajude quem sente que não há mais nada para além dos estudos.

Obrigado a todos os que colaboraram e tornaram possível a reali zação desta edição.

Margarida Próspero