UNIVERSIDADE FEDERAL DO RIO DE JANEIRO

KIM OHANNA PIMENTA INADA

EFFECT OF TECHNOLOGICAL PROCESSES ON PHENOLIC COMPOUNDS CONTENTS OF JABUTICABA (MYRCIARIA JABOTICABA) PEEL AND SEED AND INVESTIGATION OF THEIR METABOLISM IN HUMANS.

RIO DE JANEIRO 2018 Kim Ohanna Pimenta Inada

EFFECT OF TECHNOLOGICAL PROCESSES ON PHENOLIC COMPOUNDS CONTENTS OF JABUTICABA (MYRCIARIA JABOTICABA) PEEL AND SEED AND INVESTIGATION OF THEIR ELLAGITANNINS METABOLISM IN HUMANS.

Tese de Doutorado apresentada ao Programa de Pós-Graduação em Ciências de Alimentos, Universidade Federal do Rio de Janeiro, como requisito parcial à obtenção do título de Doutor em Ciências de Alimentos

Orientadores: Profa. Dra. Mariana Costa Monteiro Prof. Dr. Daniel Perrone Moreira

RIO DE JANEIRO 2018

DEDICATION

À minha família e às pessoas maravilhosas que apareceram na minha vida. ACKNOWLEDGMENTS

Primeiramente, gostaria de agradecer a Deus por ter me dado forças para não desistir e por ter colocado na minha vida “pessoas-anjo”, que me ajudaram e me apoiaram até nos momentos em que eu achava que ia dar tudo errado.

Aos meus pais Beth e Miti. Eles não mediram esforços para que eu pudesse receber uma boa educação e para que eu fosse feliz. Logo no início da graduação, a situação financeira ficou bem apertada, mas eles continuaram fazendo de tudo para me ajudar. Foram milhares de favores prestados, marmitas e caronas. Meu pai diz que fez anos de curso de inglês e espanhol, porque passou anos acordando cedo no sábado só para me levar no curso que eu fazia no Fundão. Tinha dia que eu saía do curso morta de fome e quando eu entrava no carro, tinha uma marmita com almoço, com direito até a garrafa de suco. Se eu desse carona para alguma amiga, ela também ganhava lanche. Esses são os meus pais. Quem conhece, não esquece.

À minha irmã Haynna, minha alma gêmea. Ela é a pessoa mais durona, com o coração mais mole que eu conheço. Confesso que eu nunca respeitei as placas de “não entre” e “não estou disposta” da porta do quarto dela. Eu entrava mesmo assim e ela me perguntava impaciente “o que você quer, Kim?”. E eu sempre queria tirar dúvidas, fazer mil perguntas sobre milhares de assuntos ou apenas contar uma fofoca rápida. Às vezes eu só queria olhar para a cara dela e mostrar os dentes mesmo. Porque a minha irmã é assim. Mesmo reclamona, eu posso contar com ela para tudo, desde reclamar da vida a rir dos outros. Ela é pau para toda obra, minha enciclopédia ambulante. Muito obrigada por tudo!

Ao meu namorado português, parte não japonesa da minha vida, Thiago. Por esse motivo, ele ganhou o apelido japonês “moshi”. Ele é uma das pessoas mais sensatas que eu conheço. Tão sensato que todos os amigos procuram ele para pedir conselhos. E foi com esses conselhos maravilhosos que eu cheguei até aqui. Nos momentos em que eu estava surtando e arrancando os cabelos, foi ele quem me acalmou. Todas as vezes que eu pensei em jogar tudo para o alto, ele sempre estava lá para não me deixar desistir. Se descobrissem você, você iria virar conselheiro dos alunos de pós-graduação (vamos manter isso em segredo). Moshi, meu companheiro de viagens, meu namorado há 10 anos. Muito obrigada por tudo! Amo você!

Aos meus sobrinhos Daniel e Davi. Meus japonesinhos com cara de índio, meus bebês que estão crescendo rápido demais. Obrigada por tornarem a minha vida tão mais colorida. Há pouco tempo, enquanto eu escrevia a tese, os dois chegaram no meu quarto e o Daniel me perguntou “O que você está estudando, dinda? Tese? O que é tese? É chato fazer isso?”. E aí sem nem me deixar responder, o Davi começou a falar “É muito chato! Chato, chato, chato!”. Tudo isso porque o Daniel estava entediado e o Davi queria que eu fizesse um avião de papel para ele. Como viver sem esses momentos? Tem como não morrer de amores?

À minha família, principalmente meus tios (Valesca, Rogério e Márcio) e minha irmã Poliana, que mesmo distantes da realidade da pesquisa, sempre torceram pelas minhas conquistas.

À minha amiga Julia. A gente se conheceu na graduação, viramos amigas no Mestrado e eu posso dizer que no Doutorado ganhei uma nova irmã. Enfrentamos muitos momentos juntas, desde as provas na graduação, a seleção de Mestrado da Nutrição e os concursos de professor substituto da UFRJ. Lembro que, enquanto estudávamos para o concurso de substituto, ela desistiu de fazer e não me contou, porque não queria me desanimar. Nos últimos dois meses, na correria para defender o Doutorado, ela me mandava mensagem todos os dias perguntando como eu estava. Esse é o tipo de amiga que ela é. Muito maravilhosa. Como se fosse de outras vidas (e talvez seja mesmo). Quero continuar indo dormir na sua casa só para conversar até tarde e tomar sopa de capeletti. Que a nossa amizade seja eterna e ainda tenha muita Casa do Sardo e menos bandejas (por favor!).

À minha amiga Iris. Viramos amigas no Doutorado e foi uma das melhores coisas que me aconteceu. Andar com a Iris é como ter um “stand-up comedy” andando atrás de você. Não tem como não rir das suas caras e bocas, das risadas escandalosas e das loucuras que ela inventa. Por trás de toda a palhaçada, existe uma amiga conselheira com um coração gigante. Ela é uma amiga para você chorar de rir ou chorar de chorar mesmo. Muito obrigada por me fazer rir e sorrir em tantos momentos. Você é maravilhosa.

À minha amiga Camila. A pessoa mais chorona que eu conheço. Já imagino ela no dia da defesa, chorando com a bochecha da cor de um pimentão. Por trás desse chororô todo, tem uma pessoa muito corajosa, sempre disposta a ajudar e a defender os amigos. Que você seja muito feliz em Portugal e que nunca se esqueça de nós. Muito obrigada por todo o carinho! Acho que ninguém sabe dessa parte, mas, quando eu estava na graduação, eu dizia que eu queria fazer Mestrado no Instituto de Química, na área de Alimentos. Eu sabia que queria isso, mas eu não fazia ideia da existência do PPGCAL e nunca fiz nada para conseguir isso. Nem sei de onde tirei essa ideia. E aí, o que aconteceu? Caiu uma pessoa do céu, na minha frente. Essa pessoa foi a minha orientadora Mariana, que trouxe, junto com ela, meu orientador Daniel. À minha orientadora Mariana. Sem nem me conhecer, primeiro, ela me aceitou como sua monitora e depois, como aluna de Mestrado e Doutorado. Durante todo esse período, ela nunca poupou esforços para me ajudar. Muito obrigada por todo o carinho, apoio e todo o aprendizado ao longo desses anos. Foi uma honra ser sua aluna e, além disso, ter a oportunidade de trabalhar ao seu lado. Nunca vou me esquecer das experiências com queijos e balas de goma. Vou sentir falta de botar a minha cara na janelinha da sua sala QUASE todos os dias, te enchendo de perguntas ou só para falar alguma bobeira mesmo – quase todos os dias, porque eu sou legal e nos fins de semana eu dava uma folga para ela. Obrigada por tudo, Mariana!

Ao meu orientador Daniel. Se a Mariana não me conhecia, ele me conhecia menos ainda. Mesmo assim, me recebeu de braços abertos desde o primeiro momento. No primeiro dia que nos conhecemos, fomos para a Embrapa juntos lavar quilos e mais quilos de jabuticaba. Muito obrigada por ter me ensinado tanto, mesmo nos momentos Tramontina. Fazer o estágio em docência com você foi uma experiência maravilhosa, em que eu tive a oportunidade de dar a minha primeira aula (com direito a foto e tudo) e quando realmente descobri o amor em lecionar. E, como lembrança dessas aulas, eu posso dizer: “Kimcrível” esse professor!

À minha aluna de IC Samara, que me ajudou muito na bancada, nas análises do primeiro capítulo. Ela nunca poupou esforços, fizesse chuva ou sol, estivesse cedo ou tarde. Ela estava sempre lá para me ajudar e sempre com um sorriso no rosto. Muito obrigada por tudo! Desejo todo o sucesso do mundo no caminho que você escolheu.

Aos meus alunos de IC Anderson e Ingrid. Muito obrigada por toda a ajuda no manejo de mais de 250 potes de urina. O tempo estava super corrido e apertado e vocês me ajudaram a fazer com que desse tudo certo. Muito obrigada!

Às meninas do eterno grupo BIOTA: Andressa, Ellen e Laís. Entramos no LBNA juntas e aprendemos muito durante todo esse tempo. Depois de quase 6 anos conhecendo vocês, posso dizer que estou muito orgulhosa do que cada uma conquistou até hoje. Que ainda venham muitas vitórias e conquistas! Muito obrigada por toda a ajuda e amizade durante todo esse tempo.

Aos queridos amigos que eu fiz durante todo esse tempo no LBNA: Aline (a mais nova “Granadina”), André, Bia (que sempre se lembrava de mim e trazia brownies e docinhos), Carol, Emília, Fabrício (meu conselheiro de séries, doutorado e da vida), Genilton (meu companheiro de viagem), Isabele, Vanessa (Jane), Tamirys (minha aluna de IC, que virou minha amiga querida), Nathália Ferrari (minha “miga” science, inglesa e blogueirinha) e Nívea. Muito obrigada pelos desabafos diários, pela ajuda no laboratório e por tornarem meus dias muito mais alegres. O doutorado acabou, mas espero que as nossas festinhas e reuniões nunca acabem.

Aos outros alunos do LBNA: Ana Rafaela, Bia Ripper, Camila, Dani, Desiree, Nathália Martins, Suellen, Talita e Yuri. Pessoas que muitas vezes me ajudaram na bancada, fizeram companhia na hora do almoço ou animaram as festinhas do laboratório. Obrigada, pessoal!

Às professoras do LABAFs: Lili, Denise, Maria Lúcia e Ana Luisa. Muito obrigada por todos as dicas e conselhos durante os journals e fora deles também. Vocês todas são professoras e seres humanos maravilhosos. Uma verdadeira inspiração para mim.

Às professoras maravilhosas Vera e Maria Lúcia. Minhas mães científicas, que me deram a oportunidade na iniciação científica. À Flávia, que, na época, ainda era mestranda e foi a minha primeira orientadora na bancada. À Chris, artista nata, compositora, cantora, produtora de clipes e palhaça demais, que agora é professora da UFPR. Vou lembrar de vocês para sempre, com um carinho enorme. À Paola, uma grande amiga que tive a sorte de conhecer durante a minha primeira IC.

Aos alunos do LABAFs: Bianca, Luan e Christian e à técnica Giselle. Muito obrigada por toda a ajuda durante todo esse tempo.

Ao professor Francisco Tomás-Barberán, mais conhecido no CEBAS-CSIC como Pachi. Não há quem não se encante por ele. Além de ser extremamente inteligente, é muito atencioso, educado, paciente e divertido. Só de olhar os espectros de absorção, ele já sabia qual poderia ser o composto e arriscava até a posição da conjugação. Apesar de estar sempre muito ocupado, ele não deixava se abater. Andava pelos corredores sempre apressado, mas estava sempre dando pulinhos (como uma criança faz quando anda na rua). Foi muito atencioso durante o meu período de estágio no CEBAS e me deu até algumas palestras particulares. Posso dizer que foi uma “sorte” imensa ter te conhecido e ter tido a oportunidade de trabalhar com você. Professor, muito obrigada por todo o aprendizado durante esse período e por ter me recebido tão bem.

Aos grandes amigos que fiz no CEBAS-CSIC, Serene, Xiao, Gang, Maria, Paco e Pilar. Obrigada por tornarem a minha estadia em Murcia e meus dias no CEBAS uma experiência inesperada e surpreendente. Jamais imaginaria que em apenas dois meses faria uma amizade tão profunda e sincera. Muito obrigada pela ajuda, preocupação, passeios e almoços juntos. Sinto muita falta da companhia de vocês!

Aos alunos e funcionários do CEBAS-CSIC. Ao Alberto, que me ajudou a fazer as análises de proantocianidinas e elagitaninos. Ao Adrian, que me ensinou a usar os equipamentos do laboratório e de quebra, ainda aprendi várias palavras novas em espanhol. Ao Thiago da USP que, assim como eu, estava no CEBAS fazendo um estágio durante o doutorado. Muito obrigada pela ajuda na bancada, pelas conversas e almoços. Ao Carlos, Lola, Rocío, Angela e Raul, muito obrigada pela ajuda no laboratório e pelas conversas. Só tenho a agradecer muito a todos vocês que me ajudaram durante esse período.

Às maravilhosas Dora e Nilka, cearenses que conheci em Murcia. Muito coincidentemente, elas também são professoras da área de alimentos na UFC. Não sei o que teria sido das minhas primeiras semanas em Murcia sem vocês. Pessoas generosas, com o coração enorme e muito divertidas. Me ajudaram em um momento de perrengue enorme. Amei conhecer vocês! Muito obrigada pelos passeios, conversas e pela companhia maravilhosa. Obrigada por tudo, meninas.

A todos os meus voluntários, que se dedicaram à pesquisa e se sacrificaram fazendo a restrição alimentar e coletando urina por 48 horas. Esse estudo não existiria sem a ajuda de vocês. Muito obrigada!

Aos alunos e professores com quem eu tive a oportunidade de trabalhar em projetos paralelos ao meu doutorado: Paula Duarte, Laura, Lauro, Julio, Elaine e Nathália Moura. Muito obrigada por essa chance de trabalhar e aprender tanto com vocês.

Aos professores do LBNA, Alexandre e Tatiana, muito obrigada pelas sugestões e aprendizados durante os seminários de quarta-feira.

Aos funcionários do LABIM, Douglas e Rui, muito obrigada pela ajuda na liofilização das minhas amostras.

Ao grupo de caronas Méier – Fundão, seus criadores, motoristas e aos “caronistas”, que viraram meus amigos. Depois de 6 anos indo e voltando da Ilha do Fundão em ônibus lotados, com engarrafamentos quilométricos e com a ocorrência de assaltos diários, eu não aguentaria passar mais 4 anos aqui. Por causa disso, eu quase desisti, ainda no primeiro ano de Doutorado. Então, eu não posso deixar de agradecer a vocês. Muito obrigada!

RESUMO

INADA, Kim Ohanna Pimenta. Effect of technological processes on phenolic compounds contents of jabuticaba (Myrciaria jaboticaba) peel and seed and investigation of their ellagitannins metabolism in humans. Rio de Janeiro, 2018. Tese (Doutorado em Ciência de Alimentos) – Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, 2018.

A jabuticaba (Myrciaria jaboticaba) é uma berry brasileira que apresenta altos teores de compostos fenólicos, principalmente antocianinas e elagitaninos. Recentemente, diversos estudos relataram os efeitos benéficos da jabuticaba in vitro e in vivo. No entanto, até onde se sabe, o metabolismo dos compostos fenólicos da jabuticaba em humanos, especificamente dos elagitaninos, ainda não foi investigado. Os metabotipos de elagitaninos (A, B ou 0), ou seja, a classificação de cada indivíduo de acordo com sua habilidade em metabolizar esses compostos e produzir metabólitos microbianos específicos (urolitinas), foram descritos para a população europeia, sendo associados à composição da microbiota intestinal e ao estado nutricional. Na jabuticaba, a maior parte dos compostos fenólicos está concentrada na casca e semente (CSJ), partes da fruta que não são usualmente consumidas. Assim, em vistas ao seu aproveitamento como ingrediente potencialmente funcional, torna-se relevante investigar a aplicação de processos tecnológicos que podem afetar os teores de compostos fenólicos da CSJ. Dessa maneira, o presente estudo teve como objetivos principais: 1) avaliar o efeito da alta pressão hidrostática, da liofilização e da secagem em estufa sobre os teores de compostos fenólicos da CSJ; 2) investigar, pela primeira vez, o metabolismo de elagitaninos em adultos brasileiros, por meio do consumo da CSJ. Em relação aos processos tecnológicos aplicados à CSJ, ao contrário do esperado, a pressurização não aumentou o teor total de compostos fenólicos. A liofilização e a secagem em estufa a 75 °C acarretaram na produção de pós com perfis polifenólicos distintos, sendo o primeiro mais rico em antocianinas e o segundo em elagitaninos. No que se refere ao metabolismo de elagitaninos da CSJ, 35 voluntários (19 eutróficos e 16 com sobrepeso ou obesidade) concluíram o estudo clínico. Após o consumo da CSJ liofilizada, 63% dos participantes relataram a ocorrência de diarreia, causada provavelmente pela elevada dose de elagitaninos consumida. A maioria dos voluntários (83%) excretaram urolitinas (1,77 a 522 µmol), sendo a variabilidade interindividual observada possivelmente associada a diferenças na composição da microbiota colônica. Se, por um lado, a excreção de urolitinas não foi afetada pelo gênero e pelo estado nutricional, os indivíduos que relataram diarreia apresentaram excreção significativamente menor desses metabólitos. Assim como observado em estudos realizados na população europeia, não somente os três metabotipos (A, B e 0) foram encontrados na população brasileira, como a maior prevalência de metabotipo B também foi observada em indivíduos com sobrepeso e obesidade (37,5%), em comparação aos eutróficos (21,1%). No entanto, a prevalência do metabotipo 0 (17,1%) foi maior que a descrita na literatura (~10%), provavelmente em decorrência da diarreia. De fato, a distribuição dos metabotipos nos indivíduos sem esse sintoma (69,2% A, 23,1% B e 7,7% 0) foi similar à já descrita para populações saudáveis. Esse efeito colateral inesperado reforça o papel fundamental da microbiota intestinal na metabolização dos elagitaninos. Assim, os resultados desse estudo mostraram que ambos os métodos de desidratação permitiram a obtenção de pó de CSJ com potenciais propriedades funcionais e que, apesar de diferenças culturais entre as populações brasileira e europeia, a composição de sua microbiota intestinal responsável por metabolizar os elagitaninos parece ser similar.

Palavras-chave: Jabuticaba, compostos fenólicos, antocianinas, elagitaninos, ácido elágico, liofilização, secagem em estufa, alta pressão hidrostática, urolitina, metabotipo, sobrepeso, obesidade.

ABSTRACT

INADA, Kim Ohanna Pimenta. Effect of technological processes on phenolic compounds contents of jabuticaba (Myrciaria jaboticaba) peel and seed and investigation of their ellagitannins metabolism in humans. Rio de Janeiro, 2018. Tese (Doutorado em Ciência de Alimentos) – Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, 2018.

Jabuticaba (Myrciaria jaboticaba) is a Brazilian berry, which presents high phenolic compounds contents, mainly anthocyanins and ellagitannins. Recently, several studies have reported the beneficial effects of jabuticaba in vitro and in vivo. However, to the best of our knowledge, the metabolism of jabuticaba phenolic compounds in humans, specifically the ellagitannins, has not yet been investigated. Elagitannins metabotypes (A, B or 0), i.e., the classification of each individual according to their ability to metabolize these compounds and to produce specific microbial metabolites (), were described for the European population, being associated with gut microbiota composition and nutritional status. In jabuticaba, most of phenolic compounds are concentrated in its peel and seed (JPS), parts of the fruit that are not usually consumed. Therefore, with the intention of its use as a potential functional ingredient, it becomes relevant to investigate the application of technological processes that can affect the phenolic compounds contents of JPS. Thus, the main objectives of this study were: 1) to evaluate the effect of high hydrostatic pressure, freeze-drying and oven- drying on JPS phenolic compounds contents; 2) to investigate, for the first time, the metabolism of elagitannins in Brazilian adults, through JPS consumption. In relation to the technological processes applied to JPS, contrary to our expectations, pressurizing did not increase the total phenolic compounds contents. Freeze- and oven-drying at 75 °C resulted in the production of powders with distinct polyphenolic profiles, the former being richest in anthocyanins and the latter in ellagitannins. Regarding the metabolism of JPS ellagitannins, 35 volunteers (19 normoweight and 16 overweight or obese) accomplished the clinical study. After the consumption of freeze-dried JPS, 63% of volunteers reported the occurrence of diarrhea, probably caused by the high dose of ellagitannins ingested. Most of the volunteers (83%) excreted urolithins (1.77 to 522 μmol), and the large interindividual variability observed was possibly associated with differences in the gut microbiota composition. If, on the one hand, gender and nutritional status did not affect urolithins excretion, volunteers that reported diarrhea had significantly lower excretion of these metabolites. As observed in studies conducted in Europe, not only the three metabotypes (A, B and 0) were found in the Brazilian population, but also a higher prevalence of metabotype B was observed in overweight and obese individuals (37.5%), in comparison to normoweight (21.1%). However, the prevalence of metabotype 0 (17.1%) was higher than that described in the literature (~ 10%), probably due to diarrhea. In fact, the distribution of metabotypes in subjects without this symptom (69.2% A, 23.1% B and 7.7% 0) was similar to that already described for healthy populations. This unexpected side effect reinforces the essential role of intestinal microbiota in ellagitannins metabolism. Thus, the results of this study showed that both dehydration methods yield a JPS powder with potential functional properties and that, despite cultural differences between European and Brazilian populations, their gut microbiota composition may be similar in terms of ellagitannins metabolism.

Key-words: Jabuticaba, phenolic compounds, anthocyanins, ellagitannins, , freeze- drying, oven-drying, high hydrostatic pressure, , metabotypes, obese, overweight.

PUBLICATION ARISING FROM THIS THESIS

Scientific conference abstracts:

LACERDA, E. C. Q.; INADA, K.O.P.; NUNES, S.; FINOTELLI, P.; TORRES, A. G.; MONTEIRO, M. C.; PERRONE, D. Characterization of jabuticaba (Myrciaria jaboticaba) and jussara (Euterpe edulis) powders produced by different drying technologies. 7th International Conference on and Health, Tours, France, 2015.

NUNES, S.; INADA, K.O.P.; PERRONE, D.; MONTEIRO, M. C. Obtenção de farinha de jabuticaba (Myrciaria jaboticaba) com elevado potencial bioativo através de secagem em estufa. 7° Semana de Integração Acadêmica da UFRJ, Rio de Janeiro, Brasil, 2016.

INADA, K.O.P.; PERRONE, D.; MONTEIRO, M. Polyphenols of jabuticaba (Myrciaria jaboticaba), a Brazilian berry, are extensively metabolized in humans: a pilot clinical study. 8th International Conference on Polyphenols and Health, Quebec City, Canada, 2017.

OTHER PUBLICATION DURING PhD THESIS PERIOD

Scientific journal papers:

INADA, K.O.P.; DUARTE, P.A.; LAPA, J.; MIGUEL, M.A.L.; MONTEIRO, M. Jabuticaba (Myrciaria jaboticaba) juice obtained by steam-extraction: phenolic compound profile, antioxidant capacity, microbiological stability, and sensory acceptability. Journal of Food Science and Technology, v. 55, p. 52-61, 2018.

SOARES, E.R.; MONTEIRO, E.B.; DE BEM, G.F.; INADA, K.O.P.; TORRES, A.G.; PERRONE, D.; SOULAGE, C.O.; MONTEIRO, M.C.; RESENDE, A.C.; MOURA-NUNES, N.; COSTA, C.A.; DALEPRANE, J.B. Up-regulation of Nrf2-antioxidant signaling by Açaí (Euterpe oleracea Mart.) extract prevents oxidative stress in human endothelial cells. Journal of Functional Foods, v. 37, p. 107-115, 2017.

Scientific conference abstracts:

SUEMITSU, L.Y.; INADA, K.O.P.; PERRONE, D.; MONTEIRO, M. C.; MELO, L. Simultaneous optimization of jabuticaba nectar acceptance based on pulp and sucrose concentrations using surface response methodology. 11th Pangborn Sensory Science Simposium, Gothenburg, Suécia, 2015.

SUEMITSU, L.Y.; MELO, L.; PERRONE, D.; MONTEIRO, M. C.; INADA, K.O.P. Acceptance of ready-to-drink jabuticaba nectar added with lyophilized jussara pulp. 11 SLACA - Simpósio Latino Americano de Ciência de Alimentos, Campinas, Brasil, 2015.

SUEMITSU, L.Y.; INADA, K.O.P.; MONTEIRO, M. C.; PERRONE, D.; MELO, L. Otimização da aceitação sensorial de néctar de jabuticaba em relação às concentrações de polpa e sacarose. XXXVI Jornada Giulio Massarani de Iniciação Científica, Tecnológica, Artística e Cultural UFRJ, Rio de Janeiro, Brasil, 2015.

DUARTE, P.A.; INADA, K.O.P.; MONTEIRO, M.C. Estabilidade química e de compostos bioativos do suco de jabuticaba obtido pela extração por arraste a vapor. XXXVI Jornada Giulio Massarani de Iniciação Científica, Tecnológica, Artística e Cultural UFRJ, Rio de Janeiro, Brasil, 2015.

SILVA, J.L.C.E.; GUIMARAES, T.E.S.; DUARTE, P. A.; INADA, K.O.P; MONTEIRO, M. C.; MIGUEL, M.A.L. Controle de qualidade e caracterização microbiológica de suco de jabuticaba (Myrciaria jaboticaba) obtido por arraste a vapor. XXXVI Jornada Giulio Massarani de Iniciação Científica, Tecnológica, Artística e Cultural UFRJ, Rio de Janeiro, Brasil, 2015.

FERNANDES, P.O.; INADA, K.O.P.; PERRONE, D.; MONTEIRO, M. C. Avaliação da estabilidade química de polpa de jabuticaba (Myrciaria jaboticaba) processada por alta pressão hidrostática. 7° Semana de Integração Acadêmica da UFRJ, Rio de Janeiro, Brasil, 2016.

LIST OF FIGURES

Literature Review Figure 1. Jabuticaba tree in flowering (A) and harvest (B) periods………...... 29 Figure 2. Jabuticaba fruit (Myrciaria jaboticaba) and their fractions………………. 30 Figure 3. Chemical structure of skeleton and its subclasses……………… 33 Figure 4. Chemical structure of flavan-3-ols……………………………………….. 34 Figure 5. Changes in conformation of anthocyanins at different pH values and the consequent changes in color……………………………………………… 35 Figure 6. Chemical structure of derivatives of benzoic acids……………………….. 36 Figure 7. Chemical structure of derivatives of cinnamic acids……………………... 36 Figure 8. Tellimagrandin II, a hydrolysable which contains both HHDP and gallic acids………………………………………………………………... 50 Figure 9. HHDP () undergoes spontaneous rearrangement in ellagic acid……………………………………………... 51 Figure 10. Hydrolysis of vescalagin, an C-glycoside, yields ellagic acid and vescalin………………………………………………………………. 52 Figure 11. Chemical structures of ellagitannins reported in jabuticaba by several studies……………………………………………………………………. 62 Figure 12. Schematic diagram of basic equipment design for high hydrostatic pressure processing of foods……………………………………………... 64 Figure 13. Cell structures of E. coli and S. aureus untreated and high hydrostatic pressure treated at 500 MPa for30 min…………………………………… 65 Figure 14. Onion treated for 3 minutes with 100 MPa (A and D), 300 MPa (B and E) and 600 MPa (C and F)…………………………………………………… 66 Figure 15. Phase diagram of pure water……………………………………………… 78 Figure 16. Schematic diagram of a typical shelf freeze-dryer...……………………… 79 Figure 17. Catabolism of ellagitannins ( and vescalagin) and ellagic acid……………………………………………………………………….. 81 Figure 18. Production of urolithins from EA by gut microbiota and urolithin metabotypes……………………………………………………………… 83 Figure 19. Subjects stratification according to urolithins production after ET and EA intake……………………………………………………………………... 85 Chapter 1 Figure 1. Pareto charts and surface response charts of high hydrostatic pressure processing conditions that affected the phenolic composition and antioxidant activity of jabuticaba peel and seeds according to experimental design results...... 101 Figure 2. Pareto charts and surface response charts of oven-drying conditions that affected the phenolic composition and antioxidant activity of jabuticaba peel and seeds according to experimental design results…………………. 109 Figure 3. Comparing the extremes of dehydration conditions (freeze- and oven- drying at 75 °C/22 h) on the main classes of phenolic compounds from jabuticaba peel and seeds…………………………………………………. 111

Chapter 2 Figure 1. Total urinary excretion (µmol) of metabolites at different intervals after acute intake of jabuticaba peel and seeds…………………………………. 130 Figure 2. Urolithins urinary excretion (µmol) by volunteers (n = 35) grouped according gender (A), nutritional status (B) and diarrhea occurrence (C)... 132 Figure 3. Distribution of urolithins metabotypes in different groups of the study population………………………………………………………………... 136

LIST OF TABLES

Literature Review Table 1. Non-ellagitannins phenolic compounds detected in jabuticaba fractions and its derived products and extracts by different studies……………….. 39 Table 2. Ellagitannins, ellagic acid and its derivatives detected in jabuticaba fractions and its derived products and extracts by different studies……... 54 Table 3. Effect of high hydrostatic pressure (HHP) on food microstructure and phenolic compounds extraction and contents evaluated by different studies……………………………………………………………………. 67 Table 4. Effect of oven-drying (OD) and freeze-drying (FD) on phenolic compounds contents evaluated by different studies………………………. 73 Table 5. Urolithin urinary metabotypes distribution in different human trials…….. 87

Chapter 1 Table 1. Phenolic compounds contents and antioxidant activity of unprocessed and HHP processed jabuticaba peel and seed…………………………………. 100 Table 2. Identification of , ellagic acid derivatives and ellagitannins by HPLC-DAD-MS-Q of freeze-dried and oven-dried jabuticaba peel and seeds……………………………………………………………………… 105 Table 3. Phenolic compounds contents, antioxidant activity, instrumental color and water activity of freeze-dried and oven-dried jabuticaba peel and seed 106 Table 4. Identification of flavan-3-ols monomers and its phloroglucinol adducts in freeze-dried jabuticaba peel and seed by HPLC-ESI-MS-MS after proanthocyanidins phloroglucinolysis…………………………………… 114

Chapter 2 Table 1. Soluble and insoluble phenolic compounds contents in the amount of jabuticaba peel and seed powder (20 g) consumed by the volunteers…….. 124 Table 2. Characteristics of the volunteers participating in the study (n = 35)…...... 126 Table 3. Urinary excretion of metabolites (µmol) at different intervals after acute consumption of jabuticaba peel and seed…………………………………. 128 Table 4. Urinary recovery (%) of soluble and insoluble ellagitannins and ellagic metabolites stratified according to the level of urolithins produced………. 134

LIST OF ABBREVIATIONS

AA Antioxidant activity aw Water activity ABTS 2,2’-Azino-bis (3-ethylbenzo-thiazoline-6-sulfonic acid) diammonium salt ANOVA Analysis of variance CAPES Coordenação de Aperfeiçoamento de Pessoal de Nível Superior CNPq Conselho Nacional de Desenvolvimento Científico e Tecnológico DAD Diode array DWB Dry weight basis ESI Electrospray ionization FAPERJ Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro FD Freeze-drying FRAP Ferric Reducing Antioxidant Power HAT Hydrogen Atom Transfer HHDP Hexahydroxydiphenic acid HPLC High Performance Liquid Chromatography HPP High pressure processing IsoUro Isourolithin JPS Jabuticaba peel and seeds JPSP Jabuticaba peel and seeds powder MPa Megapascal MS Mass spectrometry MSD Mass selective detector MW Molecular weight OD Oven-drying PPO Polyphenoloxidase PTFE Polytetrafluoroethylene PVDF Polyvinylidene Fluoride PVPP Polyvinylpolypyrrolidone Q Quadrupole SET Sigle Electron Transfer SIM Selective ion monitoring TEAC Trolox Equivalent Antioxidant Capacity TPTZ 2,4,6-tripyridyl-S-triazine TROLOX 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid Uro Urolithin

SUMMARY

Thesis introduction…………………………………………………………………….. 26

Thesis objectives………………………………………………………………………... 27

Literature Review……………………………………………………………………… 28 1. Jabuticaba……………………………………………………………………………. 29 1.1 Phenolic compounds composition…………………………………………………… 31 1.2.1 Non-ellagitannins………………………………………………………………….. 32 1.2.2 Ellagitannins………………………………………………………………………. 50 2. Effects of technological processes on phenolic compound contents……………….. 64 2.1 High Hydrostatic Pressure…………………………………………………………... 64 2.2 Dehydration methods: oven-drying and freeze-drying………………………………. 72 3. Metabolism of ellagitannins……………………………………………………… 80

Chapter 1……………………………………………………………………………….. 90 1. Introduction………………………………………………………………………….. 91 2. Material and methods……………………………………………………………….. 92 2.1 Standards and chemicals…………………………………………………………….. 92 2.2 Samples...... 93 2.3 High Hydrostatic Pressure processing of jabuticaba peel and seed…………………. 93 2.4 Oven drying of jabuticaba peel and seed…………………………………………….. 93 2.5 Freeze drying of jabuticaba peel and seed…………………………………………… 94 2.6 Phenolic compounds analysis………………………………………………………... 94 2.6.1 Phenolic acids, flavonols and anthocyanins extraction and quantification by HPLC-DAD……………………………………………………………………………… 94 2.6.2 Ellagitannins, ellagic acid derivatives and gallic acid extraction and quantification by HPLC-DAD-MS………………………………………………………. 95 2.6.3 Proanthocyanidins extraction and identification by HPLC-ESI-MS-MS…………... 96 2.7 Antioxidant activity…………………………………………………………………... 97 2.8 Instrumental color measurement…………………………………………………….. 97 2.9 Water activity………………………………………………………………………... 97 2.10 Polyphenoloxidase activity…………………………………………………………. 97 2.11 Statistical analyses…………………………………………………………………. 98 3. Results and Discussion………………………………………………………………. 98 3.1 Jabuticaba peel and seed did not exhibit polyphenoloxidase activity………………... 98 3.2 Pressurization was ineffective in increasing phenolic compounds contents and antioxidant activity of jabuticaba peel and seed…………………………………………. 99 3.3 Freeze-drying of jabuticaba peel and seed led to higher anthocyanins contents, while oven-drying led to higher ellagitannins contents………………………………………… 104 3.4 Analysis of freeze-dried jabuticaba peel and seed after phloroglucinolysis indicates a diverse profile of proanthocyanidins…………………………………………………... 113 4. Conclusion……………………………………………………………………………. 114 Acknowledgments……………………………………………………………………… 115

Chapter 2……………………………………………………………………………….. 116 1. Introduction………………………………………………………………………….. 117 2. Material and methods……………………………………………………………….. 119 2.1 Standards and chemicals…………………………………………………………….. 119 2.2 Production of jabuticaba powder……………………………………………………. 119 2.3 Ellagitannins, ellagic acid derivatives and gallic acid analysis by HPLC-DAD-MS 119 2.4 Clinical study design………………………………………………………………… 120 2.5 Urolithin metabolites analysis by HPLC-DAD-MS………………………………….. 121 2.6 Statistical analysis…………………………………………………………………… 122 3. Results and Discussion………………………………………………………………. 123 3.1 Ellagitannins, ellagic acid derivatives and gallic acid are mainly found in the insoluble fraction of jabuticaba peel and seed powder…………………………………... 123 3.2 Jabuticaba ellagitannins are mainly metabolized to urolithin glucuronides by colonic microbiota………………………………………………………………………. 125 3.3 Excretion of urolithins showed high interindividual variability and was affected by diarrhea occurrence, but not by gender or nutritional status……………………………. 130 3.4 Up to 36% of ellagitannins and ellagic acid metabolites were recovered in urine within 48 hours of jabuticaba peel and seed powder consumption………………………. 133 3.5 Distribution of metabotypes in Brazilians was similar to that of European populations, being affected by both nutritional status and diarrhea occurrence………… 135 4. Conclusion……………………………………………………………………………. 137

Acknowledgments……………………………………………………………………… 138

Thesis General Conclusion…………………………………………………………….. 139

Future Perspectives…………………………………………………………………….. 140

References………………………………………………………………………………. 141

Annexes…………………………………………………………………………………. 159

26

THESIS INTRODUCTION

Jabuticaba (Myrciaria jaboticaba) is a Brazilian berry, which presents high phenolic compounds contents, specially anthocyanins and ellagitannins. In recent years, several studies have reported the beneficial effects of jabuticaba in vitro and in vivo, both in animal models, and more recently, in humans. However, to the best of our knowledge, the metabolism of jabuticaba phenolic compounds, specifically the ellagitannins, in humans has not been yet investigated. Elagitannins metabotypes (A, B or 0), i.e., the classification of each individual according to their ability to metabolize these compounds and to produce specific microbial metabolites (urolithins), were described for the European population, being associated with gut microbiota and nutritional status. In jabuticaba, most of phenolic compounds are concentrated in jabuticaba peel and seed, fractions that are not usually consumed. Therefore, with the intention of its use as a potential functional ingredient and considering that some technological processes modifies the phenolic profile, it becomes relevant to investigate the effect of high hydrostatic pressure and dehydration methods on phenolic compounds contents of jabuticaba peel and seed. For this purpose, the thesis was divided into two chapters. Chapter one was dedicated to the investigation of different technological processes (high hydrostatic pressure, freeze-drying and oven-drying) on the chemical composition and physical characteristics of jabuticaba peel and seed, with emphasis on phenolic compounds. Chapter two was dedicated to the investigation of metabolism of ellagitannins and ellagic acid from jabuticaba, in normoweight, and overweight and obese Brazilian adults. It is important to note that this was the first study that evaluated the metabolism of ellagitannins in a non-European population.

27

THESIS OBJECTIVES

General:

The general objective of this study was to evaluate the effect of different technological processes on the physical and chemical characteristics of jabuticaba peel and seed, with emphasis on phenolic compounds contents and to investigate the metabolism of their elagitannins and ellagic acid in normoweight, and overweight and obese humans.

Specifics:

Chapter 1: Investigate the effect of high hydrostatic pressure on phenolic compounds contents and antioxidant activity of jabuticaba peel and seed; Investigate the effect of dehydration methods, oven-drying and freeze-drying, on phenolic compounds contents, antioxidant activity, color and water activity of jabuticaba peel and seed; Evaluate the polyphenoloxidase activity of jabuticaba peel and seed.

Chapter 2: Determine the soluble and insoluble phenolic compounds contents of jabuticaba peel and seed; Assess urinary excretion of ellagitannins metabolites over 48 hours in normoweight, and overweight and obese adults after the acute consumption of jabuticaba peel and seed; Stratify the studied population according to the level and profile of urolithins excreted; Calculate the urinary recovery of ellagitannins metabolites; Calculate the distribution of metabotypes in the study population.

28

LITERATURE REVIEW

29

1. Jabuticaba

Jabuticaba is Brazilian berry belonging to the Myrtaceae family and Myrciaria genus, that is also termed as Plinia. It is cultivated from Brazil’s south to north, but the highest production occurs in the southeast states of São Paulo, Rio de Janeiro, Minas Gerais and Espírito Santo (Ascheri et al., 2006; Citadin et al., 2010; Sasso et al., 2010). Several species have been described in the literature, but only three are naturally dispersed and cultivated in Brazil, including Myrciaria trunciflora O. Berg (Cabinho jabuticaba), Myrciaria cauliflora (Mart.) O. Berg (Paulista, Ponhema or Assu jabuticaba) and Myrciaria jaboticaba (Vell.) O. Berg (Sabará jabuticaba) (Citadin et al., 2010; Wu et al., 2013). Jabuticaba tree exibits medium to large size, with 6 to 9 meters of height, and it is characterized by presenting flowers and fruits that are born directly from the trunks and main branches (Figure 1), giving it an ornamental characteristic. Jabuticaba tree from Paulista variety is known for its large size and high production capacity, exhibiting large fruits. On the other hand, Sabará variety presents a medium size tree and is the most appreciated and cultivated, producing small and sweet fruits (de Jesus et al., 2004; Lima et al., 2008).

Figure 1. Jabuticaba tree in flowering (A) and harvest (B) periods.

Jabuticaba fruit is a globular-type berry with diameter of up to 3.5 cm (Figure 2). The peel is thin and fragile and exhibits dark purple to black color, which varies according to its ripeness stage, and covers a pulp with sweet and slightly acid and tange taste and appreciable flavor, which contains from 1 to 4 seeds (de Jesus et al., 2004; Lima et al., 2008; Wu et al., 2013). 30

Figure 2. Jabuticaba fruit (Myrciaria jaboticaba) and their fractions (Inada et al., 2015).

Jabuticaba is still considered a homegrown orchard plant, with small production limited to some regions of the country, where poor families collect the in natura fruits and sell them along roadsides. This informal activity presents great economic and social relevance, as it provides additional income to these families during jabuticaba harvest period (Citadin et al., 2010). However, jabuticaba exhibits a short shelf life, of up to three days after harvesting, due to the intense water loss, microbial deterioration and pulp fermentation. Thus, although jabuticaba is popular throughout the country, its high perishability impairs the commercialization of the in natura fruit (Agostini et al., 2009). Nevertheless, this fruit presents great market potential, since it is very appreciated for the production of artisanal products, such as juices, jams, jellies, vinegars, liqueurs and wine (Citadin et al., 2010). According to Citadin et al. (2010), this fruit can conquer markets, as long as basic and technological research is performed in this culture. Regarding jabuticaba chemical composition, some studies have reported that this fruit is a source of vitamins, minerals and fibers (Inada et al., 2015), and presents high antioxidant activity and phenolic compounds contents, especially anthocyanins and ellagitannins (Pereira et al., 2017; Plaza et al., 2016; Inada et al., 2015; Wu et al., 2012). However, most of these phenolic compounds are concentrated in jabuticaba peel and seeds, fractions that are not usually consumed, and that represent about 40% of the fruit's weight (Morales et al., 2016; Inada et al., 2015). For this reason, jabuticaba peel and seeds could be used for the development of natural 31 dyes, functional ingredients and dietary supplements by the food and pharmaceutical industries (Morales et al., 2016; Inada et al., 2015; Gurak et al., 2014). In fact, several studies have used these fractions in the development of food products, such as yogurts (Alves et al., 2013), petit Suisse cheese (Pereira et al., 2016), ferment dessert (Neta et al., 2018), jelly (Dessimoni-Pinto et al., 2011), functional tea (Da Silva et al., 2017) and sausages (Baldin et al., 2018, 2016; De Almeida et al., 2015). Thus, considering the potential applications of jabuticaba peel and seeds in the development of food products, it is relevant to evaluate the effect of technological processes that can increase or maintain the phenolic compounds contents, aiming to add value to these fruit fractions that are usually discarded. In addition, several studies have reported the beneficial effects of jabuticaba and its extracts in in vitro (Silva et al., 2016; Caloni et al., 2015; Wang et al., 2014; Leite-Legatti et al., 2012; Reynertson et al., 2006) and animal models (Baseggio et al., 2018; Batista et al., 2018; Lamas et al., 2018; Moura et al., 2018; Quartrin et al., 2018; Batista et al., 2017; De Souza et al., 2017; Hsu et al., 2016; Batista et al., 2014; Araújo et al., 2014; Dragano et al., 2013) and, more recently, in clinical trials (Balisteiro et al., 2018; Plaza et al., 2016). However, to the best of our knowledge, the metabolism of jabuticaba phenolic compounds, specifically ellagitannins and ellagic acid, in humans, has not been yet investigated.

1.1 Phenolic compounds composition

Several studies have observed that jabuticaba presents high contents of phenolic compounds, which are mainly concentrated in their peel and seeds (Da Silva et al., 2017; Pereira et al., 2017; Plaza et al., 2016; Inada et al., 2015; Lima et al., 2011). The dark purple to black color of jabuticaba peel indicates the presence of high anthocyanins contents, which have been described in some studies as the major phenolic compounds of this fruit, with contents of up to 3,222 mg/100 g, on dry weight basis (dwb) (Plaza et al., 2016; Inada et al., 2015; Wu et al., 2012; Reynertson et al., 2008). On the other hand, some studies have also reported the presence of ellagitannins in jabuticaba (Neves et al., 2018; Pereira et al., 2017; Moura et al., 2018; Plaza et al., 2016; Alezandro et al., 2013; Wu et al., 2012). Among these studies, Pereira et al. (2017) observed that ellagitannins were the major compounds, with contents of up to 5,822 mg/100 g, on dwb, suggesting that ellagitannins contents may have been underestimated in the other studies. Corroborating this data, some studies reported high levels of ellagic acid, after acid hydrolysis (Abe et al., 2011; Alezandro et al., 2013), which indicates the presence of high 32 contents of elagitannins in this fruit. Thus, considering the recent relevance of jabuticaba ellagitannins, this item was divided into non-ellagitannins and ellagitannins.

1.2.1 Non-ellagitannins

Phenolic compounds are secondary metabolites of plants which, although are not essential for the plant growth and development, they have many functions, including the protection against herbivores, microbial infections and UV radiation, attractant of pollinators and seed-dispersing animals, among others (Del Rio et al., 2012). They present at least one aromatic ring attached to one or more hydroxyl groups (Crozier et al., 2009; Vermerris & Nicholson, 2006), ranging from simple structures with low molecular weight to large and complex molecules, presenting more than 8,000 different structures already described (Del Rio et al., 2012; Crozier et al., 2009). The name "polyphenols" is also used to designate these compounds. Nonetheless, this term may be misleading since it implies that all of them are phenolic polymers. However, some substances, such as gallic acid, has only one phenolic ring (Vermerris & Nicholson, 2006). For this reason, some authors also use the name “(poly)phenolic compounds” (Del Rio et al., 2012). Due to the great structural variability, phenolic compounds can be classified according to the number and arrangement of their carbon atoms as and non-flavonoids (Del Rio et al., 2012; Crozier et al., 2009). Flavonoids are phenolic compounds with 15 carbons, which present 2 aromatic rings (A and B) connected by a three carbons bridge (C6–C3–C6), forming an oxygenated heterocycle (C) (Figure 3) (Del Rio et al., 2012; Crozier et al., 2009; Manach et al., 2004). The main subclasses of flavonoids found in diet are flavonols, flavones, flavan-3-ols, anthocyanidins, flavanones and isoflavones (Figure 3) (Del Rio et al., 2012; Crozier et al., 2009; Manach et al., 2004). However, other minority subclasses can also be found in diet, such as dihydroflavonols, flavan-3,4-diols, coumarins, chalcones, dihydrochalcones and aurones (Figure 3) (Del Rio et al., 2012; Crozier et al., 2009). The basic skeleton of flavonoids may have numerous substituents, such as hydroxyl and sugars. In fact, most flavonoids exist in the glycosides forms, rather than the aglycones (Crozier et al., 2009). Among flavonoids subclasses, flavonols are the most widespread in nature (Crozier et al., 2009; Manach et al., 2004). Its main dietary representatives are kaempferol, quercetin, isorhamnetin and myricetin, which are mainly found in glycoside forms (Del Rio et al., 2012; Crozier et al., 2009). Since the biosynthesis of these compounds is stimulated by sunlight, the flavonols are mainly found in the fruit skins and aerial leaves (Manach et al., 2004). On the 33 other hand, flavones, such as apigenin and luteolin, are less widespread in nature. The only important edible sources of flavones are celery, parsley and some herbs (Crozier et al., 2009; Manach et al., 2004).

Figure 3. Chemical structure of flavonoid skeleton and its subclasses (Crozier et al., 2009)

Flavan-3-ols are the most structurally complex subclass of flavonoids, ranging from simple monomers of (+)-catechin and (-)-epicatechin, which can be hydroxylated to form (epi)gallocatechins and can undergo esterification with gallic acid yielding several complex structures (Figure 4), including dimeric, oligomeric and polymeric proanthocyanidins, which are also known as condensed (Del Rio et al., 2012; Crozier et al., 2009). Proanthocyanidins can occur as polymers of up to 50 units of flavan-3-ol monomers. When composed exclusively by units of (epi)catechin are called procyanidins, the most abundant type of proanthocyanidins in nature. On the other hand, if proanthocyanidins are composed by 34

(epi)gallocatechins are termed as prodelphinidins (Karonen et al., 2007). The proanthocyanidins form complexes with salivary proteins being responsible for the characteristic astringency of several fruits (e.g. apples, berries, grapes, pears, peaches) and beverages (e.g. wine, beer, tea, cider), and for the bitterness of chocolate (Manach et al., 2004). However, because of the wide variety and complexity of chemical structures, it is difficult to estimate the content of polymeric proanthocyanidins in food (Karonen et al., 2007).

Figure 4. Chemical structure of flavan-3-ols (Adapted from Crozier et al., 2009)

Anthocyanidins are the aglycone forms of anthocyanins, pigments widely distributed in nature, which confer red, blue and purple color to flowers and fruits (Castañeda-Ovando et al., 2009). In plant tissues, these compounds are mainly found conjugated to sugars, especially glucose, which often occurs at carbon 3 (Crozier et al., 2009). In addition, anthocyanins can also be esterified with different organic acids, such as citric and malic, and phenolic acids (Manach et al., 2004). The conjugation of anthocyanidins with these compounds increases the molecule stability (Manach et al., 2004). Due to the wide variety of their conjugated forms, to date, more than 600 anthocyanins have been reported (Zhang et al., 2014), of which the most common anthocyanidins are cyanidin, pelargonidin, delphinidin, peonidin, petunidin and malvidin (Crozier et al., 2009). Anthocyanins are very unstable and susceptible to degradation. Their stability is affected by several factors, such as pH, storage temperature, light, oxygen, metal ions, among others. They can exist in different forms according to the pH, which changes its color (Zhang et al., 2014) (Figure 5). However, the instability of anthocyanins, special at neutral to alkaline pH limits their use as colorants (Zhang et al., 2014). 35

Figure 5. Changes in conformation of anthocyanins at different pH values and the consequent changes in color (Zhang et al., 2014).

To conclude, among flavonoids, flavanones are present in high concentrations in citrus fruits, where the most common compounds are hesperetin-7-O-rutinoside and naringenin-7-O- rutinoside (Crozier et al., 2009; Manach et al., 2004). On the other hand, isoflavones, such as daidzein and genistein, are found almost exclusively in leguminous plants, with the highest concentrations occurring in soybean (Crozier et al., 2009). Among the non-flavonoids, two main classes of phenolic acids can be found: derivatives of benzoic and cinnamic acids (Manach et al., 2004). Hydroxybenzoic acids are characterized by presenting a carboxyl group substituted on a phenol, such as the compounds gallic, p- hydroxybenzoic, 3,4-dihydroxybenzoic, salicylic and vanillic acids (Figure 6) (Vermerris & Nicholson, 2006). Among phenolic acids, gallic acid is the commonest compound (Crozier et al., 2009). In relation to hydroxycinnamic acids, the most common compounds present a C6– C3 skeleton, including caffeic, ferulic and sinapic acids (Figure 7) (Crozier et al., 2009; Vermerris & Nicholson, 2006). In addition, cinnamic acids are commonly found in plants as esters of quinic, shikimic and tartaric acids, such as chlorogenic acid, which is an ester of caffeic and quinic acids (Figure 7) (Vermerris & Nicholson, 2006).

36

Figure 6. Chemical structure of derivatives of benzoic acids (Adapted from Vermerris & Nicholson, 2006)

Figure 7. Chemical structure of derivatives of cinnamic acids (Adapted from Vermerris & Nicholson, 2006)

37

Several studies have characterized the non-ellagitannins phenolic compounds present in the different fractions of jabuticaba fruit, as well as in its derived products and extracts (Table 1). Among these compounds, anthocyanins, flavonols, flavanones, flavones, hydroxycinnamic acids and hydroxybenzoic acids were reported. Anthocyanins were the major flavonoids detected in jabuticaba, with the highest concentrations found in the peel (334 to 3,222 mg/100 g on dwb) (Table 1). Among these, only two anthocyanins were reported: cyanidin-3-O-glucoside and delphinidin-3-O-glucoside. On the other hand, a large variety of flavonols was described (Table 1), including quercetin, myricetin, kaempferol and isorhamnetin and their glycoside forms. Besides, some forms of quercetin were found both glycosylated and conjugated to phenolic acids (e.g. quercetin- caffeoyl-hexoside) (Table 1). Among flavonoids, the subclass of flavones was the minor one, in which luteolin was the only compound detected, at low concentrations (0.004 mg/100 g, dwb) (Table 1). For flavanones, the major compound was hesperitin-7-O-rutinoside, which was detected only in the pulp, while naringenin and 3,5,7-trihydroxyflavanone were detected at low concentrations in the whole fruit (0.004 to 0.05 mg/100 g, dwb). Considering the flavan-3-ols, some studies reported the presence of (+)-catechin, (-) epicatechin, (-)-epicatechin gallate and some of their polymeric forms, the procyanidins A2, B1 and B2 (Table 1). Among these compounds, the major ones were the flavan-3-ols monomeric forms, with the highest amounts found in jabuticaba peel extracts (8.2 to 117.5 mg/100 g on dwb). On the other hand, the polymeric forms were found in low concentrations in jabuticaba peel tea (0.012 mg/100 mL) and pulp (1.53 to 4.63 mg/100 g, dwb). However, it is noteworthy that due to the wide variety of proanthocyanidins molecules found in nature, the analysis of polymeric proanthocyanidins is difficult (Karonen et al., 2007). The large number of isomers leads to compounds coelution and large unresolved peaks in chromatographic analysis (Lazarus et al., 1999). In addition, with the increase of polymerization degree, the ionization efficiency decreases, making it difficult to interpret the compounds MS spectra (Karonen et al., 2007). Considering that berries are usually rich in proanthocyanidins (Basanta et al., 2016; Fracassetti et al., 2013; Tarascou et al., 2010), it is possible that the proanthocyanidins of jabuticaba were under-identified and underestimated. Among the phenolic acids, seven hydroxycinnamic acids were reported in jabuticaba, at low concentrations, including trans-cinnamic, ferulic, caffeic, chlorogenic and p-, o- and m- coumaric acids. In addition, six hydroxybenzoic acids were also described, including benzoic, 3,4-dihydroxybenzoic, syringic and gallic acids, and the gallotannin pentagalloyl-hexose. 38

Among these, gallic acid was the major compound detected with the highest concentrations found in jabuticaba seed (422 mg/100 g, dwb).

39

Table 1. Non-ellagitannins phenolic compounds detected in jabuticaba fractions and its derived products and extracts by different studies. Compounds Fruit fraction, derived product or extract Contents References Anthocyanins 81 mg/100 g, dwb Reynertson et al. (2008) Delphinidin-3-O-glucoside 73.6 mg/100 g, dwb Wu et al. (2012) Whole fruit 23.5 mg/100 g, dwb Alezandro et al. (2013) 48 mg/100 g, dwb Inada et al. (2015) 271 mg/100 g, dwb Lima et al. (2011) 635.3 mg/100 g, dwb Leite et al. (2011) 634.75 mg/100g, dwb Leite-Legatti et al. (2012) Peel 269 mg/100 g, dwb Inada et al. (2015) 356.3 mg/100 g, dwb Plaza et al. (2016) 61.6 mg/100 g, dwb Pereira et al. (2017) Lima et al. (2011) Pulp ND Inada et al. (2015) Seed 11.8 mg/100 g, dwb Inada et al. (2015) Peel and seed 157 mg/100 g, dwb Inada et al. (2015) Peel tea 0.05 mg/100 mL Da Silva et al. (2017) Juice (steam extraction) 1.3 mg/100 mL Inada et al. (2018b) Non- fermented pomace extract 0.013 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 0.016 mg/100 g, dwb Whole fruit extracts 2.3-4.6 mg/100 mL Moura et al. (2018) (continued on next page) 40

(continued) Cyanidin-3-O-glucoside 433 mg/100 g, dwb Reynertson et al. (2008) 298 mg/100 g, dwb Wu et al. (2012) Whole fruit 123 mg/100 g, dwb Alezandro et al. (2013) 280 mg/100 g,dwb Inada et al. (2015) 1,964 mg/100 g, dwb Leite et al. (2011) 2,598 mg/100 g, dwb Lima et al. (2011) 1,964 mg/100 g, dwb Leite-Legatti et al. (2012) 1,737 mg/100 g, dwb Batista et al. (2013) Peel 1,261 mg/100 g, dwb Inada et al. (2015) 2,866 mg/100 g, dwb Plaza et al. (2016) 1,301 mg/100 g, dwb Batista et al. (2017) 272 mg/100 g, dwb Pereira et al. (2017) 7 mg/100 g, dwb Lima et al. (2011) Pulp 0.4 mg/100 g, dwb Inada et al. (2015) 2.01 mg/100 g, dwb Dantas et al. (2019) Seed 58 mg/100 g, dwb Inada et al. (2015) Peel and seed 707 mg/100 g, dwb Inada et al. (2015) Peel tea 0.9 mg/100 mL Da Silva et al. (2017) (continued on next page)

41

(continued) Cyanidin-3-O-glucoside Juice (clarified) 14.7 mg/100 mL Balisteiro et al. (2018) Juice (pulp) 0.079 mg/100 mL Inada et al. (2018a) Juice (steam extraction) 6.77 mg/100 mL Inada et al. (2018b) Non- fermented pomace extract 0.205 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 0.188 mg/100 g, dwb Whole fruit extracts 11.7-19.0 mg/100 mL Moura et al. (2018) Flavonols 4 mg/100 g, dwb Reynertson et al. (2008) Quercetin 0.8 – 1.3 mg/100 g, fwb Hoffmann-Ribani et al. (2009) 0.56 mg/100 g, fwb Abe et al. (2011) Whole fruit 115.7 mg/100 g, dwb Wu et al. (2012) 0.4 mg/100 g, dwb Inada et al. (2015) 5.2 mg/100 g, dwb Seraglio et al. (2018) 1.2 mg/100 g, dwb Betta et al. (2018) 4.8-55 mg/100 g, dwb Batista et al. (2013) Peel 3.5 mg/100 g, dwb Inada et al. (2015)

NQ Neves et al. (2018) Seed 0.4 mg/100 g, dwb Inada et al. (2015) Peel and seed 1.6 mg/100 g, dwb Inada et al. (2015) Peel tea 0.01 mg/100 mL Da Silva et al. (2017) (continued on next page)

42

(continued) Quercetin Juice (commercial) 3.3 mg/100 g, dwb Wu et al. (2012) Juice (steam extraction) 0.048 mg/100 mL Inada et al. (2018b) Quercetin-3-O-glucoside 42 mg/100 g, dwb Wu et al. (2012) Whole fruit 0.9 mg/100 g, dwb Seraglio et al. (2018) 4.3-11.2 mg/100 g, dwb Betta et al. (2018) Pulp ND Dantas et al. (2019) Juice (commercial) 11.4 mg/100 g, dwb Wu et al. (2012) Peel NQ Neves et al. (2018) Quercetin-7-O-glucoside Whole fruit 4.9 mg/100 g, dwb Wu et al. (2012) Juice (commercial) 3.5 mg/100 g, dwb Wu et al. (2012) Quercetin-3-O-rhamnoside 11 mg/100 g, dwb Reynertson et al. (2008) Whole fruit 38.8 mg/100 g, dwb Wu et al. (2012) Peel 60.7 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 11.1 mg/100 g, dwb Wu et al. (2012) Non- fermented pomace extract 17 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 17 mg/100 g, dwb Whole fruit extracts 1.4-1.9 mg/100 mL Moura et al. (2018) Quercetin-3-O-rutinoside 21 mg/100 g, dwb Reynertson et al. (2008) Whole fruit 77 mg/100 g, dwb Inada et al. (2015) (continued on next page)

43

(continued) Quercetin-3-O-rutinoside Peel 247 mg/100 g, dwb Inada et al. (2015) 7.1 mg/100 g, dwb Inada et al. (2015) Pulp ND Dantas et al. (2019) Seed 241 mg/100 g, dwb Inada et al. (2015) Peel and seed 195 mg/100 g, dwb Inada et al. (2015) Peel tea 0.34 mg/100 mL Da Silva et al. (2017) Juice (pulp) 0.175 mg/100 mL Inada et al. (2018a)

Juice (steam extraction) 2.81 mg/100 mL Inada et al. (2018b) Quercetin derivatives Whole fruit 2.1 mg/100 mL Alezandro et al. (2013) Juice (clarified) 0.53 mg/100 mL Balisteiro et al. (2018) Quercetin-3-O-galactoside Peel NQ Neves et al. (2018) Quercetin-3-O-rhamnoside Peel NQ Neves et al. (2018) Quercetin-pentoside Peel NQ Neves et al. (2018) Quercetin-galloyl-pentoside Peel NQ Neves et al. (2018) Quercetin-caffeoyl-hexoside Peel NQ Neves et al. (2018) Quercetin-p-coumaroyl -hexoside Peel NQ Neves et al. (2018) Quercetin-feruloyl -hexoside Peel NQ Neves et al. (2018) (continued on next page)

44

(continued) Myricetin 2 mg/100 g, dwb Reynertson et al. (2008) Whole fruit ND Hoffmann-Ribani et al. (2009) 0.8 mg/100 g, dwb Inada et al. (2015) Peel 5.1 mg/100 g, dwb Inada et al. (2015) Pulp 0.9 mg/100 g, dwb Inada et al. (2015) Seed 4.5 mg/100 g, dwb Inada et al. (2015) Peel and seed 3 mg/100 g, dwb Inada et al. (2015) Peel tea 0.006 mg/100 mL Da Silva et al. (2017) Juice (pulp) 0.008 mg/100 mL Inada et al. (2018a) Juice (steam extraction) 0.035 mg/100 mL Inada et al. (2018b) Myricetin-3-O-rhamnoside Whole fruit 18 mg/100 g, dwb Wu et al. (2012) 3.5 mg/100 g, dwb Inada et al. (2015) 20 mg/100 g, dwb Inada et al. (2015) Peel NQ Neves et al. (2018) Pulp 0.7 mg/100 g, dwb Inada et al. (2015) Seed 1.8 mg/100 g, dwb Inada et al. (2015) Peel and seed 20 mg/100 g, dwb Inada et al. (2015) (continued on next page)

45

(continued) Myricetin-3-O-rhamnoside Juice (commercial) 9.7 mg/100 g, dwb Wu et al. (2012) Juice (pulp) 0.028 mg/ 100 mL Inada et al. (2018a) Juice (steam extraction) 0.325 mg/100 mL Inada et al. (2018b) Whole fruit extracts 2.7-2.8 mg/100 mL Moura et al. (2018) Myricetin-3-O-glucoside Peel NQ Neves et al. (2018) Myricetin-3-O-galactoside Peel NQ Neves et al. (2018) Kaempferol ND Hoffmann-Ribani et al. (2009) Whole fruit ND Abe et al. (2011) 0.03 mg/100 g, dwb Seraglio et al. (2018) Kaempferol-3-O-glucoside Pulp 2.29 mg/100 mL, dwb Dantas et al. (2019) Peel tea 0.062 mg/100 mL Da Silva et al. (2017) Isorhamnetin 0.08 mg/100 g, dwb Seraglio et al. (2018) Whole fruit 0.16-0.18 mg/100 g, dwb Betta et al. (2018) Isorhamnetin-3-O-glucoside Peel tea 0.01 mg/100 mL Da Silva et al. (2017) Flavon-3-ols ND Abe et al. (2011) (+)-Catechin ND Wu et al. (2012) Whole fruit ND Seraglio et al. (2018) 0.81-1.39 mg/100 g, dwb Betta et al. (2018) Pulp 5.85 mg/100 g, dwb Dantas et al. (2019) (continued on next page)

46

(continued) (+)-Catechin Peel tea 0.013 mg/100 mL Da Silva et al. (2017) 10.5-117.5 mg/100 g, Peel extracts Machado et al. (2018) dwb ( - )-Epicatechin Abe et al. (2011) Whole fruit ND Seraglio et al. (2018) Peel tea 0.006 mg/100 mL Da Silva et al. (2017) Peel extracts 8.2-35.9 mg/100 g, dwb Machado et al. (2018) ( - )-Epicatechin gallate ND Seraglio et al. (2018) Whole fruit 0.09-0.16 mg/100 g, dwb Betta et al. (2018) Peel tea 0.008 mg/100 mL Da Silva et al. (2017) Peel extracts 0.0-84.5 mg/100 g, dwb Machado et al. (2018) Procyanidin A2 Pulp ND Dantas et al. (2019) Peel tea 0.012 mg/100 mL Da Silva et al. (2017) Procyanidin B1 Pulp 1.53 mg/100 g, dwb Dantas et al. (2019) Procyanidin B2 Pulp 4.63 mg/100 g, dwb Dantas et al. (2019) Flavanones Pulp 13.81 mg/100 g, dwb Dantas et al. (2019) Hesperitin-7-O-rutinoside Naringenin Seraglio et al. (2018) Whole fruit 0.04 mg/100 g, dwb Betta et al. (2018) Pulp ND Dantas et al. (2019) (continued on next page) 47

(continued) 3,5,7-trihydroxyflavanone 0.04 mg/100 g, dwb Seraglio et al. (2018) Whole fruit 0.05 mg/100 g, dwb Betta et al. (2018) Flavones Whole fruit 0.004 mg/100 g, dwb Seraglio et al. (2018) Luteolin Hydroxycinnamic acids Peel tea 0.015 mg/100 mL Da Silva et al. (2017) p-coumaric acid Peel extracts 0.7-8.3 mg/100 g, dwb Machado et al. (2018) Whole fruit 0.36 mg/100 g, dwb Seraglio et al. (2018) o-coumaric acid Peel extracts 0.0-1.5 mg/100 g, dwb Machado et al. (2018) m-coumaric acid Whole fruit 0.3 mg/100 g Inada et al. (2015) Peel 0.2 mg/100 g Inada et al. (2015) Pulp 0.6 mg/100 g Inada et al. (2015) Seed 0.6 mg/100 g Inada et al. (2015) Peel and seed 0.4 mg/100 g Inada et al. (2015) Juice (pulp) 0.053 mg/100 mL Inada et al. (2018a)

Juice (steam extraction) 0.089 mg/100 mL Inada et al. (2018b) Trans-cinnamic acid Whole fruit 1.4 mg/100 g Inada et al. (2015) Peel 0.5 mg/100 g Inada et al. (2015) Pulp 0.4 mg/100 g Inada et al. (2015) (continued on next page)

48

(continued) Trans-cinnamic acid Peel and seed 0.6 mg/100 g Inada et al. (2015) Juice (pulp) 0.1 mg/100 mL Inada et al. (2018a) Ferulic acid Peel extracts 0.0-0.9 mg/100 g, dwb Machado et al. (2018) Whole fruit 0.2 mg/100 g, dwb Seraglio et al. (2018) Caffeic acid Whole fruit 0.03 mg/100 g, dwb Seraglio et al. (2018) Chlorogenic acid 0.19 mg/100 g, dwb Seraglio et al. (2018) Whole fruit 1.3-2.1 mg/100 g, dwb Betta et al. (2018) Hydroxybenzoic acids Peel tea 0.015 mg/100 mL Da Silva et al. (2017) Benzoic acid 3,4-dihydroxybenzoic acid 4.4 mg/100 g Inada et al. (2015) Whole fruit 1.3 mg/100 g Seraglio et al. (2018) 1.2-1.4 mg/100 g Betta et al. (2018) Peel 16 mg/100 g Inada et al. (2015) Peel and seed 8.4 mg/100 g Inada et al. (2015) Syringic acid Peel extracts 0.0-1.9 mg/100 g, dwb Machado et al. (2018) Whole fruit 0.1 mg/100 g, dwb Seraglio et al. (2018) Gallic acid 50.7 mg/100 g, dwb Wu et al. (2012) 205.9 mg/100 g, dwb Inada et al. (2015) Whole fruit 4.2 mg/100 g, dwb Seraglio et al. (2018) 2.4-2.6 mg/100 g, dwb Betta et al. (2018) (continued on next page) 49

(continued) Gallic acid 4.0-49.9 mg/100/g, dwb Batista et al. (2013) Peel 153.2 mg/100 g, dwb Inada et al. (2015) 4.9 mg/100 g, dwb Pereira et al. (2017) 4.5 mg/100 g, dwb Inada et al. (2015) Pulp 3.5 mg/100 g, dwb Pereira et al. (2017) 8.04 mg/100 g, dwb Dantas et al. (2019) 422 mg/100 g, dwb Inada et al. (2015) Seed 23.6 mg/100 g, dwb Pereira et al. (2017) Peel and seed 307 mg/100 g, dwb Inada et al. (2015) Peel tea 0.43 mg/100 mL Da Silva et al. (2017) Juice (commercial) 4.1 mg/100 g, dwb Wu et al. (2012) Juice (pulp) 0.33 mg/100 mL Inada et al. (2018a) Juice (steam extraction) 0.13 mg/100 mL Inada et al. (2018b) Peel extracts 1.8-2.7 mg/100 g, dwb Machado et al. (2018) Pentagalloyl-hexose Peel 32.8 mg/100 g, dwb Plaza et al. (2016) ND – not detected. NQ – not quantified.

50

1.2.2 Ellagitannins

Tannins are phenolic compounds with high molecular weight (of up to 20,000 Da), which present the ability to form strong complexes with carbohydrates and proteins (Montes- Ávila et al., 2017; Serrano et al., 2009). These compounds are highly hydroxylated and exhibit a wide variety of chemical structures (Montes-Ávila et al., 2017). They can be classified according to their monomeric units into three main groups: hydrolysable tannins, proanthocyanidins and condensed tannins (Montes-Ávila et al., 2017; Serrano et al., 2009). Hydrolysable tannins are composed by esters of organic acids with a sugar moiety and comprise both ellagitannins and gallotannins. If the acid component is gallic acid, the compound is termed galotannin. On the other hand, if the organic acid is hexahydroxydiphenic acid (HHDP), the compound is called ellagitannin (Montes-Ávila et al., 2017; Serrano et al., 2009). However, most ellagitannins are a mixture of esters containing both HHDP and gallic acid (Figure 8) (Serrano et al., 2009). Besides, the sugar component is usually glucose, although fructose, xylose and sucrose can also be found (Montes-Ávila et al., 2017; Serrano et al., 2009).

Figure 8. Tellimagrandin II, a which contains both HHDP and gallic acids (Plaza et al., 2016)

51

The term "hydrolysable tannins" is due to the fact that they can be easily hydrolyzed in the presence of acids, bases or tannases enzymes in their monomeric forms (Montes-Ávila et al., 2017). Under these conditions, the ester bonds are hydrolyzed, releasing gallic acid and/or HHDP. HHDP is a dimeric compound composed by two units of gallic acid linked to each other through its aromatic carbon atoms (Khanbabaee & Ree, 2001). When HHDP is released from ellagitannins, it undergoes spontaneous rearrangement in ellagic acid (Figure 9) (Clifford & Scalbert, 2000; Okuda, 1995). However, it is worth mentioning that some ellagitannins, such as vescalagin and , are not completely hydrolysable. These compounds are C- glycoside, which present a C-C coupling the galloyl or HHDP residue to the sugar, and for this reason, they can only be partially hydrolyzed (Khanbabaee & Ree, 2001, Okuda, 1995). Thus, the hydrolysis of vescalagin and castalagin yields one molecule of ellagic acid and vescalin (Figure 10) or , respectively (Yamada et al., 2018). Nevertheless, for historical reasons these compounds were also classified as hydrolysable tannins (Khanbabaee & Ree, 2001).

Figure 9. HHDP (hexahydroxydiphenic acid) undergoes spontaneous rearrangement in ellagic acid (Adapted from Quideau & Feldman, 1995)

52

Figure 10. Hydrolysis of vescalagin, an ellagitannin C-glycoside, yields ellagic acid and vescalin (source: www.foodb.ca).

While gallotannins present limited distribution in nature, elagitannins are found in several plant families, with more than 1,000 compounds identified (Montes-Ávila et al., 2017). The large structural variability of ellagitannins is due to the several possibilities of linkage among gallic acid and HHDP residues with the sugar moiety (Landete, 2011). In addition, these residues can also bind to each other by their aromatic carbons and/or phenolic oxygen atoms (Khanbabaee & Ree, 2001). Nevertheless, despite the large number of ellagitannins already found in nature (Montes- Ávila et al., 2017), the correct identification and quantification of these compounds is limited. This is due to the great complexity and structural variability of ellagitannins, their sensitivity to hydrolysis during extraction, and the lack of commercial standards (García-Villalba et al., 2015; Serrano et al., 2009). In addition, part of the ellagitannins present in the food matrix is insoluble and is found covalently bound to cell wall and/or macromolecules, which may underestimate their contents (García-Villalba et al., 2015; Arranz et al., 2009). For these reasons, several studies have subjected food samples to acid hydrolysis in order to quantify the free ellagic acid released from soluble and insoluble ellagitannins (García-Villalba et al., 2015; Alezandro et al., 2013; Abe et al., 2011; Pinto et al., 2008, Hakkinen et al., 2000). Quantification of ellagic acid and other derivative compounds released after hydrolysis can provide important information on 53

ellagitannins contents and chemical structure of the analyzed samples (García-Villalba et al., 2015). Ellagitannins and ellagic acid have been reported in teas (Tomás-Barberán & Yang, 2018), nuts (Pelvan et al., 2018; Grace et al., 2016; Slatnar et al., 2015), (García- Villalba et al., 2015; Fischer et al., 2011) and berries (e.g. strawberry, red raspberry, cloudberry, blackberry) (Abe et al., 2011; Pinto et al., 2008, Hakkinen et al., 2000). Among berries, several studies have characterized the ellagitannins, ellagic acid and derivatives compounds present in the different fractions of jabuticaba, and its derived products and extracts (Table 2). Free ellagic acid and its glycosylated forms were detected in jabuticaba (Table 2). Some studies have quantified the soluble ellagic acid of different jabuticaba fractions (4 to 422 mg/100 g, on dry weight basis (dwb)), while other studies performed an acid and/or alkaline hydrolysis in order to depolymerize the ellagitannins and quantify the compounds present in the insoluble fraction of jabuticaba (5.3 a 9,566 mg/100 g, dwb) (Table 2). The high contents of ellagic acid found after hydrolysis in some studies, indicate the presence of high amounts of ellagitannins in jabuticaba (Alezandro et al., 2013; Abe et al., 2011). In fact, other studies reported high contents of ellagitannins in this fruit. Vescalagin and castalagin were the major elagitannins detected in jabuticaba, with the highest concentrations found in the seeds (953 to 2,817 mg/100 g on dwb) (Table 2). In addition, the ellagitannins , casuariin, , , , and II, , and some isomers of strictinin, pedunculagin and sanguiin H-6 and H-10 were also detected (Table 2). Besides, a new ellagitannin named as cauliflorin was recently described in the jabuticaba peel and pulp (Pereira et al., 2017). The chemical structures of ellagitannins reported in jabuticaba are shown in Figure 11.

54

Table 2. Ellagitannins, ellagic acid and its derivatives detected in jabuticaba fractions and its derived products and extracts by different studies. Compounds Fruit fraction, derived product or extract Contents References Soluble EA Juice (commercial) 15.6 mg/100 g, dwb Wu et al. (2012) (without extraction) Juice (pulp) 0.14 mg/100 mL Inada et al. (2018a) Juice (steam extraction) 3.1 - 14.3mg/100 mL Inada et al. (2018b) Soluble EA 52 mg/100 g, dwb Reynertson et al. (2008) (organic solvent extraction)a 15.4 mg/100 g, dwb Wu et al. (2012) Whole fruit 15.4 - 26 mg/100 g, dwb Alezandro et al. (2013) 34 mg/100 g, dwb Inada et al. (2015) NQ Neves et al. (2018) 19.5 mg/100 g, dwb Alezandro et al. (2013) 348 mg/100 g, dwb Batista et al. (2013) Peel 178 mg/100 g, dwb Inada et al. (2015) 142.8 mg/100 g, dwb Plaza et al. (2016) 17.3 - 20 mg/100 g, dwb Pereira et al. (2017) 5.3 mg/100 g, dwb Inada et al. (2015) Pulp 8.3 - 10 mg/100 g, dwb Pereira et al. (2017) 83 mg/100 g, dwb Inada et al. (2015) Seed 55.9 - 67 mg/100 g, dwb Pereira et al. (2017) Peel and seed 99 mg/100 g, dwb Inada et al. (2015) (continued on next page)

55

(continued) Soluble EA Non- fermented pomace extract 12 mg/100 g, dwb Morales et al. (2016) (organic solvent extraction)a Fermented pomace extract 83 mg/100 g, dwb Soluble EA 6 mg/100 g, fwb Abe et al. (2011) Whole fruit (SPE extraction)b 19 - 40 mg/100g, dwb Alezandro et al. (2013) Peel 10.8 - 24 mg/100 g, dwb Abe et al. (2011) Pulp 2.2 - 35.8 mg/100 g, dwb Abe et al. (2011) Seed 118 - 422 mg/100 g, dwb Abe et al. (2011) Juice (clarified) 8.21 mg/100 mL Balisteiro et al. (2018) Low tannin whole fruit extract 19.6 mg/100 mL Moura et al. (2018) High tannin whole fruit extract 42.8 mg/100 mL Total EA 311 mg/100 g, fwb Abe et al. (2011) Whole fruit (after acid hydrolysis)c 5,050 - 9,566 mg/100 g, dwb Alezandro et al. (2013) 2,250 - 4,395 mg/100 g, dwb Abe et al. (2011) Peel 1,546 mg/100 g, dwb Batista et al. (2013) Pulp 460 - 3,680 mg/100 g, dwb Abe et al. (2011) Seed 3,761 - 9173 mg/100 g, dwb Abe et al. (2011) Juice (clarified) 16.15 mg/100 mL Balisteiro et al. (2018) Low tannin whole fruit extract 22.7 mg/100 mL Moura et al. (2018) High tannin whole fruit extract 329.6 mg/100 mL (continued on next page)

56

(continued) Total EA (after alkaline and acid Whole fruit 193 mg/100 g, dwb Inada et al. (2015) hydrolyses)d Peel 276 mg/100 g, dwb Inada et al. (2015) Pulp 5.3 mg/100 g, dwb Inada et al. (2015) Seed 244 mg/100 g, dwb Inada et al. (2015) Peel and seed 258 mg/100 g, dwb Inada et al. (2015) 28.5 mg/100 g, dwb Wu et al. (2012) EA pentoside Whole fruit NQ Neves et al. (2018) Peel 55.1 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 8.1 mg/100 g, dwb Wu et al. (2012) Non- fermented pomace extract 25 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 4 mg/100 g, dwb EA hexoside Whole fruit NQ Neves et al. (2018) Non- fermented pomace extract 83 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 6 mg/100 g, dwb EA rhamnoside isomers Whole fruit NQ Neves et al. (2018) EA-acetyl-rhamnoside isomers Whole fruit NQ Neves et al. (2018) EA-galloyl-pentoside Whole fruit NQ Neves et al. (2018) EA-valeryl- rhamnoside isomers Whole fruit NQ Neves et al. (2018) (continued on next page) 57

(continued) EA-caprylyl- rhamnoside Whole fruit NQ Neves et al. (2018) Methyl EA pentoside Whole fruit NQ Neves et al. (2018) Methyl EA hexoside Whole fruit NQ Neves et al. (2018) Methyl EA rhamnoside Whole fruit NQ Neves et al. (2018) Methyl EA-acetyl-rhamnoside Whole fruit NQ Neves et al. (2018) isomers Methyl EA-valeryl- rhamnoside Whole fruit NQ Neves et al. (2018) isomers Methyl EA-caprylyl- rhamnoside Whole fruit NQ Neves et al. (2018) dilactone Whole fruit 20.1 mg/100 g, dwb Wu et al. (2012) Juice (commercial) 20.5 mg/100 g, dwb Wu et al. (2012) Non- fermented pomace extract 0.0 - 41 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 11 - 13 mg/100 g, dwb HHDP-galloyl-glucose Whole fruit extracts NQ Moura et al. (2018) (Strictinin) HHDP-galloyl-glucose Whole fruit 8.8 mg/100 g, dwb Wu et al. (2012) (not specified) Peel 77.g mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 5.2 mg/100 g, dwb Wu et al. (2012) (continued on next page)

58

(continued) HHDP-galloyl-glucose Non- fermented pomace extract 70 - 86 mg/100 g, dwb Morales et al. (2016) (not specified) Fermented pomace extract 11 - 17 mg/100 g, dwb Vescalagin Whole fruit 28.7 mg/100 g, fwb Silva et al. (2016) Peel 4.1 - 9 mg/100 g, dwb Pereira et al. (2017) Pulp 18.7 - 103 mg/100 g, dwb Pereira et al. (2017) Seed 953 - 2,817 mg/100 g, dwb Pereira et al. (2017) Whole fruit extracts NQ Moura et al. (2018) Castalagin Whole fruit 78.4 mg/100 g, fwb Silva et al. (2016) Peel ND Pereira et al. (2017) Pulp 52 - 151 mg/100 g, dwb Pereira et al. (2017) Seed 1,757 - 2,553 mg/100 g, dwb Pereira et al. (2017) Whole fruit extracts NQ Moura et al. (2018) Bis-HHDP-glucose (Casuariin) Whole fruit 5.9 mg/100 g, dwb Wu et al. (2012) Peel 75.1 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 4.1 mg/100 g, dwb Wu et al. (2012) Whole fruit extracts NQ Moura et al. (2018) Bis-HHDP-glucose 11.3 mg/100 g, dwb Wu et al. (2012) Whole fruit (Pedunculagin) 9.8 mg/100 g, fwb Silva et al. (2016) 111.3 mg/100 g, dwb Plaza et al. (2016) Peel 21 - 335 mg/100 g, dwb Pereira et al. (2017) (continued on next page) 59

(continued) Bis-HHDP-glucose Pulp 24 - 753 mg/100 g, dwb Pereira et al. (2017) (Pedunculagin) Seed 117- 452 mg/100 g, dwb Pereira et al. (2017) Juice (commercial) 3.9 mg/100 g, dwb Wu et al. (2012) Whole fruit extracts NQ Moura et al. (2018) Bis-HHDP-glucose (not specified) Non- fermented pomace extract 35 mg/100 g, dwb Morales et al. (2016) Fermented pomace extract 13 mg/100 g, dwb Cauliflorin Peel 15 - 126 mg/100 g, dwb Pereira et al. (2017) Pulp 16 - 40 mg/100 g, dwb Pereira et al. (2017) Seed ND Pereira et al. (2017) Galloyl-bis-HHDP-glucose Whole fruit 34.3 mg/100 g, dwb Wu et al. (2012) (Casuarinin) Peel 185.5 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 24.9 mg/100 g, dwb Wu et al. (2012) Whole fruit extracts NQ Moura et al. (2018) Galloyl-bis-HHDP-glucose Whole fruit 34.4 mg/100 g, dwb Wu et al. (2012) (Casuarictin) Peel 212.6 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 70.8 mg/100 g, dwb Wu et al. (2012) Whole fruit extracts NQ Moura et al. (2018) (continued on next page) 60

(continued) HHDP-digalloylglucose Whole fruit 21.7 mg/100 g, dwb Wu et al. (2012) (Tellimagrandin I) Peel 51.3 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 7.3 mg/100 g, dwb Wu et al. (2012) Whole fruit extracts NQ Moura et al. (2018) HHDP-trigalloylglucose Whole fruit 30.5 mg/100 g, dwb Wu et al. (2012) (Tellimagrandin II) Peel 74.8 mg/100 g, dwb Plaza et al. (2016) Juice (commercial) 9.7 mg/100 g, dwb Wu et al. (2012) Sanguiin H-10 isomers Whole fruit 9.3 - 96 mg/100 g, dwb Alezandro et al. (2013) Peel 19.5 - 35 mg/100 g, dwb Alezandro et al. (2013) Pulp 1.6 - 5.4 mg/100 g, dwb Alezandro et al. (2013) Seed 12.6 - 38.6 mg/100 g, dwb Alezandro et al. (2013) Sanguiin H-6 isomers Whole fruit 1 - 84.8mg/100 g, dwb Alezandro et al. (2013) Peel 1.9 - 8.9mg/100 g, dwb Alezandro et al. (2013) Pulp 1.8 - 6.6mg/100 g, dwb Alezandro et al. (2013) Seed 3 - 175 mg/100 g, dwb Alezandro et al. (2013) Lambertianin C Whole fruit 6.9 - 24.7 mg/100 g, dwb Alezandro et al. (2013) Peel 10.4 mg/100 g, dwb Alezandro et al. (2013) (continued on next page)

61

(continued) Lambertianin C Pulp 1.8 mg/100 g, dwb Alezandro et al. (2013) Seed 2.8 mg/100 g, dwb Alezandro et al. (2013) EA – ellagic acid. NQ – not quantified. DWB – dry weight basis. FWB – fresh weight basis. aExtraction was performed with organic solvents. bExtraction was performed with methanol:water (70:30, v/v) and then, the obtained extract was passed through solid phase extraction (SPE) columns in order to retain ellagitannins. In this case, the authors considered the sum of ellagic acid and its glycosides. cThe samples were submitted to acid hydrolysis in order to depolymerize ellagitannins in ellagic acid. dThe samples were extracted with organic solvent and then, the extract residue was subsequent submitted to alkaline and acid hydrolyses in order to quantify the insoluble phenolic compounds.

62

Figure 11. Chemical structures of ellagitannins reported in jabuticaba by several studies (source: www.foodb.ca). 63

Figure 11. Chemical structures of ellagitannins reported in jabuticaba by several studies (continued) (source: www.foodb.ca). The chemical structure of cauliflorin was drawn in the software Chemsketch 2018.1.1 (ACD, Toronto, Canada). 64

2. Effects of technological processes on phenolic compound profile

2.1 High hydrostatic pressure

High hydrostatic pressure (HHP) is a non-thermal conservation method, which destroy microorganisms and inactivate enzymes, while preserving the food sensorial and nutritional characteristics (Inada et al., 2018a; Silva & Sulaiman et al., 2018; Liu et al., 2016; Barba et al., 2013; Varela-Santos et al., 2012; Yaldagard et al 2008; Laboissière et al., 2007). A HHP system is composed of a vessel containing the pressure transmitting liquid, a pressurizing chamber with a closing piston and a pressure generating system (pump) (Figure 12) (Hogan et al., 2005; Huppertz et al., 2002). The HHP processing consists of subjecting foods to high pressure levels ranging from 50 to 1,000 MPa (Hogan et al., 2005). Packaged food is inserted into the chamber, which is closed and pressurized. When the desired pressure is reached, the pump or piston is stopped, the valves are closed, and the chamber is maintained pressurized. After the end of the pressurizing time, depressurizing is performed quickly (Hogan et al., 2005; Huppertz et al., 2002).

Figure 12. Schematic diagram of basic equipment design for high hydrostatic pressure processing of foods (Adapted from Huppertz et al., 2002).

As previously mentioned, several studies have shown that HHP destroy microorganisms (Inada et al., 2018a; Błaszczak et al., 2017; González-Cebrino et al., 2013; Buzrul et al., 2008; Lavinas et al., 2008). The cell damage caused by HHP varies according to the applied pressure and time levels. The cell membrane is the main target for HHP, which causes changes in the 65

membrane permeability and consequently loss of cell integrity, disruption and leaking (Figure 13) (Huang et al., 2014; Yang et al., 2012; Rendueles et al., 2011; Bravim et al., 2010).

Figure 13. Cell structures of E. coli and S. aureus untreated and high hydrostatic pressure treated at 500 MPa for 30 min. Untreated E. coli (a); high hydrostatic pressure treated E. coli (a’); Untreated S. aureus (b); high hydrostatic pressure treated S. aureus (b’). The microstructure was evaluated by transmission electron microscopy (Yang et al., 2012).

As described for microorganisms, some studies have also observed that HHP can affect the microstructure of plant foods (Figure 14). Changes in foods microstructure include the disturbance of cell walls, membranes and organelles integrity, which causes the release of phenolic compounds from the intracellular to the intercellular space, increasing the extractability of phenolic compounds (Xi & Luo, 2016; Vásquez-Gutiérrez et al., 2013; Vásquez-Gutiérrez et al., 2011). Therefore, in addition to its application as a conservation technique, HHP has been used to extract bioactive compounds in some food matrices, being even denominated as “high hydrostatic pressure extraction” (Pinela et al., 2018; Corrales et al., 2009; Jun et al., 2009; Prasad et al., 2009), being even more effective than other conventional extraction methods (Briones-Labarca et al., 2015; Jun et al., 2009). The results of some studies that investigated the effect of HHP on the microstructure of plant foods, extraction of phenolic compounds and their contents are summarized in Table 3. 66

Figure 14. Onion treated for 3 minutes with 100 MPa (A and D), 300 MPa (B and E) and 600 MPa (C and F). is: intercellular space; dc: distorted cell; t: tonoplast; fis: flooded intercellular space; bcw: broken cell wall. The microstructure was evaluated by cryo scanning electron microscopy (Vásquez-Gutiérrez et al., 2013).

In general, some studies observed that the effect of HHP is both condition and compound specific, varying not only according to the HHP conditions applied, but also with the individual phenolic compounds analyzed (Table 3) (Inada et al., 2018a; Pinela et al., 2018; Corrales et al., 2009). However, although several studies have shown an increase in phenolic compounds contents after pressurization of different food matrices (Fernández-Jalao et al., 2018; Torres-Ossandón et al., 2018; Briones-Labarca et al., 2015; Vásquez-Gutiérrez et al., 2013; Corrales et al., 2009), some studies have observed that HHP maintained or decreased the contents of these compounds (Table 3) (Saikaew et al., 2019; Yuan et al., 2018; Jez et al., 2018; Blaszczak et al., 2017; Cao et al., 2011). Changes in the foods microstructure with consequent increase in the polyphenols extractability is important, because it may increase the bioavailability of these compounds. In fact, Rodríguez-Roque et al. (2015) observed that HHP increased the bioaccessibility of phenolic compounds from fruit juices and related beverages, suggesting that pressurization may increase not only the extractability of bioactive compounds, but also their intestinal absorption. For this reason, it is relevant to investigate the effect of HHP on the phenolic compounds contents of plant foods. 67

Table 3. Effect of high hydrostatic pressure (HHP) on food microstructure and phenolic compounds extraction and contents evaluated by different studies. Food matrix HHP condition HHP effects References Food microstructure - Both HHP conditions significantly affected the integrity 200 MPa for 1 min Persimmon fruit of cell walls and membranes, releasing tannins into the Vásquez-Gutiérrez et al. (2011) 400 MPa for 6 min intercellular space. Greater cell damage and tannins release were observed at 400 MPa for 6 min. Food microstructure and polyphenols contents 100 MPa for 3 min - This condition did not cause serious external damage in the cells and did not affect total phenolic contents. - Compression and deformation of cells was noticeable. 300 MPa for 3 min - Increased total phenols in 1.6-fold. Onions Vásquez-Gutiérrez et al. (2013) - Increasing HHP time lead to progressive distortion of cell walls and release of cellular contents to intercellular 600 MPa for 1, 2 and spaces. 600 MPa for 3 min lead to greater loss of cell 3 min integrity, increased deformation and fragmented cell walls. - Increased total phenols of up to 1.7-fold. Food microstructure and polyphenols extraction Radix 100, 200, 300, 400, - The surface was destroyed and the texture was crumbled. Angelica sinensis 500 MPa for 10 min Xi & Luo (2016) Severe damage on cell walls and organelles was observed. (continued on next page)

68

(continued) Food matrix HHP condition HHP effects References Food microstructure and polyphenols extraction Radix 100, 200, 300, 400, - HHP increased ferulic acid extraction yield and the Xi & Luo (2016) Angelica sinensis 500 MPa for 10 min efficiency of extraction. Polyphenols extraction - A selective anthocyanin extraction (monoglucosides and acylglucosides) was achieved by varying pressure intensity. All HHP conditions increased anthocyanins 200, 400, 600 MPa monoglucosides contents from 47% (400 and 600 MPa) for 30 min Grape skin to 130% (200 MPa). On the other hand, only the Corrales et al. (2009) pressurization at 600 MPa increased acylated glucoside anthocyanins (54%). The other conditions did not affect these latter compounds contents. 600 MPa for 30, 60 - Holding time did not influence anthocyanins extraction and 90 min yield. - When the pressure increased from 100 to 600 MPa, the 100, 200, 300, 400, extraction yields of total polyphenols increased from 15% Green tea leaves 500, 600 MPa for 1, 4, Jun et al. (2009) to 30%. On the other hand, HHP holding time did not 7 and 10 min affect total polyphenols contents. (continued on next page)

69

(continued) Food matrix HHP condition HHP effects References - In comparison to conventional extraction methods 100, 200, 300, 400, (ultrasound, heat reflux and extraction at room Green tea leaves 500, 600 MPa for 1, 4, Jun et al. (2009) temperature), HHP provided higher extraction yields, 7 and 10 min requiring shorter time. - Increasing HHP pressure levels (200 to 500 MPa), at 30 200, 300, 400, 500 min, increased contents from 5.9 to 9.6-fold. Longan fruit pericarp MPa for 2.5, 5, 15 and Prasad et al. (2009) - Increasing HHP holding time (2.5 to 30 min) did not 30 min affect corilagin contents. - Increasing HHP times, from 5 to 15 min, led to higher total polyphenols contents (of up 112%) and total 500 MPa for 5,10,15 flavonoid contents (of up to 277%). Chilean papaya seeds Briones-Labarca et al. (2015) min - HHP extraction proved to be more effective than traditional extraction methods (ultrasound and conventional extraction). (continued on next page)

70

(continued) Food matrix HHP condition HHP effects References - In general, the recovery of phenolic compounds was maximized when high pressures and short extraction times were applied. 0.1 to 600 MPa for 1.5 Watercress - HHP selective extracted individual phenolic Pinela et al. (2018) to 33.5 min compounds. However, the optimal HHP conditions for the extraction of total phenolic compounds were 600 MPa for 3.1 min. Polyphenols contents 400, 500, 600 MPa for - Pressurization at 400 MPa for 5, 10 and 20 min Strawberry pulp 5, 10, 15, 20 and decreased total phenolic compounds contents analyzed Cao et al. (2011) 25 min by HPLC of up to 10%, while the other conditions maintained these compounds contents. - All HHP conditions decreased total phenolic 200, 400, 600 MPa for compounds contents. On the other hand, in general, HHP Aronia juice Blaszczak et al. (2017) 15 min maintained or increased the individual phenolic compounds analyzed by HPLC. - All HHP conditions lead to a substantial loss in 450, 550, 650 MPa for Tomato purée individual decreased phenolic compounds contents Jez et al. (2018) 5, 10 and 15 min analyzed by HPLC. (continued on next page) 71

(continued) Food matrix HHP condition HHP effects References 200, 400, 600 MPa for - HHP did not affect total phenolic compounds and total Aronia berry purée Yuan et al. (2018) 2.5 and 5 min anthocyanins contents. Combined methods: - Supercritical fluid extraction and pressurization at 400 HHP (300, 400, 500 MPa and 500 MPa for 3 and 5 min increased total Cape gooseberry pulp MPa for 1, 3 and 5 polyphenols contents of up to 23%, while other Torres-Ossandón et al. (2018) min) and supercritical conditions decreased these compounds contents of up to fluid extraction 33%. - HHP affected phenolic compounds depending on the apple geographical origin. In Spanish apples, HHP at 400 400, 500, 600 MPa for MPa increased total flavonols (30%), but maintained total Apples Fernández-Jalao et al. (2018) 5 min phenolic compounds determined by HPLC. On the other hand, in Italian apples, pressurization at 600 MPa increased total polyphenols (54%). - The HHP effects on phenolic composition of jabuticaba 200, 500 MPa for 5 juice was both compound and condition specific. Jabuticaba juice and 10 min and 350 However, considering the total phenolic compounds Inada et al. (2018a) MPa for 7.5 min contents by HPLC, pressurization led to an average increase of 35% of total phenolic compounds contents. 250 to 700 MPa for30 - HHP at all conditions decreased total phenolic Purple waxy corn Saikaew et al. (2019) and 45 min compounds, flavonoids and anthocyanins contents. 72

2.2 Dehydration methods: oven-drying and freeze-drying

Dehydration is one of the earliest conservation methods, which reduces water activity of food and consequently decreases microbial growth, as well as the occurrence of chemical reactions, extending shelf life (Lewick, 2006). Besides, by removing most of the water present in foods, it leads to the production of a dry powder highly concentrated on bioactive compounds, such as phenolic compounds (Samoticha et al., 2016; De Torres et al., 2010; Wojdyło et al., 2009). Among the existing dehydration methods, oven-drying (also known as convective drying or hot air drying) is one of the most popular due to its high drying efficiency and low cost (Lewick, 2006; Ratti, 2001). In oven-drying, the food is subjected to a hot air flow with low moisture content, which transfers the heat mainly through convection. As a result, the food moisture is removed by transferring from the solid material to the drying medium (air) (Barta, 2006). However, the long exposure of food to high temperatures and oxygen can lead to oxidation of phenolic compounds, as well as color change (Zhou et al., 2017; Samoticha et al., 2016; Michalska et al., 2016; Wojdylo et al 2014; Wojdylo et al., 2009). On the other hand, freeze-drying (also known as lyophilization or sublimation drying) consists of removing food water by sublimation (primary drying) and desorption (secondary drying) (Kharaghani et al., 2017; Stapley, 2008). Since this method is performed at low temperature and pressure, it promotes less degradation of bioactive compounds and color (Samoticha et al., 2016; Michalska et al., 2016; Mphahlele et al 2016; Wojdylo et al 2014; De Torres et al., 2010). However, although freeze-drying usually leads to high quality products, this method presents low efficiency of drying, as well as high-energy consumption and cost (Ratti, 2001). The results of some studies that investigated the effect of oven-drying and freeze- drying on phenolic compounds contents are summarized in Table 4.

73

Table 4. Effect of oven-drying (OD) and freeze-drying (FD) on phenolic compounds contents evaluated by different studies. Food matrix Conditions Observed effects References - FD did not affect the total phenolic compounds contents - Fresh in comparison to fresh sample, while OD increased these - OD (80 °C for 2 h + compounds of up to 29%. Tomatoes varieties Chang et al. (2006) 60 °C for 6 h) - In addition, FD and OD increase total flavonoids contents - FD of up to 72% and 89%, respectively, in comparison to fresh sample. - OD (70 °C for 54 h) - Total phenolic compounds contents of OD sample was Pumpkin Que et al. (2008) - FD 4.6 times higher than FD sample. - Fresh - In comparison to fresh sample, FD did not affect non- - OD (70 °C for 550 anthocyanins phenolic compounds and anthocyanins Strawberry Wojdyło et al. (2009) min) analyzed by HPLC, while OD led to a decrease in these - FD compounds contents of 36% and 68%, respectively. - In comparison to fresh sample, FD promoted lower losses - Fresh of total anthocyanins and flavonols analyzed by HPLC Grape skin - OD (60 °C for 24 h) De Torres et al. (2010) (18% and 27%, respectively) than OD (37% and 34%, - FD respectively). - Fresh - FD did not affect the total phenolic compounds contents Muscadine pomace - OD (70 °C and 80 °C) in comparison to fresh sample, while OD decreased the Vashisth et al. (2011) - FD contents of up to 18%. (continued on next page) 74

(continued) Food matrix Conditions Observed effects References - In comparison to fresh sample, drying processes caused significant reduction of total anthocyanins > total phenolics > proanthocyanidins. OD at 80 °C led to highest degradation of these compounds (100%, 64% and 47%, - Fresh respectively), while FD led to lowest losses (54%, 42%, Camu-camu - OD (50 °C and 80 °C) 16%, respectively). De Azevêdo et al. (2014) - FD - Drying processes also resulted in decrease of individual phenolic compounds analyzed by HPLC in comparison to fresh samples. OD at 50 °C led to higher degradation of these compounds, while no differences were observed between OD at 80 °C and FD. - OD sample presented higher contents of individual (hydroxybenzoic and hydroxycinnamic acids) and total - OD (65 °C for 24 h) phenolic compounds analyzed by HPLC (3.8 mg/kg) than Pumpkin Aydin & Gocmen (2015) - FD FD sample (2.2 mg/kg). - OD sample presented a higher phenolic compounds bioaccessibility in comparison to FD. (continued on next page)

75

(continued) Food matrix Conditions Observed effects References - FD samples presented higher contents of total phenolic compounds analyzed by HPLC (1,352 mg/100 g) than OD samples (1,142 mg/100 g). Among these compounds, FD - OD (60 °C and 70 presented higher contents of anthocyanins and (+)- Plum °C) Michalska et al. (2016) catechin. - FD - On the other hand, no differences were found between OD and FD samples regarding flavonols, phenolic acids and polymeric proanthocyanidins determined by HPLC. - In comparison to FD, OD led to a decrease in total phenolic compounds, tannins and flavonoids contents. - Considering individual compounds analyzed by HPLC, - OD (40 °C, 50 °C OD also led to a reduction of rutin, (+)-catechin, Pomegranate peel and 60 °C) Mphahlele et al. (2016) epicatechin and hesperidin contents, in comparison to FD. - FD However, OD at 60 °C increased contents (25%), while p-coumaric acid was only detected in samples after OD at 50 °C and 60°C. (continued on next page)

76

(continued) Food matrix Conditions Observed effects References - In comparison to FD, OD at 50 °C, 60 °C and 70 °C progressively decrease total flavan-3-ols contents - OD (50 °C, 60 °C and determined by HPLC (23% to 55%). On the other hand, Jujube fruit cultivars 70 °C) Wojdyło et al. (2016) OD at 50 °C led to an increase of 13% in flavonols - FD contents, while OD at 60 °C and 70 °C decreased these compounds of up to 31%. - In comparison to fresh sample, FD promoted higher losses of soluble phenolic compounds analyzed by HPLC (58%) than OD (37%). On the other hand, OD led to a higher loss of insoluble phenolics (48%) than FD (29%). - Fresh Therefore, equivalent losses of total phenolics (42%) were Guava - OD (55 °C for 22 h) Nunes et al. (2016) observed. - FD - Both processes modified the distribution of soluble and insoluble phenolics. The authors suggested that OD promoted a release of insoluble phenolics of food matrix, mainly flavonoids. (continued on next page)

77

(continued) Food matrix Conditions Observed effects References - In comparison to fresh sample, OD promoted higher - Fresh losses of total phenolic compounds (34%) and Chockeberries - OD (50 °C and 80 °C) Samoticha et al. (2016) anthocyanins (77%) than FD, which led to decrease in - FD these compounds contents of 9% and 43%, respectively. - Fresh - No differences were found in total phenolic compounds - OD (50 °C, 60 °C, 70 contents between FD and OD at 50 °C, while the other Blackcurrant pomace Michalska et al. (2017) °C, 80 °C and 90°C) conditions led to a progressively decrease in these - FD compounds of up to 54%. - In comparison to OD, FD sample contained higher - OD (55 °C for 72 h) Plums cultivars contents of total anthocyanins and chlorogenic acid Vandgal et al. (2017) - FD determined by HPLC.

78

Freeze-drying principle can be better understood by referring to a pressure-time diagram of pure water (Figure 15). At the pressure and temperature values of 6.1 hPa and 0 °C, respectively, the solid, liquid and gaseous states of water coexist in equilibrium, which is known as triple point (Figure 15). If the pressure is greater than 6.1 hPa, by increasing the temperature, the water passes from solid to liquid states and from liquid to gaseous states. On the other hand, below this pressure value (< 6.1 hPa), the water passes directly from the solid to the gaseous state (sublimation) (Figure 15), which is the basis of freeze-drying dehydration (Kharaghani et al., 2017).

Figure 15. Phase diagram of pure water. Letters indicate the different steps in the freeze-drying process: liquid water freezes into ice (A–B); pressure reduction (B–C) and temperature increased (B–C) enable the water sublimation and desorption (Kharaghani et al., 2017).

In Figure 15 it is also possible to observe the required steps to perform the lyophilization. The first step consists of reducing the temperature to freeze the sample (Figure 15; AB). Following the sample freezing, the pressure is reduced (Figure 15; BC) and the temperature is increased (Figure 15; CD), which allows sublimation of the frozen water. This step is known as primary drying and it is often the longest drying stage, resulting in a product with approximately 5% to 10% of unfrozen water, which is trapped inside the food matrix. In parallel with the frozen water sublimation, desorption of the unfrozen or bound water occurs, which is called as secondary drying and results in a product with 1% to 2% of water (Kharaghani et al., 2017). 79

The freeze-dryer equipment presents four main components: drying or vacuum chamber, condenser, vacuum pump and compressor (Figure 16) (Kharaghani et al., 2017; Stapley, 2008). The vacuum or drying chamber (Figure 16a) is where the food dehydration happens. According to the type of drying chamber, freeze-dryers can be classified into manifold dryers, which are more common in laboratory-scale equipment, into shelf dryers, mainly found in industries, and a combination of both types (Kharaghani et al., 2017). The condenser (Figure 16b) operates at temperatures between - 40 °C and - 60 °C and condenses the water vapor released from the food. The vacuum pump (Figure 16c), in turn, is responsible for reducing the chamber pressure and removing the non-condensable gases, since these gases slow the transfer rate of the water vapor from the food to the condenser (Stapley, 2008). A common misleading is to think that the vacuum pump is used to remove the water vapor. However, it is known that the water is effectively removed by the condenser (Stapley, 2008). Finally, the compressor cools the ice condenser (Kharaghani et al., 2017). Besides these components, it is worth mentioning the presence of heat suppliers, which favor the sublimation of food water. In the case of laboratory-scale freeze-dryer, the heat is supplied by convection from the surrounding air through the flask wall, in an uncontrolled method of heating. On the other hand, in industrial- scale freeze-dryers, the heat is supplied by conduction, radiation, electric resistance, microwave or infrared (Stapley, 2008; Considine, 2005).

Figure 16. Schematic diagram of a typical shelf freeze-dryer. Drying chamber (a); Ice condenser (b); Vacuum pump (c); Compressor (d). (Kharaghani et al., 2017).

80

3. Metabolism of ellagitannins

Several studies indicate that ellagitannins (ET) are not absorbed as such (González- Barrio et al., 2010, Espín et al., 2007; Mertens-Talcott et al., 2006; Cerdá et al., 2005a; Seeram et al., 2004). Only two studies found the intact punicalagin in rat plasma after its administration in high doses during 37 days (Cerdá et al., 2003a; Cerdá et al., 2003b). Thus, it is assumed that these molecules must undergo hydrolysis to ellagic acid (EA) (González-Barrio et al., 2011a; Cerdá et al., 2003b), which, in turn, is extensively metabolized by gut microbiota into urolithins (García-Villalba et al., 2013; Gonzáles-Barrio et al., 2011; Cerdá et al., 2005b; Cerdá et al., 2003b), metabolites that are more bioavailable and biologically active than their precursors (Tomás-Barberán et al., 2017; Landete, 2011). However, it remains unclear if ETs are efficiently hydrolyzed in EA under the physiological conditions of the gastrointestinal tract or whether the involvement of gut microbiota is necessary (Tomás-Barberán et al., 2017; Landete, 2011; Daniel et al., 1991). In addition, it is not known if urolithins can be produced directly from ET or if it is necessary the previous release of EA (Tomás-Barberán et al., 2017). In vitro studies showed that ET are stable under gastric conditions (pH range from 2.0 to 3.8) (González-Sarrías et al., 2015; Daniel et al., 1991), while incubation of ET in small and large intestines pH (pH 7 and 8) led to a significantly hydrolysis of ET in EA (Daniel et al., 1991), demonstrating that gastrointestinal tract pH is able to hydrolyze ET. In addition, Daniel et al (1991) observed that incubation of ET with contents from rat’s large intestine led to a significant increase in ET hydrolysis, which suggests that besides pH, components from large intestine, such as gut microbiota, may influence the hydrolysis of these compounds. The action of gut microbiota on ET hydrolysis has been associated with the synthesis of tannase enzymes by these bacteria (Figure 17) (Tomás-Barberán et al., 2017). These enzymes, also named as tannin acyl hydrolase (E.C. 3.1.1.20), catalyze the hydrolysis of tannins ester bonds (Lekha & Lonsane, 1997), releasing hexahydroxydiphenic acid, which suffers spontaneous lactonization to EA (Tomás-Barberán et al., 2017). However, González-Barrio et al (2010) evaluated the metabolism of ET in ileostomized individuals, which present functional loss of colon and consequently loss of colonic microbiota and observed an EA recovery of 241% in the ileal fluid, indicating that ET hydrolysis does not seems to depend on the colonic microbiota. 81

Figure 17. Catabolism of ellagitannins (punicalagin and vescalagin) and ellagic acid (adapted from Tomás-Barberán et al., 2017 and Landete, 2011).

Although non-bioavailable ET are hydrolyzed in EA, this latter compound is poorly absorbed, reaching maximum plasma concentrations (Cmax) from 0.06 μmol/L to 0.11 μmol/L (González-Sarrías et al., 2015; Mertens-Talcott et al., 2006; Seeram et al., 2006; Seeram et al., 2004) following the consumption of doses ranging from 12 mg to 524 mg of EA and from 130 to 330 mg of ET. Among these studies, González-Sarrías et al. (2015) observed that the consumption of higher doses of free EA (130 mg ET + 524 mg EA), in comparison to higher doses of ET (279 mg ET + 25 mg of EA) did not increase the bioavailability of EA. The low bioavailability of EA is probably due to its high insolubility in aqueous medium, especially at low pH values (Daniel et al., 1991; González-Sarrías et al., 2015), as well as its ability to bind and accumulate in the intestinal epithelium (Whitley et al., 2006) and its ability to the formation of complexes with calcium and magnesium ions in the intestine (Seeram et al., 2006). Additionally, Whitley et al. (2006) observed that some EA concentrations (5 μM to 25 μM) 82

inhibit the action of the organic anion transporter, an intestinal transporter, which exhibit high affinity for EA and are probable transporters of this compound, which may reduce its absorption. In addition, another factor that may limit EA bioavailability is the fact that free EA is a minor compound in the diet and needs to be released from non-bioavailable ET (González- Sarrías et al., 2015). Although poorly absorbed, EA is found in plasma, exhibiting a time to reach maximal concentration (Tmax) from 0.98 h to 2.58 h (González-Sarrías et al., 2015; Seeram et al., 2008; Seeram et al., 2006; Mertens-Talcott et al., 2006). The rapid detection of EA in plasma and urine suggests that its absorption occurs in the stomach. In fact, Espín et al. (2007) observed the presence of EA metabolites in bile and urine of pigs and its absence in the intestinal tissues, suggesting that EA is directly absorbed in the first portions of the gastrointestinal tract. Several studies have detected free EA (González-Sarrías et al., 2015; Seeram et al., 2008; Mertens- Talcott et al., 2006; Seeram et al., 2006; Seeram et al., 2004) and dimethylellagic acid (DMEAG) (Mertens-Talcott et al., 2006) in plasma. On the other hand, in urine were found EA (García-Muñoz et al., 2014), DMEAG (Ludwig et al., 2015; García-Muñoz et al., 2014; Nuñez- Sanchez et al., 2014), methyl ellagic acid (MEA) (Nuñez-Sanchez et al., 2014), hexahydroxydiphenic acid (HHDP) and the glucuronidated forms of DMEAG (García-Muñoz et al., 2014; Nuñez-Sanchez et al., 2014) and MEA (Nuñez-Sanchez et al., 2014). In the intestine, the non-absorbed EA undergoes lactone ring cleavage, decarboxylation and subsequent dihydroxylations reactions until the formation of different urolithins. These reactions are catalyzed, respectively, by bacterial lactonases, decarboxylases and dehydroxylases (Tomás-Barberán et al., 2017). In order to propose the catabolic routes and the time course production of urolithins, García-Villalba et al (2013) incubated EA with gut microbiota of healthy human volunteers. The results of this study suggested that EA lactone ring opening, and its subsequent decarboxylation led to the production of the pentahydroxy- urolithin (urolithin-M5) (Figure 18), which is the key intermediate in the production of different urolithins. Uro-M5, in turn, undergoes dihydroxylations leading to the production of three different tetrahydroxy-urolithins (uro-D, uro-M6 and uro-E). These latter urolithins are transformed into trihydroxy-urolithins (uro-C and uro-M7), which finally led to the main metabolites detected in the human organism, the dihydroxy-urolithins, uro-A and isourolithin- A, and the monohydroxy-urolithin, uro-B. It is worth mentioning that these results are corroborated by earlier studies performed in vitro (González-Barrio et al., 2011a) and in vivo, in pigs (Espín et al., 2007), which proposed a preliminary pathway on the conversion of ellagitannins in some urolithins. 83

Figure 18. Production of urolithins from ellagic acid by gut microbiota and urolithin metabotypes. ( ) Metabotype 0; ( ) Metabotype A; ( ) Metabotype B. (a) Lactonase/decarboxylase enzyme. Bacterial sources: Gordonibacter urolithinfaciens (DSM 27213T) and Gordonibacter pamelaeae (DSM 19378T). (b) Dehydroxylase enzyme. Bacterial sources: G. urolithinfaciens (DSM 27213T) and G. pamelaeae (DSM 19378T). (c) Dehydroxylase enzymes. Bacterial sources: G. urolithinfaciens (DSM 27213T) and G. pamelaeae (DSM 19378T). (d-h) Dehydroxylase enzymes. Bacterial sources: Unknown. (i) Cytochrome P450 (CYP) enzyme (tentative) (j) Dehydroxylase enzyme (suggested). (Tomás-Barberán et al., 2017) 84

Even though in vitro studies have reported the conversion of ET and EA into different aglycone urolithins (García-Villalba et al., 2013; González-Barrio et al., 2011a; Cerdá et al., 2005b), in vivo studies have shown that urolithins are mainly excreted as glucuronidated forms (Romo-Vaquero et al., 2015; Nuñez-Sanchez et al., 2014; Truchado et al., 2012; González- Barrio et al., 2011a; González-Barrio et al., 2010), which suggests that after absorption, urolithins are conjugated to glucuronic acid in the intestine or liver (González-Barrio et al., 2011a). In fact, Espín et al (2007) evaluated the metabolism of ET in pigs and did not detect conjugated forms of urolithins in the intestinal lumen, whereas these compounds were found in intestinal tissue and bile, confirming the previous hypothesis. On the other hand, few studies reported the presence of sulfated forms of urolithins in urine (Nuñez-Sanchez et al., 2014) and bile (Espín et al., 2007), which indicates that urolithins are preferentially conjugated to glucuronic acid (Espín et al., 2017). Additionally, several in vivo studies have reported that the major urolithins excreted in urine are the final metabolites, uro-A, uro-B and isouro-A (Romo- Vaquero et al., 2015; Nuñez-Sanchez et al., 2014; Truchado et al., 2012; González-Barrio et al., 2011a; González-Barrio et al., 2010). As described above, the intermediate urolithins undergo dihydroxylation reactions until the production of the final metabolites (García-Villalba et al., 2013). These reactions increase the molecules lipophilic character, facilitating their intestinal absorption and, consequently, increasing their bioavailability (Espín et al., 2007), which may explain these findings. Several studies have reported that urolithins are excreted in urine from 7 h to 16 h (Ludwig et al., 2015; González-Barrio et al., 2010; Seeram et al., 2006; Cerdá et al., 2005a) after the consumption of ET and EA food sources and persist in the organism of up to 92 h (Truchado et al 2012). The long time required for these compounds to be detected in the urine is due to their extensive metabolization by the gut microbiota (García-Villalba et al., 2013; González-Barrio et al., 2011a). In addition, the persistence of these compounds in the organism is probably due to the enterohepatic circulation, where the metabolites are absorbed, reach the liver and can be released again in the intestine through bile secretion, where they can be reabsorbed (Espín et al., 2007). In fact, Espin et al., 2007 detected a large diversity of urolithins in pig’s gallbladder after the consumption of ET. Studies performed in vivo reported a large interindividual variability in the amounts of urolithins produced by humans (García-Muñoz et al., 2014; Nuñez-Sanchez et al., 2014; Truchado et al., 2012; González-Sarrías et al., 2010; Cerdá et al., 2005a) and even by pigs (Espín et al., 2007). According to these authors, since urolithins are colonic microbiota metabolites, the great interindividual variability is probably due to differences in the gut 85

microbiota composition. In addition, some in vitro studies incubated ET and EA with humans fecal samples and also observed a great interindividual variability in the amount of urolithins produced (García-Villalba et al., 2013; González-Barrio et al., 2011a; Cerdá et al., 2005b), which corroborates the previous hypothesis. The large interindividual variability in ET metabolites production observed in some studies, led to the stratification of the studied populations according to the level of urolithin production as non-producers or low, medium and high producers (Figure 19) (García-Muñoz et al., 2014; Nuñez-Sanchez et al., 2014; Truchado et al., 2012; González-Sarrías et al., 2010; Cerdá et al., 2005a). It is important to note that in the group of non-producers there may be subjects in whom the excreted urolithins were not detected because they are in concentrations below the limit of analytical detection and they may also be considered as low producers (Tomás-Barberán et al., 2017). Another way to stratify the subjects is according to the ability of non-producers to start producing urolithins after a chronic consumption of ET and EA food sources, which was observed in some clinical trials (González-Sarrías et al., 2017; Li et al., 2015; Puuponen-Pimia et al., 2013). In these studies, non-producing individuals who started to produce ET metabolites were classified as responders (Figure 19), while the other subjects were classified as non-responders.

Figure 19. Subjects stratification according to urolithins production after ellagitannins and ellagic acid consumption. *Urolithins, if present, are below the limit of detection (Adapted from Tomás-Barberán et al., 2017).

In addition to the stratification of individuals according to the level of urolithins produced, recently, Tomás-Barberán et al. (2014) proposed a classification of subjects into three metabotypes or phenotypes, according to their ability to produce ET metabolites, as well as the profile of urolithins produced (Figure 19). Thus, individuals were categorized into metabotype A (uro-A producers), metabotype B (uro-A, uro-B and isouro-A producers) and metabotype 0 (non-producers of these urolithins) (Figure 18). 86

The three metabotypes were observed in all the studies performed so far, regardless of the food source or amount of EA consumed, the number of subjects (n), age, and nutritional and health status of the studied population (Table 5). However, several studies have shown that some of these factors may affect metabotypes distribution (Table 5). Considering the studies performed on healthy volunteers (Cortés-Martín et al., 2018; Selma et al., 2018; Tomás- Barberán et al., 2014; Truchado et al., 2012), metabotypes distribution is, on average, 70% of A, 20% of B and 10% of 0. However, a higher proportion of metabotype B was observed in healthy overweight and obese individuals (~ 30%) (Selma et al., 2018; Tomás-Barberán et al., 2014; González-Sarrías et al., 2017), as well as in subjects with diseases associated with gut dysbiosis, such as metabolic syndrome and colorectal cancer (41% to 50%) (Nuñez-Sanchez et al., 2014; Puuponen-Pimia et al., 2013; Tulipani et al., 2012). However, Cortés-Martín et al (2018) observed that metabotypes distribution was only associated with body mass index (BMI) in individuals aged between 5 to 40 years. These authors observed that this distribution is mainly affected by age – increasing age led to an increase in metabotype B concomitant to a decrease of metabotype A, while the metabotype 0 remains constant (Table 5). On the other hand, it is worth mentioning that metabotype A, was the most prevalent not only in normoweight-healthy individuals (Selma et al., 2018; Tomás-Barberán et al., 2014; Truchado et al., 2012), but also in subjects with diseases such as benign hyperplasia or prostate cancer, which are not associated with gut dysbiosis (González-Sarrías et al., 2010). These results suggest that factors associated with dysbiosis, such as weight gain, increased age and some diseases, such as metabolic syndrome and colorectal cancer, may increase the proportion of isourolithin-A and urolithin-B producing bacteria (Cortés-Martín et al., 2018; Selma et al., 2018; Selma et al., 2016; Tomás-Barberán et al., 2014).

87

Table 5. Urolithin urinary metabotypes distribution in different human trials (adapted from Tomás-Barberán et al., 2014). Food and EA contents Age BMI Metabotypes distribution References N Health status after hydrolysis (years) (kg/m2) A B 0 Tomás-Barberán et al.1 Walnuts (162.8 mg/day) 20 (10 M, 10 W) 21-55 23.8±2.3 Healthy 65% 20% 15% Tomás-Barberán et al.1 PE (220 mg) 20 (10 M, 10 W) 18-23 20.5±2.1 Healthy 80% 10% 10% Truchado et al.2 Strawberry (150 mg) 20 (8 M, 12 W) 25-30 NA Healthy 80% 15% 5% Selma et al.3 Walnuts (162.8 mg/day) 20 (11 M, 9 W) 33.6±10.2 22.8±1.4 Healthy 70% 20% 10% Walnuts (102.5 mg/day) Cortés-Martín et al.4 159 (82 M, 77 W) 5-10 18.9±1.3 Healthy 81% 8% 11% or PJ (75.8 mg/day) Walnuts (102.5 mg/day) Cortés-Martín et al.4 139 (65 M, 74 W) 11-14 21.5±0.9 Healthy 79% 14% 7% or PJ (75.8 mg/day) Walnuts (102.5 mg/day) Cortés-Martín et al.4 127 (63 M, 64 W) 15-19 23.8±2.7 Healthy 77% 14% 10% or PJ (75.8 mg/day) Walnuts (123 mg/day) Cortés-Martín et al.4 219 (61 M, 158 W) 20-72 27.4±3.3 Healthy 52% 41% 7% or PE (218.4 mg/day) Tomás-Barberán et al.1 PE (110 mg/day) 49 (32 M, 17 W) 40-65 30.4±3.4 Healthy OW/OB 60% 30% 10% Selma et al.3 PE (55 mg/day) 49 (32 M, 17 W) 45.7±6.7 31.4±3.2 Healthy OW/OB 57% 31% 12% González-Sarrías et al.5 PE (72.9 mg/day) 49 (32 M, 17 W) 40-65 30.4±3.4 Healthy OW/OB 63% 31% 6% Selma et al.3 Mixed nuts (81.4 mg/day) 50 (28 M, 22 W) 51.8±8.3 31.4±3.2 MetS 50% 41% 9% Puuponen-Pimia et al.6 Berries (222 mg/day) 20 (10 M, 10 W) 50-65 31.8±4.4 MetS 25% 50% 25% EA: ellagic acid. PE: pomegranate extract. PJ: pomegranate juice. M: men. W: women. NA: not available. OW/OB: overweight/obese. MetS: metabolic syndrome. References: 1. Tomás-Barberán et al (2014); 2. Truchado et al. (2012); 3. Selma et al. (2018); 4. Cortés-Martín et al (2018); 5. González-Sarrías et al. (2017); 6. Puuponen-Pimia et al (2013); 7. Tulipani et al (2012); 8. González-Sarrías et al. (2010); 9. Nuñez-Sanchez et al (2014).(continued on next page)

88

(continued) Food and ellagic acid Age BMI Metabotypes distribution References N Health status contents after hydrolysis (years) (kg/m2) A B 0 Tulipani et al.7 Mixed nuts (81.4 mg/day) 20 (13 M, 9 W) 31-63 31.0±2.9 MetS 50% 41% 9% Walnuts (190 mg/day) Prostate cancer or González-Sarrías et al.8 28 M 56-90 27.7±3.6 60% 15% 25% or PJ (165.7 mg/day) hyperplasia Nuñez-Sanchez et al.9 PE (110 mg/day) 26 (14 M, 12 W) 52-89 28.9±3.9 Colorectal cancer 42% 42% 16% EA: ellagic acid. PE: pomegranate extract. PJ: pomegranate juice. M: men. W: women. NA: not available. OW/OB: overweight/obese. MetS: metabolic syndrome. References: 1. Tomás-Barberán et al (2014); 2. Truchado et al. (2012); 3. Selma et al. (2018); 4. Cortés-Martín et al (2018); 5. González-Sarrías et al. (2017); 6. Puuponen-Pimia et al (2013); 7. Tulipani et al (2012); 8. González-Sarrías et al. (2010); 9. Nuñez-Sanchez et al (2014).

89

Some clinical studies have shown that classification of subjects into metabotypes helps to explain the controversial biological effects of some phenolic compounds consumption in health (Tomás-Barberán et al., 2016). For soy isoflavones, several studies have shown that individuals which produce equol, an isoflavone metabolite, respond better to the beneficial effects of soy consumption in comparison to non-equol producers, presenting a vascular function improvement (Hazim et al., 2016; Kreijkamp-Kaspers et al., 2005), as well as a reduced risk to develop coronary heart diseases (Liu et al., 2014; Zhang et al., 2012). On the other hand, the association between urolithins metabotypes and bioactivity has recently been described in the literature. Metabotype B individuals present increased risk for cardiovascular diseases (Selma et al., 2018; González-Sarrías et al., 2017), while metabotype A appear to present a protective factor against these diseases (Selma et al., 2018). Moreover, González- Sarrías et al. (2017) reported that the consumption of a pomegranate extract rich in ellagitannins, by obese individuals, presented a hypolipidemic effect, but only in metabotype B individuals, suggesting that this metabotype seems to respond better to the beneficial effects of ellagitannins consumption. The results of these studies can partially explain the absence of phenolic compounds health claims and show the relevance of considering the metabotype to evaluate polyphenols-rich foods bioactivity. The distribution of isoflavone metabotypes, as equol producers or non-producers, has already been described in different populations, in which equol-producers correspond to about 50% of the eastern population (Guo et al., 2010; Song et al., 2006; Morton et al., 2002; Arai et al., 2000), against only 30% of western population (Gardana et al., 2009; Fuhrman et al., 2008; Song et al., 2006). These differences are probably due to differences between eating habits and lifestyle of these populations, which may affect the gut microbiota composition and consequently, the metabotypes distribution (Tomás-Barberán et al., 2016). On the other hand, although several studies have already evaluated the urolithin metabotypes distribution in the European population (Table 5), so far, there are no reports on the study of metabotype distribution in other populations. Thus, considering the differences observed in isoflavone metabotypes in different populations, the characterization of urolithin metabotypes in other continents and countries, such as Brazil, is relevant.

90

CHAPTER 1:

Study on the effect of pressurization and of drying methods on phenolic compounds profile of jabuticaba (Myrciaria jaboticaba) peel and seed

Manuscript to be published in Food Chemistry

91

1. Introduction

Jabuticaba (Myrciaria jaboticaba) is a Brazilian berry, which presents a whitish pulp with a sweet to slightly tangy taste, containing 1 to 4 seeds and covered by a dark purple to black peel. Although jabuticaba is highly appreciated for its taste and flavor, the high perishability of the fruit hampers its commercialization, which is limited to some regions of the country, where the fruit is mainly consumed in natura, as well as artisanal products such as juices, jellies, liqueurs and fermented beverages (Inada et al., 2015). In recent years, several studies have reported the beneficial effects of jabuticaba in vitro (Wang et al., 2014; Leite- Legatti et al., 2012; Reynertson et al., 2006) and in vivo, in animal models (Moura et al., 2018; Batista et al., 2017; Araújo et al., 2014; Dragano et al 2013), and more recently, in humans (Balisteiro et al., 2018; Plaza et al., 2016). In addition, regarding jabuticaba chemical composition, some studies have reported that this fruit is a source of vitamins, minerals and fibers (Inada et al., 2015), and presents high antioxidant activity and phenolic compounds contents, especially anthocyanins and ellagitannins (Pereira et al., 2017; Plaza et al., 2016; Inada et al., 2015; Wu et al., 2012). However, most of these compounds are concentrated in jabuticaba peel and seed, fractions that are not usually consumed, and that represent about 40% of the fruit’s weight (Inada et al., 2015). Thus, considering both the potential beneficial health effects of jabuticaba consumption and its perishability, it is relevant to investigate the application of technological processes, such as high hydrostatic pressure and dehydration methods, on phenolic compounds contents of jabuticaba peel and seed. High hydrostatic pressure (HHP) is a non-thermal conservation method, which consists of subjecting food to HHP levels from 100 to 1000 MPa, in order to destroy microorganisms and inactivate enzymes, while preserving the food sensorial and nutritional characteristics (Inada et al., 2018a; Silva & Sulaiman et al., 2018; Liu et al., 2016; Barba et al., 2013; Varela- Santos et al., 2012; Yaldagard et al., 2008). Moreover, some studies have shown that HHP can affect the microstructure of food plant tissues, increasing the extractability of bioactive compounds (Xi & Luo, 2016; Vásquez-Gutiérrez et al., 2013; Vásquez-Gutiérrez et al., 2011; Inada et al., 2018a). Therefore, in addition to is application as conservation technique, HHP has been used to extract bioactive compounds in some food matrices, being even denominated as “high hydrostatic pressure extraction” (Xio & Luo, 2016; Corrales et al., 2009; Jun et al., 2009). In addition, Rodríguez-Roque et al. (2015) observed that HHP increased the bioaccessibility of phenolic compounds from fruit juices and related beverages, suggesting that pressurization may increase not only the extractability of bioactive compounds, but also their intestinal absorption. 92

These studies show the relevance of HHP processing of polyphenols-rich foods, such as jabuticaba. Dehydration is a conservation method that reduces water activity of food and consequently decreases microbial growth, as well as the occurrence of chemical and enzymatic reactions, extending shelf life (Lewick, 2006). Besides, by removing most of the water present in foods, it leads to the production of a dry powder highly concentrated on bioactive compounds, such as phenolic compounds. Among the existing dehydration methods, oven drying (also known as convective drying or hot air drying) is one of the most popular dehydration methods due to its high drying efficiency and low cost (Lewick, 2006; Ratti, 2001). However, the long exposure of food to high temperature and oxygen can lead to oxidation of phenolic compounds, as well as color change (Zhou et al., 2017; Samoticha et al., 2016; Michalska et al., 2016; Wojdylo et al., 2014; Wojdylo et al., 2009). On the other hand, freeze drying is a method performed at low temperature and pressure, in which the water is removed by sublimation, promoting less degradation of bioactive compounds and color. Although freeze drying usually leads to high quality products (Samoticha et al., 2016; Michalska et al., 2016; Mphahlele et al., 2016; Wojdylo et al., 2014; Torres et al., 2010), this method presents low efficiency of drying, as well as high energy consumption and cost (Ratti, 2001). Thus, the aim of this study was to evaluate the effect of different processing technologies (high hydrostatic pressure, oven-drying and freeze-drying) on the chemical composition and physical characteristics of jabuticaba peel and seed, with emphasis on phenolic compounds.

2. Material and methods

2.1 Standards and chemicals

2,4,6-tris(2-pyridyl)-S-triazine (TPTZ), 2,2’-azino-bis (2-ethylbenzothiazoline-6- sulfonic acid) diammonium salt (ABTS), potassium persulfate, (±)-6-hydroxy-2,5,7,8- tetramethylchromane-2-carboxylic acid (Trolox), polyvinylpyrrolidone, phloroglucinol, catechol and sodium acetate were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA). Iron (II) sulfate, Triton X-100® and hydrochloric acid were purchased from Merck KGaA (Darmstadt, Germany). L(+)-ascorbic acid was purchased from Acros-Organics, Thermo Fischer Scientific Inc. (Waltham, USA) Anthocyanins and non-anthocyanins standards were purchased respectively from Indofine Chemical Co. (Hillsborough, NJ, USA) and Sigma- 93

Aldrich Chemical Co. (St. Louis, MO, USA). All solvents were HPLC grade. HPLC grade water (Milli-Q system, Millipore, Bedford, MA, USA) was used throughout the experiments.

2.2 Samples

Jabuticaba fruits (Myrciaria jaboticaba, cv. Sabará) were purchased at Rio de Janeiro’s agricultural trading center. Fruits were selected, washed and sanitized in sodium hypochlorite 100 ppm solution for 15 min. Jabuticaba peel and seed (JPS) were separated in a Bonina® 0.25 df horizontal depulper (NPC Equipamentos®, Itabuna, Brazil), subsequently packed in aseptic polyethylene bags and stored at -20 °C.

2.3 High Hydrostatic Pressure processing of jabuticaba peel and seed

Fifty grams of JPS were thawed, transferred to a small 5-layer nylon bag with oxygen barrier and vacuum-sealed (TecMaq®, São Paulo, Brazil). Bags were subjected to HHP processing in a pilot equipment (Bras Solution Ltd., Rio de Janeiro, Brazil), using a mixture of distilled water and a concentrated carboxylate-based synthetic fluid (2:1, v/v) as pressure transmitting medium. HHP experiments were performed according to a 22 full factorial design with a central point in duplicate, totaling 6 experimental runs. Independent variables were pressure (200, 350 and 500 MPa) and time of processing (1, 5.5 and 10 min). Response variables were soluble phenolic compounds contents determined by HPLC-DAD and antioxidant activity (AA) measured by FRAP (Ferric Reducing Antioxidant Power) and TEAC (Trolox Equivalent Antioxidant Capacity) assays.

2.4 Oven drying of jabuticaba peel and seed

JPS were thawed and dehydrated in a forced air circulation oven (Tecnal®, Piracicaba, Brazil). Dehydration was performed according to a 22 full factorial design with a central point in duplicate, totaling 6 experimental runs. Independent variables were temperature (55, 65 and

75 °C) and time (14, 18 and 22 h). Response variables were water activity (aw), phenolic compounds contents determined by HPLC-DAD and HPLC-DAD-MS and AA measured by FRAP and TEAC assays.

94

2.5 Freeze drying of jabuticaba peel and seed

JPS were thawed and dehydrated in a freeze dryer (Thermo Fischer Scientific®, Waltham, USA) at -50 °C and 0.065 mbar for 72 h.

The following analysis were performed in freeze-dried JPS: water activity (aw), phenolic compounds by HPLC-DAD, HPLC-DAD-MS and HPLC-ESI-MS-MS, and AA measured by FRAP and TEAC assays.

2.6 Phenolic compounds analysis

2.6.1 Phenolic acids, flavonols and anthocyanins extraction and quantification by HPLC-DAD

Extraction of soluble phenolic acids, flavonols and anthocyanins from unprocessed, HHP-processed, freeze-dried and oven-dried JPS was performed in triplicate, according to Inada et al. (2015). Samples were extracted for 10 min with 20 mL of cold ethanol 80% and centrifuged (2,500 g, 5 min, 10 °C). The supernatant was collected, and the residue re- extracted. Supernatants were combined, the solvent was removed, and the dry residue was reconstituted in water. The extract was filtered through a 0.45 µm PTFE (Analítica®, São Paulo, Brazil) prior to HPLC analysis. The liquid chromatography system (Shimadzu®, Japan) included two parallel pumps LC-20AD, automatic injector SIL-20AHT, system controller CBM-20A, degasser DGU-20A5 and diode-array detector (DAD) SPD-M20A. Chromatographic separation of anthocyanins was achieved using a reverse phase column C18 (5 μm, 150 mm × 4.6 mm, Phenomenex®), according to the adapted methodology of Inada et al. (2015). The mobile phase consisted of a gradient of 1% formic acid and 2% acetonitrile in water (eluent A) and 1% formic acid and 2% acetonitrile in methanol (eluent B), at a flow rate of 1.0 mL/min. Prior to injection, the column was equilibrated with 23% B. After injection of sample, solvent composition was kept constant until 1 min, increased to 29% B in 2 min, to 33% B in 4 min, to 48% B in 6 min, to 85% B in 8 min, to 95% B in 10 min and then decreased to 23% B in 11 min. Between injections, 10 min intervals were used to re-equilibrate the column with 23% B. Anthocyanins were monitored by DAD at 530 nm. The injection volume was 10 µL. Chromatographic separation of phenolic acids and flavonols was achieved using a reverse phase column C18 (5 μm, 250 mm × 4.6 mm, Phenomenex®), according to the adapted methodology of Inada et al. (2015). The mobile phase consisted of a gradient of 0.3% formic 95

acid and 1% acetonitrile in water (eluent A) and 1% acetonitrile in methanol (eluent B), at a flow rate of 1.0 mL/min. Prior to injection, the column was equilibrated with 18.2% B. After sample injection, solvent composition changed 20.2% B in 1 min, to 43.4% B in 18 min, 85.9% B in 23 min and kept constant until 30 min. Between injections, 10 min intervals were used to re-equilibrate the column with 18.2% B. These phenolic compounds were monitored by DAD from 190 to 370 nm. The injection volume was 10 µL. Identification was performed by comparison with retention time and absorption spectrum of the respective standard. Quantification was performed by external calibration. Data were acquired by Lab Solutions software (Shimadzu Corporation®, version 5.82 SP1, 2008- 2015).

2.6.2 Ellagitannins, ellagic acid derivatives and gallic acid extraction and quantification by HPLC-DAD-MS

Extraction of ellagitannins and ellagic acid derivatives compounds from freeze-dried and oven-dried JPS was performed in triplicate, according to the adapted methodology of García-Villalba et al. (2015). One hundred and fifty milligrams of samples were extracted for 1 min with 10 mL of 70% methanol and centrifuged (4,696 g, 10 min, 20 °C). The supernatant was collected and filtered through a 0.22 µm PVDF filter (Merck KGaA®, Darmstadt, Germany). The liquid chromatography system was an Agilent 1200 HPLC equipped with a diode array detector and a single quadrupole mass spectrometer (6120 Quadrupole, Agilent Technologies, Santa Clara, CA, USA). Chromatographic separation of compounds was achieved using a Poroshell 120 EC-C18 column (2.7 μm, 100 mm × 3 mm, Agilent Technologies) operating at 25 °C. The mobile phase consisted of a gradient of 1% aqueous formic acid (eluent A) and acetonitrile (eluent B), at a flow rate of 0.5 mL/min. Prior to injection, the column was equilibrated with 5% B. After injection of sample, solvent composition increased to 18% B at 7 min, 28% B at 17 min, 50% at 22 min and 90% at 27 min, which was maintained up to 28 min. The initial conditions were re-established at 29 min and kept under isocratic conditions up to 35 min. The injection volume was 5 μL. Ellagitannins and ellagic acid derivatives were monitored at 280 and 360 nm, respectively. Mass spectrometry with single Q analyzer was used to confirm the identification. Optimal ESI-MS parameters using nitrogen as nebulizer gas were: capillary voltage 3500 V; drying gas flow 10 L/min; nebulizer pressure 40 psi, drying temperature 300 °C. MS spectra were acquired in negative 96

ionization mode and measured in selective ion monitoring (SIM) mode. Identification of all compounds was carried out by their typical spectral properties and molecular masses. Gallic acid, ellagic acid, vescalagin and castalagin were also identified by the retention time of their respective standards. Quantification of gallic acid, ellagic acid, vescalagin and castalagin was carried out using calibration curves of respective standards. The other ellagic acid derivatives and ellagitannins were quantified using calibration curves of ellagic acid and vescalagin, respectively.

2.6.3 Proanthocyanidins extraction and identification by HPLC-ESI-MS-MS

Extraction of proanthocyanidins from freeze-dried JPS was performed according to the methodology of Kennedy and Jones (2001), using acid catalysis in the presence of phloroglucinol. Phloroglucinol reagent consisted of a 0.1 M HCl methanolic solution, containing 50 g/L phloroglucinol and 10 g/L ascorbic acid. Fifty milligrams of freeze-dried JPS were dissolved in 800 µL of phloroglucinol reagent. The reaction mix was vortexed and incubated at 50 °C for 20 min, under stirring. To stop the reaction, the sample was placed in ice and 1 mL of 40 mM sodium acetate solution was added. The sample was centrifuged (3,650 g, 10 min, 20 °C) and filtered through a 0.22 µm PVDF filter (Merck KGaA®, Darmstadt, Germany) prior to injection in HPLC. The liquid chromatography system was an Agilent 1100 HPLC equipped with a MSD Trap 1100 Series (Agilent). Chromatographic separation of compounds was achieved according to adapted methodology of Karonen et al. (2007), using a Pursuit XRs C18 column (5 μm, 250 mm × 4 mm, Agilent Technologies) operating at 25°C. The mobile phase consisted of a gradient of 1% aqueous formic acid (eluent A) and acetonitrile (eluent B), at a flow rate of 0.8 mL/min. Prior to injection, the column was equilibrated with 3% B. After injection of sample, solvent composition increased to 9% B at 5 min, 16% B at 15 min, 50% at 45 min and 90% at 47 min, which was maintained up to 52 min. The initial conditions were re-established at 52 min and kept under isocratic conditions up to 57 min. The injection volume was 8 μL. The MS detector operated in negative ion-mode. The trap interface and ion optics settings were the following: spray potential 65 psi; nebulization gas (nitrogen) relative flow value 11; capillary temperature 350 °C. Full-scan mass spectra were acquired scanning the range 50–1200 m/z. The products obtained after phloroglucinolysis (phloroglucinol adducts and flavan-3-ol-monomers) were identified by their elution order and MS spectra.

97

2.7 Antioxidant activity

The AA of the soluble phenolic compounds extract (described in item 2.6.1) were performed by FRAP (Benzie and Strain, 1996) and TEAC assays (Re, Pellegrini, Proteggente, Pannala, Yang and Rice-Evans, 1999), with slight modifications. Results were expressed as mmol of Fe2+ equivalents or mmol of Trolox equivalents per 100 g on dry weight basis (dwb). Each extract was analyzed in triplicate.

2.8 Instrumental color measurement

Instrumental color parameters of freeze-dried and oven-dried JPS were measured in triplicate using a Konica Minolta CR-400 colorimeter (Konica Minolta, Tokyo, Japan). The equipment was set to illuminant D65 (2° observer angle) and calibrated using a standard white reflector plate. The CIELab color space was used to determine the color components: L* (brightness or lightness; 0 = black, 100 = white), a* (-a*, greenness; +a*, redness) and b* (-b*, blueness; +b*, yellowness). The total color difference (ΔE*) was calculated as follows:

∗ ∗ 2 ∗ ∗ 2 ∗ ∗ 2 ∆E* = √(퐿푓 − 퐿표) + (푎푓 − 푎표) + (푏푓 − 푏표)

where L*, a*, and b* were the color coordinates of freeze-dried (f) and oven-dried (o) samples.

2.9 Water activity

Water activity (aw) of freeze-dried and oven-dried JPS was measured directly at 25 °C using the LabMaster-aw analyzer (Novasina, Pfáffikon, Switzerland).

2.10 Polyphenoloxidase activity

The polyphenoloxidase (PPO) activity of JPS was evaluated by a colorimetric method, according to Moreno et al. (2013). Two grams of dehydrated samples or eight grams of unprocessed samples were extracted, in an ice bath, with 30 mL of 0.2 M sodium phosphate buffer (pH 6.5) containing 2% of polyvinylpyrrolidone (w/v) and two drops of Triton X-100, using an Ultraturrax extractor T18 BASIC (IKA®, Staufen, Germany), at 14,000 rpm for a total 98

of three min with one min intervals. The extract was centrifuged (4,696 g, 10 min, 4 °C) and the supernatant filtered with paper filter (Whatman no. 1). Three hundred microliters of the PPO extract were added to three milliliters of 0.15 M catechol solution in 0.05 M sodium phosphate buffer (pH 6.5). The increase in absorbance from 0 to 5 min was measured at 420 nm.

2.11 Statistical analyses

Data were expressed as mean ± standard deviation. Experimental design matrixes were generated and analyzed using Statistica software, version 7.0 (StatSoft Inc., Tulsa, OK). Analysis of variance (one-way ANOVA) followed by Tukey’s multiple comparison post-test was performed for comparing different processing conditions applied in JPS, using GraphPad Prism software for Windows, version 5.04 (GraphPad Software, San Diego, CA). Pearson correlation analysis was performed to evaluate associations between variables, using Statistica software, version 7.0 (StatSoft Inc., Tulsa, OK). Results were considered significant when p < 0.05.

3. Results and Discussion

3.1 Jabuticaba peel and seed did not exhibit polyphenoloxidase activity

PPO (EC 1.14.18.1) is an enzyme that catalyzes the oxidation of phenolic compounds present in plant foods to quinones. Several studies have demonstrated that different food technological processes can reduce or even increase this enzyme activity (Silva & Sulaiman, 2018; Liu et al., 2016; Sulaiman et al., 2015; Akissoé et al., 2003; Raynal et al., 1989). Thus, even if a technological process maintains or increases the contents of phenolic compounds of a food, it is important to investigate whether this food exhibits PPO activity, otherwise polyphenols oxidation may occur during storage. For this reason, it is important to evaluate the effect of processing technologies on this enzyme. However, until now no study had evaluated the polyphenoloxidase activity of jabuticaba peel and seeds, which was performed for the first time in the present study. Jabuticaba whole fruit exhibited a small increase in absorbance at 420 nm, indicative of PPO activity. However, JPS did not exhibit PPO activity, suggesting that PPO was localized in the pulp. Lopes et al. (2016) observed higher PPO activity in mango peel (Tommy Atkins variety) than in its pulp. On the other hand, other studies reported that while some varieties of 99

apple and grapes presented higher PPO activity in the peel, others exhibited greater activity in the pulp (Troiani et al., 2003; Janovitz-Klapp et al., 1989). Considering that Vieites et al. (2011) and Daiuto et al. (2010) observed PPO activity in the whole fruit, to confirm our results, we assayed the PPO activity of fresh apple, since this fruit is known to present high activity (Janovitz-Klapp et al., 1989). In addition to the determination of PPO activity of fresh apple extract, two other tests were performed: a positive control, in which the apple extract was mixed with JPS extract, and a negative control, in which the apple extract was heated at 100 °C for 30 minutes to reduce PPO activity. In all three tests, PPO activity was observed, being the highest and lowest activities observed for the apple extract and the negative control, respectively, as expected (results not shown). These results confirm that while jabuticaba pulp shows PPO activity, its peel and seed do not.

3.2 Pressurization was ineffective in increasing phenolic compounds contents and antioxidant activity of jabuticaba peel and seed

In general, unprocessed and pressurized samples presented nine phenolic compounds, including two hydroxybenzoic acid derivatives (gallic and ellagic acids), one hydroxycinnamic acid derivative (m-coumaric acid), four flavonols (myricetin-3-O-rhamnoside, quercetin-3-O- rutinoside, myricetin and quercetin) and two anthocyanins (delphinidin-3-O-glucoside and cyanidin-3-O-glucoside) (Table 1). Myricetin was not detected in some pressurized samples (200 MPa/10 min, 350 MPa/5.5 min and 500 MPa/10 min), while m-coumaric acid was not detected in the unprocessed sample (Table 1). This phenolic compounds profile is in accordance to previously published data (Inada et al., 2015; Wu et al., 2012). HHP factors affected only m-coumaric acid and myricetin-3-O-rhamnoside contents. It is important to note that they were minor phenolic compounds of JPS, contributing, on average, to 0.5% and 0.04% of total phenolic compounds, respectively (Table 1). m-Coumaric acid was negatively affected by time and by the interaction of pressure and time (at higher pressure values, increasing processing time led to lower contents, whereas at lower pressure values the opposite occurred) (Figure 1). Myricetin-3-O-rhamnoside was negatively affected by pressure, time (Figure 1) and positively affected by the interaction of these variables (Figure 1).

100

Table 1. Phenolic compounds contents and antioxidant activity of unprocessed and HHP processed jabuticaba peel and seed.1 Phenolic compounds (mg/100g dwb2) Unprocessed 200 MPa/1 min 200 MPa/10 min 350 MPa/5.5 min 500 MPa/1 min 500 MPa/10 min Phenolic acids Gallic acid 127±4a 81±1c 91±8c 113±7b 34±2d 45±4d m-Coumaric acid ND3 0.4±0.1a 0.4±0.0a 0.3±0.0b 0.4±0.0a 0.3±0.0b Ellagic acid 229±12a 215±9a 212±4a 178±37a 172±3a 171±2a Flavonols Myricetin-3-O-rhamnoside 7.7±0.3a 8.2±0.3a 3.5±0.3c 4.9±0.3b 4.5±0.1b 5.0±0.1b Quercetin-3-O-rutinoside 117±6a 88±4abc 63±1c 84±19bc 96±2abc 116±15ab Myricetin 1.2±0.0c 1.9±0.1a ND* ND* 1.5±0.1b ND* Quercetin 1.1±0.1b 1.7±0.2a 0.4±0.0c 0.7±0.2c 0.6±0.0c 0.3±0.0c Anthocyanins Delphinidin-3-O-glucoside 73±3a 75±2a 15±1d 39±4b 30±3c 19±1d Cyanidin-3-O-glucoside 725±22ab 818±28a 346±8c 644±56b 582±29b 444±22c Total 1,283±47a 1,290±44a 732±22d 1,064±123b 921±38bc 799±44cd Antioxidant activity FRAP (mmol Fe2+/100 g dwb) 98±3a 49±2c 39±2cd 67±7b 32±0d 38±1cd TEAC (mmol Trolox/100 g dwb) 188±4a 114±1b 76±1c 109±2b 71±3c 67±3c 1Results expressed as mean±SD for triplicates. 2DWB - dry weight basis. 3ND - not detected. Phenolic compounds contents were determined by HPLC-DAD. Different superscript letters in the line indicate significant difference between mean values (One-way ANOVA followed by Tukey multiple comparison post hoc test, p<0.05). 101

Figure 1. Pareto charts and surface response charts of high hydrostatic pressure processing conditions that affected the phenolic composition and antioxidant activity of jabuticaba peel and seeds according to experimental design results. In these Pareto charts, the standardized effects that reached the dotted line (p = 0.05) are statistically significant. (P) = pressure; (T) = time of processing; (P) by (T) = interaction between pressure and time of processing. 102

In comparison to unprocessed sample, pressurizing at 200 MPa for 1 min was the only condition that did not show difference in total phenolic compounds contents (Table 1). Higher pressure and longer time resulted in decreases from 17% (350 MPa for 5.5 min) to 43% (200 MPa for 10 min) in total phenolic compounds contents (Table 1). In general, studies reported maintenance or increase of total phenolic compounds even when higher pressure and longer time conditions were applied. An increase of up to 66% of total polyphenols contents were observed after pressurization of onions (300 MPa for 3 min and 600 MPa for 1 to 5 min) (Vásquez-Gutiérrez et al., 2013), blueberry juice (200 MPa for 5 to 15 min and 400 MPa for 15min) (Barba et al., 2013) and pomegranate juice (450 MPa for 90 s to 150 s and 550 MPa for 30 s) (Varela-Santos et al., 2012). However, these studies quantified total polyphenols by spectrophotometric methods that are prone to many interferes, differently from the HPLC technique used in the present study. In that sense, Jez et al., (2018) observed that intense HHP processing conditions (450, 550 and 650 MPa for 5, 10 and 15 min) applied to tomato led to decreases of up to 57% in total polyphenols contents. These results demonstrate the relevance of using specific methods, such as HPLC, to evaluate the effect of food processing technologies on their bioactive compounds contents. Although pressurization at 200 MPa for 1 min did not influence the total polyphenols contents in comparison to the unprocessed sample (Table 1), the contents of minor compounds changed (increase in myricetin and quercetin and decrease in gallic acid). In addition, m- coumaric acid, which was not found in the unprocessed sample, was detected after pressurization. While pressurization at all other conditions led to decreases in total phenolic compounds contents, some compounds such as ellagic acid, quercetin-3-O-rutinoside and cyanidin-3-O-glucoside were not affected and myricetin contents even increased in some HHP conditions (Table 1). These results demonstrate that the effect of HHP on phenolic compounds contents depends not only on the pressure and time conditions, but also on the compound analyzed. In fact, Vega-Gálvez et al. (2014) reported that syringic acid contents decreased after pressurization of gooseberry pulp in some conditions (300 MPa for 1 and 3 min) and increased in others (400 MPa for 5 min and 500 MPa for 1 min), while trans-ferulic acid presented the opposite behavior. Liu et al. (2016) pressurized blue honeysuckle berry and observed that as HHP increased, from 200 to 600 MPa, cyanidin-3-O-glucoside contents increased, while delphinidin-3-O-rutinoside and delphinidin-3-p-coumaroylglucoside contents decreased. Inada et al. (2018a) reported that effect of HHP on phenolic composition of jabuticaba juice was both compound and condition specific and led to an increase of ellagic, trans-cinnamic and m- coumaric acids, myricetin, quercetin-3-O-rutinoside and cyanidin-3-O-glucoside contents, as 103

well as a decrease of gallic acid and myricetin-3-O-rhamnoside after some pressure and time conditions. The increase of specific minor phenolic compounds (m-coumaric acid, myricetin and quercetin) after some pressurizing conditions (Table 1) is probably due to an increment of polyphenols extractability. In fact, HHP has been used in some studies as a method of phenolic compounds extraction from different food matrices, such as a type of gingseng (Angelica sinensis) (Xi and Luo, 2016), green tea leaves (Jun et al., 2009), and grape skins (Corrales et al., 2009), proving to be even more efficient than other conventional extraction methods, such as ultrasound and heat reflux extraction Jun et al. (2009). This increased extractability of phenolic compounds after HHP processing may be associated to microstructural changes in food matrix. Some studies analyzed HHP processed foods by electron microscopy and observed that pressurization generates intense cellular damage, making the tissue porous, suggesting that this intense injury of tissue allows the release of food matrix phenolic compounds, increasing the efficiency of their extraction (Xi and Luo, 2016; Vásquez-Gutiérrez et al., 2011). Moreover, in addition to the hypothesis of the increased extractability of phenolic compounds after pressurization, the increase in quercetin contents after pressurization at 200 MPa for 1 min (Table 1) might be related to the degradation of quercetin-3-O-rutinoside through hydrolysis of its glycosidic bond by HHP, since it was already described that pressurization can accelerate some chemical reactions, such as hydrolysis (Tauscher, 1995). The degradation of some phenolic compounds observed after HHP processing is usually associated to the activation of some enzymes, such as peroxidase (EC 1.11.1.x) and PPO (EC 1.14.18.1), which can catalyze the oxidation of phenolic compounds. Some studies reported the activation of peroxidases and PPO in berries, carrot, apples and pears after some conditions of HHP processing (Liu et al., 2016; Sulaiman et al., 2015; Garcia-Palazon et al., 2004; Anese et al., 1995). However, as discussed in the previous section, no PPO activity was found for JPS, suggesting that the activation of this enzyme would not be the factor responsible for phenolic compounds degradation after HHP processing. AA measured by TEAC, but not by FRAP, was negatively affected by pressure and time and positively affected by the interaction of these variables (Figure 1). When compared to the unprocessed sample, all pressurized samples showed decreases in AA, ranging from 32% (350 MPa for 5.5 min) to 67% (500 MPa for 1 min) for FRAP, and from 39% (200 MPa for 1 min) to 64% (500 MPa for 10 min) for TEAC (Table 1). Similar results were reported in the literature by Blaszczak et al. (2017), who reported a decrease of up to 8% in TEAC values after pressurization of aronia juice. In addition, Jez et al. (2018) also observed a reduction of up to 104

35% in FRAP values after pressurization of tomato under different processing conditions. On the other hand, differently from the present study, a previous study of the present research group (Inada et al., 2018a) reported an increase of up to 46% in FRAP values after HHP processing of jabuticaba juice, as well as a decrease in TEAC values of up to 14%, but only at pressures higher than 200 MPa. TEAC values positively correlated with total phenolic compounds contents (r > 0.79, p = 0.03), while FRAP values only correlated to non-anthocyanins compounds contents (r > 0.76, p = 0.05). Although TEAC presented a positive correlation with delphinidin-3-O-glucoside (r > 0.81, p = 0.03), the minor anthocyanin of jabuticaba, neither of the methods significantly correlated with total anthocyanins contents (p > 0.05). These results suggest that although anthocyanins were the major compounds (61.5%) in JPS samples, they were not the major contributors to AA. However, it is worth mentioning that ellagitannins, found in high concentrations in jabuticaba (Pereira et al., 2017; Plaza et al., 2016), but not determined in unprocessed and HHP processed samples, could contribute to AA. In fact, Plaza et al (2016) reported that hydrolysable tannins were more relevant to the AA of jabuticaba than anthocyanins.

3.3 Freeze-drying of jabuticaba peel and seed led to higher anthocyanins contents, while oven- drying led to higher ellagitannins contents.

Twelve ellagitannins, four ellagic acid derivatives and gallic acid were detected in freeze-dried and oven-dried samples by HPLC-DAD-MS (Table 2). The contents of these compounds, in addition to four flavonols, one hydroxycinnamic acid derivative and two anthocyanins are presented in Table 3. and trans-cinnamic acid were only found in the freeze-dried sample, while HHDP-digalloyl-glucose, di-HHDP-galloyl- glucose and myricetin were only detected in oven-dried samples. This profile is in accordance with published data (Neves et al., 2018; Pereira et al., 2017; Plaza et al., 2016; Inada et al., 2015; Wu et al., 2012).

105

Table 2. Identification of gallic acid, ellagic acid derivatives and ellagitannins by HPLC-DAD- MS-Q of freeze-dried and oven-dried jabuticaba peel and seeds. 1 - Compound RT (min) [M – H] max (nm) Gallic acid 1.883 169 270 Ellagic acid derivatives Ellagic acid pentoside 9.043 433 256, 366 Ellagic acid pentoside 10.009 433 254, 360 Ellagic acid 10.673 301 254, 368 Valoneic acid dilactone 11.587 469 254, 372 Ellagitannins Vescalagin 2.784 933 244 Di-HHDP-glucose isomer 3.185 783 244 Vescalagin/Castalagin isomer 3.726 933 244 Castalagin 4.053 933 244 Di-HHDP-glucose isomer 5.131 783 246 HHDP-digalloyl-glucose 6.083 785 270 Di-HHDP-galloyl-glucose 7.025 935 274 Di-HHDP-glucose isomer 7.151 783 274 Di-HHDP-glucose isomer 9.306 783 260 Vescalagin/Castalagin isomer 9.482 933 272 Vescalagin/Castalagin isomer 10.514 933 256 Vescalagin/Castalagin isomer 12.056 933 244 1Retention time. Identification of gallic acid, ellagic acid, vescalagin and castalagin was performed by comparison with retention time, absorption spectrum and molecular mass in negative ionization mode of the respective standard. The other compounds were identified by the molecular mass and the typical absorption spectrum.

106

Table 3. Phenolic compounds contents, antioxidant activity, instrumental color and water activity of freeze-dried and oven-dried JPS*.1 Phenolic compounds (mg/100g dwb2) Freeze-dried 55 oC/14 h 55 oC/22 h 65 oC/18 h 75 oC/14 h 75 oC/22 h Gallic acid derivatives Gallic acida 50±0c 68±1bc 88±1b 55±17c 157±3a 102±4b Ellagic acid derivatives Ellagic acida,b 314±6a 237±14b 272±15ab 244±27b 287±11ab 272±19ab Valoneic acid dilactoneb 32±0 ND3 ND ND ND ND Ellagic acid pentoside isomersb 31±1a 19±2b 21±2b 20±1b 18±1b 21±1b Total 377±7a 256±16b 293±16b 264±28b 306±12b 293±20b Ellagitannins Vescalagina,c 740±12c 818±39c 846±56c 802±39c 1442±47b 1805±75a Castalagina 898±5a 781±28a 808±61a 541±52b 507±40b 561±3b Vescalagin/Castalagin isomersc 543±14a 251±9e 320±14cd 311±15d 488±12b 348±19c Di-HHDP-glucose isomersc 240±6b 190±6c 171±8c 148±6d 308±11a 245±15b HHDP-digalloyl-glucosec ND 26±1b 26±2b 21±3c 36±1a 26±2b Di-HHDP-galloyl-glucosec ND 74±3a 117±10a 125±7a 136±1a 130±3a Total 2,421±38b 2,140±86bc 2,287±151b 1,948±121c 2,917±113a 3,115±117a Flavonols Myricetin-3-O-rhamnosidea 4.6±0.4ab 3.4±0.2b 4.4±0.4ab 3.0±1.0 5.3±0.2a 3.6±0.1b Quercetin-3-O-rutinosidea 16±1c 58±3a 36±0b 30±4bc 24±2c 27±2c Myricetina ND 1.8±0.0a ND ND ND ND Quercetina 1.1±0.0b 0.9±0.0b 1.1±0.1b 1.0±0.1b 1.7±0.1a 1.6±0.1a Total 22±1d 64±3a 42±1b 34±5c 31±2c 32±2c (continued on next page) 107

Table 3. Phenolic compounds contents, antioxidant activity, instrumental color and water activity of freeze-dried and oven-dried JPS.1 Phenolic compounds (mg/100g dwb2) Freeze-dried 55 oC/14 h 55 oC/22 h 65 oC/18 h 75 oC/14 h 75 oC/22 h Hydroxycinnamic acid derivatives trans-Cinnamic acida 0.3±0.0 ND ND ND ND ND Anthocyanins Delphinidin-3-O-glucosidea 70±0a 21±0c 25±2c 22±2c 46±3b 24±2c Cyanidin-3-O-glucosidea 904±10a 335±12c 283±8d 337±21c 442±26b 253±24d Total 974±11a 355±12c 308±9cd 359±23c 488±29b 277±26d Total phenolic compounds 3845±57a 2883±118b 3018±178b 2661±195b 3898±159a 3819±166a Antioxidant activity FRAP (mmol Fe2+/100 g dwb) 109±7a 48±1c 44±1c 51±3c 76±2b 54±0c TEAC (mmol Trolox/100 g dwb) 45±5b 54±2a 34±0c 46±2b 52±0ab 25±1c Instrumental color L* 50.11±0.00a 38.67±0.01c 38.94±0.01b 35.96±0.01d 34.99±0.00e 35.91±0.01d a* 18.41±0.02a 10.98±0.01f 12.49±0.02e 13.93±0.46d 15.61±0.01b 14.43±0.04cd b* 4.15±0.02e 6.03±0.02a 5.00±0.00c 5.26±0.07b 4.97±0.02c 4.78±0.02d E* 13.77 12.67 14.89 15.40 14.76 Water activity 0,140e 0,620a 0,485b 0,237c 0,195d 0,127e 1Results expressed as mean±SD for triplicates. *JPS – jabuticaba peel and seeds. 2 DWB - dry weight basis. 3ND - not detected. Gallic acid, ellagic acid derivatives and ellagitannins were determined by HPLC-DAD-MS-Q and identified according to the information described in Table 2. The other compounds were determined by HPLC-DAD and identified by comparison with retention time and absorption spectrum of the respective standards. Quantification was carried out using calibration curves of arespective standards, bellagic acid and cvescalagin. Different superscript letters in the line indicate significant difference between mean values (One-way ANOVA followed by Tukey multiple comparison post hoc test, p<0.05). 108

Oven-drying factors affected only cyanidin-3-O-glucoside and vescalagin contents, which are major phenolic compounds in JPS, in addition to antioxidant activity measured by TEAC. Vescalagin was positively affected by temperature, while cyanidin-3-O-glucoside and TEAC values were negatively affected by time (Figure 2). Similar to our results, Samoticha et al (2016) reported that degradation of chokeberry phenolic compounds was more associated with drying time than temperature. On the other hand, as observed in the present study, Mphahlele et al. (2016) reported that dehydration of pomegranate peel at higher temperatures (60 °C) increased its ellagitannins contents in comparison to both freeze-drying and oven- drying at lower dehydration temperatures (40 °C and 50 °C). These results suggest that oven- drying at higher temperatures may release the insoluble ellagitannins from the food matrix.

109

Figure 2. Pareto charts and surface response charts of oven-drying conditions that affected the phenolic composition and antioxidant activity of jabuticaba peel and seeds according to experimental design results. In these Pareto charts, the standardized effects that reached the dotted line (p = 0.05) are statistically significant. Temp (°C) = temperature; (°C) by (min) = interaction between temperature and time of processing. 110

In comparison to the freeze-dried sample, oven-drying at temperatures of up to 65 °C resulted in an average decrease of 26% of total phenolic compounds contents (Table 3). The decrease of phenolic compounds contents at 55 °C was mainly due to a decrease in anthocyanins (66%, on average) and ellagic acid derivatives (27%, on average). Oven drying at 65 °C also resulted in a significantly decrease of ellagitannins (20%) (Table 3). The temperature and oxygen exposure associated to oven-drying leads to oxidation of phenolic compounds. Similar to our results, some studies show that dehydration of different berries by hot air (50 °C to 90 °C) leads to decreases from 14% to 56% of total phenolic compounds contents, in comparison to freeze-drying (Michalska et al., 2017; Michalska et al., 2016; Samoticha et al., 2016; Wojdylo et al., 2014; De Torres et al., 2010; Wojdylo et al., 2009). Many authors report that anthocyanins are the class of phenolic compounds most sensitive to degradation during oven- drying (Michalska et al., 2016; Wojdylo et al., 2014; De Torres et al., 2010; Wojdylo et al., 2009). In fact, Zhou et al. (2017) observed that vacuum drying of mulberry led to lower degradation of anthocyanins when compared to convective drying, corroborating this hypothesis. Although drying at 75 °C did not change total phenolic compounds contents in comparison to freeze-drying, the phenolic compounds profile profoundly modified: anthocyanins and ellagic acid derivatives contents decreased by up to 68% and 32%, respectively, whereas ellagitannins and gallic acid contents increased by up to 29% and 214%, respectively (Table 3). Therefore, if we compare the most extreme oven-drying condition (75 °C for 22 h) with freeze-drying, samples showed similar total phenolic compounds contents but completely different relative contents of ellagitannins and anthocyanins (Figure 3). Freeze- dried JPS was richer in anthocyanins (26%), while oven-dried JPS was richer in ellagitannins (85%).

111

4 0 0 0 E A derivatives A n tho cy an ins

3 0 0 0 Ellagitannins

)

s

b

(

g

0 2 0 0 0

0

1

/

g m 1 0 0 0

0 F re e ze drie d 7 5 °C /2 2 h Figure 3. Comparing the extremes of dehydration conditions (freeze- and oven-drying at 75 °C/22 h) on the main classes of phenolic compounds from jabuticaba peel and seeds. Oven-drying at 75 °C/22 h did not affect total polyphenols contents (Student’s unpaired t-test, p < 0.05), but modified phenolics profile.

In comparison to the freeze-dried sample, AA measured by TEAC increased (20%), decreased (up to 44%) or did not change depending on the process conditions, whereas AA measured by FRAP decreased (up to 60%) at all tested conditions (Table 3). Similar to our results, studies reported a decrease of up to 63.5% of FRAP and TEAC values after dehydration of berries at different temperatures of up to 70 °C (Michalska et al., 2016; Samoticha et al., 2016; Wojdylo et al., 2014). Moreover, TEAC values did not change when black currant pomace was oven drying for temperatures of up to 90 °C (Michalska et al., 2017). In dried samples, TEAC values did not correlate with any of the analyzed phenolic compounds, whereas FRAP values positively correlated to total anthocyanins (r > 0.95, p = 0.001) and total ellagic acid derivatives (r > 0.89, p = 0.007). Correlations were also observed between FRAP values and individual phenolic compounds, such as cyanidin-3-O-glucoside (r > 0.95, p = 0.001), delphinidin-3-O-glucoside (r > 0.97, p < 0.001) and ellagic acid pentoside isomers (r > 0.80, p = 0.03). These results corroborate the hypothesis from a previous work from our group (Inada et al., 2015) that FRAP assay is more sensitive than TEAC assay to evaluate the AA of phenolic compounds from jabuticaba. The differences among the AA assays can be justified by their mechanisms of action. While FRAP assay is based on SET mechanism (single electron transfer) and is not capable to detect HAT (hydrogen atom transfer) -acting antioxidant compounds, TEAC assay is based both on SET and HAT mechanisms and is able to detect a larger range of compounds (Prior et al., 2005). Besides, as HAT-based assays 112

involves a kinect approach, it allows the estimation of trapping capacity of most reactive compounds. On the other hand, SET is a thermodynamic measure and therefore, provide an overall picture of the effective efficiency of oxidation/reduction of all antioxidants present in the sample, including the ‘slow-acting’ ones, which are not detected by kinetic methods (Prior et al., 2005). Oven-drying significantly affected the instrumental color of JPS (Table 3), in comparison to the freeze-dried sample, leading to a visible color change (ΔE* > 12.7). Luminosity (L*) decreased at all oven-drying conditions by up to 30%, a* values decreased by up to 40% and b* values increased by up to 45%. Therefore, when compared to freeze-drying, oven-drying led to darker and less red and less blue samples. The reduction of red and blue colors is probably a consequence of anthocyanins degradation during oven-drying was expected, since these pigments are known to provide purple, red and blue colors in several plants. In fact, cyanidin-3-O-glucoside and delphinidin-3-O-glucoside contents positively correlated with a* values (r > 0.81, p ≤ 0.027), while delphinidin-3-O-glucoside contents negatively correlated with b* values (r = - 0.75, p = 0.05). Although we also observed a correlation between anthocyanins contents and L* values (r > 0.77, p ≤ 0.041), the darker color of oven-dried samples may not only be explained by the degradation of these pigments. The darker color developed during heating is more probably associated with non-enzymatic browning reactions, mainly the Maillard reaction. The products formed during browning reactions may also help explaining the lack of correlation between TEAC values and phenolic compounds in JPS dried samples. One may argue that these molecules, which show antioxidant activity, could have influenced the response of the TEAC assay. Similar to our results, some studies also reported a decrease in luminosity (L*) from 7.6 % to 17.7%, in oven dried plum and camu-camu (50, to 80 °C) (Michalska et al., 2016; De Azevêdo et al., 2014), as well a decrease in a* values, from 36 % to 79 %, after dehydration of chockeberry and sour cherries (50 °C to 70 °C), in comparison to freeze drying (Samoticha et al., 2016; Wojdylo et al., 2014). Considering their water activity, all dried JPS powders showed potentially long shelf- life ( 0.62, Table 3). Water activity is a critical factor that affects food stability, since lower values reduces microorganism’s growth, as well as chemical and enzymatic reactions occurrence. According to Defraeye (2017), to ensure stability of dried fruits, their dehydration must be carried out until the water activity reaches a value of 0.6. Increasing oven drying conditions of temperature and time led to a gradual reduction of this parameter (Table 3), reaching 0.127 after oven drying at 75 °C for 22 h, which in turn was not statistically different from the freeze-dried sample (0.140). 113

3.4 Analysis of freeze-dried jabuticaba peel and seed after phloroglucinolysis indicates a diverse profile of proanthocyanidins

Proanthocyanidins are oligomers or polymers composed by flavan-3-ol subunits. The wide variety of proanthocyanidins molecules found in nature is due to the high variability in their chemical structure and number of flavan-3-ol units, as well as to differences in the location and stereochemistry of the inter-flavonoid bonds (Karonen et al., 2007). For this reason, the analysis of polymeric proanthocyanidins is extremely difficult – the large number of positional and optical isomers leads to compounds coelution and therefore many unresolved peaks in chromatographic analysis (Lazarus et al., 1999). In addition, with increasing polymerization degree, ionization efficiency decreases, making it difficult to detect compounds and interpret their MS spectrum (Karonen et al., 2007). Therefore, Kennedy & Jones (2001) developed a method to analyze proanthocyanidins following acid-catalyzed cleavage in the presence of phloroglucinol. In this method, proanthocyanidins are depolymerized, releasing their flavan-3- ols monomers, as well as their extension subunits, which are trapped by the nucleophilic phloroglucinol to generate analyzable and stable flavan-3-ol-phloroglucinol adducts. This is the first study that performed the characterization of flavan-3-ols monomers in jabuticaba following proanthocyanidins phloroglucinolysis. Despite the identification of proanthocyanidins performed in the present study in freeze-dried JPS (Table 4; Annex 1), their chromatographic coelution with anthocyanins and ellagitannins hindered their quantification and the determination of their polymerization degree.

114

Table 4. Identification of flavan-3-ols monomers and its phloroglucinol adducts in freeze-dried jabuticaba peel and seed by HPLC-ESI-MS-MS after proanthocyanidins phloroglucinolysis. Retention MS1 Compound MS2 fragments time (min) [M – H]- Gallocatechin adduct 5.3 429 124, 176, 260, 302 Catechin adduct 5.5 413 124, 176, 260, 302 Gallocatechin 7.0 305 124, 164, 178, 218, 260 Epigallocatechin gallate adduct 8.2 581 174, 260, 302, 318, 410, 428 Epicatechin adduct 8.5 413 124, 160, 174, 216, 260, 286 Catechin 11.5 289 178, 204, 244, 270 Epicatechin gallate adduct 13.2 565 244, 286, 394, 412 Epicatechin gallate 21.4 441 168, 288, 330 Full-scan mass spectra were acquired between 50–1200 m/z. Main MS2 fragments are shown in bold.

Three flavan-3-ol monomers (catechin, gallocatechin and epicatechin gallate) were identified as terminal subunits and five phloroglucinol adducts (catechin, gallocatechin, epicatechin, epicatechin gallate and epigallocatechin gallate) were detected as extension subunits (Table 4). The presence of different flavan-3-ols as terminal and extension subunits indicates a high diversity of proanthocyanidins in jabuticaba. The presence of catechin, epicatechin and epicatechin gallate as extension subunits (Table 4) suggests the occurrence of procyanidins, while the presence of gallocatechin and epigallocatechin gallate as extension subunits suggest the presence of prodelphinidins. Procyanidins and prodelphinidins are the two major proanthocyanidins classes found in foods (Lazarus et al., 1999). Corroborating these results, procyanidins B1 e B2 were found in jabuticaba pulp (Dantas et al., 2019) and procyanidin A2 was detected in jabuticaba peel (Da Silva et al., 2017). In addition, Alezandro et al (2013) reported that jabuticaba seed and peel contained approximately 13.0 and 7.0 mg/100 g of total proanthocyanidins and that these molecules were mainly present as monomers, dimers and polymers.

4. Conclusion

Contrary to our expectations, HHP processing was not effective in increasing total phenolic compounds contents and antioxidant activity of jabuticaba peel and seed. Freeze- drying and oven-drying yielded powders with distinct polyphenolic profile, the former richer 115

in anthocyanins and the latter richer in ellagitannins. Thus, considering that freeze-drying is a high cost method in comparison to oven-drying, the choice between these drying methods to obtain these functional ingredients depends on the purpose of the final product. If the intention is to use this ingredient as a natural food colorant (e.g. to produce biscuits), freeze-drying would be best choice. On the other hand, if there is no need to preserve anthocyanins and the main goal is a final product with high ellagitannins contents, oven-drying at 75 °C would be a better option. Considering that jabuticaba peel and seed showed a very rich and diverse phenolic compounds profile, including phenolic acids, flavonoids, ellagitannins and proanthocyanidins, their metabolism and potential bioactivity in humans merits further studies.

Acknowledgments

The financial support of Brazilian funding agencies (FAPERJ, CNPq and CAPES) is greatly acknowledged.

116

CHAPTER 2:

Metabolism of ellagitannins from jabuticaba (Myrciaria jaboticaba) is similar to that of other food sources and urolithin metabotype distribution of Brazilian individuals

resembles that of European populations.

Manuscript to be published in Journal of Functional Foods

117

1. Introduction

The beneficial effects associated with fruits and vegetables consumption are in part due to the presence of phenolic compounds, which have been described to present protective effects against non-communicable chronic diseases such as cancer, cardiovascular disease and diabetes in in vitro and in vivo studies (García-Niño & Zazueta, 2015; Kruger et al., 2014; Strathearn et al., 2014; Sancho & Pastore, 2012; Punithavathi et al. 2011; Wang & Stoner, 2008). However, an increasing number of clinical trials have reported limited and controversial results on the effects of phenolic compounds on human health (Sahebkar et al., 2016a; Sahebkar et al., 2016b; Schar et al., 2015; Vlachojannis et al., 2015). Among the factors that can lead to controversial results, the high metabolic variability between individuals stands out, which in turn has been mainly associated with differences in gut microbiota composition (Espín et al., 2017; González- Sarrías et al., 2017; Manach et al., 2017; Tomás-Barberán et al., 2016). The large interindividual variability in phenolic compounds metabolism resulted in the classification of individuals in metabotypes, according to their ability to produce specific microbial metabolites (Espín et al, 2017; Tomás-Barberán et al., 2016). For soy isoflavones, for example, the classification of individuals into equol producers and non-producers has already been described in several studies with different populations (Hazim et al., 2016; Tomás- Barberán et al., 2016; Atkinson et al., 2005; Frankenfeld et al., 2005; Frankenfeld et al., 2004). However, for other phenolic compounds, this classification is still preliminary or non-existent. Recently, three ellagitannin metabotypes were described: metabotype A (producers of urolithin- A), metabotype B (producers of urolithin-A, isourolithin-A and urolithin-B) and metabotype 0 (non-producers of these urolithins) (Tomás-Barberán et al., 2014). Some studies have reported that the distribution of urolithins metabotypes in the population is associated with gut microbiota composition and may also be affected by age, nutritional status and presence of diseases associated with dysbiosis such as metabolic syndrome and colorectal cancer (Cortés- Martín et al., 2018; Selma et al., 2018; González-Sarrías et al., 2017; Selma et al., 2016; Tomás- Barberán et al., 2014). It is noteworthy that urolithin metabotypes were only studied in European populations so far (Cortés-Martín et al., 2018; Selma et al., 2018; González-Sarrías et al., 2017; Selma et al., 2016; Tomás-Barberán et al., 2014), and that no data were found in the literature regarding the Brazilian population. Some clinical studies demonstrate that stratification of subjects into metabotypes helps explaining the controversial biological effects found in dietary interventions with phenolic compounds from food sources. It has already been reported that equol-producers respond better 118

to soybean and/or isoflavones beneficial effects when compared to non-producers (Kreijkamp- Kaspers et al., 2005; Zhang et al., 2012; Liu et al., 2014; Hazim et al., 2016). The relation between urolithin metabotypes and ellagitannins bioactivity was recently described by González-Sarrías et al. (2017), who observed that the chronic consumption of a pomegranate extract (rich in ellagitannins) by obese humans promoted a hypolipidemic effect only in metabotype B individuals, demonstrating that this metabotype seems to respond better to ellagitannins benefits. The results of these studies may partly explain the absence of phenolic compounds health claims and demonstrate the relevance of considering the metabotype when investigating foods bioactivity. Among the foods described in the literature for presenting high contents of ellagitannins and ellagic acid , jabuticaba (Myrciaria jaboticaba), a Brazilian berry, shows great marketing potential due to its high productivity and desirable flavor (Pereira et al., 2017; Plaza et al., 2016; Inada et al., 2015; Wu et al., 2012). The fruit presents a whitish pulp with sweet and slightly tange taste, containing from one to four seeds, covered by a dark purple to black peel (Inada et al., 2015). Jabuticaba peel and seed are usually not consumed and together represent about 40% of fruit weight. However, as these fractions present high fiber and phenolic compounds contents (Neves et al., 2018; Pereira et al., 2017; Inada et al., 2015), their use as functional ingredients would be of interest by food and pharmaceutical industries (Inada et al., 2015). In fact, several studies have reported the beneficial effects of jabuticaba and its extracts in in vitro (Wang et al., 2014; Leite-Legatti et al., 2012; Reynertson et al., 2006) and animal models (Moura et al., 2018; Batista et al., 2017; Araújo et al., 2014; Dragano et al., 2013). More recently, the effect of acute consumption of jabuticaba on glucose metabolism and plasma antioxidant activity was investigated in a clinical trial (Plaza et al., 2016). However, to the best of our knowledge, the metabolism of jabuticaba phenolic compounds, specifically ellagitannins and ellagic acid, in humans has not been yet investigated. In this context, the present study aimed at investigating the metabolism of ellagitannins from jabuticaba in Brazilian adults. In addition, the distribution of urolithin metabotypes according to nutritional status was evaluated and compared to other populations.

119

2. Material and methods

2.1 Standards and chemicals

Standards of urolithins (urolithin-A, urolithin-B, urolithin-M6, urolithin-M7, isourolithin-A, urolithin-A 3-glucuronide, urolithin-A 8-glucuronide, isourolithin-A 9- glucuronide, isourolithin-A 3-glucuronide and urolithin-B glucuronide) were previously synthetized (García-Villalba et al., 2016) and were available in the Food and Healthy laboratory at CEBAS-CSIC, Murcia, Spain. Urolithin-D and urolithin-C were purchased from Dalton Pharma Services (Toronto, Canada). Urolithin-A sulfate and urolithin-B sulfate were obtained as described by González-Sarrías et al. (2013). Standards of ellagic acid, gallic acid and the internal standard 6,7-dihydroxycoumarin were purchased from Sigma-Aldrich (St. Louis, Mo, USA). Authentic standards of vescalagin and castalagin were kindly provided by Prof S. Quideau, Univ. Bordeaux, France. All solvents were HPLC grade. HPLC grade water (Milli-Q system, Millipore, Bedford, MA, USA) was used throughout the experiments.

2.2 Production of jabuticaba powder

Jabuticaba fruits (Myrciaria jaboticaba, cv. Sabará) were purchased at Rio de Janeiro’s agricultural trading center. Fruits were selected, washed and sanitized in sodium hypochlorite 100 ppm solution for 15 min. Jabuticaba peel and seed were separated in a Bonina® 0.25 df horizontal depulper (NPC Equipamentos, Itabuna, Brazil), subsequently packed in aseptic polyethylene bags and stored at -20 °C. Jabuticaba peel and seed were thawed, freeze-dried (Tecnal®, Piracicaba, Brazil) at -98 °C and 0.025 mbar for 72 h, and milled to obtain jabuticaba peel and seed powder (JPSP).

2.3 Ellagitannins, ellagic acid derivatives and gallic acid analysis by HPLC-DAD-MS

For soluble phenolic compounds extraction, 150 mg of JPSP was extracted, in triplicate, with 10 mL of 70% methanol for 1 min, and centrifuged (4,696 g, 10 min, 20 °C). The supernatant was collected and filtered through a 0.22 µm PVDF filter (Millipore) (García- Villalba et al., 2015). For total phenolic compounds extraction, which comprises soluble and insoluble polyphenols, 5 mL of 4 M HCl were added to 50 mg of JPSP, in triplicate, vortexed for 1 min 120

and incubated at 90 °C for 4 hours. After incubation, samples were cooled to room temperature and pH was adjusted to 2.5 with 10 M NaOH solution. Then, samples were centrifuged (4,696 g, 10 min, 20 °C), the supernatant was collected, the volume was adjusted to 10 mL with Milli-Q water, and filtered through a 0.22 µm PVDF filter (Millipore). The pellet was redissolved with 10 mL of methanol:DMSO (50:50, v/v), vortexed for 2 min and extracted in an ultrasound bath for 15 min. After centrifugation (4,696 g, 10 min, 20 °C), the supernatant was collected and filtered through a 0.22 µm PVDF filter (Millipore). The liquid chromatography system was an Agilent 1200 HPLC equipped with a diode array detector and a single quadrupole (6120 Quadrupole, Agilent Technologies, Santa Clara, CA, USA). Chromatographic separation of compounds was achieved using a Poroshell 120 EC- C18 column (2.7 μm, 100 mm × 3 mm, Agilent Technologies) operating at 25 °C. The mobile phase consisted of a gradient of 1% aqueous formic acid (eluent A) and acetonitrile (eluent B), at a flow rate of 0.5 mL/min. Prior to injection, the column was equilibrated with 5% B. After injection of sample, solvent B composition increased to 18% B at 7 min, 28% B at 17 min, 50% at 22 min and 90% at 27 min which was maintained up to 28 min. The initial conditions were re-established at 29 min and kept under isocratic conditions up to 35 min. The injection volume was 5 μL. Ellagitannins and ellagic acid derivatives were monitored at 280 nm and 360 nm, respectively. Optimal ESI-MS parameters using nitrogen as nebulizer gas were: capillary voltage of 3,500 V; drying gas flow of 10 L/min; nebulizer pressure of 40 psi, drying temperature of 300 oC. MS spectra were acquired in negative ionization mode and measured in selective ion monitoring (SIM) mode. Identification of all compounds was carried out by their typical spectral properties and molecular masses. Gallic acid, ellagic acid, vescalagin and castalagin were also identified by the retention time of their respective standards. Quantification of gallic acid, ellagic acid, vescalagin and castalagin was carried out using calibration curves of respective standards. The other ellagic acid derivatives and ellagitannins were quantified using calibration curves of ellagic acid and vescalagin, respectively.

2.4 Clinical study design

This project was approved by the research ethics committee of Clementino Fraga Filho Hospital from Federal University of Rio de Janeiro, Brazil (number 1.343.292). The volunteers signed an informed consent form (Annex 2). Inclusion criteria were age between 20 and 45 years, body mass index (BMI) between 18.5 and 25.0 kg/m2 (normoweight) and higher than 27 kg/m2 (overweight and obese), and 121

regular bowel function (from one evacuation each two days to two daily evacuations). Exclusion criteria were chronic and inflammatory bowel diseases, use of probiotics and antibiotics in the three months prior to the study, regular use of drugs and/or nutritional supplements, smoking, alcoholism, pregnant or lactating women, and allergy to artificial strawberry gelatin. Two hundred and sixty-five participants completed an on-line questionnaire, from which 102 were eligible. Of those, 67 were excluded and 35 completed the study. Of excluded participants, 63 gave up participating, 1 did not accomplished the phenolic compounds restriction, 1 did not like the gelatin taste and the other 2 felt nauseated after its consumption. Volunteers were instructed, both verbally and in writing (Annex 3), to restrict ellagitannins food sources, as berries and pomegranate and its derived products in the 96 hours prior to the study. Volunteers were also instructed to start a total restriction of phenolic compounds food sources (i.e. fruits, vegetables, whole grains, oleaginous fruits, cacao, coffee and beer) in the 48 hours prior to the study and until the end of the 48 h urine collection. Dietary records of the phenolic compounds restriction period were obtained from the volunteers. The study consisted in the acute consumption of JPSP using a gelatin dessert as vehicle. A dessert containing 20 g of JPSP was prepared. One gram of carboxymethylcellulose was dissolved in 100 mL of warm water (~50 °C). After its complete dissolution, 20 g of artificial strawberry gelatin and 4 g of unflavored gelatin were dissolved. Finally, JPSP was carefully incorporated into the preparation. After 10-hour overnight fasting, volunteers were instructed to consume the jabuticaba dessert in up to 15 min along with 150 mL of water. In the following eight hours, volunteers consumed at every two hours, standardized meals consisting of two slices of white bread, 20 g of processed cheese and 150 mL of water. Urine samples were collected at baseline (total urine from the previous night until the consumption of jabuticaba dessert) and at 0-4, 4-8, 8-12, 12-24, 24-36 and 36-48 hours intervals after consumption. Urine volume at each collection interval was measured. Urine samples were freeze-dried until analysis.

2.5 Urolithin metabolites analysis by HPLC-DAD-MS

Before analysis, urine samples were reconstituted with Milli-Q water, vortexed, centrifuged at 14,324 x g for 10 min at 10 °C (Sigma 1-16K, Sigma Laborzentrifugen, Osterode an Harz, German) and filtered through a 0.22 µm PVDF filter (Millipore). 122

The liquid chromatography system was an Agilent 1200 HPLC equipped with a diode array detector and a single quadrupole (6120 Quadrupole, Agilent Technologies, Santa Clara, CA, USA). Chromatographic separation of urolithins was achieved using a Poroshell 120 EC- C18 column (2.7 μm, 100 mm × 3 mm, Agilent Technologies) operating at 25 °C, according to García-Villalba et al. (2016). The mobile phase consisted of a gradient of 0.5% aqueous formic acid (eluent A) and acetonitrile (eluent B), at a flow rate of 0.5 mL/min. Prior to injection, the column was equilibrated with 5% B. After injection of sample, the solvent B composition increased to 18% B at 7 min, 28% B at 17 min, 50% at 22 min and 90% at 27 min which was maintained up to 28 min. The initial conditions were re-established at 29 min and kept under isocratic conditions up to 33 min. The injection volume was 5 μL. Urolithins were detected and quantified by DAD at 305 nm. Optimal ESI-MS parameters using nitrogen as nebulizer gas were: capillary voltage 3500 V; drying gas flow 10 L min−1; nebulizer pressure 45 psi, drying T 300 oC. MS spectra were acquired in negative ionization mode and measured in selective ion monitoring (SIM) mode. Identification of all compounds was carried out by their spectral properties and molecular masses (García-Villalba et al., 2016).

2.6 Statistical analysis

Data were expressed as mean ± standard deviation. Since data did not present a normal distribution (Shapiro-Wilk and Kolmogorov-Smirnov tests), non-parametric statistical tests were used. The mean excretion of urinary metabolites among different intervals was compared using Friedman test followed by Dunn’s multiple comparisons post-hoc test. The mean excretion of urinary metabolites among different volunteers groups (normoweight versus overweight/obese; women versus men; without diarrhea versus with diarrhea) was compared using Mann-Whitney test. Statistical analyses were performed using GraphPad Prism software for Windows, version 5.04 (GraphPad Software, San Diego, CA, USA). Results were considered significant when p < 0.05.

123

3. Results and Discussion

3.1 Ellagitannins, ellagic acid derivatives and gallic acid are mainly found in the insoluble fraction of jabuticaba peel and seed powder

Phenolic compounds can be found in plant foods as soluble and insoluble forms. The insoluble compounds are covalently bound to plant cell wall components or to macromolecules, not being extracted by organic solvents (Acosta-Estrada et al., 2014). Therefore, insoluble phenolic compounds are usually underestimated, as their quantification is difficult (Arranz et al., 2009). Considering that the correct quantification of ellagitannins and ellagic derivatives, including their insoluble forms, is important to study the metabolism of these compounds in the human body, two extractions were performed. One of them was an acid hydrolysis to release these insoluble forms from the food matrix. The other extraction used an organic solvent to quantify the soluble forms of these phenolic compounds. In summary, the soluble forms of gallic acid, ellagic acid derivatives and ellagitannins of JPSP were quantified after extraction with organic solvent, whereas the soluble and insoluble forms of the same compounds were quantified together after acid hydrolysis. Therefore, the contents of insoluble forms were estimated by the difference between both extractions. One should note that individual compounds quantified in these extractions may not be directly compared, since the former promotes depolymerization of most ellagitannins into free ellagic acid and therefore does not allows determination of original compounds (García-Villalba et al., 2015). The contents of soluble and insoluble phenolic compounds in 20 g of JPSP, which was the amount offered to the volunteers is presented in Table 1. A total of 569.6 mg of soluble phenolic compounds was found, being ellagitannins the most abundant class (85%), followed by ellagic acid and its conjugates (13%) and gallic acid (2%). Castalagin, vescalagin and other non-identified ellagitannins (quantified with the castalagin standard) comprised 90% of ellagitannins; free ellagic acid corresponded to 84% of ellagic acid derivatives. This phenolic profile was similar to that described in the literature by different authors for jabuticaba (Plaza et al., 2016; Pereira et al., 2017; Neves et al., 2018). The sum of soluble and insoluble phenolic compounds in JPSP was 1,492.7 mg, being ellagic acid derivatives the most abundant class (69%), followed by ellagitannins (16%) and gallic acid (15%). Ellagic acid corresponded to 71% of ellagic acid derivatives, followed by valoneic acid dilactone isomers (28%); the only ellagitannins were di-hexahydroxydiphenoyl (HHDP)-glucose isomers. Insoluble compounds contents were 923.1 mg, which corresponded 62% of the total phenolic compounds (Table 1). 124

Differently from the present study, García-Villalba et al. (2015) observed that most pomegranate ellagitannins and ellagic acid derivatives were soluble (73%). Although insoluble phenolic compounds are considered to be less bioavailable (Acosta-Estrada et al., 2014), they may be metabolized into highly bioactive metabolites by gut microbiota and absorbed locally.

Table 1. Soluble and insoluble phenolic compounds contents1 in the amount of jabuticaba peel and seed powder (20 g) consumed by the volunteers. Phenolic compounds Content (mg/20 g) Soluble phenolics2 Gallic acid 10.0 ± 0.0 Ellagic acid 62.8 ± 1.2 Valoneic acid dilactone 6.4 ± 0.1 Ellagic acid pentoside isomers 6.2 ± 0.2 Vescalagin 148.1 ± 2.3 Castalagin 179.6 ± 1.0 Other unknown ellagitannins (quantified 108.6 ± 2.9 as castagalin) Di-HHDP-glucose isomers 47.9 ± 1.3 Soluble + insoluble phenolics3 Gallic acid 224.1 ± 9.6 Ellagic acid 734.3 ± 19.3 Valoneic acid dilactone isomers 293.3 ± 12.0 Ellagic acid hexoside isomers 6.3 ± 0.3 Di-HHDP-glucose isomers 234.7 ± 8.0 Total soluble compounds 569.6 ± 9.0 Total soluble + insoluble compounds 1,492.7 ± 49.2 Total insoluble compounds4 923.1 ± 40.2 1Results expressed as mean ± SD for triplicates. Gallic acid, ellagic acid, vescalagin and castalagin were quantified using their authentic standards. Valoneic acid dilactone and ellagic acid pentoside isomers were quantified using ellagic acid standard. Vescalagin/castalagin isomers and di- hexahydroxydiphenoyl (HHDP)-glucose isomers were quantified using vescalagin standard. 2Soluble compounds were determined after sample extraction with methanol 70%. 3Soluble + insoluble compounds were determined after sample hydrolysis with HCl 4N and extraction with methanol:DMSO (50:50, v/v). 4Insoluble compounds were estimated as the difference between soluble + insoluble and soluble compounds contents.

125

3.2 Jabuticaba ellagitannins are mainly metabolized to urolithin glucuronides by gut microbiota

Thirty-five volunteers concluded this study: 19 were classified as normoweight (BMI 23±2 kg/m2), being 10 women and 9 men, with a mean age of 26 years; 16 were overweight or obese (BMI 34±6 kg/m2), being 7 women and 9 men, with a mean age of 29 years (Table 2). 22 volunteers (63%) reported the occurrence of diarrhea about 10 hours after JPSP consumption. Although according to reports of the Brazilian popular medicine, jabuticaba peel could be used to treat diarrhea (Borges et al., 2014), the laxative effect of its peel has been described by Araújo et al (2014). These authors reported that rats which consumed jabuticaba peel had increased fecal volume and more humid feces, associated with its high fiber contents. The dose administered in the present study (20 g of JPSP) contained approximately 7 g of fibers (Inada et al., 2015), which is below the daily intake recommendations for adults of 25 g for women and 38 g for men (Otten et al., 2006). Therefore, the fiber consumed by the volunteers through JPSP was probably not the cause of diarrhea occurrence, although it may have contributed to this side effect. A hypothesis that could explain this laxative effect may be associated with the presence of hydrolysable tannins, which have been recently suggested as natural substitutes of laxative medical drugs. In fact, Hsieh et al. (2016) reported that the ellagitannin strictinin, extracted from red tea, accelerated the intestinal transit of rats, and Kim et al. (2016) reported that a gallotannin-rich extract exhibited laxative effect in constipated rats. In this sense, the high dose of phenolic compounds administered in the present study (1,493 mg of ET, EA and gallic acid) in comparison to that of other studies that also investigated the metabolism of ellagitannins (5.4 to 514 mg) (Cortés-Martín et al., 2018; González-Sarrías et al 2015; Ludwig et al., 2015; Truchado et al., 2012; González-Barrio et al., 2010; Cerdá et al., 2005a), may have contributed to the occurrence of diarrhea by most volunteers.

126

Table 2. Characteristics of the volunteers participating in the study (n = 35) Weight status Gender BMI1 (kg/m2) Age (years) Female (n = 10) 21.7 ±2.0 25.3±4.9 Normoweight (n = 19) Male (n = 9) 23.9 ±0.7 26.2±3.5 All volunteers 22.7 ±1.8b 25.7±4.2a Female (n = 7) 35.2±7.2 29.4±6.3 Overweight/Obese (n = 16) Male (n = 9) 33.0±5.3 28.2±6.9 All volunteers 34.0±6.1a 29.2±6.4a 1Body mass index. Different superscript letters in the same column indicate significant differences between mean values of normoweight and overweight/obese volunteers (Mann Whitney test, p < 0.05).

At baseline, no ellagitannin metabolites were detected in the baseline urine of 34 volunteers, indicating that the 96-h dietary restriction was adequate to washout these compounds from the organism. One volunteer, however, excreted a very low amount (1.2 μmol) of urolithin-A 3-glucuronide in urine. According to its dietary record, 80 h before JPSP consumption this volunteer consumed nuts, a known source of ellagitannins, which metabolites can remain in the human body for up to 92 h (Truchado et al., 2012). In urine collected at the three following intervals (0-4, 4-8 and 8-12 h) no ellagitannin metabolites were detected, suggesting that all metabolites excreted thereafter by this volunteer derived from JPSP and not from any other food. Therefore, its excretion data were not excluded from this study. After JPSP consumption, 13 metabolites were detected in urine: urolithin (uro) A, isourolithin (isoUro) A, , urolithin C, urolithin M6, urolithin M7, urolithin A 3- glucuronide, urolithin A sulfate, isourolithin A 3-glucuronide, isourolithin A 9-glucuronide, urolithin B glucuronide, urolithin B sulfate and ellagic acid (Table 3). With the exception of uro-M6 and uro-M7, all other metabolites have already been reported in urine after consumption of ellagitannin food sources (González-Sarrías et al., 2015; Romo-Vaquero et al., 2015; Ludwig et al., 2015; García-Muñoz et al., 2014; Nuñez-Sanchez et al., 2014; Seeram et al., 2006). Until now, uro-M6 and uro-M7 have only been reported in animal feces (González-Barrio et al., 2012; González-Barrio et al., 2011b; Ito et al., 2008), human feces (Romo-Vaquero et al., 2015; García-Villalba et al., 2013) and human colon tissues (Nuñez-Sanchez et al 2014) and, to the best of our knowledge, this study is the first to report their presence in urine (Table 3). Uro-M6 and uro-M7 detection in urine samples is probably associated to the high dose of ellagitannins administered in this study in comparison to the other studies (Cortés-Martín et al., 2018; González-Sarrías et al 2015; Ludwig et al., 2015; Truchado et al., 2012; González-Barrio et al., 127

2010; Cerdá et al., 2005a), which probably resulted in higher production of these metabolites, leading to urinary concentrations above the analytical detection limit. After 48 h of JPSP consumption, uro-A 3-glucuronide was both the most abundant metabolite (74.9 µmol), representing 66% of total metabolites, and the most prevalent one (74%), being excreted by 26 volunteers (Table 3). In fact, the other three most abundant metabolites were all glucuronides: isoUro-A 3-glucuronide (16.0 µmol; n = 7), isoUro-A 9- glucuronide (7.7 µmol; n = 10) and uro-B glucuronide (6.0 µmol; n = 6). Even though uro-A was excreted at lower concentrations than glucuronides, it was the second most prevalent (69%) metabolite, being excreted by 24 volunteers. The higher concentrations of urolithin glucuronides when compared to their aglycones was expected (Romo-Vaquero et al., 2015; Nuñez-Sanchez et al 2014; Truchado et al., 2012; González-Barrio et al., 2010), since these compounds undergo conjugation reactions in the small intestine and liver, aimed at increasing their hydrophilic character to facilitate excretion (Manach et al., 2004).

128

Table 3. Urinary excretion of metabolites (µmol) at different intervals after acute consumption of jabuticaba peel and seed1 Compounds 0-4 h 4-8 h 8-12 h 12-24 h 24-36 h 36-48 h Total (0-48 h) 1.9±5.2b 21.2±33.1a 29.0±40.4a 22.7±37.3a 74.9±90.8 Uro-A 3-glucuronide ND ND (n = 11) (n = 22) (n = 24) (n = 22) (n = 26) 0.01±0.03a 0.03±0.19a 0.8±3.5a 0.4±1.7a 1.2±5.1 Uro-A sulfate ND ND (n = 1) (n = 1) (n = 3) (n = 2) (n = 6) 0.1±0.2b 0.4±0.7ab 2.9±6.7a 0.8±1.7ab 4.1±8.1 Uro-A ND ND (n = 5) (n = 10) (n = 20) (n = 11) (n = 24) 0.1±0.6a 3.4±11.8a 8.9±26.4a 4.0±14.8a 16.0±51.0 IsoUro-A 3-glucuronide ND ND (n = 1) (n = 7) (n = 6) (n = 5) (n = 7) 0.1±0.4a 1.6±4.8a 4.6±14.8a 1.4±5.7a 7.7±21.0 IsoUro-A 9-glucuronide ND ND (n = 3) (n = 9) (n = 6) (n = 7) (n = 10) 0.01±0.05a 0.2±0.7a 1.3±3.9a 0.4±1.4a 1.9±5.4 IsoUro-A ND ND (n = 2) (n = 3) (n = 8) (n = 3) (n = 8) 0.1±0.7a 0.4±1.5a 1.7±5.0a 3.7±13.5a 6.0±18.6 Uro-B glucuronide ND ND (n = 1) (n = 3) (n = 6) (n = 6) (n = 6) 0.1±0.8 0.1±0.8 Uro-B sulfate ND ND ND ND ND (n = 1) (n = 1) 0.05±0.21a 0.1±0.4a 0.2±0.6 Uro-B ND ND ND ND (n = 2) (n = 2) (n = 2) 0.01±0.05a 0.1±0.4a 0.3±0.9a 0.04±0.22a 0.5±1.2 Uro-C ND ND (n = 2) (n = 5) (n = 4) (n = 1) (n = 8) 0.01±0.03a 0.01±0.08a 0.02±0.09 Uro-M6 ND ND ND ND (n = 1) (n = 1) (n = 2) 0.04±0.22 0.04±0.22 Uro-M7 ND ND ND ND ND (n = 1) (n = 1) 0.06±0.11a 0.01±0.03b 0.07±0.12 Ellagic acid ND ND ND ND (n = 12) (n = 4) (n = 13) 1Results expressed as mean±SD. Different superscript letters in the line indicate significant difference between mean values of each compound contents excreted in different intervals (Friedman test followed by Dunn’s multiple comparison post hoc test, p<0.05). 129

Minor metabolites, both in terms of abundance and prevalence, were isoUro-A (1.9 µmol; n = 8), uro-A sulfate (1.2 µmol; n = 6), uro-C (0.5 µmol; n = 8), uro-B sulfate (0.1 µmol; n = 1), uro-B (0.2 µmol; n = 2), ellagic acid (0.07 µmol; n = 13), uro-M7 (0.04 µmol; n = 1) and uro-M6 (0.02 µmol; n = 2). Therefore, if on the one hand, urolithin glucuronides were major urinary metabolites, on the other hand, urolithin sulfates were minor metabolites (Table 3), which is in agreement with literature reports (Nuñez-Sanchez et al 2014; Espín et al., 2007). It has already been reported that the type of conjugation appears to vary according to phenolic compounds class and to the dose consumed (Manach et al., 2004). Thus, our results together with other reported in the literature (Nuñez-Sanchez et al., 2014; Espín et al., 2007), suggests that urolithins are preferentially conjugated to glucuronic acid, regardless of ellagitannins ingested dose and food source. The low urinary excretion of uro-M6, uro-M7 and uro-C, combined with their high concentrations reported in feces (Romo-Vaquero et al., 2015; Nuñez- Sanchez et al 2014; González-Barrio et al., 2012; González-Barrio et al., 2011b) suggests that these gut metabolites are poorly absorbed, probably due to their hydrophilicity (Tomás- Barberán et al., 2017; Espín et al., 2007). These compounds may undergo sequential dihydroxylations until the formation of the more lipophilic uro-A, uro-B and isoUro-A (Espín et al., 2007), which were the most abundant metabolites, as previously discussed. Ellagic acid was the only compound detected within 8 h of urine collection (Table 3). It should be noticed that among all metabolites, ellagic acid was the only compound already found in JPSP (Table 1). Moreover, ellagic acid reaches peak plasma concentrations within 1 to 2 h (González-Sarrías et al., 2015; Seeram et al., 2006), suggesting its rapid absorption, mainly in stomach (Espín et al., 2007). Among the other compounds detected in urine (Table 3), some were excreted only at isolated intervals (uro-B, uro-B sulfate, uro-M6 and uro-M7), whereas all other metabolites were excreted in all urine intervals between 8-48 h. The lowest excretions were observed at the intervals of 0-4 h (0.06 μmol), 4-8 h (0.01 μmol) and 8-12 h (2.3 μmol) (Figure 1). These results are in accordance with the literature, which reported the excretion of urolithins only 8 h after the consumption of ellagitannins’ food sources (Ludwig et al., 2015; Truchado et al., 2012; González-Barrio et al., 2010; Cerdá et al., 2005a). The highest excretions of metabolites were observed at the intervals of 12-24 h (27.1 μmol), 24-36 h (49.7 μmol) and 36-48 hours (33.5 μmol), which is supported by other studies that observed prolonged urinary excretion of urolithins after ellagitannins consumption, at intervals of up to 92 h (Ludwig et al., 2015; Truchado et al., 2012; González-Barrio et al., 2010; Seeram et al., 2006; Cerdá et al., 2005a). Moreover, considering that Truchado et al. (2012) observed that more than 50% of total urolithins were excreted after 48 h of ellagitannins consumption, and 130

that a high concentration of metabolites was detected in our last urine collection interval (36- 48 h) (Figure 1), it is expected that urolithins would be still detected after this collection.

Figure 1. Total urinary excretion (µmol) of metabolites at different intervals after acute intake of jabuticaba peel and seeds. Results expressed as mean±SD. Different letters indicate significant difference between mean values of total amounts of metabolites excreted in different intervals (Friedman test followed by Dunn’s multiple comparison post hoc test, p<0.05).

The long time required for the urinary excretion of urolithins is associated with their metabolization from ellagitannins and ellagic acid by the gut microbiota (Tomás-Barberán et al., 2017; Selma et al., 2014; García-Villalba et al., 2013). This association was directly investigated by González-Barrio et al (2010), who observed that ileostomized individuals did not produce urolithins after consumption of ellagitannins. Additionally, the prolonged permanence of these compounds in the organism seems to be related to enterohepatic circulation (Espín et al., 2007).

3.3 Excretion of urolithins showed high interindividual variability and was affected by diarrhea occurrence, but not by gender or nutritional status

Twenty-nine volunteers (83%) excreted urolithins amounts ranging from 1.77 μmol to 522 μmol. Due to this high interindividual variability, volunteers were divided into tertiles, 131

according to urolithins excretion amount, resulting in their classification as low producers (1.77 μmol to 69.4 μmol), medium producers (> 69.4 μmol to 121.5 μmol) or high producers (> 121.5 μmol to 522 μmol). The remaining six volunteers (17%) did not excrete urolithins and therefore were classified as non-producers (Figure 2). The high interindividual variability of urolithins urinary excretion has already been reported in humans (García-Muñoz et al., 2014; Nuñez- Sanchez et al., 2014; Truchado et al., 2012; González-Sarrías et al., 2010; Cerdá et al., 2005a), and is probably due to differences in microbiota composition. In fact, this association was investigated by Cerdá et al. (2005b), who observed variability in urolithins production when punicalagin and ellagic acid were incubated with feces from six different volunteers. García- Villalba et al. (2013) reported similar results when ellagic acid was incubated with feces from two volunteers, which showed significant differences among specific bacterial groups. Urolithins excretion was not statistically affected by neither gender nor nutritional status (Figure 2A and B). However, volunteers who reported diarrhea episodes presented significantly lower urolithins excretion (67.2 μmol) than those who did not present this side effect (189.7 μmol) (Figure 2C). This suggested that diarrhea reduced contact time between gut microbiota and ellagitannins due to increased intestinal motility and consequent evacuation, probably resulting in lower production of these gut metabolites by these volunteers. The fact that five out of six non-producers reported diarrhea strengthens this hypothesis. Finally, one would expect a higher urolithins excretion if volunteers in general did not show this side effect.

132

Figure 2. Urolithins urinary excretion (µmol) by volunteers (n = 35) grouped according gender (A), nutritional status (B) and diarrhea occurrence (C). Volunteers were divided into tertiles, according to urolithins excretion amount, resulting in their classification as low producers (1.77 μmol to 69.4 μmol), medium producers (> 69.4 μmol to 121.5 μmol) or high producers (> 121.5 μmol to 522 μmol). Volunteers that did not excrete urolithins were classified as non-producers. The asterisk indicates statistical difference (Mann Whitney test, p < 0.05). NW = normoweight; OW/OB = overweight/obese.

133

3.4 Up to 36% of ellagitannins and ellagic acid metabolites were recovered in urine within 48 hours of jabuticaba peel and seed powder consumption

Based on the ellagitannins and ellagic acid contents present in the amount of JPSP consumed by volunteers (Table 1), the urinary recovery of their metabolites was calculated (Table 4). Although gallic acid was quantified in JPSP, this compound was not considered when calculating urinary recovery since it is not a precursor of urolithins. Therefore, considering total soluble and insoluble ellagitannins and ellagic acid derivatives (1,268 mg) (Table 1), the average urinary recovery was 4.2% (Table 4), ranging from 0.1% to 16%. Low producers showed urinary recovery of 1.0% (0.1% to 2.1%), medium producers 2.8% (2.3% to 3.3%) and high producers 8.7% (4.2 to 16.1%). These results were similar to those described in the literature after consumption of different doses (122 to 514 mg) of ellagitannins and ellagic acid present in raspberry (7.0% to 8.6%) (Ludwig et al., 2015; González-Barrio et al., 2010) and pomegranate juice (7%) (Yang et al., 2016), in urine collected for up to 56 h. However, our results were lower than those described after consumption of 62 mg of ellagic acid present in strawberry (up to 99%) (Truchado et al., 2012), which was associated to a longer time of urine collection (92 h). As previously stated, our volunteers showed high metabolites concentrations in the last urine collection interval (36-48 h) (Figure 1), suggesting that urolithins would be excreted in urine longer periods, which probably underestimated urinary recovery. Cerdá et al. (2005a) reported that ellagitannins from strawberries, raspberries, wine and walnuts showed urinary excretions of 2.8%, 3.4%, 6.5% and 16.6%, respectively, highlighting the influence of the food matrix on compounds release and consequently their bioavailability. Furthermore, urinary recoveries ranging from 14.8% to 23.5% were observed after the consumption of pomegranate extracts rich in ellagitannins and their derivatives (González- Sarrías et al., 2015), reinforcing the hypothesis of food matrix effect. Thus, the consumption of foods with higher contents of soluble ellagitannins would possibly lead to higher bioavailability of ellagitannins. Considering that 44% of jabuticaba ellagitannins and ellagic acid derivatives were soluble (Table 1), the average urinary recovery calculated based on this content (560.6 mg) was 9.5%, ranging from 0.1% to 36.5% (Table 4). We can assume that these results better reflect the urinary recovery of jabuticaba ellagitannins and ellagic acid derivatives already available for metabolization.

134

Table 4. Urinary recovery (%) of soluble and insoluble ellagitannins and ellagic metabolites stratified according to the level of urolithins produced.1 Urinary recovery (%) of soluble Urinary recovery (%) of Total metabolites excreted (mg) and insoluble precursors2 soluble precursors3 12.6±8.7 1.0±0.7 2.3±1.6 Low producers (n = 10) (0.7 – 26.6) (0.1 – 2.1) (0.1 – 4.7) 35.1±5.3 2.8±0.4 6.3±1.0 Medium producers (n = 9) (29.7 – 42.0) (2.4 – 3.3) (5.3 – 7.5) 110.3±49.8 8.8±4.0 19.7±8.9 High producers (n = 10) (53.8 – 204.6) (4.3 – 16.3) (9.6 – 36.5) 53.3±51.8 4.2±4.1 9.5±9.3 All volunteers (n = 29) (0.7 – 204.6) (0.1 – 16.3) (0.1 – 36.5) 1Results expressed as mean±SD for triplicates. The total amount of urolithins excreted was divided into tertiles, resulting in the classification of subjects as low producers (1.77 µmol – 69.40 µmol), medium producers (> 69.40 µmol – 121.50 µmol) and high producers (> 121.50 µmol – 522.06 µmol). 2Soluble + insoluble precursors include the ellagitannins and ellagic acid derivatives determined after sample hydrolysis with HCl and extraction with organic solvents, as described in material and methods.3Soluble compounds precursors include the ellagitannins and ellagic acid derivatives determined after sample extraction with organic solvents, as described in material and methods. The urinary recovery (%) was calculated as the ratio between the sum of all the metabolites detected and the total of soluble and insoluble ellagitannins and ellagic acid derivatives (1,259 mg) or the total of soluble ellagitannins and ellagic acid derivatives (560 mg). 135

Another factor that may affect urinary recovery of ellagitannins and ellagic acid derivatives is their chemical structure. Since ellagitannins need to be hydrolyzed to free ellagic acid prior to their metabolization into urolithins (Tomás-Barberán et al 2017), a higher urinary recovery after the consumption of ellagic acid (23.5%), in comparison to ellagitannins (14.5%) was observed by González-Sarrías et al. (2015). It is also important to note that vescalagin and castalagin, the main jabuticaba ellagitannins (Table 1), are C-glycoside ellagitannins, which C- C bonds are not hydrolyzed by microbial tannases (Tomás-Barberán et al., 2017). Thus, while the hydrolysis of most ellagitannins can lead to the formation of several molecules of free ellagic acid, vescalagin or castalagin would only release one molecule of free ellagic acid and one molecule of ellagic acid C-glycoside upon hydrolysis (Tomás-Barberán et al., 2017).

3.5 Distribution of metabotypes in Brazilians was similar to that of European populations, being affected by both nutritional status and diarrhea occurrence

Three metabotypes related to ellagitannins and ellagic acid metabolism have been recently described: metabotype A (producers of urolithin-A), metabotype B (producers of urolithin-A, isourolithin-A and urolithin-B) and metabotype 0 (non-producers of urolithins) (Tomás-Barberán et al., 2014). The relative distribution of these metabotypes has already been described in European populations, mainly from Spain (Cortés-Martín et al., 2018; Tomás- Barberán et al., 2014). To the best of our knowledge, non-European populations have not yet been studied. The present study is the first to investigate urolithin metabotypes in non-European individuals, specifically from the city of Rio de Janeiro, Brazil. The three aforementioned metabotypes were observed in our study (Figure 3; Annex 4 and 5), as expected, since they were observed in all clinical trials conducted in Europe, regardless of demographic characteristics (age, gender and BMI), health status and source or amount of ellagitannins or ellagic acid consumed (Tomás-Barberán et al., 2014). In spite of their existence, the relative distribution of these three metabotypes may be influenced by some of the factors mentioned above (Cortés-Martín et al., 2018; González-Sarrías et al., 2017; Selma et al., 2016; Tomás- Barberán et al., 2014).

136

Figure 3. Distribution of urolithins metabotypes in different groups of the study population. 137

Considering all volunteers of the present study (n= 35), 54.3% were classified as metabotype A, 28.6% as metabotype B and 17.1% as metabotype 0 (Figure 3A). The distribution of metabotypes between women (Figure 3B) and men (Figure 3C) did not differ, in accordance to the literature (Cortés-Martín et al., 2018). The prevalence of both metabotypes B and 0 was higher than that observed for healthy populations, which usually presents an average distribution of 20% and 10% of metabotypes B and 0, respectively (Tomás-Barberán et al., 2014). The higher prevalence of metabotype 0 is probably due to diarrhea occurrence, as previously discussed, since 83% of metabotype 0 volunteers reported this symptom. In fact, if the volunteers are stratified according to diarrhea occurrence, the relative distribution of metabotypes among those that did not report this symptom was 69.2%, 23.1% and 7.7% of metabotypes A, B and 0, respectively (Figure 3D), which is very similar to that described in the literature for healthy populations (Tomás-Barberán et al., 2014). Metabotypes distribution among individuals that reported diarrhea was 45.5%, 31.8% and 22.7% of A, B and 0, respectively (Figure 3E), reinforcing the impact of this side-effect on the metabolization of jabuticaba ellagitannins. When volunteers were stratified according to their nutritional status, normoweight individuals showed a higher prevalence of metabotype 0 (26.3%) and a lower prevalence of metabotype B (21.1%) (Figure 3F) in comparison to overweight and obese (6.3% and 37.5%, respectively) (Figure 3G). Metabotype A distribution was similar between normoweight (52.6%) and overweight and obese volunteers (56.3%). The higher prevalence of metabotype B in overweight and obese subjects (~ 30%) in comparison to normoweight (10% to 20%) has already been reported in the literature (Selma et al., 2016; Tomás-Barberán et al., 2014). As previously discussed, these differences in metabotypes distribution are probably due to differences in microbiota composition (Selma et al., 2016). In fact, in addition to obesity, a higher prevalence of metabotype B was also reported in individuals with dysbiosis (Selma et al., 2018; Tomás-Barberán et al., 2014), suggesting that this condition favors the growth of isourolithin-A- and urolithin-B-producing gut bacteria rather than urolithin-A-producing gut bacteria.

4. Conclusion

This is the first study of the ellagitannins metabolism in non-European individuals, specifically in Brazilians, through the consumption of JPSP. A high interindividual variability in terms of ellagitannins metabolism and urolithins excretion was observed, which may be 138

associated to differences in gut microbiota. Furthermore, this investigation made possible the description of urolithin metabotypes in non-European individuals, which resembled that of Europeans. This similarity was also observed in terms of a higher prevalence of metabotype B in overweight and obese individuals, in comparison to normoweight. However, despite cultural and ethnical differences between Europeans (mostly from Caucasian origin) and Brazilians (mixed ethnical background, from European, African and indigenous ancestors), our results suggest that their gut microbiota composition may be similar in terms of ellagitannins metabolism. The high dose of ellagitannins consumed through jabuticaba peel and seed powder probably caused the occurrence of diarrhea in 63% of volunteers, which led to a reduction in urolithins urinary excretion and ellagitannins urinary recovery, as well as an increase in the prevalence of metabotypes B and 0 in the volunteers. This unexpected side-effect reinforces the pivotal role of gut microbiota on the metabolization of ellagitannins. These results concerning both the metabolism of ellagitannins from jabuticaba and the distribution of urolithin metabotypes in Brazilian individuals are crucial for future studies on the bioactivity of jabuticaba or of other ellagitannins food sources in the Brazilian population.

Acknowledgments The financial support of Brazilian funding agencies (FAPERJ, CNPq and CAPES) is greatly acknowledged.

139

THESIS GENERAL CONCLUSION

Considering the effects of technological processes on polyphenols contents of jabuticaba peel and seed, high hydrostatic pressure was not effective in increasing total phenolic compounds contents, while freeze-drying and oven-drying yielded powders with distinct polyphenolic profile, the former richer in anthocyanins and the latter richer in ellagitannins. Thus, the choice of the drying method should consider not only processing costs, but also the purpose of the final product. If the objective is to obtain a powder with higher anthocyanins contents for further application as a natural colorant in the development of food products, freeze-drying is the most appropriate method. On the other hand, if the aim is to obtain a final product with functional properties, considering that both anthocyanins and elagitannins have already been described to present potential beneficial health effects, both methods could be used. However, in this latter circumstance, if we take into account processing costs and drying efficiency, oven-drying is preferable. Regarding the metabolism of jabuticaba ellagitannins, this study was the first to investigate the metabolism of these compounds in non-European individuals, specifically in Brazilians, through the consumption of jabuticaba peel and seed. As observed in studies conducted in Europe, not only the three metabotypes (A, B and 0) were found in the Brazilian population, but also a higher prevalence of metabotype B was observed in overweight and obese individuals, in comparison to normoweight. Additionally, in the present study, the unexpected occurrence of diarrhea as a side-effect by some volunteers reduced the production of urolithins, as well as affected the metabotypes distribution, reinforcing the pivotal role of gut microbiota on ellagitannins metabolism. As a whole, these results suggest that despite cultural and eating habits differences between European and Brazilian populations, their gut microbiota composition may be similar in terms of ellagitannins metabolism. 140

FUTURE PERSPECTIVES

 To analyze the gut microbiota composition from volunteers that consumed jabuticaba peel and seed and associate these results to the nutritional state, gender, metabotypes distribution and occurrence of diarrhea;

 Considering that ellagic acid is the main precursor of urolithins and that this compound would have to be released from ellagitannins prior to metabolization, it would be relevant to investigate the effect of acid hydrolysis of jabuticaba peel and seed on the depolymerization of ellagitannins to ellagic acid, as well as on urolithins bioavailability;

 To investigate the effect of jabuticaba peel and seed chronic consumption on metabolic parameters, inflammatory biomarkers and oxidative stress in obese and dyslipidemic adults.

141

REFERENCES

142

ABE, L. T.; LAJOLO, F. M.; GENOVESE, M. I. Potential dietary sources of ellagic acid and other antioxidants among fruits consumed in Brazil: jabuticaba (Myrciaria jaboticaba (Vell.) Berg). Journal of the science of food and agriculture, v. 92, n. 8, p. 1679–1687, 2011.

ACOSTA-ESTRADA, B. A.; GUTIÉRREZ-URIBE, J. A.; SERNA-SALDÍVAR, S. O. Bound phenolics in foods, a review. Food chemistry, v. 152, p. 46–55, 2014.

AGOSTINI, J. DA S. et al. Atmosfera modificada e condições de armazenamento nas características físico-químicas de jabuticabas da cultivar ‘paulista’. Ciência Rural, v. 39, n. 9, p. 2601–2608, 2009.

AKISSOÉ, N. et al. How blanching and drying affect the colour and functional characteristics of yam (Dioscorea cayenensis-rotundata) flour. Food Chemistry, v. 82, n. 2, p. 257–264, 2003.

ALEZANDRO, M. R. et al. Comparative study of chemical and phenolic compositions of two species of jaboticaba: Myrciaria jaboticaba (Vell.) Berg and Myrciaria cauliflora (Mart.) O. Berg. Food Research International, v. 54, n. 1, p. 468–477, 2013.

ALVES, A. P. DE C. et al. Flour and anthocyanin extracts of jaboticaba skins used as a natural dye in yogurt. International Journal of Food Science & Technology, v. 48, n.10, p. 2007- 2013, 2013.

ANESE, M. et al. Effect of high-pressure treatments on peroxidase and polyphenoloxidase activities. Journal of Food Biochemistry, v. 18, p. 285–293, 1995.

ARAI, Y. et al. Comparison of isoflavones among dietary intake, plasma concentration and urinary excretion for accurate estimation of phytoestrogen intake. Journal of Epidemiology, v. 10, n. 2, p. 127–135, 2000.

ARAÚJO, C. R. R. et al. Myrciaria cauliflora peel flour had a hypolipidemic effect in rats fed a moderately high-fat diet. Journal of medicinal food, v. 17, n. 2, p. 262–267, 2014.

ARRANZ, S. et al. High contents of nonextractable polyphenols in fruits suggest that contents of plant foods have been underestimated. Journal of Agricultural and Food Chemistry, v. 57, n. 16, p. 7298–7303, 2009.

ASCHERI, D. P. R.; ACHERI, J. L. R.; CARVALHO, C. W. P. DE. Caracterização da farinha de bagaço de jabuticaba e propriedades funcionais dos extrusados. Ciência e Tecnologia de Alimentos, v. 26, n. 4, p. 897–905, 2006.

ATKINSON, C.; FRANKENFELD, C. L.; LAMPE, J. W. MINIREVIEW Gut Bacterial Metabolism of the Soy Isoflavone Daidzein: Exploring the Relevance to Human Health. Exp Bioi Mad, v. 230, p. 155–170, 2005.

AYDIN, E.; GOCMEN, D. The influences of drying method and metabisulfite pre-treatment onthe color, functional properties and phenolic acids contents and bioaccessibility of pumpkin flour. LWT - Food Science and Technology, v. 60, n. 1, p. 385–392, 2015. 143

BALDIN, J. C. et al. Effect of microencapsulated Jabuticaba (Myrciaria cauliflora) extract on quality and storage stability of mortadella sausage. Food Research International, v. 108, p. 551–557, 2018.

BALISTEIRO, D. M. et al. Effect of clarified Brazilian native fruit juices on postprandial glycemia in healthy subjects. Food Research International, v. 100, p. 196–203, 2017.

BARBA, F. J.; ESTEVE, M. J.; FRIGOLA, A. Physicochemical and nutritional characteristics of blueberry juice after high pressure processing. Food Research International, v. 50, n. 2, p. 545–549, 2013.

BASEGGIO, A. M. et al. Jaboticaba peel extract decrease autophagy in White adipose tissue and prevents metabolic disorders in mice fed with a high-fat diet. PhamaNutrition, v. 6, p. 147-156, 2018.

BARTA, J. Fruit drying principles. In: Hui, Y. H. (Ed). Handbook of Fruits and Fruit Processing. Hoboken: Blackwell Publishing, 2006, 81-93.

BASANTA, M. F. et al. Antioxidant Japanese plum (Prunus salicina) microparticles with potential for food preservation. Journal of Functional Foods, v. 24, p. 287–296, 2016.

BATISTA, Â. G. et al. Jaboticaba (Myrciaria jaboticaba (Vell.) Berg.) peel increased triglycerides excretion and hepatic lipid peroxidation in high-fat-fed rats. Revista de Nutrição de Campinas, v. 26, n. 5, p. 571–581, 2013.

BATISTA, Â. G. et al. Intake of jaboticaba peel attenuates oxidative stress in tissues and reduces circulating saturated lipids of rats with high-fat diet-induced obesity. Journal of Functional Foods, v. 6, p. 450–461, dez. 2014.

BATISTA, Â. G. et al. Jaboticaba berry peel intake prevents insulin-resistance-induced tau phosphorylation in mice. Molecular Nutrition and Food Research, v. 61, n. 10, p. 1–10, 2017.

BATISTA, Â. G. et al. Jaboticaba berry peel intake increases short chain fatty acids production and prevent hepatic steatosis in mice fed high-fat diet. Journal of Functional Foods, v. 48, n. June, p. 266–274, 2018.

BENZIE, I. F. F.; STRAIN, J. J. The ferric reducing ability of plasma (FRAP) as a measure of “antioxidant power”: the FRAP assay. Analytical biochemistry, v. 239, n. 1, p. 70–76, 1996.

BETTA, F. D. et al. Phenolic Compounds Determined by LC-MS/MS and In Vitro Antioxidant Capacity of Brazilian Fruits in Two Edible Ripening Stages. Plant Foods for Human Nutrition, v. 73, n. 4, p. 302–307, 2018.

BŁASZCZAK, W.; AMAROWICZ, R.; GÓRECKI, A. R. Antioxidant capacity, phenolic composition and microbial stability of aronia juice subjected to high hydrostatic pressure processing. Innovative Food Science and Emerging Technologies, v. 39, p. 141–147, 2017.

BORGES, L. L.; CONCEIÇÃO, E. C.; SILVEIRA, D. Active compounds and medicinal properties of Myrciaria genus. Food Chemistry, v. 153, p. 224–233, 2014. 144

BRAVIM, F. et al. High hydrostatic pressure and the cell membrane: Stress response of Saccharomyces cerevisiae. Annals of the New York Academy of Sciences, v. 1189, p. 127– 132, 2010.

BRIONES-LABARCA, V. et al. High hydrostatic pressure and ultrasound extractions of antioxidant compounds, sulforaphane and fatty acids from Chilean papaya (Vasconcellea pubescens) seeds: Effects of extraction conditions and methods. LWT - Food Science and Technology, v. 60, n. 1, p. 525–534, 2015.

BUZRUL, S. et al. Inactivation of Escherichia coli and Listeria innocua in kiwifruit and pineapple juices by high hydrostatic pressure. International journal of food microbiology, v. 124, n. 3, p. 275–8, 2008.

CALLONI, C. et al. Jaboticaba (Plinia trunciflora (O. Berg) Kausel) fruit reduces oxidative stress in human fibroblasts cells (MRC-5). Food Research International, v. 70, p. 15–22, 2015.

CAO, X. et al. Effects of high hydrostatic pressure on enzymes, phenolic compounds, anthocyanins, polymeric color and color of strawberry pulps. Journal of the science of food and agriculture, v. 91, n. 5, p. 877–85, 2011.

CASTAÑEDA-OVANDO, A. et al. Chemical studies of anthocyanins: A review. Food Chemistry, v. 113, n. 4, p. 859–871, 2009.

CERDÁ, B. et al. Repeated oral administration of high doses of the punicalagin to rats for 37 days is not toxic. Journal of Agricultural and Food Chemistry, v. 51, n. 11, p. 3493–3501, 2003a.

CERDÁ, B. et al. Evaluation of the bioavailability and metabolism in the rat of punicalagin, an antioxidant polyphenol from pomegranate juice. European Journal of Nutrition, v. 42, n. 1, p. 18–28, 2003b.

CERDÁ, B.; TOMÁS-BARBERÁN, F. A.; ESPÍN, J. C. Metabolism of antioxidant and chemopreventive ellagitannins from strawberries, raspberries, walnuts, and oak-aged wine in humans: Identification of biomarkers and individual variability. Journal of Agricultural and Food Chemistry, v. 53, n. 2, p. 227–235, 2005a.

CERDÁ, B. et al. Identification of urolithin A as a metabolite produced by human colon microflora from ellagic acid and related compounds. Journal of Agricultural and Food Chemistry, v. 53, n. 14, p. 5571–5576, 2005b.

CHANG, C. H. et al. Comparisons on the antioxidant properties of fresh, freeze-dried and hot- air-dried tomatoes. Journal of Food Engineering, v. 77, n. 3, p. 478–485, 2006.

CITADIN, I.; DANNER, M. A.; SASSO, S. A. Z. Jabuticabeiras. Revista Brasileira de Fruticultura, v. 32, n. 2, p. 343–656, 2010.

CLIFFORD, M. N.; SCALBERT, A. Review Ellagitannins – nature, occurrence and dietary burden. Journal of the Science of Food and Agriculture, v. 80, p. 1118-1125, 2000. 145

CONSIDINE, G.D. Van Nostrand’s Scientific Encyclopedia. Hoboken: John Wiley & Sons. Inc, 2005. 1856 p.

CORRALES, M. et al. Extraction of anthocyanins from grape skins assisted by high hydrostatic pressure. Journal of Food Engineering, v. 90, n. 4, p. 415–421, 2009.

CORTÉS-MARTÍN, A. et al. The gut microbiota urolithin metabotypes revisited: the human metabolism of ellagic acid is mainly determined by aging. Food & Function, v. 9, n.8, 4100- 4106, 2018.

CROZIER, A.; JAGANATH, I. B.; CLIFFORD, M. N. Dietary phenolics: Chemistry, bioavailability and effects on health. Natural Product Reports, v. 26, n. 8, p. 1001–1043, 2009.

DA SILVA, J. K. et al. Functional tea from a Brazilian berry: Overview of the bioactive compounds. LWT - Food Science and Technology, v. 76, p. 292–298, 2017.

DAIUTO, E. R. et al. Qualidade pós colheita dos frutos de jabuticaba tratada por hipotermia. Agronomia Tropical, v. 60, p. 231–240, 2010.

DANIEL, E. M. et al. The effects of pH and rat intestinal contents on the liberation of ellagic acid from purified and crude ellagitannins. Journal of Natural Products, v. 54, n. 4, p. 946– 952, 1991.

DANTAS, A. M. et al. Bioaccessibility of phenolic compounds in native and exotic frozen pulps explored in Brazil using a digestion model coupled with a simulated intestinal barrier. Food Chemistry, v. 274, p. 202–214, 2019.

DE ALMEIDA, P. L. et al. Effect of jabuticaba peel extract on lipid oxidation, microbial stability and sensory properties of Bologna-type sausages during refrigerated storage. Meat Science, v. 110, p. 9–14, 2015.

DE AZEVÊDO, J. C. S. et al. Dried camu-camu (Myrciaria dubia H.B.K. McVaugh) industrial residue: A bioactive-rich Amazonian powder with functional attributes. Food Research International, v. 62, p. 934–940, 2014.

DE JESUS, N. et al. Caracterização de quatro grupos de jabuticabeira, nas condições de jaboticabal-SP. Revista Brasileira de Fruticultura, v. 26, n. 3, p. 482–485, 2004.

DE SOUZA, C. G. et al. Radical scavenger capacity of jabuticaba fruit (myrciaria cauliflora) and its biological effects in hypertensive rats. Oxidative Medicine and Cellular Longevity, v. 2017, 2017.

DE TORRES, C. et al. Effect of freeze-drying and oven-drying on volatiles and phenolics composition of grape skin. Analytica Chimica Acta, v. 660, n. 1–2, p. 177–182, 2010.

DEFRAEYE, T. When to stop drying fruit: Insights from hygrothermal modelling. Applied Thermal Engineering, v. 110, p. 1128–1136, 2017. 146

DEL RIO, D. et al. Dietary (poly)phenolics in human health: structures, bioavailability, and evidence of protective effects against chronic diseases. Antioxidants & Redox Signaling, v. 18, n. 14, p. 1818–1892, 2013.

DESSIMONI-PINTO, N. A. V. et al. Jaboticaba peel for jelly preparation: an alternative technology. Ciência e Tecnologia de Alimentos, v.31, n. 4, p. 864–869, 2011.

DRAGANO, N. R. V et al. Freeze-dried jaboticaba peel powder improves insulin sensitivity in high-fat-fed mice. The British journal of nutrition, v. 110, n. 3, p. 447–55, 2013.

ESPÍN, J. C. et al. Iberian pig as a model to clarify obscure points in the bioavailability and metabolism of ellagitannins in humans. Journal of Agricultural and Food Chemistry, v. 55, n. 25, p. 10476–10485, 2007.

ESPÍN, J. C.; GONZÁLEZ-SARRÍAS, A.; TOMÁS-BARBERÁN, F. A. The gut microbiota: A key factor in the therapeutic effects of (poly)phenols. Biochemical Pharmacology, v. 139, p. 82–93, 2017.

FERNÁNDEZ-JALAO, I.; SÁNCHEZ-MORENO, C.; DE ANCOS, B.. Effect of high- pressure processing on flavonoids, hydroxycinnamic acids, dihydrochalcones and antioxidant activity of apple ‘Golden Delicious’ from different geographical origin. Innovative Food Science and Emerging Technologies. https://doi.org/10.1016/j.ifset.2018.06.002

FISCHER, U. A.; CARLE, R.; KAMMERER, D. R. Identification and quantification of phenolic compounds from pomegranate (Punica granatum L.) peel, mesocarp, aril and differently produced juices by HPLC-DAD-ESI/MSn. Food Chemistry, v. 127, n. 2, p. 807– 821, 2011.

FRACASSETTI, D. et al. Ellagic acid derivatives, ellagitannins, proanthocyanidins and other phenolics, vitamin C and antioxidant capacity of two powder products from camu-camu fruit (Myrciaria dubia). Food Chemistry, v. 139, n. 1–4, p. 578–588, 2013.

FRANKENFELD, C.L. et al. Familial correlations, segregation analysis, and nongenetic correlates of soy isoflavone-metabolizing phenotypes. Experimental Biology and Medicine, v. 229, n. 9, p. 902-913, 2004.

FRANKENFELD, C. L. et al. High concordance of daidzein-metabolizing phenotypes in individuals measured 1 to 3 years apart. The British journal of nutrition, v. 94, n. 6, p. 873– 876, 2005.

FUHRMAN, B. J. et al. Equol status modifies the association of soy intake and mammographic density in a sample of postmenopausal women. Cancer Epidemiology Biomarkers and Prevention, v. 17, n. 1, p. 33–42, 2008.

GARCÍA-MUÑOZ, C. et al. Diversity of urinary excretion patterns of main ellagitannins’ colonic metabolites after ingestion of tropical highland blackberry (Rubus adenotrichus) juice. Food Research International, v. 55, p. 161–169, 2014. 147

GARCÍA-NIÑO, W. R.; ZAZUETA, C. Ellagic acid: Pharmacological activities and molecular mechanisms involved in liver protection. Pharmacological Research, v. 97, p. 84–103, 2015.

GARCIA-PALAZON, A. et al. The effects of high hydrostatic pressure on β-glucosidase, peroxidase and polyphenoloxidase in red raspberry (Rubus idaeus) and strawberry (Fragaria x ananassa). Food Chemistry, v. 88, n. 1, p. 7–10, 2004.

GARCÍA-VILLALBA, R. et al. Time course production of urolithins from ellagic acid by human gut microbiota. Journal of Agricultural and Food Chemistry, v. 61, n. 37, p. 8797– 8806, 2013.

GARCÍA-VILLALBA, R. et al. Validated Method for the Characterization and Quantification of Extractable and Nonextractable Ellagitannins after Acid Hydrolysis in Pomegranate Fruits, Juices, and Extracts. Journal of Agricultural and Food Chemistry, v. 63, n. 29, p. 6555– 6566, 2015.

GARCÍA-VILLALBA, R.; ESPÍN, J. C.; TOMÁS-BARBERÁN, F. A. Chromatographic and spectroscopic characterization of urolithins for their determination in biological samples after the intake of foods containing ellagitannins and ellagic acid. Journal of Chromatography A, v. 1428, p. 162–175, 2016.

GARDANA, C.; CANZI, E.; SIMONETTI, P. The role of diet in the metabolism of daidzein by human faecal microbiota sampled from Italian volunteers. Journal of Nutritional Biochemistry, v. 20, n. 12, p. 940–947, 2009.

GONZÁLEZ-BARRIO, R. et al. Bioavailability of anthocyanins and ellagitannins following consumption of raspberries by healthy humans and subjects with an ileostomy. Journal of Agricultural and Food Chemistry, v. 58, n. 7, p. 3933–3939, 2010.

GONZÁLEZ-BARRIO, R. et al. Colonic catabolism of ellagitannins, ellagic acid and raspberry anthocyanins: in vivo and in vitro studies. Drug Metabolism and Disposition, v. 39, n. 9, p. 1680–1688, 2011a.

GONZÁLEZ-BARRIO, R. et al. UV and MS identification of urolithins and nasutins, the bioavailable metabolites of ellagitannins and ellagic acid in different mammals. Journal of Agricultural and Food Chemistry, v. 59, n. 4, p. 1152–1162, 2011b.

GONZÁLEZ-BARRIO, R. et al. Metabolism of oak leaf ellagitannins and urolithin production in beef cattle. Journal of Agricultural and Food Chemistry, v. 60, n. 12, p. 3068–3077, 2012.

GONZÁLEZ-CEBRINO, F. et al. Changes after high-pressure processing on physicochemical parameters, bioactive compounds, and polyphenol oxidase activity of red flesh and peel plum purée. Innovative Food Science & Emerging Technologies, v. 20, p. 34–41, 2013.

GONZÁLEZ-SARRÍAS, A. et al. Occurrence of urolithins, gut microbiota ellagic acid metabolites and proliferation markers expression response in the human prostate gland upon consumption of walnuts and pomegranate juice. Molecular Nutrition and Food Research, v. 54, n. 3, p. 311–322, 2010. 148

GONZÁLEZ-SARRÍAS, A. et al. The gut microbiota ellagic acid-derived metabolite urolithin a and its sulfate conjugate are substrates for the drug efflux transporter breast cancer resistance protein (ABCG2/BCRP). Journal of Agricultural and Food Chemistry, v. 61, n. 18, p. 4352– 4359, 2013.

GONZÁLEZ-SARRÍAS, A. et al. Identifying the limits for ellagic acid bioavailability: a crossover pharmacokinetic study in healthy volunteers after consumption of pomegranate extracts. Journal of Functional Foods, v. 19, p. 225-235, 2015.

GONZÁLEZ-SARRÍAS, A. et al. Clustering according to urolithin metabotype explains the interindividual variability in the improvement of cardiovascular risk biomarkers in overweight- obese individuals consuming pomegranate: A randomized clinical trial. Molecular Nutrition and Food Research, v. 61, n. 5, p. 1–14, 2017.

GRACE, M. H. et al. Chemical composition, antioxidant and anti-inflammatory properties of pistachio hull extracts. Food Chemistry, v. 210, p. 85–95, 2016.

GUO, K. et al. Daidzein-metabolising phenotypes in relation to serum lipids and uric acid in adults in Guangzhou, China. British Journal of Nutrition, v. 104, n. 1, p. 118–124, 2010.

GURAK, P. D. et al. Jaboticaba Pomace Powder Obtained as a Co-product of Juice Extraction: A Comparative Study of Powder Obtained from Peel and Whole Fruit. Food Research International, v. 62, p. 786–792, 2014.

HÄKKINEN, S. H. et al. Ellagic acid content in berries: Influence of domestic processing and storage. European Food Research and Technology, v. 212, n. 1, p. 75-80, 2000.

HAZIM, S. et al. Acute benefits of the microbial-derived isoflavone metabolite equol on arterial stiffness in men prospectively recruited according to equol producer phenotype: A double-blind randomized controlled trial. American Journal of Clinical Nutrition, v. 103, n. 3, p. 694–702, 2016.

HOFFMANN-RIBANI, R.; HUBER, L. S.; RODRIGUEZ-AMAYA, D. B. Flavonols in fresh and processed Brazilian fruits. Journal of Food Composition and Analysis, v. 22, n. 4, p. 263–268, 2009.

HOGAN, E.; KELLY, A. L.; SUN, D.-W. High Pressure Processing of Foods : An Overview. Emerging technologies for food processing, p. 3–32, 2005.

HSIEH, S. K. et al. Antibacterial and laxative activities of strictinin isolated from Pu’er tea (Camellia sinensis). Journal of Food and Drug Analysis, v. 24, n. 4, p. 722–729, 2016.

HSU, J. D. et al. Myrciaria cauliflora extract improves diabetic nephropathy via suppression of oxidative stress and inflammation in streptozotocin-nicotinamide mice. Journal of Food and Drug Analysis, v. 24, n. 4, p. 730–737, 2016.

HUPPERTZ, T.; KELLY, A. L.; FOX, P. F. Effects of high pressure on constituents and properties of milk. International Dairy Journal, v. 12, n. 7, p. 561–572, 2002. 149

INADA, K. O. P. et al. Screening of the chemical composition and occurring antioxidants in jabuticaba (Myrciaria jaboticaba) and jussara (Euterpe edulis) fruits and their fractions. Journal of Functional Foods, v. 17, p. 422–433, 2015.

INADA, K. O. P. et al. High hydrostatic pressure processing affects the phenolic profile, preserves sensory attributes and ensures microbial quality of jabuticaba (Myrciaria jaboticaba) juice. Journal of the Science of Food and Agriculture, v. 98, n. 1, p. 231–239, 2018a.

INADA, K. O. P. et al. Jabuticaba (Myrciaria jaboticaba) juice obtained by steam-extraction: phenolic compound profile, antioxidant capacity, microbiological stability, and sensory acceptability. Journal of Food Science and Technology, v. 55, n. 1, p. 52–61, 2018b.

ITO, H.; IGUCHI, A.; HATANO, T. Identification of urinary and intestinal bacterial metabolites of ellagitannin in rats. Journal of Agricultural and Food Chemistry, v. 56, n. 2, p. 393–400, 2008.

JANOVITZ-KLAPP, A.; RICHARD, F.; NICOLAS, J. Polyphenoloxidase from apple, partial purification and some properties. Phytochemistry, v. 28, n. 11, p. 2903–2907, 1989.

JEŻ, M. et al. The impact of high pressure processing on the phenolic profile, hydrophilic antioxidant and reducing capacity of purée obtained from commercial tomato varieties. Food Chemistry, v. 261, p. 201–209, 2018.

JUN, X. et al. Characterization of polyphenols from green tea leaves using a high hydrostatic pressure extraction. International journal of pharmaceutics, v. 382, n. 1–2, p. 139–143, 2009.

KARONEN, M. et al. Reversed-phase HPLC-ESI/MS analysis of birch leaf proanthocyanidins after their acidic degradation in the presence of nucleophiles. Analysis, v. 18, n. 5, p. 378–386, 2007.

KENNEDY, J. A.; JONES, G. P. Analysis of proanthocyanidin cleavage products following acid-catalysis in the presence of excess phloroglucinol. Journal of Agricultural and Food Chemistry, v. 49, n. 4, p. 1740–1746, 2001.

KHANBABAEE, K.; REE, T. VAN. Tannins: Classification and definition. Natural Product Reports, v. 18, n. 6, p. 641–649, 2001.

KHARAGHANI, A. et al. Freeze-drying. In: Hogdson R.H. (Ed). Ullmann’s Encyclopedia of Industrial Chemistry. Weinheim: Wiley-VCH Verlag GmbH & Co, 2017, 1-47.

KIM, J. E. et al. Gallotannin-enriched extract isolated from galla rhois may be a functional candidate with laxative effects for treatment of loperamide-induced constipation of SD Rats. PLoS ONE, v. 11, n. 9, 2016.

KREIJKAMP-KASPERS, S. et al. Randomized controlled trial of the effects of soy protein containing isoflavones on vascular function in postmenopausal women. The American Journal of Clinical Nutrition, v. 81, n. 1, p. 189–195, 2005. 150

KRUGER, M. J. et al. Proanthocyanidins, anthocyanins and cardiovascular diseases. Food Research International, v. 59, p. 41–52, 2014.

LABOISSIÈRE, L. H. E. S. et al. Effects of high hydrostatic pressure (HHP) on sensory characteristics of yellow passion fruit juice. Innovative Food Science & Emerging Technologies, v. 8, n. 4, p. 469–477, 2007.

LAMAS, C. A. et al. Jaboticaba extract prevents prediabetes and liver steatosis in high-fat-fed aging mice. Journal of Functional Foods, v. 47, n. May, p. 434–446, 2018.

LANDETE, J. M. Ellagitannins, ellagic acid and their derived metabolites: a review about source, metabolism, functions and health. Food Research International, v. 44, n. 5, p. 1150– 1160, 2011.

LAVINAS, F. C. et al. Effect of High Hydrostatic Pressure on Cashew Apple (Anacardium occidentale L.) Juice Preservation. Journal of Food Science, v. 73, n. 6, p. M273–M277, 2008.

LAZARUS, S. A. et al. High-performance liquid chromatography/mass spectrometry analysis of proanthocyanidins in foods and beverages. Journal of Agricultural and Food Chemistry, v. 47, p. 3693–3701, 1999.

LEITE, A. V et al. Antioxidant potential of rat plasma by administration of freeze-dried jaboticaba peel (Myrciaria jaboticaba Vell Berg). Journal of agricultural and food chemistry, v. 59, n. 6, p. 2277–2283, 2011.

LEITE-LEGATTI, A. V. et al. Jaboticaba peel: Antioxidant compounds, antiproliferative and antimutagenic activities. Food Research International, v. 49, n. 1, p. 596–603, 2012.

LEKHA, P. K., LONSANE, B. K. Production and application of tannin acyl hydrolase: state of art. Advances in Applied Microbiology, v. 44, 215-260, 1997.

LEWICKI, P. P. Design of hot air drying for better foods. Trends in Food Science and Technology, v. 17, n. 4, p. 153–163, 2006.

LIMA, A. D. J. B. et al. Caracterização química do fruto jabuticaba (Myrciaria cauliflora Berg) e de suas frações. Archivos Latinoamericanos de Nutricion, v. 58, n. 4, p. 416–421, 2008.

LIMA, A. D. J. B. et al. Anthocyanins, pigment stability and antioxidant activity in jabuticaba [Myrciaria cauliflora (Mart.) O. Berg]. Revista Brasileira de Fruticultura, v. 33, n. 3, p. 877– 887, 2011.

LIU, Z. M. et al. Cardiovascular risks in relation to daidzein metabolizing phenotypes among Chinese postmenopausal women. PLoS ONE, v. 9, n. 2, 2014.

LIU, S. et al. Effects of high hydrostatic pressure on physicochemical properties, enzymes activity, and antioxidant capacities of anthocyanins extracts of wild Lonicera caerulea berry. Innovative Food Science and Emerging Technologies, v. 36, p. 48–58, 2016. 151

LOPES, M. M. A. et al. Low fluence pulsed light enhanced phytochemical content and antioxidant potential of “Tommy Atkins” mango peel and pulp. Innovative Food Science and Emerging Technologies, v. 33, p. 216–224, 2016.

LUDWIG, I. A. et al. New insights into the bioavailability of red raspberry anthocyanins and ellagitannins. Free Radical Biology and Medicine, v. 89, p. 758–769, 2015.

MACHADO, G. H. A. et al. Antibacterial activity and in vivo wound healing potential of phenolic extracts from jaboticaba skin. Chem Biol Drug Des, v. 92, n. 1, p. 1333–1343, 2018.

MANACH, C. Polyphenols: food sources and bioavailability. American Journal of Clinical Nutrition, v. 79, n. 5, p. 727–747, 2004.

MANACH, C. et al. Addressing the inter-individual variation in response to consumption of plant food bioactives: Towards a better understanding of their role in healthy aging and cardiometabolic risk reduction. Molecular Nutrition and Food Research, v. 61, n. 6, p. 1–16, 2017.

MERTENS-TALCOTT, S. U. et al. Absorption, metabolism, and antioxidant effects of pomegranate (Punica granatum L.) polyphenols after ingestion of a standardized extract in healthy human volunteers. Journal of Agricultural and Food Chemistry, v. 54, n. 23, p. 8956–8961, 2006.

MICHALSKA, A. et al. Physicochemical properties of whole fruit plum powders obtained using different drying technologies. Food Chemistry, v. 207, p. 223–232, 2016.

MICHALSKA, A. et al. Effect of different drying techniques on physical properties, total polyphenols and antioxidant capacity of blackcurrant pomace powders. LWT - Food Science and Technology, v. 78, p. 114–121, 2017.

MONTES-ÁVILA, J.; LÓPEZ-ANGULO, G.; DELGADO-VARGAS, F. Tannins in fruits and vegetables: Chemistry and biological functions. In: Yahia, E. M (Ed.). Fruit and Vegetable : Chemistry and Human Health. Hoboken: Wiley-Blackwell, 2017, p. 221– 268.

MORALES, P. et al. Non-fermented and fermented jabuticaba (Myrciaria cauliflora Mart.) pomaces as valuable sources of functional ingredients. Food Chemistry, v. 208, p. 220–227, 2016.

MORENO, J. et al. Influence of ohmic heating/osmotic dehydration treatments on polyphenoloxidase inactivation, physical properties and microbial stability of apples (cv. Granny Smith). Innovative Food Science and Emerging Technologies, v. 20, p. 198–207, 2013.

MORTON, M. S. et al. Phytoestrogen concentrations in serum from Japanese men and women over forty years of age. American Society for Nutritional Sciences, v. 132, n. 10, p. 3168– 3171, 2002. 152

MOURA, M. H. C. et al. Phenolic-rich jaboticaba (Plinia jaboticaba (Vell.) Berg) extracts prevent high-fat-sucrose diet-induced obesity in C57BL/6 mice. Food Research International, v. 107, n. January, p. 48–60, 2018.

MPHAHLELE, R. R. et al. Effect of drying on the bioactive compounds, antioxidant, antibacterial and antityrosinase activities of pomegranate peel. BMC Complementary and Alternative Medicine, v. 16, n. 1, p. 1–12, 2016.

NETA, M. C. A. et al. Fermented dessert with whey, ingredients from the peel of Jabuticaba (Myrciaria cauliflora) and an indigenous culture of lactobacillus plantarum: Composition, microbial viability, antioxidant capacity and sensory features. Nutrients, v. 10, n. 9, p. 1–19, 2018.

NEVES, N. DE A. et al. Flavonols and ellagic acid derivatives in peels of different species of jabuticaba (Plinia spp.) identified by HPLC-DAD-ESI/MSn. Food Chemistry, v. 252, n. January, p. 61–71, 2018.

NUNES, J. C. et al. Effect of drying method on volatile compounds, phenolic profile and antioxidant capacity of guava powders. Food Chemistry, v. 197, p. 881–890, 2016.

NUÑEZ-SÁNCHEZ, M. A. et al. Targeted metabolic profiling of pomegranate polyphenols and urolithins in plasma, urine and colon tissues from colorectal cancer patients. Molecular Nutrition and Food Research, v. 58, n. 6, p. 1199–1211, 2014.

OKUDA, T.; YOSHIDA, T.; HATANO, T. Hydrolyzable tannins and related polyphenols. In: Berlinck et al. (Ed.). Progress in the Chemistry of Organic Natural Products. Berlin:Springer-Verlag, 1995, p. 1–117.

OTTEN, J. J., Hellwig, J. P., Meyers, L. D. Dietary reference intakes: The essential guide to nutrient requirements. Washington: National Academies of Sciences, 2006. 1344 p.

PELVAN, E. et al. Phenolic profiles and antioxidant activity of Turkish Tombul hazelnut samples (natural, roasted, and roasted hazelnut skin). Food Chemistry, v. 244, p. 102–108, 2018.

PEREIRA, E. P. R. et al. Effect of incorporation of antioxidants on the chemical, rheological, and sensory properties of probiotic petit suisse cheese. Journal of Dairy Science, v. 99, n. 3, p. 1762–1772, 2016.

PEREIRA, L. D. et al. Polyphenol and ellagitannin constituents of jabuticaba (Myrciaria cauliflora) and chemical variability at different stages of fruit development. Journal of Agricultural and Food Chemistry, v. 65, n. 6, p. 1209–1219, 2017.

PINELA, J. et al. Cold extraction of phenolic compounds from watercress by high hydrostatic pressure: Process modelling and optimization. Separation and Purification Technology, v. 192, p. 501–512, 2018. 153

PINTO, M. DA S.; LAJOLO, F. M.; GENOVESE, M. I. Bioactive compounds and quantification of total ellagic acid in strawberries (Fragaria x ananassa Duch.). Food Chemistry, v. 107, n. 4, p. 1629–1635, 2008.

PLAZA, M. et al. Characterization of antioxidant polyphenols from Myrciaria jaboticaba peel and their effects on glucose metabolism and antioxidant status: A pilot clinical study. Food Chemistry, v. 211, p. 185–197, 2016.

PRASAD, K. N. et al. High pressure extraction of corilagin from longan (Dimocarpus longan Lour.) fruit pericarp. Separation and Purification Technology, v. 70, n. 1, p. 41–45, 2009.

PRIOR, R. L.; WU, X.; SCHAICH, K. Standardized methods for the determination of antioxidant capacity and phenolics in foods and dietary supplements. Journal of agricultural and food chemistry, v. 53, n. 10, p. 4290–4302, 2005.

PUNITHAVATHI, V. R. et al. Antihyperglycaemic, antilipid peroxidative and antioxidant effects of gallic acid on streptozotocin induced diabetic Wistar rats. European Journal of Pharmacology, v. 650, n. 1, p. 465–471, 2011.

PUUPPONEN-PIMIA, R. et al. Effects of ellagitannin-rich berries on blood lipids, gut microbiota, and urolithin production in human subjects with symptoms of metabolic syndrome. Molecular Nutrition & Food Research, v. 57, n. 12, p. 2258-2263, 2013.

QUATRIN, A. et al. The hepatoprotective effect of jaboticaba peel powder in a rat model of type 2 diabetes mellitus involves the modulation of thiol/disulfide redox state through the upregulation of glutathione synthesis. Journal of Nutrition and Metabolism, v. 2018, p. 1- 13, 2018.

QUE, F. et al. Comparison of hot air-drying and freeze-drying on the physicochemical properties and antioxidant activities of pumpkin (Cucurbita moschata Duch.) flours. International Journal of Food Science and Technology, v. 43, n. 7, p. 1195–1201, 2008.

QUIDEAU, S.; FELDMAN, K. S. Ellagitannin Chemistry. Chemical Reviews, v. 96, n. 1, p. 475–504, 1996.

RATTI, C. Hot air and freeze-drying of high-value foods: A review. Journal of Food Engineering, v. 49, n. 4, p. 311–319, 2001.

RAYNAL, J.; MOUTOUNET, M.; SOUQUET, J. MARC. Intervention of Phenolic Compounds in Plum Technology. 1. Changes during Drying. Journal of Agricultural and Food Chemistry, v. 37, n. 4, p. 1046–1050, 1989.

RE, R. et al. Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free radical biology & medicine, v. 26, n. 9/10, p. 1231–1237, 1999.

RENDUELES, E. et al. Microbiological food safety assessment of high hydrostatic pressure processing: A review. LWT - Food Science and Technology, v. 44, n. 5, p. 1251–1260, 2011. 154

REYNERTSON, K. A. et al. Bioactive depsides and anthocyanins from jaboticaba (Myrciaria cauliflora). Journal of natural products, v. 69, n. 8, p. 1228–1230, 2006.

REYNERTSON, K. A. et al. Quantitative analysis of antiradical phenolic constituents from fourteen edible Myrtaceae fruits. Food Chemistry, v. 109, n. 4, p. 883–890, 2008.

RODRÍGUEZ-ROQUE, M. J. et al. Impact of food matrix and processing on the in vitro bioaccessibility of vitamin C, phenolic compounds, and hydrophilic antioxidant activity from fruit juice-based beverages. Journal of Functional Foods, v. 14, p. 33–43, 2015.

ROMO-VAQUERO, M. et al. Interindividual variability in the human metabolism of ellagic acid: Contribution of Gordonibacter to urolithin production. Journal of Functional Foods, v. 17, p. 785–791, 2015.

SAHEBKAR, A. et al. Effects of supplementation with pomegranate juice on plasma C-reactive protein concentrations: A systematic review and meta-analysis of randomized controlled trials. Phytomedicine, v. 23, n. 11, p. 1095–1102, 2016a.

SAHEBKAR, A. et al. Lipid profile changes after pomegranate consumption: A systematic review and meta-analysis of randomized controlled trials. Phytomedicine, v. 23, n. 11, p. 1103– 1112, 2016b.

SAIKAEW, K. et al. Effect of high-pressure processing on colour, phytochemical contents and antioxidant activities of purple waxy corn (Zea mays L. var. ceratina) kernels. Food Chemistry, v. 243, n. May 2017, p. 328–337, 2018.

SAMOTICHA, J.; WOJDYŁO, A.; LECH, K. The influence of different the drying methods on chemical composition and antioxidant activity in chokeberries. LWT - Food Science and Technology, v. 66, p. 484–489, 2016.

SANCHO, R. A. S.; PASTORE, G. M. Evaluation of the effects of anthocyanins in type 2 diabetes. Food Research International, v. 46, n. 1, p. 378–386, abr. 2012.

SASSO, S. A. Z.; CITADIN, I.; DANNER, M. A. Propagação de jabuticabeira por estaquia. Revista Brasileira de Fruticultura, v. 32, n. 2, p. 577–583, 2010.

SCHAR, M. Y. et al. Orange juice-derived flavanone and phenolic metabolites do not acutely affect cardiovascular risk biomarkers: A randomized, placebo-controlled, crossover trial in men at moderate risk of cardiovascular disease1-5. American Journal of Clinical Nutrition, v. 101, n. 5, p. 931–938, 2015.

SEERAM, N. P.; LEE, R.; HEBER, D. Bioavailability of ellagic acid in human plasma after consumption of ellagitannins from pomegranate (Punica granatum L.) juice. Clinica Chimica Acta, v. 348, n. 1–2, p. 63–68, 2004.

SEERAM, N. P. et al. Pomegranate juice ellagitannin metabolites are present in human plasma and some persist in urine for up to 48 hours. The Journal of nutrition, v. 136, n. May, p. 2481– 2485, 2006. 155

SEERAM, N. P. et al. Pomegranate juice and extracts provide similar levels of plasma and urinary ellagitannin metabolites in human subjects. Journal of Medicinal Food, v. 11, n. 2, p. 390–394, 2008.

SELMA, M. V. et al. Description of urolithin production capacity from ellagic acid of two human intestinal Gordonibacter species. Food and Function, v. 5, n. 8, p. 1779–1784, 2014.

SELMA, M. V. et al. The human gut microbial ecology associated with overweight and obesity determines ellagic acid metabolism. Food and Function, v. 7, n. 4, p. 1769–1774, 2016.

SELMA, M. V. et al. The gut microbiota metabolism of pomegranate or walnut ellagitannins yields two urolithin-metabotypes that correlate with cardiometabolic risk biomarkers: Comparison between normoweight, overweight-obesity and metabolic syndrome. Clinical Nutrition, v. 37, n. 3, p. 897–905, 2018.

SERAGLIO, S. K. T. et al. Nutritional and bioactive potential of Myrtaceae fruits during ripening. Food Chemistry, v. 239, p. 649–656, 2018.

SERRANO, J. et al. Tannins: Current knowledge of food sources, intake, bioavailability and biological effects. Molecular Nutrition and Food Research, v. 53, p. 310–329, 2009.

SILVA, F. V. M.; SULAIMAN, A. Polyphenoloxidase in fruit and vegetables: inactivation by thermal and non-thermal processes. Reference Module in Food Science. Amsterdã:Elsevier, 2018. 15 p.

SILVA, R. M. et al. Protective effect and induction of DNA repair by Myrciaria cauliflora seed extract and pedunculagin on cyclophosphamide-induced genotoxicity. Mutation Research - Genetic Toxicology and Environmental Mutagenesis, v. 810, p. 40–47, 2016.

SLATNAR, A. et al. Identification and quantification of phenolic compounds in kernels, oil and bagasse pellets of common walnut (Juglans regia L.). Food Research International, v. 67, p. 255–263, 2015.

SONG, K. BIN et al. Prevalence of daidzein-metabolizing phenotypes differs between Caucasian and Korean American women and girls. The Journal of nutrition, v. 136, n. 5, p. 1347–51, 2006.

STAPLEY, A. Freeze drying. In: Evans, J. A. (Ed). Frozen Food Science and Technology. Hoboken: Blackwell Publishing Ltd., 2008, 248-275.

STRATHEARN, K. E. et al. Neuroprotective effects of anthocyanin- and proanthocyanidin- .s disease. Brain research, v. 1555, p. 60–77, 2014׳rich extracts in cellular models of Parkinson

SULAIMAN, A. et al. Modeling the polyphenoloxidase inactivation kinetics in pear, apple and strawberry purees after high pressure processing. Journal of Food Engineering, v. 147, p. 89– 94, 2015.

TARASCOU, I. et al. The hidden face of food phenolic composition. Archives of Biochemistry and Biophysics, v. 501, n. 1, p. 16–22, 2010. 156

TAUSCHER, B. Pasteurization of food by hydrostatic high pressure: chemical aspects. Zeitschrift für Lebensmittel-Untersuchung und -Forschung, v. 200, n. 1, p. 3–13, 1995.

TOMÁS-BARBERÁN, F. A. et al. Ellagic acid metabolism by human gut microbiota: Consistent observation of three urolithin phenotypes in intervention trials, independent of food source, age, and health status. Journal of Agricultural and Food Chemistry, v. 62, n. 28, p. 6535–6538, 2014.

TOMÁS-BARBERÁN, F. A.; SELMA, M. V.; ESPÍN, J. C. Interactions of gut microbiota with dietary polyphenols and consequences to human health. Current Opinion in Clinical Nutrition and Metabolic Care, v. 19, n. 6, p. 471–476, 2016.

TOMÁS-BARBERÁN, F. A. et al. Urolithins, the rescue of “old” metabolites to understand a “new” concept: Metabotypes as a nexus among phenolic metabolism, microbiota dysbiosis, and host health status. Molecular Nutrition and Food Research, v. 61, n. 1, 2017.

TOMÁS-BARBERÁN, F. A.; YANG, X. Tea is a significant dietary source of ellagitannins and ellagic acid. Journal of Agricultural and Food Chemistry, doi: 10.1021/acs.jafc.8b05010, 2018.

TORRES, C. DE et al. Effect of freeze-drying and oven-drying on volatiles and phenolics composition of grape skin. Analytica Chimica Acta, v. 660, n. 1–2, p. 177–182, 2010.

TORRES-OSSANDÓN, M. J. et al. Effects of high hydrostatic pressure processing and supercritical fluid extraction on bioactive compounds and antioxidant capacity of Cape gooseberry pulp (Physalis peruviana L.). The Journal of Supercritical Fluids, v. 138, p. 215– 220, 2018.

TROIANI, E. D. P.; TROPIANI, C. T.; CLEMENTE, E. Peroxidase (POD) and polyphenoloxidase (PPO) in grape (Vitis vinifera L). Ciência e Agrotecnologia, v. 27, p. 635– 642, 2003.

TRUCHADO, P. et al. Strawberry processing does not affect the production and urinary excretion of urolithins, ellagic acid metabolites, in humans. Journal of Agricultural and Food Chemistry, v. 60, n. 23, p. 5749–5754, 2012.

TULIPANI, S. et al. Urolithins are the main urinary microbial-derived phenolic metabolites discriminating a moderate consumption of nuts in free-living subjects with diagnosed metabolic syndrome. Journal of Agricultural and Food Chemistry, v. 60, n. 36, p. 8930–8940, 2012.

VANGDAL, E. et al. Effects of the drying technique on the retention of phytochemicals in conventional and organic plums (Prunus domestica L.). LWT - Food Science and Technology, v. 85, p. 506–509, 2017.

VARELA-SANTOS, E. et al. Effect of high hydrostatic pressure (HHP) processing on physicochemical properties, bioactive compounds and shelf-life of pomegranate juice. Innovative Food Science & Emerging Technologies, v. 13, p. 13–22, 2012. 157

VASHISTH, T.; SINGH, R. K.; PEGG, R. B. Effects of drying on the phenolics content and antioxidant activity of muscadine pomace. LWT - Food Science and Technology, v. 44, n. 7, p. 1649–1657, 2011.

VÁZQUEZ-GUTIÉRREZ, J. L. et al. Changes in the microstructure and location of some bioactive compounds in persimmons treated by high hydrostatic pressure. Postharvest Biology and Technology, v. 61, n. 2–3, p. 137–144, 2011.

VÁZQUEZ-GUTIÉRREZ, J. L. et al. Changes in the structure and antioxidant properties of onions by high pressure treatment. Food & function, v. 4, n. 4, p. 586–591, 2013.

VEGA-GÁLVEZ, A. et al. High hydrostatic pressure effect on chemical composition, color, phenolic acids and antioxidant capacity of Cape gooseberry pulp (Physalis peruviana L.). LWT - Food Science and Technology, v. 58, n. 2, p. 519–526, 2014.

VERMERRIS, W.; NICHOLSON, R. Phenolic Compound Biochemistry. Dordrecht:Springer, 2006. 283 p.

VIEITES, R. L. et al. Caracterização físico-química, bioquímica e funcional da jabuticaba armazenada sob diferentes temperaturas. Revista Brasileira de Fruticultura, v. 33, n. 2, p. 362–375, 2011.

VLACHOJANNIS, C. et al. A critical evaluation of the clinical evidence for pomegranate preparations in the prevention and treatment of cardiovascular diseases. Phytotherapy Research: PTR, v. 29, n. 4, p. 501–508, 2015.

WANG, L.-S.; STONER, G. D. Anthocyanins and their role in cancer prevention. Cancer letters, v. 269, n. 2, p. 281–90, 8 out. 2008.

WANG, W. H. et al. Evaluation of the antioxidant activity and antiproliferative effect of the jaboticaba (Myrciaria cauliflora) seed extracts in oral carcinoma cells. BioMed Research International, v. 2014, p. 1–7, 2014.

WHITLEY, A. C.; SWEET, D. H.; WALLE, T. Site-specific accumulation of the cancer preventive dietary polyphenol ellagic acid in epithelial cells of the aerodigestive tract. Journal of Pharmacy and Pharmacology, v. 58, n. 9, p. 1201–1209, 2006.

WOJDYLO, A.; FIGIEL, A.; OSZMIAN, J. Effect of drying methods with the application of vacuum microwaves on the bioactive compounds, color, and antioxidant activity of strawberry fruits. Journal of agricultural and food chemistry, v. 57, n. 4, p. 1337–1343, 2009.

WOJDYLO, A.; FIGIEL, A.; OSZMIAN, J. Effect of Drying Methods with the Application of Vacuum Microwaves on the Bioactive Compounds , Color , and Antioxidant Activity of Strawberry Fruits. Journal of agricultural and food chemistry, v. 57, n. 4, p. 1337–1343, 2009.

WOJDYŁO, A. et al. Effect of convective and vacuum-microwave drying on the bioactive compounds, color, and antioxidant capacity of sour cherries. Food and Bioprocess Technology, v. 7, n. 3, p. 829–841, 2014. 158

WOJDYŁO, A. et al. Chemical composition, antioxidant capacity, and sensory quality of dried jujube fruits as affected by cultivar and drying method. Food Chemistry, v. 207, p. 170–179, 2016.

WU, S. et al. Metabolite profiling of jaboticaba (Myrciaria cauliflora) and other dark-colored fruit juices. Journal of agricultural and food chemistry, v. 60, n. 30, p. 7513–7525, 2012.

WU, S. B. et al. Bioactive and marker compounds from two edible dark-colored Myrciaria fruits and the synthesis of jaboticabin. Journal of agricultural and food chemistry, v. 61, n. 17, p. 4035–43, 2013.

XI, J.; LUO, S. The mechanism for enhancing extraction of ferulic acid from Radix Angelica sinensis by high hydrostatic pressure. Separation and Purification Technology, v. 165, p. 208–213, 2016.

YALDAGARD, M.; MOTAZAVI, S. A.; TABATABAIE, F. The principles of ultra high pressure technology and its application in food processing/preservation: a review of microbiological and quality aspects. African Journal of Biotechnology, v. 7, p. 2739–2767, 2008.

YAMADA, H. et al. Structural Revisions in Natural Ellagitannins. Molecules, v. 23, n. 8, p. 1901, 2018.

YANG, B. et al. Inactivation of foodborne pathogens in raw milk using high hydrostatic pressure. Food Control, v. 28, n. 2, p. 273–278, 2012.

YANG, J. et al. Soy protein isolate does not affect ellagitannin bioavailability and urolithin formation when mixed with pomegranate juice in humans. Food Chemistry, v. 194, p. 1300– 1303, 2016.

YUAN, B. et al. High pressure processing (HPP) of aronia berry purée: Effects on physicochemical properties, microbial counts, bioactive compounds, and antioxidant capacities. Innovative Food Science and Emerging Technologies, v. 47, p. 249–255, 2018.

ZHANG, X. et al. Urinary isoflavonoids and risk of coronary heart disease. International Journal of Epidemiology, v. 41, n. 5, p. 1367–1375, 2012.

ZHANG, Y.; BUTELLI, E.; MARTIN, C. Engineering anthocyanin biosynthesis in plants. Current opinion in plant biology, v. 19, p. 81–90, 2014.

ZHOU, M. et al. Degradation kinetics of cyanidin 3-O-glucoside and cyanidin 3-O-rutinoside during hot air and vacuum drying in mulberry (Morus alba L.) fruit: A comparative study based on solid food system. Food Chemistry, v. 229, p. 574–579, 2017.

159

ANNEXES

160

Annex 1 – Proanthocyanins MS spectra and MS2 fragmentation pattern

MS spectra (A) and MS2 fragmentation pattern (B) of proanthocyanidins after phloroglucinolysis. Compounds identified: gallocatechin aduct (1; m/z 429), catechin aduct (2; m/z 413), gallocatechin (3; m/z 305), Epigallocatechingallate aduct (4; m/z 581), epicatechin aduct (5; m/z 413), catechin (6; m/z 289), epicatechin gallate aduct (7; m/z 565), epicatechin gallate (8; m/z 441).

161

Annex 1 – Proanthocyanins MS spectra and MS2 fragmentation pattern (continued)

MS spectra (A) and MS2 fragmentation pattern (B) of proanthocyanidins after phloroglucinolysis. Compounds identified: gallocatechin aduct (1; m/z 429), catechin aduct (2; m/z 413), gallocatechin (3; m/z 305), Epigallocatechingallate aduct (4; m/z 581), epicatechin aduct (5; m/z 413), catechin (6; m/z 289), epicatechin gallate aduct (7; m/z 565), epicatechin gallate (8; m/z 441).

162

Annex 1 – Proanthocyanins MS spectra and MS2 fragmentation pattern (continued)

MS spectra (A) and MS2 fragmentation pattern (B) of proanthocyanidins after phloroglucinolysis. Compounds identified: gallocatechin aduct (1; m/z 429), catechin aduct (2; m/z 413), gallocatechin (3; m/z 305), Epigallocatechingallate aduct (4; m/z 581), epicatechin aduct (5; m/z 413), catechin (6; m/z 289), epicatechin gallate aduct (7; m/z 565), epicatechin gallate (8; m/z 441).

163

Annex 1 – Proanthocyanins MS spectra and MS2 fragmentation pattern (continued)

MS spectra (A) and MS2 fragmentation pattern (B) of proanthocyanidins after phloroglucinolysis. Compounds identified: gallocatechin aduct (1; m/z 429), catechin aduct (2; m/z 413), gallocatechin (3; m/z 305), Epigallocatechingallate aduct (4; m/z 581), epicatechin aduct (5; m/z 413), catechin (6; m/z 289), epicatechin gallate aduct (7; m/z 565), epicatechin gallate (8; m/z 441).

164

Annex 1 – Proanthocyanins MS spectra and MS2 fragmentation pattern (continued)

MS spectra (A) and MS2 fragmentation pattern (B) of proanthocyanidins after phloroglucinolysis. Compounds identified: gallocatechin aduct (1; m/z 429), catechin aduct (2; m/z 413), gallocatechin (3; m/z 305), Epigallocatechingallate aduct (4; m/z 581), epicatechin aduct (5; m/z 413), catechin (6; m/z 289), epicatechin gallate aduct (7; m/z 565), epicatechin gallate (8; m/z 441).

165

Annex 2– Consent Form

Universidade Federal do Rio de Janeiro Instituto de Nutrição Josué de Castro

TERMO DE CONSENTIMENTO LIVRE E ESCLARECIDO (TCLE)

Nome:______Contato: ______

Trabalho de pesquisa: “Metabolismo da jabuticaba em adultos com excesso de peso”

Você está sendo convidado (a) a participar como voluntário (a) de uma pesquisa cujo objetivo é estudar as mudanças que componentes naturalmente presentes na jabuticaba (compostos fenólicos) sofrem no corpo de seres humanos. Esse estudo é importante para entender melhor as ações da jabuticaba na saúde. Este estudo está sendo realizado pela nutricionista Kim Inada, sob orientação da Prof. Drª Mariana Monteiro do Instituto de Nutrição e do Prof. Dr. Daniel Perrone, do Instituto de Química, ambos da UFRJ. Participando desse estudo, dois dias antes e durante o dia de urina você será orientado a: - Beber 2 litros de água por dia; - Anotar todas as bebidas e alimentos consumidos; - Não consumir frutas, hortaliças, cereais integrais, oleaginosas (exemplo: nozes, castanhas, amendoim), cacau, chá, café, vinho e cerveja. Após passar a noite anterior, durante 10 horas, sem consumir alimentos, será oferecida uma gelatina com adição de casca e semente de jabuticaba em pó. Amostras de urina serão coletadas antes e nos períodos de 0-2; 2-4; 4-6; 6-8; 8-12; 12-24 horas após comer a gelatina. Amostras de fezes serão coletadas 48 horas antes do consumo da gelatina. Todos os dados fornecidos são considerados confidenciais, garantindo totalmente a sua privacidade. Os dados não serão divulgados de forma a possibilitar a sua identificação. Esclarecemos que o procedimento de coleta de urina não acarretará desconforto ou riscos para você. Lembramos que é muito importante a lavagem das mãos com água e sabão após a coleta da urina. Esclarecemos, ainda, que não há benefício direto para o participante, e que você será reembolsado de despesas relacionadas à pesquisa, tais como alimentação e transporte necessários para sua participação no estudo. Você terá acesso a todos os resultados dos exames de sangue e urina realizados. A sua participação na pesquisa não é obrigatória e sua recusa não acarretará nenhum prejuízo. Além disso, você poderá se retirar da pesquisa a qualquer momento, sem qualquer tipo de aborrecimento, sem que isso lhe traga qualquer prejuízo ou punição, sem necessidade de justificativa. Garantimos indenização diante de eventuais danos decorrentes da pesquisa. Em qualquer etapa do estudo, você terá acesso direto com os responsáveis pela pesquisa das seguintes formas: • Profa. Mariana Monteiro: email: [email protected]. Endereço: Av. Carlos Chagas Filho, 373, Centro de Ciências da Saúde, bloco J, segundo andar, sala 16. Cidade Universitária. Rio de Janeiro. Telefone de contato: 3938-6449. • Prof. Daniel Perrone: email: [email protected]. Endereço: Av. Athos da Silveira Ramos, 149, Centro de Tecnologia, bloco A, sala 528A. Cidade Universitária. Rio de Janeiro. Telefone de contato: 3938-8213. • Kim Inada: email: [email protected]. Telefone de contato: 3938-7351 / 99635-0151.

(continued on next page)

166

Annex 2– Consent Form (continued)

Se você tiver alguma consideração ou dúvida sobre a ética da pesquisa, entre em contato com o Comitê de Ética em Pesquisa (CEP) do Hospital Universitário Clementino Fraga Filho/HUCFF/UFRJ. O CEP é um órgão institucional que tem por objetivo proteger o bem-estar dos participantes de um estudo clínico por meio da avaliação dos projetos de pesquisa que envolvam a participação de seres humanos. O CEP está localizado na Rua Professor Rodolpho Paulo Rocco, nº 255, sala 01D-46/1º andar. Cidade Universitária/Ilha do Fundão. Rio de Janeiro. Telefone (21) 3938-2480, atendimento de segunda a sexta-feira, das 8 às 16 horas ou através do email: [email protected]. Acredito ter sido suficientemente informado (a) a respeito das informações sobre o estudo acima citado que li ou que foram lidas para mim. Eu discuti com a nutricionista Kim Inada sobre a minha decisão em participar nesse estudo. Ficaram claros para mim quais são os propósitos do estudo, os procedimentos a serem realizados, seus desconfortos e riscos, as garantias de confidencialidade e de esclarecimentos permanentes. Ficou claro também que minha participação é isenta de despesas e que tenho garantia de acesso a tratamento hospitalar quando necessário. Concordo voluntariamente em participar deste estudo e poderei retirar o meu consentimento a qualquer momento, sem penalidades ou prejuízos e sem a perda de atendimento nesta Instituição ou de qualquer benefício que eu possa ter adquirido. Eu receberei uma cópia desse Termo de Consentimento Livre e Esclarecido (TCLE) e a outra ficará com o pesquisador responsável por essa pesquisa. Além disso, estou ciente de que eu e o pesquisador responsável deveremos rubricar todas as folhas desse TCLE e assinar na última folha.

______Rio de Janeiro, ____/____/____

Nome do sujeito Assinatura do sujeito da

da pesquisa pesquisa

Declaro que obtive de forma apropriada e voluntária o Consentimento Livre e Esclarecido do voluntário para participação no estudo.

______Rio de Janeiro, ____/____/____

Nome do pesquisador Assinatura do pesquisador

responsável responsável

167

Annex 3 – Instructions for volunteers

ESTUDO DO METABOLISMO DOS COMPOSTOS DA JABUTICABA Bem-vindo ao nosso estudo! Para que o estudo seja realizado adequadamente, é essencial que você siga algumas etapas que serão descritas aqui. Se você tiver qualquer dúvida, não hesite em entrar em contato comigo.

1° ETAPA: COLETA DE FEZES  QUANDO? A qualquer momento, até o dia do estudo na UFRJ.  O QUE FAZER? Coletar as fezes no frasco universal fornecido pelo pesquisador. Seguir as instruções da coleta de fezes na página 4. A amostra de fezes deverá ser entregue no dia do estudo na UFRJ.

2° ETAPA: REGISTRO ALIMENTAR E RESTRIÇÃO DO CONSUMO DE FRUTAS VERMELHAS E ROXAS  QUANDO? Nas 96 horas (4 dias) anteriores ao estudo até o término da coleta de urina em casa.  O QUE FAZER? REGISTRO ALIMENTAR: você deve anotar todos os alimentos que você consumir, bem como suas quantidades e horários de consumo. Além disso, durante esse tempo, você NÃO poderá consumir frutas vermelhas e arroxeadas (morango, uva, açaí, cereja, romã, cranberry, amora, mirtilo, dentre outras), bem como suas bebidas derivadas (sucos, refrescos, vinhos, etc).

3° ETAPA: REGISTRO ALIMENTAR E RESTRIÇÃO DO CONSUMO DE OUTROS ALIMENTOS  QUANDO? Nas 48 horas (2 dias) anteriores ao estudo até o término da coleta de urina em casa.

(continued on next page) 168

Annex 3 – Instructions for volunteers (continued)

 O QUE FAZER? REGISTRO ALIMENTAR (= 2° etapa) e você NÃO poderá consumir os seguintes alimentos:  Alimentos de origem vegetal, por ex: - Frutas frescas ou secas (nenhum tipo); - Verduras (ex. alface, agrião, rúcula, couve, brócolis, couve-flor); - Legumes (ex. abóbora, berinjela, abobrinha, chuchu, maxixe, pimentão, etc); - Raízes e tubérculos (ex. batata, batata-doce, beterraba, cenoura, aipim, nabo, rabanete, etc.); - Leguminosas (ex. feijões, lentilha, ervilha, grão de bico, soja, tremoço, etc.); - Cereais integrais (ex. pão integral, arroz integral, macarrão integral, farinhas integrais, etc.); - Oleaginosas (ex. amendoim, nozes, castanha do pará, castanha de caju, amêndoas, avelã, etc.) - Vinho (nenhum tipo – ex. branco, tinto, seco, etc.); - Cerveja (nenhum tipo – ex. claras ou escuras); - Chás (nenhum tipo – ex. preto, verde, camomila, erva-cidreira, erva-doce); - Cacau (ex. chocolate em barra – ex. branco, ao leite, meio amargo com diferentes concentrações de cacau; achocolatados em pó – Nescau, Toddy, cacau em pó, etc.); - Café; - Produtos alimentícios contendo/ produzidos a partir destes ingredientes – exemplos:  Sucos de frutas e hortaliças e outras bebidas que contém algum destes ingredientes (ex. chás gelados com sabor de fruta, refrigerantes de frutas, etc);  Doce de frutas em calda, geleias, frutas cristalizadas ou secas (ex. damasco, ameixa).  Pães do tipo panettone e colomba (contém frutas cristalizadas) ou pães com adição de frutas (ex. pão de cenoura);  Bebidas lácteas e iogurte com adição de frutas;

(continued on next page)

169

Annex 3 – Instructions for volunteers (continued)

 Leites achocolatados e produtos com adição de chocolate (ex. Frappucino mocha ou frappucino com adição de chocolate, Toddynho, Nescau pronto para beber, chocolate quente, bombons, brigadeiro, outros docinhos ou sobremesas com adição de chocolate);  Bebidas à base de café (ex. cappuccino, frapuccino, etc.) ou sobremesas com adição de café (ex. tiramisu).  Barrinhas de cereais (contém cereais integrais e oleaginosas);

ALIMENTOS QUE SERÃO PERMITIDOS: - Carnes de boi, frango, peixe e frutos do mar (desde que em preparações que não contenham os produtos da lista anterior); - Arroz branco, pão branco e outros cereais refinados (ex. preparações com farinha de trigo branca, macarrão e massa não integral); - Leite e derivados (queijos, requeijão, iogurte, bebidas lácteas – sem adição de frutas); - Doces sem adição de alimentos da lista anterior (ex. leite condensado e doce de leite);

4° ETAPA: JEJUM NOTURNO E COLETA DE URINA EM CASA  QUANDO? No dia anterior ao estudo, à noite.

 O QUE FAZER? Você deverá realizar um jejum noturno de 10 horas. A partir do momento em que você iniciar o jejum, você deverá iniciar também a coleta de urina em casa até o dia seguinte. Essa urina deverá ser armazenada na geladeira.

5° ETAPA: COLETA DE URINA NA UFRJ  QUANDO? No dia e data marcados.

(continued on next page) 170

Annex 3 – Instructions for volunteers (continued)

 O QUE FAZER? Você deverá comparecer ao laboratório da UFRJ com: o pote da coleta da urina na noite anterior e o frasco com as fezes congeladas. Lá você será instruído sobre o que deve fazer nesta etapa do estudo, permanecendo na UFRJ por 8 horas.

6° ETAPA: CONTINUAÇÃO DA COLETA DE URINA EM CASA  QUANDO? Após o dia do estudo na UFRJ.

 O QUE FAZER? Após a coleta de urina que você será orientado a fazer no laboratório da UFRJ, você levará alguns potes para que você possa dar continuidade à coleta de urina EM CASA por mais 40 horas. Atenção às instruções da coleta de urina: 1. Colete toda a urina que você fizer até o final do estudo. Caso você se esqueça e urine no vaso sanitário, por favor me avise, pois, o estudo deverá ser interrompido. 2. Cada intervalo de coleta de urina deve ser realizada no pote correspondente! Veja as informações na tampa do pote. (Obs, nunca troque as tampas dos potes!) 3. Sempre que coletar a urina, armazene-a na geladeira.

7° ETAPA: ENTREGA DOS POTES DE URINA NA UFRJ Ao final da coleta, você deverá entregar os potes de urina de volta para o pesquisador no laboratório da UFRJ.

(continued on next page) 171

Annex 3 – Instructions for volunteers (continued)

INSTRUÇÕES PARA COLETA DE FEZES (1° etapa):

 Para amostras de consistência pastosa ou petrificada:  Para a coleta utilizar um penico, comadre ou plástico. Não deixe as fezes entrarem em contato com a água do vaso sanitário ou urina.  Coletar uma amostra das fezes com o coletor do frasco e colocar imediatamente no frasco universal. (Se você for conseguir, pode fazer a coleta diretamente no frasco universal).  Armazenar imediatamente em congelador.

 Para amostras de consistência aquosa ou liquefeitas:  A coleta deve ser realizada diretamente no frasco coletor.  Armazenar imediatamente em congelador.

172

Annex 4– Chromatogram of urolithins excreted in the urine of a subject metabotype A

LC-DAD (A) and LC-MS (B) chromatograms of urolithins excreted in the urine of a metabotype A subject. Compounds identified: uro-A-3-glucuronide (1; m/z 227), uro-A (2; m/z 403).

173

Annex 5– Chromatogram of urolithins excreted in the urine of a subject metabotype B

LC-DAD (A) and LC-MS (B) chromatograms of urolithins excreted in the urine of a metabotype B subject. Compounds identified: uro-A-3-glucuronide (1; m/z 403), iso-uro- A-3-glucuronide (2; m/z 403), uro-B-glucuronide (3; m/z 387), isouro-A (1; m/z 227), uro- A (1; m/z 227).