Universidade Federal Do Rio De Janeiro André Mesquita Magalhães

Home , Calendic acid, Punicic acid, Rumenic acid

Universidade Federal do Rio de Janeiro

André Mesquita Magalhães Costa

POMEGRANATE SEED OIL (PUNICA GRANATUM L.): CHEMICAL COMPOSITION AND PRODUCTION OF VALUE- ADDED MICROPARTICLES BY COMPLEX COACERVATION

RIO DE JANEIRO 2018 André Mesquita Magalhães Costa

POMEGRANATE SEED OIL (PUNICA GRANATUM L.): CHEMICAL COMPOSITION AND PRODUCTION OF VALUE- ADDED MICROPARTICLES BY COMPLEX COACERVATION

Tese de Doutorado apresentada ao Programa de Pós-graduação em Ciência de Alimentos, Instituto de Química, Universidade Federal do Rio de Janeiro, como requisito parcial à obtenção do título de Doutor em Ciência de Alimentos.

Orientadores: Prof. Dr. Alexandre Guedes Torres Drª Renata Valeriano Tonon

Rio de Janeiro 2018

Dedication

Aos meus pais, minha irmã e meus avós.

Ao amor infinito, carinho e toda dedicação.

Devo tudo a vocês. ACKNOWLEDGEMENTS

Agradeço imensamente a todos que me ajudaram na realização desse trabalho, principalmente meus orientadores, Alexandre e Renata, minha família, meus companheiros de laboratório, meus amigos e acima de tudo DEUS.

“Os nossos dias de luta

São as nossas noites de glória.”

Charlie Brown Jr.

“Você não sabe o quanto caminhei

Pra chegar até aqui.”

Cidade Negra RESUMO

ÓLEO DE SEMENTE DE ROMÃ (PUNICA GRANATUM L.): COMPOSIÇÃO QUÍMICA E PRODUÇÂO DE MICROPARTÍCULAS COM VALOR AGREGADO POR COACERVAÇÃO COMPLEXA

André Mesquita Magalhães Costa

Orientador: Prof. Dr. Alexandre Guedes Torres

Co-orientadora: Drª Renata Valeriano Tonon

O óleo de semente de romã (OSR) é um óleo funcional com uma composição singular em compostos bioativos, principalmente isômeros do ácido linolênico conjugado (cLnA). O encapsulamento é uma tecnologia capaz de aumentar a estabilidade do OSR e dessa forma viabilizar a sua adição em alimentos. Este trabalho apresentou 3 grandes objetivos: 1) determinar a composição físico-química detalhada de OSR comerciais provenientes da Turquia e Israel, e identificar possíveis tendências regionais relacionadas à composição em compostos bioativos (fenólicos totais, tocoferóis totais, cLnA total, β-caroteno); 2) encapsular o OSR por coacervação complexa utilizando o sistema whey protein-goma Arábica e investigar a influência da concentração de polímeros totais (Cp) e razão material de parede:óleo (WM:Oil) nas propriedades físicas, químicas e morfológicas das micropartículas; e 3) avaliar a estabilidade oxidativa em duas condições de estocagem (temperatura ambiente por 90 dias e 60 °C por 10 dias seladas com/sem vácuo) e as propriedades tecnológicas da melhor formulação de micropartícula. A aplicação das micropartículas em um potencial veículo (bebida instantânea de café com leite adicionada de micropartículas de OSR) para o consumo de OSR também foi avaliada. A caracterização dos OSR comerciais demonstrou similaridades entre as amostras quanto aos índices de qualidade, perfil de classes lipídicas e teor de ácido punícico (590 mg/g óleo). O conteúdo total de cLnA e teor total de tocoferóis foi influenciado pela origem das amostras. OSRs apresentaram um baixo teor de β-caroteno e perfis variados de compostos fenólicos e voláteis. Todas amostras apresentaram uma baixa estabilidade oxidativa, provavelmente pelo alto teor de cLnAs altamente oxidáveis. Segundo a análise de componente principal, a composição em bioativos das amostras promoveu uma discriminação preliminar baseada na origem. Em seguida, o OSR foi encapsulado, nove formulações foram testadas segundo um planejamento rotacional composto central, onde as variáveis independentes foram Cp (2,2-7,8%) e razão WM:Oil (0,5-5,0), e as variáveis dependentes, retenção de óleo, eficiência de microencapsulamento (EM), conteúdo de ácido punícico, índice de peróxido, conteúdo de umidade, atividade de água e tamanho de partícula. Cp e razão WM:Oil influenciaram a retenção de óleo, EM, conteúdo de ácido punícico, conteúdo de umidade e tamanho de partícula. Valores intermediários de Cp e WM:Oil produziram micropartículas com alta retenção de óleo e o maior conteúdo de ácido punícico, neste sentido a formulação do ponto central foi escolhida (Cp = 5%; WM:Oil = 2,75). Por fim, o processo de coacervação complexa foi comparado com um método de encapsulamento mais tradicional, spray drying. As micropartículas coacervadas apresentaram estabilidade oxidativa equivalente às micropartículas não coacervadas por 60 dias e a estocagem à vácuo reduziu a taxa de oxidação do OSR. A alta solubilidade e estabilidade térmica das micropartículas coacervadas possibilitam a sua aplicação no desenvolvimento de alimentos. A adição das micropartículas coacervadas no produto desenvolvido alterou apenas o conteúdo de umidade do mesmo. A composição físico-química detalhada do OSR foi determinada e o mesmo foi encapsulado com sucesso por coacervação complexa, se mantendo estável por 60 dias, possibilitando assim sua aplicação no desenvolvimento de produtos alimentícios com potencial funcionalidade.

Palavras chaves: análise de alimentos, composição de alimentos, óleo de semente de romã, ácido punícico, ácidos fenólicos, gama-tocoferol, compostos voláteis, origem geográfica, variação na composição química, planejamento experimental, coacervação complexa, whey protein, goma Arábica, spray drying, isômeros de cLnA, estabilidade oxidativa, estocagem, hidroperóxidos, propriedades tecnológicas e alimento funcional.

ABSTRACT

POMEGRANATE SEED OIL (PUNICA GRANATUM L.): CHEMICAL COMPOSITION AND PRODUCTION OF VALUE-ADDED MICROPARTICLES BY COMPLEX COACERVATION

André Mesquita Magalhães Costa

Supervision: Prof. Dr. Alexandre Guedes Torres

Co-supervision: Drª Renata Valeriano Tonon

Pomegranate seed oil (PSO) is a functional edible oil with a unique composition of bioactive compounds, especially conjugated linolenic acids (cLnA) isomers. The microencapsulation technique could be applied to increase PSO’s stability, enabling food addition. This work had three major objectives: 1) perform a detailed physical-chemical characterization of commercial PSOs from Turkey and Israel, and identify possible regional trends related with bioactive compounds (total phenolics, total cLnAs, total tocopherols and β-carotene) composition; 2) encapsulate PSO by complex coacervation utilizing the whey protein-gum Arabic system and investigate the influence of total polymer concentration (Cp) and wall material:oil ratio (WM:Oil) on microparticles’ physical, chemical and morphological characteristics; and 3) evaluate the oxidative stability in two storage conditions (ambient temperature during 90 days and 60 °C during 10 days with/without vacuum) and technological properties of the best PSO microparticle’s formulation. Microparticles’ application on a potentional vehicle (instant coffee latte drink added with PSO’s microparticles) for PSO consumption, was also investigated. Commercial PSOs characterization indicated that samples were similar concerning quality indexes, lipid classes and punicic acid content (ca. 590 mg/g oil). Total cLnA content and total tocopherols were influenced by samples origin. PSOs showed low contents of β-carotene and a very distinct profiles of phenolics and volatile compounds. Samples showed low oxidative stability, possibly because of cLnA’s oxidizability. Bioactive compounds data provided a marginal discrimination of PSO’s geographical origin, by principal component analysis. Next step was PSO encapsulation, nine formulations were tested according to a rotatable central composite design. Independent variables were: Cp (2.2-7.8%) and WM:Oil ratio (0.5-5.0). Oil retention, microencapsulation efficiency (ME), punicic acid content, peroxide value, moisture content, water activity and particle size were analyzed as responses. Cp and WM:Oil ratio affected oil retention, ME, punicic acid content, moisture content and particle size. Intermediate values of Cp and WM:Oil ratio produced microparticles with the highest punicic acid content and 100% oil retention, therefore central point formulation (Cp = 5%; WM:Oil = 2.75) was selected to continue in the study. In the third phase of the study, complex coacervation process was compared throughout the study with spray drying, a more traditional encapsulation method. At ambient temperature, coacervate microparticles showed similar oxidative stability as non- coacervate microparticles up to day 60. These results might be associated with coacervate microparticles’ more porous structure demonstrated by morphological analysis. At 60 °C, storage under reduced O2 concentration diminished microparticles’ oxidation ratio. Coacervate microparticles were applicable to wide range of products, due to its high solubility and thermal stability. Microparticles addition only affected product’s moisture. PSO detailed physical-chemical composition was determined and successfully stabilized by complex coacervation, showing an overall 60 days stability, enabling it application in the development of potential functional food products.

Keywords: Food analysis, food composition, pomegranate seed oil, punicic acid, phenolic acids, tocopherol, volatile compounds, geographical origin, chemical composition variation, experimental design, complex coacervation, whey protein, gum Arabic, spray drying, cLnA isomers, oxidative stability, storage, hydroperoxides, technological properties and functional food.

LIST OF FIGURES

Section Page

Chapter 1 Figure 1. Punica granatum cv ‘Bhagua’ structures. A) Pomegranate plant; B) Flowering stem; C) Developing fruit; D) Matured fruit; E) Fruit split open; white arrow: outside peel; black arrow: white septal membranes; F) Arils. ------26 Figure 2. Most important pomegranates’ cultivars. ------27 Figure 3. Structure of the cLnA isomers present in the pomegranate seed oil. ------31 Figure 4. Structures of Tocopherols presents in PSO. ------33 Figure 5. Structures of the most common phytosterols present in plants. ------34 Figure 6. Structures of sex steroids identified in PSO. ------35 Figure 7. cLnA metabolism ------41 Figure 8. Stages of Linoleic acid oxidation. ------45 Figure 9. Hydroperoxides decomposition reaction. ------46 Figure 10. Particles physical structure. ------53 Figure 11. Schematic diagram of coacervates formation by protein-polysaccharide interactions. ------55 Figure 12. Production of microencapsulated oils by complex coacervation utilizing the gelatin-gum Arabic as example ------57 Figure 13. Spray drying system ------58

Chapter 2 Figure 1. Plots of principal component analysis with the following variables: contents of total phenolics, total tocopherols, β-carotene and total cLnA. Principal component 1 (PC 1) and principal component 2 (PC 2) explained 91.3% of the total variance in this data matrix ------112

Chapter 3 Figure 1. Influence of polimer concentration (Cp) and ratio wall material:oil (WM:Oil) in coacervate yield (in parenthesis) of the experimental design formulations. * A dye (oil red) was used in formulation 1 and 2 to demonstrate the affinity of the coacervate phase to the PSO. ------129 Figure 2. Wet capsules’ micrographs of the experimental’s design formulations obtained by optical microscopy. Formulations: A) 1; B) 2; C) 3; D) 4; E) 5; F) 6; G) 7; H) 8; I) 9; J1) Coacervate; J2) Coacervate; L) Formulation 9 produced at pH 7. Wet capsules were prepared in pH 3,7, whey protein: gum Arabic ratio (2:1) and reaction time of 10 min. A, B, C, D, E, F, G, H, I, J2 and M: magnification of 400×; J2: magnification of 1000×. ------131 Figure 3. Response surface of: A) oil yield; B) punicic acid content; C) span; and D) Influence of the microencapsulation process on PSO’s conjugated linolenic acids (cLnA) isomers profile. ------139 Figure 4. Dried microparticles micrographs of the experimental’s design formulations obtained by scanning electron microscopy (SEM). Formulations: A) 1; B) 2; C) 3; D) 4; E) 5; F) 6; G) 7; H) 8; I) 9. White arrows indicates cracks and open pores. ------141 Figure 5. X-ray photoelectron spectroscopy (XPS) results. A) XPS survey spectra of wall material (WM), whey protein (WP), gum Arabic (GA), elected formulation (F9) and formulation with the lowest microencapsulation efficiency (F7); B) Elemental composition of WM, GA, WP, F9 and F7; and XPS C1s high-resolution spectra of C) WM, D) WP, E) GA, F) F9, G) F7. ------144

Chapter 4 Figure 1. PSO’s microparticles stability during 90 days of storage at ambient temperature. Moisture content (%) A); water activity (aw) B); Peroxide value (meq O2/kg oil) C); Volatile compounds content (µg/g of microparticle) D); Total cLnA content (mg/100mg of microparticle) E); γ-Tocopherol variation (%) F). * Significant different from day 0, α indicate significant difference from non- coacervate microparticles at same time. (p< 0.05; repeated measures ANOVA). Coacervate microparticles □) and non-coacervate microparticles ■). ------167 Figure 2. PSO’s microparticles profile of volatile compounds and cLnA isomers during 90 days of storage at ambient temperature. Volatile fraction profile: coacervate microparticles A) and non-coacervate microparticles B). cLnA isomer profile: coacervate microparticles C) and non-coacervate microparticles D). ------171 Figure 3. Correlation between tocopherols content and oxidative indicators during stability test at ambient temperature. cLnA content and tocopherols content A); Peroxide value and tocopherols content B); Volatile compounds content and tocopherols content C) (Pearson’s correlation). ------177 Figure 4. Coacervate microparticles’ structure during storage under ambient temperature. Time 0: A) and B); Time 60 days: C) and D); and Time 90 days: E) and F). Bar 20 µm: A), C) and E); bar 10 µm: B), D) and F). White arrows indicate microparticles fragments. ------180 Figure 5. Non-coacervate microparticles’s structure during storage under ambient temperature. Time 0: A) and B); Time 90 days: C) and D). Bar 20 µm: A), C); bar 10 µm: B), D). ------181 Figure 6. PSO’s microparticles stability during 10 days of storage at 60°C. Moisture content (%) A); water activity (aw) B); Peroxide value (meq O2/kg oil) C); Volatile compounds content (µg/g of microparticle) D) and Total cLnA content (mg/100mg of microparticle) E).* Significant different from day 0, α indicate significant difference from non-coacervate microparticles at same time; β indicate significant difference between the sealing method (under O2 vs under vaccum) in the same treatment and at same time (p < 0.05; repeated measures ANOVA). ------182 Figure 7. PSO’s microparticles profile of volatile compounds and cLnA isomers during 10 days of storage at 60 °C. Volatile fraction profile: coacervate microparticles A); coacervate microparticles under vacuum B); non-coacervate microparticles C) and non-coacervate microparticles under vacuum D). cLnA isomer profile: coacervate microparticles E); coacervate microparticles under vacuum F); non-coacervate microparticles G) and non-coacervate microparticles under vacuum H). ------189 Figure 8. Coacervate microparticles’s structure during storage under 60 °C. Time 0: A) and B); Time 5 days: sealed C) and sealed under vacuum D); and Time 10 days: sealed E), F) and sealed under vacuum G), H), I). Bar 20 µm: A), D), E) and G); bar 10 µm: B) and H) and bar 2 µm: F) and I). White arrows indicate microparticles fragments and open pores. ------192 Figure 9. Non-coacervate microparticles’s structure during storage under 60 °C. Time 0: A) and B); Time 5 days: sealed C), D) and sealed under vacuum E), F); and Time 10 days: sealed G), H) and sealed under vacuum I). Bar 20 µm: A), C), E), G) and I); bar 10 µm: B), D), F) and H). ------193

APPENDIX A Figure 1. Partial GC-MS Chromatogram of DMOX derivatives from a representative sample of PSO. Peaks were assigned as follows: 1: Punicic acid; 2: Alpha-elostearic acid; 3: Catalpic acid, 4: unidentified cLnA; 5: Beta-eleostearic acid and 6: unidentified cLnA. GC column: Omegawax-320, 30 m. ------216 Figure 2. MS spectra of punicic acid DMOX derivative. ------217 Figure 3. MS spectra of the cLnA isomers DMOX derivatives of commercial PSO. 1: Punicic acid; 2: Alpha-elostearic acid; 3: Catalpic acid, 4: unidentified cLnA; 5: Beta-eleostearic acid and 6: unidentified cLnA. GC column: Omegawax-320, 30m. 221 Figure 4. A) GC-FID chromatogram of FAME derivatives from a representative sample of PSO. B) Partial GC-FID chromatogram with emphasis on cLnA isomers elution region from a representative sample of PSO. C) Partial GC-FID chromatogram with emphasis on cLnA isomers elution region from a representative sample of PSO fortified with Beta-eleostearic standard. IS: Internal standard; GC column: Omegawax-320, 30 m. ------225

APPENDIX B Figure 1. GC-FID chromatogram of the standard 37 FAME mix; GC column: Omegawax-320, 30 m. ------227 Figure 2. HPLC-ELSD chromatograms of lipid classes analysis. A) HPLC-ELSD chromatogram of the standard (soybean oil), 1: monoacylglicerols, diacylglycerols and free fat acids elution region, and 2: triacylglycerols elution region; B) HPLC- ELSD chromatogram from a representative sample of PSO; HPLC column: C18, 250 × 4.6 m. ------228 Figure 3. GC-FID chromatograms of volatile composition analysis. A) Partial GC- FID chromatogram of the standard (C7-C30 saturated alkanes). B) GC-FID chromatogram from a representative sample of PSO; GC column: 007-5, 30m. ------229 Figure 4. HPLC-DAD chromatograms of phenolic compounds analysis. A) HPLC- DAD chromatogram of the standard (containing 17 phenolic compounds, other possible compounds were injected alone); B) HPLC-DAD chromatogram from a representative sample of PSO; 1: gallic acid, 2: 3,4-dihidroxy-phenylacetic acid, 3: 5-caffeoylquinic acid, 4: p-hydroxy-benzoic acid, 5: 4-hydroxyphenylacetic acid, 6: vanillic acid, 7: caffeic acid, 8: syringic acid, 9: 2,4-hydroxy-benzoic acid, 10: p- coumaric acid, 11: ferulic acid, 12: rutin, 13: hydroxycinnamic, 14: salicylic acid, 15: rosmarinic acid, 16: naringenin, 17: kaempferol, 18: m-coumaric acid, 19: trans-cinnamic acid, and 20: quercetin; HPLC column: C18, 250 × 4.6 m. ------230 Figure 5. HPLC-FLU chromatograms of tocopherols quantification. A) HPLC-FLU chromatogram of the standard (1:α-tocopherol, 2:β-tocopherol, 3:γ-tocopherol, and 4: δ-tocopherol); B) HPLC-FLU chromatogram from a representative sample of PSO.; HPLC column: Zorbax-SIL, 250 × 4.6 m. ------231 Figure 6. HPLC-DAD chromatograms of carotenoids quantification. A) HPLC- DAD chromatogram of the standard (1:β-carotene); B) HPLC-DAD chromatogram from a representative sample of PSO.; HPLC column: Zorbax-SIL, 250 × 4.6 m. ---- 232

LIST OF TABLES

Section Page

Chapter 1 Table 1. Fatty acid profile of commercial cold pressed pomegranate seed oil from Turkey. ------30 Table 2. PPAR isoforms and majors functionalities. ------36 Table 3. Factors that affect positively or negatively the lag phase duration. ------43 Table 4. Oxidation markers of fatty acid methyl esters. ------51

Chapter 2 Table 1. Quality indices (peroxide value, acid value, p-anisidine value, and refractive index) and oxidative stability index (OSI; by Rancimat®) of commercial cold-pressed pomegranate seed oils. ------98 Table 2. Fatty acid (mg/g oil; g/100 g of fatty acids)and major lipid classes (Area%) composition of commercial cold-pressed pomegranate seed oils. ------101 Table 3. Volatile compounds in commercial samples of cold-pressed pomegranate seed oil ------104 Table 4. Odor description of volatile compounds in commercial cold-pressed pomegranate seed oils. ------106 Table 5. Phenolic compounds, tocopherols and carotenoids composition (mg/100 g oil) of commercial cold-pressed pomegranate seed oils. ------110

Chapter 3 Table 1. Oil retention, microencapsulation efficiency (ME), punicic acid content, peroxide value, moisture, water activity (aw), particle size, for the 11 trial of the experimental design. ------133 Table 2. Coded second-order regression coefficients, F values and determination coefficients (R2) for oil yield, microencapsulation efficiency (ME), punicic acid content and particle size scattering index (Span). ------134

Chapter 4 Table 1. Microencapsulation efficiency (ME) and particle size results of the coacervate and non-coacervate microparticles. ------166 Table 2. Volatile compounds identified in the PSO’ microparticles during 90 days at ambient temperature. ------172 Table 3. Volatile compounds identified in the PSO’ microparticles during 10 days at 60 °C sealed in atmospheric condition or with reduction of O2. ------185 Table 4. PSO’s microparticles and instant coffee latte drink added with PSO’s microparticles technological properties (hygroscopicity, solubility, wettability, bulk density, moisture water activity (aw) and thermal stability). ------201

APPENDIX A Table 1. Characteristic fragments (m/z) of selected conjugated trienoic fatty acids. - 217

LIST OF ABREVIATIONS

ANOVA: Analysis of Variance;

AOCS: American Oil Chemists' Society; cLA: Conjugated linoleic acid; cLnA: Conjugated linolenic acid;

Cp: Total polymer concentration;

DMOX: Dimethyloxazoline fatty acid derivatives;

ELSD: Evaporative light-scattering detector;

FAD: Fatty acid dessaturases;

FADX: Fatty acid conjugases;

FAME: Fatty acid methyl esters;

FID: Flame ionization detector;

GA: Gum Arabic;

GC: Gas chromatography;

HPLC: High pressure liquid chromatography;

IBD: Inflammatory bowel disease;

LRI: Linear retention index;

ME: Microencapsulation efficiency;

MS: Mass spectrometry;

NEC: Necrotizing colitis;

PCA: Principal component analysis;

PDA: Photometric diode array detector; pHc: critical pH; pHopt: optimum pH; pI: Isoelectric point;

PSO: Pomegranate seed oil; PPAR: Peroxisome proliferator-activated receptors;

PV: Peroxide value;

RF: Fluorescence detector;

SEM: Scanning electron microscopy;

SPME: Solid phase microextraction;

UV: Ultraviolet;

WP: Whey protein;

WPI: Whey protein isolate;

XPS: X-ray photoelectron spectroscopy;

SUMMARY

Section Page

Chapter 1 1. Introduction ------21 2. Objectives ------23 3. Thesis structure ------24 4. Literature review ------25 4.1 Pomegranate seed oil ------25 4.1.1 Pomegranate ------25 4.1.1.1 Plant description ------26 4.1.1.2 Cultivars ------26 4.1.2 Pomegranate chemical composition ------27 4.1.3 Economical application ------28 4.1.4 Pomegranate seed oil production ------28 4.1.5 Pomegranate seed oil chemical and phytochemical composition ------29 4.1.5.1 Fatty acid profile ------29 4.1.5.2 Volatile fraction composition ------32 4.1.5.3 Tocopherols ------32 4.1.5.4 Phenolics compounds ------33 4.1.5.5 Sterols ------34 4.1.5.6 Sex sterols ------35 4.1.6 Pomegranate seed oil bioactivity 35 4.1.6.1 Effects on body weight, fat metabolism, blood glucose, insulin 37 sensivity and blood lipids ------4.1.6.2 Anti-cancer activity ------38 4.1.6.3 Anti-inflammatory activity ------39 4.1.6.4 Antioxidant activity ------39 4.1.7 cLnA metabolism ------41 4.1.8 PSO toxicity ------42 4.1.9 PSO food application ------42 4.2. Lipid oxidation ------43 4.2.1 Concept ------43 4.2.2 Mechanisms ------44 4.2.3 cLnA oxidation ------46 4.2.4 Methods for measuring oxidative rancidity in oils ------47 4.2.4.1 Sensory analysis ------47 4.2.4.2 Primary products ------47 4.2.4.2.1 Oxygen ------47 4.2.4.2.2 Fatty acids quantification ------48 4.2.4.2.3 Peroxide value (PV) ------48 4.2.4.2.4 Conjugated dienes ------49 4.2.4.3 Secondary products ------49 4.2.4.3.1 2-Thiobarbituric acid value (TBA) ------49 4.2.4.3.2 p-Anisidine value (p-AnV) ------49 4.2.4.3.3 TOTOX value ------50 4.2.4.3.4 Volatile compounds ------50 4.2.4.3.5 Volatile acids ------52 4.2.4.3.6 Triacylglycerol dimers and higher oligomers ------52 4.3. Encapsulation ------52 4.3.1 Concept ------52 4.3.2 Complex coacervation ------54 4.3.2.1 Concept ------54 4.3.2.2 Hardening methods ------57 4.3.2.2.1 Spray drying ------58 4.3.2.3 Whey protein-gum Arabic system ------58 4.3.2.3.1 Wall material ------60 4.3.2.3.1.1 Whey protein ------60 4.3.2.3.1.2 Gum Arabic ------60 4.3.2.3.2 Encapsulation of oils with whey protein-gum Arabic system - 61 4.4. Microparticles characterization ------61 4.4.1 Microencapsulation efficiency ------61 4.4.2 Particle size and distribution ------61 4.4.3 Surface morphology ------63 4.4.4 Thermal characterization ------63 4.4.5 Moisture content, water activity and hygroscopicity ------64 4.4.6 Solubility and wettability ------64 4.4.7 Density ------64 4.5 Experimental Design ------64 5. References ------66

Chapter 2 Abstract ------87 1. Introduction ------88 2. Materials and Methods ------89 2.1 Samples and Reagents ------89 2.2 Pomegranate Seed Oil Quality Indexes and Oxidative Stability ------90 2.3 Fatty acid composition by GC-FID ------90 2.4 Lipid classes distribution by HPLC-ELSD ------91 2.5 Volatile Compounds by SPME-GC-MS/FID ------91 2.6 Bioactive Compounds in Pomegranate Seed Oil ------93 2.6.1 Phenolic compounds by HPLC-PDA ------94 2.6.2 Tocopherols, chlorophylls and carotenoids by HPLC-FLU/PDA ------95 2.7 Statistical analysis ------96 3. Results and Discussion ------97 3.1 Quality indexes and oxidative stability ------97 3.2 Chemical composition ------99 3.3 Bioactive compounds ------108 3.4 PSO samples grouping: Multivariate data analysis ------111 4. Conclusions ------112 Acknowledgements ------113 5. References ------113

Chapter 3 Abstract ------120 1. Introduction ------121 2. Material and Methods ------123 2.1 Materials ------123 2.2 Production of the PSO’s microparticles by complex coacervation ------123 2.3 Drying of the PSO’s microparticles ------124 2.4 Wet microparticles analysis ------124 2.4.1 Coacervation yield ------124 2.4.2 Morphology ------124 2.5 Dried microparticles analysis ------125 2.5.1 Oil retention ------125 2.5.2 Microencapsulation efficiency (ME) ------125 2.5.3 Punicic acid content ------125 2.5.4 Peroxide Value ------126 2.5.5 Moisture content ------126 2.5.6 Water activity (aw) ------126 2.5.7 Particle size ------127 2.5.8 Microparticles Morphology ------127 2.5.9 X-ray photoelectron spectroscopy (XPS) of microparticles ------127 2.6 Statistical analysis ------128 3. Results and Discussion ------128 3.1 Wet microparticles ------128 3.2 Dried microparticles ------132 3.3 Selection of the most attractive formulation ------142 3.4 XPS analysis ------142 4. Conclusion ------145 Acknowledgements ------145 5. References ------146

Chapter 4 Abstract ------153 1. Introduction ------154 2. Material and Methods ------155 2.1 Materials ------155 2.2 Production of the PSO’s microparticles by complex coacervation ------156 2.3 Drying of the PSO’s microparticles ------156 2.4 Physical-chemical characterization ------157 2.4.1 Microencapsulation efficiency (ME) ------157 2.4.2 Particle size ------157 2.5 Stability test ------158 2.5.1 Moisture and water activity (aw) ------158 2.5.2 Encapsulated oil extraction ------159 2.5.3 Peroxide Value ------159 2.5.4 cLnA isomer profile and quantification ------159 2.5.5 Volatile Compounds by SPME-GC-MS ------160 2.5.6 Tocopherols by HPLC-FLU ------161 2.5.7 Morphology ------162 2.6 PSO’s microparticles technological properties ------162 2.6.1 Hygroscopicity ------162 2.6.2 Solubility ------163 2.6.3 Wettability ------163 2.6.4 Bulk Density ------163 2.6.5 Moisture and water activity (aw) ------163 2.6.6 Thermal stability ------164 2.7 Application of the PSO’s microparticles in an instant coffee latte drink ------164 2.7.1 Production of the instant coffee latte drink ------164 2.7.2 Products technological properties ------164 2.8 Statistical analysis ------165 3. Results and Discussion ------165 3.1 Physical-chemical characterization ------165 3.2 Stability test ------167 3.3 Technological properties ------194 3.4 Instant coffee latte drink added with PSO’s microparticles ------198 4. Conclusion ------202 Acknowledgements ------202 5. References ------203

Chapter 5 1. Global Conclusions ------213 2. Suggestions for future studies ------215

APPENDIX A ------216

APPENDIX B ------227

APPENDIX C ------233

Chapter 1

Introduction, Objectives, Thesis Structure and Literature Review

1. Introduction

Pomegranate (Punica granatum L.) is a shrub tree native from Iran, but highly adaptable to different climate conditions. Pomegranate’s fruit has three major parts: the outside peel, the inside peel (white septal membranes) and the edible arils, which are juice- containing sacs formed by pomegranate’s seeds and pulp. Pomegranate’s pulp is sweet and highly appreciated for its flavor and high contents of phenolic acids and anthocyanins. The major part of pomegranate production is allocated to the industry for transformation into edible products (juice, nectars, jams and jellies). However during this process a large quantity of byproducts rich in bioactive compounds, especially seed and peels, is produced (Teixeira da Silva et al., 2013).

The seeds are a rich source of lipids, which correspond to 12-20% of total seed weight (Elfalleh et al., 2011; Fernandes et al., 2015). Pomegranate seed oil (PSO) is known for its high contents of tocopherols, sterols and phenolic compounds. However the described functional activities, such as cytotoxic effect, antioxidant properties, modulation of the immune system and anti-diabetic properties, seems to be associated with PSO’s singular fatty acid profile, composed chiefly by conjugated linolenic acid isomers (cLnA), specially punicic acid (~ 70% of total fatty acids in the oil) (Elfalleh et al., 2011; Fernandes et al., 2015; Shabbir et al., 2017).

cLnA is a collective term for the positional and geometric isomers of linolenic acid (C18:3), characterized by the presence of three conjugated double bounds, usually in positions Δ9,11,13, and Δ8,10,12 and with varying combinations of geometrical configurations, cis or trans (Cao et al., 2006). Although cLnA isomers are emerging as a potential bioactive nutrient, studies in humans evaluating the health benefits of its acute or chronic consumption are still scarce, because these compounds, especially punicic acid, are restricted to PSO and Trichosanthes kirilowii seed oil (Shabbir et al., 2017). Thus the developing of food products using PSO as a functional ingredient could enable the design of future studies assessing the health effects of oil consumption. Nevertheless, direct addition of PSO aiming at supplement food products is limited by oil’s hydrophobic nature and high susceptibility of cLnA isomers to lipid oxidation when exposed to oxygen (Yang, Cao, Chen, & Chen, 2009). Therefore before food application, PSO should be protected in order to preserve its physical and chemical stability, avoiding oxidative rancidity and nutritional losses. 22

Microencapsulation is a “packing” technique in which an active ingredient is covered by a wall material, being often used to protect unstable molecules from interaction with other components and the adjacent environment during food processing and storage (Gouin, 2004). Encapsulation processes, such as spray drying and freeze drying, have already been applied on PSO to increase its oxidative stability, producing microparticles with appealing technological properties, such as controlled release, high core retention and solubility (Goula & Adamopoulos, 2012; Gupta, Ghosh, Maiti, & Ghosh, 2012; Sahin-Nadeem & Özen, 2014). Nevertheless, to the best of our knowledge, microencapsulation of PSO by complex coacervation has not been studied yet. Complex coacervation is a physical-chemical method that has been successfully used to encapsulate hydrophobic materials (Eratte et al., 2015; Weinbreckt, Minorf, & Kruif, 2004). It consists of a liquid-liquid phase separation phenomenon that occurs when electrostatically opposite charged biopolymers, usually a protein and a polysaccharide, are subjected to specific conditions, producing aggregates (coacervates) that promptly deposit on the oil droplets (Weinbreckt & Minorf, 2004). Compared to other encapsulation techniques, complex coacervation is able to produce microparticles with higher microencapsulation efficiency utilizing high core load and low wall material concentration (Gouin, 2004).

The stabilization of PSO by a low cost encapsulation technology, as well as the characterization and the stability investigation of the produced microparticles might contribute to the development of a stable and versatile food ingredient capable to be applied in a wide range of food products. Possibly this approach would aloud the design of new foodstuff with added PSO, increasing the penetration of this nutrient in population nutritional habits. Furthermore, this initiative aims at add value to a byproduct of the juice industry. Therefore, the objectives of the present work were as follows:

2. Objectives

2.1 Main objective

• Perform a detailed physical-chemical characterization of pomegranate seed oil (PSO) and investigate the stability of a designed potential functional food ingredient produced by complex coacervation.

2.2 Specific objectives

Chapter 2 • Perform a detailed chemical characterization of commercial PSO samples from Turkey and Israel, with emphasis on fatty acid profile, lipid classes, volatile compounds profile and bioactive compounds (phenolics, tocopherols and carotenoids), as well as quality indexes and oxidative stability; • Identify possible regional trends in commercial PSOs’ bioactive compounds profiles (total phenolics, total tocopherols, β-carotene and total cLnA) by applying Principal component (PCA);

Chapter 3 • Apply a experimental design to evaluate the influence of total polymer concentration

(Cp) (2.17-7.82%) and wall material:oil ratio (WM:Oil ratio) (0.48-5.01) on the microencapsulation of PSO by complex coacervation, using the whey protein-gum Arabic system. Oil retention, microencapsulation efficiency, punicic acid content, peroxide value, moisture content, water activity and particle size were analyzed as responses;

• Investigate the influence of Cp and WM:Oil ratio on microparticles morphology (wet and dried) and coacervation yield; • Evaluate microparticles surface chemical composition; • Select the best formulation of microparticle based in the experimental design responses;

Chapter 4 • Investigate microparticles’ oxidative stability in two storage condition: ambient temperature during 90 days and 60 °C during 10 days with/without vacuum; • Determinate microparticles’ technological properties; 24

• Propose a vehicle for PSO’s microparticles consumption (food product) and evaluate the influence of microparticles addition on product original technological properties; • During the stability test and food design stages, the complex coacervation process was compared throughout the study with spray drying, a more utilized encapsulation method.

3. Thesis structure

The present thesis is divided in 5 chapters, as follows:

The present Chapter (Chapter 1) is dedicated to thesis’ introduction, objectives, structure and literature review concerning the principal topics discussed in this work;

Chapter 2 is the accept manuscript submitted to Food Composition and Analysis (Costa, A. M. M., Silva, L. O., & Torres, A. G. 2019. Chemical composition of commercial cold-pressed pomegranate (Punica granatum) seed oil from Turkey and Israel, and the use of bioactive compounds for samples’ origin preliminary discrimination. Journal of Food Composition and Analysis, 75, 8–16). This chapter was dedicated to chemical composition investigation of three commercial pomegranate seed oils (PSO) from Turkey and Israel, with emphasis on PSO’s bioactivity components and quality parameters, such as: fatty acid profile, lipid classes, volatile compounds, quality indexes, oxidative stability and bioactive compounds (phenolics, tocopherols and carotenoids). Furthermore, Principal component analysis (PCA) was used to identify regional trends in PSOs, based on bioactive compounds profile;

Chapter 3 was structured according to Food Chemistry’s guidelines aiming at submission in a near future. This chapter was dedicated to the evaluation of the influence of total polymer concentration (Cp) and wall material:oil ratio (WM:Oil) on the microencapsulation of pomegranate (Punica granatum) seed oil (PSO) by complex coacervation (CC), using whey protein-gum Arabic as wall materials and spray drying as hardening/recuperation step;

Chapter 4 was dedicated to the assessment of the oxidative stability (ambient temperature during 90 days and 60 °C during 10 days with/without vacuum) and technological properties of the best PSO microparticles formulation produced by complex 25

coacervation. A vehicle for microparticles consumption, instant coffee latte drink added with PSO’s microparticles, was developed and the influence of microparticles addition on product’s technological properties was also evaluated. Moreover, complex coacervation process was compared throughout the study with spray drying, a more utilized encapsulation method;

Chapter 5 was dedicated to thesis’ Global conclusions and suggestions for future studies.

4. Literature review

4.1 Pomegranate seed oil

4.1.1 Pomegranate

The pomegranate is a member of the Punicaceae family, owning a scientific name (Punica granatum), which is an allusion to seeded apple, Pomum (apple) and granatus (grainy). Pomegranate’s use as raw fruit or herbal healer is described in ancient cultures and mythologies, such as manuscripts of Hippocrates, Pliny, Soranus and Dioscorides, and even in the Bible. However, there is now an increase awareness of pomegranate’s different parts potentional bioactivity (Teixeira da Silva et al., 2013).

Pomegranate is native to Persia and surrounding areas, it is believed that this plant was initially farmed in central Asia, specifically Iran, from where it spread worldwide (Teixeira da Silva et al., 2013). Commercial production is concentrated in semi-arid and arid climates, but once established pomegranate shows a high tolerance to different weathers, soil qualities and water availability (Melgarejo, 2003). Currently the biggest world producers are India, China and Iran, which are also major consumers of this production, followed by Turkey, Spain, Tunisia and Azerbaijan. Whereas in the last decades, United States, Israel, South Africa, Peru, Chile and Argentina have emerged as new markets of production and commerce (Cambici, 2001). In Brazil, commercial production is recent in the northeast, central-west and southeast of the country (Líder agronomia, 2012).

4.1.1.1 Plant description

Pomegranate plants are shrubs or small trees (5-10 m), however dwarf plants are also frequent (1-2 m). It shows smooth stem with dark grey bark and sometimes spiny branches. Leaves are bright green and flowers are brightly-red with terminal or axillary disposition (Lawrence, 1951). The fruit has three major parts: the outside peel, the inside peel (white septal membranes) and the edible arils, which are brightly-red juice-containing sacs formed by pomegranate’s seeds and pulp. Depending of the pomegranate variety the arils represents 49 to 66% in fruit weight (Fernandes et al., 2015) (Figure 1).

A B C

E F D

Figure 1. Punica granatum cv ‘Bhagua’ structures. A) Pomegranate plant; B) Flowering stem; C) Developing fruit; D) Matured fruit; E) Fruit split open; white arrow: outside peel; black arrow: white septal membranes; F) Arils. (Adapted from: Rana, Narzary, & Ranade, 2010)

4.1.1.2 Cultivars

There are more than 500 cultivars of pomegranate identified, however depending of the harvesting region pomegranates with similar genotype characteristics might be 27

recognized by different names (IPGRI, 2001). The cultivar Wonderful is one of the most farmed in United States, being the principal commercial cultivar in this region. The fruit posses a brightly-red coloration in the outside peel and arils. Moreover, qualities, such as: high contents of soluble sugar (19%) and arils per 100 g of fruit (61%), makes this cultivar very valuable in the juice industry (Fernandes et al., 2015). On the other hand, in the western Europe, especially Spain, Molar de Elche and Valenciana are the major cultivated cultivars. Although with a less brightly-red coloration, these variations are also highly appreciated in the juice industry (Figure 2).

Molar Del Elche

Valenciana

Wonderful

Figure 2. Most important pomegranates’ cultivars (Adapted from: Fernandes et al., 2015)

4.1.2 Pomegranate chemical composition

Pomegranate fruit is a large berry and its outer peel and edible part represents 31-50% and 54-66% of fruit weight, respectively (Fadavi, Barzegar, Azizi, & Bayat, 2005; Fernandes et al., 2015).

Pomegranate peels are rich in phenolics compounds, such as: flavonoids, ellagitannins (especially Punicalin and Punicalagin) and proanthocyanidin. Complex polysaccharides and minerals (K, Ca, Mg, P and Na) have already been described in this pomegranate’s part (Fadavi et al., 2005; Viladomiu, Hontecillas, Lu, & Bassaganya-Riera, 2013). 28

The juice is a recognized source of phenolics compounds, mainly anthocyanins, which gives the color to the fruit. Moreover ascorbic acid, tocopherols, coenzyme Q10 and simple sugars (fructose and glucose) are also present (Fadavi et al., 2005; Sreekumar, Sithul, Muraleedharan, Azeez, & Sreeharshan, 2014; Viladomiu et al., 2013).

The seeds are a rich source of lipids, which correspond to 12-20% of total seeds weight (Elfalleh et al., 2011; Fernandes et al., 2015). The most important phytochemicals presents in the seeds are: linolenic acid isomers, tocopherols, phytosterols and sex steroids (Caligiani, Bonzanini, Palla, Cirlini, & Bruni, 2010; Khoddami, Man, & Roberts, 2014). Pomegranate seed oil chemical composition will be described in details later.

4.1.3 Economical application

The major part of pomegranate production is allocated to the industry for transformation into edible products, such as: juice, nectars, jams and jellies. However during this process a large quantity of pomegranate waste, composed chiefly by pomegranate’s seed and peels, is produced. The large amount of by-products from this economical activity arises concerning due to environment impact and the lost of components with high nutritional value and therapeutic properties. In the past years, aiming at increasing the process sustainability, many researchers have evaluated the potential application and bioactive properties of pomegranate’s discarded parts (He et al., 2011; Pande & Akoh, 2009; Sreekumar et al., 2014; Viladomiu et al., 2013). In this context, the pomegranate seed oil (PSO), an oil with a unique fatty acid profile and described anti-obesity, anti-cancer, anti-inflammatory and antioxidant activities, has emerged as a functional oil (Shabbir et al., 2017).

4.1.4 Pomegranate seed oil production

The commercial production of PSO largely relies on the extraction by cold pressing, an environmental friendly process, which shows extraction efficiency around 7% in a dry- weight basis (Khoddami et al., 2014). Other extraction methods such as: superheated solvent, supercritical CO2, microwave irradiation, Soxhlet, among other, have been already applied in PSO production, but only in laboratory scale (Abbasi, Rezaei, Emamdjomeh, & Ebrahimzadeh Mousavi, 2008; Abbasi, Rezaei, & Rashidi, 2008; Eikani, Golmohammad, & Homami, 2012; Kýralan, Gölükcü, & Tokgöz, 2009)

4.1.5 Pomegranate seed oil chemical and phytochemical composition

4.1.5.1 Fatty acid profile

PSO is composed mainly by unsaturated fatty acids (Punicic acid < Linoleic acid < Oleic acid), therefore it is highly recommended for human consumption when compared to other vegetable oils, due to low contents of pro-atherogenic fatty acids such as C14:0 and C16:0, and high contents of health promoting fatty acids (Costa, Silva, & Torres, 2019). The fatty acid profile of a commercial cold pressed PSO from Turkey is described in Table 1, nevertheless is important emphasize that this parameter is highly influenced by pomegranate’s cultivar and origin (Costa et al., 2019; Elfalleh et al., 2011; Kýralan et al., 2009) .

Table 1. Fatty acid profile of commercial cold pressed pomegranate seed oil from Turkey Fatty acids Commercial cold pressed PSO from Turkey

Saturated fatty acids (mg/g oil) 16:0 19.9 + 0.56 18:0 16.4 + 0.75 20:0 5.36 + 0.48

Unsaturated fatty acids (mg/g oil) 18:1 n-9 40.0 + 1.56 18:1 n-7 3.62 + 0.15 18:2 n-6 46.7 + 1.72 20:1 n-9 6.95 + 0.87

Conjugated linolenic acid isomers (mg/g oil) Punicic 606.2 + 54.1a α-Eleostearic 60.4 + 5.64 Catalpic 47.9 + 3.59 β-Eleostearic 14.1 + 1.70

Fatty acids groups (g/100 g fatty acids) cLnA 84.3 + 0.62a UFA 95.3 + 0.19 PUFA 89.6 + 0.37 MUFA 5.71 + 0.18

cLnA: conjugated linolenic acid; MUFA: monounsaturated fatty acids; PUFA: polyunsaturated fatty acids; UFA: unsaturated fatty acids. (Adapted from: Costa et al., 2019)

The high content of conjugated linolenic acid (cLnA) isomers in the PSO, gives to this oil a unique fatty composition. Although other species’ seed oils also contains punicic acid, namely Ecballium elaterium (22%), Fevillea trilobata (30%), Trichosanthes kirilowii (40%) T. anguina (43%), T. bracteata (42%), Mormordica balsamica (50%), T. nervifolia (52%), PSO is by far the richest natural source (60-70%) (Holic et al., 2018; Joh, Kim, & Christie, 1995).

cLnA is a collective term for the positional and geometric isomers of linolenic acid possessing three conjugated double bounds, usually in positions Δ8,10,12; Δ9,11,13, and Δ10,12,14 and with varying combinations of geometrical configurations, cis or trans. As mentioned before, Punicic acid (c9,t11,c13-18:3) is the major cLnA, followed by α- Eleostearic (c9,t11,t13-18:3) acid, Catalpic acid (t9,t11,c13-18:3) and β-Eleostearic (t9,t11,t13-18:3) (Cao et al., 2006) (Figure 3). A drastic modification in this profile might indicate thermal abuse during the extraction process or storage (Sassano et al., 2009).

Figure 3. Structure of the cLnA isomers present in the pomegranate seed oil. (Adapted from: Tran et al., 2010)

Pomegranate seeds are able to accumulate cLnA isomers due to an unique cellular machinery composed by fatty acid desaturases (FADs) and fatty acid conjugases (FADXs). In the seeds, more specifically in the cells’ endoplasmatic reticulum, Oleic acid is first desaturated to Linoleic acid (c9,c12-18:2), then α-Linolenic (c9,c12,c15-18:3) by the sequential action of FAD2 and FAD3, respectively. The successive formation of cLnAs is 32

catalized by FADXs. These enzymes act in the conversion of the double bounds of Linoleic acid in position 12 into 2 conjugated double bonds in positions 11 and 13, producing the cLnA isomers. In the oil these bioactive fatty acids are distributed majorly in position sn-2 of the triacylglycerol molecules, thus suggesting a good bioavailability (Holic et al., 2018; Kaufman & Wiesman, 2007).

4.1.5.2 Volatile fraction composition

PSO volatile composition is yet to be fully unraveled, recent findings indicate the presence of alcohols, aldehydes, ketones, esterers and carboxylic acids. Additionally, as expected, PSO volatile profile is highly influenced by pomegranate’s cultivar and origin ( (Costa et al., 2019; Karaman et al., 2015; Khoddami et al., 2014).

The investigation of PSO volatile fraction might be a promising tool to indicate oil identity, oxidative stability and consumer acceptance.

4.1.5.3 Tocopherols

Tocopherols are the most important antioxidants naturally occurring in vegetable oils. This type of antioxidant behaves like a chain-breaking electron donor antioxidant by competing with the substrate for peroxyl oxidation reactions, thus stopping lipid peroxidation chain reaction. Tocopherols react more promptly to peroxyl radicals than lipids, forming a resonance stabilized radical that does not propagate the chain reaction (Reische, Lillard, & Eitenmiller, 2008).

Tocopherols in PSO can occur in 4 different homologs (α,β,γ and δ), which are discriminate by the extent of methylation of the chroman ring containing a saturated phytyl side chain (Figure 4) (O’Keefe, 2008). γ-Tocopherol is consistently reported as the most concentrated tocol in PSO (Caligiani et al., 2010; Costa et al., 2019; Fernandes et al., 2015;. Liu, Xu, Gong, He, & Gao, 2012; Melo et al., 2016), however δ- and β-Tocopherols have already been identified as major tocols (Caligiani et al., 2010). Total Tocopherols’ contents in PSO can vary largely ranging from 135 to 4561 mg/100g oil depending of the pomegranate variety and extraction method (Caligiani et al., 2010; Costa et al., 2019; Fernandes et al., 2015; Liu et al., 2012; Melo et al., 2016).

Figure 4. Structures of Tocopherols presents in PSO. (Adapted from: O’Keefe, 2008)

PSO is a concentrated source of tocopherols, showing much higher contents than highly consumed vegetable oils, such as: corn (104 mg/100g), sunflower (61 mg/100g), olive (17.8 mg/100g) and flaxseed (53.4 mg/100g) (Akil et al., 2015; Carpenter, 1979; Melo et al., 2016)

4.1.5.4 Phenolics compounds

Phenolics profile of pomegranate’s peels, mesocarp, arils, seeds, leaves, and juice are well defined (Fischer et al., 2011; Pande & Akoh, 2009), however the determination of these compounds in the oil still scarce. The restrict studies which evaluated the total phenolics contents in PSO utilized a low selectivity spectrophotometric assay (Abbasi, Rezaei, Emamdjomeh, et al., 2008; Amri et al., 2017; Khoddami et al., 2014; Verardo et al., 2014).

Recently, Costa et al., (2019) demonstrated the presence of phenolics acids, quercetin and narigenin, as well as a highly compounds variability between samples of commercial cold pressed PSOs from Turkey and Israel, after evaluation of phenolics profile by HPLC-DAD. In this same work anthocyanins were not observed in the oil. In line with this results, Pande & Akoh (2009) also related phenolics acids and quercetin in pomegranate seeds from distinct pomegranates cultivars grown in Georgia.

4.1.5.5 Sterols

The sterols derived from plants are called phytosterols and by definition are compounds that retain some or all of the carbon atoms of squalene (a triterpene composed by 30 carbons) in its side chain and partitions nearly completely into the ether layer when it is shaken with equal volumes of ether and water. This definition excludes steroid hormones and bile acids (Nes & Mckean, 1977).

Although plants show a restrict number of majors phytosterols, plants’ sterol profile can be very complex. The following phytosterols: Campesterol, Stigmasterol, β-Sitosterol, Sitosterol, Δ-5-avenasterol and Citrostadienol (Figure 5), have already been identified in PSO (Amri et al., 2017; Caligiani et al., 2010; Fernandes et al., 2015; Verardo et al., 2014). Previous studies consistently showed β-Sitosterol as the major phytosterols in PSO. Depending of the pomegranate’s cultivar PSO’s phytosterols concentration can vary up to 10 folds (363-1,521 mg/100g) (Caligiani et al., 2010; Fernandes et al., 2015; Verardo et al., 2014).

From the nutritional point of view, phytosterols determination is an important matter as these compounds show protective effects against cardiovascular diseases, obesity, diabetes and cancer (Bradford & Awad, 2007).

Figure 5. Structures of the most common phytosterols present in plants. (Adapted from: Lipid library, 2018) 35

4.1.5.6 Sex sterols

Sex sterols are compounds from the same class of sterols, therefore they possess the same definition described above and are also synthesized from a Squalene molecule. Studies evaluating PSO sex sterols are still scant, but the restrict data indicates the presence of Estrone, 17-α-Estradiol, Estriol and Testosterone (Lansky & Newman, 2007).

Figure 6. Structures of sex steroids identified in PSO. (Adapted from: Lansky & Newman, 2007)

4.1.6 Pomegranate seed oil bioactivity

As previous reported PSO bioactivity is associated with oil’s high contents of cLnA isomers, especially punicic acid. cLnA mechanisms of action are thru activation of peroxisome proliferator-activated receptors (PPARs), a recognized molecular targets for drugs against type 2 diabetes, and promising new targets for the treatment and prevention of inflammatory disorders. PPARs are ligand-induced transcription factors that belong to the nuclear hormone receptor superfamily with 48 members identified in the human genome. These molecular targets regulate gene expression by biding with Retinoid X Receptor (RXR) 36

as a heterodimeric partner to specific DNA sequence elements named Peroxisome Proliferator Response Element (PPRE) (Viladomiu et al., 2013). PPARs regulate the expression of a wide range of genes involving in biologic functions such as inflammation, immunity and specially metabolism, therefore they are potent modulators of lipid and carbohydrate metabolism (Lansky & Newman, 2007). There are three PPAR isoforms (α, β/δ and γ) that shows different functionality and tissue distribution. The distinct functionality of the PPAR isoforms are described in Table 2.

Table 2. PPAR isoforms and majors functionalities. PPAR isoform Functionality Reference

α Lipid metabolism Chinetti et al., 2001

β or δ Lipid oxidation Park, Vogelstein, & Kinzler, 2001

Cell proliferation

γ Carbohydrate metabolism Chinetti et al., 2001

Modulate inflammation

(Adapted from: Viladomiu et al. 2013)

PPARα is associated with the clearance of circulating or cellular lipids by regulating gene expression involved in lipid metabolism in liver and skeletal muscle. PPARβ/δ modulates lipid oxidation and cell proliferation. Finally. PPARγ modulate inflammatory response, by antagonizing the activity of proinflammatory transcription factors, such as nuclear factor kappa-ligh-chain-enhancer of activated B cells (NF-κβ), signal transducer and activator of transcription (STAT), activator protein (AP)-1 and proinflammatory interleucines. cLnA PPAR activation and consequently gene expression is isomer specific, as isomers with different positional and geometrical configurations shows distinct activation abilities. Based in the potentional health benefits of PPARs activation, naturally occurring agonists might be an promising strategy to treat and prevent a wide range of diseases, such as: diabetes, obesity and intestinal inflammation (Viladomiu et al., 2013);

4.1.6.1 Effects on body weight, fat metabolism, blood glucose, insulin sensivity and blood lipids

cLA isomers anti-obesity properties have been already extensively described (Khanal, 2004) and due to structural similarities among conjugated fatty acids, cLnA isomers anti- obesity action is also a matter of interest. An in vitro study with mature adipocytes, demonstrated a dose dependent activation of PPARα and γ by a mixture of cLnA isomers (punicic acid and rumelenic acid). Treated adipocytes showed a decrease of triacylglycerol accumulation, therefore authors proposed that this result was consequence of a higher lipid mobilization in adipose tissue, as main enzymes involved in this process, such as hormone- sensitive lipase and adipose triglyceride lipase, showed higher gene expression (Miranda et al., 2011).

Studies with animal model and oral administration of PSO shows controversial results regarding weight loss. Vroegrijk et al. (2011) demonstrated weight loss in mice after providing a high-fat diet with 1% of PSO during 12 weeks. In line with these results, Hontecillas, Diguardo, Duran, Orpi, & Bassaganya-Riera (2008) demonstrated a decrease in abdominal fat deposition in mice after administration o catalpic acid (1 g/ 100 g diet) during 78 days in a high-fat diet. On the other hand, Arao et al. (2004) and Hontecillas, O’Shea, Einerhand, Diguardo, & Bassaganya-Riera (2009) did not observed weight loss or fat deposition reduction after PSO supplementation. Nevertheless, all previous studies are consistent regarding PSO health benefits related to obesity comorbidities, such as: diabetes type 2, hypertension and dyslipidemia. Previous works indicated improvement of insulin sensivity and reduction of fasting plasma glucose, as well as an increase in high-density lipoprotein, a decrease of plasma triglyceride levels and an alleviation of hepatic triacylglycerols accumulation (Arao et al., 2004; Hontecillas et al., 2008; Vroegrijk et al., 2011). Additionally, an improvement in the obesity-related inflammation was also reported by suppression of proinflammatory transcription factors, such as NF-κβ (Hontecillas et al., 2009).

Studies with human subjects are scarce, but the consumption of PSO on hyperlipidaemic subjects have been already evaluated by Mirmiran, Fazeli, Asghari, Shafiee, & Azizi (2010). This trial consisted of a double-blind placebo-controlled randomized clinical trial, where the PSO group received 400 mg of PSO twice a day during 4 weeks. PSO 38

consumption promoted encouraging effects on lipid profiles, such as triglycerides and triglycerides:HDL-C ratio.

4.1.6.2 Anti-cancer activity

PSO anti-cancer properties have already been confirmed in a wide range of cell lines, such as: breast, skin, prostate, colon, and monocytic leukemia cells (Hora, Maydew, Lansky, & Dwivedi, 2003; Kim et al., 2002; Kohno et al., 2004; Lansky, Harrison, Froom, & Jiang, 2005; Suzuki et al., 2001). PSO exert anti-proliferative effect by different mechanisms, Kohno et al. (2004) evaluated the protective effect of PSO in different dosages in colon cancer induced by azoxymethane and observed an inhition of cancer incidence and multiplicity. Authors attributed these results to a higher expression of PPARγ and a reduction of PUFA peroxidation in cell membranes of treated cells with azoxymethane. Additionally, an increase of cLA isomers in colonic mucosa lipids after PSO administration was reported, suggesting that cLnA exert anti-cancer properties by cLA. On the other hand, Suzuki et al. (2001) associated the cytotoxicity of cLna isomers in human monocytic leukemia cells to an increase of lipid peroxidation, this hypothesis was confirmed by a decrease of cytotoxicity after butylated hydroxytolune (antioxidant) adition. Another important finding reported in this study was the isomer specific anti-cancer properties, as punicic acid (9c,11t,13c-cLnA), α- eleostearic acid (9c,11t,13t-cLnA) and catalpic acid (9t,11t,13c-cLnA) showed a higher cytotoxicity than calendic acid (8t,10t,12c-cLnA).

Ornithine decarboxylase (ODC) inhibition have already been associated in PSO preventive effects against skin cancer induced by 12-O-tetradecanoylphorbol 13-acetate (TPA) (Hora et al., 2003). ODC catalyzes formation of polyamines that regulate cell growth processes, enzyme inhibition could hinder cancer proliferation (Lansky & Newman, 2007). Another mechanism reported was the ability of cLnA isomers to inhibit the growth of macrophage-like leukemia cells thought cell cycle arrest and apoptosis ( Liu & Leung, 2015).

When working with human prostate cancer, punicic acid was also efficient in inhibit cell invasion ( Lansky et al., 2005). Regarding androgen-dependent prostate cancer, punicic acid demonstrated potent growth suppression by an antiandrogenic effect and a pro-apoptotic mechanism that down-regulate antiapoptotic proteins (Bc12) (Gasmi & Sanderson, 2010) 39

Nevertheless, is important emphasize that other minor compounds presents in PSO, such as: elllagitannins and tocopherols might act synergistically with cLnA isomers in the aforementioned anti-cancer properties.

4.1.6.3 Anti-inflammatory activity

Inflammation is a natural process, but when its resolution is time delayed or the inflammation response is excessive, this could lead to cancer and immune-associated diseases. cLnA immune modulators properties in cell assays and animal studies indicated a mechanism mediated by PPARγ stimulation (Bassaganya-Riera et al., 2011; Boussetta et al., 2009; Coursodon-Boyiddle et al., 2012; Hontecillas et al., 2009; Lewis et al., 2011; Yamasaki et al., 2006). Punicic acid prevents or ameliorates experimental inflammatory bowel disease (IBD) in mice or rats by TNF-α suppression (Bassaganya-Riera et al., 2011; Boussetta et al., 2009). These results were confirmed with mouse genotype PPARγ null were protective effect of punicic acid was abrogated. In line with this results, α-eleostearic acid also showed improvement in IBD mediated thru PPARγ mechanisms (Lewis et al., 2011). PSO also showed encouraging protective effects in necrotizing colitis (NEC), a severe intestinal inflammation, which is the major morbidity and mortality cause of premature newborn. Rats fed with formula with PSO (1.5%) showed a decrease of NEC incidence, as well as improvement of enterocyte proliferation and normalization of proinflammatory mediators (IL- 6, IL-8, IL-12, IL-23 and TNF-α) (Coursodon-Boyiddle et al., 2012). Additionally, according to Hontecillas et al. (2009) PSO was efficient in suppress obesity-related inflammation by the TNF-α suppression and Yamasaki et al. (2006) demonstrated improvement in B-cell function after PSO consumption, both results were exhibit in mice.

The anti-inflammatory effects of cLnA isomers have not been evaluated in humans.

4.1.6.4 Antioxidant activity

Currently, naturally occurring antioxidant have been preferred by consumers, as synthetic antioxidants have been associated with cancer promotion and endrocrine disruption (Yanishlieva & Marinova, 2001). cLnA isomers antioxidant properties were confirmed in an in vitro study which evaluated modifications on human diabetic and non-diabetic: plasma, low-density lipoprotein and erythrocyte membrane lipids after α-eleostearic addition (0.05% and 0.1%). As expected, diabetic blood samples showed higher peroxidation and α- eleostearic, in a dose dependent manner, meliorated oxidative-status in plasma, low-density 40

lipoprotein and erythrocyte membrane lipids (Dhar et al., 2007). Authors suggested that cLnA might be an important nutrient in the prevention of cardiovascular diseases. Punicic acid antioxidant activity was reported in chemically induced colon carcinogenesis in rats (Kohno et al., 2004). PSO administration reduced PUFA peroxidation in cell membranes of treated cells with azoxymethane. In agreements with these results, PSO reduced oxidative stress markers (malondialdehyde, coenzyme Q10 and thiol groups) in kidney and heart of streptozotocin-induced diabetic rats. These findings suggested that cLnA antioxidant effects are attributed to the isomers scavengings properties, moreover main cLnA oxidation products, dimmers and polymers, are less reactive than hydroperoxides (Suzuki, Abe, & Miyashita, 2004).

Nevertheless, is important emphasize that cLnA antioxidant effects are dose dependent, as studies in humans and animal model showed pro-oxidant effects. Yuan, Sinclair, Xu, & Li (2009) showed a pro-oxidant activity after administration of 3 g/day punicic acid to 15 healthy young humans during 28 days. In rats, a dose of 0.6 g/kg of punicic acid lead to antioxidant effects, while a higher doses 1.2 g/kg were attributed with pro-oxidant activity (Mukherjee, Bhattacharyya, Ghosh, & Bhattacharyya, 2002).

Besides the aforementioned bioactive properties, PSO has also been suggested as a preventive strategy against osteoporosis and neurodegenerative disorders, such as Alzheimer’s diseases (Mizrahi et al., 2014; Spilmont et al., 2013). Nevertheless these findings need further confirmations in humans’ trials and the elucidation of the molecular events that leads to these outcomes.

The reported studies supported the potentional PSO bioactivity and its use as a functional ingredient. Shinohara et al., (2012) based in an animal model study, suggested a consumption of 2-3g cLnA/day aiming at health promotion. This extrapolation should be evaluated with caution, as rodents shows different nutrients metabolisms and absorption than humans. Therefore encouraging outcomes obtained in works performed in animals or in cell lines, should always be confirmed in human trials. Additionally, investigations in this topic might subsidize the establishment of effective PSO consumption associated with health benefits.

4.1.7 cLnA metabolism

cLnA isomers posses lower absorption than linolenic acid and cLA (Tsuzuki et al., 2006). Studies in rodent and humans indicate that part of the absorbed cLnA is quickly converted to cLA by an enzymatic Δ13-saturation (Figure 7) in small intestine, liver and kidney, indicating that cLnA might exert its bioactive effects by cLA transformation (De Melo et al., 2016; Tsuzuki et al., 2006; Yuan et al., 2009). Accumulation of non-converted cLnA isomers was observed in adipose tissue and in red blood cells membranes after consumption of PSO and Trichosanthes kirilowii seed oil, respectively

Figure 7. cLnA metabolism (Adapted from: Shabbir et al. 2017)

PSO might be used as a much concentrated naturally occurring cLA source, as PSO shows contents of cLnA varying to 600-800 mg /g of oil, while the principal natural sources of cLA, butter (2.5 mg/g of lipid) and whole milk (8.4 mg/g of lipid), demonstrate very lower contents (Costa, Silva, & Torres, 2019; Nunes & Torres, 2010). Regarding cLA supplements, which are equimolar mixtures of cLA-c9,t11, and cLA-t10,c12, PSO is also advantageous and safer, because cLnA isomers are not transformed to the isomer cLA-t10,c12. Deleterious effects (liver enlargment, hepatic steatosis, hyperinsulinemia and decreased serum levels of leptin) associated with cLA supplementation have been associated with t10,c12 isomer (Clément et al., 2002; Riserus, Arner, Brismar, & Vessby, 2002; Tsuboyama-Kasaoka et al., 2000).

4.1.8 PSO toxicity

Pomegranate is an ancient fruit and its parts have been used as a medical food in the Middle East for millennia. No adverse side effects have been associated with PSO consumption or administration in the aforementioned studies. A toxicological evaluation of pomegranate seed oil indicated that a consumption high as 4.3 g PSO/kg body weigh/day showed no observable adverse effect (Meerts et al., 2009).

4.1.9 PSO food application

PSO consumption is very discreet, in this sense the development of food products with PSO would be a plausible vehicle for its consumption and subsidize studies evaluating the health benefits associated with its intake. However, direct addition of PSO aiming at supplement food products is limited by its hydrophobic nature and by the high susceptibility of cLnA isomers to lipid oxidation when exposed to oxygen (Yang et al., 2009). Thus, these functional lipids should be protected in order to preserve their physical and chemical stability, avoiding oxidative rancidity and nutritional losses.

Few attempts to increase PSO stability have already been performed by microencapsulation techniques, such as: spray drying and freeze drying, but to the best of our knowledge PSO food application have not been evaluated yet (Goula & Adamopoulos, 2012; Gupta et al., 2012; Sahin-Nadeem & Özen, 2014).

Food enrichment with cLA isomers have been more studied and due to similarities in chemical structures, these studies could provide important information regarding cLnA expected behavior/stability in foods. The evaluation of cLA stability in food matrices focused in processing and storage (time and temperature). cLA content in a naturally enriched milk was not influenced by UHT processing (142 °C, 2 sec; 4.68 g/100g of fatty acidbefore processing vs

4.68 g/100g of fatty acidafter processing) (Jones et al., 2005). Whereas Campbell, Drake, & Larick (2003) related a decrease of cLA-c9,t11 after HTST pasteurization (77.2 °C, 16 sec) in fluid milk enriched with cLA (2%). These results imply an association between time-temperature binomial and cLA stability.

Regarding stability during storage refrigeration, cLA isomers stability can vary largely (2-10 weeks) depending of the product. cLA isomer showed a higher stability in milk, butter and cheese (8-10 week) than in fluid milk and yoghurts (~ 2 weeks) (Campbell et al., 2003; 43

Lynch et al., 2005; Mallia et al., 2008; Rodríguez-Alcalá & Fontecha, 2007; Serafeimidou, Zlatanos, Kritikos, & Tourianis, 2013).

4.2. Lipid oxidation

4.2.1 Concept

Lipid oxidation is a generic term used to describe chemicals alterations resulted from the interactions between lipids and oxygen. In the last stage of lipid peroxidation, the fatty acids esterified in the triacylglycerols and phospholipids are degraded, producing low weight volatiles molecules, although high molecular weight substances (polymers) are also described. These compounds might show toxicity and promote flavor and coloration transformations, as well as nutritional losses. The presence of lipid oxidation products in foods can make it inappropriate for human consumption or less attractive. Volatiles compounds, even in small amounts, especially fatty acids having n-3 unsaturation (ppb), are able to compromise the sensory attributes of oil containing foods. (Costa, 2014; McClements & Decker, 1996).

The kinetics of lipid oxidation in foods is composed by an initial phase (lag) that is followed by a phase which the oxidation rate increases exponentially (log). The duration of the lag phase is very important for the food industry, because during this period product quality is the highest, due to non-detected rancidity (Costa, 2014; McClements & Decker, 1996). The factors that influence lag phase duration are described in Table 1.

Table 3. Factors that affect positively or negatively the lag phase duration Positively Negatively - Low temperatures - Presence of unsaturated fatty acids - Chelates - High temperatures - Absence of water - Presence of metals - Absence of microorganisms - Presence of water - Presence of antioxidant - Presence of microorganisms - Absence of antioxidant (Adapted from: McClements & Decker, 1996)

4.2.2 Mechanisms

Lipid oxidation can be deflagrated by singlet oxygen or triplet oxygen, for singlet oxygen oxidation occurrence photosensitizer (chlorophyll, pheophytins, synyhetic colorants, riboflavin, among other) must be present, therefore oxidation mediated by oxygen triplet is the most studied. In this sense, only lipid oxidation caused by triplet oxygen will be addressed in this section (Costa, 2014; Kim & Min, 2008) .

Triplet oxygen oxidation has 3 stages: initiation, propagation and termination, which are demonstrated in Figure 8, using Linoleic acid as example. Initiation stage is defined as the production of free alkyl radicals, heat, light, metals and reactive oxygen species act as initiator of the process. The formation of alkyl radicals (R•) by hydrogen abstraction occurs in carbons, where the abstraction energy is lower. The energy required to break carbon-hydrogen bounds is highly influenced by fatty acid double bounds configurations, because adjacent doubles bounds decrease the carbon-hydrogen strength by withdrawing electrons, facilitating hydrogen abstraction. In the example, C11 has the lowest abstraction energy due to proximity to double bond in C9 and C12. Yang et al. (2009) related a lower oxidative stability of cLnA when compared with other highly oxidizable fatty acids (cLA, Linoleic acid and Linolenic acid), which might be explained by cLnA’s 3 conjugated double bounds. After production of the alkyl radical, the double bond is rearranged, in order to stabilize the carbon chain. Conjugated double bounds could be formed when chain reorganization occurs in polyunsaturated fatty acids (Costa, 2014; Kim & Min, 2008).

The alkyl radical reacts with atmospheric oxygen (triplet oxygen) producing the peroxyl radical (ROO•). The peroxyl radical easily abstract hydrogen from other fatty acid, producing hydroperoxides (ROOH), the primary product of lipid peroxidation, and another alkyl radical. The alkyl radical can start a chain reaction or the so-called propagation step by abstracting a hydrogen from another fatty acid (Costa, 2014; Kim & Min, 2008). 45

Figure 8. Stages of Linoleic acid oxidation. (Adapted from: Kim & Min, 2008)

When there is no more fatty acid available to hydrogen abstraction, the radicals combine with each other forming polymeric products, and the hydroperoxides degraded into other compounds. In the last stage, the termination step, hydroperoxides are decomposed in volatile (aldehydes, ketones, esters, carboxylic acids and alcohols) and non-volatile (short chain hydrocarbons) compounds. Hydroperoxides decomposition process was described in Figure 9. 46

Figure 9. Hydroperoxides decomposition reaction (Adapted from: Kim & Min, 2008).

4.2.3 cLnA oxidation

cLnA oxidation mechanisms still unraveled, therefore there is no established oxidation marker of cLnA isomers. The few data available indicate that cLnA containing oils are highly oxidizable and the hydroperoxides are not the major primary oxidation product, as polymer formation is verified in the beginning of the oxidation process (Miller & Claxton, 1928; L. Yang et al., 2009)

On the other hand cLA oxidation mechanisms have been extensively described and due to structures similarities might provide interesting information regarding the expected primary and secondary cLnA’s oxidation products. Previous studies involving cLA isomers indicate that conjugated fatty acids shows a distinct pattern of oxidation. According to Allen, Jackson, & Kummerow (1949) and Márquez-Ruiz, García-Martínez, Holgado, & Velasco, (2014) cLA autoxidation generates less hydroperoxides than its non-conjugated isomer (Linoleic acid) and, consistently with was observed with cLnA, polymer formation is related in the beginning of the oxidation process. Furthermore, García-Martínez, Márquez-Ruiz, 47

Fontecha, & Gordon (2009) evaluated the volatile profile of a cLA rich oil and suggested heptanal as a oxidation marker.

4.2.4 Methods for measuring oxidative rancidity in oils

Lipid oxidation promotes physical-chemicals and sensory modifications in oils, these modifications can be measure in order to determine samples’ oxidative stage. However, there is not a standard method suitable to all source of sample and able to detect all possible alterations. In this sense, the available methods are divided in 2 groups, as following: methods to determine primary products of lipid oxidation and methods to assess secondary products. In order to accurately assess sample’s oxidative status different methods should be applied, and primary and secondary products should be quantified as well. Primary products provide information regarding off-flavor precursors, while secondary products show a direct association with compounds related to sample’s sensory characteristics and acceptability. Therefore, method election must be based in its capacity to predict product’s shelf life and its acceptance by consumer (Costa, 2014; Shahidi & Wanasundara, 2008). The most used methods to evaluate oxidative rancidity in oils are as follows:

4.2.4.1 Sensory analysis

Sensory methods based in the evaluation of the oil, or oil containing food organoleptic characteristics provide the best correlation with consumer acceptability. Methods’ sensibility is dependent of tasters training degree. Therefore the high influence of tasters in the results and the high costs associated with training are major limitations of these types of methods (Frankel, 2005).

4.2.4.2 Primary products

4.2.4.2.1 Oxygen

Methods involving oxygen absorption are based in the fact that lipid oxidation can be translated in a measurable oxygen consumption (Jadhav et al., 1996). The kinetic study of oxygen consumption, inherent to lipid oxidation, enables the determination of the initiation phase or induction period, due to an increase of oxygen consumption for hydroperoxides formation in the early stages of autoxidation. There are described manometric, weight gain and chromatographic techniques to assess oxygen consumption, however the weight gain method is the most used because of instrumentation low cost and simple sample preparation. 48

Nevertheless this technique shows certain disadvantages, such as: 1) the weighing frequency hinders monitoring of fast kinetics; 2) long analysis times for tests in low or moderate temperatures; 3) non-reproducible results due to samples discontinuous heating; 4) intense human participation; and 5) results are highly affected by sample size, shape of container and temperature (Shahidi & Wanasundara, 2008; L. Silva, Pinto, Carrola, & Paiva-Martins, 2010).

4.2.4.2.2 Fatty acids quantification

The association between fatty acids composition and the extent of oxidation is only applicable in polyunsaturated fatty acids, especially marine oils and conjugated fatty acids, because in these oils, the fatty acids profile is influenced by the oxidation process right in the beginning. Fatty acid content is determined by gas chromatography and in order to obtain reliable results, sample’s oil extraction must be complete (Costa et al., 2015; Shahidi & Wanasundara, 2008; Silva, Borges, & Ferreira, 1999).

4.2.4.2.3 Peroxide value (PV)

The peroxide value is a classical method for quantification of hydroperoxides. The official AOCS method quantifies samples’ PV based on the reduction of the hydroperoxides - group (ROOH) with iodide ion (I ). The content of liberated iodine (I2) is proportional to the samples’ peroxide content. A standardized solution of sodium thiosulfate (Na2S2O3) and a starch indicator is used for titration in order to determine the released I2. Chemical reactions involved in PV assessment are given below (Equation 1 and 2) (AOCS, 2012):

+ + ROOH + 2H + 2KI → I2 + ROH + H2O + 2K [Equation 1]

I2 + 2Na2S2O3 → Na2S4O6 + 2NaI [Equation 2]

Nevertheless this method shows potential drawbacks such as: 1) absorbtion of iodine at unsaturation sites of fatty acids; 2) oxygen present in the solution might liberate iodine; 3) results are highly influenced by peroxides structure and reactivity, reaction time and temperature, as well as variations in the procedure; and 4) requires a large amount of lipids. In order to improve the aforementioned drawbacks, different methods were developed. In this context a fast and sensitive spectrophotometric method proposed by Shantha & Decker (1994) which also requires low amounts of lipids have been highly utilized to determine microencapsulated oils oxidative stability. This method is based on the spectrophotometric 49

quantification of ferric thiocyanate (a pink chromophore) which is produced by oxidation of ferrous to ferric ions caused by samples’ peroxides (Frankel, 2005).

At early stages of lipid oxidation, the rate of hydroperoxides formation is higher than your decomposition, however this patterns is reversed in more advanced stages. In this sense, PV must be monitored regularly in order to define oil’s oxidative stage (Frankel, 2005).

4.2.4.2.4 Conjugated dienes

The oxidation of polyunsaturated fatty acids promotes the formation of hydroperoxides and rearrange of double bonds with consequent formation of conjugated dienes and trienes. The latter are quantified spectrophotometrically as they exhibit an intense absorption at 234 nm (dienes) and 268 nm (trienes). However this method might not be applied to PSO as it naturally posses high contents of cLnAs isomers (Shahidi & Wanasundara, 2008).

4.2.4.3 Secondary products

Primary oxidation products are transient and do not provoke a major impact on oil’s quality, as they are colorless, odorless and tasteless. On the other hand, quantification of secondary products provide more reliable results regarding sample’s oxidative status, because these compounds are more stable and with pronounced odor.

4.2.4.3.1 2-Thiobarbituric acid value (TBA)

The TBA is also a classical and widely used method to determine secondary products of lipid peroxidation. This method is based on the colorimetric reaction between 2- Thiobarbituric acid and oxidation products of polyunsaturated fatty acids, especially malonaldehyde. TBA is a fast method, however its lack of specificity, as proteins, products of non-enzymatic browning and sugars are able to produce colorimetric complexes with TBA (AOCS, 2012).

4.2.4.3.2 p-Anisidine value (p-AnV)

This spectrophotometric method assesses the amount of aldehydes (principally 2- alkenal and 2,4-alkadienals) in oils by a reaction with p-anisidine reagent under acidic condition, forming yellowish products. p-AnV is defined as 100 times the optical density measured at 350 nm in a 1.0 cm cell of a solution containing 1.0 g of oil in 100 mL of mixture 50

of solvent and reagent (Shahidi & Wanasundara, 2008). According to Silva et al. (1999) oils with good quality should exhibit p-AnV under 10.

4.2.4.3.3 TOTOX value

TOTOX value is the combination of PV and p-AnV (Equation 3).

[Equation 3] TOTOX = 2PV + p-AnV

Where: PV: Peroxide value; p-AnV : p-Anisidine value.

TOTOX combine information regarding the history (p-AnV) of the oil and its present (PV). Nevertheless results should be interpreted with caution, as there is no scientific basis in the association of 2 variables with different dimensions (Shahidi & Wanasundara, 2008).

4.2.4.3.4 Volatile compounds

Volatile compounds determination by GC is a method capable to identify/quantify secondary lipid oxidation products such as: alcohols, aldehydes, ketones, carboxylic acids, esters and hydrocarbons. This method monitors compounds associated with flavor development in oxidized lipids. Besides a good correlation with flavor scores in sensory analyses, GC analyses also provide a sensitive method to detect low levels of oxidation. This characteristic is very important when dealing with highly unsaturated fatty acids which volatile products shows undesired flavor attributes in very low concentrations (0.04-2.5 ppm) (Frankel, 2005). Established oxidation markers of recognized fatty acids were described in Table 4. When compared with sensory analysis this technique shows the following advantages: 1) low cost, as there are no tasters involved, and 2) higher sensitivity and low variability, because these parameters are not influenced by tasters’ subjectivity.

Volatile fraction analysis is suitable for a wide range of oil containing food as it is “solventless” method and exhibits a simple sample preparation procedure. Among the different extraction techniques utilized for volatile compounds extraction, the solid-phase microextraction (SPME) is the most utilized due to low cost, convenience and reproducibility (Costa et al., 2015; García-Martínez et al., 2009; Vichi, Pizzale, Conte, Buxaderas, & López- Tamames, 2003). In the SPME, samples’ volatiles are extracted and concentrated in a fiber 51

composed by a selective stationary phase, then the entrapped volatile compounds are inject in a gas chromatograph coupled to a mass detector for identification, or a flame ionization detector (FID) for quantification (Kataoka, Lord, & Pawliszyn, 2000).

Table 4. Oxidation markers of fatty acid methyl esters

Oleic acid Linoleic acid Linolenic acid

Aldehyde

Octanal Pentanal Propanal Nonanal Hexanal Butanal 2-Decenal 2-Octenal 2-Butenal Decanal 2-Nonenal 2-Pentenal 2,4-Decadienal 2-Hexenal 3,6-Nonadienal Decatrienal

Carboxylic acid

Methyl heptanoate Methyl heptanoate Methyl heptanoate Methyl octanoato Methyl octanoate Methyl octanoate Methyl 8-oxooctanoate Methyl 8-oxooctanoate Methyl nonanoate Methyl 9-oxononanoate Methyl 9-oxononanoate Methyl 9-oxononanoate Methyl 10-oxodecnoate Methyl 10-oxodecanoate Methyl 10-oxodecanoate Methyl 10-oxo-8-decenoate Methyl 11-oxo-9-undecenoate

Alcohol

1-Heptanol 1-Pentanol 1-Octene-3-ol

Hydrocarbon Heptane Pentane Ethane Octane Pentane

(Adapted from: Choe & Min, 2006)

4.2.4.3.5 Volatile acids

The evaluation of volatile acids content can be performed by conductimetry or by the Rancimat® method. This method allows the determination of oils’ oxidative stability automatically under standard conditions. It is widely used in the oils and fats industry and is one of the most important parameters for quality assessment. The analysis consists of the indirect quantification of low molecular weight volatile acids formed during heating of oils at high temperatures and under constant aeration. These compounds get trapped in the water, altering and increasing its conductivity. In this way it is possible to obtain the time where there is a sharp increase in samples’ oxidative rate (induction time) (Silva et al., 1999).

4.2.4.3.6 Triacylglycerol dimers and higher oligomers

As described before conjugated fatty acids exhibits a different pattern of oxidation, as polymer formation is verified in the beginning of the oxidation process (Márquez-Ruiz et al., 2014; Miller & Claxton, 1928). Polymer formation can be investigated by quantification of triacylglycerols dimers and higher oligomers utilizing high performance size exclusion chromatography coupled with a refractive index detector (Márquez-Ruiz et al., 2014).

4.3. Encapsulation

4.3.1 Concept

Encapsulation methods have attracted great attention in the last decades, due to an increasing use of bioactive and unstable compounds in the food industry. Microencapsulation is a “packing” technique in which an active ingredient (core), is covered by a wall material, being often used to protect unstable molecules from interaction with other components and the adjacent environment during food processing and storage, or aiming controlled release (Gouin, 2004). The encapsulated substance can be designed as core or active ingredient, while the covering material is commonly described as wall material, wall system, encapsulant matrix and carrier. The core and the wall material can be composed by more than one ingredient, and for some encapsulation techniques, the encapsulant matrix might even been formed by 2 layers (Gharsallaoui, Roudaut, Chambin, Voilley, & Saurel, 2007; Gouin, 2004). Particles obtained through encapsulation may be classified according in it sizes: macro (> 5000 µm); micro (1.0-5000 µm) and nano (<1.0 µm) (King, 1995). 53

Regarding particles physical structure, there are 2 major classifications: single-core (capsule) or multiple-core (particle) microparticles (Figure 10). Capsules are typically produced by complex coacervation, fluidized bed drying, droplet co-extrusion and molecular inclusion, in this configuration the core is covered by the wall system, forming a reservoir system. The latter is characteristic of spray drying process, and the core material is dispersed throughout the wall system, showing a void in the central area (Jafari, Assadpoor, He, & Bhandari, 2008).

Figure 10. Particles physical structure. (Adapted from: Jafari et al., 2008).

The major advantages of the encapsulation are: 1) protect from adverse ambient conditions (light, moisture, oxygen, UV radiation); 2) protect from interaction with other food components during production and storage; 3) provide core controlled release (active ingredient delivery in the right place at the right time); 4) mask or preserve flavors; and 5) increase substances versatility by transforming them in a “easy handle” free-flowing power suitable for application in a wide range food products (Jafari, Assadpoor, He, et al., 2008).

The encapsulation methods are classified in: 1) physical (spray drying, spray cooling and chilling, fluidized bed coating, extrusion, freeze drying and co-crystallization); chemical (molecular inclusion and interfacial polymerization); and 3) physicochemical (coacervation, organic phase separation and liposome entrapment). The election of the encapsulation methods relies in economic factors, core sensibility and the desired microparticles application (Jafari, Assadpoor, He, et al., 2008).

Encapsulation techniques such as spray drying and freeze drying have already been applied to PSO aiming at increasing its stability (Goula & Adamopoulos, 2012; Gupta et al., 2012; Sahin-Nadeem & Afşin Özen, 2014). Goula & Adamopoulos (2012) and Sahin- Nadeem & Afşin Özen (2014) microencapsulated PSO by spray drying in wall systems 54

composed by skimmed milk powder and blends of starch derivates/ whey protein, respectively, and produced microparticles with good oxidative stability and physical-chemical properties (high values of oil retention, microencapsulation efficiency and solubility). Furthermore, Sahin-Nadeem & Afşin Özen (2014) related cLnA isomerization, promoted by the drying process. While Gupta et al. (2012) produced microparticles with controlled release, they encapsuled PSO by freeze drying in a wall material composed by sodium alginate/trehalose and calcium caseinate.

Nevertheless complex coacervation, a very traditional encapsulation method used for lipids encapsulation, has not been tested in PSO yet. Comparing to freeze drying and spray drying, complex coacervation is able to deliver microparticles with good technological properties, especially high microencapsulation efficiency, utilizing low contents of wall system. The complex coacervation process usually uses a total concentration of wall system until 7% (Eratte, Wang, Dowling, Barrow, & Adhikari, 2014; Santos, Bozza, Thomazini, & Favaro-Trindade, 2015; Weinbreckt et al., 2004), whereas microparticles produced by freeze drying and spray drying are formulated with at least 10% of wall material (Carneiro, Tonon, Grosso, & Hubinger, 2013; Costa et al., 2015; Goula & Adamopoulos, 2012; Sahin-Nadeem & Afşin Özen, 2014).

4.3.2 Complex coacervation

4.3.2.1 Concept

By definition coacervation is the separation into two liquid phases in colloidal systems, forming a phase rich in soluble or insoluble colloid (complexes), and another composed chiefly by solvent. Complex coacervation occurs when two oppositely charged colloids interact (Burgess, 1994). The two oppositely charged polyelectrolytes interact by electrostatic interactions forming complexes, that, depending of the strength of the electrostatic interactions, might be soluble or solid-like precipitates (Figure 11). During this process, the encapsulation takes place, as the coacervate phase adsorbed onto oil droplets, producing the microparticles (Eratte et al., 2014). 55

Figure 11. Schematic diagram of coacervates formation by protein-polysaccharide interactions. (Adapted from: Wu, Degner, & McClements, 2014)

The study of attractive interactions between proteins and polysaccharides has been investigated since 1991. The most used proteins for complex coacervation processes are gelatin and whey protein, however recent studies have focused in plant derived proteins, such as: pea, soy and flaxseed, due to consumers concerns (Eratte et al., 2014; Gharsallaoui, Yamauchi, Chambin, Cases, & Saurel, 2010; Ifeduba & Akoh, 2016; Kaushik, Dowling, McKnight, Barrow, & Adhikari, 2016; Liu, Low, & Nickerson, 2010a). Regarding the polysaccharides, they can be divided in 2 groups: 1) anionic (gum Arabic, alginate, pectin and carrageenan); and 2) cationic (chitosan) (Kruif, Weinbreck, & De Vries, 2004; Huang, Sun, Xiao, & Yang, 2012).

Complexes formation is influenced by several parameters such as: physical-chemical characteristic of involved polymers (charge density, molar mass), total polymer concentration and ratio, solution conditions (pH, ionic strength, type of ions), temperature and shear. Solution pH is one of the most important parameters to induce polymers interactions, as in pH under isoelectric point (pI), proteins exhibit a positive charge, while anionic polysaccharides usually are negatively charged in low pH (Weinbreckt et al., 2004).

The formation of protein-polysaccharide complexes is chiefly pH dependent and follows a two-step nucleation and growth-type kinetics process with the formation of soluble and insoluble complexes. Initially, protein-polysaccharide soluble complexes are formed near 56

the protein’s isoelectric point, the pH where this phenomenon occurs is denoted pHc and is marked by a slight increase of solution’s turbidity. pH reduction is followed by macroscopic phase separation with formation of insoluble protein-polysaccharide complexes at pHϕ1. The critical pH (pHc) indicates initial non-covalent attraction between polymers, while pHϕ1 demonstrate the pH where polymer net charge is opposite. Conditions for complexation are optimal at pHopt, which is the pH were both polymers reach their electrical equivalence.

Nevertheless, complex coacervation ceases at pH lower than pHϕ2, which is the pH where polymers shows similar net charges, as polysaccharides become protonated (Turgeon, Schmitt, & Sanchez, 2007; Weinbreck, Tromp, & de Kruif, 2004).

The production of microencapsulated oils by complex coacervation involves the following steps (Figure 12) (Xiao, Liu, Zhu, Zhou, & Niu, 2014a):

1) Dissolution- Both polymers are dissolved separately in the solvent (water). Protein dissolution usually occurs at a temperature above the gelling point, and at a pH above the pI;

2) Emulsification- The hydrophobic material (core) is dispersed in the polymer with emulsification properties (protein). When a stable emulsion is formed, the polysaccharide is added;

3) Coacervation/encapsulation- induction of protein-polysaccharide electrostatic interaction by lowering the solution pH below proteins’ pI, forming an insoluble polymer rich phase and another phase composed chiefly by solvent. During this process, the encapsulation takes place, as the coacervate phase adsorbed onto oil droplets, producing the microparticles. For some systems such as gelatin-gum Arabic, a gelation step, deflagrated by a controlled cooling bellow protein’s gelling temperature, is necessary for coacervate layer deposition on oil droplet. Proteins with lower gelling point (< 30 °C), such as fish gelatin and whey protein, dissolution, emulsification and coacervation/encapsulation steps can be carried out in room temperature, not requiring a colling step for wall deposition.

4) Hardening and recuperation- complex coacervate microparticles structure is pH dependent, therefore in order to harden/stabilize microparticles’ structure a cross-linking agent is used. Complex coacervation process occur in solution, in this sense, after production microparticles must be recovered from reaction media, by drying methods such as: oven, spray drying and freeze drying. 57

Recuperation

Microparticles formation

Figure 12. Production of microencapsulated oils by complex coacervation utilizing the gelatin-gum Arabic as example. (Adapted from: Manjanna, Shivakumar, & Kumar, 2010)

4.3.2.2 Hardening methods

Microparticles hardening is the limiting phase of complex coacervation, the process itself is inexpensive and does not requires sophisticated instrumentation, however it last step (Hardening) hinders its application due to safety, longs times of processing and economical reasons. Glutaraldehyde is the most popular cross-linker agent, this compound promotes non- soluble networks, via the reaction of aldehyde with amino groups in protein. However glutaraldehyde is considered toxic and depending of country’s legislation this compound cannot be applied in food. Transglutaminase and tannic acid are safe options for food application. Transglutaminase are naturally occurring enzymes which catalyses the formation of isopeptide bonds between the amino acids lysine and glutamine presents in the protein. Nevertheless, Transglutaminase is more expensive than other hardening agents and process duration is longer. Tannic acid is a plant polyphenol capable to complex with protein through hydrogen bonds and hydrophobic interactions. In this context the spray drying process emerges as a dynamic, low cost and safe hardening/recuperation method (Gouin, 2004; Lamprecht, Schäfer, & Lehr, 2001; Xiao, Liu, Zhu, Zhou, & Niu, 2014b).

4.3.2.2.1 Spray drying

The spray drying or atomization is a continuous process (Figure 13), where a liquid is transformed in a free-flowing powder, almost instantaneously. The method consist essentially of fluid atomization in a drying chamber where a hot air flow is injected concurrently with the liquid, promoting a fast sample dehydration, which allows samples’ low temperature maintenance during the whole process. In this sense, the spray drying is suitable for unstable compounds, such as PSO (Ré, 1998; Reineccius, 2004).

Microparticles structure hardening occurs by water evaporation transforming the coacervate phase in a stable barrier, protecting the core from interaction with the surround ambient (Reineccius, 2004).

Figure 13. Spray drying system (Adapted from: Reineccius, 2004)

4.3.2.3 Whey protein-gum Arabic system

Different protein-polysaccharide systems have been utilized for nutrients microencapsulation, such as: gelatin-gum Arabic, gelatin-carrageenan, gelatin-carboxymethyl cellulose, soybean protein-gum Arabic, soybean protein-pectin, pea protein-pectin, flaxseed protein-flaxseed gum, among other. Each system has its optimum conditions for complex coacervation (total polymer concentration and ratio, solution conditions (pH, ionic strength, 59

type of ions), temperature and shear) that affect greatly microparticles physical-chemical properties. (Gharsallaoui et al., 2010; Kaushik et al., 2016; Shuanghui Liu, Low, & Nickerson, 2009; Xiao et al., 2014a).

The is one of the most studied protein-polysaccharide system due to whey protein high emulsification power and low gelling point enabling microparticles production at room temperature. These advantages give dynamism to the process, by removing cooling steps from the production flowchart (Eratte et al., 2014; Weinbreck, de Vries, Schrooyen, & de Kruif, 2003; Weinbreck et al., 2004).

Factors affecting protein-polysaccharide interaction in this system have already been extensively evaluated (Eratte et al., 2014; Weinbreck, de Vries, Schrooyen, & Kruif, 2003; Weinbreck et al., 2004; Weinbreck, Rollema, Tromp, & De Kruif, 2004). Total polymer concentration influences largely the complex coacervation, as the process releases in solution counterions and H2O molecules. A high concentration of counterions is capable of suppress the coacervation process by screening the charges of polymers, and an increase of polymers concentration, raises proportionately counterions release. According to Weinbreck et al. (2003) complex coacervation is favored in total polymer concentration until 20%, but at high values (above 12.5) there is a decrease in the process yield. In this study, the same trend was observed with solution ionic strength, solutions with high ionic strength, analogous to total polymer concentration, are able to screen polymers charges. In this sense, polymers are usually dissolved in distilled or deionized water without adding salts (Eratte et al., 2014; Weinbreckt et al., 2004).

The ratio protein-polysaccharide (2:1) showed the best result with the fast phase separation as reported previously by Weinbreck et al. (2004), while Eratte et al. (2014) related

(3:1) ratio as the best. Each ratio has it pHopt, ratio (2:1) = pHopt 4.0, ratio (3:1) = pHopt 3.75, due to distinct molecules interaction sites.

Solution pH has a major impact in the process, previous studies indicate that coacervation occur in a pH range between 3-4.5, which is under whey proein pI, and the pHopt is highly influenced by the aforementioned factor (total polymer concentration, ionic strength, ratio protein-polysaccharide, among other) (Eratte et al., 2014; Weinbreck et al., 2003; Weinbrreck, Tromp, & de Kruif, 2004). 60

In short, the coacervation process is influenced by a wide range of factor and although some are already established, other might be tentatively evaluated. In this sense, always when a new core is tested, preliminary analysis should be performed in order to optimize the process.

4.3.2.3.1 Wall material

4.3.2.3.1.1 Whey protein

Whey protein is a non-toxic wall material highly utilized due to the following functional properties: water solubility, emulsifying and foaming capacity. Whey proteins are a by-product of cheese manufacturing and represents 20% of proteins in milk (Wijayanti, Bansal, & Deeth, 2014).

Whey protein is composed by β-lactoglobulin (β-Lg: 3.2g/L), α-lactalbumina (α-La: 1.2g/L), bovine serum albumin (BSA: 0.4 g/L) and immunoglobulins (0.7 g/L). β-Lg has a pI _ + at pH 5.2 and α-La at pH 4.1, thus under this pH the protein carries a positive charge ( NH3 ). Whey protein interactions with gum Arabic is dominated by β-Lg, because of the higher concentration (Raikos, 2010; Xiao et al., 2014a).

4.3.2.3.1.2 Gum Arabic

Gum Arabic is the exudation from certain acacia trees originated from semi-arid zones in the sub-Saharan Africa. Gum Arabic is an odorless, insipid, safe and highly soluble material widely used for micropencapsulation processes that exhibits emulsifying properties and low viscosity in high concentrations (Dror, Cohen, & Yerushalmi-Rozen, 2006).

Gum arabic is composed by 3 fractions: the major one consists of a highly branched polysaccharide of β-(1→3) galactose backbone with linked branches of arabinose and rhamnose, which finished in glucuronic acid (found as magnesium, potassium and calcium salts). The second one is composed by a high molecular weight arabinogalactan-protein complex (GAGP-GA glycoprotein), where arabinogalactan are covalently linked by hydroxyproline and serine residues through the protein chain. The last and minor fraction is another glycoprotein with different amino acids composition than GAGP-GA glycoprotein. Gum Arabic is a weak polyelectrolyte which is negatively charged (_COO-) above pH 2.2, because carboxyl groups dissociation is suppressed in low pH (< 2.2) (Burgess & Carless, 1984; Dror et al., 2006). 61

4.3.2.3.2 Encapsulation of oils with whey protein-gum Arabic system

Weinbreckt et al. (2004) successfully encapsulated different flavors (lemon and orange) by complex coacervation utilizing the whey protein-gum Arabic system. Microparticles produced showed a high payload (> 80%) and were fit for food application in Gouda cheese.

In line with these results, Eratte et al. (2014) also encapsulated tuna oil by complex coacervation utilizing whey protein-gum Arabic and compared the performance of spray drying and freezing drying as hardening/recuperation steps. Spray drying microparticles had a better oxidative stability, due to lower non-encapsulated oil content and a lack of open pores on particles’ surface. Recently, Eratte et al. (2015) utilized the same system to co-encapsulate tuna oil and probiotic bacteria, the co-encapsulation significantly increased probiotic viability and again microparticles were oxidative stable.

4.4. Microparticles characterization

The investigation of food properties is very important to optimize processes, functionality and cost reduction. Regarding microencapsulates substances, its physical, chemical, morphological and technological properties have attracted much attention from the food and pharmaceutical industry. Microparticles’ physical, chemical, morphological and technological properties are important indicators of oxidative stability, thermal stability, microbiological safety and suitability for food application. Particles density, porosity, stickiness, flow, particle size and distribution are considered physical properties. On the other hand, chemical properties are related to powder composition and interactions with other food components. There is not a clear definition for technological properties, but some authors use this term as a synonymous for physical and chemical characteristics that indicate food application such as: solubility, wettability, thermal stability, hygroscopicity among others. Finally morphological properties are related to particles’ shape and surface’s topography and integrity (Barbosa-Cánovas & Juliano, 2005).

4.4.1 Microencapsulation efficiency

Microencapsulation efficiency (ME) is one of the most important properties in encapsulation process involving lipids. ME determination is more critical than core retention assessment, as the first is more specific, indicating the oil internalized by the wall system, while the last refers to microparticles total oil content. The non-encapsulated oil or surface oil is free for ambient oxidants interactions (O2, light, food components), as there is no protecting barrier. As result, microparticles ME has a great influence on microparticles oxidative stability and previous studies concerning oils encapsulation consistently demonstrated an association between low ME and high oxidative ratios (Eratte et al., 2014; Kaushik et al., 2016; S. Liu, Low, & Nickerson, 2010b; Tonon, Grosso, & Hubinger, 2011). Additionally, microparticles solubility, dispersability and wettability are also affected negatively by high contents of surface oil (Vega & Roos, 2006) .

Factors such as oil:wall system ratio, total polymer concentration and hardening/recuperation method influence microparticles ME in complex coacervation processes (Eratte et al., 2014; Lamprecht et al., 2001; Weinbreckt & Minorf, 2004). Usually non-encapsulated oil is determined by solvent extraction, but more sophisticated techniques, such a X-ray photoelectron spectroscopy (XPS), have been used to analyze microparticles surface composition, as solvent extraction might remove to some extent encapsulated oil (Jafari, Assadpoor, Bhandari, & He, 2008).

XPS is able to determine particles surface (within the first 10 nm) elemental composition by measuring the kinetics energies of electrons ejected from the sample (Van der Heide, 2012). Jafari, Assadpoor, Bhandari, & He (2008) microencapsulated fish oil by spray drying in different wall systems composed of whey protein, modified starch and maltodextrin, they demonstrated that both methods, XPS and solvent extraction, were well correlated. Recently Eratte et al., (2015) successfully applied XPS in the evaluation of surface composition of tuna oil and probiotic bacteria microparticles produced by complex coacervation utilizing whey protein-gum Arabic system.

4.4.2 Particle size and distribution

Particle size and distribution affects directly microparticles functional properties, such as: solubility, dispersibility, wettability and core release (Pierucci, Andrade, Baptista, Volpato, & Rocha-Leão, 2006; Reineccius, 2004). Particles with a narrow size distribution are 63

important for controlling core sustained release and decomposition. The following parameters influence microparticles’ particle sizes: oil:wall system ratio, total polymer concentration and hardening/recuperation method (Eratte et al., 2014; Lamprecht et al., 2001; Weinbreckt & Minorf, 2004). Particle size and distribution is commonly assessed by the scattering pattern of a transverse laser light technique.

4.4.3 Surface morphology

The investigation of microparticles surface morphology provides important information regarding product’s functionality, core stability and flowability (Rosenberg, Kopelman, & Talmon, 1990; Sheu & Rosenberg, 1998). Microparticles integrity and porosity affects directly oxidative stability of lipids’ cores, as particles exhibiting open pores and cracks shows less protection against oxidants interactions, and thus demonstrate higher core degradation (Eratte et al., 2014; Jimenez, García, & Beristain, 2004). Furthermore, microparticles characteristics such as: open pores, cracks, invagination and roughness influence negatively product flowability (Rosenberg, Kopelman, & Talmon, 1990).

Microstructure of dried microparticles is affected by wall material properties, total polymer concentration, ratio oil:wall material, hardening/recuperation method and storage conditions (Eratte et al., 2014; Sheu & Rosenberg, 1998; Yang, Gao, Hu, Li, & Sun, 2015). Generally, this parameter is assessed by scanning electron microscopy (SEM), which enables the evaluation of microparticles’ internal and external morphology (Schugens et al., 1994).

4.4.4 Thermal characterization

Microparticles thermal characterization is assessed by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC).

TGA is an analytical technique used to determine microparticles’ thermal stability by monitoring samples’ weight changes during heating at constant rate. TGA provides information regarding the temperature (Tonset) at microparticles’ materials start to degrade.

Tonset is important for food applications, as it indicates the maximum temperature allowed during product processing, production and storage (Jain & Sharma, 2012).

DSC is a thermoanalytical technique in which the difference in the amount of heat required to increase the temperature of a sample and reference is measured as a function of temperature. By performing DSC analysis is possible to determine microparticles glass 64

transition temperature (Tg), a very important microparticles property for maintenance of core stability during storage. Tg is defined as the temperature at which an amorphous sytem (wall material) changes from a glassy to a rubbery state. In the glassy state wall system shows a high viscosity and molecular movements is highly limited, therefore diffusion of O2 or other ambient component thru the wall system is almost nonexistent (Bhandari & Howes, 1999; Kodre, Attarde, Yendhe, Patil, & Barge, 2014).

4.4.5 Moisture content, water activity and hygroscopicity

Moisture content, water activity and hygroscopicity are microparticles properties related with water content and water absorption, the maintenance of low values of these parameters during storage are crucial to core protection, microparticles’ technological properties and microbiological safety (Beristain, Azuara, & Vernon-Carter, 2002; Reineccius, 2004).

4.4.6 Solubility and wettability

Microparticles solubility and wettability are direct related to powder rehydration and instantization, therefore these parameters indicate particles suitability for addition in beverages. Solubility and wettability are influenced by powder particle size, bulk density, porosity, surface charge and presence of hydrophobic components on microparticles surface (Bae & Lee, 2008; De Barros Fernandes, Marques, Borges, & Botrel, 2014; Vega & Roos, 2006).

4.4.7 Density

Density is an important factor related to microparticles storage and commercialization logistics. In dried products, density is expressed as bulk density which is a relation (g/cm3) between sample mass and volume occupied in a graduated cylinder after tapping (Tonon, Pedro, Grosso, & Hubinger, 2012; Goula & Adamopoulos, 2004).

4.5 Experimental Design

A wide range of factors affect the complex coacervation process, influencing process yield and microparticles physical-chemical characteristics. As result, aiming at optimize the process and product quality, reduce costs and increase process dynamism, an increasing 65

number of researchers have utilized experimental design techniques to achieve these goals (Rodrigues & Lemma, 2014).

The experimental design methodology associated with responses surfaces evaluation, provides reliable information regarding factors that affects significantly the process and possible interactions between them, reducing empiricism and time consuming experiments related with techniques involving trial and error (Rodrigues & Lemma, 2014). According to Calado & Montgomery (2003) response surfaces are applied when response variables are influenced by a large number of independent factor, aiming at response variables optimization.

Among the distinct types of experimental design, the rotatable central composite design (RCCD) is one of the most used to optimize microparticles’ physical-chemical characteristics in encapsulation processes. The RCCD is able to verify plane curvature and due to the presence of second order terms, it provides the maximum, or the lowest response of a given dependent variable when is possible (Calado & Montgomery, 2003)

The application of experimental design, especially RCCD, to optimize encapsulation processes of different oils by spray drying has been extensively described. Usually, the influence o process parameters (inlet air temperature, homogenization) and feed emulsion composition (solids content, oil:wall system ratio, oil content) on process yield and microparticles physical-chemical properties, especially ME, were evaluated by RCCD (Ahn et al., 2008; De Barros Fernandes et al., 2014; Frascareli, Silva, Tonon, & Hubinger, 2011; Tonon, Grosso, et al., 2011). Ahn et al. (2008) utilized central composite design to optimize ME in microencapsulated sunflower oil. In this study, they were able to produce microparticles with high ME based in a equation proposed by the experimental design containing the following independent variables: oil concentration, proportion of milk protein isolates to wall system and emulsifier concentration. Additionally, De Barros Fernandes et al. (2014) produced microparticles of rosemary essential oil in a wall system composed of maltodextrin and modified starch in optimized conditions of wettability, hygroscopicity, total oil and retention oil, utilizing a feed emulsion composition established by the RCCD.

Regarding complex coacervation, Yang et al. (2015), encapsulated poppy-seed in gelatin-gum Arabic system and by experimental design evaluated the effect of total polymer concentration, oil:wall system ratio, pH and emulsifier concentration on ME. pH was the factor that most influenced ME, followed by concentration of emulsifier, then total polymer 66

concentration and finally oil:wall system ratio. Moreover, Koupantsis, Pavlidou, & Paraskevopoulou (2014) encapsulated β-pinene in milk proteins-CMC complexes and applied RCCD to investigate the influence of protein-polysaccharide ratio and core concentration on microparticles physical-chemical properties (ME and encapsulation yield). As result microparticles with high protein-polysaccharide ratio and high core content showed the best protection capacity. Nevertheless, is important to emphasize that the utilization of statistical techniques to optimize microencapsulation processes by complex coacervation still scarce.

5. References

Abbasi, H., Rezaei, K., Emamdjomeh, Z., & Ebrahimzadeh Mousavi, S. M. (2008). Effect of various extraction conditions on the phenolic contents of pomegranate seed oil. European Journal of Lipid Science and Technology, 110(5), 435–440. http://doi.org/10.1002/ejlt.200700199

Abbasi, H., Rezaei, K., & Rashidi, L. (2008). Extraction of essential oils from the seeds of pomegranate using organic solvents and supercritical CO2. JAOCS, Journal of the American Oil Chemists’ Society, 85(1), 83–89. http://doi.org/10.1007/s11746-007-1158- x

ABIC (Associação Brasileira da Indústria do Café) (2018). Indicadores da Indústria do café. Available in : http://abic.com.br/estatisticas/indicadores-da-industria/indicadores-da- industria-de-cafe-2017/. Access in 10 october 2018.

Ahn, J.-H., Kim, Y.-P., Lee, Y.-M., Seo, E.-M., Lee, K.-W., & Kim, H.-S. (2008). Optimization of microencapsulation of seed oil by response surface methodology. Food Chemistry, 107(March 2008), 98–105. http://doi.org/10.1016/j.foodchem.2007.07.067

Akil, E., Castelo-Branco, V. N., Costa, A. M. M., do Amaral Vendramini, A. L., Calado, V., & Torres, A. G. (2015). Oxidative Stability and Changes in Chemical Composition of Extra Virgin Olive Oils After Short-Term Deep-Frying of French Fries. Journal of the American Oil Chemists’ Society, 92(3), 409–421. http://doi.org/10.1007/s11746-015- 2599-2

Allen, R. R., Jackson, A., & Kummerow, F. A. (1949). Factors which affect the stability of highly unsaturated fatty acids. I. Differences in the oxidation of conjugated and 67

nonconjugated linoleic acid. Journal of the American Oil Chemists’ Society, 26(8), 395– 399. http://doi.org/10.1007/BF02749501

American Oil Chemists’ Society. (2012). Official methods and recommended practices of the AOCS (6th ed.). Chicago.

Amri, Z., Lazreg-Aref, H., Mekni, M., El-Gharbi, S., Dabbaghi, O., Mechri, B., & Hammami, M. (2017). Oil characterization and lipids class composition of pomegranate seeds. BioMed Research International, 2017. http://doi.org/10.1155/2017/2037341

Arao, K., Wang, Y.-M., Inoue, N., Hirata, J., Cha, J.-Y., Nagao, K., & Yanagita, T. (2004). Dietary effect of pomegranate seed oil rich in 9cis, 11trans, 13cis conjugated linolenic acid on lipid metabolism in obese, hyperlipidemic OLETF rats. Lipids in Health and Disease, 3, 24. http://doi.org/10.1186/1476-511X-3-24

Arruda, A. C., Paula, V., Minim, R., Aurélio, M., Ferreira, M., Minim, L. A., … Soares, C. F. (2009). Justificativas e motivações do consumo e não consumo de café. Ciência e Tecnologia de Alimentos, 29(4), 754–763. http://doi.org/10.1590/S0101- 20612009000400009

Bae, E. K., & Lee, S. J. (2008). Microencapsulation of avocado oil by spray drying using whey protein and maltodextrin. Journal of Microencapsulation, 25(8), 549–60. http://doi.org/10.1080/02652040802075682

Barbosa-Cánovas, G. V., & Juliano, P. (2005). Physical and chemical properties of food powders. In C. Onwulata (Ed.), Encapsulated and powdered foods (pp. 39–71). Boca Raton, FL: Taylor & Francis.

Bassaganya-Riera, J., DiGuardo, M., Climent, M., Vives, C., Carbo, A., Jouni, Z. E., … Hontecillas, R. (2011). Activation of PPARγ and δ by dietary punicic acid ameliorates intestinal inflammation in mice. British Journal of Nutrition, 106(6), 878–886. http://doi.org/10.1017/S0007114511001188

Beristain, C. I., Azuara, E., & Vernon-Carter, E. J. (2002). Effect of Water Activity on the Stability to Oxidation of Spray-Dried Encapsulated Orange Peel Oil Using Mesquite Gum ( Prosopis Juliflora ) as Wall Material. Journal of Food Science, 67(1), 206–211. http://doi.org/10.1111/j.1365-2621.2002.tb11385.x 68

Bhandari, B. ., & Howes, T. (1999). Implication of glass transition for the drying and stability of dried foods. Journal of Food Engineering, 40(1–2), 71–79. http://doi.org/10.1016/S0260-8774(99)00039-4

Boussetta, T., Raad, H., Lettéron, P., Gougerot-Pocidalo, M.-A., Marie, J.-C., Driss, F., & El- Benna, J. (2009). Punicic acid a conjugated linolenic acid inhibits TNFalpha-induced neutrophil hyperactivation and protects from experimental colon inflammation in rats. PloS One, 4(7), e6458. http://doi.org/10.1371/journal.pone.0006458

Bradford, P. G., & Awad, A. B. (2007). Phytosterols as anticancer compounds. Molecular Nutrition & Food Research, 51(2), 161–170. http://doi.org/10.1002/mnfr.200600164

Burgess, D. J. (1994). Complex Coacervation: Micro-Capsule Formation. In P. Dubin, J. Bock, R. M. Davies, D. N. Schulz, & C. Thies (Eds.), Macromolecular Complexes in Chemistry and Biology (1st ed., pp. 285–299). Berlin: Springer-Verlag.

Burgess, D. J., & Carless, J. E. (1984). Microelectrophoretic studies of gelatin and acacia for the prediction of complex coacervation. Journal of Colloid And Interface Science, 98(1), 1–8. http://doi.org/10.1016/0021-9797(84)90472-7

Calado, V., & Montgomery, D. (2003). Planejamento de experimentos usando o Statistica (1a). Rio de Janeiro: E-Papers Serviços Editoriais.

Caligiani, A., Bonzanini, F., Palla, G., Cirlini, M., & Bruni, R. (2010). Characterization of a Potential Nutraceutical Ingredient: Pomegranate (Punica granatum L.) Seed Oil Unsaponifiable Fraction. Plant Foods for Human Nutrition, 65(3), 277–283. http://doi.org/10.1007/s11130-010-0173-5

CAMBICI (CÂMARA BRASIL-ISRAEL DE COMÉRCIO E INDÚSTRIA). Anuário 2011: Agronegócio. 2011. Available in: http://www.cambici.org.br/download/anuario/2011/Agronegocio.pdf. Access in: 14 october 2018

Campbell, W., Drake, M. A., & Larick, D. K. (2003). The Impact of Fortification with Conjugated Linoleic Acid (CLA) on the Quality of Fluid Milk. Journal of Dairy Science, 86(1), 43–51. http://doi.org/10.3168/jds.S0022-0302(03)73582-6

Cao, Y., Gao, H. L., Chen, J. N., Chen, Z. Y., & Yang, L. (2006). Identification and 69

characterization of conjugated linolenic acid isomers by Ag+-HPLC and NMR. Journal of Agricultural and Food Chemistry, 54(24), 9004–9009. http://doi.org/10.1021/jf0616199

Carneiro, H. C. F., Tonon, R. V., Grosso, C. R. F., & Hubinger, M. D. (2013). Encapsulation efficiency and oxidative stability of flaxseed oil microencapsulated by spray drying using different combinations of wall materials. Journal of Food Engineering, 115(4), 443–451. http://doi.org/10.1016/j.jfoodeng.2012.03.033

Carpenter, A. P. (1979). Determination of tocopherols in vegetable oils. Journal of the American Oil Chemists Society, 56(7), 668–671. http://doi.org/10.1007/BF02660070

Chinetti, G., Lestavel, S., Bocher, V., Remaley, A. T., Neve, B., Torra, I. P., … Staels, B. (2001). PPAR-α and PPAR-γ activators induce cholesterol removal from human macrophage foam cells through stimulation of the ABCA1 pathway. Nature Medicine, 7(1), 53–58. http://doi.org/10.1038/83348

Choe, E., & Min, D. B. (2006). Comprehensive Reviews in Food Science and Food Safety Mechanisms and Factors for Edible Oil Oxidation. Comprehensive Reviews in Food Science and Food Safety, 5(4), 169–186. http://doi.org/10.1111/j.1541- 4337.2006.00009.x

Clément, L., Poirier, H., Niot, I., Bocher, V., Guerre-Millo, M., Krief, S., … Besnard, P. (2002). Dietary trans-10,cis-12 conjugated linoleic acid induces hyperinsulinemia and fatty liver in the mouse. Journal of Lipid Research, 43(9), 1400–1409. http://doi.org/10.1194/jlr.M20008-JLR200

Costa, A. M. M. (2014). Encapsulamento de ácido linoleico conjugado (CLA) por spray drying. Rio de Janeiro, 102p. Dissertação (Mestrado)- Instituto de Nutrição Josué de Castro, Universidade Federal do Rio de Janeiro.

Costa, A. M. M., Nunes, J. C., Lima, B. N. B., Pedrosa, C., Calado, V., Torres, A. G., & Pierucci, A. P. T. R. (2015). Effective stabilization of CLA by microencapsulation in pea protein. Food Chemistry, 168. http://doi.org/10.1016/j.foodchem.2014.07.016

Costa, A. M. M., Silva, L. O., & Torres, A. G. (2019). Chemical composition of commercial cold-pressed pomegranate (Punica granatum) seed oil from Turkey and Israel, and the use of bioactive compounds for samples’ origin preliminary discrimination. Journal of 70

Food Composition and Analysis, 75, 8–16. http://doi.org/10.1016/j.jfca.2018.09.004

Coursodon-Boyiddle, C. F., Snarrenberg, C. L., Adkins-Rieck, C. K., Bassaganya-Riera, J., Hontecillas, R., Lawrence, P., … Dvorak, B. (2012). Pomegranate seed oil reduces intestinal damage in a rat model of necrotizing enterocolitis. Am J Physiol Gastrointest Liver Physiol, 303, 744–751. http://doi.org/10.1152/ajpgi.00248.2012.-Pomegranate

De Barros Fernandes, R. V., Marques, G. R., Borges, S. V., & Botrel, D. A. (2014). Effect of solids content and oil load on the microencapsulation process of rosemary essential oil. Industrial Crops and Products, 58, 173–181. http://doi.org/10.1016/j.indcrop.2014.04.025

De Kruif, C. G., Weinbreck, F., & De Vries, R. (2004). Complex coacervation of proteins and anionic polysaccharides. Current Opinion in Colloid and Interface Science, 9(5), 340– 349. http://doi.org/10.1016/j.cocis.2004.09.006

De Melo, I. L. P., De Oliveira E Silva, A. M., De Carvalho, E. B. T., Yoshime, L. T., Sattler, J. A. G., & Mancini-Filho, J. (2016). Incorporation and effects of punicic acid on muscle and adipose tissues of rats. Lipids in Health and Disease, 15(1), 1–12. http://doi.org/10.1186/s12944-016-0214-7

Dhar, P., Chattopadhyay, K., Bhattacharyya, D., Roychoudhury, A., Biswas, A., & Ghosh, S. (2007). Antioxidative effect of conjugated linolenic acid in diabetic and non-diabetic blood: an in vitro study. Journal of Oleo Science, 56(1), 19–24. http://doi.org/JST.JSTAGE/jos/56.19 [pii]

Dror, Y., Cohen, Y., & Yerushalmi-Rozen, R. (2006). Structure of gum Arabic in aqueous solution. Journal of Polymer Science: Part B: Polymer Physics, 3265–3271. http://doi.org/10.1002/polb

Eikani, M. H., Golmohammad, F., & Homami, S. S. (2012). Extraction of pomegranate (Punica granatum L.) seed oil using superheated hexane. Food and Bioproducts Processing, 90(1), 32–36. http://doi.org/10.1016/j.fbp.2011.01.002

Elfalleh, W., Ying, M., Nasri, N., Sheng-Hua, H., Guasmi, F., & Ferchichi, A. (2011). Fatty acids from Tunisian and Chinese pomegranate (Punica granatum L.) seeds. International Journal of Food Sciences and Nutrition, 62(3), 200–206. http://doi.org/10.3109/09637486.2010.526932 71

Eratte, D., McKnight, S., Gengenbach, T. R., Dowling, K., Barrow, C. J., & Adhikari, B. P. (2015). Co-encapsulation and characterisation of omega-3 fatty acids and probiotic bacteria in whey protein isolate-gum Arabic complex coacervates. Journal of Functional Foods, 19, 882–892. http://doi.org/10.1016/j.jff.2015.01.037

Eratte, D., Wang, B., Dowling, K., Barrow, C. J., & Adhikari, B. P. (2014). Complex coacervation with whey protein isolate and gum arabic for the microencapsulation of omega-3 rich tuna oil. Food Funct., 5(11), 2743–2750. http://doi.org/10.1039/C4FO00296B

Fadavi, A., Barzegar, M., Azizi, M. H., & Bayat, M. (2005). Note. Physicochemical Composition of Ten Pomegranate Cultivars (Punica granatum L.) Grown in Iran. http://doi.org/10.1177/1082013205052765

Fernandes, L., Pereira, J. A., Lopéz-Cortés, I., Salazar, D. M., Ramalhosa, E., & Casal, S. (2015). Fatty acid, vitamin E and sterols composition of seed oils from nine different pomegranate (Punica granatum L.) cultivars grown in Spain. Journal of Food Composition and Analysis, 39, 13–22. http://doi.org/10.1016/j.jfca.2014.11.006

Fischer, U. A., Carle, R., & Kammerer, D. R. (2011). 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, 127(2), 807– 821. http://doi.org/10.1016/j.foodchem.2010.12.156

Frankel, E. N. (2005). Methods to determine extent of oxidation. In E. N. Frankel (Ed.), Lipid Oxidation (2nd ed., pp. 99–128). Cambridge, UK: Woodhead Publishing.

Frascareli, E. C., Silva, V. M., Tonon, R. V, & Hubinger, M. D. (2011). Food and Bioproducts Processing Effect of process conditions on the microencapsulation of coffee oil by spray drying. Food and Bioproducts Processing, 90(3), 413–424. http://doi.org/10.1016/j.fbp.2011.12.002

García-Martínez, M. C., Márquez-Ruiz, G., Fontecha, J., & Gordon, M. H. (2009). Volatile oxidation compounds in a conjugated linoleic acid-rich oil. Food Chemistry, 113(4), 926–931. http://doi.org/10.1016/j.foodchem.2008.08.020

Gasmi, J., & Sanderson, J. T. (2010). Growth inhibitory, antiandrogenic, and pro-apoptotic effects of punicic acid in LNCaP human prostate cancer cells. Journal of Agricultural 72

and Food Chemistry, 58(23), 12149–12156. http://doi.org/10.1021/jf103306k

Gharsallaoui, A., Roudaut, G., Chambin, O., Voilley, A., & Saurel, R. (2007). Applications of spray drying in microencapsulation of food ingredients: An overview. Food Research International, 40(9), 1107–1121. http://doi.org/10.1016/j.foodres.2007.07.004

Gharsallaoui, A., Yamauchi, K., Chambin, O., Cases, E., & Saurel, R. (2010). Effect of high methoxyl pectin on pea protein in aqueous solution and at oil/water interface. Carbohydrate Polymers, 80(3), 817–827. http://doi.org/10.1016/j.carbpol.2009.12.038

Gouin, S. (2004). Microencapsulation: Industrial appraisal of existing technologies and trends. In Trends in Food Science and Technology (Vol. 15, pp. 330–347). http://doi.org/10.1016/j.tifs.2003.10.005

Goula, A. M., & Adamopoulos, K. G. (2004). Spray drying of tomato pulp: Effect of feed concentration. Drying Technology, 22(10), 2309–2330. http://doi.org/10.1081/LDRT- 200040007

Goula, A. M., & Adamopoulos, K. G. (2012). A method for pomegranate seed application in food industries: Seed oil encapsulation. Food and Bioproducts Processing, 90(4), 639– 652. http://doi.org/10.1016/j.fbp.2012.06.001

Gupta, S. S., Ghosh, S., Maiti, P., & Ghosh, M. (2012). Microencapsulation of conjugated linolenic acid-rich pomegranate seed oil by an emulsion method. Food Science and Technology International, 18(6), 549–558. http://doi.org/10.1177/1082013211433078

He, L., Xu, H., Liu, X., He, W., Yuan, F., Hou, Z., & Gao, Y. (2011). Identification of phenolic compounds from pomegranate (Punica granatum L.) seed residues and investigation into their antioxidant capacities by HPLC-ABTS+ assay. Food Research International, 44(5), 1161–1167. http://doi.org/10.1016/j.foodres.2010.05.023

Holic, R., Xu, Y., Caldo, K. M. P., Singer, S. D., Field, C. J., Weselake, R. J., & Chen, G. (2018). Bioactivity and biotechnological production of punicic acid. Applied Microbiology and Biotechnology. http://doi.org/10.1007/s00253-018-8883-y

Hontecillas, R., Diguardo, M., Duran, E., Orpi, M., & Bassaganya-Riera, J. (2008). Catalpic acid decreases abdominal fat deposition, improves glucose homeostasis and upregulates PPAR α expression in adipose tissue. Clinical Nutrition, 27(5), 764–772. 73

http://doi.org/10.1016/j.clnu.2008.07.007

Hontecillas, R., O’Shea, M., Einerhand, A., Diguardo, M., & Bassaganya-Riera, J. (2009). Activation of ppar γ and α by punicic acid ameliorates glucose tolerance and suppresses obesity-related inflammation. Journal of the American College of Nutrition, 28(2), 184– 195. http://doi.org/10.1080/07315724.2009.10719770

Hora, J. J., Maydew, E. R., Lansky, E. P., & Dwivedi, C. (2003). Chemopreventive Effects of

Pomegranate Seed Oil on Skin Tumor Development in CD 1 Mice. Journal of Medicinal Food, 6(3), 157–161. http://doi.org/10.1089/10966200360716553

Huang, G. Q., Sun, Y. T., Xiao, J. X., & Yang, J. (2012). Complex coacervation of soybean protein isolate and chitosan. Food Chemistry, 135(2), 534–539. http://doi.org/10.1016/j.foodchem.2012.04.140

I Ré, M. (1998). MICROENCAPSULATION BY SPRAY DRYING. Drying Technology, 16(6), 1195–1236. http://doi.org/10.1080/07373939808917460

Ifeduba, E. A., & Akoh, C. C. (2016). Microencapsulation of stearidonic acid soybean oil in Maillard reaction-modified complex coacervates. Food Chemistry, 199, 524–532. http://doi.org/10.1016/j.foodchem.2015.12.011

IPGRI (International Plant Genetic Resources Institute, Rome, Italy). Regional report CWANA 19992000. 2001. Disponível em: http://www.ipgri.cgiar.org/publications/pdf/821.pdf. Access in: 20 october 2018

Jadhav, S. J., Nimbalkar, S. S., Kulkarni, A. D., Madhavi, D. L., Rajalakshmi, D., & Narasimhan, S. (1996). Lipid oxidation in biological and food systems. In D. L. Madhavi, S. S. Deshpande, & D. K. Salunkhe (Eds.), Food antioxidants: Technological, Toxicological, and Health Perpectives (pp. 5–63). New York: Marcel Dekker Inc.

Jafari, S. M., Assadpoor, E., Bhandari, B., & He, Y. (2008). Nano-particle encapsulation of fish oil by spray drying. Food Research International, 41(2), 172–183. http://doi.org/10.1016/j.foodres.2007.11.002

Jafari, S. M., Assadpoor, E., He, Y., & Bhandari, B. (2008). Encapsulation efficiency of food flavours and oils during spray drying. Drying Technology. http://doi.org/10.1080/07373930802135972 74

Jain, S., & Sharma, M. P. (2012). Application of thermogravimetric analysis for thermal stability of Jatropha curcas biodiesel. Fuel, 93, 252–257. http://doi.org/10.1016/J.FUEL.2011.09.002

Jimenez, M., García, H. S., & Beristain, C. I. (2004). Spray drying microencapsulation and oxidative stability of conjugated linoleic acid. European Food Research and Technology, 219(6), 588–592. http://doi.org/10.1007/s00217-004-0992-4

Joh, Y. G., Kim, S. J., & Christie, W. W. (1995). The structure of the triacylglycerols, containing punicic acid, in the seed oil of Trichosanthes kirilowii. JAOCS, Journal of the American Oil Chemists’ Society, 72(9), 1037–1042. http://doi.org/10.1007/BF02660718

Jones, E. L., Shingfield, K. J., Kohen, C., Jones, A. K., Lupoli, B., Grandison, A. S., … Yaqoob, P. (2005). Chemical, Physical, and Sensory Properties of Dairy Products Enriched with Conjugated Linoleic Acid. Journal of Dairy Science, 88(8), 2923–2937. http://doi.org/10.3168/jds.S0022-0302(05)72973-8

Karaman, S., Karasu, S., Tornuk, F., Toker, O. S., Geçgel, Ü., Sagdic, O., … Gül, O. (2015). Recovery potential of cold press byproducts obtained from the edible oil industry: Physicochemical, bioactive, and antimicrobial properties. Journal of Agricultural and Food Chemistry, 63(8), 2305–2313. http://doi.org/10.1021/jf504390t

Kataoka, H., Lord, H. L., & Pawliszyn, J. (2000). Applications of solid-phase microextraction in food analysis. Journal of Chromatography A. http://doi.org/10.1016/S0021- 9673(00)00309-5

Kaufman, M., & Wiesman, Z. (2007). Pomegranate oil analysis with emphasis on MALDI- TOF/MS triacylglycerol fingerprinting. Journal of Agricultural and Food Chemistry, 55(25), 10405–10413. http://doi.org/10.1021/jf072741q

Kaushik, P., Dowling, K., McKnight, S., Barrow, C. J., & Adhikari, B. (2016). Microencapsulation of flaxseed oil in flaxseed protein and flaxseed gum complex coacervates. Food Research International, 86, 1–8. http://doi.org/10.1016/j.foodres.2016.05.015

Khanal, R. (2004). Potential health benefits of conjugated linoleic acid (CLA): A review. Asian Australasian Journal of Animal …, 17(9), 1315–1328. http://doi.org/dx.doi.org/10.5713/ajas.2004.1315 75

Khoddami, A., Man, Y. B. C., & Roberts, T. H. (2014). Physico-chemical properties and fatty acid profile of seed oils from pomegranate (Punica granatum L.) extracted by cold pressing. European Journal of Lipid Science and Technology, 116(5), 553–562. http://doi.org/10.1002/ejlt.201300416

Kim, H. J., & Min, D. B. (2008). Chemistry of Lipid Oxidation. In C. C. Akoh & D. B. Min (Eds.), Food Lipids: Chemistry, Nutrition and Biotechnology (3rd ed., pp. 299–318). Boca Raton: CRC Press.

Kim, N., Mehta, R., Yu, W., Neeman, I., Livney, T., Amichay, A., … Lansky, E. (2002). Chemopreventive and adjuvant therapeutic potential of pomegranate (Punica granatum) for human breast cancer. Breast Cancer Research and Treatment, 71, 203–217. Retrieved from https://s3.amazonaws.com/academia.edu.documents/40038713/0046351f070910950b000 000.pdf20151115-68247- bt2ncz.pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1518804993 &Signature=S1gGu2un8avv%2Fb1VbFBmbcKA4Vc%3D&response-content- disposition=inline%3B filenam

King, A. H. (1995). Encapsulation of food ingredients - a review of available technology, focusing on hydrocolloids. Encapsulation and Controlled Release of Foof Ingredients, 590, 26–39.

Kodre, K. V., Attarde, S. R., Yendhe, P. R., Patil, R. Y., & Barge, V. U. (2014). Differential Scanning Calorimetry: A Review. Research & Reviews: Journal of Pharmaceutical Analysis, 3(3), 11–22. http://doi.org/10.1080/09546553.2011.638347

Kohno, H., Suzuki, R., Yasui, Y., Hosokawa, M., Miyashita, K., & Tanaka, T. (2004). Pomegranate seed oil rich in conjugated linolenic acid suppresses chemically induced colon carcinogenesis in rats. Cancer Science, 95(6), 481–486. http://doi.org/10.1111/j.1349-7006.2004.tb03236.x

Koupantsis, T., Pavlidou, E., & Paraskevopoulou, A. (2014). Flavour encapsulation in milk proteins – CMC coacervate-type complexes. Food Hydrocolloids, 37, 134–142. http://doi.org/10.1016/j.foodhyd.2013.10.031

Kýralan, M., Gölükcü, M., & Tokgöz, H. (2009). Oil and conjugated linolenic acid contents 76

of seeds from important pomegranate cultivars (Punica granatum L.) Grown in Turkey. JAOCS, Journal of the American Oil Chemists’ Society, 86(10), 985–990. http://doi.org/10.1007/s11746-009-1436-x

Lamprecht, a, Schäfer, U., & Lehr, C. M. (2001). Influences of process parameters on preparation of microparticle used as a carrier system for omega - 3 unsaturated fatty acid ethyl esters used in supplementary nutrition. Journal of Microencapsulation, 18(3), 347– 357. http://doi.org/10.1080/02652040010000433

Lansky, E., Harrison, G., Froom, P., & Jiang, W. (2005). Pomegranate (Punica granatum) pure chemicals show possible synergistic inhibition of human PC-3 prostate cancer cell invasion across Matrigel. Investigational New Drugs, 23(2), 121–122.

Lansky, E. P., & Newman, R. A. (2007). Punica granatum (pomegranate) and its potential for prevention and treatment of inflammation and cancer. Journal of Ethnopharmacology, 109(2), 177–206. http://doi.org/10.1016/j.jep.2006.09.006

Lawrence, G.H.M, 1951. Taxonomy of Vascular Plants. Macmillan and Co, New York, pp. 628–629

Lewis, S. N., Brannan, L., Guri, A. J., Lu, P., Hontecillas, R., Bassaganya-Riera, J., & Bevan, D. R. (2011). Dietary α-eleostearic acid ameliorates experimental inflammatory bowel disease in mice by activating Peroxisome proliferator-activated receptor-γ. PLoS ONE, 6(8), e24031. http://doi.org/10.1371/journal.pone.0024031

Líder Agronomia. Romã. (2012). Available in: http://www.lideragronomia.com.br/2012/07/roma.html. Access in: 14 october 2018.

Lipid Libraray. (2018). Available in: http://lipidlibrary.aocs.org/Analysis/content.cfm?ItemNumber=40384. Access in: 10 october 2018

Liu, G., Xu, X., Gong, Y., He, L., & Gao, Y. (2012). Effects of supercritical CO2 extraction parameters on chemical composition and free radical-scavenging activity of pomegranate (Punica granatum L.) seed oil. Food and Bioproducts Processing, 90(3), 573–578. http://doi.org/10.1016/j.fbp.2011.11.004

Liu, S., Low, N. H., & Nickerson, M. T. (2009). Effect of pH, salt, and biopolymer ratio on 77

the formation of pea protein isolate gum arabic complexes. Journal of Agricultural and Food Chemistry, 57(4), 1521–1526. http://doi.org/10.1021/jf802643n

Liu, S., Low, N. H., & Nickerson, M. T. (2010a). Entrapment of flaxseed oil within gelatin- gum Arabic capsules. JAOCS, Journal of the American Oil Chemists’ Society, 87, 809– 815. http://doi.org/10.1007/s11746-010-1560-7

Liu, S., Low, N. H., & Nickerson, M. T. (2010b). Entrapment of flaxseed oil within gelatin- gum Arabic capsules. JAOCS, Journal of the American Oil Chemists’ Society, 87(7), 809–815. http://doi.org/10.1007/s11746-010-1560-7

Liu, W. N., & Leung, K. N. (2015). Jacaric acid inhibits the growth of murine macrophage- like leukemia PU5-1.8 cells by inducing cell cycle arrest and apoptosis. Cancer Cell International, 15(1), 90. http://doi.org/10.1186/s12935-015-0246-5

Lynch, J. M., Lock, A. L., Dwyer, D. A., Noorbakhsh, R., Barbano, D. M., & Bauman, D. E. (2005). Flavor and Stability of Pasteurized Milk with Elevated Levels of Conjugated Linoleic Acid and Vaccenic Acid*. Journal of Dairy Science, 88(2), 489–498. http://doi.org/10.3168/jds.S0022-0302(05)72711-9

Mallia, S., Piccinali, P., Rehberger, B., Badertscher, R., Escher, F., & Schlichtherle-Cerny, H. (2008). Determination of storage stability of butter enriched with unsaturated fatty acids/conjugated linoleic acids (UFA/CLA) using instrumental and sensory methods. International Dairy Journal, 18(10–11), 983–993. http://doi.org/10.1016/j.idairyj.2008.05.007

Manjanna, K. M., Shivakumar, B., & Kumar, T. M. P. (2010). Microencapsulation: An Acclaimed Novel Drug-Delivery System for NSAIDs in Arthritis. Critical Reviews™ in Therapeutic Drug Carrier Systems, 27(6), 509–545. http://doi.org/10.1615/CritRevTherDrugCarrierSyst.v27.i6.20

Márquez-Ruiz, G., García-Martínez, M., Holgado, F., & Velasco, J. (2014). Effectiveness of α-, γ- and δ-Tocopherol in a CLA-Rich Oil. Antioxidants, 3(1), 176–188. http://doi.org/10.3390/antiox3010176

McClements, D. J., & Decker, E. A. (1996). Lipids. In S. Domadoran, K. L. Parkin, & O. R. Fennema (Eds.), Fennema’s Food Chemistry (4th ed., pp. 155–216). Boca Raton. 78

Meerts, I. A. T. M., Verspeek-Rip, C. M., Buskens, C. A. F., Keizer, H. G., Bassaganya- Riera, J., Jouni, Z. E., … van de Waart, E. J. (2009). Toxicological evaluation of pomegranate seed oil. Food and Chemical Toxicology, 47(6), 1085–1092. http://doi.org/10.1016/j.fct.2009.01.031

Melgarejo, P. (2003) Tratado de fruticultura para zonas aridas y semoaridas. II. Algarrobo, granado y jinjolero. Mundi-prensa, Madrid.

Melo, I. L. P. de, Carvalho, E. B. T. de, Silva, A. M. de O. e, Yoshime, L. T., Sattler, J. A. G., Pavan, R. T., & Mancini-Filho, J. (2016). Characterization of constituents, quality and stability of pomegranate seed oil (Punica granatum L.). Food Science and Technology (Campinas), 36(1), 132–139. http://doi.org/10.1590/1678-457X.0069

Miller, A. B., & Claxton, E. (1928). Atmospheric Oxidation of Methyl and Glycol Esters of Beta–Eleostearic Acid. Industrial & Engineering Chemistry, 20(1), 43–48. http://doi.org/10.1021/ie50217a021

Mirmiran, P., Fazeli, M. R., Asghari, G., Shafiee, A., & Azizi, F. (2010). Effect of pomegranate seed oil on hyperlipidaemic subjects: A double-blind placebo-controlled clinical trial. British Journal of Nutrition, 104(3), 402–406. http://doi.org/10.1017/S0007114510000504

Mizrahi, M., Friedman-Levi, Y., Larush, L., Frid, K., Binyamin, O., Dori, D., … Gabizon, R. (2014). Pomegranate seed oil nanoemulsions for the prevention and treatment of neurodegenerative diseases: the case of genetic CJD. Nanomedicine: Nanotechnology, Biology, and Medicine, 10, 1353–1363. http://doi.org/10.1016/j.nano.2014.03.015

Mukherjee, C., Bhattacharyya, S., Ghosh, S., & Bhattacharyya, D. K. (2002). Dietary Effects of Punicic Acid on the Composition and Peroxidation of Rat Plasma Lipid. Journal of Oleo Science, 51(8), 513–522. http://doi.org/10.5650/jos.51.513

Nes, W. R., & Mckean, M. L. (1977). Biochemistry of steroids and other isopentenoids. Baltimore: University Park Press.

Nunes, J. C., & Torres, A. G. (2010). Fatty acid and CLA composition of Brazilian dairy products, and contribution to daily intake of CLA. Journal of Food Composition and Analysis, 23(8), 782–789. http://doi.org/10.1016/j.jfca.2010.03.023 79

O’Keefe, S. F. (2008). Nomenclature and classification of lipids. In C. C. Akoh & D. B. Min (Eds.), Food Lipids: Chemistry, Nutrition and Biotechnology (3rd ed., pp. 3–35). Boca Raton, FL.

Pande, G., & Akoh, C. C. (2009). Antioxidant capacity and lipid characterization of six georgia-grown pomegranate cultivars. Journal of Agricultural and Food Chemistry, 57(20), 9427–9436. http://doi.org/10.1021/jf901880p

Park, B. H., Vogelstein, B., & Kinzler, K. W. (2001). Genetic disruption of PPARdelta decreases the tumorigenicity of human colon cancer cells. Proceedings of the National Academy of Sciences of the United States of America, 98(5), 2598–2603. http://doi.org/10.1073/pnas.051630998

Pierucci, A. P. T. R., Andrade, L. R., Baptista, E. B., Volpato, N. M., & Rocha-Leão, M. H. M. (2006). New microencapsulation system for ascorbic acid using pea protein concentrate as coat protector. Journal of Microencapsulation, 23(6), 654–62. http://doi.org/10.1080/02652040600776523

Raikos, V. (2010). Effect of heat treatment on milk protein functionality at emulsion interfaces. A review. Food Hydrocolloids, 24(4), 259–265. http://doi.org/10.1016/j.foodhyd.2009.10.014

Rana, T. ., Narzary, D., & Ranade, S. A. (2010). Systematics and taxonomic disposition of the genus Punica L. In R. Chandra (Ed.), Pomegranate. Fruit Veg. Cereal Sci. Biotechnol (2nd ed., pp. 19–25).

Reineccius, G. A. (2004). The Spray Drying of Food Flavors. Drying Technology, 22(6), 1289–1324. http://doi.org/10.1081/DRT-120038731

Reische, D. W., Lillard, D. A., & Eitenmiller, R. R. (2008). Antioxidants. In C. C. Akoh & D. B. Min (Eds.), Food Lipids: Chemistry, Nutrition and Biotechnology (3rd ed., pp. 409– 430). Boca Raton, FL: CRC Press.

Riserus, U., Arner, P., Brismar, K., & Vessby, B. (2002). Treatment With Dietary trans10cis12 Conjugated Linoleic Acid Causes Isomer-Specific Insulin Resistance in Obese Men With the Metabolic Syndrome. Diabetes Care, 25(9), 1516–1521. http://doi.org/10.2337/diacare.25.9.1516 80

Rodrigues, M. I., & Lemma, A. F. (2014). Considerações Iniciais. In M. I. Rodrigues & A. F. Lemma (Eds.), Planejamento de Experimentos e Otimização de Processo (3a, pp. 1–7). Campinas, SP: Casa do Espírito Amigo Fraternidade Fé e Amor.

Rodríguez-Alcalá, L. M., & Fontecha, J. (2007). Hot topic: Fatty acid and conjugated linoleic acid (CLA) isomer composition of commercial CLA-fortified dairy products: evaluation after processing and storage. Journal of Dairy Science, 90, 2083–2090. http://doi.org/10.3168/jds.2006-693

Rosenberg, M., Kopelman, I. J., & Talmon, Y. (1990). Factors affecting retention in spray drying microencapsulation of volatile materials. Journal of Agricultural and Food Chemistry, 38(5), 1288–1294. http://doi.org/10.1021/jf00095a030

Sahin-Nadeem, H., & Afşin Özen, M. (2014). Physical properties and fatty acid composition of pomegranate seed oil microcapsules prepared by using starch derivatives/whey protein blends. European Journal of Lipid Science and Technology, n/a-n/a. http://doi.org/10.1002/ejlt.201300355

Sahin-Nadeem, H., & Özen, M. A. (2014). Physical properties and fatty acid composition of pomegranate seed oil microcapsules prepared by using starch derivatives/whey protein blends. European Journal of Lipid Science and Technology, 116(7), 847–856. http://doi.org/10.1002/ejlt.201300355

Santos, M. G., Bozza, F. T., Thomazini, M., & Favaro-Trindade, C. S. (2015). Microencapsulation of xylitol by double emulsion followed by complex coacervation. Food Chemistry, 171, 32–39. http://doi.org/10.1016/j.foodchem.2014.08.093

Sassano, G., Sanderson, P., Franx, J., Groot, P., Van Straalen, J., & Bassaganya-Riera, J. (2009). Analysis of pomegranate seed oil for the presence of jacaric acid. Journal of the Science of Food and Agriculture, 89(6), 1046–1052. http://doi.org/10.1002/jsfa.3552

Schugens, C., Laruelle, N., Nihant, N., Grandfils, C., Jérome, R., & Teyssié, P. (1994). Effect of the emulsion stability on the morphology and porosity of semicrystalline poly l-lactide microparticles prepared by w/o/w double emulsion-evaporation. Journal of Controlled Release, 32(2), 161–176. http://doi.org/10.1016/0168-3659(94)90055-8

Serafeimidou, A., Zlatanos, S., Kritikos, G., & Tourianis, A. (2013). Change of fatty acid profile, including conjugated linoleic acid (CLA) content, during refrigerated storage of 81

yogurt made of cow and sheep milk. Journal of Food Composition and Analysis, 31, 24– 30. http://doi.org/10.1016/j.jfca.2013.02.011

Shabbir, M. A., Khan, M. R., Saeed, M., Pasha, I., Khalil, A. A., & Siraj, N. (2017). Punicic acid: A striking health substance to combat metabolic syndromes in humans. Lipids in Health and Disease, 16(1), 1–9. http://doi.org/10.1186/s12944-017-0489-3

Shahidi, F., & Wanasundara, U. N. (2008). Methods for measuring oxidative rancity in fats and oils. In C. C. Akoh & D. B. Min (Eds.), Food Lipids: Chemistry, Nutrition and Biotechnology (3rd ed., pp. 387–403). Boca Raton: CRC Press.

Shantha, N. C., & Decker, E. A. (1994). Rapid, sensitive, iron-based spectrophotometric methods for determination of peroxide values of food lipids. Journal of AOAC International, 77(October), 421–424.

Sheu, T.-Y., & Rosenberg, M. (1998). Microstructure of microcapsules consisting of whey proteins and carbohydrates. Journal of Food Science, 63(3), 491–494.

Shinohara, N., Tsuduki, T., Ito, J., Honma, T., Kijima, R., Sugawara, S., … Ikeda, I. (2012). Jacaric acid, a linolenic acid isomer with a conjugated triene system, has a strong antitumor effect in vitro and in vivo. Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids, 1821(7), 980–988. http://doi.org/10.1016/j.bbalip.2012.04.001

Silva, F. A. M., Borges, M. F. M., & Ferreira, M. A. (1999). Métodos para avaliação do grau de oxidação lipídica e da capacidade antioxidante. Quimica Nova, 22(1), 94–103. http://doi.org/10.1590/S0100-40421999000100016

Silva, L., Pinto, J., Carrola, J., & Paiva-Martins, F. (2010). Oxidative stability of olive oil after food processing and comparison with other vegetable oils. Food Chemistry, 121(4), 1177–1187. http://doi.org/10.1016/j.foodchem.2010.02.001

Spilmont, M., Léotoing, L., Davicco, M.-J., Lebecque, P., Mercier, S., Miot-Noirault, E., … Coxam, V. (2013). Pomegranate seed oil prevents bone loss in a mice model of osteoporosis, through osteoblastic stimulation, osteoclastic inhibition and decreased inflammatory status☆. The Journal of Nutritional Biochemistry, 24, 1840–1848. http://doi.org/10.1016/j.jnutbio.2013.04.005 82

Sreekumar, S., Sithul, H., Muraleedharan, P., Azeez, J. M., & Sreeharshan, S. (2014). Pomegranate fruit as a rich source of biologically active compounds. BioMed Research International, 2014. http://doi.org/10.1155/2014/686921

Suzuki, R., Abe, M., & Miyashita, K. (2004). Comparative study of the autoxidation of TAG

containing conjugated and nonconjugated C 18 PUFA. Journal of the American Oil Chemists’ Society, 81(6), 563–569. http://doi.org/10.1007/s11746-006-0942-3

Suzuki, R., Noguchi, R., Ota, T., Abe, M., Miyashita, K., & Kawada, T. (2001). Cytotoxic effect of conjugated trienoic fatty acids on mouse tumor and human monocytic leukemia cells. Lipids, 36(5), 477–482. http://doi.org/10.1007/s11745-001-0746-0

Teixeira da Silva, J. A., Rana, T. S., Narzary, D., Verma, N., Meshram, D. T., & Ranade, S. A. (2013). Pomegranate Biology And Biotechnology: A Review. Scientia Horticulturae, 160, 85–107. http://doi.org/10.1016/j.scienta.2013.05.017

Tonon, R. V, Grosso, C. R. F., & Hubinger, M. D. (2011). Influence of emulsion composition and inlet air temperature on the microencapsulation of flaxseed oil by spray drying microencapsulation of flaxseed oil by spray drying. Food Research International, 44(1), 282–289. http://doi.org/10.1016/j.foodres.2010.10.018

Tran, H. N. A., Bae, S.-Y., Song, B.-H., Lee, B.-H., Bae, Y.-S., Kim, Y.-H., … Newman, R. A. (2010). Pomegranate ( Punica granatum ) Seed Linolenic Acid Isomers: Concentration-Dependent Modulation of Estrogen Receptor Activity. Endocrine Research, 35(1), 1–16. http://doi.org/10.3109/07435800903524161

Tsuboyama-Kasaoka, N., Takahashi, M., Tanemura, K., Kim, H. J., Tange, T., Okuyama, H., … Ezaki, O. (2000). Conjugated linoleic acid supplementation reduces adipose tissue by apoptosis and develops lipodystrophy in mice. Diabetes, 49(9), 1534–1542. http://doi.org/10.2337/diabetes.49.9.1534

Tsuzuki, T., Kawakami, Y., Abe, R., Nakagawa, K., Koba, K., Imamura, J., … Miyazawa, T. (2006). Conjugated linolenic acid is slowly absorbed in rat intestine, but quickly converted to conjugated linoleic acid. The Journal of Nutrition, 136(8), 2153–2159.

Turgeon, S. L., Schmitt, C., & Sanchez, C. (2007). Protein–polysaccharide complexes and coacervates. Current Opinion in Colloid & Interface Science, 12(4–5), 166–178. http://doi.org/10.1016/J.COCIS.2007.07.007 83

Van der Heide, P. (2012). Introduction. In X-Ray Photoelectron Spectroscopy (pp. 1–13). Hoboken, New Jersey: John Wiley & Sons, Inc.

Vega, C., & Roos, Y. H. (2006). Invited Review: Spray-Dried Dairy and Dairy-Like Emulsions—Compositional Considerations. Journal of Dairy Science, 89(2), 383–401. http://doi.org/10.3168/jds.S0022-0302(06)72103-8

Verardo, V., Garcia-Salas, P., Baldi, E., Segura-Carretero, A., Fernandez-Gutierrez, A., & Caboni, M. F. (2014). Pomegranate seeds as a source of nutraceutical oil naturally rich in bioactive lipids. Food Research International, 65(PC), 445–452. http://doi.org/10.1016/j.foodres.2014.04.044

Vichi, S., Pizzale, L., Conte, L. S., Buxaderas, S., & López-Tamames, E. (2003). Solid-Phase Microextraction in the Analysis of Virgin Olive Oil Volatile Fraction: Modifications Induced by Oxidation and Suitable Markers of Oxidative Status. Journal of Agricultural and Food Chemistry, 51(22), 6564–6571.

Viladomiu, M., Hontecillas, R., Lu, P., & Bassaganya-Riera, J. (2013). Preventive and prophylactic mechanisms of action of pomegranate bioactive constituents. Evidence- Based Complementary and Alternative Medicine, 2013, doi: 10.1155/2013/789764. http://doi.org/10.1155/2013/789764

Vroegrijk, I. O. C. M., van Diepen, J. A., van den Berg, S., Westbroek, I., Keizer, H., Gambelli, L., … Voshol, P. J. (2011). Pomegranate seed oil, a rich source of punicic acid, prevents diet-induced obesity and insulin resistance in mice. Food and Chemical Toxicology, 49(6), 1426–1430. http://doi.org/10.1016/j.fct.2011.03.037

Weinbreck, F., de Vries, R., Schrooyen, P., & de Kruif, C. G. (2003). Complex coacervation of whey proteins and gum arabic. Biomacromolecules, 4(2), 293–303. http://doi.org/10.1021/bm025667n

Weinbreck, F., Rollema, H. S., Tromp, R. H., & De Kruif, C. G. (2004). Diffusivity of whey protein and gum arabic in their coacervates. Langmuir, 20(14), 6389–6395. http://doi.org/10.1021/la049908j

Weinbreck, F., Tromp, R., & de Kruif, C. (2004). Composition and Structure of Whey Protein / Gum Arabic Coacervates. Biomacromolecules, 5, 1437–1445. Retrieved from http://pubs.acs.org/doi/abs/10.1021/bm049970v 84

Weinbreckt, F., & Minorf, M. (2004). Microencapsulation of oils using whey protein / gum arabic coacervates. Journal of Microencapsulation, 21(6), 667–679. http://doi.org/10.1080/02652040400008499

Weinbreckt, F., Minorf, M., & Kruif, C. G. (2004). Microencapsulation of oils using whey protein / gum arabic coacervates. Journal of Microencapsulation, 21(6), 667–679. http://doi.org/10.1080/02652040400008499

Weinbrreck, F., Tromp, R. H., & de Kruif, C. G. (2004). Composition and structure of whey protein/gum arabic coacervates. Biomacromolecules, 5(4), 1437–1445. http://doi.org/10.1021/bm049970v

Wijayanti, H. B., Bansal, N., & Deeth, H. C. (2014). Stability of Whey Proteins during Thermal Processing: A Review. Comprehensive Reviews in Food Science and Food Safety, 13(6), 1235–1251. http://doi.org/10.1111/1541-4337.12105

Wu, B. C., Degner, B., & McClements, D. J. (2014). Soft matter strategies for controlling food texture: Formation of hydrogel particles by biopolymer complex coacervation. Journal of Physics Condensed Matter, 26(46). http://doi.org/10.1088/0953- 8984/26/46/464104

Xiao, Z., Liu, W., Zhu, G., Zhou, R., & Niu, Y. (2014a). A review of the preparation and application of flavour and essential oils microcapsules based on complex coacervation technology. Journal of the Science of Food and Agriculture, 94(8), 1482–1494. http://doi.org/10.1002/jsfa.6491

Xiao, Z., Liu, W., Zhu, G., Zhou, R., & Niu, Y. (2014b). Production and characterization of multinuclear microcapsules encapsulating lavender oil by complex coacervation. Flavour and Fragrance Journal, 29(3), 166–172. http://doi.org/10.1002/ffj.3192

Yamasaki, M., Kitagawa, T., Koyanagi, N., Chujo, H., Maeda, H., Kohno-Murase, J., … Yamada, K. (2006). Dietary effect of pomegranate seed oil on immune function and lipid metabolism in mice. Nutrition, 22(1), 54–59. http://doi.org/10.1016/j.nut.2005.03.009

Yang, L., Cao, Y., Chen, J. N., & Chen, Z. Y. (2009). Oxidative stability of conjugated linolenic acids. Journal of Agricultural and Food Chemistry, 57(10), 4212–4217. http://doi.org/10.1021/jf900657f 85

Yang, X., Gao, N., Hu, L., Li, J., & Sun, Y. (2015). Development and evaluation of novel microcapsules containing poppy-seed oil using complex coacervation. Journal of Food Engineering, 161, 87–93. http://doi.org/10.1016/j.jfoodeng.2015.03.027

Yuan, G., Sinclair, A. J., Xu, C., & Li, D. (2009). Incorporation and metabolism of punicic acid in healthy young humans. Molecular Nutrition and Food Research, 53(10), 1336– 1342. http://doi.org/10.1002/mnfr.200800520

Chapter 2

Chemical composition of commercial cold-pressed pomegranate (Punica granatum) seed oil from Turkey and Israel, and the use of bioactive compounds for samples’ origin preliminary discrimination

Authors: Costa, André M. M.; Silva, Laís O.; Torres, Alexandre G.

Abstract

This study investigated the chemical composition of three commercial pomegranate seed oils (PSO) from Turkey and Israel, with emphasis on PSO’s bioactivity and quality, as follows: fatty acid profile, lipid classes, volatile compounds, quality indexes, oxidative stability and bioactive compounds (phenolics, tocopherols and carotenoids). Principal component analysis (PCA) was used to identify regional trends in the PSO, based on bioactive compounds profile. PSO samples were similar concerning quality indexes and lipid classes. Punicic acid content (ca. 590 mg/g oil) was similar among all samples however, total conjugated linolenic acid

(cLnA) content was higher in PSOIsrael_A (87.0 g/100 g). Phenolic acids, quercetin and naringenin were the phenolic compounds present in PSO samples. γ-Tocopherol was the major tocol, especially in PSOTurkey_batches_1/2 (ca. 1143mg/100 g), but samples showed low carotenoids contents. PSO showed low oxidative stability, possibly because of cLnA’s oxidizability. The volatile fraction of PSOs varied between samples and was composed chiefly by alcohols, aldehydes, ketones, esters and carboxylic acids. Bioactive compounds data provided a marginal discrimination of PSO’s geographical origin, by PCA analysis. The high contents of minor components in PSO add to its potential bioactivity attributed to cLnA.

Keywords

Food analysis; food composition; pomegranate seed oil; punicic acid;phenolic acids; gamma- tocopherol;volatilecompounds; geographical origin; chemical composition variation

Chemical compounds studied in this article:

-eleostearic acid (CID- 5281115); -tocopherol (CID-14985); -carotene (CID- 5280489);- eleostearic acid (CID- 5282820); catalpic acid (CID- 5385589); -tocopherol (CID- 92094);- tocopherol (CID- 92729); punicic acid (CID-5281126); vanillic acid (CID- 8468). 88

1. Introduction

Pomegranate (Punica granatum) is a shrub tree native from Iran, but highly adaptable to varying climate conditions. Pomegranate is greatly appreciated for its brightly-red and very succulent pulp fruit. Currently, pomegranate is cultivated worldwide, but the largest producers are the Mediterranean countries. The fruit has three major parts: the outside peel, the inside peel (white septal membranes) and the edible arils, which are juice-containing sacs formed by pomegranate’s seeds and pulp. Pomegranate’s pulp is sweet and highly appreciated for its flavor and high contents of bioactive compounds, especially phenolic acids. Apart from being consumed fresh, the principal commercial use of the arils are in the juice industry and the seeds are a byproduct of this activity (Teixeira da Silva et al., 2013).

The seeds are a rich source of lipids, which correspond to 12-20% of total seed weight (Elfalleh et al., 2011; Fernandes et al., 2015). Pomegranate seed oil (PSO) presents high contents of conjugated linolenic acids (cLnA), which is a collective term for the positional and geometric isomers of linolenic acid possessing three conjugated double bounds, usually in positions Δ9,11,13, and Δ8,10,12 and with varying combinations of geometrical configurations, cis or trans (Cao et al., 2006). PSO has a unique fatty acid profile, being the oil most concentrated in the cLnA isomer punicic acid, with over 60 % of this fatty acid. cLnA isomers are associated with health benefits such as cytotoxic effect in mouse tumor cells, antioxidant properties in rat tissues and modulation of the immune system, glucose uptake and lipid metabolism by the activation of peroxisome proliferator-activated receptors (PPARs), among others (Dharet al., 1999; Suzuki et al., 2001; Viladomiu et al., 2013). Other bioactive compounds described in PSO are tocopherols, sterols and phenolic compounds (Fernandes et al., 2015; Khoddami et al., 2014; Pande & Akoh, 2009).

Although PSO is emerging as a functional food there is scant data about the detailed chemical composition of cold pressed PSO available in the consumer’s market, in order to assess how PSO bioactivity would compare to other highly consumed edible oils. These results could also provide valuable data to establish how health benefits of PSO are related to its composition, and eventually stimulate its consumption. To the best of our knowledge, there are three published articles (Caligiani et al., 2010; Khoddami et al., 2014; Melo et al., 2016) that evaluated commercial PSO’s oil quality, oxidative stability, fatty acid and volatile profiles and bioactive compounds content (tocopherols, phythosterols and total phenolics content by spectrophotometric assay), but these studies left gaps concerning contents of 89

bioactive compounds, such as carotenoids and phenolics profile. Commercial PSO phenolics profile and carotenoids content (by HPLC) and volatiles profile (by GC-MS-FID) are still unpublished, and these data could bring more awareness to the potential bioactive properties of PSO, or could help in the development of PSO based products.

In this sense, this study aimed at performing a detailed chemical characterization of three commercial PSO samples from different origins, with emphasis on a detailed fatty acid profile, lipid classes, volatile compounds profile and bioactive compounds (phenolics, tocopherols and carotenoids), as well as quality indexes and oxidative stability. The present work also aimed at identifying possible regional trends in the PSO’s bioactive compounds profile (total phenolics, total tocopherols, β-carotene and total cLnA) by applying principal component (PCA) analysis.

2. Materials and Methods

2.1 Samples and Reagents

Cold pressed commercial pomegranate seed oils (PSO) were obtained from Turkey (Turkey, batches 1 and 2) and Israel (Israel Brand A; and Israel Brand B,batches 1 and 2). Tocopherols (α-, β-, γ-, δ-) (≥ 99% purity), fatty acid methyl esters (37 FAME mix), glyceryl triheptadecanoate (≥ 99% purity), bromobenzene methanolic solution, hydrocarbons standards mixture (C7-C30 saturated alkanes), and Folin-Ciocalteau’s reagent were from Sigma-Aldrich Chemical Co. (St. Louis, MO, US) and Supelco Co. (Bellefonte, PA, US). cLnA standards (punic acid, α-eleostearic acid, catalpic acid and β-eleostearic acid) (> 98% purity) were from Larodan AB (Solna, Sweden). Anthocyanins (> 97% purity) and non-anthocyanins (> 85% purity) standards were, respectively, from Indofine Chemical Co. (Hillsborough, NJ, US) and Sigma-Aldrich Chemical Co (St. Louis, MO, US). Carotenoids were isolated by open column chromatography from natural food sources (Rodriguez-Amaya & Kimura, 2004), as follows: α- and β-carotene from carrot, lutein, and zeaxanthin from corn, and lycopene from tomato. Standards of chlorophylls a and b were obtained from spinach (Yentsch & Menzel, 1963). All pigments’ standards showed purity grade >95%, determined by HPLC-PDA. All solvents used were HPLC grade from Tedia (Goiania, GO, Brazil). HPLC grade water (Milli-Q system, Millipore, Danvers, MA, US) was used throughout the experiments.

2.2 Pomegranate Seed Oil Quality Indexes and Oxidative Stability

The peroxide and p-anisidine values were determined according to Official Methods of AOCS, respectively, methods Cd 8-53(3) and Cd 18-90(11)(American Oil Chemists’ Society, 2012).Acid value was determined by potentiometric titration in an automatic titrator model G20 (Mettler Toledo,Urdorf, Switzerland) (Greco-Duarte, Cavalcanti-Oliveira, Da Silva, Fernandez-Lafuente, & Freire, 2017). Briefly, oils’ acidity was analyzed by titrating with 0.4 M NaOH the PSO commercial samples (100 µL) dissolved in 45 mL acetone:ethanol (1:1, v/v) until pH 11.0. Results were expressed as % punicic acid. Refractive index was determined directly in a portable digital refractometer (ATAGO, N-3E; Tokyo, Japan). Oxidative stability was determined in a Rancimat® apparatus (Metrohm 743; Metrohm Co., Herisau, Switzerland). PSO (3 g) was heated at 80 °C with a 20.0 L/h air flow until a sharp increase in the water conductivity which corresponds to the oil stability index (OSI; h). Data were collected and processed using 743 Rancimat software (version 1.1, 2010; Metrohm Co., Herisau, Switzerland).

2.3 Fatty acid composition by GC-FID

Commercial PSO fatty acids were methylated by a base-catalyzed transesterification procedure (Kramer et al.,1997) to avoid isomerization. PSOs’ fatty acid methyl esters (FAME) were analyzed by gas chromatography coupled with flame ionization detector (GC- FID) and equipped with a split/splitless injector (GC-2010 chromatograph; Shimadzu Co., Kyoto, Japan) (Nunes & Torres, 2010). Briefly, FAME were separated in a capillary column (30 m × 0.32 mm i.d., 0.25 µm film; Omegawax-320;Supelco Co., Bellefonte, PA, US). Helium was used as the carrier gas, and column pressure was set to attain a carrier gas speed of 25.0 cm/s. Temperatures of injector and detector were 260 °C and 280 °C, respectively, and the split ratio was 1:20. The oven program was as follows: 160 °C for 2 min, temperature programmed at 2.5 °C/min to 190 °C and held for 5 min, then temperature programmed at 3.5 °C/min to 220 °C and held for 15 min. Data were collected by Lab Solutions GC software package (version 2.30.00, 2004; Shimadzu Co., Kyoto, Japan). Gas-chromatographic peaks of FAME were identified by comparison with a commercial mixture of standards (37-component FAME mix; Supelco Co., Bellefonte, PA, US) and, specifically for cLnA isomers, individual commercial isomers were used (punic acid, α-eleostearic acid, catalpic acid and β-eleostearic acid; Larodan AB,Solna, Sweden). A solution of glyceryl triheptadecanoate in hexane (1.456 mg/mL) was used as internal standard for fatty acid quantification. Results of single fatty 91

acids were expressed as mg/g oil and fatty acids groups as g/100 g fatty acid.

cLnA isomers peak identities were confirmed by gas chromatography coupled to mass spectrometry (GC-MS). Briefly, PSOs’ FAME were derivatized to dimethyloxazoline (DMOX) fatty acid derivatives (Christie, 2003)and analyzed on a GC-17A gas chromatograph equipped with a QP5050A mass spectrometer (Shimadzu Co., Kyoto, Japan).The capillary column used was the same and the chromatographic conditions were similar to those used for GC-FID analysis, described above. The mass spectrometer was operated in electron impact mode at 70 eV. The interface and ion source temperature were 260 °C. Analyses were performed in full scan acquisition mode, in the mass range of 20 to 700 m/z at 0.5 scan/s. Data were collected by Lab Solutions GC-MS (version 1.21, 2008; Shimadzu Co., Kyoto, Japan). Compounds were identified based on published data of DMOX derivatives of trienoic fatty acids mass spectra(Spitzer, 1996) (ANNEX 1).

2.4 Lipid classes distribution by HPLC-ELSD

Lipid classes in PSOs (MAG: monoacylglycerols; DAG: diacylglycerols; TAG: triacylglycerols and FFA: free fat acids) were analyzed by a high performance liquid chromatography (HPLC) system (Shimadzu Co., Kyoto, Japan) consisting of a quaternary pump LC-20AT, evaporative light-scattering detector ELSD-LT II, system controller CBM- 20A and degasser DGU-20A5 (Holcapek, Lísa, Jandera, & Kabátová, 2005). Samples (20 µL) dissolved in acetonitrile:iso-propanol:hexane (2:2:1, v/v/v) were separated in a C18 column (250 × 4.6 mm, 5 µm, Kromasil®, Bohus, Sweden). The mobile phase consisted of a gradient of acetonitrile (eluent A) and isopropanol (eluent B), at 1.0 mL/min flow rate. The gradient elution program started with 100% (A), changed linearly to 69% (B) at 60 min, and kept constant until 75 min, followed by a 4 min re-equilibration at 100% (A), starting at 76 min, for a total analysis time of 80 min. ELSD detector conditions were nebulizer gas (N2) pressure of 3.5 bar and nebulizing temperature 40 °C. Data were acquired by LC solution software (version 1.25, 2009; Shimadzu Co., Kyoto, Japan). Lipid classes contents were calculated by internal normalization and contents expressed as Area%.

2.5 Volatile Compounds by SPME-GC-MS/FID

analyzed by GC-FID-MS as described previously (Akil et al.,2015; Costa et al., 2015). Prior to extraction, the fiber was conditioned for 60 min in the GC injection port at 270 °C. PSOs samples (1 g) were weighed in a headspace vial, and 20 µL of internal standard (0.1 mg/mL of bromobenzene; Supelco Co.; Bellefonte, PA, US) in methanol was added, and homogenized, followed by sealing the flasks with a PTFE-lined septum. Headspace vials were placed in a glycerol bath (40 °C) for 30 min until equilibration, and septum was pierced, and the fiber was exposed to the sample headspace for 10 min.

Qualitative analysis was performed by GC-MS on a GC-17A gas chromatograph coupled to a QP5050A mass spectrometer (Shimadzu Co.; Kyoto, Japan) equipped with a split/splitless injector and a fused capillary silica column 5% phenyl /95% methylpolysiloxane (30 m × 0.32 i.d., 3 µm film; 007-5; Quadrex, Bethany, CT, US). Volatile compounds were desorbed from the SPME fiber in the injection port for 3 min at 260 °C, in splitless mode, and after 3 min sampling the split purge valve was open at 3.0 mL/min. Helium was used as the carrier gas and the column pressure was set at 49 kPa. The column oven temperature was held at 30 °C for 10 min, then increased at 3 °C/min to 200 °C and held for 10 min. The mass spectrometer was operated in electron impact mode at 70 eV, and the interface and ion source temperatures were 260 °C. Analyses were performed in full scan acquisition mode, at a mass range of 40 to 500 m/z, at 0.5 scan/s. A mixture of hydrocarbons (C7-C30 saturated alkanes; Supelco Co.; Bellefonte, PA, US) was run under the same conditions and used as standards to allow calculation of linear retention index (LRI) values for the volatile compounds (Viegas & Bassoli, 2007). Data were collected by Lab Solutions GC-MS software package (version 1.21, 2008; Shimadzu Co., Kyoto, Japan). Compounds were tentatively identified utilizing these following data: 1) Comparison of mass spectra with those of National Institute of Standards (NIST) library and calculation of similarity indexes (SI) provided by the instrument’s software (Lab Solutions GC-MS; Shimadzu Co., Kyoto, Japan); 2) comparing (LRIexperimental and LRIpublished values and calculating the variation between them (variation (%) =

(((LRIexperimental-LRIpublished)/LRIexperimental) × 100)) ; and 3) analysis was also performed in a polar column (polyethylene glycol, 30 m × 0.32 mm i.d., 0.25 µm film thickness, Omegawax- 320; Supelco Co., Bellefonte, PA, US) to provide more accuracy in the confirmation of the compounds. Based in this previous information, the criteria for identification of the volatile compounds in the non-polar column (007-5; Quadrex, Bethany, CT, US) were: 1) SI > 80% and variation between LRIexperimental and LRIpublished values < 5%; or 2) SI > 95% and no

LRIpublished values available. Whenever possible, identity was confirmed in the polar column, 93

and the criteria for identification in this case were: 1) SI > 75% and variation between

LRIexperimental/LRIpublished< 2%; or 2) SI > 85% and no LRIpublished values available.LRIpublished were obtained from free databases available on the internet (for full references, see Table 3).

Quantitative analysis of volatile compounds was performed in a GC-2010 (Shimadzu Co., Kyoto, Japan) gas chromatograph equipped with a FID detector, split/splitless injector, and the same capillary column used for qualitative analysis. Chromatographic conditions were similar to those described in section 2.5. During analysis, injector and detector temperatures were 260 and 280 °C, respectively. A mixture of C7-C30 hydrocarbons was run under the same conditions to allow calculation of LRI values for the volatile compounds and comparison with GC-MS data. Additionally, the same analysis was also performed in the polyethylene glycol column described above to provide additional LRI values for comparison with GC-MS data. Data were collected by Lab Solutions GC software package (version 2.30.00, 2004; Shimadzu Co., Kyoto, Japan). Unfortunately, not all compounds identified by SPME-GC-MS were possible to quantify, as some compounds were undetectable or unidentifiable in GC-FID analysis. A detection threshold was established, once compounds with contents under 0.02 µg/g were considered unidentified. Results were expressed in µg of the volatile compound/g oil.

2.6 Bioactive Compounds in Pomegranate Seed Oil

2.6.1 Phenolic compounds by HPLC-PDA

Non-anthocyanin phenolic compounds were extracted according to Siger et al. (2007). Briefly, these compounds were extracted from PSOs in diol solid phase extraction (SPE) cartridges (Bond Elut 2-OH, 3 mL, 500 mg; Agilent Technologies,Santa Clara, CA, US). Briefly, cartridges were conditioned (5 mL of methanol and 5 mL of n-hexane, sequentially) and 2.5 g of sample in n-hexane (1:2 w/v) were added. The cartridge was washed with 5 mL of n-hexane:ethyl acetate (9:1, v/v) and non-anthocyanin phenolic compounds were eluted with pure methanol and collected. Anthocyanins were extracted by liquid-liquid extraction with methanol (Tuberoso et al., 2007). Samples were dissolved in n-hexane (1:2 w/v) and anthocyanins were extracted with 80% aqueous methanol. Solutions were vortex-mixed for 10 min and the lower phase was removed and transferred to a round bottom flask. Extraction was repeated three times and all extracts were combined. After evaporating the solvent in a 94

Rotavapor® R-215 (BÜCHI, Flawil, Switzerland), the residue was suspended in a known volume of 80% aqueous methanol.

Analysis of non-anthocyanin phenolic compounds was performed in a HPLC system (Shimadzu Co., Kyoto, Japan) consisting of a quaternary pump LC-20AT, photodiode array detector (PDA) SPD-M20A, system controller CBM-20A and degasser DGU-20A5 according to De La Torre-Carbot et al. (2005) with adaptations. Chromatographic separation of non- anthocyanin phenolic compounds was achieved using a reversed-phase C18 column (5 μm, 250 mm × 4.6 mm, Kromasil®, Bohus, Sweden). The mobile phase consisted of a gradient of 0.3% aqueous formic acid (eluent A), methanol (eluent B) and acetonitrile (eluent C), at 1.0 mL/min flow rate. Eluent C concentration was kept constant at 1% during analysis. Prior to injection, the column was equilibrated with 24% B. The gradient elution program was as follows: 24% (B) at 8 min, 28% (B) at 18 min, 33% (B) at 30 min, 65% (B) at 60 min, 24% (B) at 60.1 min, followed by 15 min re-equilibration, for a total analysis cycle time of 75 min. Non-anthocyanin phenolic compounds were monitored by PDA, from 190 to 370 nm. Data were acquired by LC solution software (version 1.25, 2009; Shimadzu Co., Kyoto, Japan).

Analysis of anthocyanins was performed at the same HPLC system (Shimadzu Co., Kyoto, Japan) described above (section 2.6.1) and according to Brown & Shipley (2011). Chromatographic separation of anthocyanins was achieved using a reversed-phase C18 column (5 μm, 250 mm × 4.6 mm, Kromasil®, Bohus, Sweden). The mobile phase consisted of a gradient of 1% aqueous formic acid (eluent A), 1% formic acid in methanol (eluent B) and acetonitrile (eluent C), at 2.0 mL/min flow rate. Eluent C concentration was kept constant at 2% during analysis. Prior to injection, the column was equilibrated with 18% B. The gradient elution program was as follows: 18% (B) at 2 min, 32% (B) at 6 min, 52% (B) at 8 min, 18% (B) at 8.1 min, followed by 15 min re-equilibration, for a total analysis cycle time of 23 min. Anthocyanins were monitored at 530 nm. Data were acquired by LC solution software (version 1.25, 2009; Shimadzu Co., Kyoto, Japan).

Phenolic compounds were identified based on retention times of commercial standards, and peak identity was confirmed by standard spiking, and by comparison of absorption spectra with their respective standards. Commercial standards of the following phenolic compounds were injected for assessing chromatographic peaks identities in PSO: 1. anthocyanins: cyanidin chloride (3,3’,4’,5,7-pentahydroxyflavylium chloride), cyanin chloride (cyanidin-3,5-diglucoside chloride), keracyanin chloride (cyanidin-3-O-rutinoside chloride), 95

kuromanin chloride (cyanidin-3-O-glucoside chloride), malvidin chloride, pelargonidin chloride (3,4’,5,7-tetrahydroxyflavylium chloride); 2. non-anthocyanins: benzoic acid, caffeic acid, 5-caffeoylquinic acid, 3,4-dihidroxy-phenylacetic acid, ellagic acid, ferulic acid, gallic acid, 2,4-hydroxy-benzoic acid, hydroxycinnamic 4-hydroxyphenylacetic acid, hippuric acid, kaempferol, m-coumaric acid, myricetin, naringenin, p-coumaric acid, p-hydroxy-benzoic acid, quercetin, rosmarinic acid, rutin, salicylic acid, sinapic acid, syringic acid, trans- cinnamic acid and vanillic acid.Quantitative analysis was based on external calibration with commercial standards. Calibration curves ranging from 0.5 to 10.0 µg/mL were linear for all anthocyanin standards (R² > 0.99, p< 0.02; Limit of detection maximum value among compounds in this class (LODmax): 0.19 µg/mL, Limit of quantification maximum value among compounds in this class (LOQmax): 0.66 µg/mL) and non-anthocyanin phenolic compounds (R² > 0.99, p< 0.01; LODmax: 0.80µg/mL, LOQmax: 2.42 µg/mL). Analysis was performed in triplicate and results were expressed in mg/100 g.

2.6.2 Tocopherols, chlorophylls and carotenoids by HPLC-FLU/PDA

The contents of tocopherols, unesterified carotenoids and chlorophylls in PSOs were determined simultaneously in a HPLC system (Shimadzu Co., Kyoto, Japan) consisting of a quaternary pump LC-20AT, system controller CBM-20A, degasser DGU-20A5, photodiode array detector (PDA) SPD-M20A and fluorescence detector RF-10AXL according to Rahmani & Csallany, (1991) and Tan & Brzuskiewicz (1989). Compounds extraction was achieved by the method proposed by Gimeno et al. (2000). Briefly, samples (50 mg) were dissolved in n-hexane (950 µL), centrifuged (10,000 g, 5 min) and the supernatant solution was filtered through a PTFE syringe filter (0.45 µm). Chromatographic separation was achieved using a normal phase silica column (Zorbax-SIL 5 µm, 4.6 mm  250 mm; Agilent Technologies, Santa Clara, CA, US) with isocratic elution (hexane:isopropanol, 99:1, v/v) at 1.0 mL/min. Fluorescence detector was operated at 290 nm (excitation) and 330 nm (emission) for tocopherols. PDA detector scanned from 390 to 700 nm, and λ at 450 nm and 665 nm were used for quantification of carotenoids and chlorophylls, respectively. Data were acquired by LC solution software (version 1.25, 2009; Shimadzu Co., Kyoto, Japan).

Commercial standards of α-, β-, γ- and δ-tocopherols were used for tocopherols identification and quantification by external calibration. Concentrations of each tocol, carotenoid and chlorophyll standard stock solutions were determined spectrophotometrically after appropriate dilutions using the following specific extinction coefficients (λ; dL/g.cm) in 96

ethanol: α-tocopherol (292 nm; 75.8); β-tocopherol (296 nm; 89.4); γ-tocopherol (298 nm; 91.4); δ-tocopherol (298 nm; 87.3); α-carotene (450 nm; 2800); β-carotene (450 nm; 2620); lutein (450 nm; 2550); lycopene (450 nm; 3450) and zeaxanthin (450 nm; 2540); and in acetone: chlorophyll a (665 nm; 670) and chlorophyll b (665 nm; 518) (Franke et al., 2007; Minguez-Mosquera et al., 1989; Rodriguez-Amaya and Kimura, 2004). Calibration curves were linear for all tocols (0.5 to 3.0 µg/mL, R² > 0.99, p< 0.0001, LODmax: 0.12 µg/mL,

LOQmax: 0.36 µg/mL), -carotene (0.05 to 3.0 µg/mL, R² > 0.99, p< 0.001, LOD: 0.03 µg/mL, LOQ:0.1 µg/mL) and chlorophylls (6.25 to 100.0 µg/mL R² > 0.99, p< 0.001,

LODmax: 5.34 µg/mL, LOQmax: 17.81 µg/mL). Analysis was performed in triplicate, and results were expressed in mg/100 g.

2.7 Statistical analysis

All statistical comparisons were based on triplicate results, and data is presented as mean and standard deviation. Data was tested for normality and homogeneity, respectively, by the Kolmogorov-Smirnov’s test and by the test of Cochran, Hartley, and Bartlett. Although all variables presented normally distributed data, some variables did not show homogenous variance/covariance. In order to investigate differences in the chemical composition of PSO, one-way ANOVA with Tukey’s post-test was used for homogenous variables (peroxide value, acid value and contents of: fatty acids, ΣcLnA, ΣUFA, ΣPUFA, ΣMUFA and Σtocopherols) and Kruskal-Wallis used for non-homogeneous variables (p-anisidine value, OSI, lipid classes and contents of β-carotene and Σ phenolics compounds). Data were analyzed in GraphPad Prism v.6.0 (GraphPad software 2012, Inc, La Jolla, CA, US) and Statistica v.7 (Statsoft Inc. 2004, Tulsa, OK, US). Significance was established at p< 0.05.

Multivariate data analysis was used to assess if samples would be grouped by origin based on chemical composition. Firstly, variables to be included in these analyzes matrices were selected according to three criteria: the compounds’ relevance to oil bioactivity, values dispersion and to avoid missing values; in order to allow a number of variables lower than the number of observations. Therefore, the variables selected were: contents of total phenolics, total tocopherols, β-carotene and total cLnA, as they were the main bioactive compounds in PSO, and were present in all samples (zero missing values); which were analyzed by PCA, aiming to address sample grouping and its association with geographical origin of PSO. Data were analyzed with Statistica v.7 (Statsoft, Inc, US).

3. Results and Discussion

Chemical composition of commercial pomegranate seed oils (PSO) from two countries (Turkey and Israel) was determined in the present work, with emphasis on the profiles of fatty acids, major lipid classes, volatile fraction and bioactive compounds (phenolics, tocopherols and carotenoids), as well as quality indexes and oxidative stability. Data of PSO chemical composition allowed for the discrimination of samples’ geographic origin by applying principal component analysis (PCA). Our data show that PSO is a functional oil, owing to its high contents of punicic acid and -tocopherol.

3.1 Quality indexes and oxidative stability

All samples showed peroxide and acid values according to the current codes of good practice (CODEX STAN 19-1981), except for PSOTurkey_batch_1that showed higher acid value than the acceptable maximum level (Table 1). The high free fatty acid content in

PSOTurkey_batch_1 could be related to a greater lipase activity in the seeds (Khoddami et al., 2014). In contrast to previous works that showed p-anisidine values between 0.56 and 6.23 (Khoddami et al., 2014; Melo et al., 2016), high p-anisidine values were observed in all PSOs. In the present work, all samples showed p-anisidine values higher than the recommended upper limit (10.0) (F. A. M. Silva et al., 1999), and 3-to-4-fold higher than these previous reports. The high p-anisidine values observed in the present study are not easily explained by oil degradation, because all samples showed appropriate quality indexes values (CODEX STAN 19-1981). All commercial PSOs showed refractive index values (1.5091-1.5177) expected for highly unsaturated vegetable oils, such as safflower (1.467-1.470), soybean (1.466-1.470) and sunflower (1.461-1.468) (CODEX-STAN 210-1999).

PSOIsrael_A (0.22 h) showed higher OSI than PSOTurkey_batch_2 (0.10 h), while

PSOTurkey_batch_1 (0.15 h) and PSOIsrael_B_batches_1/2 (mean: 0.13 h) presented intermediate values, and undifferentiated from the former two (p> 0.05). This difference might be associated with each oil’s distinct chemical composition and especially concerning the contents of oxidizable substrates and antioxidants. In previous studies, oxidative stability of PSO by Rancimat® showed longer values, but also showed large between-sample variations, from 3.03 h (Basiri, 2015) to 0.72 h (Melo et al., 2016). We suggest that PSO’s OSI was primarily determined by its high contents of the rapidly oxidized cLnAs (Yanget al., 2009), despite the oils’ high contents of antioxidant compounds. 98

Table 1. Quality indices (peroxide value, acid value, p-anisidine value, and refractive index) and oxidative stability index (OSI; by Rancimat®) of commercial cold-pressed pomegranate seed oils (mean + SD; n = 3). Maximum Turkey Turkey Israel Israel Israel acceptable batch 1 batch 2 Brand A Brand B – batch 1 Brand B – batch 2 level* Quality index Peroxide value 2.69 + 0.26a 2.23 + 0.18ab 1.75 + 0.10b 1.19 + 0.10bc 0.91 + 0.10c 15 (mEq O2/kg oil) Acid value 4.38 + 0.03a 2.91 + 0.04b 2.32 + 0.06c 1.80 + 0.08d 2.03 + 0.5e 4 (as % punicic acid) p-anisidine value 13.8+ 0.92a 15.0 + 0.24a 18.6 + 0.23a 13.9 + 0.27a 13.8 + 0.08a - Refractive index 1.5097 + 0.00 1.5091 + 0.00 1.5177 + 0.00 1.5129 + 0.00 1.5135 + 0.00 - Oxidative stability Oil stability index, 0.15 + 0.01ab 0.10 + 0.00a 0.22 + 0.01b 0.12 + 0.00ab 0.14 + 0.01ab - OSI (h)

Different lower-case letters within the same row indicate significant differences: peroxide value, acid value and TEAC (p< 0.05; ANOVA) or p- anisidine value and OSI (p< 0.05; Kruskal-Wallis).

* Reference: CODEX STAN 19-1981. 99

3.2 Chemical composition

3.2.1 Fatty acids and lipid classes

All PSO samples showed similar fatty acid profiles, composed mainly (approximately 95%) by unsaturated fatty acids (Table 2). The high content of unsaturated fatty acids in the PSO has been previously reported by Fernandes et al., (2015) and Khoddami et al. (2014) that observed similar values of UFA varying from 92.5 to 95.2 %. These results suggest that PSO is highly recommended for human consumption when compared to other vegetable oils, because of its low contents of pro-atherogenic fatty acids such as C14:0 and C16:0, and high contents of health-beneficial fatty acids (Table 2). As expected, PSO is a rich source of the bioactive cLnA isomers, representing approximately 85% of the oils’ fatty acids. GC analysis, both with FID and MS detectors, showed eight cLnA isomers from which four (punicic, α- eleostearic, catalpic and β-eleostearic acids) were identified based on comparisons with standards’ retention times and mass spectrometric fragmentation profiles. Punic acid was the major cLnA (ca. 591 mg/g oil) isomer in all PSOs, and its content was not influenced by the origin or production batches of the oil (p> 0.05). However, the total content of cLnA varied between samples (p< 0.05), PSOIsrael_A and PSOIsrael_B_batch_2 showed the highest contents of total cLnA, and PSOTurkey the lowest. Besides that, this major cLnA was not different between production batches (p> 0.05) in the PSOTurkey and PSOIsrael_B. The results of fatty acid profile presented here are in accordance with Khoddami et al., (2014) and Melo et al., (2016) which evaluated the fatty acid composition of commercial pomegranate seed oils produced in Turkey, Iran and the USA. Our results are also in agreement with Elfalleh et al., (2011) that evaluated fatty acid composition of seeds from 21 pomegranate cultivars from Tunisia and China. They found that the fatty acids contents and profile could be affected by sample origin, similarly with total cLnA content variation in the present study. Therefore, besides PSO minor components, major ones such as fatty acids can also be affected by the pomegranate variety, climate conditions in which the plants were grown, and ripening stage at harvest.

All samples showed similar lipid classes composition (Table 2), being triacylglycerols the major lipid class. This finding is in accordance with a previous work (Kaufman & Wiesman, 2007), which showed that punicic acid in PSO is mostly in the form of triacylglycerols. In the present study, it was not possible to analyze punicic acid regio- distribution in TAG molecules (separation of 2-MAG was impractical after selective 100

hydrolysis by porcine pancreatic lipase, either by TLC or SPE; data not shown). However, this issue would be of interest in future investigations, because this localization might affect fatty acids bioavailability (Chardigny et al., 2003).

101

Table 2. Fatty acid (mg/g oil; g/100 g of fatty acids)and major lipid classes (Area%) composition of commercial cold-pressed pomegranate seed oils (mean + SD; n = 3).

Turkey Turkey Israel Israel Israel Batch 1 Batch 2 Brand A Brand B – Batch 1 Brand B – Batch 2 Saturated fatty acids(mg/g oil) 16:0 19.9 + 0.56a 18.0 + 0.56b 16.7 + 0.60b 17.4 + 0.06b 17.2 + 0.54b 18:0 16.4 + 0.75a 15.0 + 0.39bc 14.5 + 0.28c 15.4 + 0.23abc 15.7 + 0.43ab 20:0 5.36 + 0.48a 4.51 + 0.11b 4.3 + 0.10b 4.85 + 0.06ab 4.93 + 0.23ab Unsaturated fatty acids (mg/g oil) 18:1 n-9 40.0 + 1.56a 35.5 + 0.52b 31.6 + 0.41c 33.7 + 0.32b 34.3 + 0.88b 18:1 n-7 3.62 + 0.15a 3.49 + 0.10a 3.60 + 0.09a 3.41 + 0.03a 3.49 + 0.09a 18:2 n-6 46.7 + 1.72a 37.2 + 0.50b 32.7 + 0.40c 37.2 + 0.96b 37.3 + 0.94b 20:1 n-9 6.95 + 0.87a 6.56 + 0.25a 6.77 + 0.28a 6.88 + 0.38a 7.19 + 0.56a Conjugated linolenic acid isomers (mg/g oil) Punicic 606.2 + 54.1a 569.8 + 14.8a 626.9 + 24.6a 552.4 + 30.5a 602.5 + 24.4a α-Eleostearic 60.4 + 5.64ac 58.5 + 2.11a 47.8 + 2.50b 68.7 + 3.77c 67.0 + 2.32ac Catalpic 47.9 + 3.59ab 46.2 + 2.28a 30.8 + 0.60c 58.5 + 3.98b 54.0 + 2.52b β-Eleostearic 14.1 + 1.70a 15.1 + 1.11a 7.44 + 0.47b 16.3 + 0.97a 15.4 + 0.96a cLnA 1 6.42 + 1.52a 6.41 + 0.63a 6.53 + 0.67a 6.66 + 0.25a 5.62 + 0.56a cLnA 2 3.97 + 0.32a 3.24 + 0.27a 5.80 + 0.58b 4.15 + 0.40a 3.67 + 0.07a cLnA 3 4.41 + 0.35a 3.74 + 0.49a 5.71 + 0.33b 4.11 + 0.23a 3.23 + 0.23a cLnA 4 3.24 + 0.33a 2.82 + 0.07a 4.27 + 0.26b 3.42 + 0.29ab 3.24 + 0.50a Fatty acids groups (g/100 g fatty acids) cLnA 84.3 + 0.62a 85.4 + 0.27ab 87.0 + 0.32c 85.7 + 0.67b 86.3 + 0.24bc UFA 95.3 + 0.19a 95.4 + 0.14a 95.8 + 0.16a 95.5 + 0.19a 95.7 + 0.11a 102

PUFA 89.6 + 0.37a 89.9 + 0.20a 90.8 + 0.24b 90.2 + 0.36ab 90.5 + 0.15ab MUFA 5.71 + 0.18a 5.52 + 0.06ab 4.97 + 0.08c 5.29 + 0.17bd 5.14 + 0.05cd Lipid classes (Area%) Σ FFA + MAG 0.54 + 0.09a 0.52 + 0.06a 0.35 + 0.01a 0.25 + 0.05a 0.34 + 0.00a + DAG Σ TAG 99.5 + 0.09a 99.5 + 0.06a 99.6 + 0.01a 99.8 + 0.05a 99.7 + 0.00a cLnA: conjugated linolenic acid; DAG: diacylglycerols; FFA: free fatty acids; MAG: monoacylglycerols; MUFA: monounsaturated fatty acids; PUFA: polyunsaturated fatty acids; TAG: triacylglycerols; UFA: unsaturated fatty acids; .Different lower-case letters within the same row indicate significant differences in: fatty acid composition (p< 0.05; ANOVA) or lipid classes (p< 0.05; Kruskal-Wallis). 103

3.2.2 Volatile compounds

The volatile fraction of PSO was analyzed by SPME-GC-MS, which enabled identification of 24 compounds, from which 19 were quantified by SPME-GC-FID (Table 3). The PSOs showed varied volatile compounds belonging to the following six chemical classes: alcohols, aldehydes, ketones, carboxylic acids, esters and hydrocarbons. Interestingly, there was not a clear typical profile of volatile compounds in the PSOs analyzed; and each sample showed a particular composition, chiefly from the following chemical classes: alcohols and esters in PSOTurkey; alcohols, aldehydes and hydrocarbons in PSOIsrael_A; and alcohols, aldehydes, ketones and carboxylic acids in PSOIsrael_B. In the present study we were interested in the qualitative evaluation of the volatile fraction composition, and the contents shown in Table 3 are semi-quantitative data, meaning that this study was exploratory in terms of PSO volatile composition. These data (Table 3) might substantiate future investigations assessing PSO oxidative stability, in which the quantification and monitoring of specific compounds might serve as chemical markers of PSO oxidative degradation. The number of samples in the present study limits us in raising hypotheses concerning the factors determining these differences in aroma compounds. However, it seems that the oils’ origin affected its volatile fraction composition, as indicated by differences between oils from Turkey and Israel. This hypothesis deserves investigation, especially because in the case it is confirmed, volatile fraction composition could be used as marker of PSO origin. To the best of our knowledge, this was the first work to use two GC capillary-columns with different polarities and SPME- GC-FID/MS to identify and quantify the volatile compounds in PSO. The following volatile compounds: phenethyl alcohol; 2,4-nonadienal; 2,3-butanediol; 2-hexenal; butanoic acid, ethyl ester; hexanoic acid, ethyl ester; 1-hexanol; hexanal and butanedioic acid, diethyl ester have been identified in previous works in a industrial byproduct of cold pressed PSO, and pomegranate’s juice, fruit extract and liquor (Galego, Jockusch, & Da Silva, 2013; Karaman et al., 2015; Vázquez-Araújo, Koppel, Chambers IV, Adhikari, & Carbonell-Barrachina, 2011). Table 4 shows the volatiles presenting descriptive aroma notes. The odorants in the PSO were mainly alcohols, aldehydes and esters. Their characteristic volatile profiles might give PSO their distinctive flavor, which might influence on consumers’ acceptance, but this hypothesis needs confirmation.

104

Table 3. Volatile compounds in commercial samples of cold-pressed pomegranate seed oil (mean + SD; n = 3).

Israel Israel Linear Turkey Turkey Israel Brand B – Brand B – retention LRI D* batch 1 batch 2 Brand A Compounds per chemical classes SI* Columnp* batch 1 batch 2 index Ref. (%) (LRI) Contents(µg/g) Alcohols 1-Butanol, 3-methyl- 88 733 + 3.6 7361 0.5 + 6.39 + 1.64 1.27 + 0.17 ND ND ND 2,3-Butanediol 93 779 + 3.5 7691 1.3 + 18.8 + 3.26 4.92 + 0.83 1.42 + 0.15 ND ND Phenylethyl Alcohol 96 1116 + 4.4 11171 0.1 + 8.28 + 1.70 2.84 + 0.13 1.27 + 0.13 ND ND 1-Hexanol 78 867 + 3.2 8581 1.1 + NQ ND ND ND ND 1-Pentanol 97 768 + 0.0 7582 1.3 - ND ND ND 11.8 + 1.35 9.36 + 1.17 Aldehydes Benzaldehyde 82 954 + 7.3 9611 0.8 + NQ ND ND ND ND 2,4-Nonadienal 90 1217 + 1.4 12223 0.4 + ND ND 2.6 + 0.4 20.8 + 1.34 13.8 + 5.80 2-Hexenal 88 854 + 0.1 8541 0.0 - ND ND ND 2.8 + 0.31 2.12 + 0.39 2-Heptenal, (Z)- 92 958 + 0.2 9641 0.7 + ND ND ND 0.07 + 0.03 0.55 + 0.16 Hexanal 92 798 + 0.2 7971 0.1 - ND ND ND ND 3.31 + 0.57 Ketones 5-Nonanone 87 1074 + 0.2 10743 0.0 - ND ND ND NQ NQ 5-Decanone 84 1173 + 0.0 11553 1.6 - ND ND ND 2.36 + 0.17 1.74 + 0.61 Carboxylic acids Pentanoic acid 94 895 + 0.1 9111 1.7 - ND ND ND 65.3 + 4.49 49.0 + 11.7 Hexanoic acid 87 967 + 12.5 9683 0.1 - ND ND ND 6.20 + 0.51 4.34 +1.29 Esters Butanoic acid, ethyl ester 96 799 + 3.5 8001 0.1 - 9.20 + 2.57 ND ND ND ND Propanoic acid, 2-hydroxy-, 94 811 + 3.5 8033 1.1 + 4.56 + 1.26 ND ND ND ND ethyl ester, (S)- 105

Acetic acid, pentyl ester 95 843 + 41.6 8664 2.6 - NQ 2.13 + 0.32 ND ND ND 1-Butanol, 82 875 + 3.5 8802 0.5 - 5.97 + 1.68 ND ND ND ND 2-methyl-, acetate Hexanoic acid, ethyl ester 81 995 + 4.0 9882 0.7 - 2.23 + 0.62 ND ND ND ND Acetic acid, hexyl ester 89 1009 + 3.6 10123 0.3 + 2.12 + 0.58 ND ND ND ND Butanedioic acid, diethyl 86 1175 + 4.4 11854 0.9 + 1.73 + 0.35 ND ND ND ND ester Acetic acid, 2-phenylethyl 80 1259 + 4.1 12704 0.8 + NQ ND ND ND ND ester Hydrocarbons Butane, 2-methyl- 89 485 + 4.5 4763 1.9 + NQ ND ND ND ND Cyclohexene, 1-methyl-4-(1- 90 1035 + 0.3 10311 0.4 - ND ND 4.11 + 0.61 ND ND methylethenyl)-, (S)- * SI: Similarity index (from MS analysis); LRI: Linear retention index; LRI Ref: Reference linear retention index (References:1Ashraf M. El- Sayed (2014). Kovats index. In: The Pherobase: database of insect pheromones and semiochemicals. Retrieved November 10, 2016 from: http://www.pherobase.com/kovats/; 2U.S. Secretary of Commerce (2017). Kovats’ RI. In: NIST Standard Reference Data. Retrieved November 10, 2016 from: http://webbook.nist.gov; 3National Center for Biotechnology Information. Kovats Retention Index. In: PubChem Substance Database. Retrieved November 10, 2016 from: https://pubchem.ncbi.nlm.nih.gov; 4Royal Society of Chemistry (2015). Retention Index (Kovats). In: Chemical Structure. Retrieved November 10, 2016 from: http://www.chemspider.com); D(%): difference (%) between experimental LRI and published LRI; Columnp: compound identity confirmed in a polar column: +, indicates confirmed identity; -, indicates non-confirmed identity; ND: not detected; NQ: compound identified by GC-MS analysis, but not quantified by GC-FID. Detection threshold: compounds with contents under 0.02 µg/g were not tentatively identified. 106

Table 4. Odor description of volatile compounds in commercial cold-pressed pomegranate seed oils.

Israel Israel Turkey Turkey Israel Odor Compound Brand B Brand B – batch 1 batch 2 Brand A description* – batch 1 batch 2

Alcohols 1-Butanol, 3- burnt, malt, + + ND ND ND methyl whiskey 2,3-Butanediol + + + ND ND fruit, onion Phenylethyl rose, spice, honey, + + + ND ND alcohol lilac flower, green, 1-Hexanol + ND ND ND ND resin 1-Pentanol ND ND ND + + balsamic, fruit

Aldehydes burnt sugar, Benzaldehyde + ND ND ND ND almond fat, geranium, 2,4-Nonadienal ND ND + + + pungent, watermelon, wax green, leaf, rancid, 2-Hexenal ND ND ND + + apple, fat 2-Heptenal, (Z)- ND ND ND + + Hexanal ND ND ND ND + fat, grass, tallow

Ketones 5-Nonanone ND ND ND + + ⎯ 5-Decanone ND ND ND + ND ⎯

Carboxylic acid Pentanoic acid ND ND ND + ND sweat Hexanoic acid ND ND ND + + ⎯

Esters Butanoic acid, + ND ND ND ND ⎯ 107

ethyl ester Propanoic acid, 2- hydroxy-, ethyl + ND ND ND ND ⎯ ester, (S)- Acetic acid, pentyl + + ND ND ND ⎯ ester 1-Butanol, 2- + ND ND ND ND apple methyl-, acetate Hexanoic acid, apple peel, fresh, + ND ND ND ND ethyl ester fruit Acetic acid, hexyl + ND ND ND ND fruit, herb ester Butanedioic acid, + ND ND ND ND fruit, wine diethyl ester Acetic acid, 2- honey, rose, + ND ND ND ND Phenylethyl ester tobacco

Hydrocarbons Butane, 2-methyl- + ND ND ND ND ⎯ Cyclohexene, 1- methyl-4-(1- ND ND + ND ND ⎯ methylethenyl), (S)-

Other 2(3H)-Furanone, ⎯ ND ND ND + ND dihydro-5-methyl-

ND: not detected. *Reference: Terry Acree & Heinrich Arn (2004). Odors. In: Flavornet (Datu, Inc.). Retrieved November 10, 2016 from:http://www.flavornet.org/flavornet.html.

108

3.3 Bioactive compounds

Anthocyanins and chlorophylls were not observed in the PSO samples. PSO showed a very distinct phenolic profile composed by three major classes, phenolic acids, flavonols and flavanones (Table 5). PSOIsrael_A presented the largest variety of phenolic compounds, and vanillic acid was the only compound observed in all oils. PSOTurkey_batch_1 showed higher contents (p< 0.05) of phenolic compounds than PSOIsrael_B_batch_2, while the other PSO samples displayed intermediate values. Phenolic compounds profile varied between PSOTurkey_batch_1 and PSOTurkey_batch_2. Concerning PSOIsrael_B samples, it showed similar phenolics compounds’ content and profile. Variations in phenolic compounds profile and contents between brands and production batches of commercial PSO are expected, because these substances depend on several factors, either genetic or environmental (Oraket al., 2012). Differently from other pomegranate’s parts or products, such as peels, mesocarp, arils, seeds, leaves, and its juice (Fischer et al., 2011; Pande & Akoh, 2009), the PSO phenolic profile was scarcely determined. Mostly, total phenolics contents in PSO were determined by a low selectivity spectrophotometric assay (Abbasi, Rezaei, Emamdjomeh, et al., 2008; Khoddami et al., 2014). Therefore, comparisons with other studies should be made with caution because different methods were used or distinct pomegranate parts were assessed. However, Pande & Akoh (2009) determined the phenolic profile of pomegranate’s seeds by HPLC-PDA and also found ferulic and p-coumaric acids, and quercetin as major phenolic compounds. Based in the potential bioactivity of these substances and the possible use of PSO as a functional food, the phenolic compounds profile displayed in the present work might stimulate future studies on the bioavailability and bioactivity of phenolic compounds from this oil. Nevertheless, the HPLC-PDA analysis gives a preliminary assessment of the phenolic profile in PSO, and for a more thorough analysis of PSO phenolics profile LC coupled with a high-resolution mass spectrometry detector should be applied.

Three tocopherols (α, γ and δ) were identified in PSO, being γ-tocopherol the major isoform in all samples (Table 5). The PSOTurkey_batches_1/2 showed the highest content (p< 0.05) of tocopherols followed by PSOIsrael_A and PSOIsrael_B_batches_1/2. Tocopherols composition was not influenced by production batches in the present work, both in PSOTurkey and PSOIsrael_B. In previous studies (Caligiani et al., 2010; Fernandes et al., 2015; G. Liu et al., 2012; Melo et al., 2016) total tocopherols’ contents varied over 30-fold, ranging from 135 to 4561 mg/100g oil, in contrast to the present work with 3-fold variation at most, although the values found were within the expected range. Regarding the distribution of isoforms, γ-tocopherol is consistently 109

reported as the tocol with the highest contents in PSO (Fernandes et al., 2015; Liu et al.,2012; Melo et al., 2016), however δ- and β-tocopherols have already been identified as major tocols (Caligiani et al., 2010). PSOs in the present work showed much higher contents of tocopherols than other highly consumed vegetable oils, such as those of corn (104 mg/100g), sunflower (61 mg/100g), olive (17.8 mg/100g) and flaxseed (53.4 mg/100g) (Akil et al., 2015; Carpenter, 1979; Melo et al., 2016).

β-Carotene was present in small amounts in commercial PSOs, and other carotenoids were undetectable (Table 5). Low levels of carotenoids in PSO were also found by Fernandes et al. (2015), that did not find any traces of carotenoids in the seeds of nine pomegranate cultivars grown in Spain.

110

Table 5. Phenolic compounds, tocopherols and carotenoids composition (mg/100 g oil) of commercial cold-pressed pomegranate seed oils (mean + SD; n = 3).

Contents (mg/100 g) Bioactive Turkey Turkey Israel Israel Israel compounds batch 1 batch 2 Brand A Brand B – Brand B – batch batch 1 2 Phenolic compounds Phenolic Acids 3,4-Dihidroxy- 1.82 + 0.15 0.17 + 0.01 0.52 + 0.01 ND ND phenylacetic p-Hidroxy-benzoic 0.89 + 0.04 ND 0.05 + 0.00 ND ND Vanillic 2.56 + 0.29 0.32 + 0.04 0.29 + 0.01 0.16 + 0.00 0.14 + 0.00 Syringic 0.88 + 0.05 0.15 + 0.00 0.19 + 0.01 ND ND Ferulic 0.56 + 0.06 0.02 + 0.00 0.03 + 0.00 ND ND m-Coumaric 9.43 + 0.58 0.72 + 0.12 tr ND ND p-Coumaric ND ND 0.02 + 0.00 ND ND trans-Cinnamic tr ND tr tr tr Rosmarinic ND ND 2.20 + 0.3 ND ND Flavonol Quercetin 1.86 + 0.08 ND ND 0.18 + 0.02 0.13 + 0.01 Flavanone Naringenin ND ND 1.77 + 0.04 ND ND ∑ Phenolic 16.8 + 0.47a 1.38 + 0.15ab 4.92 + 0.04ab 0.34 + 0.02ab 0.28 + 0.01b compounds Tocopherols α 4.88 + 0.05 3.92 + 0.04 2.83 + 0.57 2.40 + 0.08 2.38 + 0.01  1125.8 + 58.8 1161.8 + 40.4 754.9 + 70.7 517.2 + 7.15 538.2 + 13.4  7.23 + 0.19 6.53 + 0.15 5.23 + 0.50 4.56 + 0.09 4.60 + 0.02 ∑ Tocopherols 1138.0 + 59.1a 1172.3 + 40.5a 762.6 + 71.3b 524.1 + 7.23c 545.2 + 13.4c Carotenoids β-carotene 1.49 + 0.03a 0.95 + 0.02b 1.11 + 0.12ab 1.02 + 0.01ab 1.05 + 0.01ab

ND: not detected; tr: traces. Different lower-case letters within the same row indicate significant differences (p< 0.05; ANOVA; Σ tocopherols content) or (p< 0.05; Kruskal-Wallis; Σ phenolics and carotenoids contents). Limits of detection (LOD): 3,4-dihidroxy-phenylacetic = 0.002 mg/100 g; p-hidroxy-benzoic = 0.006 mg/100 g; syringic = 0.02 mg/100 g; ferulic = 0.008 mg/100 g; m- coumaric = 0.06 mg/100 g; p-coumaric: 0.001 mg/100 g; trans-cinnamic = 0.13 mg/100 g; rosmarinic = 0.008 mg/100 g; quercetin = 0.04 mg/100 g and naringenin = 0.10 mg/100 g. Limits of quantification (LOQ): m-coumaric = 0.1 mg/100 g and trans-cinnamic = 0.39 mg/100 g.

111

3.4 PSO samples grouping: Multivariate data analysis

Multivariate data analysis was used to achieve data reduction, namely identifying the major variables contributing to data variation in the PSOs’ data matrix by Principal Component Analysis (PCA). PCA allowed the selection of the bioactive compounds: total contents of phenolics, total tocopherols, β-carotene and total cLnA, that mostly contributed to samples’ variation. After data reduction, PCA was run and resulted in the first two principal components (PC) in the data matrix, which explained 91.31% of total variance. All bioactive compounds (total contents of phenolics (0.95), total tocopherols (0.74), β-carotene (0.89) and total cLnA (0.84)) were associated to PC 1, which explained 73.87% of total variance, whereas PC 2 explained 17.44% (Figure 1). This analysis allowed preliminary discrimination of PSO by geographical origin, revealing one cluster composed by samples from Israel, which clustered apart from the PSO from Turkey. Grouping of the PSO from Israel was due to their distinctive composition of bioactive compounds (phenolics, β-carotene and total cLnA).Unlike PSOIsrael samples, it was not possible of grouping the two batches of PSOTurkey, indicating that these Turkish PSO present a lower homogeneity in bioactives’ composition. Samples’ grouping by geographical origin based on bioactive compounds composition was suggestive for Israel samples, as PSOTurkey_batches_1/2 showed a scattered dispersion pattern among PC1 and PC2 axes (Figure 1).The restricted sample number in the present study limited these analyses, hindering the use of multivariate data analysis as a tool to determine Turkey and Israel PSO identity. However, PCA provided a marginal discrimination of samples’ geographic origin as they showed a distinct bioactive compounds composition. Fatty acid profile in the seeds of Tunisian and Chinese pomegranates was also strongly related to geographical origin (Elfalleh et al., 2011). Genetic differences between pomegranates, such as varying cultivars, and also environmental factors, such as sunlight exposure, temperature, water availability, and agricultural practices, among other factors might explain sample origin discrimination based on chemical composition.

112

Figure 1. Plots of principal component analysis with the following variables: contents of total phenolics, total tocopherols, β-carotene and total cLnA. Principal component 1 (PC 1) and principal component 2 (PC 2) explained 91.3% of the total variance in this data matrix (Adapted from: (Costa et al., 2019)

4. Conclusions

As pomegranate seed oil (PSO) is an emerging functional food, in the present study we determined oil’s chemical composition, which was used to discriminate commercial PSO. In the present work we showed that the contents of bioactive compounds (phenolics, cLnA, tocopherols, and beta-carotene) allowed samples’ discrimination. However, an evident trend in samples geographical origin based on bioactive compounds was observed only in Israel samples, possibly because of potential between-variety variation in chemical composition, hindering the present investigation to develop a PSO origin identification tool. In this sense, it is worth confirming this assessment in future works with increased sample size and from distinct origins. Additionally, the present work provides valuable data concerning potential bioactivity of PSO that might subsidize future studies on the health effects of this oil’s consumption, or on its use as a functional food ingredient. PSO volatile composition seemed promising as a marker of origin and possibly processing, because it varied widely between samples, but this hypothesis needs confirmation in future studies with larger sample sizes. PSO volatile fraction showed marker components previously identified in pomegranate seeds and juice, providing valuable means of determining PSO identity. It would be interesting to 113

determine in future investigations, to what extent PSO aroma compounds arrived from the seeds (oil’s raw material) or from chemical transformations during processing and storage.

Acknowledgements

The authors greatly acknowledge the funding provided by CAPES, CNPq (Grant numbers: 432484/2016-7 and 309558/2015-8) and FAPERJ (Grant numbers: E- 26/010.001277/2015 and E-26/203.197/2015) (Brazil). AGT is a recipient of a CNPq scholarship, AMMC and LOS were recipients of CNPq PhD studentships. We are also indebted to Prof.Verônica M. A. Calado and Vanessa N. Castelo-Branco for their help with multivariate statistical analyses, to Prof. Tatiana El-Bacha for kindly providing us the cLnA standards, and to Leticia Korin Moretti for her help with laboratory analysis of oils’ quality indices.

5. References

Abbasi, H., Rezaei, K., Emamdjomeh, Z., Ebrahimzadeh Mousavi, S.M., 2008. Effect of various extraction conditions on the phenolic contents of pomegranate seed oil. Eur. J. Lipid Sci. Technol. 110, 435–440. doi:10.1002/ejlt.200700199 Akil, E., Castelo-Branco, V.N., Costa, A.M.M., do Amaral Vendramini, A.L., Calado, V., Torres, A.G., 2015. Oxidative Stability and Changes in Chemical Composition of Extra Virgin Olive Oils After Short-Term Deep-Frying of French Fries. J. Am. Oil Chem. Soc. 92, 409–421. doi:10.1007/s11746-015-2599-2 American Oil Chemists’ Society, 2012. Official methods and recommended practices of the AOCS, 6th ed. Chicago. Basiri, S., 2015. Evaluation of antioxidant and antiradical properties of Pomegranate (Punica granatum L.) seed and defatted seed extracts. J. Food Sci. Technol. 52, 1117–1123. doi:10.1007/s13197-013-1102-z Brown, P.N., Shipley, P.R., 2011. Determination of anthocyanins in cranberry fruit and cranberry fruit products by high-performance liquid chromatography with ultraviolet detection: Single-laboratory validation. J. AOAC Int. 94, 459–466. Caligiani, A., Bonzanini, F., Palla, G., Cirlini, M., Bruni, R., 2010. Characterization of a Potential Nutraceutical Ingredient: Pomegranate (Punica granatum L.) Seed Oil 114

Unsaponifiable Fraction. Plant Foods Hum. Nutr. 65, 277–283. doi:10.1007/s11130-010- 0173-5 Cao, Y., Gao, H.L., Chen, J.N., Chen, Z.Y., Yang, L., 2006. Identification and characterization of conjugated linolenic acid isomers by Ag+-HPLC and NMR. J. Agric. Food Chem. 54, 9004–9009. doi:10.1021/jf0616199 Cao, Y., Yang, L., Gao, H.L., Chen, J.N., Chen, Z.Y., Ren, Q.S., 2007. Re-characterization of three conjugated linolenic acid isomers by GC-MS and NMR. Chem. Phys. Lipids 145, 128–133. doi:10.1016/j.chemphyslip.2006.11.005 Carpenter, A.P., 1979. Determination of tocopherols in vegetable oils. J. Am. Oil Chem. Soc. 56, 668–671. doi:10.1007/BF02660070 Chardigny, J.M., Masson, E., Sergiel, J.P., Darbois, M., Loreau, O., Noël, J.P., Sébédio, J.L., 2003. The position of rumenic acid on triacylglycerols alters its bioavailability in rats. J Nutr 133, 4212–4214. Christie, W.W., 2003. Lipid analysis: Isolation, Separation, Identification and Structural Analysis of Lipids, 3rd ed. The Oily Press/ PJ Barnes and Associates, Bridgwater, England. Codex Alimentarius (1981). Standard for edible fats and oils not covered by individual standards (CODEX STAN 19-1981). Retrived March 22, 2017 from: http://www.fao.org/input/download/standards/74/CXS_019e_2015.pdf Codex Alimentarius (1999). Standard for named vegetable oils (CODEX-STAN 210-1999). Retrived March 22, 2017 from: http://www.fao.org/docrep/004/y2774e/y2774e04.htm#bm4.1 Costa, A.M.M., Nunes, J.C., Lima, B.N.B., Pedrosa, C., Calado, V., Torres, A.G., Pierucci, A.P.T.R., 2015. Effective stabilization of CLA by microencapsulation in pea protein. Food Chem. 168, 157–166. doi:10.1016/j.foodchem.2014.07.016 De La Torre-Carbot, K., Jauregui, O., Gimeno, E., Castellote, A.I., Lamuela-Raventós, R.M., López-Sabater, M.C., 2005. Characterization and quantification of phenolic compounds in olive oils by solid-phase extraction, HPLC-DAD, and HPLC-MS/MS. J. Agric. Food Chem. 53, 4331–4340. doi:10.1021/jf0501948 Dhar, P., Ghosh, S., Bhattacharyya, D.K., 1999. Dietary effects of conjugated octadecatrienoic fatty acid (9 cis, 11 trans, 13 trans) levels on blood lipids and nonenzymatic in vitro lipid peroxidation in rats. Lipids 34, 109–14. doi:10.1007/s11745- 999-0343-2 Elfalleh, W., Ying, M., Nasri, N., Sheng-Hua, H., Guasmi, F., Ferchichi, A., 2011. Fatty acids 115

from Tunisian and Chinese pomegranate (Punica granatum L.) seeds. Int. J. Food Sci. Nutr. 62, 200–206. doi:10.3109/09637486.2010.526932 Fernandes, L., Pereira, J.A., Lopéz-Cortés, I., Salazar, D.M., Ramalhosa, E., Casal, S., 2015. Fatty acid, vitamin E and sterols composition of seed oils from nine different pomegranate (Punica granatum L.) cultivars grown in Spain. J. Food Compos. Anal. 39, 13–22. doi:10.1016/j.jfca.2014.11.006 Fischer, U.A., Carle, R., Kammerer, D.R., 2011. Identification and quantification of phenolic compounds from pomegranate (Punica granatum L.) peel, mesocarp, aril and differently produced juices by HPLC-DAD–ESI/MSn. Food Chem. 127, 807–821. doi:10.1016/j.foodchem.2010.12.156 Franke, A.A., Murphy, S.P., Lacey, R., Custer, L.J., 2007. Tocopherol and Tocotrienol Levels of Foods Consumed in Hawaii. J. Agric. Food Chem. 55, 769–778. Galego, L.R., Jockusch, S., Da Silva, J.P., 2013. Polyphenol and volatile profiles of pomegranate (Punica granatum L.) fruit extracts and liquors. Int. J. Food Sci. Technol. 48, 693–700. doi:10.1111/ijfs.12014 Gimeno, E., Castellote, A.I., Lamuela-Raventós, R.M., de la Torre, M.C., López-Sabater, M.C., 2000. Rapid determination of vitamin E in vegetable oils by reversed-phase high- performance liquid chromatography. J. Chromatogr. A 881, 251–254. doi:10.1016/S0021-9673(00)00219-3 Greco-Duarte, J., Cavalcanti-Oliveira, E.D., Da Silva, J.A.C., Fernandez-Lafuente, R., Freire, D.M.G., 2017. Two-step enzymatic production of environmentally friendly biolubricants using castor oil: Enzyme selection and product characterization. doi:10.1016/j.fuel.2017.04.036 Holcapek, M., Lísa, M., Jandera, P., Kabátová, N., 2005. Quantitation of triacylglycerols in plant oils using HPLC with APCI-MS, evaporative light-scattering, and UV detection. J. Sep. Sci. 28, 1315–1333. doi:10.1002/jssc.200500088 Karaman, S., Karasu, S., Tornuk, F., Toker, O.S., Geçgel, Ü., Sagdic, O., Ozcan, N., Gül, O., 2015. Recovery potential of cold press byproducts obtained from the edible oil industry: Physicochemical, bioactive, and antimicrobial properties. J. Agric. Food Chem. 63, 2305–2313. doi:10.1021/jf504390t Kaufman, M., Wiesman, Z., 2007. Pomegranate oil analysis with emphasis on MALDI- TOF/MS triacylglycerol fingerprinting. J. Agric. Food Chem. 55, 10405–10413. doi:10.1021/jf072741q Khoddami, A., Man, Y.B.C., Roberts, T.H., 2014. Physico-chemical properties and fatty acid 116

profile of seed oils from pomegranate (Punica granatum L.) extracted by cold pressing. Eur. J. Lipid Sci. Technol. 116, 553–562. doi:10.1002/ejlt.201300416 Kramer, J.G., Fellner, V., Dugan, M.R., Sauer, F., Mossoba, M., Yurawecz, M., 1997. Evaluating acid and base catalysts in the methylation of milk and rumen fatty acids with special emphasis on conjugated dienes and total trans fatty acids. Lipids 32, 1219–1228. doi:10.1007/s11745-997-0156-3 Larick, D.K., Parker, J.D., 2001. Chromatographic Analysis of Secondary Lipid Oxidation Products, in: Wrolstad, R.E., Acree, T.E., Decker, E.A., Penner, M.H., Reid, D.S., Schwartz, S.J., Shoemaker, C.F., Smith, D.M., Sporns, P. (Eds.), Curr. Protoc. Food Analyt. Chem. John Wiley & Sons, Inc., pp. 1–9.

Liu, G., Xu, X., Gong, Y., He, L., Gao, Y., 2012. Effects of supercritical CO2 extraction parameters on chemical composition and free radical-scavenging activity of pomegranate (Punica granatum L.) seed oil. Food Bioprod. Process. 90, 573–578. doi:10.1016/j.fbp.2011.11.004 Melo, I.L.P. de, Carvalho, E.B.T. de, Silva, A.M. de O. e, Yoshime, L.T., Sattler, J.A.G., Pavan, R.T., Mancini-Filho, J., 2016. Characterization of constituents, quality and stability of pomegranate seed oil (Punica granatum L.). Food Sci. Technol. 36, 132–139. doi:10.1590/1678-457X.0069 Minguez-Mosquera, M.I., Garrido-Fernandez, J., Gandul-Rojas, B., Garrido-fernindez, J., 1989. Pigment changes in olives during fermentation and brine storage. J. Agric. Food Chem. 37, 8–11. doi:10.1021/jf00085a002 Nunes, J.C., Torres, A.G., 2010. Fatty acid and CLA composition of Brazilian dairy products, and contribution to daily intake of CLA. J. Food Compos. Anal. 23, 782–789. doi:10.1016/j.jfca.2010.03.023 Orak, H.H., Yagar, H., Isbilir, S.S., 2012. Comparison of antioxidant activities of juice, peel, and seed of pomegranate (Punica granatum L.) and inter-relationships with total phenolic, Tannin, anthocyanin, and flavonoid contents. Food Sci. Biotechnol. 21, 373– 387. doi:10.1007/s10068-012-0049-6 Pande, G., Akoh, C.C., 2009. Antioxidant capacity and lipid characterization of six georgia- grown pomegranate cultivars. J. Agric. Food Chem. 57, 9427–9436. doi:10.1021/jf901880p Rahmani, M., Csallany, A.S., 1991. Chlorophyll and β-carotene pigments in moroccan virgin olive oils measured by high-performance liquid chromatography. J. Am. Oil Chem. Soc. 68, 672–674. doi:10.1007/BF02662293 117

Rodriguez-Amaya, D.B., Kimura, M., 2004. HarvestPlus Hanbook for Carotenoid Analysis. International Food Policy Research Institute and International Center for Tropical Agriculture, Washinton, DC, US. Siger, A., Nogala-Kalucka, M., Lampart-Szczapa, E., 2007. Compounds in Cold-Pressed Plant Oils. J. Food Lipids 15, 137–149. doi:10.1111/j.1745-4522.2007.00107.x Silva, F.A.M., Borges, M.F.M., Ferreira, M.A., 1999. Métodos para avaliação do grau de oxidação lipídica e da capacidade antioxidante. Quim. Nova 22, 94–103. doi:10.1590/S0100-40421999000100016 Spitzer, V., 1996. Structure analysis of fatty acids by gas chromatography-low resolution electron impact mass spectrometry of their 4,4-dimethyloxazoline derivatives-a review. Prog. Lipid Res. 35, 387–408. Suzuki, R., Noguchi, R., Ota, T., Abe, M., Miyashita, K., Kawada, T., 2001. Cytotoxic effect of conjugated trienoic fatty acids on mouse tumor and human monocytic leukemia cells. Lipids 36, 477–482. doi:10.1007/s11745-001-0746-0 Tan, B., Brzuskiewicz, L., 1989. Separation of tocopherol and tocotrienol isomers using normal- and reverse-phase liquid chromatography. Anal. Biochem. 180, 368–373. doi:10.1016/0003-2697(89)90447-8 Teixeira da Silva, J.A., Rana, T.S., Narzary, D., Verma, N., Meshram, D.T., Ranade, S.A., 2013. Pomegranate Biology And Biotechnology: A Review. Sci. Hortic. (Amsterdam). 160, 85–107. doi:10.1016/j.scienta.2013.05.017 Tuberoso, C.I.G., Kowalczyk, A., Sarritzu, E., Cabras, P., 2007. Determination of antioxidant compounds and antioxidant activity in commercial oilseeds for food use. Food Chem. 103, 1494–1501. doi:10.1016/j.foodchem.2006.08.014 Vázquez-Araújo, L., Koppel, K., Chambers IV, E., Adhikari, K., Carbonell-Barrachina, a. a., 2011. Instrumental and sensory aroma profile of pomegranate juices from the USA: Differences between fresh and commercial juice. Flavour Fragr. J. 26, 129–138. doi:10.1002/ffj.2035 Viegas, M.C., Bassoli, D.G., 2007. Utilização do índice de retenção linear para caracterização de compostos voláteis em café solúvel utilizando GC-MS e coluna HP-Innowax. Quim. Nova. 30, 2031-2034 doi:10.1590/S0100-40422007000800040 Viladomiu, M., Hontecillas, R., Lu, P., Bassaganya-Riera, J., 2013. Preventive and prophylactic mechanisms of action of pomegranate bioactive constituents. Evidence- based Complement. Altern. Med. 2013, doi: 10.1155/2013/789764. doi:10.1155/2013/789764 118

Yang, L., Cao, Y., Chen, J.N., Chen, Z.Y., 2009. Oxidative stability of conjugated linolenic acids. J. Agric. Food Chem. 57, 4212–4217. doi:10.1021/jf900657f Yentsch, C.S., Menzel, D.W., 1963. A method for the determination of phytoplankton chlorophyll and phaeophytin by fluorescence. Deep. Res. Oceanogr. Abstr. 10, 221–231. doi:10.1016/0011-7471(63)90358-9

119

Chapter 3

Microencapsulation of pomegranate (Punica granatum L.) seed oil by complex coacervation: high bioactive compounds retention

Authors: Costa, André M. M.; Simões, Grazieli; Silva, Kelly A; Calado, Verônica; Tonon, Renata V.; Torres, Alexandre G.

120

Abstract

The objective of this study was to evaluate the influence of total polymer concentration (Cp) and wall material:oil ratio (WM:Oil) on the microencapsulation of pomegranate (Punica granatum) seed oil (PSO) by complex coacervation (CC), using whey protein-gum Arabic as wall materials and spray drying as hardening/recuperation step. Eleven formulations were tested according to a rotatable central composite design. Independent variables were: Cp (2.2-7.8%) and WM:Oil ratio (0.5-5.0). Oil retention, microencapsulation efficiency (ME), punicic acid content, peroxide value, moisture content, water activity and particle size were analyzed as responses. Microparticles’ morphology and surface composition were also evaluated. Both Cp and WM:Oil ratio affected oil retention, ME, punicic acid content, moisture content and particle size. Higher Cp and lower WM:Oil ratios promoted microparticles’ agglomeration. Surface composition analyses were in accordance with ME. Microparticles with high oil yield and punicic acid content were successfully produced by CC.

Keywords

Functional food, pomegranate seed oil, encapsulation, complex coacervation, spray drying, whey protein, gum Arabic and punicic acid. 121

1. Introduction

Pomegranate (Punica granatum) is a fruit appreciated for its pulp flavor and color that industrially is mainly destined for juice, nectar, jam and jellies production (Shabbir et al., 2017). The seeds are one of the main pomegranate’s residues (3.7 to 7.9% in fruit weight, w/w), consisting of a rich source of lipids (12 to 20%, w/w) with an interesting chemical composition in terms of bioactives compounds (Fernandes et al., 2015), but that is largely discarded. Pomegranate seed oil (PSO) is known for its high contents of tocopherols, phytosterols and phenolic compounds. However, the described functional activities, such as cytotoxic effect, modulation of the immune system and anti-diabetic properties, seem to be associated with PSO’s singular fatty acid profile, composed chiefly by conjugated linolenic acid isomers (cLnA), specially punicic acid, which corresponds to approximately 70% of total fatty acids in the oil (Elfalleh et al., 2011; Fernandes et al., 2015; Shabbir et al., 2017).

cLnA is a collective term for the positional and geometric isomers of linolenic acid (C18:3), characterized by the presence of three conjugated double bounds, usually in positions Δ9,11,13, and Δ8,10,12 and with varying combinations of geometrical configurations, cis or trans (Cao et al., 2006). Although cLnA isomers are emerging as a potential bioactive nutrient, studies in humans evaluating the health benefits of its acute or chronic consumption are still scarce, because these compounds, especially punicic acid, are restricted to PSO and Trichosanthes kirilowii seed oil (Shabbir et al., 2017). Thus the development of food products using PSO as a functional ingredient could enable the design of future studies on the health effects of its consumption. However, direct addition of PSO aiming to supplement food products is limited by its hydrophobic nature and by the high susceptibility of cLnA isomers to lipid oxidation when exposed to oxygen and light (Yang et al., 2009). Thus, these functional lipids should be protected in order to preserve their physical and chemical stability, avoiding oxidative rancidity and nutritional losses.

the best of our knowledge, microencapsulation of PSO by complex coacervation has not been studied yet. Complex coacervation is a physical-chemical method that has been successfully used to encapsulate hydrophobic materials (Eratte et al., 2015; Weinbreckt et al., 2004). It consists of a liquid-liquid phase separation phenomenon that occurs when electrostatically opposite charged biopolymers, usually a protein and a polysaccharide, are subjected to specific conditions, producing aggregates (coacervates) that promptly deposit on the oil droplets (Weinbreckt & Minorf, 2004). Compared to other encapsulation techniques, complex coacervation is able to produce microparticles with higher microencapsulation efficiency utilizing high core load and low wall material concentration (Gouin, 2004).

The process performance and the physicochemical properties of particles produced by complex coacervation can be influenced by parameters such as: total biopolymers concentration, protein:polysaccharide ratio, core:wall material ratio, pH, salt concentration and others (Weinbreckt & Minorf, 2004). For established complex coacervate systems, such as whey protein-gum Arabic and gelatin-gum Arabic, some conditions (protein:polyssacharide ratio, pH and salt concentration) are well known (De Kruif et al., 2004), while other parameters need testing when new core nutrients are used.

Whey protein-gum Arabic is a classic system used for complex coacervation, due to its good emulsifying and film-forming properties, and its suitability for temperature-sensitive compounds, because the coacervation process can be carried out at room temperature (Gharsallaoui et al., 2007; Weinbreckt & Minorf, 2004). Previous studies have already shown promising results when applying this system in the microencapsulation of tuna and sunflower oil (Eratte et al., 2015, 2014; Weinbreckt & Minorf, 2004). In this context, the objective of this study was to propose complex coacervation as a suitable system for microencapsulation of pomegranate seed oil, using whey protein and gum Arabic as wall materials, and to evaluate the influence of total polymer concentration (Cp) (2.2-7.8%) and wall material:oil ratio (WM:Oil ratio) (0.5-5.0) on the particles properties. Oil retention, microencapsulation efficiency, punicic acid content, peroxide value, moisture content, water activity and particle size were analyzed as responses. Microparticles morphology (wet and dried) and surface composition were also evaluated. It is suggested that whey protein-gum Arabic system is a suitable system for the encapsulation of PSO by complex coacervation and microparticles’ physical-chemical properties are highly influenced by total polymer concentration (Cp) and wall material:oil ratio (WM:Oil ratio). 123

2. Material and Methods

2.1 Materials

Cold-pressed commercial pomegranate seed oil (PSO) (C16:0 = 2.25%, C18:0 = 1.86%, C18:1 n-9 = 4.52%, C18:1 n-7 = 0.41%, C18:2 n-6 = 5.29%, C20:0 = 0.61%, C20:1 n- 9 = 0.78%, Total cLnA = 84.3%; Oneva Food Co®; Istanbul- Turkey) was used as core material. Whey protein isolate (WPI) (Alibra®; São Paulo, Brazil) and Gum Arabic (GA) (Instantgum BA®; Colloides Naturels, São Paulo, Brazil) were used as wall materials for particles production. A commercial mixture of fatty acid methyl esters (37-component FAME mix; Supelco, Bellefonte, PA, US) and individual cLnA commercial isomers (punic acid, α- eleostearic acid, catalpic acid and β-eleostearic acid; Larodan AB, Solna, Sweden)were used as standards for fatty acid identification in gas-chromatographic analyzes. All solvents used were HPLC grade from Tedia (São Paulo, Brazil) and all reagents used were from Merck (Darmstadt, Germany).

2.2 Production of the PSO’s microparticles by complex coacervation

A rotatable central composite design was applied to evaluate the effect of total polymer concentration (Cp) (2.2-7.8%) and wall material:oil ratio (WM:Oil ratio) (0.5-5.0) on the microencapsulation of PSO by complex coacervation, according to Table 1.

The following polynomial equation was fitted to data (Equation 1): 2 2 y = β0 + β1x1 + β2x2 + β11x1 + β22x2 + β12x1x2 [Equation 1]

where βn are constant regression coefficients, y is the response (oil retention, microencapsulation efficiency, punicic acid content, peroxide value, moisture content, water activity or particle size), and x1 and x2 are the coded independent variables (Cp and WM:Oil ratio).

The formulations of PSO’s microparticles were produced as follows: firstly, the wall materials (WPI and GA) were separately weighed and diluted in distilled water under magnetic stirring during 30 min, at room temperature, to obtain solutions with Cps described in Table 1. The solutions’ pH was adjusted to 7.0 with HCl (0.5 N) and NaOH (0.5 N), and the ratio of WPI:GA was 2:1 (Weinbreck, Tromp, & de Kruif, 2004). After dissolution, PSO was added dropwise to the WPI solution under continuous stirring (16,000 RPM) during 5 min with an Ultra-Turrax homogenizer (T25-IKA®,IKA,Wilmington, US), to produce a stable 124

emulsion. Then, GA solution was added to the previous emulsion and homogenized for 1 min at 16,000 RPM. Finally, emulsion’s pH was adjusted to 3.75 by adding HCl (1 N) (Eratte et al., 2014) aiming at inducing electrostatic interactions between WPI and GA, thus forming the wet microparticles. Microencapsulation process was carried out at 25 °C, for 10 min.

2.3 Drying of the PSO’s microparticles

Immediately after preparation, the wet microparticles were dried in a laboratory scale spray dryer (SD-06AG, Lab Plant, North Yorkshire, UK) equipped with a 0.5 mm nozzle. The drying conditions were: inlet air temperature of 180 °C, outlet air temperature of 66 ± 4 °C, air flow pressure of 2.5 bar and feed flow rate of 0.5 L/h. Dried microparticles were collected, sealed in plastic bags under vacuum, covered in aluminum foil and stored at -80 °C for further characterization.

2.4 Wet microparticles analysis

2.4.1 Coacervation yield

Coacervation yield was calculated according toEratte, Wang, Dowling, Barrow, & Adhikari (2014), with modifications. Briefly, PSO’s microparticles were prepared as described in section 2.2 and emulsions’ aliquots (10 mL) were withdrawn and transferred to test tubes. Microparticles were allowed to stand overnight to facilitate the precipitation of the gel-like coacervates, and then were carefully separated and dried at 105°C until constant weight. The coacervation yield was calculated according to Equation 2:

푀푎푠푠 표푓 푑푟푖푒푑 푝푎푟푡푖푐푙푒푠 (푔) [Equation 2] 퐶표푎푐푒푟푣푎푡푖표푛 푦푖푒푙푑 = 푀푎푠푠 표푓 푃푆푂 (푔) + 푊푃퐼 (푔) + 퐺퐴 (푔)푖푛 10 푚푙

2.4.2 Morphology

Wet microparticles morphology was evaluated in an optical microscope (Eclipse E200; Nikon®, Tokyo, Japan) coupled to a digital camera (Evolution VF, Media Cybernetics, Rockville, MD, US). Prior to observation in the microscope, microparticles were extracted from the emulsions described in section 2.2 by centrifugation (11,739 g, 10 min) and were diluted twice with Milli-Qwater (pH= 3.7).Microparticles’ images were assessed in 400× and 1000×magnifications. 125

2.5 Dried microparticles analysis

2.5.1 Oil retention

Total oil content present in the microparticles was extracted in a Soxhlet apparatus (SER 148; Velp® Scientifica, Usmate, Italy). Samples (1 g) were extracted using 70 mL petroleum ether with added BHT (0.05%, w/v) during 4 h (Goula & Adamopoulos, 2012).

After solvent evaporation with N2, the resulting oil was weighed, re-suspended with a known volume of hexane with added BHT (0.05%, w/v) and stored at -20 °C until further analyses (punicic acid content and peroxide value).

Oil retention was defined as the percentage of total oil in the final power to that initially added in the feed emulsion and was calculated as follows (Equation 3):

푀푎푠푠 표푓 푡표푡푎푙 표𝑖푙 (푔) 푂푖푙 푟푒푡푒푛푡푖표푛 = ×100 [Equation 3] 푀푎푠푠 표푓 표𝑖푙 𝑖푛 푡ℎ푒 푓푒푒푑 푒푚푢푙푠𝑖표푛 (푔)

2.5.2 Microencapsulation efficiency (ME)

Surface oil was determined according to Sankarikutty, Sreekumar, Narayanan, & Mathew (1988), with modifications proposed by Kouassi et al. (2012). Briefly, 5 mL of hexane was added to the microparticles (200 mg) and gently mixed for 5 min. Then, samples were centrifuged (753 ×g, 5 min) for separation of the supernatant containing the non- encapsulated oil, which was filtered in a filtration column composed of Celite and Na2SO4, and in a Whatman filter n°1 (Marlborought, MA, US). The filtration apparatus was washed twice with n-hexane (5 mL), and the filtrates were collected. Finally, the solvent was left to evaporate and the flasks were dried in an oven (105 °C) until constant weight. The surface oil was calculated based on the difference between the initial clean flask and the flask containing extracted oil residue. ME was calculated according to Equation 4:

푇표푡푎푙 표푖푙 (푔) − 푆푢푟푓푎푐푒 표푖푙 (푔) [Equation 4] 푀퐸 (%) = × 100 푇표푡푎푙 표푖푙 (푔)

2.5.3 Punicic acid content

PSO fatty acids extracted from the microparticles were methylated by a base-catalyzed transesterification procedure (Kramer et al.,1997) to avoid isomerization. PSOs’ fatty acid methyl esters (FAME) were analyzed by gas chromatography with a flame ionization detector 126

(GC-FID) and equipped with a split/splitless injector (GC-2010 chromatograph; Shimadzu, Japan) (A.M.M. Costa et al., 2019). Briefly, FAME were separated in a capillary column (30 m × 0.32 mm i.d., 0.25 µm film; Omegawax-320, Supelco Co. Bellefonte, PA, US). Helium was used as the carrier gas, and column pressure was set to attain a carrier gas speed of 25.0 cm/s. Temperatures of injector and detector were 260 °C and 280 °C, respectively, and the split ratio was 1:20. The oven program was as follows: 160 °C for 2 min, temperature programmed at 2.5 °C/min to 190 °C and held for 5 min, then temperature programmed at 3.5 °C/min to 220 °C and held for 15 min. Gas-chromatographic peaks of FAME were identified by comparison with a commercial mixture of standards (37-component FAME mix; Supelco, PA, Bellefonte, PA, US) and, specifically for cLnA isomers, individual commercial isomers (punic acid, α-eleostearic acid, catalpic acid and β-eleostearic acid; Larodan AB, Solna, Sweden) were used. Data were collected by Lab Solutions GC software package (version 2.30.00, 2004; Shimadzu Co., Kyoto, Japan). Fatty acid composition results were expressed as g/100 g fatty acid.

2.5.4 Peroxide Value

Microparticles’ peroxide value was evaluated spectrophotometrically (Shantha & Decker, 1994). Briefly, an aliquot of the extract obtained in section 2.5.1 was evaporated under N2 and the residual oil (10 mg) was resuspended in a chloroform/methanol (7:3) mixture (9.8 mL). For color formation, 50 µL of an ammonium thiocyanate solution was added and briefly vortexed, right after the same process was repeated with a Fe+2 solution (50 µL). Samples reacted in the dark for 5 min and absorbance was measured at 500 nm (UV- 1800; Shimadzu Corporation, Kyoto, Japan). Hydroperoxides concentrations were determined using a Fe+3 standard curve with iron concentrations varying from 1 to 35 µg. Results were expressed as mEq O2/kg oil.

2.5.5 Moisture content

Moisture content was determined in all samples with a moisture balance with infrared radiation heating (MA35Mettler Toledo, Urdorf, Switzerland).

2.5.6 Water activity (aw)

Water activity was determined in water activity analyzer (LabMaster-aw, Novasina AB, Lanchen, Switzerland). 127

2.5.7 Particle size

PSO’s microparticles were dispersed in isopropanol and analyzed in a particle size analyzer SDC- Microtrac S3500 (Microtrac, Montgovery Vile, US) by the scattering pattern of a transverse laser light. Results were reported as D[0.5] and scattering index (Span) (Equation 5), which are defined as maximum size (µm) of 50% analyzed particles and width distribution of particles size range, respectively.

퐷[0.9] − 퐷[0.1] [Equation 5] 푆푝푎푛 = 퐷[0.5]

Where: D[0.9], D[0.1] and D[0.5] are the diameters at 10, 50, and 90% cumulative volume, respectively.

2.5.8 Microparticles Morphology

PSO microparticles morphology was evaluated by scanning electron microscopy (SEM). Samples were directly deposited on carbon conductive tape on aluminum SEM stubs, and coated with a thin gold layer, using a gold-sputter (Desk V, Denton Vacuum®, Moorestown, NJ, US). The samples were analyzed using a Tescan Vega 3 SEM (Tescan®, Kohoutovice, Czech Republic) operated at 15 kV.

2.5.9 X-ray photoelectron spectroscopy (XPS) of microparticles

XPS analysis was performed on a UHV Xi ESCALAB 250 (Thermo Fisher Scientifics, US) spectrometer equipped with a hemispherical electron energy analyzer. The XPS spectra were collected using monochromatic Al Kα X-ray source (1486.6 eV) and an electron emission angle of 90° with the surface. Survey scans were recorded with 1 eV steps and 100 eV analyzer pass energy. The high-resolution C1s spectra were recorded with 0.1 eV step and 25 eV pass energy analyzer. The linearity of the energy scale was checked using Cu (932.7 eV), Ag (368.3 eV) and Au (84.0 eV) lines. Data processing was performed using Thermo Avantage software. Peak fitting was carried out with Lorentzian/Gaussian ratio of 30%/70%.

128

2.6 Statistical analysis

For the central composite design, analysis of variance (ANOVA), test of lack of fit, determination of regression coefficients and the construction of response surface (3D) graphs Statistica 7.0 (StatSoft, Tulsa, US) software package was used. Statistical comparisons were based on at least triplicate results, and all data were presented as mean and standard deviation. Comparisons between means were made by one-way ANOVA with Tukey’s post-test considering central point variance. Data were analyzed with GraphPad Prism v.6.0 (GraphPad software 2012, La Jolla, CA, US). Significance level was established at p< 0.05.

3. Results and Discussion

This work was a successfully attempt to produce versatile microparticles of PSO with high oil retention and punicic acid contents by complex coacervation. Moreover in this work, the atomization process was used as a recovery and hardening step in order to reduce the cost of the complex coacervation technique and make it less time consuming.

3.1 Wet microparticles

3.1.1 Coacervation yield

The coacervation yield of the formulations varied from 37% to 81% (Figure 1). Previous works showed similar results in the encapsulation of hydrophobic cores by complex coacervation using systems like whey protein-gum Arabic (Eratte et al., 2014), gelatin-gum Arabic (Jun-xia, Hai-yan, & Jian, 2011) and soybean protein isolate-gum Arabic (Xiao, Liu,

Zhu, Zhou, & Niu, 2014). Coacervation process releases in solution counterions and H2O molecules. In this sense, the use of higher Cp implies in a higher release of counterions, which can suppress the coacervation process by hindering polymers surface charges (the so-called screening of electrostatic interactions), thus inhibiting coacervate formation (Weinbreck et al., 2003). For this reason, polymer concentrations above 8% were not used in the present work.

129

Figure 1. Influence of polimer concentration (Cp) and ratio wall material:oil (WM:Oil) in coacervate yield (in parenthesis) of the experimental design formulations. * A dye (oil red) was used in formulation 1 and 2 to demonstrate the affinity of the coacervate phase to the PSO. 130

3.1.2 Morphology

A noticeable coacervate layer around the oil droplets was present in all formulations (Figure 2A, B, C, D, E, F, G, H and I), indicating the effectiveness of the encapsulation process. As a control, formulation 9 was observed at pH = 7, in which WPI and gum Arabic possess similar net charges, thus the coacervation process is suppressed and, as expected, there was no coacervate layer adsorbed on the oil droplets (Figure 2L). Moreover, the coacervates without the PSO were also observed, showing a transparent gel like structure (Figure 2J1, J2) (Weinbreck et al., 2004). According to Figure 2, all formulations presented multicored microparticles, which is generally observed when high homogenization degrees are applied in the emulsification step prior to the coacervation process (Eratte et al., 2014; Kaushik et al., 2016; S. Liu et al., 2010b).

Both Cp and WM:Oil ratio influenced the wet microparticles morphology. According to Figure 2, an increase in Cp in formulations with the same WM:Oil ratio (F5, F6 and F9; Figures 2E, F and I) decreased the oil droplets size, which might be attributed to a higher number of polymer molecules, that were able to completely cover the oil droplets, preventing droplet coalescence. Piacentini et al. (2013) applied the system cold water fish gelatin-gum Arabic to microencapsulate sunflower oil by complex coacervation and observed this same trend, when the polymer concentration was raised from 0.5 to 4%. While polymer concentration affected the oil droplets size, the WM:Oil ratio influenced the thickness of the coacervate layer. When comparing formulations with the same polymer concentration and distinct WM:Oil ratios (F1 vs F2; F3 vs F4; F7 vs F8 vs F9; Figure 2A, B, C, D, E, G, H, I), an increase in the WM:Oil ratio produced microparticles with a thicker coacervate layer. Microparticles with a thicker coacervate layer are more suitable to the atomization process, because a minimum length between the core and microcapsule’s surface should be maintained in order to avoid oil droplets migration to particles’ surface, reducing microparticles’ surface oil (Goula & Adamopoulos, 2012).

131

Figure 2. Wet capsules’ micrographs of the experimental’s design formulations obtained by optical microscopy. Formulations: A) 1; B) 2; C) 3; D) 4; E) 5; F) 6; G) 7; H) 8; I) 9; J1) Coacervate; J2) Coacervate; L) Formulation 9 produced at pH 7. Wet capsules were prepared in pH 3,7, whey protein: gum Arabic ratio (2:1) and reaction time of 10 min. A, B, C, D, E, F, G, H, I, J2 and M: magnification of 400×; J2: magnification of 1000×.

132

3.2 Dried microparticles

3.2.1 Experimental design

The results for each response analyzed in the central composite design are shown in Table 1. The regression coefficients for the polynomial equation, the F values and determination coefficients (R2) of the variables influenced by the independent factors are shown in Table 2. The calculated F values were higher than the tabulated ones for all the evaluated responses, except for ME, indicating that this response could not be predicted by the obtained model.

133

Table 1. Oil retention, microencapsulation efficiency (ME), punicic acid content, peroxide value, moisture, water activity (aw), particle size, for the 11 trial of the experimental design. (Mean ± SD).

F Cp Ratio Oil retention ME Punicic acid Peroxide Moisture aw Particle size (%) (WM:Oil) (%) (%) content value (%) (g /100g fatty (mEq O2/kg D[0.5] (µm) span acid) oil) 1 3 (-1) 1.15 (-1) 78.05 ± 3.23 50.00 ± 0.88 60.48 ± 0.54 19.48 ± 2.93 3.81 ± 0.28 0.410 ± 0.011 8.43 ± 0.24 3.33 ± 0.08 2 3 (-1) 4.35 (+1) 75.00 ± 0.01 55.16 ± 0.11 60.49 ± 0.74 18.53 ± 2.57 4.81 ± 0.28 0.329 ± 0.002 9.18 ± 1.53 2.24 ± 0.59 3 7 (+1) 1.15 (-1) 98.58 ± 1.06 43.17 ± 4.17 60.69 ± 1.83 12.00 ± 0.66 1.91 ± 0.42 0.353 ± 0.001 10.08 ± 0.51 1.27 ± 0.17 4 7 (+1) 4.35 (+1) 88.64 ± 4.68 60.76 ± 2.09 59.95 ± 0.37 15.57 ± 2.11 4.21 ± 0.01 0.288 ± 0.006 8.67 ± 0.25 1.29 ± 0.05 5 2.17 (-1.41) 2.75 (0) 60.33 ± 2.71 61.40 ± 0.84 59.88 ± 0.78 11.17 ± 1.04 5.30 ± 0.41 0.366 ± 0.006 8.36 ± 1.38 3.75 ± 0.12 6 7.82 (+1.41) 2.75 (0) 85.08 ± 1.11 62.92 ± 0.18 61.22 ± 0.19 8.10 ± 1.30 3.21 ± 0.28 0.219 ± 0.006 10.96 ± 0.75 1.61 ± 0.21 7 5 (0) 0.48 (-1.41) 106.19 ± 0.44 38.52 ± 1.92 62.12 ± 0.38 7.44 ± 0.51 0.4 ± 0.28 0.285 ± 0.036 9.10 ± 0.10 1.53 ± 0.07 8 5 (0) 5.01 (+1.41) 83.23 ± 3.43 67.40 ± 5.31 61.08 ± 1.71 10.26 ± 0.85 4.19 ± 0.34 0.230 ± 0.006 8.82 ± 0.44 0.99 ± 0.06 9ª 5 (0) 2.75 (0) 95.2 ± 0.08 55.46 ± 4.62 62.93 ± 2.14 10.3 ± 0.56 4.92 ± 0.71 0.283 ± 0.001 9.29 ± 0.09 1.54 ± 0.24 9ª 5 (0) 2.75 (0) 94.28 ± 3.04 56.98 ± 0.59 64.22 ± 0.5 13.88 ± 2.27 4.19 ± 0.52 0.379 ± 0.004 10.28 ± 0.12 1.41 ± 0.29 9a 5 (0) 2.75 (0) 105.41 ± 0.13 48.74 ± 0.47 63.85 ± 1.1 13.27 ± 2.45 4.38 ± 0.28 0.331 ± 0.068 10.13 ± 1.08 1.92 ± 0.06 a F: Formulation; central point; Cp: polymer concentration in the emulsion; WM: wall material aw: water activity; D[0.5]: maximum size (µm) of 50% analyzed particles; span: particle size scattering index. 134

Table 2. Coded second-order regression coefficients, F values and determination coefficients (R2) for oil yield, microencapsulation efficiency (ME), punicic acid content and particle size scattering index (Span). Coefficient Oil retention (%) ME (%) Punicic acid content Span (g /100g fatty acid) ** ** ** ** β0 98.29 53.75 63.67 1.62 * ** β1 8.66 0.12 0.19 -0.76 ** ** ** β11 -12.51 2.90 -1.73 0.55 * β2 -5.69 7.96 -0.27 -0.23 * β22 -1.44 -1.73 -1.21 -0.17 β12 -1.72 3.11 -0.19 0.28 Fcalculated 13.91 4.6 8.4 57.98 Ftabulated 5.05 5.05 5.05 5.05 R2 0.932 0.821 0.896 0.980 β0: mean; β1: Total polymer concentration linear; β11: Total polymer concentration quadratic; β2: Wall material: oil (WM:Oil) ratio linear; β22: WM:Oil ratio quadratic; β12: Total polymer concentration× WM:Oil ratio. *significant at p < 0.06; **significant at p < 0.05.

3.2.1.1 Moisture and water activity (aw)

PSO’s microparticles moisture and aw varied from 0.4-5.3% and 0.219-0.410, respectively. These values were expected for microparticles produced by spray drying (moisture ≤ 6% and aw ≤ 0.6) (Reineccius 2004; Klaypradit & Huang 2008).

3.2.1.2 Oil retention

All formulations showed oil retention between 75 and 106%, considerably higher than related in the microencapsulation of flaxseed oil and tuna oil by complex coacervation (~50%) (Eratte et al., 2014; Kaushik et al., 2016; S. Liu et al., 2010b). In these studies the systems gelatin-gum Arabic and flaxseed protein-flaxseed gum were utilized for flaxseed oil, and whey protein-gum Arabic, for tuna oil. Oil retention was significantly influenced by the

polymer concentration (Table 2). According to Figure 3A, the increase of Cp was associated

to the increase of oil yield until an maximum Cp value ranging from 5 to 7%. Previous studies (Frascareli et al., 2011; Goula & Adamopoulos, 2012; Sahin-Nadeem & Özen, 2014; Tonon,

Grosso, et al., 2011) indicated that the increase of Cp implies in an increase of emulsion viscosity. Viscosity of the coacervate layer is important to form a resistant barrier around the

oil droplet, thus favoring oil retention. In addition, this effect of Cp in the oil yield could also

be associated to the oil droplet size reduction observed in formulations with higher Cp, as 135

shown in Figure 2. Small oil droplets are more easily encapsulated by the coacervate layer (Weinbreckt et al., 2004).

Furthermore, considering the spray drying process, emulsions with higher viscosity require a short time to form a semi-permeable membrane at the surface of the drying particle. A fast crust formation could be associated to lower surface oil content and reduction of losses by evaporation during the process, as there is less time to the core to diffuse until the drying particle surface. Additionally, higher viscosities prevent circulation movements inside the drying particles, thus avoiding the diffusion of the core to the microcapsule surface (Goula & Adamopoulos, 2012;Ré, 1998).

3.2.1.3 Microencapsulation Efficiency

The encapsulation efficiency of the PSO’ microparticles varied from 38.5% to 67.4% (Table 1) and was only positively influenced by the WM:Oil ratio (Table 2). As shown in Table 1, formulations with WM:Oil ratios equal to or greater than 2.75 showed the highest values of ME. A response surface was not presented for this response, because the proposed 2 model did not show a good fit to experimental data (R = 0.82 and Fcalculated

Additionally at low WM:Oil ratios (high oil loads), the Cp utilized could have not been enough to completely cover the oil droplets and prevent oil droplets coalescence. As result, large droplets were surrounded by a thin coacervate layer, while small droplets were encapsulated by a matrix of coacervate, as shown in Figure 2.

The non-encapsulated oil or surface oil is readily oxidized, as there is no physical barrier preventing its interaction with O2 and light exposure (Ahn et al., 2008). Previous studies demonstrated that microparticles with high surface oil contents posses low oxidative stability (Kaushik et al., 2016; Kim & Kim, 2014). Although we observed that the oil load inversely affected ME, it is economically advantageous to produce microparticles with the higher oil load as possible. In this sense, the complex coacervation is an appealing 136

encapsulation technique capable of deliver microparticles with high ME while using low Cp, comparing to the encapsulation by spray drying (Eratte et al., 2014; Jun-xia et al., 2011). However, there is a discrepancy in the ME results of microparticles produced by complex coacervation in the literature. ME values can vary from 22.2% (Rutz et al., 2017) by applying the chitosan-pectin system to encapsulate palm oil, to ~ 90% in encapsulated tuna oil by WPI- gum Arabic complex coacervates (Eratte et al., 2014). ME results observed in this study were within this range.

3.2.1.4 Punicic acid content

PSO is a rich source of a bioactive class of fatty acids, namely cLnA, as demonstrated in Figure 3D. GC analysis, both with FID and MS detectors, showed eight cLnA isomers from which four (punicic, α-eleostearic, catalpic and β-eleostearic acids) were identified based on comparisons with standards’ retention times and mass spectrometric fragmentation profiles. PSO fatty acid profile was evaluated in a recent study from the group (Costa et al., 2019). Punicic acid is the major cLnA in the PSO and its content is associated to the oil functional properties (Viladomiu et al., 2013). Being PSO highly oxidizable because of its high cLnA content, the encapsulation process promoted a small but significant loss of punicic acid (Figure 3D), possibly by oxidation. In parallel, β-eleostearic acid and other minor cLnA isomers (cLnA 1, cLnA 2, cLnA 3 and cLnA 4) contents increased 1.5 folds to 3.6 folds, implying that part of the punicic acid being lost was isomerized, possibly because of the high temperatures applied in the drying process. These results are in agreement with those reported by Sahin-Nadeem and Afşin Özen (2014) that also showed loss of punicic acid and increase of β-eleostearic acid and α-eleostearic acid after microencapsulation of PSO by spray drying. cLnA isomerization was attributed to high processing temperatures, and the higher thermal stability of trans-isomers compared to cis-isomers (Giua, Blasi, Simonetti, & Cossignani, 2012).

Punicic acid contents only slightly varied between the PSO microparticles formulations (from 59.9 to 64.2 g/ 100 g oil, Table 1), however this marginally difference might result in a final product with a significant lower/higher content of punicic acid, as microparticles’ oil load can vary up to 10 folds between formulations. According to Figure 3B, there was a region of maximum punicic acid content in the microparticles, in the medium conditions of Cp and WM:Oil ratio. The punicic acid content is directly related to the oil retention and ME, because high oil retention indicates low core losses during the process, thus 137

high contents of punicic acid. Similarly, high ME values are also associated to high contents of the bioactive fatty acid, due to low amounts of non-encapsulated punicic acid in the microcapsule surface that could be lost by the oxidative process. In this sense, as expected, the punicic acid content was influenced by the factors that affected these two responses. The experimental design provided the microparticles formulation (central point) with the most adequate physical-chemical conditions (oil droplet size and coacervate layer thickness) to produce PSO’ microparticles with high punicic acid content.

3.2.1.5 Peroxide value

Peroxide value results of the PSO’s microparticles ranged from 7.44 to 19.48 mEq O2/ kg oil. Similarly to the fatty acid composition, the encapsulation process promoted a small but significant increase in the peroxide values of formulations 1, 2 and 4. According to the experimental design results, peroxide value was not significantly influenced by Cp and WM:Oil ratio. Tonon et al. (2011) demonstrated the influence of the solid content and WM:Oil ratio in the microencapsulation of flaxseed oil by spray drying, indicating that the lower oxidative stability was caused by the higher surface oil showed in formulations with low solids contents and higher core loads (low WM:Oil ratio). Although in the present work, ME results varied between PSO’s microparticles and were significantly influenced by WM:Oil ratio, no association between ME and lipid oxidation was verified. Right after PSO’s microparticles production, they were vacuum sealed and stored at -80 °C until analysis, in order to hinder surface oil oxidation. Nevertheless peroxide values results should be evaluated with caution, when used alone to assess lipid oxidation, because hydroperoxides are unstable primary product of lipid peroxidation, thus low concentrations of this products does not necessarily indicate an initial oxidation stage (Liu et al., 2010a).

3.2.1.6 Particle size

The particle size of microencapsulated oils is determined by emulsion preparation, drying procedure and types of encapsulants agents. The influence of this parameter in microcapsule’s technological properties, such as: flowability, dispersibility and solubility, is well defined, however the relation of this parameter with microparticles’ oxidative stability still a controversial matter (Goula & Adamopoulos, 2012; Reineccius, 2004; Sahin-Nadeem & Afşin Özen, 2014) PSO’s microparticles showed a small variation in the particle size (8.36- 10.96 µm), with no significant influence of any independent variable, and an overall narrow 138

particle size distribution, demonstrated by the low span values (Table 1). The results reported in this study are in accordance with different microencapsulated oils produced by spray drying (Goula & Adamopoulos, 2012; Lamprecht et al., 2001; Tonon, Grosso, et al., 2011).

According to the experimental design, only the total polymer concentration significantly affected the span value of the PSO’s microparticles (Table 2). The total polymer concentration influenced this variable in two distinct ways: until 5% of total polymer concentration, an increase in this factor negatively influenced the span, while in concentrations above this value, an increase in the total polymer positively affected the span (Figure 3C). Production of particles with low span values is important from the nutritional point of view, because narrow particle size distribution is important for controlling the sustained core release (Walton & Mumford, 1999).

139

Figure 3. Response surface of: A) oil yield; B) punicic acid content; C) span; and D) Influence of the microencapsulation process on PSO’s conjugated linolenic acids (cLnA) isomers profile. * Cp: Total polymer concentration, WM: wall material

140

3.2.1.7 Dried PSO’ microparticles morphology

Surface morphology of the PSO’s microparticles is shown in Figure 4. Most of the formulations presented spherical shape and variable sizes, which is typical of particles produced by spray drying. Microparticles’ surface topography was characterized by intense invaginations, due to the shrinkage of the particles during the drying process after rapid water evaporation (Ré, 1998). These results are in agreement with Eratte et al. (2015), Eratte et al. (2014) and Yang et al. (2015), they microencapsulated edible oils by complex coacervation and applied the atomization process as a hardening/recuperation step.

As reported in previous studies (Jimenez, García, & Beristain, 2006b; Sheu & Rosenberg, 1998; Yang et al., 2015), wall material properties and composition, WM:Oil ratio and drying conditions affect the microstructure of spray-dried microparticles. In this study, we were able to evaluate the influence of the total polymer concentration and WM:Oil on the microparticles’ surface morphology. In formulations with a WM:Oil ratio of 2.75 (Formulations 5, 6 and 9; Figure 4E, 4F and 4I), the increase of the total polymer concentration above 5% promoted noticeable aggregation and conglutination problems. Yang et al. (2015) microencapsulated poppy-seed oil by complex coacervation utilizing the gelatin- gum Arabic system and related this same trend. According to this study, an overbalance of the wall material intensified microparticles adhesion, because of feed emulsion viscosity increase. Regarding the oil load, formulations with a WM:Oil ratio of 1.15 or lower also showed aggregation and conglutination problems. These problems were more evident in Formulation 7 (Figure 4G) were the microparticle structure collapse and it is not possible to observe individual microparticles’ surface features. This could be attributed to the high surface oil content present in these formulations, which may have enhanced microparticles’ stickiness (Goula & Adamopoulos, 2012). Optimum morphological characteristics were observed for WM:Oil ratios around 2.75-4.35, while at a WM:Oil ratio of 5.01 microparticles showed the high content of cracks and pores (Figure 4H). The presence of these imperfections in formulation 8 could be explained by the high viscosity of the feed emulsion that enable a rapid formation of a semi-permeable membrane on the droplet surface and the excessive evaporation could rupture this membrane and promote the onset of cracks and pores. It is interesting modulate microencapsulation parameters to achieve a continuous phase around the core to provide lower permeability to gases, better oxidative stability and core protection. Additionally microparticles’s surface morphology strongly affects powder floatability and reconstitution properties. 141

Figure 4. Dried microparticles micrographs of the experimental’s design formulations obtained by scanning electron microscopy (SEM). Formulations: A) 1; B) 2; C) 3; D) 4; E) 5; F) 6; G) 7; H) 8; I) 9. White arrows indicates cracks and open pores.

142

3.3 Selection of the most attractive formulation

The selection of the best formulation was based in the main goal of the work, which was to produce a potentional functional ingredient rich in cLnA isomers, with the highest oil retention and minimal surface oil content. In this sense, the variables used to select the most interesting formulation were: oil retention, microencapsulation efficiency and punicic acid content. Based on these responses, the medium values of Cp and WM:Oil ratio (Cp= 5%, WM:Oil ratio: 2.75), which refer to the central point, were chosen. Although this formulation did not show the highest values of microencapsulation efficiency, it was located in the optimum area of the oil retention and punicic acid content surface areas (Figure 3A and B), showing high contents of the most important PSO nutrient. Additionally from the economical point of view, central point formulation is very promising, due to high oil load (26.6 mg of PSO/ 100 mg of microparticles). In this study, we do not utilized the experimental design data to optimize the process because the prediction model proposed for EM did not show a good fit (R2: 0.82 and

Fcalculated

3.4 XPS analysis

The concentration of different carbon species was determined by curve-fitting the high-resolution C1s spectra. The spectra of the WM, WP, GA the elected formulation (F9) and F7 (formulation with the lowest ME: 38.5%) showed carbon components with the following assignments: C1 (C-C, C-H); C2 (C-O, C-N); C3 (C=O, O-C-O, N-C=O) and C4 (O-C=O). The C1s high resolution spectra are shown in Figure 5C, D, E. F and G. The XPS analysis of oil was impractical in ultra-high vacuum as it is in liquid form.

The C1s spectrum of WP is characterized by a particularly strong C1 contribution (Figure 5C). This characteristic is not expected for a protein spectra but can refer to the presence of lipids on the surface. The GA peaks (Figure 5E) showed, as expected, a typical polysaccharide spectrum, dominated by the C-O peak. This was confirmed in the survey spectrum by a strong oxygen signal and almost no nitrogen (Figure 5A). Surface composition of WM coacervates (without oil) presented characteristics of both WP and GA (Figure 5A and C). F9 and F7 spectras combine characteristics of WP spectrum, such as prominent C1 chemical environment (Figure 5F and G) and the presence of nitrogen in the survey spectra 143

(Figure 5A), indicating that different concentrations of Cp results in variations in microparticles surface composition. XPS analysis was tentatively used to evaluate microparticles’ surface composition as this method does not modify samples’ surface composition, as the classical washing method applied earlier. Therefore, after evaluated C1-4 chemical environment, was demonstrated a relation between the O-C=O peak, pointed out as acid/ester groups, and a higher content of surface oil, as samples with encapsulated PSO showed higher contributions of C4 than the control with no PSO (WP, GA, WM; Figure 4B). This result could also be extrapolated to all formulations, because C4 peek intensity was influenced inversely by ME results (Table 1), indicating agreement between XPS results and the washing method performed earlier. Jafari, Assadpoor, Bhandari, & He (2008) microencapsulated fish oil by spray dried and also related an agreement of both methods. Moreover, XPS analysis were also capable to indicate interactions between WP and GA, as we observed an increase C3 and C4 chemical environment in WM, F9 and F7 (Figure 4B), + - probably because a higher exposure of WP (NH3 ) and GA (COO ) interaction sites on microparticles surface, caused by pH adjustment (Eratte et al., 2015).

144

Figure 5. X-ray photoelectron spectroscopy (XPS) results. A) XPS survey spectra of wall material (WM), whey protein (WP), gum Arabic (GA), elected formulation (F9) and formulation with the lowest microencapsulation efficiency (F7); B) Elemental composition of WM, GA, WP, F9 and F7; and XPS C1s high-resolution spectra of C) WM, D) WP, E) GA, F) F9, G) F7. 145

4. Conclusion

PSO was successfully microencapsulated by complex coacervation utilizing the whey protein- gum Arabic system and spray drying as a hardening/recuperation step, the formulation of the feed emulsion has a significant impact in the coacervation process and microparticles’ physico-chemical and morphological properties. Cp and WM:Oil ratio influenced the coacervation yield and wet microparticles morphology. Among the physical- chemical properties evaluated in this study, only oil retention, ME, punicic acid content, moisture content and span were influenced by experimental design factors. Most of the formulations presented interesting morphological characteristics, such as spherical shape with variable size, intense invaginations and smooth surface free of visible cracks. Formulations with high Cp and low WM:Oil ratio showed prominent agglomeration and conglutination problems. XPS results showed that surface analysis can be a powerful toll to characterize microparticles surface composition and it agrees with the extraction method applied. Central point formulation was elected due to samples’ highest punicic acid content and economical appealing indicated by formulations’ high oil retention and interesting oil load (WM:Oil: 2.75). In the present study, Coacervation process parameters manipulation enabled the design of PSO’ microparticles with physical-chemical properties that allows its application in food products. Additionally, these microparticles might subsidize future studies assessing PSO’ consumption health benefits. It would be interesting to evaluate in future studies microparticles’ stability in different condition, digestibility and the development of a functional food applying the PSO’s microparticles.

Acknowledgements

The authors greatly acknowledge the funding provide by CAPES, CNPq (Grant numbers: 432484/2016-7 and 309558/2015-8) and FAPERJ (Grant numbers: E- 26/010.001277/2015 and E-26/203.197/2015) (Brazil). AGT is are cipient of a CNPq scholarship, AMMC was a recipient of CNPq PhD studentships. We are also indebted with Brenda Duarte Gralha and Leticia Korin Moretti for their help with laboratory analysis of microparticles physical-chemical properties.

146

5. References

Cao, Y., Gao, H. L., Chen, J. N., Chen, Z. Y., & Yang, L. (2006). Identification and characterization of conjugated linolenic acid isomers by Ag+-HPLC and NMR. Journal of Agricultural and Food Chemistry, 54(24), 9004–9009. http://doi.org/10.1021/jf0616199

Drusch, S., Serfert, Y., Scampicchio, M., Schmidt-Hansberg, B., & Schwarz, K. (2007). Impact of physicochemical characteristics on the oxidative stability of fish oil microencapsulated by spray drying. Journal of Agricultural and Food Chemistry, 55(26), 11044–51. http://doi.org/10.1021/jf072536a

Eratte, D., Wang, B., Dowling, K., Barrow, C. J., & Adhikari, B. P. (2014). Complex coacervation with whey protein isolate and gum arabic for the microencapsulation of 147

omega-3 rich tuna oil. Food Funct., 5(11), 2743–2750. http://doi.org/10.1039/C4FO00296B

Goula, A. M., & Adamopoulos, K. G. (2004). Spray drying of tomato pulp: Effect of feed concentration. Drying Technology, 22(10), 2309–2330. http://doi.org/10.1081/LDRT- 200040007

I Ré, M. (1998). MICROENCAPSULATION BY SPRAY DRYING. Drying Technology, 16(6), 1195–1236. http://doi.org/10.1080/07373939808917460

Jafari, S. M., Assadpoor, E., Bhandari, B., & He, Y. (2008). Nano-particle encapsulation of fish oil by spray drying. Food Research International, 41(2), 172–183. 148

http://doi.org/10.1016/j.foodres.2007.11.002

Jimenez, M., García, H. S., & Beristain, C. I. (2006). Spray-dried encapsulation of Conjugated Linoleic Acid (CLA) with polymeric matrices. Journal of the Science of Food and Agriculture, 86(14), 2431–2437. http://doi.org/10.1002/jsfa.2636

Jun-xia, X., Hai-yan, Y., & Jian, Y. (2011). Microencapsulation of sweet orange oil by complex coacervation with soybean protein isolate/gum Arabic. Food Chemistry, 125(4), 1267–1272. http://doi.org/10.1016/j.foodchem.2010.10.063

Kim, Y., & Kim, H. (2014). Optimization of microencapsulation of seed oil by response surface methodology Optimization of microencapsulation of seed oil by response surface methodology. Food Chemistry, 107(March 2008), 98–105. http://doi.org/10.1016/j.foodchem.2007.07.067

Kouassi, G. K., Teriveedhi, V. K., Milby, C. L., Ahmad, T., Boley, M. S., Gowda, N. M., & Terry, R. J. (2012). Nano-Microencapsulation and Controlled Release of Linoleic Acid in Biopolymer Matrices : Effects of the Physical State , Water Activity , and Quercetin on Oxidative Stability, 2(March), 1–10.

Kramer, J. G., Fellner, V., Dugan, M. R., Sauer, F., Mossoba, M., & Yurawecz, M. (1997). Evaluating acid and base catalysts in the methylation of milk and rumen fatty acids with special emphasis on conjugated dienes and total trans fatty acids. Lipids, 32(11), 1219– 1228. http://doi.org/10.1007/s11745-997-0156-3

Liu, S., Low, N. H., & Nickerson, M. T. (2010a). Entrapment of flaxseed oil within gelatin- gum Arabic capsules. JAOCS, Journal of the American Oil Chemists’ Society, 87(7), 809–815. http://doi.org/10.1007/s11746-010-1560-7 149

Liu, S., Low, N. H., & Nickerson, M. T. (2010b). Entrapment of flaxseed oil within gelatin- gum Arabic capsules. JAOCS, Journal of the American Oil Chemists’ Society, 87, 809– 815. http://doi.org/10.1007/s11746-010-1560-7

Piacentini, E., Giorno, L., Dragosavac, M. M., Vladisavljević, G. T., & Holdich, R. G. (2013). Microencapsulation of oil droplets using cold water fish gelatine/gum arabic complex coacervation by membrane emulsification. Food Research International, 53(1), 362– 372. http://doi.org/10.1016/j.foodres.2013.04.012

Reineccius, G. A. (2004). The Spray Drying of Food Flavors. Drying Technology, 22(6), 1289–1324. http://doi.org/10.1081/DRT-120038731

Rutz, J. K., Borges, C. D., Zambiazi, R. C., Crizel-Cardozo, M. M., Kuck, L. S., & Noreña, C. P. Z. (2017). Microencapsulation of palm oil by complex coacervation for application in food systems. Food Chemistry, 220, 59–66. http://doi.org/10.1016/j.foodchem.2016.09.194

Sankarikutty, B., Sreekumar, M. M., Narayanan, C. S., & Mathew, A. G. (1988). Studies on microencapsulation of cardamom oil by spray drying technique. Journal of Food Science and Technology, 25(6), 325–355. Retrieved from http://cat.inist.fr/?aModele=afficheN&cpsidt=7181377

Shantha, N. C., & Decker, E. A. (1994). Rapid, sensitive, iron-based spectrophotometric methods for determination of peroxide values of food lipids. Journal of AOAC 150

International, 77(October), 421–424.

Sheu, T.-Y., & Rosenberg, M. (1998). Microstructure of microcapsules consisting of whey proteins and carbohydrates. Journal of Food Science, 63(3), 491–494.

Tonon, R. V, & Freitas, S. (2011). Spray drying of açai ( Euterpe oleraceae Mart .) juice : Effect of inlet air temperature and type of carrier agent SPRAY DRYING OF AÇAI ( EUTERPE OLERACEAE MART .) JUICE : EFFECT OF INLET AIR TEMPERATURE AND TYPE OF. Journal of Food Processing and Preservation, 35(April 2011), 691–700. http://doi.org/10.1111/j.1745-4549.2011.00518.x

Walton, D., & Mumford, C. (1999). The morphology of spray-dried particles: the effect of process variables upon the morphology of spray-dried particles. Chemical Engineering Research and Design, 77(July), 442–460. http://doi.org/http://dx.doi.org/10.1205/026387699526296

Weinbreck, F., de Vries, R., Schrooyen, P., & de Kruif, C. G. (2003). Complex coacervation of whey proteins and gum arabic. Biomacromolecules, 4(2), 293–303. http://doi.org/10.1021/bm025667n

Weinbreck, F., Rollema, H. S., Tromp, R. H., & De Kruif, C. G. (2004). Diffusivity of whey protein and gum arabic in their coacervates. Langmuir, 20(14), 6389–6395. http://doi.org/10.1021/la049908j

Weinbreckt, F., & Minorf, M. (2004). Microencapsulation of oils using whey protein / gum arabic coacervates. Journal of Microencapsulation, 21(6), 667–679. http://doi.org/10.1080/02652040400008499

152

Chapter 4

Application of oxidatively stable pomegranate seed oil (Punica

granatum L.) microparticles in an instant coffee latte drink

Authors: Costa, André M. M.; Calado, Verônica; Tonon, Renata V.; Torres, Alexandre G.

153

Abstract

The present work aimed to investigate the oxidative stability (ambient temperature during 90 days and 60 °C during 10 days with/without vacuum) and technological properties of PSO’s microparticles produced by complex coacervation. A vehicle for microparticles consumption, instant coffee latte drink added with PSO’s microparticles, was developed and the influence of microparticles addition on product’s technological properties was also evaluated. Moreover, complex coacervation process was compared throughout the study with spray drying, a more utilized encapsulation method. Stability test at ambient temperature showed an inferior protection ability of the complex coacervation method, as coacervate microparticles demonstrated higher PSO degradation, indicated by cLnA total content reduction and total volatile compounds increase at day 90. These results might be associated with appearance of open pores and cracks in coacervate microparticles. PV and γ-Tocopherol contents results were not affected by the encapsulation method. Additionally, stability parameters (PV, cLnA total content and total volatile compounds content) were associated (Pearson correlation) with microparticles’ γ-tocopherol contents, especially PV (r = -0.79). At 60 °C, coacervate microparticles showed again lower oxidative stability and storage under reduced O2 concentration decreased microparticles’ oxidation ratio, therefore the application of this sealing method might increase PSO’s microparticles shelf life. Coacervate microparticles showed high solubility and thermal stability, enabling it application in a wide range of food products. Only wettability, water activity and thermal stability were affected by the coacervation process. Regardless non coacervate microparticles higher oxidative stability, PSO encapsulation by complex coacervation was considered successfully as microparticles showed an overall stability of 60 days. Moreover, microparticles showed appealing technological properties, which enable their application in a wide range of food products and addition on the instant coffee latte drink validated microparticles’ versatility.

Keywords Pomegranate seed oil, complex coacervation, spray drying, cLnA isomers, oxidative stability, storage, hydroperoxides, volatile compounds, tocopherols, technological properties and functional food.

154

1. Introduction

Pomegranate seed oil (PSO) is a emerging functional oil with a wide range of described bioactive properties such as: cytotoxic effect, antioxidant activity, modulation of the immune system and lipid metabolism (Fernandes et al., 2015; Shabbir et al., 2017). PSO potentional functional activity is associated with its unique fatty acid composition rich in conjugated linolenic acids isomers (C18:3) (cLnA), especially punicic acid (9cis,11trans,13cis-18:3) which accounts for ~ 70% of total fatty acids in the oil (Elfalleh et al., 2011; Fernandes et al., 2015; Shabbir et al., 2017). Nevertheless strategies aiming at PSO food addition are unsuccessful because of oil’s high oxidative instability and hidrophobicity.

Microencapsulation is a “packing” technique which an active ingredient (core), is cover by a wall material (WM), avoiding core interactions with food components and the adjacent environment. The unstable molecules protection provided by the encapsulation process enables the application of these substances in food products. This technique produces free flow powders that are easily handled during storage and food processing. To the best of our knowledge, among the encapsulation methods, complex coacervation has not been applied yet to PSO.

Complex coacervation is a liquid-liquid phase separation phenomenon that occurs when electrostatically opposite charged biopolymers, usually proteins and polysaccharide, are submitted to specific conditions, producing aggregates (coacervates) that promptly deposit on to oil phase, forming a protection barrier around oil droplets (Weinbreckt & Minorf, 2004). Comparing to others encapsulation techniques, complex coacervation is capable to produce microparticles with higher microencapsulation efficiency utilizing high core load and low WM concentration (Gouin, 2004). A recent study (Chapter 3) successfully stabilized PSO by complex coacervation utilizing the whey protein-gum Arabic system, producing microparticles with appealing physical-chemical properties such as: high punicic acid content, high oil retention and spherical-shaped morphology. However, microparticles oxidative stability under distinct storage condition were not investigated, as well as it suitability for food addition. Thus following the previous work, in the present study microparticles’ stability was assessed in 2 conditions (ambient temperature and 60 °C) and the influence of atmospheric O2 on PSO oxidation was also evaluated. It is worth emphasize that cLnA’s oxidative pathways still an unexplored subject, therefore this group of fatty acids does not have established oxidation markers. Additionally, different from others unsaturated fatty 155

acids, hydroperoxides are not the principal primary oxidation products, as conjugated fatty acids demonstrated high polymers formation from the very beginning of the oxidation process (Márquez-Ruiz et al., 2014; Miller & Claxton, 1928). In this sense, in order to accurately assess PSO oxidative status different compounds involved in oxidative process were investigated, such as: hydroperoxides, cLnA isomers, volatile compounds and tocopherols.

Coffee is one of the most consumed beverages in the world and in Brazil it is part of the population dietary habits (Arruda et al., 2009). Recent data from the Brazilian association of the coffee industry (ABIC, 2017) shows that coffee’s internal consumption continues to grow. Additionally, the special coffees and capsules segment is experiencing a sharp grown. In this sense, an instant coffee latte drink was designed and proposed to act as a vehicle for PSO’s microparticles consumption. In this context, the present work aimed to investigate the oxidative stability in two storage condition (ambient temperature during 90 days; 60 °C during 10 days with/without vacuum) and technological properties of the PSO’s microparticles produced by complex coacervation. A potentional functional product (instant coffee latte drink added with PSO’s microparticles) was developed, as well the influence of microparticles addition on product’s technological properties was also evaluated. Moreover, complex coacervation process was compared throughout the study with spray drying, a more utilized encapsulation method.

2. Material and Methods

2.1 Materials

Microparticles were prepared utilizing cold-pressed pomegranate seed oil (PSO) (C16:0 = 2.25%, C18:0 = 1.86%, C18:1 n-9 = 4.52%, C18:1 n-7 = 0.41%, C18:2 n-6 = 5.29%, C20:0 = 0.61%, C20:1 n-9 = 0.78%, Total cLnA = 84.3%; Oneva Food Co®; Istanbul, Turkey) as core and Whey protein isolate (WPI) (Alibra®; São Paulo, Brazil) and Gum Arabic (GA) (Instantgum BA®; Colloides Naturels, São Paulo, Brazil) combined as wall material (WM). A commercial mixture of fatty acid methyl esters (37-component FAME mix; Supelco, Bellefonte, PA, US), individual cLnA commercial isomers (punic acid, α-eleostearic acid, catalpic acid and β-eleostearic acid; Larodan AB, Solna, Sweden) and decaheptanoic acid (Sigma-Aldrich, St Louis, MO, US) were employed for fatty acid identification and quantification. Bromobenzene and C7-C30 saturated alkanes solutions from Supelco Co 156

(Bellefonte, PA, US) were utilized in volatile compounds quantification and identification, respectively. Tocopherols (α-,β-,γ-,δ-) (≥ 99% purity) from Supelco Co (Bellefonte, PA, US) was employed for tocopherols identification. Products were composed by powered milk (Ninho, Nestle®, Minas Gerais, Brazil) and soluble coffee (Pilão®, São Paulo, Brazil). All solvents used were HPLC grade from Tedia (GO, Brazil). All reagents used were from Merck (Darmstadt, Germany).

2.2 Production of the PSO’s microparticles by complex coacervation

PSO’s microparticles were produced by complex coacervation as described previously

(Chapter 3). Briefly, microparticles were composed by a total polymer concentration (Cp) of 5%, considering a whey protein and gum arabic (WPI:GA) of 2:1 (w/w) , and wall material:oil ratio (WM:Oil ratio) of 2.75:1 (w/w). WPI and GA were weighed and diluted in distilled water to obtain solutions with 5% Cp . Firstly, WPI and GA were dissolved separately under magnetic stirring during 30 min at ambient temperature and solutions’ pH were adjusted to 7.0. After the dissolution step, PSO was added dropwise to the WPI solution under continuous stirring (16.000 RPM) during 5 minutes with an Ultra-Turrax (T25-IKA®; IKA, Wilmington, US) in order to produce a stable emulsion. GA solution was added to the previous emulsion and homogenized for 1 min at 16.000 RPMs. Finally, emulsion’s pH was adjusted to 3.75 by adding HCl (1 N) aiming at induce electrostatic interaction between WPI and GA, and formation of the wet microparticles. Microencapsulation process was carried out at 25 °C with a reaction time of 10 min and straightaway the emulsions containing the microparticles were submitted to the atomization process to obtain dry microparticles. This rapid microencapsulation procedure was tested because of the PSO low oxidative stability. In order to compare the complex coacervation process with a more traditional encapsulation method, a control identified as non-coacervate was produced by spray drying. Non-coacervate microparticles showed identical composition (Cp = 5%; WM:Oil ratio = 2.75:1) and production process was identical as coacervate microparticles, excluding only the pH adjustment to 3.75, which induced the production of the wet microparticles. Therefore, non- coacervate microparticles were feed to spray dryer right after Ultra-Turrax homogenization.

2.3 Drying of the PSO’s microparticles

Coacervate and non-coacervate emulsions were dried in a laboratory scale spray dryer (SD-06AG, Lab Plant, North Yorkshire, UK) equipped with a 0.5 mm nozzle. The drying 157

conditions were as follows: inlet temperature: 180 °C, outlet temperature: 66 ± 4 °C, air flow pressure: 2.5 bar and feed flow rate: 0.5 L/h. Dried microparticles were collected, sealed in plastic bags under vacuum, covered in aluminum foil and stored at -80 °C until further analyzes.

2.4 Physical-chemical characterization

2.4.1 Microencapsulation efficiency (ME)

Total oil content present in the microparticles was extracted in a Soxhlet apparatus (SER 148; Velp® Scientifica, Usmate, Italy). Samples (1 g) were extracted using 70 ml petroleum ether with added BHT (0.05%, w/v) during 4 h, the synthetic antioxidant was used to prevent oil oxidation throughout extraction, (Goula & Adamopoulos, 2012). After solvent evaporation with N2, the resulting oil was weighed. Surface oil was determined according to Sankarikutty, Sreekumar, Narayanan, & Mathew (1988), with modifications proposed by Kouassi et al. (2012). Briefly, 5 mL of hexane was added to the microparticles (200 mg) and gently mixed for 5 min. Then, samples were centrifuged (715.52 ×g, 5 min) for separation of the supernatant containing the non-encapsulated oil and before collection in a previously tared erlenmeyer flask (125 ml), the supernatant was filtered in a filtration column, composed of

Celite and Na2SO4, and in a Whatman filter n°1 (Marlborought, MA, US). These filtration apparatus was washed twice with hexane (5 ml), and the filtrates were collected in the former erlenmeyer flask. Finally, the solvent was left to evaporate and the flasks were dried in an oven (105 °C) until constant weight. The surface oil was calculated based on the difference between the initial clean flask and that containing the extracted oil residue. EM was calculated according the following Equation 1:

Total oil − surface oil [Equation 1] ME (%) =  100 Total oil

2.4.2 Particle size

PSO’ microparticles were dispersed in isopropanol and analyzed in a particle size analyzer (SDC- Microtrac S3500; Microtrac, Montgovery Vile- US) by the scattering pattern of a transverse laser light. Results were reported as D[0.5] and scattering index (span), which 158

are defined as maximum size (µm) of 50% analyzed particles and the particles size range, respectively. The span was calculated as follows (Equation 2):

D[0.9] − D[0.1] [Equation 2] span = D[0.5]

Where: D[0.9], D[0.1] and D[0.5] are the maximum size (µm) of 90%, 10% and 50% analyzed particles, respectively.

2.5 Stability test

Complex coacervation (coacervate microparticles) protection efficiency during storage was compared with encapsulation by spray drying (non-coacervate) utilizing the following stability parameters: moisture, water activity (aw), peroxide value (PV), cLnA isomer profile and content, volatile compounds content, tocopherols content and morphological evaluation, in two conditions:

a) Stability at ambient temperature: this condition aimed at to define PSO’s microparticles shelf life. PSO microparticles were sealed in plastic bags and protected from light during 90 days of storage at ambient temperature (23.6 ± 1.6 °C; 0.28 ± 0.02% RH). Samples were withdrawn for analyzes at days 0, 15, 30, 60 and 90. b) Stability at 60 °C: the objective of this drastic condition was to investigate the influence of O2 on microparticles oxidative stability, in this sense PSO’s microparticles were sealed in two distinct conditions: under atmospheric air and under vacuum (reduced O2 atmosphere). Samples were sealed and storage in an oven at 60 °C protected from light during 10 days. Samples were withdrawn for analyzes at days 0, 5 and 10.

2.5.1 Moisture and water activity (aw)

Moisture was determined in all samples with a moisture balance (MA35 Mettler Toledo, Urdorf, Switzerland) according to Costa et al. (2015). Water activity was evaluated in water activity analyzer (LabMaster-aw, Novasina AB, Lanchen, Switzerland) (Costa et al., 2015).

159

2.5.2 Encapsulated oil extraction

2.5.3 Peroxide Value

Microparticles’ peroxide values (PV) were evaluated spectrophotometrically according with the method proposed by Shantha & Decker (1994). Briefly, an aliquot of the extract, obtained in section 2.5.1, was evaporated under N2 and the oil residue (10 mg) was resuspended in a chloroform/methanol (7:3) mixture (9.8 mL). For color formation, 50 µL of an ammonium thiocyanate solution was added and briefly vortexed, right after the same process was repeated with a Fe+2 solution (50 µL). Samples reacted in the dark for 5 min and absorbance was measured at 500 nm (UV-1800; Shimadzu Corporation, Kyoto, Japan). Hydroperoxide concentrations were determined using a Fe+3 standard curve with iron concentrations varying from 1 to 35 µg. Results were expressed as meq O2/ kg oil.

2.5.4 cLnA isomer profile and quantification

PSO fatty acids, extracted from the microparticles by a Soxhlet apparatus (section 2.5.2), were methylated by a base-catalyzed transesterification procedure (Kramer et al.,1997) to avoid isomerization. PSOs’ fatty acid methyl esters (FAME) were analyzed by gas chromatography coupled with flame ionization detector (GC-FID) and equipped with a split/splitless injector (GC-2010 chromatograph; Shimadzu, Japan) (Costa et al., 2019). Briefly, FAME were separated in a capillary column (30 m × 0.32 mm i.d., 0.25 µm film; Omegawax-320, Supelco Co. Bellefonte, PA, US). Helium was used as the carrier gas, and column pressure was set to attain a carrier gas speed of 25.0 cm/s. Temperatures of injector and detector were 260 °C and 280 °C, respectively, and the split ratio was 1:20. The oven program was as follows: 160 °C for 2 min, temperature programmed at 2.5 °C/min to 190 °C and held for 5 min, then temperature programmed at 3.5 °C/min to 220 °C and held for 15 min. Gas-chromatographic peaks of FAME were identified by comparison with a commercial 160

mixture of standards (37-component FAME mix; Supelco, PA, Bellefonte, PA, US) and, specifically for cLnA isomers, was utilized individual commercial isomers (punic acid, α- eleostearic acid, catalpic acid and β-eleostearic acid; Larodan AB, Solna, Sweden). Data were collected by Lab Solutions GC software package (version 2.30.00, 2004; Shimadzu Co., Kyoto, Japan). cLnA isomers were estimated by internal normalization and results were expressed as g/100 g fatty acid. Total cLnA content was quantified by internal standardization. A solution of decaheptanoic acid (Sigma-Aldrich, St Louis, MO, US) in chloroform (1.91 mg/mL) was added in the extraction step (section 2.5.2) and results were expressed as total cLnA mg/ 100 mg of microparticles.

2.5.5 Volatile Compounds by SPME-GC-MS

Volatile compounds from commercial PSOs were extracted by solid phase microextraction (SPME) using a divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber (Supelco, PA, USA) (Larick & Parker, 2001) and analyzed by GC- MS as described previously (Akilet al.,2015; Costa et al., 2015). Prior to extraction, the fiber was conditioned for 60 min in the GC injection port at 270 °C. Samples (1 g) were weighed in a headspace vial, dispersed in KCl (4 mL; 3 M) and 20 µL of internal standard (0.1 mg/mL of bromobenzene; Supelco, PA, US) in methanol was added, and homogenized, followed by sealing the flasks with a PTFE-lined septum. Headspace vials were placed in a glycerol bath (40 °C) under agitation for 30 min until equilibration, septum was pierced and the fiber was exposed to the sample headspace for 10 min.

Qualitative analysis was performed by GC-MS on a Trace 1300 gas chromatograph coupled to a ISQ LT mass spectrometer (Thermo Scientific, MA, US) equipped with a split/splitless injector and a fused capillary silica column 5% phenyl/ 95% methylpolysiloxane (30 m × 0.32 i.d., 3 µm film; 007-5; Quadrex, USA). Volatile compounds were desorbed from the SPME fiber in the injection port for 3 min at 260 °C, in splitless mode, after 3 min sampling time the split purge valve was open at 3.0 mL/min. Helium was used as the carrier gas and the column pressure was set at 49 kPa. The column oven temperature was held at 30 °C for 10 min, then increased at 3 °C/min to 200 °C and held for 10 min. The mass spectrometer was operated in electron impact mode at 70 eV, and the interface and ion source temperatures were 260 °C. Analyses were performed in full scan acquisition mode, at a mass range of 40 to 500 m/z, at 0.5 scan/s. A mixture of hydrocarbons (C7-C30 saturated alkanes; Supelco, PA, US) was run under the same conditions and used as standards to allow 161

calculation of linear retention index (LRI) values for the volatile compounds (Viegas & Bassoli, 2007). Data were collected by GC-MS software package (Xcalibur, Thermo Scientific, MA, US). Compounds were tentatively identified utilizing these following data: 1) Comparison of mass spectra with those of National Institute of Standards (NIST) library and calculation of similarity indexes (Rmatch) provided by the instrument’s software; and 2) comparing (LRIexperimental and LRIpublished values and calculating the variation between them

(variation (%) = (((LRIexperimental-LRIpublished)/LRIexperimental) × 100)). Based in these prior information, the criteria for volatile compounds identification were: 1) Rmatch > 800 and variation between LRIexperimental and LRIpublished values < 5%; or 2) Rmatch > 900 (no LRIpublished values available). LRIpublished were obtained from the following websites libraries: pherobase.com; webbook.nist.gov; pubchem.ncbi.nlm.nih.gov and chemspider.com.

Quantitative analysis was tentatively performed utilizing the same GC-MS apparatus described above. Qualitative analysis identified a large number of compounds, therefore in order to keep quantification results objective and avoid misleading conclusions, compounds quantification followed the following parameters: a) compounds should be a recognized oxidation markers; and/or b) compounds concentration should be influenced by time during the stability. In this sense, The following compounds were quantified at ambient temperature: butanal; 3-methyl, pentanal; 1-pentanol; hexanal; 1-butanol, 3-methyl-, acetate; 2-heptenal; decane; phenylethyl alcohol; 2,4-nonadienal and dodecane. While, pentanal; hexanal; 2- hexenal; 2-heptenal; decane; 2-octenal; phenylethyl alcohol; 2-heptenal, 2-propyl and 2,4- nonadienal were quantified at 60°C. Volatile fraction was estimated by internal standardization utilizing bromobenzene solution (0.1 mg/mL) and results were expressed as total volatile compounds content µg/ g of microparticles.

2.5.6 Tocopherols by HPLC-FLU

The contents of tocopherols in PSO’s microparticles were determined in a HPLC system (Shimadzu®, Japan) consisting of a quaternary pump LC-20AT, system controller CBM-20A, degasser DGU-20A5, photodiode array detector (PDA) SPD-M20A and fluorescence detector RF-10AXL. An aliquot of the extract containing a know mass of PSO

(50 mg) was dried under N2 and dissolved in n-hexane (950 µL), centrifuged (10,000 g, 5 min) and the supernatant solution was filtered through a PTFE syringe filter (0.45 µm) (Gimeno et al., 2000). Chromatographic separation was achieved using a normal phase silica column (Zorbax-SIL 5 µm, 4.6 mm  250 mm; Agilent Technologies; CA, USA) with 162

isocratic elution (hexane:isopropanol, 99:1, v/v) at 1.0 mL/min. Fluorescence detector was operated at 290 nm (excitation) and 330 nm (emission) for tocopherols (Tan & Brzuskiewicz, 1989). Data was acquired by LC solution software (Shimadzu Corporation®, version 1.25, 2009).

Commercial standards of α-, β-, γ- and δ-tocopherols were used for tocopherols identification and quantification by external calibration. Concentrations of each tocol standard stock solutions were determined spectrophotometrically after appropriate dilutions using the following specific extinction coefficients (λ; dL/g.cm) in ethanol: α-tocopherol (292 nm; 75.8); β-tocopherol (296 nm; 89.4); γ-tocopherol (298 nm; 91.4); δ-tocopherol (298 nm; 87.3); (Franke, Murphy, Lacey, & Custer, 2007). Calibration curves were linear for all tocols

(0.5 to 3.0 µg/mL, R² > 0.99, p < 0.0001, LODmax: 0.12 µg/mL, LOQmax: 0.36 µg/mL) Analyses were performed in triplicate, and results were expressed in percentage variation of γ-Tocopherol during the stability test.

2.5.7 Morphology

Surface morphology of PSO’s microparticles was carried out by scanning electron microscopy (SEM). Samples of microparticles were directly deposited on carbon conductive tape on aluminum SEM stubs, and coated with a thin gold layer, using a gold-sputtering (Desk V, Denton Vacuum®, Moorestown, NJ, US). The samples were analyzed using a Tescan Vega 3 SEM (Tescan®, Kohoutovice, Czech Republic) operated at 15 kV.

2.6 PSO’s microparticles technological properties

PSO’s microparticles technological properties were determined in order to investigate the future application of these potentional functional ingredients in food products.

2.6.1 Hygroscopicity

Hygroscopicity was evaluated according to Al‐Kahtani & Hassan (1990), with some modification. Microparticles (1g) were placed at 25 °C in a desiccator, silica was replaced by NaCl saturated solution in order to submit microparticles to a ~ 75% RH. The gain in weight due to moisture adsorption was recorded at 15 min intervals during 2 h. Results were expressed as g water/ kg dry solids/ min.

163

2.6.2 Solubility

The solubility in water of PSO’s microparticles and components were determined as described by Lai & Cheng (2004) with modifications. Briefly, 1.25 g of sample (dry sample weight), was reconstituted in 15 mL of distilled water by stirring with a magnetic bar at 25 °C for 30 min. The emulsion was centrifuged at 1690.4×g for 30 min; subsequently, a supernatant aliquot (10 mL) was placed in a Petri dish and its weigh was measured after drying at 105 °C for 5 h in an oven (dry supernatant). Solubility in water was calculated by the following equations (Equation 3 and 4): Dry supernatant in 15 mL (g) = dry supernatant in 10 ml (g) × 1.5 [Equation 3] Solubility (%) = (dry supernatant in 15 mL (g)/ dry sample weight (g)) × 100 [Equation 4]

2.6.3 Wettability

Wettability of PSO’s microparticles and components were assessed by a method described by Fuchs et al. (2006), with modifications. Samples (500 mg) were sprinkled over the surface of a 1000 ml Becker containing 950 ml of distilled water at ambient temperature without agitation. The time necessary to microparticles become wet, sediment, sink, or submerse and disappear from the water’s surface was measure and expressed in seconds as wettability time.

2.6.4 Bulk Density

PSO’s microparticles bulk density and components were determined according to Goula & Adamopoulos (2004). Samples (2g) were transferred to a 50 mL graduated cylinder. Packed bulk density was calculated from the weight of power contained in the cylinder after being tapped by hand on a bench 50 times from a height of 10 cm. Results were expressed as g/cm3.

2.6.5 Moisture and water activity (aw)

PSO’s microparticles moisture and aw were assessed as described in section 2.5.1.

164

2.6.6 Thermal stability

A thermogravimetric analyser, from Perkin-Elmer Pyris 1, was used to analyze the thermal stability of the PSO’s microparticles and their components. Samples (5.0 mg) were heated from 25 °C to 800 °C (10 °C/min) under nitrogen gas flow (30 mL/min). Results were expressed as Tonset which is defined as the temperature where the derivative weight curve first starts to rise, indicating the onset point of degradation.

2.7 Application of the PSO’s microparticles in an instant coffee latte drink

2.7.1 Production of the instant coffee latte drink

A product, instant coffee latte drink, was designed to receive the PSO’s microparticles, acting as a vehicle for encapsulated PSO consumption. Previous data demonstrated that addition of 30% of microparticles in relation to coffee mass (m/m; g/g) did not compromise product’s acceptability. The instant coffee latte drink was elaborated according to ingredients’ labels instructions and the following formulations were produced:

a) product control: 200 mL of product was composed of 180 mL of water, 1g of soluble coffee and 26g of powder milk. b) product added with coacervate microparticles (189 mg of total cLnA/ 100 g of product): 200 mL of product was composed of 180 mL of water, 1g of soluble coffee, 0.3 g (51 mg of total cLnA) of coacervate microparticles (30% of soluble coffee mass) and 25.7g of powder milk (milk content was reduced after microparticles addition, in order to maintain all products with equal solids content); c) product added with non-coacervate microparticles (189 mg of total cLnA/ 100 g of product): 200 mL of product was composed of 180 mL of water, 1g of soluble coffee, 0.3 g (51 mg of total cLnA) of non-coacervate microparticles (30% of soluble coffee mass) and 25.7g of powder milk.

2.7.2 Products technological properties

The influence of PSO’s addition on product’s technological properties was assed based in the following parameters: hygroscopicity, solubility, wettability, bulk density, moisture and aw. The described analyses were performed as described in section 2.6.

165

2.8 Statistical analysis

All statistical comparisons were based on triplicate results, and data is presented as mean and standard deviation. Data was tested for normality and homogeneity, respectively, by the Kolmogorov-Smirnov’s test and by the test of Cochran, Hartley, and Bartlett. Comparisons between means were made by t-test or one-way ANOVA with Tukey’s post- test, whenever appropriate. Repeated measures ANOVA with Bonferroni’s post-test was used to investigate time-, formulation- and sealing method-dependent differences in PV, cLnA quantification and isomer profile, volatile compounds and tocopherols content during the stability test. Pearson correlation was applied to investigate associations between microparticles’ tocopherols contents and the following stability parameters: PV, cLnA content and volatile compounds content. Data was analyzed in GraphPad Prism v.6.0 (GraphPad software 2012, Inc, La Jolla, CA, US) and Statistica v.7 (Statsoft Inc. 2004, Tulsa, OK, US). Significance was established at p< 0.05.

3. Results and Discussion

3.1 Physical-chemical characterization

ME and particle size of the PSO’s microparticles (Table 1) were evaluated again in this study because the microencapsulation process might show some variation and these physical-chemical characteristics posses a great influence on microparticles oxidative stability and technological properties such as solubility and wettability. As expected coacervate microparticles showed higher ME, however this difference was not significant (p> 0.05). ME values of the coacervate microparticles were similar of the experimental design (Chapter 3, Table 1). Particle size was influenced by the encapsulation method, as coacervate microparticles showed higher values of D[0.5] and span. This result was expected, because the encapsulation process involves the production of whey protein-gum Arabic insoluble aggregates with higher particles sizes than soluble whey protein and gum arabic. Particle size results of the coacervate microparticles were influenced by a different production batch, as the present microparticles showed higher particle sizes than the ones from the experimental design (Chapter 3, Table 1).

166

Table 1. Microencapsulation efficiency (ME) and particle size results of the coacervate and non-coacervate microparticles (Mean ± SD).

Microparticles ME (%) Particle Size

D[0.5] (µm) span Coacervate 57.7 ± 2.4 11.1 ± 0.5a 2.2 ± 0.1a Non-coacervate 49.9 ± 3.8 9.9 ± 0.4b 2.0 ± 0.1b D[0.5]: maximum size (µm) of 50% analyzed particles; span: particle size scattering index. Different letter indicate significant differences (p< 0.05; t-test)

167

3.2 Stability test

3.2.1 Stability at ambient temperature

Figure 1. PSO’s microparticles stability during 90 days of storage at ambient temperature. Moisture content (%) A); water activity (aw) B); Peroxide value (meq O2/kg oil) C); Volatile compounds content (µg/g of microparticle) D); Total cLnA content (mg/100mg of microparticle) E); γ-Tocopherol variation (%) F). * Significant different from day 0, α indicate significant difference from non- coacervate microparticles at same time. (p< 0.05; repeated measures ANOVA). Coacervate microparticles □) and non-coacervate microparticles ■). 168

Moisture and water activity were used as quality parameters and were evaluated during 90 days at ambient temperature (Figure 1A and B) in order to assess any modification in these parameters that could promote PSO oxidation, compromise microparticles’ technological properties and microbiological safety. During the test period, both microparticles types showed a slight increase in both parameters and finished the study with similar values (p> 0.05) (Coacervate: moisture = 6.58%; aw = 0.41; non-coacervate: moisture

= 5.8%; aw = 0.41). PSO’s microparticles finished this stability test with values expected for microparticles produced by spray drying (moisture ≤ 6% and aw ≤ 0.6) (Reineccius 2004; Klaypradit & Huang 2008).

The Peroxide value (PV) is one of the oldest method and more utilized to evaluate lipid oxidation. This method quantify primary products of lipid oxidation, the hydroperoxides, thus it should only been applied in oils at early oxidation stages, because in highly oxidized samples these lipid oxidation products are decomposed in aldehydes, ketones, acids esters, alcohols, and short chain hydrocarbons, therefore low peroxide value results might be misleading. Additionally, when dealing with conjugated fatty acids, hydroperoxides are not the principal primary oxidation products, as these types of fatty acids demonstrated high polymers formation from the very beginning of the oxidation process (Márquez-Ruiz et al., 2014; Miller & Claxton, 1928). Therefore the evaluation of PV results should always be interpreted in association with other lipid oxidation indicator in order to determine sample’s oxidative status.

PSO’ microparticles peroxide values were evaluated at ambient temperature during 90 days (Figure 1C). Coacervate microparticles showed an increase in the PV up to day 15, after this increase, a small decreased was demonstrated and PV values were kept constant up to day 60. At the end of the study PV increased again and coacervate microparticles finished the study with PV of 36.6 ± 0.8 meq O2/ kg oil. This pattern of oxidation might be related with a heterogeneity in the degree of microparticles’ ME, promoting a different oxidation ratio between particles (Andersen, Risbo, Andersen, & Skibsted, 2000). As well as coacervate microparticles, non-coacervate microparticles’ PV showed an increase up to day 30, which was maintained until day 60 and started to rise again up to day 90. Although non-coacervate microparticles demonstrated higher PV at day 30 and 60, PSO’s microparticles started and finished the test with similar PV (p < 0.05). The influence of the encapsulation method and WM properties on microparticles oxidative stability is well know (Eratte et al., 2014; Jimenez, García, & Beristain, 2006a), however in this storage conditions, we could not notice 169

a significant difference between both encapsulation methods. PV values related in this study were in accordance with Liu, Low, & Nickerson (2010a), which showed variation of ~ 9-27 meq O2/ kg oil in flaxseed microparticles produced by complex coacervation utilizing gelatin-gum Arabic system storage at ambient temperature during 25 days. Moreover, Kaushik et al. (2016) also encapsulated flaxseed oil by complex coacervation utilizing flaxseed gum and protein, demonstrating lower PV when microparticles were storage under lower temperatures (4 °C). In this sense, comparisons with others studies should be made with caution because storage conditions, such as: temperature and relative humidity might show a great influence on microparticles’ oxidative ratio.

Volatile products produced by PSO degradation were another lipid oxidation marker evaluated during the study. Volatile compounds determination by GC is a method capable to identify/quantify secondary lipid oxidation products such as: alcohols, aldehydes, ketones, carboxylic acids, esters and hydrocarbons. This method monitors compounds associated with flavor development in oxidized lipids. Besides a good correlation with flavor scores in sensory analyses, GC analyses also provide a sensitive method to detect low levels of oxidation. This characteristic is very important when dealing with highly unsaturated fatty acids which volatile products shows undesired flavor attributes in very low concentrations (0.04-2.5 ppm) (Frankel, 2005).

All volatile compounds identified in the stability test were described in Table 2. During the stability at temperature ambient were identified a total of 42 volatile compounds, divided in the following chemical classes: esters, aldehydes, alcohols, ketones, carboxylic acids, ethers and hydrocarbons. Coacervate microparticles showed a variation on the volatile fraction composition during the stability test, up to day 30 volatile fraction was composed chiefly by hydrocarbons, esters and aldehydes. After this period until the end of this study, an increase of aldehydes and a decrease of esters was reported (Figure 2A). Based in previous results (PV and cLnA isomer profile and quantification), day 15 was not evaluated because of samples lack of oxidation alteration during this short period. Although identification of volatile compounds by GC-MS has an exploratory character and results are qualitative, we might suggest that the increase of aldehydes fraction is consequence of the oxidation process, as compounds from this chemical class are recognized oxidation markers (García-Martínez et al., 2009; Giua, Blasi, Simonetti, & Cossignani, 2013; Jimenez et al., 2006a). To the best of our knowledge, cLnA oxidation pathway still unclear and there are no specific oxidation markers, however we were able to identified traditional oxidation markers, such as: pentanal, 170

hexanal, 2-hexenal, 2-heptenal, 2-octenal, 3-6-nonadienal and 2,4-nonadienal in samples (Table 2). Nevertheless, we could not associate these results only with PSO oxidation, because WM was analyzed as control, and showed small amounts of pentanal, hexanal, heptanal and 2,4-nonadienal. The occurrence of these volatile compounds in the WM was probably due to the presence of residual milk lipids, which were oxidized. Regarding non- coacervate microparticles, volatile fraction composition was kept fairly stable during the storage (Figure 2B). Moreover it is worth pointing out that volatile compounds previously identified in the raw PSO, such as: butanoic acid, ethyl ester; 1-butanol, 3-methyl-, acetate; acetic acid, hexyl ester; hexanal; 2-hexenal; 2-heptenal; benzaldehyde; 2,4-nonadienal; 1- pentanol and phenylethyl alcohol (Costa et al., 2019), were present in both microparticles, indicating flavor transference to microparticles. From the economical point of view, microparticles’ distinct odor notes could aggregate market value to the product.

In order to overcome limitations associated with the qualitative analysis, volatiles compounds were quantified utilizing GC-MS chromatografic peak area. The quantification analysis of volatile compounds during ambient temperaturature was consistent with the qualitative analysis, as coacervate microparticles showed a significant increase of volatile compounds at the end of the study, whilst non-coacervate microparticles demonstrated a stable volatile compounds quantification troughtout 90 days of storage (Figure 1C). Results are in accordance with Costa et al. (2015) and Jimenez, García, & Beristain (2006), which sucesfully microencapsulated cLA in distinct polymeric matrices and related lower values of volatile compounds during the stability test. In the future, would be interesting to quantify PSO’s microparticles volatile compounds by GC coupled to flame ionization detector aiming at increasing results accuracy. 171

Figure 2. PSO’s microparticles profile of volatile compounds and cLnA isomers during 90 days of storage at ambient temperature. Volatile fraction profile: coacervate microparticles A) and non- coacervate microparticles B). cLnA isomer profile: coacervate microparticles C) and non- coacervate microparticles D).

172

Table 2. Volatile compounds identified in the PSO’ microparticles during 90 days at ambient temperature. Linear Coacervate Non-coacervate RT retention microparticles microparticles Compounds per chemical classes Rmatch LRIRef D*(%) (min) index Time (days) (LRI) 0 30 60 90 0 30 60 90 Ester Ethyl Acetate 909 6.71 X X X X X X sec-Butyl acetate 819 18.4 761.8 7381 3.1 X X X Isobutyl acetate 903 19.51 775.9 8001 3.1 X X X X X X X X Butanoic acid, ethyl ester 904 21.49 801.2 8123 1.4 X X X X X X X X Etil, lactate 820 22.43 814.8 8031 1.4 X X X X X Butanoic acid, 2-methyl-, ethyl 940 24.98 851.7 8531 0.2 X X X X X ester 1-Butanol, 3-methyl-, acetate 921 26.75 877.3 8662 1.3 X X X X X 1-Butanol, 2-methyl-, acetate 902 26.91 879.6 8741 0.6 X Butanoic acid, propyl ester 848 28.16 897.7 8792 2.1 X X X X X X X Acetic acid. hexyl ester 848 35.16 1011.5 10101 0.2 X X X X X X X X Butanoic acid, 3-methylbutyl 857 37.61 1055.0 10464 0.8 X X X X X X X ester Phenethyl acetate 903 48.31 1260.0 12501 0.8 X X X X X X X X Aldehyde Butanal, 3-methyl 937 9.25 X X X X X X X X Pentanal 924 12.96 X X X X X X X X Hexanal 843 21.24 797.8 7942 0.5 X X X X X X X X Furfural 878 23.61 831.8 8293 0.3 X X 2-Hexenal 914 25.1 853.4 8551 0.2 X X 2-Heptenal, (Z)- 950 31.81 956.4 9582 0.2 X X X X X X X X 173

Benzaldehyde 947 32.13 961.6 9613 0 X X X X X X X X 2-Octenal. (E)- 824 37.82 1058.7 10571 0.2 X (Z,Z)-3,6-Nonadienal 841 42.12 1137.6 11001 3.3 X X X X 2,4-Nonadienal, (E,E)- 800 45.11 1194.7 12101 1.3 X X X X X X X X 2,4-Nonadienal, (E,E)- 849 46.14 1215.4 12101 0.4 X X X X X X X X Alcohol 1-Pentanol 868 16.47 737.2 7491 1.6 X X Phenylethyl Alcohol 900 41.00 1116.2 11171 0.1 X X X X X X Ketone Cyclobutanone, 2,2,3-trimethyl- 923 19.09 770.5 7872 2.1 X 2-Hexanone 838 20.63 790.1 7881 0.3 X X X Carboxylic acid

Butanoic acid 863 20.2 784.6 7841 0.1 X X X X X Ether

Propane. 2-methoxy-2-methyl- 900 4.52 X Hydrocarbon Toluene 933 18.69 765.4 7691 0.5 X X X X X X X Styrene 914 27.68 890.7 8821 1 X X X X X X X X Nonane, 3-methyl- 876 32.71 970.9 9701 0 X X X X X X X X 3-Nonene, 3-methyl-, (E)- 814 33.63 985.8 10002 1.4 X X X X X X X Decane 948 34.41 998.4 10002 0.2 X X X X X X X X o-Cymene 941 36.05 1027.3 10201 0.7 X X X X X X X X Benzene, n-butyl- 900 37.88 1059.8 10681 0.8 X X X Decane, 2-methyl- 844 39.52 1088.8 10641 2,3 X X X X Decane, 2,3,4-trimethyl- 892 40.09 1098.9 11212 2 X X X X X X Undecane, 5-methyl- 868 41.67 1129.0 11561 2.4 X X X X X X X 174

Undecane, 3-methyl- 882 43.8 1169.7 11711 0.1 X X X X X X X X Dodecane 910 45.29 1198.1 12001 0.2 X X X X X X X X Tetradecane 907 54.74 1397.8 14001 0.2 X X X X X X X X X: indicate compound presence. * LRI: Linear retention index; LRI Ref: Reference linear retention index (References: 1Ashraf M. El-Sayed (2014). Kovats index. In: The Pherobase: database of insect pheromones and semio chemicals. Retrieved November 10, 2016 from: http://www.pherobase.com/kovats/; 2U.S. Secretary of Commerce (2017). Kovats’RI. In: NIST Standard Reference Data. Retrieved November 10, 2016 from:http://webbook.nist.gov; 3National Center for Biotechnology Information. Kovats Retention Index. In: PubChem Substance Database. Retrieved November 10, 2016 from: https://pubchem.ncbi.nlm.nih.gov; 4Royal Society of Chemistry (2015). Retention Index (Kovats). In: Chemical Structure. Retrieved November 10, 2016 from: http://www.chemspider.com); D(%): difference (%) between experimental LRI and published LRI. 175

During stability test the most important bioactive compound present in the PSO’s microparticles was determined, cLnA isomers. PSO’s microparticles showed a stable and similar (p< 0.05) total cLnA concentration until day 60 at ambient temperature, thereafter coacervate microparticles’ total cLnA concentration decreased 39% up to day 90, while non- coacervate microparticles kept total cLnA content values constant throughout the study (Figure 1E). Previous studies (Goula & Adamopoulos, 2012; Sahin-Nadeem & Afşin Özen, 2014) which encapsulated cLnA did not evaluate these fatty acid during stability, however a similar conjugated fatty acid, cLA, had demonstrated a 2 months stability at ambient temperature, when encapsulated by spray drying (Costa et al., 2015). The higher oxidative stability of non-coacervate microparticles was in accordance with volatile fraction quantification analysis. Besides that, an increase of volatile compounds and a decrease of total cLnA content were reported at day 90 in coacervate microparticles, suggesting that cLnA degradation might be consequence of an increase of lipid peroxidation. To the best of our knowledge there is no data regarding the storage stability of microencapsulated cLnA and according to Yang et al. (2009) among highly unstable fatty acids such as: linoleic acid, cLA and α-linolenic; cLnA was the most oxidizable of them all. In this sense, results found in this work were very encouraging, because the encapsulation processes were capable to stabilize a highly unstable bioactive fatty acid for at least 2 months.

cLnA bioactivity is isomer depend (Suzuki et al., 2001), therefore microparticles’s isomer profile was monitored throughout the study. Nevertheless it is worth to point it out that the microencapsulation process promotes a small but significant degradation of cLnA and isomerization, as demonstrated by our previous results (Chapter 3). Consistently with total cLnA quantification analysis, coacervate microparticles showed a fairly stable cLnA isomers profile during 60 days, and after this period, a significant decrease of punicic acid was demonstrated (Figure 2C). On the other hand, non-coacervate microparticles demonstrated a stable cLnA isomer profile during the entire study period, suggesting a better protection against oxidation provided by the spray drying encapsulation (Figure 2D).

PSO is a rich source of tocopherols, specially γ-Tocopherol. Previous data (Costa et al., 2019) showed that the commercial PSO used in this study is composed by high contents of γ-Tocopherol (1125.8 mg/ 100 g oil) and marginal contents of α- and δ-Tocopherol. In this context, during the stability test only γ-Tocopherol was monitored and, accordingly to volatile fraction analysis, day 15 was not evaluated. PSO’s microparticles showed similar variation of 176

γ-Tocopherol during stability, indicated by comparable values of γ-Tocopherol at day 0, 30 and 90 (p > 0.05) (Figure 1F). However it is important to empathize that coacervate microparticles showed a sharper decrease of γ-Tocopherol between day 30 and day 60, while non-coacervate microparticles kept this antioxidant concentration during the aforementioned period. The higher rate of γ-Tocopherol degradation observed in coacervate samples around day 60 is in accordance with the following results: total cLnA content decrease at day 90, aldehydes accumulation increase from the 60th day and total volatile compounds content at day 90 also related in coacervate microparticles. Tocopherols are capable to interrupt lipid autoxidation by acting as a chain-breaking antioxidant either in initiation or propagation step, additionally this antioxidant can also act in the decomposition step. Therefore an increase in lipid oxidation is expected after tocopherols depletion (Frankel, 2005). García-Martínez et al., (2009) and Márquez-Ruiz, García-Martínez, Holgado, & Velasco (2014) showed that tocopherols exhaustion precedes an increase in oxidation rate when working with cLA, confirming the previous hypothesis.

Aiming at to verify the associations related between tocopherols content and other stability parameters (PV, cLnA content and total volatile compounds) correlation analysis were run (Figure 3). Tocopherols contents showed a fair correlation with cLnA content (r = 0.7195) (Figure 3A), PV (r = 0.7939) (Figure 3B) and volatile compounds (r = 0.6910) (Figure 3C). One of the few articles (Miller & Claxton, 1928) that evaluated cLnA isomers oxidation showed that hydroperoxides were not the principal primary oxidation products, contrasting with our results which PV demonstrated the higher correlation with tocopherols content. cLnA oxidation still an unexplored subject and divergent results, indicates that further investigations regarding this theme should be performed. Nevertheless, the present study provides a reliable data that suggest an import protection role of tocopherols in the oxidation process.

177

Figure 3. Correlation between tocopherols content and oxidative indicators during stability test at ambient temperature. cLnA content and tocopherols content A); Peroxide value and tocopherols content B); Volatile compounds content and tocopherols content C) (Pearson’s correlation). 178

Non-coacervate microparticles showed a similar or superior oxidative stability than coacervate microparticles in all parameter evaluated during stability test. Microparticles’ non- encapsulated oil content which is free for O2 interaction did not easily explain the described results, as both samples have similar ME (Table 1), thus we proposed that this difference might be associated with WM porosity and physical integrity.

Surface morphology of the PSO’s microparticles during the stability test is shown in Figures 4 and 5. The majority of the formulations presented spherical shape and variable sizes, accordingly with particles produced by spray drying, as it resemble the initial droplet formed in the atomization step. PSO microparticles’ surface topography was characterized by intense invaginations, due to the shrinkage of the particles during the drying process after rapid evaporation of water (Ré, 1998). These results are in agreement with previous studies (Costa et al., 2015; Eratte et al., 2015, 2014; Tonon, Grosso, et al., 2011; Yang et al., 2015) which microencapsulated edible oils by complex coacervation or spray drying using WM composed by proteins and polysaccharides. At ambient temperature, coacervate microparticles started the stability test with microparticles showing spherical-shaped morphology with intense invaginations and smooth surface free of visible cracks and pores (Figure 4A and B). These described microparticles’ morphological characteristics were kept constant until day 60 (Figure 4C and D). In the end of the study, fragments of microparticles were observed, suggesting particle’s collapsation and consequent reduction of PSO protection, demonstrated by an increase of lipid peroxidation, previously showed by: an decrease of γ-Tocopherol and cLnA; and a total volatile compounds content increase.

Regarding non-coacervate microparticles, these types of microparticles also started the study with good morphological characteristics, such as: spherical-shaped morphology with intense invaginations and smooth surface free of visible cracks and pores (Figure 5A and B). However at the end of the study agglomeration problems were reported (Figure 5C and D), probably because of WM hygroscopicity, and differently from coacervate microparticles these morphological alterations were not associate with lipid oxidation increase. According to Beristain, Azuara, & Vernon-Carter (2002) as long water is absorbed without initiating WM dissolution process, the adsorbed water might promote formation of a gel-like structure capable to seal surface pores and increase protection against oxygen diffusion. We propose that the coacervation process reduced microparticles’ hygroscopicity, as the coacervation process reorganizes WPI and GA structure, producing insoluble microparticles. In this sense, coacervate microparticles should be storage in an optimum RH in order to avoid open pores 179

and cracks promoted by microparticles dehydration. However this hypothesis needs further investigation. Furthermore, is important emphasize that PSO is a highly oxidizable oil, as result, any lack of core protection by wall system, might initiate core degradation.

The retention of core materials and their protection in a microencapsulated product are related to the porosity and degree of integrity of microcapsules. In this sense, microparticles’ morphology data were successfully applied in the investigation of PSO’s microparticles distinct oxidation pattern. Additionally aiming at a more complete evaluation of microparticles structure, samples porosity was also evaluated by Brunauer-Emmett-Teller (BET) method, however microparticles' sorption isotherms did not show adequate results and results were not displayed.

180

Figure 4. Coacervate microparticles’ structure during storage under ambient temperature. Time 0: A) and B); Time 60 days: C) and D); and Time 90 days: E) and F). Bar 20 µm: A), C) and E); bar 10 µm: B), D) and F). White arrows indicate microparticles fragments.

181

Figure 5. Non-coacervate microparticles’s structure during storage under ambient temperature. Time 0: A) and B); Time 90 days: C) and D). Bar 20 µm: A), C); bar 10 µm: B), D).

182

3.2.2 Stability at 60°C

Figure 6. PSO’s microparticles stability during 10 days of storage at 60°C. Moisture content (%) A); water activity (aw) B); Peroxide value (meq O2/kg oil) C); Volatile compounds content (µg/g of microparticle) D) and Total cLnA content (mg/100mg of microparticle) E).* Significant different from day 0, α indicate significant difference from non-coacervate microparticles at same time; β indicate significant difference between the sealing method (under O2 vs under vaccum) in the same treatment and at same time (p < 0.05; repeated measures ANOVA). 183

In a more drastic condition, 60 °C for 10 days, microparticles moisture content and aw (Figure 6A and B) were evaluated in two sealing condition. A decrease in moisture content and aw was showed in both types of microparticles and in both types of sealing. All treatments finished the test with similar results (p > 0.05) (moisture ~ 3.9% and aw ~ 0.104), however samples sealed under vacuum (t = 5 day) reached equilibrium before microparticles sealed without vacuum (t = 10 day).

Coacervate microparticles showed an increase in PV up to day 5 and a decrease right after. Concerning non-coacervate microparticles, PV increased up to day 5 and kept constant until the end of the study. At this condition, microencapsulation method influenced microparticles oxidative stability, because non-coacervate microparticles showed a better oxidative stability indicated by lower values of PV. Moreover, we suggested that the inferior PV demonstrated by coacervate microparticles at day 10, was associated with an advanced oxidative stage where hydroperoxides decomposition were greater than production (Frankel, 2005). Kim & Kim (2014) evaluated the oxidative stability of microencapsulated sunflower oil in similar conditions of this study (60 °C) during 30 days and related comparable PV results. The present work is also in agreement with Wang, Shi, & Han (2018) which microencapsulated by spray drying a highly oxidizable oil composed by 40% of α-linolenic acid (peony seed oil) and showed PV result around 50 meq O2/ kg oil after storage at 60 °C during 18 days. Regarding the sealing method, as expected the abstraction of O2, an initiator of the oxidation reaction, reduced the oxidation ratio up to day 5 in coacervate microparticles and until the end of the study in non-coacervate microparticles. This result is in agreement with our suggestion that the non-coacervate microparticles posses a better oxidative stability.

GC-MS analysis identified a total of 60 volatile compounds divided in the following chemical classes: esters, aldehydes, alcohols, ketones, carboxylic acids, ethers and hydrocarbons (Table 3). As expected higher temperature increased oil oxidation, confirming the hypothesis reported at ambient temperature that showed an association between aldehydes accumulation and lipid oxidation (Figure 7A and C). Concerning the sealing method, storage without or under vacuum (reduced O2 concentration) seems not to influence microparticles oxidation, as samples showed similar volatile fraction composition (Figure 7A vs B; and C vs D). Nevertheless these results should be interpreted with caution, because of data’s qualitative characteristic and for proper volatile compounds quantification comparisons, statistical analysis should be applied. In this sense, as described before, the identified volatile compounds were tentatively quantified in order to overcome the qualitative analysis 184

limitations. Coacervate microparticles showed an increase of volatile compounds up to day 5, them microparticles’ volatile content was kept constant troughtout the study (Figure 6D). While non-coacervate microparticles demonstrated a constant increase of volatile content during the study (Figure 6D). Based on previous results that showed a higher oxidative stability of non-coacervate microparticles, we suggest that the lower content of volatile compounds observed at day 10 in coacervate microparticles, might be associated with degradation of aldehydes and hydrocarbons in products that were not quantified. However this hipothesis needs futher investigation. Concerning the sealing condition, unexpectedly storage under vacuum was not capable to reduce volatile compounds concentration during 10 days, except for non-coacervate microparticles at day 10 (Figure 6D). The present study results is in accordance with Carneiro, Tonon, Grosso, & Hubinger (2013) which microencapsulated by spray drying flaxseed oil using different combinations of wall materials and reported accumulation of propanal and hexanal around of 900 and 450 µg of compound/ g oil, respectively, after storage at 45 °C during 4 weeks.

185

Table 3. Volatile compounds identified in the PSO’ microparticles during 10 days at 60 °C sealed in atmospheric condition or with reduction of O2. Linear Coacervate microparticles Non-coacervate microparticles Compounds per chemical RT retention Time (days) Rmatch LRIRef D*(%) classes (min) index 5 10 5 10 0 0 (LRI) O2 vacuum O2 vacuum O2 vacuum O2 vacuum Ester Ethyl Acetate 909 6.71 X X sec-Butyl acetate 819 18.4 761.8 7381 3.1 X X Isobutyl acetate 903 19.51 775.9 8001 3.1 X X Butanoic acid, ethyl 904 21.49 801.2 8123 1.4 X X X X X X X ester Etil, lactate 820 22.43 814.8 8031 1.4 X X Butanoic acid, 2- 940 24.98 851.7 8531 0.2 X X X X methyl-, ethyl ester 1-Butanol, 3-methyl-, 921 26.75 877.3 8662 1.3 X X X X X X acetate 2-n-Butyl furan 914 27.68 890.7 8891 0.2 X X X X X X X X Butanoic acid, propyl 848 28.16 897.7 8792 2.1 X X ester Acetic acid, hexyl ester 848 35.16 1011.5 10101 0.2 X X X Butanoic acid, 3- 857 37.61 1055.0 10464 0.8 X X methylbutyl ester Phenethyl acetate 903 48.31 1260.0 12501 0.8 X X X X X Aldehyde Butanal 909 5.44 X X X X X X X X Butanal, 3-methyl 937 9.25 X X X X X Pentanal 924 12.96 X X X X X X X X X X 186

2-Pentenal, (E)- 894 17.83 754.5 7431 1.5 X X X X X X 2-Hexenal, (E)- 857 18.48 762.8 7511 1.5 X X 2-Butenal, 2-ethyl 853 18.49 762.9 7912 3.7 X X X X X Hexanal 843 21.24 797.8 7942 0.5 X X X X X X X X X 2-Pentenal, 2-methyl 898 22.69 818.5 8181 0.1 X X X X X X X X Furfural 878 23.61 831.8 8293 0.3 X X X X X X X X 2-Hexenal 914 25.1 853.4 8551 0.2 X X X X X X X X 4,4-Dimethylpent-2- 868 28.96 910.3 8812 3.2 X X X X enal 2-Heptenal, (Z)- 950 31.81 956.4 9582 0.2 X X X X X X X X X X Benzaldehyde 947 32.13 961.6 9613 0 X X X X X X X X X X trans-trans-2,4- 857 35.12 1010.8 10121 0.1 X X Heptadienal 2-Heptenal, 2-methyl- 888 36.34 1032.4 9902 4.1 X X X X X X X 2-Octenal, (E)- 824 37.82 1058.7 10591 0 X X X X X X X X 2,4-Octadienal, (E,E)- 903 40.7 1110.5 10872 2.1 X X X X X X X 2-Heptenal, 2-propyl- 939 44.76 1188.0 11684 1.7 X X X X X X X X 2,4-Nonadienal, (E,E)- 800 45.11 1194.7 12101 1.3 X X X X X X X X X X 2,4-Nonadienal, (E,E)- 849 46.14 1215.4 12101 0.4 X X X X X X X X X X 3-Cyclohex-1-enyl- 830 46.65 1225.9 11862 3.3 X X X X X X X prop-2-enal 2-Decenal, (E)- 907 48.48 1263.4 12631 0 X X X X X X X X 2,4-Decadienal 920 51.17 1319.9 13261 0.5 X X X X X X X X Alcohol 1-Pentanol 868 16.47 737.2 7491 1.6 X X X X Cyclopentanol, 3- 886 24.49 844.6 8492 0.2 X X X X X X methyl- 1-Octen-3-ol 857 33.21 979.0 9811 0.2 X X X X X X X X Phenylethyl Alcohol 900 41 1116.2 11171 0.1 X X X X X X X X X 187

Ketone 2-Hexanone 838 20.63 790.1 7881 0.3 X X X X X X X 5-Nonanone 890 38.68 1073.9 10731 0.1 X X X X X X X X 5-Decanone 900 43.99 1173.3 11551 1.6 X X X X Carboxylic acid Butanoic acid 863 20.2 784.6 7841 0.1 X X X X Hexanoic acid 826 32.85 973.2 9501 2.4 X X X X Ether Propane, 2-methoxy-2- 900 4.52 X methyl- 2-Pentylfuran 906 34.03 992.2 9891 0.3 X X X X X X X X Hydrocarbon Toluene 933 18.69 765.4 7691 0.5 X X X X Styrene 914 27.68 890.7 8821 1 X X Nonane 910 28.29 899.6 9001 0 X X X X X X X X Nonane, 3-methyl- 876 32.71 970.9 9701 0 X X 3-Nonene, 3-methyl-, 814 33.63 985.8 10002 1.4 X X X X X (E)- Decane 948 34.41 998.4 10002 0.2 X X X X X X X X X X o-Cymene 941 36.05 1027.3 10201 0.7 X X X X X X X X X Decane, 2-methyl- 844 39.52 1088.8 10641 2.3 X X Decane, 2,3,4- 892 40.09 1098.9 11212 2 X X X trimethyl- Undecane, 5-methyl- 868 41.67 1129.0 11561 2.4 X X Undecane, 3-methyl- 882 43.8 1169.7 11711 0.1 X X Dodecane 910 45.29 1198.1 12001 0.2 X X Tetradecane 907 54.74 1397.8 14001 0.2 X X 188

X: indicate compound presence* LRI: Linear retention index; LRI Ref: Reference linear retention index (References: 1Ashraf M. El-Sayed (2014). Kovats index. In: The Pherobase: database of insect pheromones and semio chemicals. Retrieved November 10, 2016 from: http://www.pherobase.com/kovats/; 2U.S. Secretary of Commerce (2017). Kovats’RI. In: NIST Standard Reference Data. Retrieved November 10, 2016 from:http://webbook.nist.gov; 3National Center for Biotechnology Information. Kovats Retention Index. In: PubChem Substance Database. Retrieved November 10, 2016 from: https://pubchem.ncbi.nlm.nih.gov; 4Royal Society of Chemistry (2015). Retention Index (Kovats). In: Chemical Structure. Retrieved November 10, 2016 from: http://www.chemspider.com); D(%): difference (%) between experimental LRI and published LRI. 189

Figure 7. PSO’s microparticles profile of volatile compounds and cLnA isomers during 10 days of storage at 60 °C. Volatile fraction profile: coacervate microparticles A); coacervate microparticles under vacuum B); non-coacervate microparticles C) and non-coacervate microparticles under vacuum D). cLnA isomer profile: coacervate microparticles E); coacervate microparticles under vacuum F); non-coacervate microparticles G) and non-coacervate microparticles under vacuum H). 190

Higher temperatures accelerated total cLnA degradation and both types of PSO’s microparticles finished the study with around 15% of the initial cLnA content (Figure 6E). In this storage condition, the encapsulation method did not influence samples’ oxidation pattern, as PSO’s microparticles showed similar (p > 0.05) total cLnA contents throughout the study. This lack of difference between encapsulation methods might be associated with the higher oxidation ratio in this condition which promotes fast cLnA degradation, hindering the evaluation of any noticeable difference in 10 days with 2 analyzing times, 5 days apart. Nevertheless, this hypothesis needs confirmation in the future, utilizing a higher number of analyzing times. Concerning the sealing condition, O2 reduction during the stability test slowed cLnA degradation in all analyzed times in non-coacervate microparticles and coacervate microparticles only in day 10 (Figure 6E). Based in our results, the produced PSO’s microparticles already have an appealing oxidative stability of at least two months, comparable with other microparticles composed by similar conjugated bioactive fatty acid (cLA) (Costa et al., 2015; Jimenez et al., 2006b), and we suggest that the association with storage under vacuum (reduced O2 atmosphere) might extend even more microparticles shelf life.

Isomer profile variation during stability was also assessed in this condition. The stability test at a higher temperature (60 °C) was capable to promote degradation of punicic acid and a significant isomerization, indicated by an increase of α/β-eleostearic acid and catalpic acid (p < 0.05) (Figure 7E, F, G and H). The degradation of punicic acid was expected, because this cLnA isomer is the most unstable against oxidation (Yang et al., 2009). Additionally, after a heat treatment or storage under ambient or low temperatures conjugated fatty acids isomers might isomerizate to trans- configuration because its higher thermal stability (Giua, Blasi, Simonetti, & Cossignani, 2012; Rodríguez-Alcalá & Fontecha, 2007; Sahin-Nadeem & Afşin Özen, 2014). Unlike ambient temperature, the content of the major cLnA isomer, punicic acid, during the storage at 60 °C was not influenced by the encapsulation method, as both types of PSO’s microparticles started and finished the test with similar results (p > 0.05) (Figure 6E and G). Regarding the sealing method, accordingly with total cLnA contents results, the abstraction of O2, reduced the degradation of punicic acid in both PSO’s microparticles (Figure 7E vs F and G vs H).

Samples’ γ-Tocopherol contents were not presented, because in this condition the antioxidant was completely degraded. 191

PSO’s microparticles scanning electron microscopy images were displayed in Figure 8 and 9. Major morphological alterations observed were consistent with samples higher oxidation ratios related in this condition. Coacervate microparticles showed spherical-shaped morphology with intense invaginations, but in this condition were reported high contents of particles fragments and open pores at day 5 (Figure 8C and D) and day 10 (Figure 8E, F, G, H and I). Probably the high temperature stimulated microparticles dehydration, accelerating the formation of cracks and pores in the microparticles. This same trend was demonstrated in microparticles stored in lower water activity (Jimenez et al., 2006b). Different from non- coacervate microparticles, the coacervation process reorganize WPI and GA structure, reducing microparticles hygroscopicity. On the other hand, non-coacervate microparticles did not show apparent cracks and open pores, however, as demonstrated at ambient temperature, microparticles showed noticeable agglomeration problems at day 10 (Figure 9G, H and I). We suggest that this sample did not have dehydration problems, because of the WM hygroscopicity. Moreover, based on the qualitative characteristic of the MEV analysis, we could not perform reliable comparisons between sealing methods and therefore the influence of this parameter on microparticles’ morphology was not assessed.

192

Figure 8. Coacervate microparticles’s structure during storage under 60 °C. Time 0: A) and B); Time 5 days: sealed C) and sealed under vacuum D); and Time 10 days: sealed E), F) and sealed under vacuum G), H), I). Bar 20 µm: A), D), E) and G); bar 10 µm: B) and H) and bar 2 µm: F) and I). White arrows indicate microparticles fragments and open pores.

193

Figure 9. Non-coacervate microparticles’s structure during storage under 60 °C. Time 0: A) and B); Time 5 days: sealed C), D) and sealed under vacuum E), F); and Time 10 days: sealed G), H) and sealed under vacuum I). Bar 20 µm: A), C), E), G) and I); bar 10 µm: B), D), F) and H).

194

Non-coacervate microparticles also showed a superior oxidative stability in a more drastic condition and the evaluation of microparticles surface morphology was successfully applied to justify microparticles distinct oxidative pattern. Moreover, O2 had a great influence in microparticles’ lipid peroxidation, although this trend could not be investigated by morphological analysis. Nevertheless these results should be evaluated with caution, as accelerated tests could be misleading. According to Silva et al. (1999) the oxidation mechanisms involved in high temperatures are different than in ambient temperature, therefore results might not reflect oil oxidation pattern at room temperature.

Regardless non-coacervate microparticles higher oxidative stability, PSO encapsulation by complex coacervation, an encapsulation technique not applied on PSO yet, was considered successfully as microparticles showed an overall stability of 60 days.

Moreover, we demonstrated that storage under reduced O2 concentration could delay oxidation process initiation. Therefore we suggest that storage under this condition might extent microparticles shelf life at ambient temperature.

After evaluating PSO’s microparticles oxidative stability, the next step in order to propose a vehicle for PSO consumption. Microparticles technological properties, such as: hygroscopicity, solubility, wettability, bulk density, moisture, water activity (aw) and thermal stability, were assessed, aiming at developing a product suitable for microparticles addition.

3.3 Technological properties

3.3.1 Hygroscopicity

Hygroscopicity is the capacity to absorb ambient moisture. Particles hygroscopicity is influenced by water concentration gradient between sample and the surrounding, therefore less moist powders shows high hygroscopicity values. Microparticles’ moisture is influenced by drying conditions, such as: air inlet temperature, feed rate and air flow and feed emulsion composition (Goula & Adamopoulos, 2012). Although was expected a higher hygroscopicity of non-coacervate microparticles, this was not demonstrated. However based in MEV analysis, it is suggested that storage time has a great influence in this parameter. (Table 4). These results are in agreement with Tonon, Freitas, & Hubinger (2011) which microencapsulated açai juice in different WM composed by maltodextrin and GA and showed results ranging from 0.066-0.280 g water/ kg dry solids/ min. Microparticles water adsorption 195

should be avoided, as this process could dilute microparticles WM and by particles structure degradation expose PSO to ambient oxidants. Moreover high water adsorbed concentrations could promote agglomeration, caking and microorganism proliferation. In this sense, in order to preserve microparticles oxidative and physical-chemical stability and microbiological safety, samples should be storage in relative humidity similar to their water content or under vacuum.

Coacervate microparticles lower oxidative stability was justified by samples lower hygroscopicity, however microparticles’ hygroscopicity analysis showed that the complex coacervation did not reduce hygroscopicity. In this sense, we suggest that this parameter might be influenced by time, as microparticles demonstrated similar morphological features in the beginning of the study and after the day 60, coacervate microparticles started to show physical integrity problems and non-coacervate microparticles, agglomeration. Nevertheless, this hypothesis needs further investigation.

3.3.2 Solubility

Microparticles are composed by highly soluble WM, WPI (94.9%) and GA (90.9%) (Choi, Ryu, Kwak, & Ko, 2010; Yousefi, Emam-Djomeh, & Mousavi, 2011) and therefore coacervate microparticles showed high solubility values (Table 4). These results are in accordance with previous studies (Choi et al., 2010; A M M Costa et al., 2015; Sahin-Nadeem & Özen, 2014) which microencapsulated lipophilic nutrients and observed solubility results varying from 22.4-97%. However when comparing non-coacervate microparticles (82.3%) and WM non-coacervate (95.5%) its observed a small but significant solubility reduction, probably associated by oil addition. Additionally, the coacervation process involves the production of insoluble microparticles that are consequence of WM components reconformation and analogous to oil addition, when comparing WM coacervate (85.4%) and non-coacervate (95.5%), the coacervate process promoted a small but significant decrease in this parameter. In this sense, as both particles showed similar (p > 0.05) solubility results, we suggest that the oil addition has a greater influence in this parameter. It s worth point it out that these high solubility values related broaden microparticles applicability spectrum, enabling the development of a functional beverage.

196

3.3.3 Wettability

The ability of powders to mix or absorb with water, is one of the most important physical properties associated with powders rehydration and instantization (De Barros Fernandes et al., 2014). In the present study, PSO’s microparticles necessary time to become completely wet varied from 750-1378.5 sec. Wettability is influenced by powder particle size, bulk density, porosity, surface charge and presence of hydrophobic components on microparticles surface (Bae & Lee, 2008; De Barros Fernandes et al., 2014; Vega & Roos, 2006). Accordingly was observed a decrease of wettability when the PSO was added, a highly hydrophobic substance, indicated by an increase (p < 0.05) of the time necessary to microparticles become completely wet, comparing non-coacervate microparticles and its WM (750 sec vs 519.5 sec) (Table 4). These results were consistent with Bae & Lee (2008) and De Barros Fernandes, Marques, Borges, & Botrel (2014) which showed this similar trend. On the other hand, in this study was not demonstrated influence of microparticles’ bulk density on wettability, as described by Bae & Lee (2008), probably by microparticles comparables (p > 0.05) bulk densities (Table 4). Moreover coacervate microparticles showed the lowest wettability values (1378.5 sec), suggesting that the coacervation process was the factor that influenced the most this parameter. As described before, the coacervation process produces insoluble microparticles, due to WM structural reorganization induced by pH reduction, which promotes the exposure of hydrophobic structures from WPI and GA on the microparticles surface. However this is a physical process, therefore when microparticles are submitted to a pH higher than the coacervate pH (3.75), as was used in this experiment (~7), WM gradually turn into its original configuration and microparticles started dilute in water, explaining the displayed similar solubility results. Coacervate microparticles lower wettability does not hinder its application in beverages, it just indicate the necessity of a homogenization step in order to accelerate microparticles dilution.

3.3.4 Bulk density

Bulk density is an important factor related to microparticles storage and commercialization logistics. The same mass can be stored in a smaller container, when comparing a higher density powder with a lower density powder. Therefore, microparticles showing high bulk density are more economic appealing, in order to reduce costs related with product’s storage. Microparticles’ bulk densities varied from 0.310-0.354 g/cm3 (Table 4). Results were in accordance with Bae & Lee (2008), Choi et al., (2010) and Tonon, Pedro, 197

Grosso, & Hubinger (2012), they produced microparticles by spray drying and showed bulk density values of 0.250 to 0.528 g/cm3. In this study, the complex coacervation process did not influence microparticles’ bulk density, because both types of microparticles and its components (WM coacervate/ WM non-coacervate) showed similar (p > 0.05) results (Table 4). Bulk density is influenced by drying conditions, feed emulsion composition and particle size (De Barros Fernandes et al., 2014). Timilsena, Adhikari, Barrow, & Adhikari, (2016) demonstrated the influence of microparticles shapes and sizes on bulk density, as smaller and spherical particles fits more compactly in the cylinder and consequently shows higher bulk density. In the present study, coacervate and non-coacervate microparticles displayed different particles sizes results (Table 1), but particles had similar shapes (Figure 4A and 8A). Moreover, regarding drying conditions and feed emulsion composition, both types of microparticles showed equal values in the aforementioned parameters, thus the reported results indicated that drying conditions and feed emulsion composition had a higher influence on microparticles’ bulk density than particle size.

3.3.5 Moisture and Water activity (aw)

PSO’s microparticles moisture and aw results have already been discussed in section

3.2, in this sense these results were not discussed again. PSO’s microparticles moisture and aw values were displayed in Table 4 in order to facilitate comparisons with the developed food product.

3.3.6 Thermal stability

PSO’s microparticles and its components thermal stability were assessed by thermogravimetric analysis (TGA) (Table 4). Non-coacervate microparticles (290.4 °C) and its components (WM non-coacervate = 288.4 °C, whey protein = 286.2 °C and Gum arabic = 295.1 °C) showed higher thermal stability than coacervate microparticles (242.3 °C) and WM coacervate (246.6 °C), indicating an influence of the coacervation process on microparticles thermal stability. Huang, Sun, Xiao, & Yang (2012) demonstrated distinct results studding the complex coacervation of soybean protein isolate and chitosan. In the aforementioned work, was described a small increase (7 °C) in protein thermal stability after interaction with chitosan in the coacervate aggregate. However, according to Vogt, Woell, & Argos (1997), protein thermal stability shows a direct association with hydrogen bonds interactions. Therefore we propose that the thermal stability decrease reported, is explained by WM 198

components structural reconfiguration deflagrated by pH reduction, reducing hydrogen bonds and increasing hydrophobic interaction. TGA results provided awareness concerning the possible microparticles’ applications. In this sense, we might suggest that coacervate microparticles could be added in food products thermally processed up to 242.3 °C, because it WMs is stable until the aforementioned temperature.

The technological analyses showed the great potential and versatility of the coacervate microparticles as a potentional functional ingredient. Microparticles’ high solubility and thermal stability enables its application in a wide range of food product.

3.4 Instant coffee latte drink added with PSO’s microparticles

An instant coffee latte drink was designed to receive the PSO’s microparticles. This type of beverage with a mild thermal processing during production was suitable for PSO’s microparticles applications due to its high solubility and thermal stability. Furthermore, the food matrix possess a high penetration in the Brazilian population food habit.

Product’s oxidative and physical-chemical stability, microbiological safety and even consumer acceptance is highly influenced by its technological properties. In this sense, the influence of PSO’s microparticles addition on the developed beverage was assessed.

3.4.1 Hygroscopicity

Coacervate microparticles addition did not affected product’s hygroscopicity (coacervate = 0.419 g water/ kg dry solids/ min vs control = 0.447 g water/ kg dry solids/ min) (Table 4). Although the control presented a higher content of lactose, an extremely hygroscopic carbohydrate which contents in whole milk powder accounts up to 36.6 wt% (Kim, Chen, & Pearce, 2002), the 0.3 g reduction of whole milk powder in coacervate microparticles did not promoted a product hygroscopic reduction. Additionally, the encapsulation method did not influence this parameter, as coacervate and non-coacervate microparticles addition products showed similar (p > 0.05) hygroscopicity values (Table 4). This result was expected as both types of microparticles demonstrated comparable (p > 0.05) hygroscopicity.

199

3.4.2 Solubility

PSO’s microparticles addition did not affect (p > 0.05) product solubility (Table 4), confirming our previous statement that the developed microparticles were suitable for drinks and beverages applications.

3.4.3 Wettability

As demonstrated previous (section 3.3) coacervate microparticles showed a lower wettability than non-coacervate microparticles, thus was expected that coacervate microparticles addition would affect product wettability. However, product with coacervate microparticles were not different (p > 0.05) from the control, indicating that microparticles addition of 30% did not influence this parameter. Maybe in a higher concentration, PSO’s microparticles would affect the product wettability. However this hypothesis needs further investigation. The wetting times of the developed functional coffee latte drink were within the range of 7-161 sec demonstrated in other instant products such as: milk powder, cocoa powder beverage, soy milk and green tea powder (Jinapong, Suphantharika, & Jamnong, 2008; Kim et al., 2002; Park, Imm, & Ku, 2001; Shittu & Lawal, 2007).

3.4.4 Bulk density

PSO’s microparticles addition did not affect product bulk density, as products added with the coacervate microparticles (0.514 g/cm3) showed similar (p > 0.05) results with the control (0.502 g/cm3) (Table 4). Additionally, the complex coacervation process did not influence this parameter, as coacervate and non-coacervate microparticles addition products showed similar (p > 0.05) bulk density values (Table 4). The developed functional coffee latte drink displayed bulk density results comparable with other instant products, such as milk powders (0.433-0.646 g/cm3) and yacon juice (0.47-0.69 g/cm3) (Fitzpatrick, Iqbal, Delaney, Twomey, & Keogh, 2004; Franco, Perussello, Ellendersen, & Masson, 2016).

3.4.5 Moisture and water activity (aw)

Coacervate microparticles addition promoted a small but significant increase in the product moisture content (coacervate = 4.2% vs control = 3.6%) (Table 4). Besides that, the encapsulation method affected the product final moisture. This result was not expected as PSO’s microparticles showed similar moisture (p > 0.05) (coacervate = 5.3%; non-coacervate 200

= 5.2%) (Table 4). Products’ moisture was similar with instant milk (2-4.7%) (Fitzpatrick et al., 2004; Kim et al., 2002)

Product’s aw was not affected (p > 0.05) by microparticles addition, although for PSO’s microparticles (coacervate = 0.340; non-coacervate = 0.305) (Table 4), this parameter was influenced by the encapsulation method. We propose that the lack of difference described in the developed functional coffee latte drink was associated with microparticles concentration addition (30%), maybe in higher concentrations we would observe differences.

Products’ aw was situated in the range of aw (0.2-0.3), which lipid oxidation reactions velocities are reduced (Silva et al., 1999).

201

Table 4. PSO’s microparticles and instant coffee latte drink added with PSO’s microparticles technological properties (hygroscopicity, solubility, wettability, bulk density, moisture water activity (aw) and thermal stability). (mean ± SD).

Hygroscopicity Solubility (%) Wettability (sec) Bulk density Moisture (%) aw Thermal (g water/ kg dry (g/cm3) stability solids/ min) (Tonset; °C) PSO’s microparticles Coacervate 0.240 ± 0.02a 77.9 ± 0.0a 1378.5 ± 12.0a 0.310 ± 0.015ab 5.3 ± 0.4a 0.340 ± 0.00a 242.3 ± 3.6c Non-coacervate 0.229 ± 0.00a 82.3 ± 2.0a 750 ± 14.1b 0.354 ± 0.004ac 5.2 ± 0.0a 0.305 ± 0.01b 290.4 ± 7.6b WM coacervate NA 85.4 ± 1.7a 900.5 ± 4.9c 0.262 ± 0.007b NA NA 246.6 ± 2.7c WM non-coacervate NA 95.5 ± 0.6b 519.5 ± 3.5d 0.307 ± 0.006ab NA NA 288.4 ± 1.2b Whey protein NA 94.9 ± 1.4b 7.7 ± 0.2e 0.377 ± 0.014c NA NA 286.2 ± 4.3b Gum arabic NA 90.9 ± 2.8b 1.6 ± 0.1e 0.586 ± 0.009d NA NA 295.1 ± 0.0b PSO NA NA NA NA NA NA 408.0 ± 10.0a Product Coacervate 0.419 ± 0.02ab 81.3 ± 0.3a 38.9 ± 0.1a 0.514 ± 0.018a 4.2 ± 0.1b 0.353 ± 0.01a NA Non-coacervate 0.397 ± 0.01b 82.3 ± 0.6a 31.8 ± 1.1a 0.548 ± 0.011a 3.4 ± 0.0a 0.345 ± 0.00a NA Control 0.447 ± 0.00a 83.6 ± 0.9a 33.7 ± 2.5a 0.502 ± 0.017a 3.6 ± 0.1a 0.333 ± 0.01a NA PSO: pomegranate seed oil; WM: wall material; NA: not analyzed. Different lower case letters within the same column indicate significant differences (p< 0.05; ANOVA).

202

4. Conclusion

In the present study PSO encapsulation by complex coacervation was considered successfully, as microparticles showed an overall stability of 60 days. Nevertheless, it is important emphasize that the spray drying process still a more eficient encapsulation method as non-coacervate microparticles showed a superior oxidative stability. Besides that, from the economical point of view, the spray drying method was also more attractive, due to a more dynamic process flow chart, not needing time consuming steps such as: pH reduction to 3.7 and time to polymer to interact. Stability at 60 °C demonstrated an important role of atmospheric O2 on microparticles oxidation, which was significantly reduced when PSO’s microparticles were storage under vacuum. In this sense, we propose that this sealing condition would be a viable strategy to increase microparticles’ shelf life. Coacervate microparticles showed technological properties that enable its application in a wide range of food products, therefore a vehicle for PSO’s microparticles consumption was proposed. A coffee latte drink was designed and microparticles addition (30%) did not affect major original technological properties of the product (hygroscopicity, solubility, wettability, bulk density and aw). Moreover, the present work provided a valuable data concerning cLnA possible oxidation pathways, an unexplored matter that needs further investigations. The developed potential functional food could be a feasible vehicle for encapsulated PSO consumption and might be used in future studies to investigate the health benefits associated with acute or chronic PSO consumption. In the future, would be interesting evaluate product/PSO’s microparticles liberation pattern, as well as coffee latte drink’s sensory attributes.

Acknowledgements

The authors greatly acknowledge the funding provide by CAPES, CNPq (Grant numbers: 432484/2016-7 and 309558/2015-8) and FAPERJ (Grant numbers: E- 26/010.001277/2015 and E-26/203.197/2015) (Brazil). AGT is a recipient of a CNPq scholarship, AMMC was a recipient of CNPq PhD studentships. We are also indebted with Brenda Duarte Gralha and Bruna R. A. Gaspar for their help with laboratory analysis of stability (PSO’s microparticles) and technological properties (PSO’s microparticles and instant coffee latte drink). 203

5. References

Al‐Kahtani, H. A., & Hassan, B. H. (1990). Spray Drying of Roselle (Hibiscus sabdariffa L.) Extract. Journal of Food Science, 55(4), 1073–1076. http://doi.org/10.1111/j.1365- 2621.1990.tb01601.x

Andersen, A. B., Risbo, J., Andersen, M. L., & Skibsted, L. H. (2000). Oxygen permeation through an oil-encapsulating glassy food matrix studied by ESR line broadening using a nitroxyl spin probe. Food Chemistry, 70(4), 499–508.

Choi, K.-O., Ryu, J., Kwak, H.-S., & Ko, S. (2010). Spray-dried conjugated linoleic acid encapsulated with Maillard reaction products of whey proteins and maltodextrin. Food Science and Biotechnology, 19(4), 957–965. http://doi.org/10.1007/s10068-010-0134-7

Fitzpatrick, J. J., Iqbal, T., Delaney, C., Twomey, T., & Keogh, M. K. (2004). Effect of powder properties and storage conditions on the flowability of milk powders with different fat contents. Journal of Food Engineering, 64(4), 435–444. http://doi.org/10.1016/j.jfoodeng.2003.11.011

Franco, T. S., Perussello, C. A., Ellendersen, L. N., & Masson, M. L. (2016). Effects of foam mat drying on physicochemical and microstructural properties of yacon juice powder. LWT - Food Science and Technology, 66(November), 503–513. http://doi.org/10.1016/j.lwt.2015.11.009

Franke, A. A., Murphy, S. P., Lacey, R., & Custer, L. J. (2007). Tocopherol and Tocotrienol Levels of Foods Consumed in Hawaii. Journal of Agricultural and Food Chemistry, 55(3), 769–778.

Frankel, E. N. (2005). Methods to determine extent of oxidation. In E. N. Frankel (Ed.), Lipid Oxidation (2nd ed., pp. 99–128). Cambridge, UK: Woodhead Publishing.

Fuchs, M., Turchiuli, C., Bohin, M., Cuvelier, M. E., Ordonnaud, C., Peyrat-Maillard, M. N., & Dumoulin, E. (2006). Encapsulation of oil in powder using spray drying and fluidised bed agglomeration. Journal of Food Engineering, 75(1), 27–35. http://doi.org/10.1016/j.jfoodeng.2005.03.047

Gimeno, E., Castellote, A. I., Lamuela-Raventós, R. M., de la Torre, M. C., & López-Sabater, M. C. (2000). Rapid determination of vitamin E in vegetable oils by reversed-phase high-performance liquid chromatography. Journal of Chromatography A, 881(1–2), 251–254. http://doi.org/10.1016/S0021-9673(00)00219-3

Giua, L., Blasi, F., Simonetti, M. S., & Cossignani, L. (2012). Oxidative modifications of 206

conjugated and unconjugated linoleic acid during heating. Food Chemistry. http://doi.org/10.1016/j.foodchem.2012.09.067

Giua, L., Blasi, F., Simonetti, M. S., & Cossignani, L. (2013). Oxidative modifications of conjugated and unconjugated linoleic acid during heating. Food Chemistry, 140(4), 680– 685. http://doi.org/10.1016/j.foodchem.2012.09.067

Goula, A. M., & Adamopoulos, K. G. (2004). Spray drying of tomato pulp: Effect of feed concentration. Drying Technology, 22(10), 2309–2330. http://doi.org/10.1081/LDRT- 200040007

I Ré, M. (1998). MICROENCAPSULATION BY SPRAY DRYING. Drying Technology, 16(6), 1195–1236. http://doi.org/10.1080/07373939808917460

Jimenez, M., García, H. S., & Beristain, C. I. (2006a). Spray-dried encapsulation of Conjugated Linoleic Acid ( CLA ) with polymeric matrices. Journal of the Science of Food and Agriculture, 2437(May 2005), 2431–2437. http://doi.org/10.1002/jsfa

Jimenez, M., García, H. S., & Beristain, C. I. (2006b). Spray-dried encapsulation of Conjugated Linoleic Acid (CLA) with polymeric matrices. Journal of the Science of Food and Agriculture, 86(14), 2431–2437. http://doi.org/10.1002/jsfa.2636

Jinapong, N., Suphantharika, M., & Jamnong, P. (2008). Production of instant soymilk powders by ultrafiltration, spray drying and fluidized bed agglomeration. Journal of Food Engineering, 84(2), 194–205. http://doi.org/10.1016/j.jfoodeng.2007.04.032 207

Kim, E. H. J., Chen, X. D., & Pearce, D. (2002). Surface characterization of four industrial spray-dried dairy powders in relation to chemical composition, structure and wetting property. Colloids and Surfaces B: Biointerfaces, 26(3), 197–212. http://doi.org/10.1016/S0927-7765(01)00334-4

Lai, H.-M., & Cheng, H.-H. (2004). Properties of pregelatinized rice flour made by hot air or gum puffing. International Journal of Food Science & Technology, 39(2), 201–212. http://doi.org/10.1046/j.0950-5423.2003.00761.x

Larick, D. K., & Parker, J. D. (2001). Chromatographic Analysis of Secondary Lipid Oxidation Products. In R. E. Wrolstad, T. E. Acree, E. A. Decker, M. H. Penner, D. S. Reid, S. J. Schwartz, … P. Sporns (Eds.), Curr. Protoc. Food Analyt. Chem (pp. 1–9). John Wiley & Sons, Inc.

Liu, S., Low, N. H., & Nickerson, M. T. (2010). Entrapment of flaxseed oil within gelatin- gum Arabic capsules. JAOCS, Journal of the American Oil Chemists’ Society, 87(7), 809–815. http://doi.org/10.1007/s11746-010-1560-7

Minguez-Mosquera, M. I., Garrido-Fernandez, J., Gandul-Rojas, B., & Garrido-fernindez, J. (1989). Pigment changes in olives during fermentation and brine storage. Journal of Agricultural and Food Chemistry, 37(1), 8–11. http://doi.org/10.1021/jf00085a002

Park, D. J., Imm, J. Y., & Ku, K. H. (2001). Improved dispersibility of green tea powder by microparticulation and formulation. Journal of Food Science, 66(6), 793–798. http://doi.org/10.1111/j.1365-2621.2001.tb15174.x

Rahmani, M., & Csallany, A. S. (1991). Chlorophyll and β-carotene pigments in moroccan virgin olive oils measured by high-performance liquid chromatography. Journal of the American Oil Chemists Society, 68(9), 672–674. http://doi.org/10.1007/BF02662293

Rodriguez-Amaya, D. B., & Kimura, M. (2004). HarvestPlus Hanbook for Carotenoid Analysis. Washinton, DC, US: International Food Policy Research Institute and International Center for Tropical Agriculture.

Sankarikutty, B., Sreekumar, M. M., Narayanan, C. S., & Mathew, A. G. (1988). Studies on microencapsulation of cardamom oil by spray drying technique. Journal of Food Science 209

and Technology, 25(6), 325–355. Retrieved from http://cat.inist.fr/?aModele=afficheN&cpsidt=7181377

Shittu, T. A., & Lawal, M. O. (2007). Factors affecting instant properties of powdered cocoa beverages. Food Chemistry, 100(1), 91–98. http://doi.org/10.1016/j.foodchem.2005.09.013

Tan, B., & Brzuskiewicz, L. (1989). Separation of tocopherol and tocotrienol isomers using normal- and reverse-phase liquid chromatography. Analytical Biochemistry, 180(2), 368– 373. http://doi.org/10.1016/0003-2697(89)90447-8

Timilsena, Y. P., Adhikari, R., Barrow, C. J., & Adhikari, B. (2016). Microencapsulation of chia seed oil using chia seed protein isolate-chia seed gum complex coacervates. International Journal of Biological Macromolecules, 91, 347–357. http://doi.org/10.1016/j.ijbiomac.2016.05.058

Tonon, R. V., Freitas, S. S., & Hubinger, M. D. (2011). Spray drying of açai (Euterpe oleraceae Mart.) juice: Effect of inlet air temperature and type of carrier agent. Journal of Food Processing and Preservation, 35(5), 691–700. http://doi.org/10.1111/j.1745- 4549.2011.00518.x 210

Tonon, R. V., Pedro, R. B., Grosso, C. R. F., & Hubinger, M. D. (2012). Microencapsulation of Flaxseed Oil by Spray Drying: Effect of Oil Load and Type of Wall Material. Drying Technology, 30(13), 1491–1501. http://doi.org/10.1080/07373937.2012.696227

Viegas, M. C., & Bassoli, D. G. (2007). Utilização do índice de retenção linear para caracterização de compostos voláteis em café solúvel utilizando GC-MS e coluna HP- Innowax. Química Nova. http://doi.org/10.1590/S0100-40422007000800040

Vogt, G., Woell, S., & Argos, P. (1997). Protein thermal stability, hydrogen bonds, and ion pairs. Journal of Molecular Biology, 269(4), 631–643. http://doi.org/10.1006/jmbi.1997.1042

Wang, S., Shi, Y., & Han, L. (2018). Development and evaluation of microencapsulated peony seed oil prepared by spray drying: Oxidative stability and its release behavior during in-vitro digestion. Journal of Food Engineering, 231, 1–9. http://doi.org/10.1016/j.jfoodeng.2018.03.007

Weinbreckt, F., & Minorf, M. (2004). Microencapsulation of oils using whey protein / gum arabic coacervates. Journal of Microencapsulation, 21(6), 667–679. http://doi.org/10.1080/02652040400008499

Yousefi, S., Emam-Djomeh, Z., & Mousavi, S. M. (2011). Effect of carrier type and spray drying on the physicochemical properties of powdered and reconstituted pomegranate juice (Punica Granatum L.). Journal of Food Science and Technology, 48(6), 677–84. http://doi.org/10.1007/s13197-010-0195-x

212

Chapter 5

Global conclusions and suggestions for future studies

213

1. Global Conclusions

PSO is rich in compounds with potentional bioactivity such as: cLnA isomers, especially punicic acid, tocopherols and phenolics compounds. The study of the chemical composition of three commercial PSO from Turkey and Israel, with emphasis on PSO’s bioactivity and quality (fatty acid profile, lipid classes, volatile compounds, quality indexes, oxidative stability and bioactive compounds (phenolics, tocopherols and carotenoids)) demonstrated that not only minor compounds, such as phenolics, tocopherols and carotenoids, were influenced by samples’ origin and production batch, but also major compounds, as fatty acids;

Overall, PSO samples showed similar quality index with values within the maximum acceptable level recommended by the current legislation. Oils displayed a low oxidative stability, probably because its high contents of the rapidly oxidized cLnAs, moreover this parameter was the only one which showed a significant difference between samples from distinct origins (PSOIsrael_A vs PSOTurkey_batch_2);

Although PSO samples demonstrated comparable contents of punicic acid and lipid classes, total content of cLnA isomers between samples from different origins was distinct, suggesting an influence of samples’ origin on oils major compounds, such as fatty acids;

Volatile analysis did not show a clear typical profile of volatile compounds in the PSO, as each sample showed a particular volatile composition. In this context, we proposed that the limited number of samples hider the use of this analysis as a tool for samples’ origin authenticity;

PSO bioactive compounds composition was highly affected by samples origin. Oils showed a very distinct phenolics profile composed by three major classes (phenolics acids, flavonols and flavanones). On the other hand γ-tocopherol was consistently the major tocol in all samples, with Turkish samples displaying the highest contents. PSOs were not a high source of β-carotene;

PCA analysis allowed preliminary discrimination of PSO by geographical origin based in samples’ bioactive components, revealing one cluster composed by samples from Israel;

Complex coacervation was applied in PSO to produce a stable and versatile functional food ingredient capable to be added in a wide range of foods. The complex coacervation 214

process was highly affected by total polymer concentration (Cp) and wall material:oil ratio (WM:Oil). The production of microparticles with good physical-chemical and morphological properties, such as high oil and punicic acid retention, and spherical shape with smooth surface free of visible cracks, were successfully achieved by running a rotatable central composite design to define the best formulation of PSO’ microparticles;

Cp and WM:Oil ratio showed a significant influence on coacervation process yield, wet/dried microparticles morphology and physical-chemical properties (oil retention, microencapsulation efficiency (ME), punicic acid content , moisture content, and span);

ME assed by solvent extraction is a reliable, fast and inexpensive method to evaluate microparticles’ non-encapsuled oil as the aforementioned technique values were consistent with surface analysis (XPS) results;

Central point formulation was elected due to samples’ highest punicic acid content and economical appealing indicated by formulations’ high oil retention and good oil load (WM:Oil: 2.75);

PSO encapsulation by complex coacervation was considered successfully, as microparticles showed an overall stability of 60 days. Nevertheless, it is important emphasize that the spray drying process still a more eficient encapsulation method as non-coacervate microparticles showed a superior oxidative stability. Besides that, from the economical point of view, the spray drying method was also more attractive, due to a more dynamic process flow chart, not needing time consuming steps such as: pH reduction to 3.7 and time to polymer to interact;

The present work provided a valuable data concerning cLnA possible oxidation pathways, an unexplored matter that needs further investigations. At ambient temperature, was demonstrated an increase of oxidation product before a significant cLnA degradation, tocopherols exhaustion precedes an increase in oxidation rate and tocopherols contents showed a fair correlation with cLnA content, PV and volatile compounds. Stability at 60 °C demonstrated an important role of atmospheric O2 on microparticles oxidation, which was significantly reduced when PSO’s microparticles were storage under vacuum;

The produced microparticles showed technological properties, such as high solubility and interesting thermal stability, which enabled its application in a wide range of food 215

products. The coacervate process affected microparticles wettability, water activity and thermal stability when compared with non-coacervate microparticles;

A coffee latte drink was designed based in product: suitability for microparticles addition, consumption data and convenience. Microparticles (30% coffee/microparticles; m/m) were successfully incorporated in the food product, as major original technological properties of the product were not affected after PSO’s microparticles addition (solubility, wettability, bulk density and aw);

The present study was a successfully attempt to increase the stability of an emerging functional oil by applying the complex coacervation encapsulation process and by this means propose a vehicle for PSO consumption. Furthermore, the present work provides valuable data concerning PSO’s chemical composition that might subsidize future studies on the health effects of this oil’s consumption, or on its use as a functional food ingredient.

2. Suggestions for future studies

• Increase sample size and investigate the potentional use of PSO volatile composition as a marker of origin and processing. Additionally, would be interesting evaluate to what extent PSO aroma compounds arrived from the seeds (oil’s raw material) or from chemical transformations during processing and storage;

• Asses microparticles stability in different conditions (wide range of aw, without light protection) and during a longer time. Moreover would be interesting confirm the aforementioned hypothesis that storage under vacuum might increase microparticles shelf-life at ambient temperature; • Quantify microparticles volatile fraction by gas chromatograph coupled with flame ionization detector aiming at increase results accuracy; • Test different concentrations of PSO’ microparticles addition on the product; • Evaluate microparticles/product digestibility and liberation pattern; • Determinate coffee latte drink’s sensory attributes and consumer acceptance; • Investigate the health benefits associated with acute or chronic PSO consumption.

216

APPENDIX A

1 Confirmation of cLnA isomers identities by GC-MS

GC-MS analysis showed six peaks in the cLnA retention window (Figure 1) in all samples. We identified four peaks based in previous studies (Cao et al., 2007; Kaufman & Wiesman, 2007; Spitzer, 1996), which investigated characteristic fragments in MS spectra of cLnA isomers. These fragments are shown in Table 1 and in Figure 2 there is an example of a typical MS spectrum of punicic acid-DMOX derivative.

2 1 3 5

4 6

Figure 1. Partial GC-MS Chromatogram of DMOX derivatives from a representative sample of PSO. Peaks were assigned as follows: 1: Punicic acid; 2: Alpha- elostearic acid; 3: Catalpic acid, 4: unidentified cLnA; 5: Beta-eleostearic acid and 6: unidentified cLnA. GC column: Omegawax-320, 30 m. 217

Table 1. Characteristic fragments (m/z) of selected conjugated trienoic fatty acids.

Fatty acid 2ndmost Base Molecular intense Characteristic 12 µ gaps peak ion peak 18:3 (8,10,12) 126 113 331 182/194; 208/220; 234/246 18:3 (9,11,13) 126 113 331 196/208; 222/234; 248/260 18:3 (10,12,14) 113 126 331 210/222; 236/248; 262/274 (Adapted from: in Spitzer 1997)

Figure 2. MS spectra of punicic acid DMOX derivative (Adapted from: Cao et al., 2007).

Peaks 1, 2, 3 and 5 in Figure 1 were identified as cLnA isomers with unsaturations in carbons 9, 11 and 13, from fragmentswith diagnostic importance observed in their mass spectra (Figures 6). However, the fragmentation profile does not give sufficient information to allow discrimination between these three isomers, which present identical mass spectra; owing that they present very similar structural formulae, with double bonds in the same positions, differing only in cis/trans double bond configuration. In this sense, peak identities were tentatively determined based on mass spectra combined with elution order and contents 218

profile from previous studies. Therefore, we concluded that peaks 1, 2, 3 and 5 were Punicic, Alpha-eleostearic, Catalpic and Beta-eleostearic acids, respectively.

The peaks 4 and 6 were tentatively identified as cLnA in the PSO, according to MS data. These three peaks presented characteristic fragments of cLnA isomers such as: 113, 126 and 331 m/z, but they do not exhibit the 12 amu gaps separating diagnostic fragments, indicating the double bonds localization in the hydrocarbon chain (Figure 3). Peaks’ lower concentration compromised samples’ fragmentation pattern resolution, hindering identification.

219

100.0 126 1

75.0

113 41 50.0

55 79 331 25.0 72 93 140 182 288 332 168 234 274 302 196 208 248 260 0.0 50.0 75.0 100.0 125.0 150.0 175.0 200.0 225.0 250.0 275.0 300.0 325.0

220

% 100.0 126 3

75.0

41 50.0

55 91 331 25.0 77 182 288 98 140 152 168 274 302 208 222 234 260 0.0 334 50.0 75.0 100.0 125.0 150.0 175.0 200.0 225.0 250.0 275.0 300.0 325.0 350.0

% 100.0 113 126 4

75.0 41

55 50.0 91 331

72 25.0 140 332 168 302 183 197 280 0.0 50.0 75.0 100.0 125.0 150.0 175.0 200.0 225.0 250.0 275.0 300.0 325.0 221

100.0 126 5

75.0

113 50.0 41

79 55 331 25.0 72 93 182 140 288 168 234 274 332 196 208 248 260 302 0.0 50.0 75.0 100.0 125.0 150.0 175.0 200.0 225.0 250.0 275.0 300.0 325.0

% 100.0 126 6

75.0

113

50.0

41 55 79 72 25.0 98 274 332 168 220 149 193 221 261 0.0 50.0 75.0 100.0 125.0 150.0 175.0 200.0 225.0 250.0 275.0 300.0 325.0

Figure 3. MS spectra of the cLnA isomers DMOX derivatives of commercial PSO. 1: Punicic acid; 2: Alpha-elostearic acid; 3: Catalpic acid, 4: unidentified cLnA; 5: Beta-eleostearic acid and 6: unidentified cLnA. GC column: Omegawax-320, 30m. 222

2. Confirmation of the cLnA isomers identities by GC-FID utilizing cLnA isomers

The following cLnA isomers standards: Punicic acid, Alpha-eleostearic acid, Catalpic acid and Beta-eleostearic acid, were applied to confirm the previous isomers identification in GC-MS. Isomer identification in the GC-FID was consistent with prior GC-MS analysis, as demonstrated in Figure 4. Punicic acid, Alpha-eleostearic acid, and Catalpic acid were easily identified based in standards retention times. However, for Beta-eleostearic acid, sample fortification with standard was utilized, as 5 peaks were present in the elution area of Beta- eleostearic acid in the GC-FID chromatogram (Figure 4B and C).

223

uV(x100,000) 7.0 Chromatogram A 6.5

6.0

5.5

5.0 Punicic acid/27.274 4.5

4.0

3.5

3.0 Catalpic acid/28.191 Figure 4B cLnA 3/29.113 2.5 cLnA 2/28.839

2.0

1.5 18:1n9/15.971

1.0 Alfa-eleostearic acid/27.623 I.S./12.740

0.5 18:2n6/17.498

16:0/10.381

Beta-eleostearic acid/28.980

18:0/15.372

cLnA 1/28.621

20:1n9/23.067

cLnA 4/29.365 20:0/22.440 0.0 18:1n7/16.180

-0.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 35.0 37.5 min

224

uV(x100,000) Chromatogram B 1.75

1.50 Punicic acid/27.274

1.25

1.00 Alfa-eleostearic acid/27.623

0.75 Catalpic acid/28.191

0.50 Beta-eleostearic acid/28.980

0.25

cLnA 2/28.839

cLnA 1/28.621

cLnA 3/29.113 cLnA 4/29.365

0.00

24.5 25.0 25.5 26.0 26.5 27.0 27.5 28.0 28.5 29.0 29.5 30.0 30.5 31.0 31.5min

225

uV(x100,000) 2.75Chromatogram C

2.50

2.25 Punicic acid/26.853

2.00

1.75

1.50 Alfa-eleostearic acid/27.193

1.25 Beta-eleostearic acid/28.529 Catalpic acid/27.745 1.00

0.75

0.50

0.25 cLnA 2/28.412

cLnA 3/28.630

cLnA 1/28.146 cLnA 4/28.851 0.00

-0.25 24.5 25.0 25.5 26.0 26.5 27.0 27.5 28.0 28.5 29.0 29.5 30.0 30.5 31.0 31.5 min

Figure 4. A) GC-FID chromatogram of FAME derivatives from a representative sample of PSO. B) Partial GC-FID chromatogram with emphasis on cLnA isomers elution region from a representative sample of PSO. C) Partial GC-FID chromatogram with emphasis on cLnA isomers elution region from a representative sample of PSO fortified with Beta- eleostearic standard. IS: Internal standard; GC column: Omegawax-320, 30 m. 226

3. Conclusion

GC-MS analysis identification of four cLnA isomers (Punicic acid, Alpha-eleostearic acid, Catalpic acid and Beta-eleostearic acid) was confirmed later by commercial cLnA isomers standards.

4. References

Cao, Y., Yang, L., Gao, H. L., Chen, J. N., Chen, Z. Y., & Ren, Q. S. (2007). Re- characterization of three conjugated linolenic acid isomers by GC-MS and NMR. Chemistry and Physics of Lipids, 145(2), 128–133. http://doi.org/10.1016/j.chemphyslip.2006.11.005

Spitzer, V. (1996). Structure analysis of fatty acids by gas chromatography-low resolution electron impact mass spectrometry of their 4,4-dimethyloxazoline derivatives-a review. Progress in Lipid Research, 35(4), 387–408.

227

APPENDIX B - CROMATOGRAMS FROM CHEMICAL COMPOSITION ANALYSIS PERFORMED ON THE COMMERCIALS POMEGRANATE SEED OILS

1. Fatty acid profile

uV(x10,000) 2.50 Chromatogram

2.25 /1.941

22:0/28.280 24:0/35.598

2.00

1.75

1.50

1.25 23:0/31.529

24:1n9/36.561 22:1n9/28.863

1.00

22:2/30.431

20:0/22.146 20:3n6/25.373

0.75

10:0/2.962

22:6n3/36.325

20:5n3/27.740 20:4n6/26.313

0.50 /17.339

21:0/25.149

18:0/15.128

20:2/24.318

20:1n9/22.773

20:3n3/25.824

18:1/15.708

8:0/2.386

16:0/10.197

/27.996

/27.586 /27.427

0.25 12:0/4.181

14:0/6.496

17:1/13.172

18:2n6c/17.190

17:0/12.520

11:0/3.464

18:3n3/19.601

18:3n6/18.290

16:1/10.791

13:0/5.174

15:0/8.173 14:1/7.106 0.00 15:1/8.914

-0.25 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 min

Figure 1. GC-FID chromatogram of the standard 37 FAME mix; GC column: Omegawax- 320, 30 m.

228

2. Lipid classes

mV A 900 /50.405 2

800 /54.322

700

600 /55.428

500

400

/59.225 /58.158 300

/47.153 /62.845

200 /61.803 /51.018

100 /62.672

/52.102

/60.407

/55.905

/54.846

/63.751

/54.942

/47.458

/66.046 /43.800 0 /67.221

0 10 20 30 40 50 60 70 min

1100

40.592 41.242 47.929 44.067 B 1000

900

800 49.043

700

600 52.880 500

400 41.931

300

48.488 44.618

200 54.431

51.286

45.897

49.577

55.620 56.472

34.637 58.182

100 53.422

52.338

42.771

47.254

56.113

51.838

34.131

45.213

59.788

55.076

67.183

50.484

57.058 62.955

50.113

43.282

60.513

35.174

28.396 30.343 0 20.626

0 10 20 30 40 50 60 70 min Figure 2. HPLC-ELSD chromatograms of lipid classes analysis. A) HPLC-ELSD chromatogram of the standard (soybean oil), 1: monoacylglicerols, diacylglycerols and free fat acids elution region, and 2: triacylglycerols elution region; B) HPLC-ELSD chromatogram from a representative sample of PSO; HPLC column: C18, 250 × 4.6 m.

229

3. Volatile profile

uV(x1,000,000)

Chromatogram 3.00 A

2.75

2.50

2.25

2.00

1.75

1.50

9:0/38.945

10:0/45.741

8:0/31.479 11:0/51.983

1.25 7:0/23.245

12:0/57.775 13:0/63.178

1.00 14:0/68.328

0.75 15:0/74.455

0.50

0.25

0.00

-0.25 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0 65.0 70.0 min

uV(x10,000) 5.0 Chromatogram B 4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

-0.5 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0 65.0 70.0 min

Figure 3. GC-FID chromatograms of volatile composition analysis. A) Partial GC- FID chromatogram of the standard (C7-C30 saturated alkanes). B) GC-FID chromatogram from a representative sample of PSO; GC column: 007-5, 30m.

230

4. Phenolic compounds

mAU 100 280nm,4nm (1.00)

90 4.173 A

80 44.071

70 13

28.719 47.996

40 16.612

7 31.922 18.027

30 55.879

15.791 13.024

20 10.630

8.015 4 19.023 10

9 44.424 13.967

44.925 40.138

14 61.163 0 1 2 3 5 6 8 10 11 12 15 16 17 -10 0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0 min

mAU 280nm4nm (1.00) 225 B

200

175

150

125

100

75 18 50 19 20

11 39.024 25

16.527

7.667

58.091

19.027

13.461 54.468

8.135

14.929 5.559 0 2 4 6 8 32.501 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0 60.0 min Figure 4. HPLC-DAD chromatograms of phenolic compounds analysis. A) HPLC-DAD chromatogram of the standard (containing 17 phenolic compounds, other possible compounds were injected alone); B) HPLC-DAD chromatogram from a representative sample of PSO; 1: gallic acid, 2: 3,4-dihidroxy-phenylacetic acid, 3: 5-caffeoylquinic acid, 4: p-hydroxy-benzoic acid, 5: 4-hydroxyphenylacetic acid, 6: vanillic acid, 7: caffeic acid, 8: syringic acid, 9: 2,4-hydroxy-benzoic acid, 10: p-coumaric acid, 11: ferulic acid, 12: rutin, 13: hydroxycinnamic, 14: salicylic acid, 15: rosmarinic acid, 16: naringenin, 17: kaempferol, 18: m-coumaric acid, 19: trans-cinnamic acid, and 20: quercetin; HPLC column: C18, 250 × 4.6 m.

231

5. Tocopherols

mV 500 A

450

400

350

300

4 10.045 250

200

150

100

50 3 7.350

1 2 6.928 5.171 0 0.0 2.5 5.0 7.5 10.0 12.5 min

1000 3 7.542

11.342 B

900

800

700

600

4 10.055 500

400

1 5.182 300

200

100

13.003

13.941

14.300 12.542 0

0.0 2.5 5.0 7.5 10.0 12.5 min

Figure 5. HPLC-FLU chromatograms of tocopherols quantification. A) HPLC-FLU chromatogram of the standard (1:α-tocopherol, 2:β-tocopherol, 3:γ-tocopherol, and 4: δ-tocopherol); B) HPLC-FLU chromatogram from a representative sample of PSO.; HPLC column: Zorbax-SIL, 250 × 4.6 m.

232

6. Carotenoids

mAU 450nm,0nm (1.00) 175

1 3.050 A

150

125

100

-25 0.0 2.5 5.0 7.5 10.0 12.5 min

mAU 5.0 450nm4nm (1.00)

2.804 B 4.5 1

4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

0.0 2.5 5.0 7.5 10.0 12.5 min

Figure 6. HPLC-DAD chromatograms of carotenoids quantification. A) HPLC-DAD chromatogram of the standard (1:β-carotene); B) HPLC-DAD chromatogram from a representative sample of PSO.; HPLC column: Zorbax-SIL, 250 × 4.6 m. 233

APPENDIX C - PUBLICATIONS ARISING FROM THIS THESIS

Scientific journal paper

1. COSTA, A. M. M.; SILVA, L. O.; TORRES, A. G. Chemical composition of commercial cold-pressed pomegranate (Punica granatum) seed oil from Turkey and Israel, and the use of bioactive compounds for samples’ origin preliminary discrimination. Journal of Food Composition and Analysis. v. 75, p. 8-16, 2019 (published)

Scientific conference abstracts

1. COSTA, A. M. M.; SILVA, L. O.; CASTELO-BRANCO, V. N.; NUNES, J. C.; TORRES, A. G. Influência do método de extração no perfil de ácidos graxos, composição de fitoesteróis e capacidade antioxidante de óleos de semente de romã (Punica granatum). In: Seminário Internacional Processamento de Óleos e Gorduras Tendências e Desafios, 2015, Santa Catarina;

2. CUNHA, A. L. A.; MORETTI, L. K.; COSTA, A. M. M.; SILVA, L. O.; TORRES, A. G. Determinação da qualidade inicial, capacidade antioxidante e estabilidade oxidativa de óleo de semente de romã (Punica granatum). In: XXXVIII Jornada Giulio Massarani de Iniciação Científica, Tecnológica, Artística e Cultural UFRJ, 2016, Rio de Janeiro, Brasil;

3. RANQUINI, L. G.; GRALHA, B. D.SILVA, L. O.; COSTA, A. M. M.; TORRES, A. G. A influência dos métodos de extração na qualidade e composição do óleo de semente de romã (Punica granatum) brasileira. In: XXXVIII Jornada Giulio Massarani de Iniciação Científica, Tecnológica, Artística e Cultural UFRJ, 2016, Rio de Janeiro, Brasil;

4. GRALHA, B. D.; RANQUINI, L. G.; SILVA, L. O.; COSTA, A. M. M.; TORRES, A. G. Compostos bioativos e capacidade antioxidante total de óleos de sementes de romã brasileira (Punica granatum) obtidos por diferentes métodos de extração. In: XXXVIII Jornada Giulio Massarani de Iniciação Científica, Tecnológica, Artística e Cultural UFRJ, 2016, Rio de Janeiro, Brasil; 234

5. COSTA, A. M. M.; SILVA, L. O.; TORRES, A. G. Physical-chemical characterization of commercial cold pressed pomegranate seed oils (PSO) from Turkey and Israel. In: XVII American Oil Chemists’ Society Latin American Congress and Exhibition on Fats, Oils, and Lipids, 2017, Cancun, Mexico;

6. COSTA, A. M. M.; TONON, R. V.; TORRES, A. G. Influence of wall material concentration and ratio core: wall material on the encapsulation of pomegranate seed oil by complex coacervation. In: XII Simpósio Latino Americano de Ciência de Alimentos, 2017, São Paulo, Brasil.