UNIVERSIDADE ESTADUAL PAULISTA “JÚLIO DE MESQUITA FILHO” INSTITUTO DE BIOCIÊNCIAS – RIO CLARO unesp PROGRAMA DE PÓS-GRADUAÇÃO EM CIÊNCIAS BIOLÓGICAS (BIOLOGIA VEGETAL) MECANISMOS DE TOLERÂNCIA AO AL3+ EM PLANTAS COMPARANDO ESPÉCIE DE CERRADO (STYRAX CAMPORUM), LIMOEIRO ‘CRAVO’ (CITRUS LIMONIA CV. CRAVO) E TRIGO (TRITICUM AESTIVUM) Carolina de Marchi Santiago da Silva Tese apresentada ao Instituto de Biociências do Câmpus de Rio Claro, Universidade Estadual Paulista, como parte dos requisitos para obtenção do título de Doutora em Ciências biológicas (Biologia Vegetal) Janeiro - 2018 MECANISMOS DE TOLERÂNCIA AO AL3+ EM PLANTAS COMPARANDO ESPÉCIE DE CERRADO (STYRAX CAMPORUM), LIMOEIRO ‘CRAVO’ (CITRUS LIMONIA CV. CRAVO) E TRIGO (TRITICUM AESTIVUM) Carolina de Marchi Santiago da Silva Orientador: Prof. Dr. Gustavo Habermann 1 Co-orientador: Dr. Ricardo Harakava 2 1 Depto. Botânica, Instituto de Biociências – UNESP – Rio Claro 2 Bioquímica fitopatológica – Instituto Biológico, SAA, SP Silva, Carolina de Marchi Santiago da Mecanismos de tolerância ao Al3+ em plantas comparando espécie de Cerrado (Styrax camporum), limoeiro ‘cravo’ (Citrus limonia cv. cravo) e trigo (Triticum aestivum) / Carolina de Marchi Santiago da Silva. - Rio Claro, 2018 146 f. : il., figs., gráfs., tabs. Tese (doutorado) - Universidade Estadual Paulista, Instituto de Biociências de Rio Claro Orientador: Gustavo Habermann Coorientador: Ricardo Harakava 1. Ecologia. 2. Alumínio. 3. Citros. 4. Transcriptoma. 5. qRT-PCR. 6. Plantas lenhosas do Cerrado. I. Título. 574.5 S586m Ficha Catalográfica elaborada pela STATI - Biblioteca da UNESP Campus de Rio Claro/SP - Adriana Ap. Puerta Buzzá / CRB 8/7987 0 À Clara. 1 AGRADECIMENTOS Agradeço à toda minha família, principalmente ao meu marido Raul, por todo o apoio e incentivo (muito necessário na pesquisa acadêmica). Aos amigos, obrigada por compreenderem as ausências e mesmo assim estarem presentes sempre que precisei. Ao Prof. Dr. Gustavo Habermann pela orientação eficiente mesmo a quilômetros de distância e ao grupo de pesquisa da fisiologia vegetal de Rio Claro (Anna, Brenda, Giselle, Lorena, Mariana, Marina, Matheus, Otávia), ensinando a fazer ciência de alto nível e ao Dr. Ricardo Harakava, meu co-orientador, e ao pessoal da Bioquímica Fitopatológica do Instituto Biológico (Alexander, Amanda, Camila, Cleusa, Julia, Juliana, Marcos, Mariane, Matheus, Mikhail, Orlene, Patrícia, Pedro), pelo apoio na parte de biologia molecular. Ao Dr. Peter Ryan e ao pessoal do CSIRO-Canberra (Akitomo, Chunyan, Deying, Emmanuel, Kumara, Tina) pela receptividade e pelo grande aprendizado em um laboratório de excelência mundial. Ao Instituto de Biociências da UNESP de Rio Claro pela estrutura e formação nesta nova etapa. À Fapesp pela bolsa de doutorado (Processo 2013/11370-3), incluindo período sanduíche no exterior (Processo 2016/00747-7). 2 Sumário Resumo ......................................................................................................................... 4 Abstract ........................................................................................................................ 6 Introdução geral……………………………………………………………………….. 12 I) Transcriptome analysis shows sensitiveness in Citrus limonia roots under aluminum stress ........................................................................................................................... 14 Abstract ................................................................................................................................... 15 Introduction ............................................................................................................................ 16 Material and methods ............................................................................................................. 17 Results ..................................................................................................................................... 20 Discussion ................................................................................................................................ 23 Figures and tables ................................................................................................................... 26 II) Aluminum-induced IAA biosynthesis may explain the Al susceptibility in Citrus limonia ....................................................................................................................... 38 Abstract ................................................................................................................................... 39 Introduction ............................................................................................................................ 40 Material and methods ............................................................................................................. 42 Results ..................................................................................................................................... 48 Discussion ................................................................................................................................ 49 Figures and tables ................................................................................................................... 54 III) Transcriptome analysis shows aluminum-tolerance pathways in Styrax camporum, a moderate Al-accumulator from the Cerrado vegetation ............................................... 64 Abstract ................................................................................................................................... 65 Introduction ............................................................................................................................ 66 Material and methods ............................................................................................................. 67 Results ..................................................................................................................................... 70 Discussion ................................................................................................................................ 72 Figures and tables ................................................................................................................... 75 IV) Styrax camporum, a moderate Al-accumulating plant might rely on mechanisms associated with cell elongation and auxin metabolism to cope with Al ........................ 88 Abstract ................................................................................................................................... 89 Introduction ............................................................................................................................ 90 Material and methods ............................................................................................................. 91 Results ..................................................................................................................................... 95 Discussion ................................................................................................................................ 96 Figures and tables ................................................................................................................. 100 3 V) Does TaALMT1, the major aluminium tolerance gene in wheat (Triticum aestivum L.), also confer tolerance to alkaline soils? ................................................................ 108 Abstract ................................................................................................................................. 109 Introduction .......................................................................................................................... 110 Material and methods ........................................................................................................... 112 Results ................................................................................................................................... 115 Discussion .............................................................................................................................. 117 Figures and tables ................................................................................................................. 121 Conclusões................................................................................................................ 130 Referências ............................................................................................................... 132 4 Mecanismos de tolerância ao Al3+ em plantas comparando espécie de Cerrado (Styrax camporum), limoeiro ‘cravo’ (Citrus limonia cv. cravo) e trigo (Triticum aestivum) Resumo Presente em solos ácidos, o alumínio principalmente encontrado na sua forma tóxica (Al3+) tem como principal efeito a diminuição do crescimento radicular. O Cerrado ocorre no centro-oeste do Brasil sob solos ácidos (pH <4) e com alto teor de alumínio (Al). Citrus limonia cv ‘Cravo’ Osbeck e Styrax camporum Pohl compartilham localização geográfica e características edáficas em que se estabelecem, no entanto exibem respectivamente sensibilidade e tolerância ao Al. O presente trabalho visa elucidar mecanismos envolvidos no estresse por Al3+ a nível de expressão gênica radicular. Para as duas espécies foi feita uma análise de transcriptoma para avaliar os efeitos do Al por uma perspectiva global, e a partir destes resultados alguns genes foram selecionados para avaliação ao longo de experimentos de hidroponia mantido por 60 dias. São discutidas estratégias de tolerância e sensibilidade ao Al nas duas espécies, juntamente com análises de biometria, quantificação hormonal e anatomia que corroboram o comportamento sensível de Citrus e tolerante de Styrax por diversas vias. Além disso, o último capítulo trata da dependência do pH em um importante mecanismo de tolerância ao Al em trigo envolvendo exsudação de ácido orgânico. Palavras-chave: Alumínio, Citros, Transcriptoma, qRT-PCR, Plantas lenhosas do Cerrado 5 6 Al-tolerance mechanisms in plants comparing Cerrado species (Styrax camporum), 'Rangpur' lime (Citrus limonia) and wheat (Triticum aestivum) Abstract Present in acid soils, aluminum mainly found in its toxic form (Al3+) has as main effect the decrease of root growth. Cerrado occurs in central-western Brazil under acid soils (pH <4) and high aluminum content (Al). Citrus limonia (Osbeck) and Styrax camporum (Pohl) share geographic location and edaphic characteristics in where they are settled, however exhibiting respectively Al-sensitivity and Al-tolerance. The present work aims to elucidate the mechanisms involved in Al3+ stress at root gene expression level. For both species, a transcriptome analysis was performed to evaluate the effects of Al in a global perspective, and from these results some genes were selected for evaluation in hydroponic experiments maintained for 60 days. Al tolerance and sensitivity strategies are discussed in both species, together with analyzes of biometrics, hormonal quantification and anatomy data that corroborate the Al-sensitive behavior of Citrus and Al-tolerance in Styrax by several routes. In addition, the last chapter deals with the dependence of pH on one important mechanism of Al tolerance in wheat involving organic acid exudation. Key words: Aluminum, Citrus, Transcriptome, qRT-PCR, Cerrado woody plants 7 8 Introdução geral Alternativas para aumentar a produção de alimentos têm sido estudadas em diversas áreas, já que em decorrência do crescimento acelerado da população mundial, a demanda torna-se maior. Atividades agropastoris teriam que aumentar a produção paralelamente ao crescimento populacional, mas áreas com as características de clima e solo necessárias para agricultura são escassas (Phalan et al. 2013). Quarenta por cento dos solos agriculturáveis são ácidos (pH < 5,0), sendo que na América do sul 60% dos solos têm pH abaixo de 4,0 (Von Uexküll and Mutert 1995). Além de características naturais do solo, chuvas ácidas e o uso de fertilizantes amoniacais contribuem para a acidez do solo (Kochian et al. 2004; Zheng 2010). Devido à larga ocorrência de solos ácidos em grandes proporções de áreas agricultáveis no Brasil (solos do Cerrado e Centro sul), esforços biotecnológicos para tornar possível a atividade agrária nestes solos têm sido feitos (Ryan et al. 2011). A acidez do solo pode diminuir em até 60% a produtividade, dependendo da cultura (Blair et al. 2009). Para minimizar a acidez do solo são usadas técnicas como aplicação de calcário (CaCO3 ou MgCO3 - calagem) e alternância com cultura de leguminosas, técnicas essas com eficácia duvidosa, já que o investimento em calagem ( US$400/ha/ano – Ratter et al. 1997) não é refletido na neutralização do pH em todo o perfil do solo, mas apenas superficialmente (0-20 cm prof.) (Delhaize et al. 2009). Quando pHsolo < 5, aluminossilicatos são ionizados e o alumínio passa à forma ionizada (Al3+), que é tóxica à maioria das plantas (Ulrich 1986). Outro problema de solos ácidos é a indisponibilidade de fósforo (P) para as raízes, porque o P fixa-se em compostos com ferro (Fe3+) e/ou Al3+ (Kochian et al. 2004; Zheng 2010). Altas concentrações de Al3+ no solo são percebidas (bioquimicamente) pelo ápice das raízes, que diminuem o crescimento radicular quase que instantaneamente (Delhaize et al. 2004; Sun et al. 2010). Esta diminuição do alongamento diminui a chance de absorção de água e nutrientes em sub-superfícies do solo. A presença de Al3+ no simplasto causa desordens fisiológicas celulares: aumento da concentração de cálcio (Ca2+) no citosol, que culminam em desorganização do citoesqueleto e deposição de calose (Horst et al. 2010). Na parte aérea da planta, a toxicidade do Al3+ está relacionada ao processo fotossintético (Konrad et al. 2005; Jiang et al. 2008), que tem baixo rendimento devido principalmente ao dano causado às membranas dos tilacóides por espécies reativas de oxigênio (Horst et al. 2010; Ryan et al. 2011). 9 A diminuição no alongamento da raiz é o sintoma de toxicidade de Al3+ mais estudado. O crescimento e desenvolvimento da raiz dependem fortemente da concentração e transporte (distribuição simétrica) de auxinas (Perilli et al. 2012) e etileno no tecido radicular, hormônios vegetais que atuam em sintonia neste órgão (Kang et al. 1971; Swarup et al. 2007). Em solos ácidos, a elevada concentração de Al3+ induz maior síntese de etileno; este hormônio faz com que a distribuição polar de auxina fique alterada e com isso, o alongamento celular não ocorre corretamente, deixando as raízes mais curtas (De Cnodder et al. 2005; Ponce et al. 2005; Sun et al. 2010; Basu et al. 2011) e tortas (Kopittke et al. 2008). Estes últimos autores demonstraram que o Al3+ causa rigidez da rizoderme, enquanto que células das camadas internas do órgão continuam crescendo, causando rupturas e raízes tortas. Isso sugere que se a parede celular tem sua composição alterada (pela perturbação direta de Al3+ e/ou competição por Ca2+), a extensibilidade celular é também alterada. De fato, o grau de metilação das pectinas, pela enzima pectina metil esterase e grupos carboxílicos da parede celular geram cargas negativas que normalmente são estabilizadas por Ca2+ (Carpita, N. C.; McCann 2000). O Al3+, no entanto, se presente no apoplasto, pode competir por estas cargas negativas. O Al3+ ligado à parede celular impede sua expansão, deixando-a mais rígida (Kopittke et al. 2008) e limita a expansão por causar ainda comprometimentos enzimáticos na parede (Horst et al. 2010). Frente a todos os tipos de dano que o Al3+ solúvel no solo pode causar às plantas, estas categorizam-se em três grupos: plantas que são sensíveis ao(s) (efeitos tóxicos do) Al3+; as que são resistentes, com mecanismos para exclusão de Al3+; e as que são tolerantes, com mecanismos para quelar e armazenar o Al3+ em formas menos tóxicas (Ryan et al. 2011). Os mecanismos de resistência e tolerância ao Al3+ podem ser determinados por um ou vários genes (Delhaize et al. 2012) dependendo da espécie, e pode ainda haver herança quantitativa, com fenótipos gradativos de resistência ou de tolerância (Kochian et al. 2004; Delhaize et al. 2012). Por outro lado, o bioma Cerrado que originalmente ocorre (ou ocorria) no centro-oeste do Brasil (originalmente, 23% do território nacional), encontra(va-se) sobre solos ácidos (pH < 4), com alta concentração de Al3+ e baixa disponibilidade de nutrientes, principalmente P, Ca2+, Mg2+ e Zn2+ (Haridasan 2008). Isso faz com que as espécies nativas desta vegetação sejam adaptadas a tais condições edáficas (Haridasan 2008). Assim, as espécies lenhosas do Cerrado são divididas em acumuladoras obrigatórias de Al3+ (não se desenvolvendo na ausência desse nutriente), acumuladoras facultativas e espécies não acumuladoras (Haridasan 1982; Haridasan et al. 1987). 10 Rawitscher (1948) já descrevera o sistema radicular profundo de espécies lenhosas do Cerrado. Franco (1998) e Hao et al. (2008) demonstraram indiretamente que estas espécies mostram sistema radicular profundo. Habermann and Bressan (2011) demonstraram por medidas diretas que espécies de savanas mostram raízes com 1m de comprimento após 100 dias da germinação. Assim, é possível que o Al3+ do solo do Cerrado esteja associado a alguma relação com a biossíntese, transporte e/ou ação de auxinas e/ou etileno, promovendo o crescimento em profundidade dessas raízes. Não que a presença de Al3+ seja essencial para esta característica. Mesmo porque para algumas plantas do Cerrado, o significativo crescimento de raízes não é plástico ao ambiente edáfico, mesmo se cultivadas em solos muito férteis e ausentes em Al3+ (Habermann and Bressan 2011). No entanto, espécies não-acumuladoras de Al3+ mostram crescimento de raízes significativamente maior quando cultivadas em 740 M Al3+ do que se cultivadas na ausência desse nutriente (de Souza and Habermann 2012). Porém, até hoje, a ciência agronômica não mostrou interesse em entender o metabolismo de Al3+ nas espécies do Cerrado (de Souza and Habermann 2012). Por outro lado, nas regiões norte, noroeste e central do estado de São Paulo e sul de Minas Gerais, que englobam áreas originalmente ocupadas pela borda sul da vegetação de Cerrado, a citricultura é desenvolvida com sucesso desde os anos 50 (de Souza and Habermann 2012). Porém os Citrus sp, ao contrário da vegetação nativa do Cerrado, podem exibir alta sensibilidade ao Al3+ solúvel em solos ácidos (Pereira et al. 2000; Jiang et al. 2008). Para resistência a altas concentrações de Al3+ no solo o principal mecanismo já demonstrado consiste em secretar ácidos orgânicos pelas raízes (Kochian et al. 2004); o ácido liga-se ao Al3+ e impede que se ligue à parede celular ou seja absorvido. O tipo de ácido orgânico secretado depende da espécie, e também da especificidade do substrato da enzima extrusora (Ryan et al. 2011), podendo a secreção ser constituída de um ou mais tipos de ácido. A parte aérea também deve ter uma função na resistência ao Al3+ (Wu et al. 2013), mas são necessários mais estudos nesta área. Os principais ácidos orgânicos na função de resistência à toxicidade do Al3+ são malato, citrato e oxalacetato (Kochian et al. 2004). A exsudação de citrato ocorre mais frequentemente nas plantas, talvez porque tenha maior poder para quelar o íon alumínio (Pereira et al. 2010). O citrato é também importante na distribuição de ferro nos organismos vegetais (Magalhaes 2010; Delhaize et al. 2012). Segundo Ryan et al. (2011), o fator limitante para a exsudação de ácidos orgânicos não é a síntese do composto (que é sintetizado continuamente), e sim a 11 extrusão deste. O Al3+ do solo é responsável por ativar a função destas proteínas de transporte (Delhaize et al. 2009) sendo a transdução deste sinal mediada por quinases (Kumari et al. 2008). Para a exsudação de malato na raiz existe a proteína ALMT (do inglês, aluminium activated malate transporter), exclusiva no reino Plantae. Já o citrato é secretado por um canal não específico (pode liberar outros compostos) que ocorre em todos os domínios vivos, pertencente à família MATE (do inglês, multidrug and toxic compound exudation Family) (Magalhaes 2010). O mecanismo de exsudação de oxalacetato ainda não é conhecido. No mecanismo de tolerância, algumas espécies são capazes de acumular grandes quantidades de Al3+ (Mukhopadyay et al. 2012; Negishi et al. 2012). Em plantas acumuladoras, o Al3+ é absorvido pela raiz e complexado ao oxalacetato, depois o Al3+ troca de aceptor, ligando-se ao citrato para transporte no xilema, chegando ao mesofilo, onde o aceptor volta a ser o oxalacetato (Kochian et al. 2004). Segundo Horst et al. (2010), o mecanismo de armazenar o Al3+ no simplasto em complexos menos reativos evita o contato do cátion com a parede e por isso anula o efeito de restrição no alongamento celular. O efeito do Al3+ na expressão gênica tem sido estudado analisando-se genes de síntese de malato e/ou citrato (Horst et al. 2010; Wang et al. 2013), transporte ALMT (Sasaki et al. 2004, 2006; Guo et al. 2007; Houde and Diallo 2008; Pereira et al. 2010), transporte MATE (Liu et al. 2009; Magalhaes 2010), síntese de etileno (Chandran et al. 2008; Sun et al. 2010), distribuição de auxina (Li et al. 2006; Goodwin and Sutter 2009; Sun et al. 2010), homeostase de cálcio (Chandran et al. 2008), genes relacionados à parede celular (Kumari et al. 2008; Maron et al. 2008; Horst et al. 2010), genes reguladores da expressão de outros genes (Delhaize et al. 2012), e genes envolvidos no próprio acúmulo de Al3+ (Delhaize et al. 2012). Embora a diversidade de genes estudados seja relativamente alta, as espécies nas quais são feitos estes estudos resumem-se basicamente a trigo, Arabidopsis, milho e cevada. Diante do exposto, fica claro que há espaço para entendimento do metabolismo do Al3+ em espécies cítricas e, sobretudo, em espécies em que o Al3+ não é tóxico, mas benéfico, como as espécies nativas do Cerrado. Objetivos Ao submeter plantas de limoeiro ‘Cravo’ (Citrus limonia cv. Cravo) e de Styrax camporum (Styracaceae), a diferentes concentrações de Al3+ em solução nutritiva, 12 foram identificadas diferenças no transcriptoma em situações contrastantes e a partir disso quantificada a expressão de diferentes genes relacionados ao metabolismo de extrusão de Al3+ pela raiz (com destaque para o funcionamento do canal TaALMT1 em trigo), metilação em pectinas da parede celular, e relacionados à síntese de etileno (Et) e transporte de auxinas (IAA). Foram testadas as seguintes hipóteses: i) Por se tratar de um espécie sensível ao Al, o transcriptoma de Citrus limonia deve apresentar respostas de mecanismos de tolerância a este estresse; ii) O balanço hormonal em C. limonia, considerando principalmente níveis de auxina e etileno, assim como genes associados à biossíntese e transporte destes hormônios devem sofrer alterações culminando em diminuição do crescimento radicular; iii) De modo análogo, o transcriptoma de Styrax camporum deve apresentar expressão acentuada de genes envolvidos em mecanismos de tolerância à toxicidade do Al, em comparação à C. limonia; iv) Em S. camporum canais de exsudação de ácidos orgânicos para malato e/ou citrato através dos canais ALMT e MATE devem ser induzidos em resposta ao Al durante todo o tempo do experimento levando ao crescimento normal das raízes. Além disso, diferentes respostas devem ocorrer nas raízes expostas a concentrações de Al contrastantes; v) O canal TaALMT1 nas raízes de trigo deve exibir resposta específica à faixa ácida de pH, onde o Al assume sua forma tóxica Al3+; Assim, apresenta-se a Tese de que o Al3+ beneficia o desenvolvimento radicular das espécies do Cerrado, em oposição às espécies agrícolas, cujo crescimento radicular, mediado por IAA e Et, é limitado pelo Al3+. 13 14 CAPÍTULO I Transcriptome analysis shows sensitiveness in Citrus limonia roots under aluminum stress Carolina M.S. Silva1, Alexander Banguela2, Douglas S. Domingues3, Gustavo Habermann3* 1Programa de Pós-Graduação em Ciências Biológicas (Biologia Vegetal), Universidade Estadual Paulista, UNESP, Instituto de Biociências, Departamento de Botânica, Av. 24-A, 1515; 13506-900, Rio Claro, SP, Brazil; 2Centro de P&D de Sanidade Vegetal, Laboratório de Bioquímica Fitopatológica, Instituto Biológico, Av. Conselheiro R. Alves, 1252; 04014-002, São Paulo, SP, Brazil; 3Departamento de Botânica, Universidade Estadual Paulista, UNESP, Instituto de Biociências, Av. 24-A, 1515; 13506-900, Rio Claro, SP, Brazil Tel: +0055 (19) 3526-4210; e-mail: ghaber@rc.unesp.br (*Corresponding author) 15 Abstract Under acidic conditions toxic aluminum (Al3+) decreases root growth in sensitive species. A wide range of Al-tolerance/sensibility occurs in Citrus genus, but the most common rootstock in Brazilian citriculture (‘Rangpur’ lime) has been not widely studied yet. Root apexes from ‘Rangpur’ lime (Citrus limonia) plants grown in a nutrient solution containing different Al concentrations for 60 days were analyzed by RNA-seq and some differentially expressed genes were validated by qRT-PCR. We found 99 genes up-regulated in response to Al, highlighting citrate synthesis and its exudation by MATE (multidrug and toxic compound exudation) transporter. Genes related to secondary metabolism, pectin methylesterification, auxin response, defense to biotic and abiotic stresses, cell division, suberin deposition and nitrate uptake were also induced by Al3+ in root nutrient solutions. The overview of up-regulated genes in C. limonia roots infer that this species may be considered an Al-sensitive rootstock, despite some tolerance responses induced. Key words: Al3+; Gene expression; ‘Rangpur’ lime, MATE transporter 16 Introduction In neutral soil conditions aluminum (Al) occurs mainly in Al3SiO4 form. In acidic soils (pH < 5) this complex is hydrolyzed to different Al species mainly Al3+, which is toxic to most plant species (Kochian 1995). Plants sensitive to Al (barley for example) shows direct effects such as inhibition of root growth, leading to lower root reach and impacting crop productivity (Kochian 1995; Horst et al. 2010; Sun et al. 2010). Root growth reduction induced by Al is documented in several species, i.e., wheat (Delhaize et al. 1993), maize (Ryan et al., 1993), barley (Furukawa et al. 2007), Arabidopsis (Hoekenga et al. 2003) and woody species (Brunner and Sperisen 2013), including Citrus trees (Lin and Myhre 1991). Decrease in root growth may be caused mainly by physic obstacle (cell wall become stiffened due to Al binding) or hormonal effects (increase in ethylene synthesis leads to auxins imbalance) (Sun et al. 2010). Al-tolerant species exhibits distinct strategies to cope with such disturbance. The main mechanism of resistance consists in organic acid (malate, citrate and oxaloacetate) efflux preventing Al absorption or binding (Kochian et al. 2015). Malate efflux is mediated by the aluminum activated malate transporters (ALMT) membrane channels (Sasaki et al. 2004; Sharma et al. 2016) and a non-specific channel that belongs to the multidrug and toxic compound exudation (MATE family), operates citrate efflux (Magalhaes 2010). While ALMTs occurs only in plants (Sharma et al. 2016), MATE genes are widely distributed in bacteria, fungi and mammals besides plants (Omote et al. 2006). The cell wall composition also plays a role in Al-tolerance, where less negative charged carboxylic groups due to higher pectin methylation avoid Al binding (Horst et al. 2010). The action of enzymes like pectin methylesterase (PME) leads to free pectic carboxylic groups, allowing Al to bind to pectic nets in the cell wall (Horst et al. 2010). Once Al binds to the cell wall, it becomes rigid, making cell elongation more difficult. Therefore, any enzyme that interfere in PME activity could decrease Al binding in the cell wall and, consequently, enhance Al resistance (Schmohl et al. 2000). Auxins play a central role in root growth (Cleland 2010) but its interaction with Al is still not clear. Disruption in polar auxin distribution was leading to arrest of root elongation in Arabidopsis Sun et al. (2010) and, on the other hand, Yang et al. (2014) found an Al-induced increased biosynthesis of indol-acetic acid (IAA) in this species 17 root apex. Nevertheless, in roots high concentration of auxin can cause strong inhibition of cell elongation (Rahman et al., 2007). The Citrus genus present a wide range of responses to Al, including transcriptional differences (Lin and Myhre 1991; Guo et al. 2017). Citrus limonia L. Osbeck (‘Rangpur’ lime) is widely used as rootstock in subtropical regions of the Americas, due its ability to cope with abiotic stresses, including drought (Ribeiro and Machado 2007). However, this species is considered sensitive to Al (Pereira et al. 2000; Santos et al. 2000). Although recently our group detected a CO2 assimilation rate reduction in leaves from C. limonia plants during the exposition to 1480 µm Al (Banhos et al. 2016a) its impact on the molecular mechanisms in roots are still unknown. In this way, studies that describe C. limonia response to Al, at physiological or molecular level, could enable breeding programs. High-throughput mRNA sequencing is one efficient approach to uncover the plant transcriptome in response to Al in species as Fagopyrum tataricum (Zhu et al. 2015), Anthoxanthum odoratum (Gould et al. 2015) and some Citrus species (Guo et al. 2017). This approach is essential to discover key genes that are regulated under Al toxicity. In this study, we analyzed the Rangpur lime root transcriptome in response to Al toxicity, having in mind the wide use of C. limonia as a rootstock in the Americas and the distinct molecular pathways in Citrus species to cope with Al stress, expecting to uncover candidate genes related to differential Al response within Citrus genus. Material and methods Plant materials and growth conditions This experiment followed the same rationale our group did in a previous study (Banhos et al. 2016a). We used three-month-old (five leaves) ‘Rangpur’ lime (Citrus limonia) plants (7 ± 0.5 cm-high). Plants were kept in a hydroponic system and grown directly on an aerated nutrient solution inside opaque plastic boxes (50 cm in length x 30 cm in width x 15 cm in height; 20 L). The nutrient solution, adapted from Clark (1975), consisted of the following macronutrients (in mM): N (NO3 -) 0.93; N (NH4 +) 0.40; P, 0.013; K, 0.90; Ca, 1.37; Mg, 0.33; S, 0.25; and micronutrients (in µM): Cl, 231.0; Fe (EDTA), 23.3; B, 8.33; Mn, 2.91; Zn, 0.76; Cu, 0.32; Mo, 0.31. This solution contained 0, and 1480 µM Al provided using AlCl3 6 H2O. The pH of the solution was monitored daily and 18 maintained at 4.0 ± 0.1 to keep Al as soluble as possible, and the solution was totally replaced every 15 days. Expanded polystyrene (Isopor®) 50 x 30 cm plates (2-cm thick), with eight holes (2.5 cm in diameter) each, were floated on the nutrient solution in the boxes, and the plants were fixed in these holes with polyurethane foam strips that were placed around the plant collar. The boxes stayed on benches (80 cm from the ground) inside a greenhouse with semi-controlled conditions (air temperature 28.5 ± 0.7C; relative humidity 63.3 ± 1.3%; 853.4 ± 175.1 µmol photons m-2 s-1; approximately 14h of natural photoperiod). After 60 days, a mixture of root tips (~ 0.5 cm) were sampled in 0 and 1480 µM Al conditions. The same experiment was repeated twice: the first experiment provided samples for RNA-seq and samples from the second experiment were used to validate gene expression in qRT-PCR. Determination of Al accumulation Roots were washed under deionized water and oven-dried at 60◦C to constant dry mass. These samples were ground and digested in a nitric:perchloric acids solution (v:v). The concentration of Al was quantified colorimetrically (Sarruge and Haag 1974). RNA extraction and transcriptome analysis Apexes were collected, frozen in liquid nitrogen and RNA was extracted using the RNeasy plant mini kit (Qiagen, Hilden, Germany). Total RNA was precipitated with sodium acetate and ethanol. All samples had a A260/A280 nm ratio > 1.99. Samples were sent to transcriptome sequencing (RNA-seq) at BGI Tech Solutions (Hong Kong, China). Two samples (bulk of root apexes at 0 and 1480 µM Al) were used for library construction, quality assessment (Agilent 2100 Bioanalyzer) and RNA sequencing in an Illumina HiSeq2000 equipment (100bp, pair-end sequencing). Raw reads from Illumina sequencing were submitted to quality control tests prior analysis. They were filtered using SOAPnuke (https://github.com/BGI-flexlab/SOAPnuke). Reads consisting only of adaptor sequences, reads which the percentage of unknown bases is greater than 10% and reads where the percentage of the low-quality bases (base with quality value ≤ 10) is greater than 50% in a read were discarded. 19 Clean reads were mapped to the sweet citrus genome (Xu et al. 2013) by using the Short Oligonucleotide Analysis Package (SOAPaligner/SOAP2) (Li et al. 2009), allowing up to five mismatches. Gene expression levels were calculated using RPKM (Reads per kilobase transcriptome per million mapped reads) (Mortazavi et al. 2008). Digital expression profiles were obtained based on Audic and Claverie (1997) test. Genes with false discovery rate ≤ 0.05 and the absolute value of Log2Ratio ≥ 1 were considered differentially expressed genes (DEGs). The DEGs list was submitted to KEGG (Kyoto Encyclopedia of Genes and Genomes) database analysis (Kanehisa et al. 2007). They were also submitted to Gene Ontology (GO) enrichment analysis. All DEGs were assigned to GO terms in the database (http://www.geneontology.org/), calculating gene numbers for every term, then uses hypergeometric test to find significantly enriched GO terms in the input list of DEGs, based on 'GO:TermFinder (http://smd.stanford.edu/help/GO-TermFinder/GO_TermFinder_help.shtml). We opted to highlight only most proeminent differences between DEGs and all genes from the database. qRT-PCR validation For qRT-PCR analysis, five replicates were used to quantify gene expression. Each biological replicate consisted of one C. limonia plant under control (0 µM Al) or Al condition (1480 µM Al). Each biological sample was run in three technical replicates. Total RNA was extracted from root samples using the RNeasy plant mini kit (Qiagen, Hilden, Germany). Two micrograms of total RNA were treated with TURBO DNA-freeTM (Ambion, Carlsbad, USA) and reverse transcribed using an oligo-dT primer and Super ScriptTM III (Life Technologies, Carlsbad, USA), according to the manufacturer’s protocol. cDNAs were used to qRT-PCR analyses, where amplification conditions were the same of amplification efficiency tests. We validated using quantitative PCR the transcriptional profile of seven genes (Wang et al. 2014): MATE (Cs7g01770), CS (Cs7g01170), PMEI (Cs2g01440), SAUR (orange11t04220), SLAH (Cs3g25590), Peroxidase (Cs2g28810) and Alcohol acyltransferase (Cs2g29820) since they are candidate genes related to Al tolerance and showed differential expression level in RNA-seq analysis. Based on the reference genome, we designed primers for qPCR (Table 1). As reference genes we used GAPC2 and EFα, as recommended by Mafra et al. (2012). 20 Five serial dilutions of genomic DNA from Citrus limonia were amplified using the internal primers in order to access the amplification efficiency during qPCR analysis. GoTaq® qPCR Master Mix with SYBR® fluorescence (Promega, Madison, USA) was used as the reagent in these tests, with the following cycle specifications: 95C for 5 min (1 cycle), 95C for 10 s, 58C for 30 s, 72C for 30 s (40 cycles) followed by a melting curve analysis (1% slope temperature; 60–95C), performed in a 7500 Fast real time PCR system (Applied Biosystems, Foster City, USA). Amplification efficiencies are presented in Table 1. qPCR data was analyzed based on the procedure of Pfaffl (2001). Expression was normalized using the geometric mean of the RQ values of GAPC2 and EFα. A one-way analysis of variance (Anova) was performed between plants exposed to 0 and 1480 M Al to test for differences in length, dry weight, [Al] and gene expression in roots after 60-days treatment. The Tukey test ( = 0.05) was used to conduct post-hoc comparisons to estimate the least significant differences between mean results. MATE phylogenetic tree Phylogenetic analysis and tree visualization were developed in Geneious software version 11.0 (Kearse et al. 2012). ClustalW was used for the alignment of 30 MATE proteins from plants and one from the Diplocarpon roseae fungus as tree outgroup. The phylogentic tree was generated by Bayesian inference MCMC sampling in MrBayes (Ronquist and Huelsenbeck 2003) using the LG +I +G substitution model with four rate categories, chosen under the BIC in ProTest (Abascal et al. 2005). Four chains were run for 50 million of generations and trees were sampled every 5000 generations, with the first 5 00000 discarded as burn-in. GenBank accessions of the protein sequences used were included in the supplementary data. Results Transcriptome analysis of Citrus limonia plants growth in Al conditions When compared with plants not exposed to Al, the aluminum content in this hydroponic experiment resulted in higher Al accumulation in roots. Moreover, root showed reduced length and dry weight (Figure 1). 21 We obtained 25,087,582 and 24,657,760 clean reads for 0 and 1480 µM Al libraries respectively, where most of them successfully mapped against Citrus sinensis genome (Table 2). Transcriptome analysis of root apexes from plants incubated with 0 and 1480 µM of Al presented 16,750 DEGs, being 99 genes up-regulated and 16,651 down-regulated by Al toxicity. In order to identify transcriptional mechanisms triggered by Al, we compared differentially up-regulated genes in Citrus limonia to those found in Citrus sinensis and Citrus grandis under similar conditions (Guo et al. 2017). A total of 20 genes were upregulated in Citrus limonia and at least one Citrus species (C. grandis and/or C. sinensis) (Figure 2). Only two genes share upregulation in the three species: multidrug and toxic compound extrusion channel (MATE - Cs7g01770) and nuclear transport factor 2 (NTF2 - Cs6g09500) (Supplementary Table II). DEGs annotation The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used to analyze DEGs, resulting in 209 categories. We selected a total of 11 categories (Table 3) that have a significant difference of genes present in DEGs. Citrate cycle, Oxidative phosphorylation and Calcium signaling were the more represented pathways in DEGs categories responding to Al. On the other hand, genes belonging to Cell cycle, RNA polymerase and Apoptosis categories were more significant (higher percentage) in the genome than in DEGs, which means secondary importance to the imposed treatment (Figure 3). A total of 13,587 genes from the genome and 8,538 DEGs were assigned to Gene Ontology (GO). Comparison of the 17 gene ontology categories enriched in DEGs with the whole C. sinensis genome (Figure 3), indicated that 10 categories had values twice higher in DEGs than in genome, including post replication repair, dicarboxylic acid transport and amide transmembrane transporter activity. Smaller differences can be found in categories as small ribosomal subunit and intrinsic to organelle membrane, which had only 25% increase in DEGs compared to the genome. Gene ontology analysis using the 16,750 differentially expressed genes showed Cellular metabolic processes - biological process category as the most abundant in DEGs, followed by Ion binding - cell components category and Nucleus - molecular function category, responsible for 60.4, 34.1 and 32.2 % of genes respectively (Figure 4). In general, categories belonging to biological process were more frequent (higher percentage). Categories as Organic anion transmembrane transporter activity, Ethylene 22 binding and Auxin binding which are recognized in Al-tolerance (Sun et al. 2010; Ryan et al. 2011) and also occurred in C. limonia transcripts. On Ion binding category, we could also identify genes responsible for Zinc ion binding (6.3% of genes), Cobalt ion binding (0.4% of genes) and Copper ion binding (1.6% of genes), showing that Al-response is common to other metal toxicities. On the same way, Response to stress category may include abiotic stresses as Response to salt stress (5.1% of genes), Radiation response (8.1% of genes), Heat response (2.1% of genes), Response to osmotic stress (5.4% of genes) and Response to oxidative stress (4.1% of genes). In both databases (KEGG and GO) tricarboxylic acid cycle genes were shown markedly responsive to Al stress. Citrate synthase (CS) - a member of tricarboxylic acid cycle - was found in transcriptome analysis and validated by qRT-PCR as up-regulated by Al treatment (Figure 5). Validation by qRT-PCR of DEGs directly related to Al response Transcriptome results were validated considering a group of genes related to Al response including MATE transporter, pectin methylesterase inhibitor (PMEI), CS, small auxin up-regulated RNAs (SAUR), S-type anion channel (SLAH), peroxidase and Alcohol acyltransferase. In these qRT-PCR validation CS, MATE, peroxidase and SLAH gene expression were confirmed to be up-regulated while PMEI and SAUR were down-regulated by Al stress (Figure 6). Alcohol acyltransferase gene expression resulted in opposite behavior, showing up-regulation at transcriptome analysis and down-regulation at qRT-PCR validation. After 60 days on hydroponic treatment the foldchange comparing 0 and 1480 µM Al resulted in high MATE up-regulation (215.7 foldchange) which was much more increased than SLAH (34.4 foldchange), CS (9.1 foldchange) and peroxidase (4.6 foldchange). PMEI and SAUR genes showed similar down-regulation rates (-18.9 and - 20.6 foldchange respectively), that showed more pronounced reduction than alcohol acyltransferase (-1.4 foldchange) (Figure 6). Phylogenetic analysis of Citrus limonia MATE Using C. sinensis corresponding gene as reference (Wang et al. 2014 - Cs7g01770), we obtain the MATE gene of Citrus limonia (GenBank - MF680000). Due to MATE generalist function, a phylogenetic analysis based in Bayesian inference was developed including reported MATE proteins and putative proteins, as well as the new 23 C. limonia MATE putative protein. In the rooted tree resulted from the phylogenetic analysis two well sustain clades were observed clustering Al and non-Al responsive MATE (Figure 7). C. limonia protein clustered at MATE group responsible for Al- dependent citrate exudation in roots (Figure 7). Discussion The main plant response to Al stress is the decrease in root growth, which may occur in varying magnitudes. In our experiments, we found that C. limonia roots were 44% short in length and 50% lighter in dry weight when imposed to Al treatment (Figure 1B and 1C), which is related to Al content almost seven times higher than in control conditions (Figure 1A). Citrus grandis Al content difference was lower (almost three times higher under Al conditions) thus leading to a subtle decrease in root dry weight (33% smaller under Al conditions) while no statistical difference was observed in C. sinensis root dry weight (Guo et al. 2017). Although in different Al concentration and exposure time, these results suggest that C. limonia might be even more sensitive than C. grandis to Al stress. After 60-days treatment most genes were inhibited, which indicated the inhibition of many physiological pathways. From the 99 up-regulated genes, 15 were only shared with C. grandis up-regulated genes, thus the higher similarity with C. grandis transcriptome could also be an indicative of sensitive profile in both species (Supplementary Table II). Three genes were only shared between C. limonia and C. sinensis, and two of them were shared among these three species – NTF2 and MATE (Figure 2). Shared between C. limonia and C. sinensis, alcohol acyltransferase belongs to HXXXD-type acyltransferase group, which forms compounds on suberin and cutin polymers, conferring a barrier to the roots (Kosma et al. 2012). The up-regulation of this gene in C. sinensis found by Guo et al. (2017) is more evident than in C. limonia, once in the last, it appears up-regulated in RNA-seq but down-regulated in qRT-PCR analysis, thus conferring more Al-tolerance to the C. sinensis. Transcriptional profiles of C. limonia and C. grandis were closely related. Among the 15 genes up-regulated by Al in both species we first highlight SLAH channels, involved in nitrate uptake (Wang et al. 2012) also allowing normal growth in acidic conditions (Zheng et al. 2015). Chalcone and stilbene synthase also deserve highlight due its involvement in secondary metabolism (Nopo-Olazabal et al. 2014). 24 Triggered by Al, fungus and bacteria defense-related genes were up-regulated in all the three Citrus species studied. In C. sinensis and C. grandis osmotin and thaumatin genes were respectively activated, while in C. limonia both genes were up-regulated exercising this function (Liu et al. 1994; Datta et al. 1999). Ran, a small GTP-binding protein, is imported to nucleus when complexed to NTF2 triggering various signaling pathways, including cell proliferation by controlling mitotic cycle (Vernoud et al. 2003). When NTF2 was over-expressed in Arabidopsis this import was disrupted (Zhao et al. 2006). Thus, by NTF2 up-regulation in these three Citrus species, we may assume that Al lead to less mitotic division in roots even in C. sinensis that showed no diminution in root dry weight (Guo et al. 2017). MATE channel was also up-regulated in these three Citrus species. Notwithstanding, here transcriptome data was validated by qPCR while C. sinensis and C. grandis studies validation were accomplished by quantification of organic acid (OA) secretion (Guo et al. 2017). In C. limonia we also found ABC transporters which were previously related to Al toxicity in roots, having function similar to MATE channel – detoxification of organic compounds (Yokosho et al. 2014; Chen et al. 2015). OA exudation is a common mechanism in tolerant species which consists in release citrate, malate or oxaloacetate to chelate Al3+. These compounds are intermediates in tricarboxylic acid cycle (TCA) therefore, these cycle constituents have markedly importance in Al tolerance. In fact, Dicarboxylic acid transport and TCA were among the categories more enriched in response to Al (Table 3 and Figure 4) when comparing DEGs set and genome gene ontology. Analysis of CS expression by qRT- PCR, as a transcript representative of TCA category, confirm its up-regulation under Al stress (Figure 5). However, in OA efflux the channel responsible for exudation is considered more consistent with Al resistance than OA synthesis (Ryan et al. 2011), meaning that to cope with Al stress MATE channel expression should have higher impact than CS in citrate efflux. In fact, Al-increased MATE expression was more than 20 times higher than Al- increased CS expression in C. limonia roots (Supplementary Figure I). Also, CS was not up-regulated by Al in C. sinensis nor C. grandis roots even with OA secretion increased (Guo et al. 2017). Citrus species show a wide range of Al-tolerance (Lin and Myhre 1991; Pereira et al. 2000). Due to higher ALMT gene expression and higher OA efflux C. sinensis was considered more tolerant to Al than C. grandis (Yang et al. 2012; Guo et al. 2017). Despite a markedly increase in MATE gene expression in C. limonia roots (which 25 indirectly means increasing citrate efflux – Figure 7), the inhibition on root growth make this species more Al-sensitive than other Citrus species previously studied. A possible cause for root growth decreased relies on PMEI low gene expression. Cell wall pectin undergoes partial apoplastic demethylesterification through PME action, resulting in the exposure of free pectic carboxylic groups, which could serve as binding sites for Al in the wall (Zhu et al. 2014). Potato transformants that exhibited higher expression of PME accumulate more Al, produce more callose and their root growth are more inhibited when exposed to Al than wild types (Schmohl et al. 2000). Analogously, lower PMEI expression favors Al binding in cell wall sites. This assumption can be verified by the relation between PMEI down-regulation (Figure 5) and higher Al content in roots with 1480 µM Al in nutrient solution (Figure 1 B). A second possible cause that results in reduced C. limonia root growth is related to SAUR gene expression. These genes are induced by auxins and act in apoplast acidification (Spartz et al. 2014), allowing cell wall expansion. Our data show SAUR down-regulation (more than 20 times fold change – Figure 5) under Al stress, which contributes to reduced root growth. In short, gene expression analysis of C. limonia plants exposed to Al stress, reveals that high MATE and CS expression (i.e. high citrate efflux) is not enough to allow normal root growth in this species. Moreover, suggest that PMEI and SAUR genes down-regulation could be related with the root growth inhibition observed, by the modification of cell wall affinities and auxin responses respectively. Acknowledgements We acknowledge the Brazilian National Council for Scientific and Technological Development (CNPq) for financial support (474169/2013-8 grant to GH). We extend acknowledgements to the São Paulo Research Foundation (Fapesp) for PhD scholarships granted to C.M.S. da Silva (Fapesp #2013/11370-3). We thank the Sanicitrus Nursery (Araras, São Paulo state, Brazil) for providing us with the ‘Rangpur’ lime plants. 26 Figures and tables Table 1. Specific primers designed for qPCR validation using C. limonia root cDNA. Gene ID Primer F Primer R Efficiency (%) Amplicon size (bp) MATE Cs7g01770 CATTGACACAGCATTTATCGGC CAAATGAGGTTGTAATATTAACTAATGGG 91.85 128 CS Cs7g01170 CGCTAAGCCAGATGGAGAAC ACAGTGGCACGATCTCTCAA 85.38 119 PMEI Cs2g01440 CCACAAGAACGACAGCGATA TTAGCGTAACTCGGCAAGGT 97.22 147 SAUR orange11t04220 TGGGTTCACAACTCACAAGC GAACAATACCAGGCAAACG 91.71 142 SLAH Cs3g25590 TGGTCAGTGGTTCACGAAAG TCTTTCCACCCCATGTTAGC 106.84 119 Peroxidase Cs2g28810 TGCTGGTCTTGTGAGAATGC GACTTCAAACCCTCGCAGAC 109.40 138 Alcohol acyltransferase Cs2g29820 TATTTCCAATTTGGCAACGAC CGATTTGTGAGCCATGTTG 96.45 124 EFα Cs8g16990 TCAGGCAAGGAGCTTGAGAAG GGCTTGGTGGGAATCATCTTAA 100.04 80 GAPC2 Cs2g24610 TCCTATGTTTGTTGTGGGTG GGTCATCAAACCCTCAACAA 97.24 144 27 Table 2. Summary of clean reads and mapping against reference genome from control and Al-treated ‘Rangpur’ lime roots Sample Clean reads Genome map rate (%) Perfect match Expressed genes 0 µM Al bulk 25087582 18145177 (72.33%) 8472000 22099 1480 µM Al bulk 24657760 21290207 (86.34%) 13654519 21199 28 Table 3. Genes classification according to KEGG database to metabolic pathways. DEGs pathways means differences in root gene expression after 60 days cultivated in nutritive solutions with 0 and 1480 µM Al. KEGG pathway DEGs (total mapped 8489) Genome (total mapped 14955) P value Oxidative phosphorylation 72 (0.85%) 94 (0.63%) 4.66E+00 Citrate cycle (TCA cycle) 41 (0.48%) 52 (0.35%) 0.0007 Ubiquitin mediated proteolysis 161 (1.9%) 250 (1.67%) 0.0079 Calcium signaling pathway 40 (0.47%) 58 (0.39%) 0.0387 DNA replication 48 (0.57%) 72 (0.48%) 0.0556 Starch and sucrose metabolism 33 (0.39%) 51 (0.34%) 0.1573 ABC transporters 109 (1.28%) 181 (1.21%) 0.1926 Cell cycle 123 (1.45%) 277 (1.85%) 0.9999 RNA polymerase 100 (1.18%) 240 (1.6%) 0.9999 Apoptosis 194 (2.29%) 469 (3.14%) 1 29 Figure 1. Al accumulation (A), length (B) and dry weight (C) in ‘Rangpur’ lime roots after 60 days cultivated in nutrient solution containing 0 and 1480 µM Al. Distinct letters indicates significant differences (p < 0.05) between treatments. Bars represent standard deviation. 30 Figure 2. Common up-regulated genes in roots of three Citrus species under aluminum treatment. 15 2 3 208 31 Figure 3. Gene ontology (GO) selected categories obtained for differential expressed genes (DEGs) and whole the genome. DEGs mean differences in root gene expression after 60 days cultivated in nutritive solutions with 0 and 1480 µM Al. 32 Figure 4. Gene ontology (GO) of selected differential expressed genes (different root gene expression after 60 days in nutritive solutions with 0 and 1480 µM Al). GO was divided in three groups: A – Biological process (8532 genes in total), B – Cell components (8109 genes in total) and C – Molecular function (8885 genes in total). 33 Figure 5. ‘Rangpur’ lime rootstock MATE, CS, PMEI and SAUR gene expression analysis by RNA-seq (in reads per kilobase per million mapped reads - RPKM) and real time PCR (qRT-PCR) after 60 days in nutritive solutions with 0 and 1480 µM Al. Distinct letters indicates significant differences (p < 0.05) between treatments. Bars represent S.D. 34 Figure 6 - Gene expression of multidrug and toxic compound exudation (MATE) transporter, citrate synthase (CS), pectin methyl esterase inhibitor (PMEI), small auxin up-regulated RNAs (SAUR), Peroxidase, S-type anion channel(SLAH) and Alcohol acyltransferase in root tips of ‘Rangpur’ lime plants grown in nutrient solutions containing 0 and 1480 µM Al at 60 days after planting. Bars represent standard difference. 35 Figure 7. Maximum clade credibility tree of multidrug and toxic compound extrusion (MATE) proteins. Putative Citrus limonia MATE protein (Cl-MATE) is shown in red cluster with the plants Al-responsive MATE group highlighted in red and non Al-responsive MATE in blue highlight branch. A putative MATE from the ascomycete fungus Diplocarpon roseae was used as outgroup. Posterior probability values are indicated over each branch. Phylogenetic tree was generated by Bayesian inference of 31 MATE protein sequences using the MrBayes plugin in Geneious software. GenBank accessions of the protein sequences are in the Material and methods section. 36 Supplemetary Table I – Amino acid sequences used to construct the phylogenetic tree. Responsiveness to Al based on Liu et al.(2016), except for Citrus species. Name Accession code Plant species Cluster SlMATE(MTP77) BE354224 Solanum lycopersicum Not responsive to Al MtMATE2 HM856605 Medicago truncatula AtFFT At4g25640 Arabidopsis thaliana RcEEF49069 EEF49069 Ricinus communis VvAM1 FJ264202 Vitis vinífera VvAM3 FJ264203 Vitis vinífera PtMATE XP_002302594 Populus trichocarpa BrTT12 ACJ36213 Brassica rapa AtTT12 At3g59030 Arabidopsis thaliana MtMATE1 FJ858726 Medicago truncatula NtMATE1 AB286961 Nicotiana tabacum AtDTX1 At2g04070 Arabidopsis thaliana Nt-JAT1 AM991692 Nicotiana tabacum AtADS1 At4g29140 Arabidopsis thaliana AtZF14 At1g58340 Arabidopsis thaliana AtEDS5 At4g39030 Arabidopsis thaliana Al responsive CsMATE Cs7g01770 Citrus sinensis ClMATE MF680000 Citrus limonia EcMATE1 AB725912 Eucalyptus camaldulensis GmFRD3b EU591741 Glycine max AtFRD3 At3g08040 Arabidopsis thaliana SbMATE ABS89149 Sorghum bicolor OsFRDL4 Os01g0919100 Oryza sativa OsFRDL1 AB571881 Oryza sativa HvAACT1 BAF75822 Hordeum vulgare TaMATE1B KC152457 Triticum aestivum ScFRDL1 AB571881 Secale cereale BoMATE KF031944 Brassica oleracea AtMATE At1g51340 Arabidopsis thaliana ZmMATE1 FJ015156.1 Zea mays ScFRDL2 AB571882 Secale cereale 37 Supplementary Table II – List of up-regulated genes shared with one or two Citrus species. Up-regulation shared with Citrus limonia ID Description Citrus sinensis Citrus grandis Cs3g25590 S-type anion channel SLAH1 X Cs2g29820 Alcohol acyltransferase X Cs3g20680 Chalcone and stilbene synthase family protein X orange1.1t01309 RPOB [Arabidopsis thaliana] X orange1.1t00666 - X orange1.1t03318 AT2G01021 [Arabidopsis thaliana] X Cs4g16600 Cytochrome F X orange1.1t03320 - X Cs1g26280 Alpha/beta-Hydrolases superfamily protein [Arabidopsis thaliana] X orange1.1t01308 - X orange1.1t05667 CYP71A20 [Arabidopsis thaliana] X orange1.1t05688 CYP71B37 [Arabidopsis thaliana] X Cs3g09470 Hypothetical chloroplast RF21 X orange1.1t03024 Pathogenesis-related thaumatin superfamily protein [Arabidopsis thaliana] X Cs7g01770 MATE efflux family protein [Arabidopsis thaliana] X X Cs8g03430 Basic form of pathogenesis-related protein 1-like X Cs6g09500 Nuclear transport factor 2 (NTF2) family protein [Arabidopsis thaliana] X X Cs8g07390 protein HYPER-SENSITIVITY- RELATED 4-like X Cs3g27360 Pentatricopeptide repeat-containing protein At2g01860 X Cs5g10550 Osmotin precursor [Arabidopsis thaliana] X 38 CAPÍTULO II Aluminum-induced IAA biosynthesis may explain the Al susceptibility in Citrus limonia Carolina M. S. Silva1, Mariana F. Cavalheiro1, Anna C. G. Bressan1, Brenda M. O. Carvalho1, Otavia F. A. A. Banhos1, Eduardo Purgatto2, Ricardo Harakava3, Francisco A. O. Tanaka4, Gustavo Habermann5* 1Programa de Pós-Graduação em Ciências Biológicas (Biologia Vegetal), Universidade Estadual Paulista, UNESP, Instituto de Biociências, Departamento de Botânica, Av. 24-A, 1515; 13506-900, Rio Claro, SP, Brazil; 2Departamento de Alimentos e Nutrição Experimental/NAPAN/FoRC- Food Research Center, Universidade de São Paulo, USP, Faculdade de Ciências Farmacêuticas, Av. Prof. Lineu Prestes, 580, bl 14, 05508-000, São Paulo, SP, Brazil; 3Centro de P&D de Sanidade Vegetal, Laboratório de Bioquímica Fitopatológica, Instituto Biológico, Av. Conselheiro R. Alves, 1252; 04014-002, São Paulo, SP, Brazil; 4Escola Superior de Agricultura “Luiz de Queiróz” -Universidade de São Paulo, ESALQ-USP, Departamento de Fitopatologia e Nematologia, Av. Pádua Dias, 11, 13418-900, Piracicaba, SP, Brazil; 5Departamento de Botânica, Universidade Estadual Paulista, UNESP, Instituto de Biociências, Av. 24-A, 1515; 13506-900, Rio Claro, SP, Brazil Tel: +0055 (19) 3526-4210; e-mail: ghaber@rc.unesp.br (*Corresponding author) 39 Abstract In acidic soils (pH < 5.0), aluminum (Al) occurs as Al3+, which is phytotoxic and reduces the root growth by hormonal imbalance and/or cell wall rigidity. However, the explanations for the decrease in root growth are not clear. A 60-day study was held with ‘Rangpur’ lime (Citrus limonia) plants grown in a nutrient solution containing different Al concentrations. We measured plant biometric data and used root apexes to analyze auxin (IAA) and ethylene quantification, expression in some Al-responsive genes and anatomic profiles. We found up-regulated expression of multidrug and toxic compound exudation (Cl-MATE channel), citrate synthase (Cl-CS) and pectin methylesterase inhibitor (Cl-PMEI) genes, but while Cl-PMEI was expressed at the beginning of the study, Cl-CS and Cl-MATE were up-regulated only after 60 days, suggesting a late and uncoordinated response to Al. Consequently, Al decreased the root growth and caused anatomical damage. In addition, genes related to IAA cell transport were not differentially expressed in the transcriptome analysis. High IAA, up-regulation of auxin-related small RNAs, as well as low ethylene production in the root tip suggest Al- induced IAA biosynthesis rather than Al-induced disruption in IAA distribution in root cells of this species. Key words: Al3+; Auxin; Efflux channels; Gene expression; ‘Rangpur’ lime; Root growth 40 Introduction Aluminum is the third most abundant element in the Earth’s crust and, in the soil, it naturally occurs as Al3SiO4 (Von Uexküll and Mutert 1995). Approximately 30- 45% of soils from the world’s ice-free land are acidic (pH < 5.0) (Von Uexküll and Mutert 1995) and, under this condition Al3SiO4 is hydrolyzed to Al3+, which is toxic to most plants (Kochian 1995). In plants that are sensitive to aluminum (Al), it is covalently retained in the apoplast of root cells showing direct effects such as inhibition of root growth (Kochian 1995; Horst et al. 2010; Sun et al. 2010). It can also have indirect effects as reduced shoot growth (Jiang et al. 2008) and leaf gas exchange (Chen et al. 2005; Banhos et al. 2016a). In plants not exposed to Al, ethylene upregulates auxin biosynthesis to inhibit root cell elongation (Swarup et al. 2007), the known crosstalk between ethylene and auxin for proper root growth. In plants exposed to Al, the toxicity perception was evidenced to occur at the root apex of maize (Zea mays) (Ryan et al. 1993), and the ethylene/auxin crosstalk seems to be affected. In fact, Al-induced ethylene biosynthesis could act as a signal to modify auxin distribution in roots by disrupting genes encoding proteins for polar transport of auxin, such as At-AUX1 and At-PIN2, which eventually leads to arrest of root elongation (Sun et al. 2010). Al-induced increase in auxin biosynthesis could also lead to inhibition of cell elongation because, in roots, high concentration of auxin can cause inhibition of cell elongation (Ryan et al. 1993; Rahman et al. 2007). For instance, an Al-induced increased biosynthesis of indole- acetic acid (IAA) was found in the root apex of Arabidopsis (Yang et al. 2014). Al toxicity has been demonstrated not only in Arabidopsis and maize, but also in woody species, like Citrus plants, where these mechanisms are still unclear. Citrus limonia or ‘Rangpur’ lime shows reduced root length after being exposed to 400 µM Al for 70 days (Pereira et al. 2003). This species shows significant drought resistance due to a vigorous root system, being important as rootstock in subtropical regions of the Americas (Ribeiro and Machado 2007). Thus, Al-induced decrease in root length has a tremendous impact on this species. In a range of Al-tolerant species, including crop plants (Ryan et al. 2011) and woody species (Brunner and Sperisen 2013), one of the mechanisms to cope with Al involves the efflux of organic anions (malate, citrate, oxalate and succinate), which chelates the toxic Al form (Al3+) preventing its absorption or binding to the cell wall. 41 Malate efflux is mediated by the ALMT channel - aluminum activated malate transporter (Sasaki et al. 2004), and a non-specific channel that belongs to the MATE family (multidrug and toxic compound exudation) operates citrate efflux (Magalhaes 2010). Considering organic acid exudation, two patterns are evident depending on Al response: no delayed response that corresponds to channel activation (pattern I) and response in a lag phase (pattern II), which involves gene expression activation (Ma 2000). In Citrus grandis and C. sinensis, the secretion of citrate and malate from excised roots exposed to Al was noted within hours, suggesting pattern I response in these species (Yang et al. 2011), but excised roots decrease their metabolism after a long time. In addition, to our knowledge, no gene expression of efflux channels has been evidenced in order to support such physiological data in Citrus plants. Another mechanism to cope with Al disturbance relies on cell wall composition, which can interfere in Al affinity. In the absence of Al, the action of the pectin methylesterase (PME) leads to free pectic carboxylic groups, which in the presence of Al allows it to bind to pectic nets in the cell wall (Horst et al. 2010). Once Al covalently binds to the cell wall, it becomes rigid, limiting cell elongation. Therefore, the pectin methylesterase inhibitor (PMEI) enzyme impedes PME action and could decrease Al binding in the cell wall and, consequently, enhance Al resistance (Schmohl et al. 2000). The efflux of Al-organic acid complexes, enzymatic protection of pectic carboxylic groups in the cell wall, as well as IAA biosynthesis is well documented. However, the measurement of these parameters in whole plants with their roots exposed to Al for several days is lacking in the literature. In a 60-day study, we grew young ‘Rangpur’ lime plants in a hydroponic system in the presence and absence of Al. We confirmed biometric parameters and also performed a transcriptome analysis, in which we expected to find differentially expressed genes related to the efflux of organic acid-Al complexes. We also evidenced organic acids biosynthesis and release from roots of intact whole plants. Although expecting to find MATE and ALMT gene families, as evidenced by citrate and malate effluxes in Citrus grandis and C. sinensis (Yang et al. 2011), we hypothesized that the expression of these genes is not enhanced due to the Al sensitiveness of ‘Rangpur’ lime plants (Pereira et al. 2003; Banhos et al. 2016a). We collected root tips and measured IAA and ethylene which, along with IAA polar transport genes possibly evidenced in the transcriptome and further gene expression analyses, made us predict that IAA distribution in root cells is also disrupted, as shown for Arabidopsis (Sun et al. 2010). In 42 addition, anatomical investigation of roots provided evidence to support the functional analyses. Material and methods Plant material and experimental conditions We used three-month-old and 7 ± 0.5 cm-high ‘Rangpur’ lime (Citrus limonia) plants, showing approximately five leaves, for studying the effects of Al on root development within a 60-day period. The plants were kept in a hydroponic system and grown directly on an aerated nutrient solution inside opaque plastic boxes (50 cm in length x 30 cm in width x 15 cm in height; 20 L). The nutrient solution was adapted from Banhos et al. (2016a) and shows a chemical composition based on the solution proposed by Clark (1975) (Clark 1975), having been used to test Al resistance in Citrus rootstocks (Santos et al. 2000; Banhos et al. 2016a). It consisted of the following macronutrients (in mM): N (NO3 -) 0.93; N (NH4 +) 0.40; P, 0.013; K, 0.90; Ca, 1.37; Mg, 0.33; S, 0.25; and micronutrients (in µM): Cl, 231.0; Fe (EDTA), 23.3; B, 8.33; Mn, 2.91; Zn, 0.76; Cu, 0.32; Mo, 0.31. In a previous study (Banhos et al. 2016a), we observed that 1480 µM Al causes Al-induced decrease in gas exchange rates in ‘Rangpur’ lime plants after 45 days. Therefore, besides macro and micronutrients, this solution contained 0, 370, 740, 1110 and 1480 µM Al provided through AlCl3 6 H2O. The pH of the solution was monitored daily and maintained at 4.0 ± 0.1 to keep Al as soluble as possible, and the solution was totally replaced every 15 days. The nominal chemical composition of this solution was also tested on Geochem-EZ software (Shaff et al. 2010), resulting in more than 85% free Al3+ available. We also measured Al in the solutions using the colorimetric method (Sarruge and Haag 1974), and nominal 370, 740, 1110 and 1480 µM Al supply resulted in 214.4 ± 32.6, 400.2 ± 32.8, 907.7 ± 42.9 and 981.6 ± 67.9 µM Al. Expanded polystyrene (Isopor®) 50 x 30 cm plates (2-cm thick), with eight holes (2.5 cm in diameter) each, were floated on the nutrient solution in the boxes, and the plants were fixed in these holes with polyurethane foam strips that were placed around the plant collar. The boxes stayed on benches (80 cm from the ground) inside a greenhouse with semi-controlled conditions (air temperature 28.5 ± 0.7C; relative humidity 63.3 ± 1.3%; 853.4 ± 175.1 µmol photons m-2 s-1; approximately 14h of natural photoperiod). 43 Experimental design After separating a group of plants for initial measurements (biometric data and biomass), we set up the plants in the hydroponic system, and measured the accumulated length of the main root every 7 days until 60 days after planting (DAP), when the number of leaves, leaf area per plant, shoot and main root lengths, as well as the biomass of organs, were assessed. Using plants grown in the most contrasting treatments (0 and 1480 µM Al), we collected root tips to measure the concentration of indole-acetic acid (IAA) at 1, 3, 7, 15, 30 and 60 DAP, and ethylene production only at 60 DAP due to fine and insufficient roots before this date. After having conducted a transcriptome analysis using ‘Rangpur’ lime plants grown under 0 and 1480 µM Al, in the present study we collected root tips at 1, 7, 15, 30 and 60 DAP to assess the expression of some genes that had been differentially expressed and revealed by the transcriptome analysis. These were the Cl- MATE (Multidrug and toxic compound exudation), citrate synthase (Cl-CS), Cl-SAUR 10 and Cl-SAUR 15 (Small Auxin up-regulated RNAs) and pectin methyl esterase inhibitor (Cl-PMEI). The MATE gene family encodes membrane proteins that facilitate the efflux of organic anions such as citrate (Liu et al. 2009). The SAUR gene family is responsible for acidification of the apoplast during IAA action (Spartz et al. 2014). The Cl-SAUR 10 showed close identity when compared to Gm-SAUR X10A and Cl-SAUR15 had high identity when compared to Gm-SAUR 15A, being Gm-SAUR X10A and Gm- SAUR 15A the first characterized SAUR genes, both evidenced in Glycine max (McClure et al. 1989), which share similar functions in elongating tissues responding to auxin within few minutes (Jain et al. 2006). PMEI inhibits demethylesterification of pectic carboxylic groups in walls of root cells, which occurs through the action of pectin methylesterase (PME) (Giovane et al. 2004). To check whether organic acids (OAs) were being synthesized and released by the roots, we cultivated a group of plants showing the same age and height as described above in the same nutrient solutions as previously described (Banhos et al. 2016a), containing 0 and 1480 µM Al. After seven days, we quantified the contents of oxalate, succinate, malate, and citrate in roots tips and also released in the solutions. In addition, we collected root tips at 7, 15, 30 and 60 DAP to conduct an anatomical analysis of plants grown under 0 and 1480 µM Al in order to check anatomical disorders during the period of exposure to Al. 44 Biometric parameters The lengths (cm) of roots (from the plant collar to the root tip) and stems (from the plant collar to the shoot apex) were measured with a ruler, and the number of leaves was counted. At 60 DAP (and also at 0 DAP for initial measurements) plants were separated into leaves, stems (plus petioles) and roots. The leaf area (cm2) was measured with an area meter (LI-3100C, LI-COR, USA). The organs were dried at 60°C until constant mass to obtain the biomass (g) of organs and total plant biomass. Indole-acetic acid analysis Extraction and purification Root apexes (0.5 cm in length) were washed in deionized water and stored at - 80C prior to analysis. Samples (420 µg fresh weight) were extracted using isopropanol:acetic acid (95:5; v:v) solution at 4C. A solution containing 0.5µg of a labelled standard [13C6 ]-IAA (Cambridge Isotopes, Inc.) was added as internal standard (IS). After 2h, the solvent was separated by centrifugation at 14,000 g for 10 min at 4C. The supernatant was concentrated to 50 µL, acidified to 2.5 pH, and IAA was extracted three times with ethyl acetate, according to Ludwig-Müller et al. (2008) (Ludwig-Müller et al. 2008). After purification, the solution was evaporated to dryness and re-suspended in 30 µL ethyl acetate for methylation with diazomethane. GC-MS analysis Quantification of free IAA was performed using gas chromatography coupled with mass spectrometry: a Hewlett Packard 6890 gas chromatograph (GC) coupled to an HP5973 mass selective detector. The GC column was HP-1701 (30 m, 0.25 mm I.D., 0.5 μm film thickness), with helium as the carrier gas at flow rate of 1 mL min-1. The temperature program for analysis was 140C for 1 min, followed by increase of 5C min-1 up to 210C, 1 min hold, increase to 280oC at 20oC min-1 then 1 min hold at 280C. The ions were monitored at m/z 130 and 189, for the endogenous IAA, and m/z 136 and 195, for the labeled IS. The ratio 130:136 was used to calculate the endogenous amount of IAA. Ethylene analysis Root tips (~1 cm in length) of ~0.5 g were excised and placed into 5.0 mL glass vials containing 0.7% (m/v) agar (Sun et al. 2010). After 2h, 1 mL of the headspace was 45 taken from the vials using a syringe (Hamilton, Gastight, Nevada, USA), and injected into a gas chromatograph - GC (Trace 2000 GC, Thermo Fisher Scientific, USA). The GC was equipped with a 1.8 m Poropack N column at 120°C, flame ionization detector (FID) at 120°C, injector at 120°C and used nitrogen as the carrier gas (30 mL min-1). Gene expression analysis We used quantitative PCR to analyze the expression of Cl-MATE, Cl-CS, Cl- SAUR 10, Cl-SAUR 15 and Cl-PMEI, which had been differentially expressed in the presence of Al as revealed by the transcriptome analysis. Based on RNA-seq data, we designed primers to Cl-MATE (forward CATTGACACAGCATTTATCGGC and reverse CAAATGAGGTTGTAATATTAACTAATGGG), Cl-CS (forward CGCTAAGCCAGATGGAGAAC and reverse ACAGTGGCACGATCTCTCAA), Cl- PMEI (forward CCACAAGAACGACAGCGATA and reverse TTAGCGTAACTCGGCAAGGT), Cl-SAUR 10 (forward TGGGTTCACAACTCACAAGC and reverse TGAACAATACCAGGCAAACG) and Cl-SAUR 15 (forward AGGCGTGCTCTTATGGTTTC and reverse TTCTGAAAGGATGGGTGCTT). As reference genes we used GAPC2 (forward TCCTATGTTTGTTGTGGGTG and reverse GGTCATCAAACCCTCAACAA) and EFα (forward TCAGGCAAGGAGCTTGAGAAG and reverse GGCTTGGTGGGAATCATCTTAA), which were proposed by Mafra et al. (2012). Serial dilutions of genomic DNA from Citrus limonia were amplified using the internal primers in order to access the amplification efficiency during qPCR analysis. GoTaq qPCR Master Mix (Promega, Madison, USA) was used as the reagent in these tests, with the following procedural specifications: 95C for 5 min (1 cycle), 95C for 10 s, 58C for 30 s, 72C for 30 s (40 cycles) followed by a melting curve analysis (1% slope temperature; 60–95C). Amplification efficiencies were 91.85% for Cl-MATE, 85.38% for Cl-CS, 97.22% for Cl-PMEI, 91.71% for Cl-SAUR 10, 91.55% for Cl-SAUR 15, 97.24% for Cl-GAPC2 and 100.04% for Cl-EFα. For quantitative analysis of gene expression, total RNA was extracted from root samples using the RNeasy plant mini kit (Qiagen, Hilden, Germany). Total RNA (2 µg) was treated with RNase-free TURBO DNase (Ambion, Carlsbad, USA) and transcribed in reverse to cDNA using an oligo-dT primer and Super Script III, according to the manufacturer’s protocol (Life Technologies, Carlsbad, USA). cDNAs were submitted to gene expression analysis using GAPC2 and the EFα genes in order to normalize the cycle threshold (Ct) values. The reagent and cycles used for the qRT-PCR analyses 46 were the same as those used for the amplification efficiency tests. In this protocol, every reaction was performed in triplicate. Ct value of each sample, determined by the mean of the three technical replicates, was converted into relative quantities (RQ) using the function RQ=EΔCt. ΔCt is the difference between the lowest Ct value across all samples for the evaluated gene and the Ct value of the given sample. A normalization factor (NF) for each sample was calculated by the geometric mean of the RQ values of GAPC2 and EFα. Normalized- relative quantity (NRQ) of each sample was calculated as the ratio of the sample RQ and the appropriate NF. Individual fold change values were determined by dividing the sample NRQ by mean values of NRQ that were obtained from the calibrator, i.e., root samples of plants not exposed to Al. Following this, fold change in the control group always shows a mean value of 1. Organic acids analysis The plants were cultivated in Falcon® tubes containing the nutrient solutions (45 mL) with 0 and 1480 µM Al. Aeration was performed twice a day and pH was maintained at 4.0. The plants were maintained in laboratory conditions (25  1 C, photoperiod of 12h and photosynthetic photon flux density of 600 µmol m-2 s-1, which returns saturating CO2 assimilation rates). When necessary, the solutions were completed to 45 mL with deionized water. Organic acids synthesized by the roots (internal OAs) were extracted by osmosis and alkaline gradient, immersing the root tips in 1.5 mL Na2CO3 40 mM for 24h. Then, we collected 1 mL of the extractor solution and dried at 80°C. Exudated OAs (released by the roots) were concentrated after drying both solutions (0 and 1480 µM Al) at 80°C. Samples were esterified according to Fischer method (methylation) (Fischer and Speier 1895). To internal OAs samples, we added 400 μL of methanol (HPLC/Spectro) and 100 μL of sulfuric acid (7N) resulting in a 0.5 mL extracting solution. To exudated OAs samples, we added 700 μL of methanol (HPLC/Spectro) and 300 μL of sulfuric acid (7N) resulting in 1 mL extracting solution. After shaking, the samples were kept for 15 min at 70°C for catalyzing the reaction. After adding 1 mL of hexane (HPLC/Spectro) 100 μL of apolar phase were collected and analyzed using a gas chromatograph coupled to a mass spectrometer (GC-MS) system (GC- 2010/GCMSQP2010 Plus, Shimadzu, Japan), with an automatic sample injector (AOC- 20i). 47 In the GC-MS, we used a 30 m-length and 250 μm-diameter fused-silica microcolumn (RTX-5MS, Restek), and analytical ultra-pure helium (99.9999%, White Martins) was used as carrier gas. The injector temperature was 250°C (Splitless mode) and the injection volume was 1 μL. Column gas flow was maintained at 41 cm/s. The initial column temperature was 50°C with a 4 min step. After that, at a 10°C/min rate, it achieved 70°C. Then, it was increased to 250°C at a 25°C/min rate, and maintained for 0.8 min, completing 14 min running. Mass detector was a simple quadrupole type with 70eV electronic impact ionization. The GC-MS interface temperature was 250°C and 230°C to the ionizer. The detector potential was relative to tunning, with a 40 to 450 m/z detection range (scanner mode). Anatomical analysis Root tips (~ 0.5 cm in length and 1 mm in diameter) were collected and immediately fixed in Karnovsky solution (Karnovsky 1965). The samples were dehydrated in increasing ethanol series [30, 50, 70, 90 (one hour each), and 100% (three times, one hour each)], then infiltrated with resin (Historesin, Leica instruments, Germany) and ethanol 100%, at a ratio of 1:1, overnight. After 24h, samples were infiltrated with pure resin, reserved overnight and then polymerized in blocks. Longitudinal sections were obtained with a rotary microtome and mounted on permanent glass slides that were immersed (5 min) in a toluidine blue solution (pH 4.5) for staining (at room temperature) structures containing nucleic acids and lignin (O’brien et al. 1964). Then, the glass slides were washed under tap water to remove excess of dye and dried with a clean cotton cloth. All sections were observed under light microscope (DMLB, Leica Microsystems, Wetzlar, Germany) and the images were captured with a digital camera (DFC-290, Leica Microsystems, Germany) functionally attached to the DMLB. The anatomical study was based on consecutively-sliced longitudinal sections of root tips containing the root cap, the quiescent center, and elongation, division and maturation zones. Data analysis Biometric data and the biomass of organs were measured using eight replications (plants). Concentration of IAA and ethylene production in the root tips was evaluated using three plants. For gene expression and OAs quantification in root tips and in the solution, five samples were used as replications. 48 A one-way analysis of variance (Anova) was performed between plants exposed to 0, 370, 740, 1110 and 1480 µM Al to test for differences in accumulated main root length at every evaluation date. The Tukey test ( = 0.05) was used to conduct post-hoc comparisons to estimate the least significant differences between mean results. Biometric data and biomass of organs were subjected to student’s t-test ( = 0.05) to verify differences between 0 and 60 DAP within each Al concentration. This same test was used to verify differences in IAA concentration, ethylene production, expression of genes and OAs synthesis/release between 0 and 1480 µM Al on distinct evaluation dates. Results Plants showed reduced plant growth, fewer leaves and decreased root development with the increase of Al concentration in the nutrient solution (Fig. 1). Differences in root growth started after 14 DAP. At 21 and 28 DAP, plants exposed to Al showed similar values that were lower than those of plants not exposed to Al. From 35 DAP until 56 DAP plants exposed to 370 µM Al emerged as an intermediate treatment, showing values that were lower than those of plants not exposed to Al but higher than those of plants exposed to 740, 1110 and 1480 µM Al (Fig. 2). When compared to initial values (0 DAP), the increment of leaf area (Fig. 3a), number of leaves (Fig. 3b), shoot length (Fig. 3c) and main root length (Fig. 3d) within 60 days reduced considerably with the increase of Al concentration in the nutrient solution. Plant biomass exhibited the same response pattern, showing reduced values of leaf (Fig. 4a), shoot (Fig. 4b), root (Fig. 4c), and total biomass (Fig. 4d) with the increase of Al concentration in the nutrient solution. Plants exposed to 1480 µM Al showed higher IAA concentration in their root tips when compared to plants not exposed to Al, mainly at 3, 7 and 15 DAP (Fig. 5). At 60 DAP, ethylene produced in the root tips of plants exposed to 1480 µM Al was five times lower than that of plants not exposed to Al (Fig 6). When exposed to 1480 µM Al, the gene responsible for Cl-MATE channel was 10.4 and 1.5 times more expressed at 1 and 7 DAP, respectively (Fig. 7a). However, at 60 DAP, Cl-MATE was 215 times more expressed in plants exposed to Al when compared to plants not exposed to Al (Fig. 7a). Citrus limonia citrate synthase gene expression was down-regulated at 15 DAP (1.6 times less expressed) and 30 DAP (5.3 times less expressed) in the presence of Al in relation to plants not exposed to Al. 49 However, Cl-CS was considerably up-regulated (9.1 times more transcripts) in the presence of Al at 60 DAP (Fig. 7b). Cl-PMEI gene expression, however, showed up- regulation at 7 DAP (1.8 times more transcripts), and then became down-regulated in response to Al, reducing to 3.4 (15 DAP), 10 (30 DAP) and 20 times less transcripts (60 DAP) in relation to plants not exposed to Al (Fig. 7c). The same up-regulation followed by down-regulation was observed for Cl-SAUR 10, which was 2.8 times more expressed at 7 DAP and reduced to 25 times less transcripts at 60 DAP (Fig. 8a). Cl-SAUR 15 followed a similar response, showing down-regulation at 15 DAP (2 times less transcripts) and 60 DAP (33.3 times less transcripts) (Fig. 8b). Roots tips OAs contents were found in all plants after seven DAP. However, in relation to plants not exposed to Al, only succinate was significantly reduced in roots of plants exposed to Al (Fig. 9a). After seven days, oxalate, succinate, and citrate were detected in the nutrient solutions and were higher in the solutions containing Al. Citrate was not found in the solution without Al, but approximately 2µmol of it was measured in the solution containing Al (Fig. 9b). In plants not exposed to Al, cells with regular size and rectangular shape were noted in the cortex (Fig. 10b), while in plants exposed to 1480 µM Al thick a non- uniform round shape cells were found in the cortex (Fig. 10d), after 15 days of Al exposure. This anatomical pattern was reflected in the shape of the extreme root tip as plants exposed to Al showed a wider root tip (Fig. 10c, 10g and 10k) when compared to a sharp arrow-like root tip in plants not exposed to Al (Fig. 10a, 10e and 10i). When exposed to Al, more cellulosic deposition was observed on cell walls in the cortex and lignin in the vascular cylinder, as evidenced by stronger blue-green stains (Fig. 10d). The longer the exposure time to Al, the more apparent became this cellulosic cortical deposition as well as lignin deposition in the vascular cylinder (Fig. 10d, 10h, 10l). Anatomical ruptures could be observed on root epidermis of plants exposed to Al, mainly after 30 and 60 days of Al exposure (Fig. 10g and 10k), indicating an irregular pattern of epidermal cell division and elongation, which contrasted with plants not exposed to Al (Fig. 10e and 10f). Discussion ‘Rangpur’ lime plants are sensitive to Al (Santos et al. 2000; Pereira et al. 2003; Banhos et al. 2016a) and here we show that, although they possess genes associated with Al resistance, as revealed by the transcriptome analysis (RNA-seq), their 50 expression is uncoordinated in this species. Cl-MATE and Cl-CS were conspicuously up-regulated when the roots were exposed to 1480 µM Al, but only at 60 DAP. We confirmed that ‘Rangpur’ lime plants had OAs inside their roots independently of the Al presence. This can be related to intense respiration activity (Krebs’ cycle), as evidenced by the presence of mitochondria in root tips of soybean plants (Xu et al. 2010). After seven days, except for malate, OAs were significantly higher in the solution containing Al, indicating that OAs, including citrate, are secreted by this species in response to Al. Citrate efflux is mediated by MATE (Furukawa et al. 2007), and a cluster analysis of the MATE sequence found showed highest phylogenetic proximity to MATE genes related with this same function in different species (data not shown). CS and MATE can be both up-regulated by Al (Xu et al. 2010), but in transformant tobacco plants CS expression did not lead to the efflux of citrate by MATE channels (Delhaize et al. 2001), while in barley citrate exudation responds to Al (Zhao et al. 2003), indicating there is no consistent pattern in the expression of MATE and CS in response to Al. In the present study, Cl-CS and Cl-MATE were concomitantly up-regulated at 60 DAP, indicating a consistent pattern in the expression of these two genes in response to Al. In soybean plants, only a 16% increase in citrate synthase activity occurred upon Al treatments after 6h (Xu et al. 2010). In Vigna umbellata, the citrate efflux mediated by MATE also occurred 6h following Al exposure (Liu et al. 2013). In the present study, Cl-CS was not significantly up-regulated at 7 DAP (Fig. 7b) corroborating the similar citrate concentration in root tips of both treatments (Fig. 9a). However, approximately 2 µmol citrate was found in the solution containing Al, at 7 DAP (Fig. 9b). These values may be comparable to the citrate secretion from excised roots of Citrus grandis and C. sinensis that reached a peak of 0.7 µmol citrate at 24h after exposure to 500 µM Al (Yang et al. 2011), which decreased after this period perhaps because excised roots reduce their metabolism after a long time. In the case of the whole root system of ‘Rangpur’ lime plants, the conspicuous up-regulation of Cl-MATE and Cl-CS only at 60 DAP may be a delayed response to mitigate the effects of Al, and investigation of OAs concentration in root tips and released in the solution at 60 DAP merits further investigation after refining the method of OAs detection presented here. Such delayed response to Al can be also supported by anatomical damage caused by Al and observed from 15 DAP until the end of the study (Fig. 10). On the other hand, Cl-PMEI was up-regulated at 7 DAP. This indicates that pectic carboxylic groups in the cell wall are protected from demethylesterification processes through the action of PME, at least 7 days after exposure to Al. In fact, pectin 51 undergoes partial apoplastic demethylesterification through the action of PME, resulting in the exposure of free pectic carboxylic groups, which could serve as binding sites for Al in the wall (Zhu et al. 2014). Additionally, potato transformants that exhibited higher expression of PME accumulate more Al, produce more callose and their root growth is more inhibited when exposed to Al than wild types (Schmohl et al. 2000). Demethylesterification through the action of PME also occurs within hours (Franco et al. 2002). Therefore, up-regulation of a gene encoding inhibitors of this process at 7 DAP may not be considered a fast response, even in a 60-day experiment. In addition, this up-regulation was inverted, becoming down-regulated at 15, 30 and 60 DAP (Fig. 7c). Taken together, Cl-PMEI up-regulation at 7 DAP (but down-regulation from 15 to 60 DAP), while Cl-MATE and Cl-CS increased expression only at 60 DAP could indicate that these mechanisms, which are related to Al resistance in insensitive plants, are not only delayed (Cl-CS and Cl-MATE) but uncoordinated in ‘Rangpur’ lime plants, and this may explain the Al sensitivity in this species. We confirmed that Al interferes with root elongation and root biomass production in ‘Rangpur’ lime plants. Plants not exposed to Al showed accumulated root length that was 10 times higher than that of plants exposed to 370 µM Al, and 16 times higher when compared to plants exposed to 740, 1110 and 1480 µM Al, at 56 DAP (Fig. 2). These expected symptoms, however, were associated with high concentration of IAA in root tips of plants exposed to Al, mainly at 3, 7 and 15 DAP (Fig. 5). In Arabidopsis, Al was also found to induce a localized enhancement of auxin biosynthesis in the root apex (Yang et al. 2014). Differently from stems, in cells (of the elongation zone) of root tips, high concentration of IAA can cause strong inhibition of cell elongation (Rahman et al. 2007; Cleland 2010). IAA concentration in root tips of plants exposed to Al was three- (7 DAP) and two-fold higher (15 DAP) than that of plants not exposed to Al (Fig. 5). Intriguingly, the expression of Cl-SAUR 10 was also up- regulated at 7 DAP, when a peak of IAA concentration occurred in plants exposed to Al after which this gene became down-regulated until the end of the study, similar to what occurred with Cl-SAUR 15 gene expression. This suggests that at 7 DAP auxin action was close to its maximum activity as SAUR genes are responsible for acidification of apoplast (Spartz et al. 2014). Alkalinization of the apoplast, on the other hand, is associated with inhibition of root cell elongation under elevated auxin concentration (Lüthen and Böttger 1993; Evans et al. 1994; Hager 2003). Thus, the up-regulation of Cl-SAUR 10 occurring together with the peak of IAA concentration in roots of plants exposed to Al demonstrates that IAA and Cl-SAUR 10 gene expression were initially 52 acting towards root cell elongation, but after that, Cl-SAUR 10 and Cl-SAUR 15 gene expressions were down-regulated while IAA concentration was kept high until 15 DAP (Fig. 5), which might have inhibited cell elongation. These responses of Cl-SAUR 10 (up-regulation peaked at 7 DAP) and Cl-SAUR 15 contrast with those of Cl-CS and Cl- MATE that peaked only at 60 DAP, reinforcing that such genes related with Al resistance are uncoordinated in ‘Rangpur’ lime plants. In addition, the transcriptome analysis of plants grown in the presence and absence of Al did not reveal any differential expression of genes responsible for membrane proteins associated with auxin polar transport, such as AUX1, AUX2 and PINs. Therefore, in the root tips of ‘Rangpur’ lime plants grown under 1480 µM Al, an increase in IAA concentration, and not an imbalance in IAA distribution, is likely to trigger the inhibition of root cell elongation and, consequently, root growth. This is evidenced by round shaped cells in the cortex of roots exposed to Al (Fig. 10d), indicating that these cells divide in a periclinal manner, contrasting with the rectangular cells in the cortex of roots not exposed to Al (Fig. 10b) that evidences an anticlinal cell division benefiting the normal growth in length. Using Arabidopsis, Sun et al. (2010) argues that Al-induced ethylene biosynthesis is likely to act as a signal to change auxin distribution in roots by disrupting At-AUX1 and At-PIN2-mediated auxin transport. Therefore, we infer that this disruption in auxin distribution leads to asymmetrical auxin distribution and, consequently, twisted root shape. Our anatomical analysis, however, indicated thick non-uniform round shape cells at both sides of the cortex of roots exposed to Al, resulting in a thick growth pattern at the extreme root tip of plants exposed to Al, where no twisted roots were observed on any evaluation date. This reinforces that Al-induced disruption in auxin distribution is unlikely to occur in ‘Rangpur’ lime plants when exposed to Al. In addition, although measured only at 60 DAP, ethylene production in plants exposed to Al was 1000 times lower than that in plants not exposed to Al. Thus, besides its role in disrupting IAA distribution that could not be evidenced in plants exposed to Al, such low ethylene production also corroborates the fact that leaves of these plants did not experience Al-induced abscission. This may be an evidence of a distinct mechanism for Al disruption of root cell elongation in this species, although measurement of ethylene production over time is needed in future studies with this species when disturbed by Al. The ruptures in the epidermis of roots exposed to 1480 µM Al have also been observed in cowpea exposed to 600 µM Al for 48h (Kopittke et al. 2008; Blamey et al. 53 2011). These authors argue that these ruptures result from an increase in cell wall rigidity in the outer layers (epidermis and outer portions of the cortex) due to higher Al accumulation in these layers, as observed in barley (Ma et al. 2004) and maize (Stass et al. 2006). According to this, Al reduces elongation in outer layers while cells of the inner root layers continue to elongate, which possibly causes ruptures. We found irregular pattern of epidermal cell division and elongation in roots of plants exposed to Al, contrasting with those of plants not exposed to Al, which could corroborate the explanation given by Kopittke et al. (2008) and Blamey et al. (2011). However, we could not anatomically evidence that the inner root layers are suggestive of regular (anticlinal) cell division, as evidenced by non-rectangular round shaped cells in the whole cortex. Therefore, identifying distinct concentrations of Al in different portions of the root tip of ‘Rangpur’ lime plants exposed to Al is needed in future studies. For example, using Al-specific x-ray coupled to scanning electron microscope (SEM/EDS) could enable us to detect the presence of Al in different parts of leaves from Al- accumulating and non-accumulating species (Bressan et al. 2016). In this study, we demonstrate that although possessing genes associated with Al resistance, ‘Rangpur’ lime plants are, in fact, sensitive to Al due to late expression of Cl-MATE, Cl-CS and Cl-PMEI, and also uncoordinated responses of these genes when the roots are exposed to Al in a 60-day study. In addition, we revealed that an increase in IAA concentration, rather than an imbalance in IAA distribution, as we expected, is associated with the inhibition of root cell elongation in this species. Acknowledgements We acknowledge the Brazilian National Council for Scientific and Technological Development (CNPq) for financial support (474169/2013-8 grant to GH), for an undergraduate scholarship to M.F. Cavalheiro and for research productivity fellowships granted to G. Habermann (Grant 308902/2014-9) and E. Purgatto (Grant 305458/2013-2). We extend acknowledgments to the São Paulo Research Foundation (Fapesp) for PhD scholarships granted to C.M.S. da Silva (Fapesp #2013/11370-3), A.C.G. Bressan (Fapesp #2014/14386-0), a MSc scholarship to B.M.O. Cravalho (Fapesp #2016/14216-3) and a grant to E. Purgatto (Fapesp #2013/07914-8), and also to the Coordination for Improvement of Graduate Personnel (Capes) for a PhD scholarship granted to O.F.A.A. Banhos. We thank the Sanicitrus Nursery (Araras, São Paulo state, Brazil) for providing us with the ‘Rangpur’ lime plants. 54 Figures and tables Fig. 1 General view of the root system of ‘Rangpur’ lime plants grown for 60 days in nutrient solutions containing 0, 370, 740, 1110 and 1480 µM Al 55 Fig. 2 Accumulated length of the main root of ‘Rangpur’ lime plants grown for 56 days in nutrient solutions containing 0, 370, 740, 1110 and 1480 µM Al. For each evaluation date, different letters represent significant differences (P < 0.05) between Al concentrations 56 Fig. 3 Leaf area, number of leaves, shoot and root length of ‘Rangpur’ lime plants at 0 and 60 days after planting (DAP), grown in nutrient solutions containing 0, 370, 740, 11