SÃO PAULO STATE UNIVERSITY “JÚLIO DE MESQUITA FILHO” FACULTY OF PHARMACEUTICAL SCIENCES and UNIVERSITY OF AVEIRO DEPARTMENT OF CHEMISTRY CASSAMO USSEMANE MUSSAGY Production of carotenoids by yeast Rhodotorula glutinis CCT-2186 and their extraction using alternative solvents Thesis submitted for the degree of Doctor of Philosophy São Paulo/Aveiro 2021 CASSAMO USSEMANE MUSSAGY Production of carotenoids by yeast Rhodotorula glutinis CCT-2186 and their extraction using alternative solvents Thesis carried out under “Joint PhD” agreement, presented to the Postgraduate Program in Biosciences and Biotechnology Applied to Pharmacy, Faculty of Pharmaceutical Sciences - UNESP (Brazil) as part of the requirements for the degree of Doctor in Biosciences and Biotechnology Applied to Pharmacy and Department of Chemistry - UA (Portugal) for the degree of Doctor in Chemical Engineering. Supervisors: Prof. Dr. Jorge Fernando Brandão Pereira Prof. Dr. João Araújo Pereira Coutinho São Paulo/Aveiro 2021 CASSAMO USSEMANE MUSSAGY Produção de carotenoides pela levedura Rhodotorula glutinis CCT-2186 e sua extração empregando solventes alternativos Tese em regime de cotutela, apresentada ao Programa de Pós-Graduação em Biociências e Biotecnologia aplicadas à Farmácia da Universidade Estadual Paulista “Júlio De Mesquita Filho” para obtenção do grau de Doutor em Biociências e Biotecnologia Aplicadas á Farmácia e ao Departamento de Química da Universidade de Aveiro para à obtenção do grau de Doutor em Engenharia Química. Orientadores: Prof. Dr. Jorge Fernando Brandão Pereira Prof. Dr. João Araújo Pereira Coutinho São Paulo/Aveiro 2021 بسم الله الرحمن الرحيم To: Kylie Glory Mussagy… Acknowledgments Firstly, I thank Allah for having provided me all the circumstances for carrying out this work and for placed people in my way who direct and indirectly assisted me during this period. I would like to express my deep gratitude to Professor Jorge F. B. Pereira and Professor João A. P. Coutinho, my research supervisors, for their patient guidance, enthusiastic encouragement and useful critiques to perform and improve the research work. I would also like to thank Valéria C. Santos Ebinuma, for her co-supervision and valuable suggestions during the planning and development of this research work. I would like to express my very great appreciation to the colleagues from UNESP-Brazil and Path Group-Portugal. To all the friendship I’ve made at Brazil and Portugal. The technicians from UNESP and UA laboratories. I am particularly grateful for the assistance given by Prof. Dr. Nivaldo Boralle from Chemistry Institute (UNESP) (allowed me to carry out NMR analyses), Kiki Kurnia (Indonesia) for COSMO-RS analyses, Fabiane Farias and Marcos Mafra from Federal University of Paraná (Brazil). I would like to thank the Brazilian financial agency CAPES-PRINT and Ministry of Science and Technology, High Education and Technical Vocational Training of Mozambique through the HEST Project-World Bank for financial support in Portugal and Brazil during the PhD. Finally, my special thanks are extended to my family in Mozambique, to my wife and my daughter Kylie Glory Mussagy for all love, encouragement throughout my study, patience and support during my years of dedication during the PhD period and also because they accepted the challenge of trust me even during my absence and difficult moments. To all who directly or indirectly contributed to this work: Thank you very much! Muito obrigado! Gracias! Merci beaucoup! Khanimambo! Abstract This thesis consists of the investigation and optimization of sustainable strategies to enhance carotenoids production from yeast Rhodotorula glutinis CCT- 2186 (R. glutinis) and their extraction by means of ionic liquids (ILs) and bio-based solvents. The interest on carotenoids such as β-carotene, torulene and torularhodin relies on the plethora of relevant properties of these products and their commercial value for food, feed, cosmetic and pharmaceutical industries. However, to make carotenoids more accessible in a sustainable way, their production from microbial sources is of paramount importance. Concerning the production of carotenoids from R. glutinis yeast there are still processual challenges, such as the improvement of production yields and development of efficient and sustainable extraction platforms. Initially, different statistical experimental designs were applied to improve carotenoids production and the best bioprocess was scaled-up to a 5 L stirred-tank bioreactor. Since the carotenoids are produced intracellularly, requiring appropriate cell-disrupting and extraction methodologies for their recovery, subsequently, the development of more benign and effective extraction/purification platforms was evaluated. A comprehensive study using aqueous solutions of ionic liquids (ILs) for solid-liquid extraction (SLE) processes was carried out, following the carotenoids purification using a three-phase partitioning system composed of aqueous solutions of ILs and inorganic salts. To gather additional information on the phase separation mechanisms, aqueous biphasic systems (ABS) composed of ILs and inorganic salts were determined and characterized. Afterward, the potential of bio-based solvents was evaluated, with the purpose of designing a more efficient and ecofriendly extraction process for recovery of the intracellular carotenoids from R. glutinis. In this study it was designed and optimized an integrated downstream platform using a ternary mixture of bio- based solvents (ethyl acetate/ethanol/water) with isolation and polishing of carotenoids as well as the recycling of the solvents. The overall sustainability of the proposed technology was assessed in terms of solvents recyclability and carotenoids polishing, and the environmental impact of the platform through a life cycle assessment (carbon footprint). This thesis demonstrates the importance of combined organic and inorganic nitrogen sources to supplement the nutritional media for cultivation of R. glutinis and production of carotenoids, and that ILs and mixed bio-based solvents can be used to design simple, efficient and sustainable platforms for the recovery of intracellular carotenoids from microbial biomass. Keywords: Rhodotorula glutinis; carotenoids; lipids; ionic liquids, biosolvents; production; extraction. Resumo A crescente busca por produtos naturais e os desafios de sua produção industrial, tem motivado pesquisadores a investigar estratégias para o incremento da produção a partir de fontes microbianas e o desenvolvimento de procedimentos sustentáveis e eficientes para a sua extração. Os carotenoides naturais como o β-caroteno, toruleno e torularodina além de serem poderosos antioxidantes naturais, são corantes que vem merecendo destaque devido às suas propriedades biológicas com aplicações relevantes nas indústrias alimentar, cosmética e farmacêutica. No presente trabalho estudou-se um processo integrado que engloba o cultivo da levedura Rhodotorula glutinis CCT- 2186 (R. glutinis) para o incremento da produção de carotenoides e a extração destes produtos empregando líquidos iónicos (LIs) e bio-solventes como solventes alternativos. Inicialmente, diferentes planejamentos fatoriais foram aplicados para incrementar a produção de carotenoides e para a melhor condição foi realizado um aumento de escala em biorreator de 5 L. Posteriormente, estudos usando soluções aquosas de LIs em processos de extração sólido-líquido foram realizados, seguidos pelo processo de purificação empregando um sistema de partição trifásico composto por soluções aquosas de LIs e sais inorgânicos. Para reunir informações adicionais sobre os mecanismos de separação das fases, sistemas aquosos bifásicos (SAB) compostos por LIs e sais inorgânicos foram determinados e caracterizados. O potencial dos bio-solventes para a recuperação de carotenoides de células de R. glutinis foram também avaliados, a fim de se obter um processo de extração/purificação sustentável, mais eficiente e ecologicamente correto. Este último estudo permitiu a compreensão e otimização de uma abordagem downstream sustentável para a extração usando misturas ternárias (acetato de etilo/etanol/água) com posterior isolamento e polimento de carotenoides, e reciclagem dos solventes da mistura. A sustentabilidade geral da tecnologia proposta neste trabalho foi avaliada pelo impacto ambiental do processo alternativo em termos de avaliação do ciclo de vida (pegada de carbono). Esta tese revela que o uso combinado de fontes de nitrogênio orgânico e inorgânico é recomendado para a suplementação do meio de cultura para o cultivo de R. glutinis e que o uso de LIs e bio-solventes como agentes para o rompimento da parede celular pode ser um procedimento simples, eficiente, sustentável e viável para a recuperação de carotenoides intracelulares da biomassa microbiana. Palavras-chave: Rhodotorula glutinis; carotenoides; lipídios, líquidos iônicos, bio- solventes, produção; extração. Rights and permissions The permissions to re-use texts, figures, and tables from previously published articles from the author were secured from: Publisher: Springer Nature and Copyright Clearance Center • Paper 1: Mussagy CU, Winterburn J, Santos-Ebinuma VC, Pereira JFB (2019) Production and extraction of carotenoids produced by microorganisms, Appl. Microbiol. Biotechnol. 103 (3): 1095-1114. doi: 10.1007/s00253-018-9557-5. License number 4920240850449 (Oct 01, 2020). Publisher: American Chemical Society • Paper 2: Mussagy CU, Santos-Ebinuma VC, Gonzalez-Miquel M, Coutinho JAP, Pereira JFB (2019) Protic ionic liquids as cell-disrupting agents for the recovery of intracellular carotenoids from yeast Rhodotorula glutinis CCT- 2186, ACS Sustainable Chem. Eng 7 (19): 16765-16776. doi: 10.1021/acssuschemeng.9b04247. The author contacted Dr. Rhea M. Williams (Senior Managing Editor of ACS Sustainable Chemistry & Engineering) by e-mail, which granted written permission for use of the respective article. Publisher: Royal Society of Chemistry • Paper 3: Mussagy CU, Santos-Ebinuma VC, Kurnia KA, Carvalho P, Dias ACR, Coutinho JAP, Pereira JFB (2020) Integrative platform for the selective recovery of intracellular carotenoids and lipids from Rhodotorula glutinis CCT-2186 yeast using mixtures of bio-based solvents. Green Chem. 22: 8478-8494. doi: 10.1039/D0GC02992K. RSC have copyright policies that grant automatic permission for re-use the paper in your thesis if you are the author: (…) you do not need to request permission to reuse your own figures, diagrams, etc, that were originally published in a RSC publication (…), permission should be requested for use of the whole article or chapter except if reusing it in a thesis (...). Publisher: Elsevier • Paper 4: Mussagy CU, Tabanez NL, Farias FO, Kurnia KA, Mafra MR, Pereira JFB (2020) Determination, characterization and modeling of aqueous biphasic systems composed of propylammonium-based ionic liquids and phosphate salts, Chem. Phys. Lett. 754: 137623. doi: 10.1016/j.cplett.2020.137623. • Paper 5: Mussagy CU, Guimarães AA, Rocha LV, Winterburn J, Santos- Ebinuma VC, Pereira JFB (2021), Improvement of carotenoid production from Rhodotorula glutinis CCT-2186, Biochem. Eng. J. 165: 107827. doi: 10.1016/j.bej.2020.107827. Elsevier have copyright policies that grant automatic permission for re-use the paper in your thesis if you are the author: (…) as the author of this Elsevier article, you retain the right to include it in a thesis or dissertation, provided it is not published commercially. Permission is not required, but please ensure that you reference the journal as the original source (…) Contents Notation .......................................................................................................................................................... I List of symbols ............................................................................................................................................ I List of abbreviations ................................................................................................................................... II List of figures ............................................................................................................................................. V List of tables ............................................................................................................................................. IX 1. MOTIVATION AND THESIS STRUCTURE .......................................................................................... 1 1.1 Motivation ...................................................................................................................................... 2 1.2 Objectives ..................................................................................................................................... 2 1.3 Thesis Structure ............................................................................................................................ 3 2. THEORETICAL INTRODUCTION ........................................................................................................ 8 2.1 Introduction .................................................................................................................................10 2.2 Structure, classification, and biosynthesis of carotenoids ..........................................................12 2.3 Microbial fermentation processes to produce carotenoids .........................................................15 2.4 New genetic engineering approaches to increase carotenoid production yields ........................19 2.5 Extraction methods for the recovery of intracellular carotenoids ................................................20 2.6 Metabolites extracted during the carotenoid extraction ..............................................................31 2.7 Analytical techniques for characterization and quantification of carotenoids .............................32 2.8 Carotenoid properties and biological functions ...........................................................................33 2.9 Applications and market ..............................................................................................................37 2.10 Conclusions .................................................................................................................................42 2.11 References ..................................................................................................................................43 3. IMPROVEMENT OF CAROTENOIDS PRODUCTION FROM Rhodotorula glutinis .......................58 3.1 Introduction .................................................................................................................................61 3.2 Experimental section ...................................................................................................................62 3.3 Results and discussion ...............................................................................................................68 3.4 Conclusion ..................................................................................................................................86 3.5 References ..................................................................................................................................87 4. RECOVERY OF CAROTENOIDS FROM R. glutinis YEAST USING PROTIC IONIC LIQUIDS ......91 4.1 Introduction .................................................................................................................................94 4.2 Experimental section ...................................................................................................................96 4.3 Results and discussion .............................................................................................................102 4.4 Conclusions ...............................................................................................................................129 4.5 References ................................................................................................................................130 5. DETERMINATION, CHARACTERIZATION AND MODELING OF AQUEOUS BIPHASIC SYSTEMS COMPOSED OF PROTIC IONIC LIQUIDS ...............................................................................................135 5.1 Introduction ...............................................................................................................................137 5.2 Experimental section .................................................................................................................138 5.3 Results and discussion .............................................................................................................144 5.4 Conclusions ...............................................................................................................................165 5.5 References ................................................................................................................................166 6. RECOVERY OF CAROTENOIDS FROM R. glutinis YEAST USING MIXTURES OF BIO-BASED SOLVENTS ................................................................................................................................................170 6.1 Introduction ...............................................................................................................................173 6.2 Experimental section .................................................................................................................176 6.3 Results and discussion .............................................................................................................187 6.4 Conclusions ...............................................................................................................................215 6.5 References ................................................................................................................................216 7. FINAL REMARKS AND FUTURE WORK ........................................................................................223 7.1 Final Remarks ..................................................................................................................................224 7.2 Future Work .....................................................................................................................................225 LIST OF PUBLICATIONS AND AWARDS ...............................................................................................227 APPENDIX .................................................................................................................................................231 I Notation List of symbols %EE Percentage of extraction efficiency ∆Ghyd Gibbs free energy of hydration A Merchuck correlation constant B Merchuck correlation constant C Merchuck correlation constant GE Excess free Gibbs energy (kJ.mol-1) HVDW Van der Waals energy (kJ.mol-1) hE Excess enthalpy (kJ.mol-1) HHB Hydrogen bonding energy (kJ.mol-1) HMF Electrostatic - misfit energy (kJ.mol-1) K Partition coefficient Kow Octanol-water partition coefficient pKa Acid constant dissociation T Temperature (ºC, K) δ Root mean-square deviation η Dynamic viscosity (mPa.s) λmax Maximum absorption ρ Density (g.cm-3) σ Charge distribution II List of abbreviations [BMIM][BF4] 1-butyl-3-methylimidazolium tetrafluoroborate [BMIM][PF6] 1-butyl-3-methylimidazolium hexafluorophosphate [DEAPA][Ac] 3-diethylamino propylammonium acetate [DEAPA][But] 3-diethylamino-propylammonium butanoate [DEAPA][Hex] 3-diethylamino-propylammonium hexanoate [DEAPA][Pro] 3-diethylamino propylammonium propanoate [DMAPA][Ac] 3-dimethylamino-1-propylammonium acetate [DMAPA][But] 3-dimethylamino-1-propylammonium butanoate [DMAPA][Hex] 3-dimethylamino-1-propylammonium hexanoate [DMAPA][Pro] 3-dimethylamino-1-propylammonium propanoate [Emim][DBP] 1-ethyl-3-methylimidazolium dibutyl-phosphate [Emim]+ 1-ethyl-3-methylimidazolium [PA][Ac] Propylammonium acetate [PA][But] Propylammonium butanoate [PA][Hex] Propylammonium hexanoate [PA][Pro] Propylammonium propanoate 2-MeTHF 2-methyl tetrahydrofuran AA Arachidonic acid ABS Aqueous biphasic systems Ac Acetic acid AILs Aprotic ionic liquids ANOVA Analysis of variance But Butyric acid CAGR Compound annual growth rate CH Cyclohexane CO2 Carbon dioxide EPA Eicosapentaenoic acid COSMO-RS COnductor-like Screening MOdel for Real Solvent DCW Dry cell weight DEAPA 3-diethylamino-propylamine DES Deep Eutectic Solvents III DHA Docosohexaenoic acid DMAPA 3-dimethylamino-1-propylamine DMAPP Dimethylallyl pyrophosphate DMSO Dimethyl sulfoxide DNS 3,5-dinitrosalicylic acid EAE Enzyme-assisted extraction EFSA European Food Safety Authority EtOAc Ethyl acetate EtOH Ethanol EU European Union FDA Food and Drug Administration FPP C15-farnesyl pyrophosphate GGDP C20-geranylgeranyl diphosphate GGPP Geranylgeranyl pyrophosphate GPP C10-geranyl pyrophosphate H2O Water Hex Hexanoic acid HPLC High-performance liquid chromatography HSPs Hansen Solubility Parameters IBA 3-indole butyric acid ILs Ionic liquids IPA Isopropanol IPP Isopentenyl pyrophosphate KF Karl Fischer LLE Liquid-liquid equilibria LLE Liquid-liquid extraction MeOH Methanol MEP 2-C-methyl-D-erythritol-4-phosphate MS Mass spectrometry MVA Mevalonic acid NMR Nuclear magnetic resonance NRTL Non-Random Two-Liquid OF Objective function IV PA Propylamine PILs Protic ionic liquids PLE Pressurized liquid extraction POME Palm oil mill effluent Pro Propanoic acid R. glutinis Rhodotorula glutinis CCT-2186 RI Resolution of identity standard ROS Reactive oxygen species RP-HPLC Reversed-phase high-performance liquid chromatography SEM Scanning electron microscopy SFE Supercritical fluid extraction SLR Solid-liquid ratio STL Slope of the tie-line THF Tetrahydrofuran TLC Thin-layer chromatography TLs Tie line TLL Tie-line length TMS Tetramethylsilane TPP Three-phase partitioning TZVP Triple- ζ valence polarized basis set UAE Ultrasound-assisted extraction US United States VOCs Volatile organic solvents VVM Air volume/medium volume/min YPD Yeast Extract-Peptone-Dextrose V List of figures Figure 1.1. Schematic representation of the thesis structure, grouped by chapters and corresponding published manuscripts. ................................................................................................................................... 4 Figure 2.1. Chemical structures of major carotenoids produced by microorganisms .................................13 Figure 2.2. Carotenoid biosynthesis through the MVA pathway in microorganisms. Boxes indicate the name of carotenoids and their respective color (adapted from Clotault et al. 2012). ............................................14 Figure 2.3. Conventional and non-conventional techniques for cell disruption and carotenoid extraction (adapted from Saini and Keum 2017). .........................................................................................................23 Figure 2.4. Symmetric and asymmetric cleavage of β-carotene (adapted from Ziouzenkova et al. 2007). 35 Figure 2.5. a) Physical (adapted from Krinsky 1989) and b) chemical (adapted from El-Agamey et al. 2004) mechanisms to neutralize the effects of ROS. .............................................................................................36 Figure 2.6. Carotenoids and their industrial applications ............................................................................38 Figure 2.7. Analysis and perspective of global carotenoid market for the period between 2007 and 2022. CAGR, compound annual growth rate (adapted from McWilliams 2018). ...................................................40 Figure 3.1. Pareto chart for the effects of variables: (1) glucose, (2) KH2PO4, (3) MgSO4, (4) nitrogen source (a: NH4NO3 and b: asparagine) and (5) pH under β- carotene, torularhodin and torulene production by R. glutinis, at 30 °C, 170 rpm for 72 h, according to the 25-1 fractional factorial design. ..................................75 Figure 3.2. Response surface and Pareto Chart for studies of the effects of independent variables x1: pH and x2: nitrogen source on y1: β-carotene (a), y2: torularhodin (b) and y3: torulene (c) production by R. glutinis CCT-2186 in 72 h in an orbital shaker at 30 °C and 170 rpm. .....................................................................80 Figure 3.3. Dry cell weight (g/L) (-■-), glucose (g/L) (-▲-), lipids (g/L) (--) and production of β-carotene (- ○-), torularhodin (-◇-) and torulene (-□-) by R. glutinis CCT-2186 in 5 L stirred-tank bioreactor at 30 oC, 1 vvm, 300 rpm for 72 h. The error bars in some cases, are smaller than the markers. ................................83 Figure 4.1. a- HPLC analysis of the three major carotenoids of R. glutinis CCT-2186 (1: torularhodin; 2: torulene; 3: β-carotene). b- Analytical TLC of R. glutinis cells' extracts: a- β-carotene standard, b- Torularhodin standard, c- Sample (cells' extracts). c- UV absorption spectra of the three major pigments of R. glutinis. ...................................................................................................................................................104 VI Figure 4.2. Recovery of β-carotene (yellow), torularhodin (red), and torulene (pink) using DMSO and aqueous solutions of PILs (90% v/v) at a wet cell concentration of 0.2 g/mL after 1 h stirring (30 rpm) at 25 °C. The error bars represent 95% confidence levels for the mean of three independent assays. ............106 Figure 4.3. Linear relationship between log Kow of the anions versus (a) β-carotene, (b) torularhodin, and (c) torulene recovery at a wet cell concentration of 0.2 g/mL after 1 h of stirring (30 rpm) at 25 °C for different PIL-based cation families: [PA][X] (red square), [DMAPA][X] (green triangle), and [DEAPA][X] (blue diamond).The error bars represent 95% confidence levels for the mean of three independent assays. ..108 Figure 4.4. Viscosity (mPa.s) of different aqueous solutions of PILs as a function of temperature. .........111 Figure 4.5. Effect of temperature [(triangle) 25 °C, (diamond) 45 °C, and (circle) 65 °C] as a function of PIL concentration [75, 80, 85, and 90% (v/v)] on the release of β-carotene, torularhodin, and torulene from R. glutinis wet cells (0.2 g/mL) after 1 h of stirring (30 rpm). The error bars represent 95% confidence levels for the mean of three independent assays but, in some cases, are smaller than the markers. ......................113 Figure 4.6. Effect of the temperature [(triangle) 25 °C, (diamond) 45 °C, and (circle) 65 °C] as a function of solid-liquid ratio, SLR (0.05, 0.1, 0.2, and 0.5 g/mL wet cells), on the release of β-carotene, torularhodin, and torulene using different solutions of PILs at 90% (v/v) after 1 h of stirring (30 rpm). The error bars represent 95% confidence levels for the mean of three independent assays, but in some cases, they are smaller than the markers. ...........................................................................................................................116 Figure 4.7. Recovery of β-carotene (yellow), torularhodin (red), and torulene (pink) using DMSO (control) and an aqueous solution of [Hex]--based PILs (90% v/v) at a wet cell concentration of 0.2 g/mL after 1 h of stirring (30 rpm) at 25, 45, and 65 °C. The error bars represent 95% confidence levels for the mean of three independent assays. ..................................................................................................................................119 Figure 4.8. Scanning electron microscopy (SEM) images of R. glutinis CCT-2186 cells (or cell debris) (×9000) at different conditions: (a) 25 °C, (b) 45 °C, and (c) 65 °C. (1) Without treatment and after treatment with (2) DMSO, (3) [PA][Hex], (4) [DMAPA][Hex], and (5) [DEAPA][Hex] at 90% (v/v) and 0.2 g/mL wet cells. ....................................................................................................................................................................121 Figure 4.9. Diagram of the integrative process for the extraction of intracellular carotenoids using [DEAPA][Hex] solution (90% v/v), the recycling of the PIL using a three-phase partitioning system (TPP) by adding K3PO4 aqueous solution, and the polishing of the carotenoids. .....................................................125 Figure 4.10. Fourier transform infrared spectroscopy with an attenuated total reflectance (FTIR-ATR) of carotenoids with DMSO extraction (control) (▬) and reused [DEAPA][Hex]: 1st reuse (▬) 2nd reuse (▬) and 3rd reuse (▬). Wavenumber (cm-1) in the x axis and transmittance (%) in the y axis. ...............................128 VII Figure 5.1. Ternary phase diagrams of the systems a) IL + K3PO4 + H2O and b) IL + K2HPO4 + H2O determined at 298 (± 1) K and atmospheric pressure. [X] corresponds to one of the following cations: [DEAPA]+ (▲); [DMAPA]+ (♦); [PA]+ (■). Note that different scales were employed for clarity of results. ..145 Figure 5.2. Ternary phase diagrams of the systems a) [DEAPA][X] + K3PO4 + H2O and b) [DEAPA][X] + K2HPO4 + H2O determined at 298 (± 1) K and atmospheric pressure. [X] corresponds to one of the following anions: [Hex]- (▲), [But]- (●) and [Pro]- (■). Note that different scales were employed for clarity of results. ....................................................................................................................................................................146 Figure 5.3. Ternary phase diagrams of the systems a) [X][Pro] + salt + H2O, b): [X][But] + salt + H2O and c) [X][Hex] + salt + H2O determined at 298 (± 1) K and atmospheric pressure. The experimental binodal curves of the systems composed of K3PO4 are represented by full symbols, and K2HPO4 by open symbols. [X] corresponds to [DEAPA]+ (green triangle) or [DMAPA]+ (red diamond) cation. ...................................149 Figure 5.4. Ternary phase diagrams of the systems a) [DEAPA][Hex] + K3PO4 + H2O and b) [DEAPA][Hex] + K2HPO4 + H2O determined at 298 K (●), 308 K (■) or 318 K (▲), and atmospheric pressure. Note that different scales were employed for clarity of results. .................................................................................150 Figure 5.5. Ternary phase diagram of the ABS composed of [DEAPA][Hex] + K3PO4 + H2O, at 298.15 K and atmospheric pressure: binodal curve data (); tie line experimental data (); total composition (■); fitted binodal curve using Equation (5.1) (−). ......................................................................................................151 Figure 5.6. Acid:Base molar ratio between the IL forming carboxylic acid ([Pro]-, [But]- and [Hex]) and propylamine (■ - [DMAPA]+; ▲ - [DEAPA]+) on the initial mixture composition (dashed lines) and in the IL (top)-rich phase of ABS composed of K2HPO4 (a) or K3PO4 (b) + IL + H2O. .............................................155 Figure 5.7. Experimental and calculated (by NRTL model) tie-lines of systems composed of a) [DEAPA][Hex] + Salt + H2O and b) [DMAPA][Hex] + Salt+ H2O systems at T= 298 (± 1) K and atmospheric pressure. .....................................................................................................................................................159 Figure 5.8. Representation of σ-profiles for: a- phosphate salts (K3PO4 and K2HPO4); b- ILs cations ([PA]+, [DMAPA]+ and [DEAPA]+) and c- ILs anions ([Pro]-, [But]-, and [Hex]-). ....................................................161 Figure 5.9. Partial molar excess enthalpies of salt anion in the ternary system of K2HPO4 (1) + IL (2) + H2O (3) at 298.15 K predicted using COSMO-RS. [X] corresponds to [PA][HEX] (■), [DMAPA][Pro] (♦), [DMAPA][But] (▲), [DMAPA][Hex] (◊), [DEAPA][Pro] (●), [DEAPA][But] (∆), and [DEAPA][Hex] (○). ......164 Figure 6.1. Recovery yields (% w/w) of (a) carotenoids (β-carotene [yellow], torularhodin [red], and torulene [pink]) and (b) lipids using pure biosolvent (MeOH- methanol, EtOH- ethanol, IPA- Isopropanol, EtOAc- ethyl acetate, 2-MeTHF- 2-methyl tetrahydrofuran and CH- cyclohexane) at a R. glutinis wet biomass VIII concentration of 0.2 g/mL and after 1 h stirring (300 rpm) at 65 °C. The error bars represent 95% confidence levels for the mean of three independent assays.......................................................................................189 Figure 6.2. Recovery yields (% w/w) of (a) carotenoids (β-carotene [yellow], torularhodin [red], and torulene [pink]) and (b) lipids using aqueous (25% w/w of H2O) mixtures of CH (25% w/w) (+ MeOH, EtOH, IPA and 2-Me-THF) (50% w/w) and EtOAc (25% w/w) (+ MeOH, EtOH, IPA and 2-Me-THF) (50% w/w) at R. glutinis wet cell concentration of 0.2 g/mL after 1 h stirring (300 rpm) at 65 °C. The conventional Bligh and Dyer mixture of chloroform (25% w/w) + MeOH (50% w/w) + H2O (25% w/w) was used as control. The error bars represent 95% confidence levels for the mean of three independent assays. ..........................................192 Figure 6.3. Ternary phase diagram of EtOAc/EtOH/H2O, solvent mixtures composition (% w/w) and respective recovery yields (% w/w) of (a) β-carotene, (b) torularhodin, (c) torulene and (d) lipids at a R. glutinis wet cell concentration of 0.2 g/mL after 1 h stirring (300 rpm) at 65 °C. The gray curve represents the solubility binodal curve that separates the monophasic (colored zone) and the biphasic region (white zone). The results represent 95% confidence levels for the mean of three independent assays. ............196 Figure 6.4. Correlation plot between the interaction energies (X-axis) of solvents mixtures and the recoveries (% w/w) of of a) (■) β-carotene, b) (■) torulene, c) (■) torularhodin, d) (■) margaric acid, e) (■) glyceryl-1,3-dilinoleate and f) (■) trilinolein. ...............................................................................................202 Figure 6.5. Ternary phase diagram of EtOAc/EtOH/H2O system, mixture points at biphasic regions studied (a to f), with respective tie-lines (---), and partition coefficient (Kcar) of (a) β-carotene, (b) torularhodin, (c) torulene and (d) partition coefficient of lipids (Klip) at a R. glutinis wet cell concentration of 0.2 g/mL after 1 h stirring (300 rpm) at 65 °C. The results represent 95% confidence levels for the mean of three independent assays. .......................................................................................................................................................205 Figure 6.6. Diagram of the integrative process for the production and extraction of intracellular carotenoids and lipids from R. glutinis biomass using ethyl EtOAc/EtOH/H2O mixture, including the recycling of the biosolvents and the polishing of the carotenoids and lipids. ......................................................................211 Figure 6.7. The carbon footprint of the integrative platform (production, extraction, purification and polishing stages) for the production of (a) β-carotene, (b) torularhodin, (c) torulene, and (d) lipids from R. glutinis wet biomass through three different Scenarios: Scenario 1 – evaporation of EtOAc and EtOH/H2O-rich phases for polishing the carotenoids and recycling of solvents, and cold acetone precipitation for the separation of proteins from lipids; Scenario 2 - evaporation of EtOAc-rich phase, and the use of EtOH/H2O-rich phase in an second LLE procedure for the separation of proteins from lipids), and a Scenario 3 corresponding to a conventional extraction process using the Bligh and Dyer method (control). ............................................213 IX List of tables Table 2.1. Examples of “microbial” carotenoids found in nature and their producing sources. ...................16 Table 2.2. Companies producing carotenoids by biotechnological routes. .................................................41 Table 3.1. Variables and factor levels employed in the two 25-1 fractional factorial designs for studying β- carotene, torularhodin and torulene production by R. glutinis, at 30 °C, 170 rpm for 72 h. .........................64 Table 3.2. Variables and factor levels used in the 22 central composite design for studying β-carotene, torularhodin and torulene production by R. glutinis, at 30 °C, 170 rpm for 72 h. .........................................64 Table 3.3. Matrix of 25-1 fractional factorial design for study of β-carotene, torularhodin and torulene production by R. glutinis using NH4NO3 as nitrogen source, at 30 °C, 170 rpm for 72 h.* ..........................69 Table 3.4. Analysis of variance (ANOVA) applied to the regression models according to 25-1 fractional factorial design to evaluate the influence of pH, NH4NO3, glucose, KH2PO4 and MgSO4 on carotenoid production by R. glutinis, at 30 °C, 170 rpm for 72 h. ..................................................................................71 Table 3.5. Matrix of 25-1 fractional factorial design for study of biomass, lipid content, β-carotene, torularhodin and torulene production by R. glutinis using asparagine as nitrogen source, at 30 °C, 170 rpm for 72 h. ..72 Table 3.6. Analysis of variance (ANOVA) applied to the regression models according to 25-1 fractional factorial design that evaluated the influence of pH, asparagine, glucose, K2HPO4 and MgSO4 on carotenoid production by R. glutinis, at 30 °C, 170 rpm for 72 h. ..................................................................................73 Table 3.7. Experimental design matrix and experimental results for the 22 central factorial design, on carotenoid production by R. glutinis, at 30 °C, 170 rpm for 72 h. ................................................................78 Table 4.1. Recovery of β-Carotene, torularhodin, and torulene using fresh and reused [DEAPA][Hex] solution (90% v/v) at a wet cell concentration of 0.2 g/mL after 1 h of stirring (30 rpm) at 25 °C (Stage 1) and residual carotenoids remaining in the IL solution after the recycling procedure (Stage 3)a ................126 Table 5.1. Name, acronym and chemical structure of the ILs and corresponding ability to form ABS with tripotassium phosphate (K3PO4) and dipotassium phosphate (K2HPO4) aqueous solutions at 298.15 K and atmospheric pressure. ................................................................................................................................139 Table 5.2. Equilibrium data of the ABS composed of propylammonium-based IL (100wIL) + K3PO4 (100ws) + H2O at 298 K and atmospheric pressure.a ..............................................................................................153 X Table 5.3. Equilibrium data of the ABS composed of propylammonium-based IL (100wIL) + K2HPO4 (100ws) + H2O at 298 K and atmospheric pressure.a ..............................................................................................154 Table 5.4. Data binary interaction parameters of NRTL model and root mean square deviations (δ) for systems composed by K3PO4 (1) + IL (2) + water (3) at 298 K. .................................................................157 Table 5.5. Data binary interaction parameters of NRTL model and root mean square deviations (δ) for systems composed by K2HPO4 (1) + IL (2) + water (3) at 298 K. ..............................................................158 Table 6.1. Recovery yields (% w/w) of β-Carotene, torularhodin, torulene and lipids using fresh and reused EtOAc/EtOH/H2O mixture at a R. glutinis wet cell concentration of 0.2 g/mL after 1 h of stirring (30 rpm) at 65 °C. ..........................................................................................................................................................209 1 1. MOTIVATION AND THESIS STRUCTURE 2 1.1 Motivation Rhodotorula glutinis is capable to synthesize several industrial/commercial high- added value microbial carotenoids, such as β-carotene, torulene and torularhodin. The global market of carotenoids reached US$1.5 billion in 2017 with an expected market of US$2.0 billion by 2022. The growing consumer interest in “natural products”, due to the concerns with synthetic pigments, has made the microbial production of carotenoids more favorable and a sustainable alternative. Often, the production of microbial carotenoids using R. glutinis yeast cells uses nitrogen from inorganic sources. In fact, the cultivation conditions are essential to drive the metabolic pathways and maximize the production of carotenoids yields, being still necessary to optimize them, for example, by evaluating how the use of organic nitrogen sources can impact the carotenogenesis. In addition, since carotenoids are produced in the intracellular environment of R. glutinis yeast, the design of suitable cell-disruption technologies and integration with further downstream operation are crucial for their recovery in commercial purity requirements. Traditionally, R. glutinis yeasts are disrupted by the application of conventional methods using environmentally non-favorable volatile organic solvents (VOCs), which, despite of the high recovery yields of intracellular carotenoids, present health and environmental concerns. The commercial importance of the microbial carotenoids and their growing market, together with the need of sustainable, biocompatible, and optimized platform to obtain carotenoids from microbial sources motivated me to develop this Thesis and to accomplish all the objectives detailed below. 1.2 Objectives The work done during my PhD aimed to increase the production of carotenoids, namely, β-carotene, torulene and torularhodin, from R. glutinis yeast and to design integrated downstream platforms for their recovery using ionic liquids (ILs) and bio-based solvents as environmentally-friendly and biocompatible alternatives. To fulfill the main objective of the thesis, the following specific objectives have been proposed and successfully achieved: • To improve the production of β-carotene, torulene and torularhodin using R. glutinis CCT 2186 yeast, by applying statistical optimization designs; 3 • To evaluate the capability of protic ionic liquids to disrupt (permeabilize) the cell wall of R. glutinis yeast cells and to increase the recovery of intracellular carotenoids; • To determine and characterize new ternary phase diagrams of aqueous biphasic systems (ABS) composed of ILs and inorganic salts; • To evaluate the use of bio-based solvents mixtures as alternative platforms to permeabilize R. glutinis yeast cells and to improve the selective recovery of β- carotene, torulene, torularhodin and lipids; • To develop integrated and circular processes for the recovery and polishing of carotenoids and recycling of the solvents; • To evaluate the environmental sustainability and impact of the integrated downstream platform. 1.3 Thesis Structure To contextualize the work of this Thesis project, first, it is important to mention that this research was conducted under the cotutelle agreement between the São Paulo State University (UNESP) and University of Aveiro (UA), and for that reason, during the past four years I developed the experimental work in the two institutions, namely, 3 years at UNESP and 1 year at UA. The obtained results were already published in a significant amount of different international scientific journals, namely, one review and four original articles. Thus, this thesis is divided in five main chapters, each one corresponding to a different published manuscript. To elucidate the structure of this Thesis and the association with respective manuscripts a schematic representation is depicted in Figure 1.1. 4 Figure 1.1. Schematic representation of the thesis structure, grouped by chapters and corresponding published manuscripts. 5 The Chapter 2 - “Theoretical Introduction” includes a brief explanation of the main scientific concepts behind the research purposes of this Thesis, followed by a comprehensive overview of the recent biotechnological developments in carotenoid production using microorganisms. A state-of-art of the hot topics in the field are properly addressed, from carotenoid biosynthesis to the current technologies involved in their extraction, and even highlighting the recent advances in the marketing and application of microbial carotenoids. The Chapter 2 is based on the critical review published in the Applied Microbiology and Biotechnology (Mussagy et al., 2019, doi: 10.1007/s00253- 018-9557-5). After the general introduction, in the Chapter 3 – “Improvement of carotenoids production from Rhodotorula glutinis” the results on the optimization of carotenoid production are presented, in which a response surface methodology was applied to maximize the yield and productivity of carotenoids by Rhodotorula glutinis strain CCT- 2186 using different nitrogen sources. In that research, two statistical experimental designs were applied to enhance carotenoid production: a first 25-1 fractional factorial design to evaluate the influence of independent variables pH, nitrogen source, glucose, KH2PO4 and MgSO4 concentrations; and a second 22 central factorial design to optimize the pH and nitrogen sources. After the optimization, the process was scaled-up to a 5 L stirred-tank bioreactor. This chapter is based on the manuscript recently published in the Biochemical Engineering Journal (Mussagy et al., 2021, doi: 10.1016/j.bej.2020.107827). Chapter 4: Recovery of carotenoids from R. glutinis yeast using protic ionic liquids includes the research published in the renowned ACS Sustainable Chemistry & Engineering (Mussagy et al., 2019, doi: 10.1021/acssuschemeng.9b04247). This research evaluated the use of protic ionic liquids (PILs) aqueous solutions to permeabilize the yeast cells and to improve the extraction of intracellular carotenoids. Twelve highly concentrated aqueous solutions of ammonium-based PILs were investigated, evaluating the influence of the relative ion’s hydrophobicity, PIL concentration, solid-liquid ratio, water content and temperature. In addition, an integrated platform was developed to allow the carotenoids recovery, together with the solvent recycling and carotenoids polishing, by using a three-phase partition system. https://doi.org/10.1007/s00253-018-9557-5 https://doi.org/10.1007/s00253-018-9557-5 https://www.sciencedirect.com/science/article/pii/S1369703X20303818 https://pubs.acs.org/doi/10.1021/acssuschemeng.9b04247 6 From the results obtained in the Chapter 4, and to explore the applications of ammonium-based PILs as potential and alternative solvents for the industrial liquid-liquid extraction (LLE) of other products, their ability to form aqueous biphasic systems (ABS) in the presence of phosphate salts (tripotassium phosphate and dipotassium hydrogen phosphate) was then evaluated. The ternary phase diagrams, tie-lines, and respective tie- line lengths were determined. The compositions of the coexisting phases were experimentally determined and correlated using the Non-Random Two-Liquid (NRTL) model for the activity coefficient. In addition, the predictive model COnductor-like Screening MOdel for Real Solvent (COSMO-RS) was used for a better understanding of the phase separation phenomena, predicting the interaction energy in term of excess enthalpy. The results obtained are presented in the Chapter 5 – “Determination, characterization and modeling of aqueous biphasic systems composed of protic ionic liquids”, which is based on the published manuscript Mussagy et al., 2020, doi: 10.1016/j.cplett.2020.137623) in Chemical Physical Letters, Chapter 6 – “Recovery of carotenoids from R. glutinis yeast using mixtures of bio-based solvents” includes all promising results on the development of integrative platform for the recovery of intracellular carotenoids using mixtures of solvents obtained from renewable sources (biosolvents). In fact, this chapter includes the promising research published in the highly recognized Green Chemistry journal (Mussagy et al., 2020, doi: 10.1039/D0GC02992K). A complete study was carried out, starting with a screening of solid-liquid extractions using pure and solvent mixtures, and further optimization of the best biosolvent mixture (i.e., ethyl acetate/ethanol/water) was carried out, covering the entire ternary phase diagram (viz. SLE at monophasic region and liquid- liquid extraction (LLE) at biphasic region). A full understanding of the solvation mechanisms towards carotenoids and lipids extraction using different solvent mixtures compositions was achieved with COnductor-like Screening MOdel for Real Solvent (COSMO-RS), and aiming at the circularity of the entire process, the LLE platform was integrated with following polishing and recycling operations, evaluating the carotenoids and lipids extraction performance from the reuse of mixed solvents in up to three consecutive stages. At the end, the environmental sustainability and the impact of the https://www.sciencedirect.com/science/article/pii/S0009261420305388 https://pubs.rsc.org/en/content/articlehtml/2020/gc/d0gc02992k 7 proposed technology were addressed by analyzing the carbon footprint of each integrative platform. Finally, in the Chapter 7 of this Thesis, I will present the main conclusions, final remarks and future perspectives based on the overall achievements of this research, with a particular emphasis in the high performance of the integrated downstream platforms using ILs and bio-based solvents mixtures. 8 2. THEORETICAL INTRODUCTION 9 Based on the manuscript Production and extraction of carotenoids produced by microorganisms Mussagy CU, Winterburn J, Santos-Ebinuma VC and Pereira JFB Applied Microbiology and Biotechnology, 103(3): 1095-1114, 2019. Abstract Carotenoids are a group of isoprenoid pigments naturally synthesized by plants and microorganisms, which are applied industrially in food, cosmetic, and pharmaceutical product formulations. In addition to their use as coloring agents, carotenoids have been proposed as health additives, being able to prevent cancer, macular degradation, and cataracts. Moreover, carotenoids may also protect cells against oxidative damage, acting as an antioxidant agent. Considering the interest in greener and sustainable industrial processing, the search for natural carotenoids has increased over the last few decades. In particular, it has been suggested that the use of bioprocessing technologies can improve carotenoid production yields or, as a minimum, increase the efficiency of currently used production processes. Thus, this review provides a short but comprehensive overview of the recent biotechnological developments in carotenoid production using microorganisms. The hot topics in the field are properly addressed, from carotenoid biosynthesis to the current technologies involved in their extraction, and even highlighting the recent advances in the marketing and application of microbial carotenoids. It is expected that this review will improve the knowledge and understanding of the most appropriate and economic strategies for a biotechnological production of carotenoids. Keywords: Carotenoids, microorganisms, production, extraction, biotechnology, market. 10 2.1 Introduction Carotenoids are a group of yellow, orange-red-pigmented polyisoprenoids, synthesized by plants, algae, cyanobacteria, bacteria, and fungi (Heba et al. 2015; Saini and Keum 2017; Rodriguez-Concepcion et al. 2018). These compounds, because of their large structural and functional versatility, are of utmost importance in nature (Esteban et al. 2015). Carotenoids play an important role in light harvesting and energy transfer during photosynthesis and in the protection of the photosynthetic apparatus against photo- oxidative damage (Henríquez et al. 2016), neutralizing free radicals, acting as antioxidant agents, and preventing oxidative damage to cells (Johnson and Schroeder 1996; Vachali et al. 2012). Although carotenoids exhibit a multitude of health beneficial and interesting properties, they are mainly known for their natural coloring characteristics, being the main molecules responsible for the pigmentation and colors of plants and microbial biomass. The presence of bright colors in nature has always captured the interest of scientists. The earliest studies focusing on carotenoids date back to the beginning of the nineteenth century, particularly related to the natural colors of different plants (Takaichi et al. 2006). In 1831, Henrich W.F. Wackenroder isolated β-carotene from carrot juice for the first time (Wackenroder 1831), following which many other carotenoids were discovered, isolated, and properly characterized. In 2017, approximately 1117 natural carotenoids from 683 sources (archaea 8; bacteria 170; and eukaryotes 505) have been described (Yabuzaki 2017). A large number of carotenoids have been proposed for, or already used in, a wide range of industrial applications, from the most traditional food and cosmetic uses to the more recent pharmaceutical uses. When applied in the food industry, carotenoids are almost exclusively used as additives, in which more than 2500 additives are intentionally added to foods, to maintain and improve organoleptic properties or even to extend their shelf-life (Carocho et al. 2014). The consumption of carotenoids either in foods or as a nutritional supplement can exert positive effects on health, as a precursor of vitamin A, preventing degenerative or age-related diseases as retinoid-dependent signaling, helping with cell communication and regulating gene expression (Sy et al. 2015). Although widely used in food formulations, many carotenoids used in the industry are artificial (synthetic colorants obtained through chemical synthesis), which are mainly 11 added to make the food more attractive and, thus, stimulate its consumption. However, the widespread use of synthetic colorants has generated discussions among scientific researchers and world health organizations, regarding the future human health impacts of these compounds. The regulatory agencies, i.e., the Food and Drug Administration (FDA) in the United States (US) and the European Food Safety Authority (EFSA) in the European Union (EU), have to approve the color additive before its application in food, drugs, cosmetics, and many medical devices. The tests performed by the regulatory agencies have shown the undesirable characteristics of several synthetic colorants, and as a result, the number of color additives approved by the regulatory agencies has reduced in the last years (Torres et al. 2016). Furthermore, the consumer’s conscience has been changing, with concerns about artificial food additives driving a consumer-led need for natural color- ants, which may be healthier than synthetic colorants. Consequently, the food industry is replacing artificial coloring agents from their products and focuses on the research and development of most stable and functional natural colorants (Zaccarim et al. 2018). In addition to synthetic colorants, natural carotenoids are already in commercialization, usually, those extracted from plant sources, as well as some produced via biotechnological routes (Valduga et al. 2009a). Nowadays, the industrial interest for microbial carotenoids has been increasing, particularly due to the low production area requirements (compared to plant sources), processing independent of climatic changes and seasonality, and soil composition (Valduga et al. 2009b). Furthermore, with improvements in biotechnology and bioprocessing technologies, carotenoid microbial bioprocessing can be fully controlled, increasing the production yields and reducing the overall processing costs (for example, using low-cost substrates and reducing the processing losses) (Cardoso et al. 2017a, 2017b). Because of their lipophilic characteristics, the majority of the microbial carotenoids are intracellular. So, besides upstream processing studies, downstream processing is also of utmost importance. Several studies have been working in the optimization of carotenoid extraction methods, aiming to increase recovery yields. In general, it is observed that the choice of the most efficient method for the extraction of carotenoids is dependent not only on the carotenoid characteristics (mainly its polarity) but also of the producer’s characteristics. In general, the chemical methods using organic solvents are the most 12 applied, but the number of studies regarding the use of alternative and sustainable methods, as for example, using green solvents or supercritical fluids, has been growing (Saini and Keum 2018). Considering the growing interest in microbial carotenoids, recently, many studies and reviews were published in the field (Gong and Bassi 2016; Markou and Nerantzis 2013; Henríquez et al. 2016; Minhas et al. 2016; Saini and Keum 2018; Ventura et al. 2017). In general, the published reviews always addressed facts, such as their origin, related products, and applications. However, there are several aspects of these biomolecules production, such as the integration of upstream and downstream processing, that have not been highlighted. Thus, in this review, we discuss the important concepts of carotenoid microbial production and extraction as well as their commercialization and market applications. 2.2 Structure, classification, and biosynthesis of carotenoids Carotenoids are lipophilic isoprenoids that can be classified according to their chemical and nutritional characteristics. Chemically, they are classified as carotenes and xanthophylls. The first class, carotenes, are the most well-known, containing carbon and hydrogen atoms in the chemical structure, as for example, α-carotene, β-carotene, γ- carotene, δ-carotene, and torulene. The second class, xanthophylls, in addition to carbon and hydrogen, also contains oxygen in their chemical structure, such as torularhodin, astaxanthin, and canthaxanthin (Cataldo et al. 2018; Delgado et al. 2016; Colmán et al. 2016; Mata-Gómez et al. 2014; Avalos and Carmen 2005). The chemical structure of the most common carotenoids of each class are depicted in Figure 2.1. 13 Figure 2.1. Chemical structures of major carotenoids produced by microorganisms According to its nutritional properties, carotenoids are usually classified as pro- vitamin A, i.e., β-carotene, β-cryptoxanthin, and α-carotene, or non-provitamin A, i.e., lycopene, lutein, zeaxanthin (Olson 1999; Maldonade et al. 2007; Toti et al. 2018), and ketocarotenoids, such as canthaxanthin and astaxanthin (Jayaraj et al. 2008). The biosynthesis of microbial carotenoids is derived from acetyl CoA, obtained from fatty acids via the β-oxidation pathway in the microorganism mitochondria (Lovisa and Kalluri 2018). Subsequently, the biosynthesis of terpenoids occurs, following the mevalonic acid (MVA) pathway, from which the microorganisms derive C5 isoprenoid precursors isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) via the MVA, or, depending on the organism, through the 2-C-methyl-D-erythritol-4-phosphate (MEP) pathway (Liang et al. 2017). Then, IPP is condensed with DMAPP, generating C10- geranyl pyrophosphate (GPP), and further elongating to C15-farnesyl pyrophosphate (FPP) and C20-geranylgeranyl diphosphate (GGDP) (Gharibzahedi et al. 2013), as summarized in the top of the scheme depicted in Figure 2.2. 14 Figure 2.2. Carotenoid biosynthesis through the MVA pathway in microorganisms. Boxes indicate the name of carotenoids and their respective color (adapted from Clotault et al. 2012). As an example, in the biosynthesis of bacterial carotenoids, the first carotenoid (phytoene) is formed from two geranylgeranyl pyrophosphate (GGPP) molecules catalyzed by phytoene synthase. Therefore, depending on the biocatalytic reactions, i.e., cyclization, substitution, elimination, addition, and rearrangement, the phytoene molecule can result in different molecular structures of carotenoids (Britton et al. 1995). As schematized in Figure 2.2, when phytoene is desaturated by phytoene desaturase, the 15 linear, all-trans lycopene is formed. Through lycopene cyclization, the lycopene β-cyclase introduces two β-ionone end-groups into the chemical structure, forming the well-known β-carotene. Consequently, if the carotene rings of β-carotene undergo hydroxylation re- actions, these are converted in xanthophylls, i.e., through two enzymatic reactions by β- carotene hydroxylase, the β-carotene is converted in zeaxanthin. Finally, violaxanthin ca- rotenoids can even be formed, through an enzymatic reaction involving zeaxanthin epoxidase (Liang et al. 2017). 2.3 Microbial fermentation processes to produce carotenoids Recently, the growing interest in natural carotenoids has been forcing their industrial production through the fermentation processes. Particularly, one of the most studied biotechnology fields aims to increase the microbial carotenoids productivity yields, optimizing simply the fermentation processes conditions or through more complex, but efficient, cell engineering and synthetic biology approaches. This section provides a summary of some of the key published works related with the increase of carotenoids productivity yields, through the optimization of the microbial fermentation processes conditions, such as the nutritional composition of the culture medium, pH, temperature, luminosity, aeration rate, and agitation. The choice of the best process conditions is of utmost importance since this affect not only the microbial cell growth but also the specificity of carotenoid biosynthesis (Mezzomo and Ferreira 2016). As shown in Table 2.1, several works studied carotenoid production using different microbial sources, particularly, microalgae, yeast, and bacteria. Independent of the microbial source, a wide array of carotenoids was successfully produced, ranging from the most well-known β-carotene and lutein to the less common astaxanthin and canthaxanthin. The production of a specific type of carotenoid, or a mixture of carotenoids, is a result not only from the microorganism species but also from the different production strategies employed, as briefly summarized in the following paragraphs. 16 Table 2.1. Examples of “microbial” carotenoids found in nature and their producing sources. Source Microorganism Pigment References Microalgae Dunaliella salina, Scenedesmus almeriensis, Coelastrella striolata var. multistriata β-carotene Rabbani et al.1998; Demming-Adams et al. 2002; Macías- Sánchez et al. 2010; Abe et al. 2005. Haematococcus pluvialis, Chlorella vulgaris, Coelastrella striolata var. multistriata Astaxanthin, Cantaxanthin, Lutein El-Baky et al. 2003; Demming-Adams et al. 2002; El-Baky et al. 2011; Demming-Adams et al. 2002; Abe et al. 2005. Undaria pinnatifida, Hijikia fusiformis, Laminaria japónica, Sargassum sp. and Fucus sp. Fucoxanthin Billakanti et al. 2013; Yan et al. 1999; Xiao et al. 2012; Heo and Jeon 2009; Zaragozá et al. 2008. Scenedesmus almeriensis Lutein Macías-Sánchez et al. 2010. Yeast Phycomyces blakesleeanus, Blakeslea trispora, Mucor circinelloides, Rhodosporidium sp., Sclerotium rolfsii, Sclerotinia sclerotiorum, Sporidiobolus pararoseus, Ustilago maydis, Aspergillus giganteus, Cercospora nicotianae, Penicillium sp., Aschersonia aleyroides, Xanthophyllomyces dendrorhous, Rhodotorula glutinis, Rhodotorula rubra NRRL Y-168 β-carotene Cerdá-Olmedo 1987; Avalos and Cerdá-Olmedo 2004; Fraser et al. 1996; Navarro et al. 1995; Miguel et al. 1997; Georgiou et al. 2001b; Georgiou et al. 2001a; Han et al. 2012; Estrada et al. 2010; El-Jack et al. 1988; Daub and Payne 1989; Han et al. 2005; Van-Eijk et al. 1979 Johnson 2003; Bhosale and Gadre 2001; Kot et al. 2016; Yoo et al. 2016; Valduga et al. 2009b; Shih and Hang 2006; Malisorn and Suntornsuk 2009; Tinoi et al. 2005. Rhodotorula glutinis, Sporobolomyces ruberrimus H110; Xanthophyllomyces dendrorhous Torularhodin, Torulene Buzzini and Martini 1999; Kot et al. 2016; Yoo et al. 2016; Valduga et al. 2009b; Cardoso et al. 2016; Tinoi et al. 2005; Fang and Wang 2002. Phaffia rhodozyma Astaxanthin Johnson et al. 1980 Bacteria Rhodococcus maris, Mycobacterium brevicaie, Rhodococcus maris Cantaxanthin Valduga et al. 2009b; Johnson and Schroeder 1995. Mycobacterium acticola Astaxanthin Johnson and Schroeder 1995. Arthrobacter glacialis, Arthrobacter sp. M3, A. arilaitensis Re117 Decaprenoxanthin Giuffrida et al. 2016; Monnet et al. 2010. G. alkanivorans strain 1B, Bacillus circulans Astaxanthin, Cantaxanthin, Lutein Silva et al. 2016; Fang and Wang 2002. 17 Valduga et al. (2009a) evaluated the effects of various chemical agents, such as acetic acid, mevalonic acid, β-ionone, and diphenylamine, in the increase of the carotenogenesis using yeasts of the Rhodotorula genus. It was observed that acetic acid (0.05 to 1% v/v) had no significant influence on the cellular growth and total production of carotenoids (β-carotene, torularhodin, and torulene) from R. glutinis and R. mucilaginosa. However, the addition of β-ionone (10−3 mol/L) after 70 h of fermentation had a negative effect on both yeast cultivations, reducing the cell density from 5.7 to 4.9 g/L, and carotenoid production from 1.98 to 1.70 mg/L. Conversely, the authors have shown that the addition of differing quantities of mevalonic acid (0.05, 0.1, and 0.2% v/v), while having no effect on the microbial growth, enhanced carotenoids production yields around 35 and 120% using R. glutinis and R. mucilaginosa yeasts, respectively. More recently, Cardoso et al. (2016) showed that the red yeast Sporobolomyces ruberrimus H110 was able to use raw glycerol (from biodiesel production) as a carbon source for carotenoid production, achieving high cellular growth (0.51 g/L) and productivity (0.0064 g/L h). Interestingly, compared to the fermentation process with pure glycerol, the use of raw glycerol increased both carotenoid concentration (approx. 27%) and productivity (1.5-fold). The authors also studied the addition of individual fatty acids (palmitic, stearic, oleic, and linoleic acids) to pure glycerol, observing that these have a favorable effect on carotenoid production, increasing, from 15 to 25%, the maximum carotenoid concentration, and, from 1.6 to 2.0-fold, the productivity rates. Similarly, to the addition of raw glycerol, the presence of palmitic and oleic acids also favored the torularhodin biosynthesis (proportion close to 66%). This is a clear example of the importance of balancing the nutritional content of the fermentation media, which can be properly adjusted, for example adding fatty acids as additives, to direct microbial carotenoid biosynthesis and, consequently, to improve the production yield of a specific carotenoid. As aforementioned, temperature and pH are also two important parameters for microbial growth and consequent pigment synthesis. For example, Shih and Hang (1996) highlight that acidic pH values, between 3.4 and 4.5, inhibit R. rubra cell growth and β- carotene production. However, the authors also observed that through a slight increase of the initial pH to 5.0, even maintaining the media slightly acidic, the maximum cell 18 concentration (0.131 mg/L) and β-carotene production yield (1041 μg/L) can be enhanced. Regarding the effect of temperature, Malisorn and Suntornsuk (2009) shown that the R. glutinis optimal growth is 30 °C, with approximately 2.3 g/L of cells, resulting in a consequent production of 0.178 mg/L of β-carotene. Although most of the studies focused on the optimization of temperature, pH, and nutritional content of the culture media, other processing parameters have been also studied. Tinoi et al. (2005) evaluated the influence of the shaking rate using shaker flasks to produce carotenoids by R. glutinis. It was identified that a balance of the agitation speed is of utmost importance, since at low shaking rates (100 to 150 rpm), the cell growth is reduced, probably because of the low availability of nutrients on the cell surface, while at high agitation rates, some disruption of cells can occur, reducing their viability. In this section, some approaches to improve carotenoid productivity yields or to adjust the production for a specific microorganism were discussed. However, instead of a single strain fermentation process, the production of carotenoids can be performed using consortia or mixtures of two microorganisms. As an example, Fang and Wang (2002) studied the production of astaxanthin, in a 1.5 L bioreactor, using a mixed culture of the yeast Xanthophyllomyces dendrorhous (formerly, Phaffia rhodozyma) and the bacterium Bacillus circulans. The process was carried out in a two-stage batch fermentation, i.e., first stage in which yeast fermentation occurred and a second one in which the bioreactor was subsequently inoculated with B. circulans. In the first stage, using solely the X. dendrorhous cells, after 72 h, a total of 9.01 mg/L of astaxanthin were produced. The second stage was started with inoculation with B. circulans in the bioreactor, and after 144 h of incubation, the production of astaxanthin was slightly increased (10.07 mg/L). Although only a 10% increase after the incubation of the second microbial strain was observed, the use of these consortia is interesting as B. circulans has a lytic enzyme activity of the yeast cell walls, providing a highest extraction yield of total carotenoids (over 96%) during the second fermentation stage. Summing up, a successful production of many carotenoids can be easily attained through the fermentation of several microorganisms. However, to maintain efficient production yields and to conduct a specific biosynthesis pathway, careful control of the 19 nutritional and processing parameters is essential. Moreover, independently of the upstream bioprocessing, after the fermentation, the carotenoids will remain inside the cells (in the biomass content), requiring proper integration with further downstream processing stages for the efficient extraction and recovery of the microbial carotenoids. 2.4 New genetic engineering approaches to increase carotenoid production yields One of the strategies to reduce production costs and increase yields is the development of bioengineered hyper-producing strains. Metabolic engineering is the improvement of cellular properties through the modification of specific biochemical reactions, with the use of recombinant DNA technology (Park et al. 2007). Moreover, genetic engineering in non-carotenoid-producing microorganisms is naturally a very useful tool, since it allows the manipulation of “well-known” (i.e., well-defined and understood metabolic pathways) microorganisms to enhance carotenoid productivity yields (Ye and Bhatia 2012). Since the beginning of this century, several authors have made use of metabolic engineering tools in yeasts, such as Saccharomyces cerevisiae and Candida utilis, which were successfully modified by inserting carotenogenic genes from Erwinia uredovora, Agrobacterium aurantiacum, and Xanthophyllomyces dendrorhus to produce carotenoids, such as, β-carotene, lycopene, or astaxanthin (Misawa and Shimada 1998; Miura et al. 1998; Bhataya et al. 2009; Ungureanu et al. 2013). Another successful example was the production of lycopene using the yeast Yarrowia lipolytica, a natural non-producer of carotenoids, simply by introducing two genes, phytoene synthase and phytoene desaturase. After the integration of the heterologous genes crtB and crtI, the transformants appeared orange in color, indicating lycopene formation. This transformation increases the specific lycopene content, reaching a yield of 16 mg/g (dry cell weight) (Matthäus et al. 2014). Pichia pastoris, another non-carotenogenic yeast, was designed and constructed by adding two plasmids pGAPZA-EBI* and pGAPZA-EBI*L containing the genes encoding lycopene and β-carotene. The results obtained by Araya- Garay and collaborators (2012) showed that the recombinant strain produced both 20 lycopene and β-carotene, reaching 1.141 and 339 μg/g (dry biomass), respectively (Araya-Garay et al. 2012). These successful examples demonstrate that the use of genetic engineering can be beneficial for the increase of carotenoid production yields, appearing as effective strategies to improve the production of specific microbial carotenoids and, thus, meeting the world demand for carotenoids in animal feed, cosmetics, food, beverages, and pharmaceutical industries. 2.5 Extraction methods for the recovery of intracellular carotenoids The microbial production of carotenoids is intracellular, and, like other intracellular bioproducts, after the fermentation, a series of downstream operation units are included for carotenoid recovery and processing. In general, in the first clarification stage, i.e., using conventional filtration or centrifugation operations, the cellular biomass, which contains the intracellular carotenoids, is separated from the supernatant. Further, to facilitate the release of intracellular carotenoids, the recovered cells are disrupted, applying at least one of many different physical, chemical, and/or biological cell-disrupting methods. After the partial or total disintegration of the cell structure, the intracellular carotenoids are then extracted and separated from the cell debris. It is important to note that both cell disruption and extraction stages can be integrated into a single operation unit or carried out through different operation units as, for example, combining a chemical pre-treatment of the cells with a further Soxhlet extraction. Afterward, further downstream processing stages for the saponification and separation of a specific carotenoid can also be carried out. Among all the downstream processing steps, the extraction and recovery stages are the critical ones, which are briefly discussed in the next paragraphs. Considering the wide range of carotenoid producers and their cellular variety and complexity, the choice of the most adequate method(s) appears as a key to obtain a complete cell disruption or a selective cell-membrane permeabilization and, consequent, carotenoid release. Gram-positive cells have an inner membrane and strong cell wall, while Gram- negative cells have both inner and outer membranes (less rigid than Gram-positive cells). These bacterial cells are yet more fragile than yeast and microalgal cells, which are 21 composed of dynamic, complex, and rigid cell walls. As expected, bacterial cell disruption is easier than the disruption of yeast and microalgal microorganisms. Moreover, as recently highlighted by Saini and Keum (2018), usually carotenoids are strongly associated with other intracellular macromolecules (for example, proteins and fatty acids), hindering their mass transfer process. Therefore, if the intention is to extract intracellular carotenoids from robust cells, most intense methods should be selected, as for example, cell cooking, cryogenic grinding, and/or using chemical agents (acids, base, surfactants, or volatile organic solvents (VOCs)). Conversely, the extraction of an intracellular product from Gram-negative cells can be simply achieved by ultrasonication or a freezing-thawing processes. In addition to the microbial cell characteristics, the relevant carotenoid properties must also be considered, namely, (a) the hydrophobic nature of these biomolecules and (b) the oxidative properties of carotenoids, which can be reduced in the presence of heat, light, acids, and long extraction times (Saini and Keum 2018). Further, independent of the cell disruption procedure, the effective disintegration of cells (total or partial) is always a prerequisite for the efficient extraction of intracellular carotenoids, particularly as disruption facilitates the entry of the solvent and the subsequent carotenoid solubilization. Recently, Saini and Keum (2018) have completely revised the conventional and non-conventional extraction procedures applied in the recovery of target carotenoids, as schematized in Figure 2.3. Considering the carotenoid characteristics, particularly due to the carotenoid hydrophobicity, most traditional extraction processes use volatile organic solvents (VOCs) as solubilizing agents. Although VOC-based processes allow high extraction yields, they exhibit several human health and environmental risks (Salar-García et al. 2017). Thus, to over-come some of these concerns in the last few years, several researchers have been searching for novel alternative and efficient techniques, particularly, (a) replacing the VOCs with greener, biocompatible, and less toxic solvents, such as supercritical fluids, biosolvents, or ionic liquids (Yara-Varón et al. 2016) and (b) reducing the amount of solvent required through the combination of the chemical extraction with novel physical (microwave and ultrasound- assisted extractions) or biocatalytic (enzyme-assisted extraction) procedures. As shown in the scheme of Figure 2.3, VOCs are mostly used in conventional techniques, such as atmospheric liquid extraction with maceration or Soxhlet extraction, but depending of the 22 solvent type, they can be associated with some of the non-conventional procedures, such as ultrasound-assisted extraction (UAE) or enzyme-assisted extraction (EAE), while non- conventional solvents are mainly applied in novel processing techniques, since they are more biocompatible, like biosolvents and green solvents, or technique-specific, i.e., supercritical fluid extraction (SFE). 23 Figure 2.3. Conventional and non-conventional techniques for cell disruption and carotenoid extraction (adapted from Saini and Keum 2017). 24 As previously highlighted, most of the academic studies and industrial processes use VOCs to extract carotenoid. These solvents are particularly interesting because of their high carotenoid solubilizing potential, as well as their cell-disrupting capability, through permeabilizing walls and membranes. The organic solvent penetrates the microbial cells, dissolving the intracellular carotenoid molecules according to the characteristics of the extractant (or permeabilizing agent). Frequently, non-polar solvents, such as hexane, petroleum ether, or tetrahydrofuran (THF), are excellent choices for the extraction of non-polar carotenoids, whereas polar solvents, like dimethyl sulfoxide (DMSO), acetone, ethanol, and ethyl acetate, are more suitable for the extraction of carotenoids with polar characteristics (Saini and Keum 2018). The influence of different VOCs on the recovery of carotenoids is being largely studied, for example, Valduga et al. 2009b, evaluated the extraction capability of different combinations of VOCs (acetone, petroleum ether, methanol, hexane, ethyl acetate, DMSO, chloroform) with liquid N2, through the conventional procedure using VOC atmospheric liquid extraction with successive macerations of cellular biomass from Sporidiobolus salmonicolor CBS 2636. Interestingly, the maximum concentration of total carotenoids (253.8 μg/g) was obtained in a combined treatment, using liquid N2 and DMSO, to disrupt the cell, followed by a liquid extraction with an acetone/methanol (7:3 v/v) organic solution. Similarly, Park et al. (2007) also evaluated the effect of different VOCs (DMSO, petroleum ether, acetone, chloroform, and hexane) in the recovery of carotenoids from R. glutinis cells. In their work, instead of maceration, the authors first lyophilized the yeast cells and then added each organic solvent mixture directly to the biomass. The integration of liquid extraction and lyophilization was an effective procedure to recover the total carotenoids produced from R. glutinis cells, with the lowest extraction capability obtained with hexane (0.19 mg/g) and the highest recovery yields with both DMSO (0.23 mg/g) and petroleum ether (0.24 mg/g) solvents. As shown in Figure 2.3, some of the most effective and innovative approaches combine conventional VOCs with non-conventional techniques, even at the industrial scale. In several countries, for example, the industrial extraction of 25 food-based carotenoids involves the use of commercial enzyme preparations in combination with organic solvents, such as hexane and ethyl acetate (Lavecchia and Zuorro 2008). A more complex approach combining conventional Soxhlet extraction using VOCs (with DMSO and acetone) and non-conventional ultrasonication was recently evaluated as an alternative for the extraction of carotenoids from R. mucilaginosa. This innovative approach allowed an increase of the concentration of total carotenoids, 317.6 μg/g (i.e., equivalent to 91.46 (μg/g) of β-carotene, 152.44 (μg/g) of torulene, and 73.04 (μg/g) of torularhodin) recovered under milder processing conditions (at 25 °C) (Cheng and Yang 2016). The combination of VOCs and non-conventional techniques proved to be more efficient than the direct application of solvents in the liquid extraction of the carotenoids. Although VOCs are widely used for the extraction of biomolecules due to advantages like high vapor pressures (easily evaporated at room temperature), low cost, and high commercial availability, they exhibit serious disadvantages regarding the bioproducts contamination, low biodegradability, and high atmospheric toxicity (“greenhouse” effect) (Datta and Philip 2018). Therefore, several alternatives have been proposed, varying from the so-called “green” solvents or biosolvents to the supercritical fluids. The search for alternative (non-conventional) solvents intends to minimize the environmental impacts and to increase the sustain-ability and biocompatibility of the overall carotenoid manufacturing process. “Green” solvents are a general classification for more environmentally friendly solvents, i.e., those that comply, at least, with most of the 12 principles of green chemistry (Anastas and Warner 1998). The biosolvents are, in general, obtained from renewable resources, like wood, starch, fruits, and vegetable oils, or from petrochemical products that are non-toxic and biodegradable (Yara-Varón et al. 2016). Yara-Varón et al. (2016) evaluated the capability of several biosolvents (cyclopentyl methyl ether, dimethyl carbonate, ethyl acetate, isopropyl alcohol, and 2-methyltetrahydrofuran) as possible substitutes for hexane in the extraction of microbial carotenoids, using two predictive models, the solute-solvent Hansen Solubility Parameters (HSPs) and Conductor-like Screening Model for Realistic Solvation (COSMO-RS). The use of predictive methods is a valuable tool to 26 understand the molecular interaction of solvents with carotenoids, avoiding extensive experimental studies, allowing a solubility scale of different carotenoids in a wide range of solvents to be easily obtained. The “green” solvents were effective for the recovery of β-carotene, particularly cyclopentyl methyl ether and 2- methyltetrahydrofuran, which gave extraction yields higher than those obtained with conventional solid–liquid extraction by maceration using hexane. Other classes of compounds that have been largely regarded as “green” solvents are ionic liquids (ILs). ILs are commonly defined as salts with a melting point below 100 °C, obtained through the combination of different organic cations and organic or inorganic anions (Chatel et al. 2014). Due to their ionic nature, ILs are wide versatile compounds, exhibiting adjustable solvent properties with an adaptability that is virtually impossible for any other class of other molecular solvents (Feldmann and Ruck 2017). As, for example, through the choice of a cation-anion combination, it is possible to design a suitable solvent, possessing specific conductivity, hydrophobicity, polarity, and solubility, based on the nature of the target solute (Kumar et al. 2017). Moreover, in the last few years, several families of ILs have been classified as eco-friendly in nature, due to the low vapor pressures (negligible volatilities), non-flammability, and high chemical and thermal stabilities (Oliveira et al. 2016). Regarding carotenoids extraction, some studies have already demonstrated the effectiveness of ILs (ILs from the imidazolium-, pyridinium-, and ammonium- based families), as solvents and permeabilizing agents, to extract carotenoids (like astaxanthin) from Haematococcus pluvialis microalgae (Praveenkumar et al. 2015) or from non-microbial origin (Saini and Keum 2018). For example, Praveenkumar et al. (2015) used a series of imidazolium- and pyridinium-based ILs as alternative solvents for the extraction of astaxanthin from Haematococcus pluvialis microalgae, using a simple liquid-liquid extraction procedure at room temperature. The addition of ILs damaged and deconstructed the cyst cell wall, facilitating the release of the astaxanthin. The highest extraction capability was obtained (19.5 pg of astaxanthin per cell) with 1-ethyl-3- methylimidazolium ethyl sulfate, in a very short extraction time (1 min of exposure 27 time); a process 82% more efficient than the conventional procedure using ethyl acetate and French-press-cell homogenization. As shown, it seems that ILs are promising, highly efficient, and biocompatible alternatives for the extraction of carotenoids. However, only a few studies have reported the extraction of microbial carotenoids using ILs. Additional studies are essential to fully validate the effectiveness of ILs as extractive agents of microbial carotenoids, but it is important to highlight that, because of the wide range of cation– anion combinations, these further studies should focus on ILs with eco-friendlier characteristics, i.e., low environmental impact and toxicities, high biodegradability, and those that can be easily obtained from renewable sources. Similarly, another class of “green” solvents that are of interest to the scientific community are Deep Eutectic Solvents (DES), in particular due to their low toxicities and reduced adverse environmental effects. These characteristics resulted in a rapid increase in the number of applications using DES for the extraction of bioactive compounds (Zainal-Abidin et al. 2017). Very recently, a pioneering work was performed by Cicci et al. (2017), in which intracellular biomolecules were recovered from microalgal biomass of Scenedesmus dimorphus by combining DES (composed of 1,2-propanediol, choline chloride, and water, in a 1:1:1 molar ratio) and UAE. DES solvents were effective in extracting a large number of intracellular carotenoids, with approximately 0.11% of carotenoids being recovered, as a proportion of the total biomolecules extracted. Although the use of DES is quite new, and no further studies have focused on the extraction of microbial carotenoids, DES-based processes were already successfully applied in carotenoid recovery from other sources (animal and vegetal). For example, astaxanthin was obtained from shrimp carotenoids using DES (Zhang et al. 2014). After evaluating different conditions, the combination of ultrasound processing with DES as solvents was established as the most efficient platform for the extraction of astaxanthin, achieving extraction yields (146 g/g), higher than an ultrasound method with ethanol as solvent (102 g/g) (Zhang et al. 2014). Lee and Row (2016) also studied the extraction of astaxanthin from Portunus trituberculatus (marine crab) using DES-based processes. The authors observed that the 28 astaxanthin extraction yields are enhanced 155% using DES (composed of methyl- tri-phenyl-phosphonium bromide and 1,2-butanediol, in 1:4 molar ratio) as additives in an acetone-based extraction procedure (73.49 mg/g), in comparison to the use of IL, 1-ethyl-3-methylimidazolium bromide (47.30 mg/g) as additive (Lee and Row 2016). These examples clearly demonstrate that DES have strong potential to be used as alternative solvents for the recovery of microbial carotenoids. Particularly, natural DES, constituted by amino acids, organic acids, sugars, or choline derivatives, fully accomplish most of the green chemistry principles (Paiva et al. 2014), and we believe that they can be next generation of solvents in biocompatible carotenoid processes. Contrarily to ILs and DES, supercritical fluids are one of the non-conventional solvents mostly studied in literature for the extraction of microbial carotenoids, with supercritical fluid extraction (SFE) being a widely known and applied technique, even at the industrial scale. This process uses non-flammable, non-toxic, and recyclable solvents under conditions close to the critical point as an extractant of non-polar carotenoids (Saini and Keum 2018). Usually, carbon dioxide (CO2) or ethanol are used as solvents (Johner and Meireles 2016). SFE is free of toxic waste, does not require post-processing for solvent removal, and does not cause thermal degradation of the biomolecules (Mezzomo and Ferreira 2016). Lim et al. (2002) compared the extraction efficiency between acetone-based conventional liquid extraction and SFE using carbon dioxide (CO2) to recover astaxanthin from Phaffia rhodozyma red yeasts. The highest astaxanthin extraction yield (90%) was attained using CO2 (50 g), with the temperature being a key parameter in the extraction. At 40 °C and 500 bar, an increase of the concentration of astaxanthin by about threefold, reaching thirteenfold at 60 °C, was observed, in comparison with a conventional liquid extraction using acetone at same temperature conditions. Another successful example of the use of SFE (using CO2) for the extraction of microbial carotenoids was reported by Sajilata et al. 2010. In that work, the authors carried out SFE using methanol as an entrainer (i.e., modifier) to extract zeaxanthin from dried bacterial biomass of Paracoccus zeaxanthinifaciens, obtaining a maximum recovery of 65% of the total zeaxanthin content using 3 mL of 29 methanol per gram of lyophilized biomass, at 300 bar and 40 °C. Macı́as-Sánchez et al. (2005) have also extracted carotenoids from Nannochloropsis gaditana microalgae biomass using SFE with CO2 and methanol, comparing the yields with conventional liquid extraction using methanol as the extractant. SFE was carried out at the micro-scale at 60 °C and 400 mbar, and, after 180 min of processing, approximately, 0.343 μg/mg of the total intracellular carotenoids was recovered. Unfortunately, it was observed that the extraction of total carotenoids with methanol was more efficient than SFE, recovering approximately 0.8 μg/mg of the total carotenoids. This suggests that the non-conventional techniques is not always more effective than the simplest and traditional recovery approaches using VOCs. Therefore, before implementing SFE processing, it is fundamental to balance the eco-friendly and extraction advantages with the non-favorable characteristics, like low efficiencies, carotenoid degradation at the operating conditions required (high temperatures and pressures), equipment cost, and/or high-energy consumption, of the SFE. Here, we highlight that, specifically, CO2-based SFE has clear biotechnological advantages for the extraction of microbial carotenoids in comparison with the majority of the conventional procedures, which are namely, (a) low toxicity of CO2, (b) overall cost–benefit of the SFE, and (c) separation and polishing of the recovered carotenoids. Other non-conventional techniques have been also successfully applied as alternatives for the extraction of microbial carotenoids, such as ultrasound-assisted extraction (UAE) (Dey and Rathod 2013; Goula et al. 2017; Parniakov et al. 2015; Gu et al. 2008; Singh et al. 2015). For example, Gu et al. (2008) evaluated the capability of UAE for the recovery of intracellular carotenoids from R. sphaeroides bacteria. In that study, solid-liquid solutions (50 of solvent per g of dried biomass) using acetone as extractant were prepared and then subjected to ultrasonic processing at 500 W. After the sonication, approximately 664 μg/g of the total intracellular carotenoids was recovered. Singh et al. 2015, using response surface methodology, optimized a series of UAE processing parameters [solvent (acetone)/CDW ratio of 67.38 μL/mg, power 27.82% (total power 500 W), pulse length of 19.7 s, and extraction time of 13.48 min], achieving extraction yields of 30 zeaxanthin (11.2 mg/g) and β-carotene (4.98 mg/g) from the green microalgae Chlorella saccharophila (Singh et al. 2015). The UAE methods using VOCs as extractant agents can significantly increase the carotenoid extraction yield when compared to the conventional techniques, but a proper optimization of several factors, such as ultrasonic power, intensity, temperature, and sample/solvent ratio, is of utmost importance. In summary, several techniques and solvents can be used for breaking cells and extracting intracellular carotenoids, but the efficiency of each method is always dependent on a combination of factors, namely, the microbial biomass, carotenoid nature, and operation conditions. Therefore, the choice of the appropriate method must consider the cost-effectiveness, environmental safety, processing efficiency, and reproducibility. We believe that the use of “green” and biocompatible solvents will overcome some of the environmental and processing drawbacks, particularly, if combined with non-conventional and innovative procedures, appearing thus as more efficient and environmentally friendly platforms for the recovery of a wide range of microbial carotenoids. However, it is essential to create effective and economical integrative platforms for recycling the solvents used in carotenoid extraction. VOCs, for example, are volatile organic solvents that can be easily and efficiently recovered through distillation, but in the case of ILs and DES, the recyclability appears yet as the greatest challenge to be overcome. However, as recently reviewed by Ventura et al. (2017), the development of effective strategies for ILs and DES recycling and carotenoid isolation is already in progress. Particularly, the authors highlight that the integration of the extraction stages with further aqueous biphasic systems (ABS) units can be a promising alternative for solvent recycling or simply by adding anti- solvents, which can allow the carotenoids to crystalize or precipitate (Ventura et al. 2017). In any case, additional studies are up most of the importance to transform these ILs and DES-based extraction processes as realistic environmental and economical sustainable platforms for an industrial recovery of microbial carotenoids. 31 2.6 Metabolites extracted during the carotenoid extraction In the previous section, we focused on microbial carotenoid extraction. However, during the extraction processes, other intracellular microbial metabolites, like fatty acids, lipids, proteins, carbohydrates, among others, can also be co- extracted, increasing the complexity of further downstream processing stages required for their separation and subsequent carotenoid purification. For example, during the extraction of carotenoids, such as astaxanthin or β- carotene, from microalgae using VOCs, other essential fatty acids are simultaneously recovered. In the literature, fatty acid extraction from Porphyridium cruentum, Isochrysis galbana, and other microalgae (Medina et al. 1995; Giménez Giménez et al. 1997; Molina Grima et al. 2003) resulted in the simultaneous recovery of fatty acids, such as eicosapentaenoic acid (EPA), docosohexaenoic acid (DHA), and arachidonic acid (AA) (Molina Grima et al. 2003). Likewise, the co-extraction of carotenoids and proteins is widely common, in that case, certain particularities must be considered, such as the use of wet biomass (Román et al. 2002). Powls and Britton extracted a violaxanthin-binding protein from a photosynthetic route using hot methanol obtained from biomass of microalgae Scenedesmus obliquus (Powls and Britton 1977). More recently, Cicci et al. (2017) have shown that DES can be used as biocompatible solvents to extract carotenoids, carbohydrates, proteins, lipids, and pigments (chlorophyll) from Scenedesmus dimorphus. As briefly highlighted, several intracellular metabolites can be recovered during the extraction of the carotenoids. Thus, after the extraction stage, a full characterization of the extract is needed, characterizing not only the target carotenoids but also identifying which metabolites are co-extracted. Consequently, depending on the extract composition, th