Dissertations / Theses on the topic 'Fermentation techniques'

L'objectif de la thèse a été le développement et/ou l'amélioration de méthodes destinées à l'évaluation de la présence des mycotoxines dans les bières européennes, et de suivre le devenir de certaines mycotoxines pendant le processus de fabrication de la bière. Cinq mycotoxines ou groupes de mycotoxines de grande importance pour l'agriculture ont été étudiées : les quatre principales aflatoxines (AFB1 AFG1, AFB2, AFG2) l'Ochratoxine A (OTA), et la Deoxynivalenol (DON), la Zearalenone (ZEA) et les Fumonisines (FBs) B1 et B2 (FB1 et FB2). Les céréales utilisées pour la fabrication de la bière peuvent être contaminées simultanément par plusieurs mycotoxines. C'est pour cette raison qu'un objectif majeur de la thèse a été la mise en place de méthodes permettant la détection simultanée de plusieurs mycotoxines. Au cours de ce travail la détermination simultanée des mycotoxines a porté sur AFs et OTA ainsi que sur FBs et ZEA. De plus, la DON a été déterminée par une méthode immuno-enzymatique (Enzyme Linked Immunosorbent Assay : ELISA) et avec une méthode de chromatographie en Phase gazeuse couplée avec spectrométrie de masse (CPG-SM). Une comparaison de ces méthodes pour la détermination de l'OTA dans le vin est également présentée afin de démontrer l'applicabilité d'une méthode basée sur la purification par immuno-affinité, pour une autre boisson alcoolique susceptible d'être affectée par l'OTA<br>This thesis reports on the quantitative assessment of some of the major mycotoxins (Aflatoxins (AFs) Ochratoxin (OTA) Fumonisis (FBs) Zearalenone (ZEA) and Deoxynivalenol (DON)) in European beers. The simultaneous determination of AFs and OTA was possible through the development of new or improved methods. The pre-column derivatization techniques, bromination and UV-irradiation specific to AFB1 and AFG1 were evaluated and compared. A method allowing the simultaneous determination of FBs and ZEA was also developed. Alternative derivatizatin partners to o-phthaldialdehyde (OPA) for the pre-column derivatization of the FBs were investigated. N-acety-cysteine (N-AC) and thioglycerol (TG) were found to be good alternatives to 2 mercaptoethanol (2-ME) which is toxic and its derivatives are unstable. Don was determined using a screening method. Enzyme-Linked Immunosorbent Assay (ELISA) confirmed by Gas Chromatography-Mass Spectrometry (GC-MS). Incidence rates for AFs and OTA in beer were 0% and 67 % respectively (n=99). Mean contamination levels were 0. 026 ng mL-1 for OTA ranging from &lt;0. 007 to 0. 15 ng mL-1. Statistical analysis showed no significant influence of alcohol content, year of production and fermentation type on OTA content

Au cours des campagnes de vinification en 1984 et 1985, différentes biomasses bactériennes (leuconostoc oenos) ont été utilisées dans la pratique sur des vins rouges (Pinot noir, Gamay) et des vins blancs (Chardonnay). La réactivation préalable des bactéries lactiques lyophilisées est indispensable avant l'inoculation des vins. Les succès obtenus, surtout en 1985 dans des cas particulièrement difficiles montrent bien l'efficacité de ce procédé pour le déclenchement de la fermentation malolactique. Les supplémentations des vins en autolysat ou broyat de levures peuvent également favoriser la fermentation malolactique. Les acides aminés du vin, plus particulièrement l'arginine, constituent des substrats importants pour la croissance des bactéries lactiques.

Physical Science is one of the subjects in which learners perform poorly in most Namibian schools. For instance, in ║Kharas region where the study was conducted, from 2010 to 2016, the overall performance has ranged between 20-40%. Furthermore, according to the results analysis report from the Directorate of National, Examination and Assessments (DNEA), from 2012 to 2016, the ||Kharas region was ranked 11th, 12th, 13th, 14th and 10th in the regional rankings. From the Examiner’s reports, the most poorly answered questions in the National Senior Certificate examinations are those that require learners to explain and show understanding. Additionally, the reports highlighted that learners lack practical skills or lack exposure to experiments and demonstrations to enhance their understanding (Namibia, 20142015). It is against this background that this study sought to explore how teachers mediate learning of experimental techniques using fermentation and distillation of a traditional brew. In particular, the study addressed the following sub-questions:1) what are Grade 11 Physical Science teachers’ perceptions and experiences on the inclusion of indigenous knowledge during Science lessons? 2) What factors influence Grade 11 Physical Science teachers’ perceptions and experiences on the inclusion of indigenous knowledge? 3) How do Grade 11 Physical Science teachers teach experimental techniques conventionally- fermentation and distillation? 4) How does the inclusion of a practical demonstration on making Ombike enable or constrain Grade 11 Physical Science teachers when explaining concepts on experimental techniques - fermentation and distillation? Thus, the focus was on conceptual understanding through linking content to the context of the learners. The study was underpinned by an interpretive paradigm whose focus is on understanding people’s worldviews. Within the interpretive paradigm, a mixed methods (Quant-QUAL methods) case study approach was employed in two sequential phases with Physical Science teachers at one secondary school in Keetmanshoop. Data were collected using document analysis (to strengthen my context). In the first phase, a questionnaire with closed-ended and open-ended questions was used to gather quantitative and qualitative data from 17 Physical Science teachers. Additionally, I conducted interviews with two Physical Science teachers from one Secondary School in Keetmanshoop who were selected based on their expertise, teaching experience and their qualifications. Quantitative data were analysed using frequency tables and graphs. For phase two, lesson and the practical demonstration observations were videotaped and transcribed inductively. Thematic analysis was used to analyse the qualitative data. The study was informed by Vygotsky’s (1978) socio-cultural theory with a focus on the mediation of learning, social interactions and learner engagement within the learners’ Zone of Proximal Development (ZPD) in conjunction with Shulman’s (1987) theory on pedagogical content knowledge (PCK). The findings of the study indicated that the majority of the Physical Science teachers were aware of and supported the inclusion of IK; on the other hand the teachers indicated some challenges that come with IK integration. Some teachers acknowledged the importance of IK, such as the enhancement of understanding, proving of theories, and arousal of interest and context relevance. However, some teachers felt that there are misconceptions in IK that can be carried into the classroom, IK resources like textbooks are not available, and IK is not integrated in the syllabus and examinations and lack of teacher training inter alia.The study recommends that the Ministry of Education, Arts and Culture especially the Division of Curriculum Planning and Development (NIED) needs to explicitly include IK in the Syllabus documents and suggest the inclusion of IK in the textbooks to the publishers to address these challenges. In order to enhance the teaching and learning with IK in Science, it is critical that teachers’ training workshops are conducted and IK resources are designed so that IK integration can be successfully implemented. For that to happen, it is recognized that the teachers’ PCK should be regularly upgraded through Continuous Professional Development (CPD) by the Subject Advisors (SA) so that the teachers are consistently guided on appropriate teaching methodologies and pedagogies about the inclusion of IK.

Cinq techniques electro-membranaires ont ete considerees pour recuperer l'acide tartrique contenu dans les solutions de detartrage chimique des cuves de vinification, plus ou moins riches en polyphenols. Tout d'abord, en vue d'eliminer les grosses molecules phenoliques susceptibles d'encrasser les membranes, diverses voies de traitement prealable des solutions de detartrage, par adsorption et par oxydation chimique, ont ete examinees. La preconcentration des solutions trop diluees peut ensuite etre effectuee a la temperature ambiante avec un empilement d'electrodialyse (montage n1) sans modifier la nature des substances. Avec un agencement de membranes echangeuses d'anions et de membranes echangeuses de cations, utilisant 3 circuits hydrauliques, l'empilement d'electro-metathese (montage n2) ou la cellule d'electro-electrodialyse (ou electrolyse a membrane) (montage n3) obtenus permettent de recuperer separement l'acide tartrique pur en solution concentree et la base. La production concomitante d'une solution d'acide tartrique et d'une solution de soude est alors entravee par formation de precipites tartriques qui resuident le taux de recuperation de l'acide. Cette insolubilite est mise a profit dans un empilement d'electro-precipitation (montage n4), qui se revele rapidement inutilisable. L'extraction du seul acide tartrique est alors effectuee soit par substitution ionique a partir d'ions provenant du catholyte d'une nouvelle cellule, soit par electro-hydrolyse dans un empilement comportant des membranes bipolaires (montage n5 et 6). La solution d'acide tartrique pur obtenue, avec un taux de recuperation et un rendement electrique rentables, a une concentration suffisant pour une utilisation ulterieure en agro-alimentaire, tandis que la solution basique est reutilisable. L'ensemble de ces resultats montre qu'il est possible de traiter les solutions ayant servi au detartrage chimique pour produire de l'acide tartrique dans le cadre du rejet zero

La demande croissante pour des chocolats de qualité supérieure a généré un besoin de diversification des arômes dans les gammes de chocolats proposées par les producteurs. La qualité du chocolat découle fortement du potentiel aromatique du cacao, qui est le résultat des réactions biochimiques qui ont lieu pendant la culture, la récolte, le traitement post-récolte et la transformation des fèves de cacao. Les composés volatils et non volatils du cacao contribuent à la perception sensorielle finale du chocolat. Au cours de la transformation, les arômes caractéristiques du chocolat se développent principalement pendant la fermentation, le séchage, la torréfaction et, dans une moindre mesure, le conchage. Bien que la fermentation du cacao ait un impact important sur sa qualité, et qu'elle soit étudiée depuis plusieurs décennies, il s'agit toujours d'un processus empirique et non maîtrisé. Les conditions de torréfaction et de conchage ont également un impact significatif sur la transformation des précurseurs d'arômes obtenus au cours de la fermentation, principalement en raison des réactions de Maillard qui ont lieu au cours de ces étapes. Dans un objectif de compréhension des mécanismes de formation de la qualité aromatique et sensorielle du chocolat, des essais de fermentation, de torréfaction et de conchage, tous dans des conditions spécifiques, ont été réalisés. Les caractéristiques, chimiques et sensorielles des fèves de cacao, tout au long de leur transformation en chocolat, ont ensuite été étudiées. Une attention particulière a été portée sur l'impact de la fermentation sur le développement des composés aromatiques et de leurs précurseurs. L'étude de l'utilisation de starters levuriens pour les fermentations s'est révélée particulièrement intéressante. Elle a permis de caractériser son impact et celui du temps de fermentation sur les différences observées dans la taxonomie microbienne, fongique et bactérienne, dans la masse en fermentation. En effet, la composition du microbiote peut influencer considérablement la composition volatile et non volatile des fèves, ce qui se traduit par des différences dans les profils aromatiques perçus au cours de l'analyse sensorielle des liqueurs et des chocolats obtenus. Enfin, des modèles prédictifs ont été développés afin de prédire les arômes des produits du cacao sur la base de leur composition chimique, en tenant compte des conditions de traitement auxquelles ont été soumises les fèves. Globalement, l'étude vise à mieux comprendre la formation des arômes dans le cacao et fournit des outils importants pour la production de chocolats capables de présenter des notes sensorielles fines, uniques et très recherchées, afin de mieux répondre à la demande croissante des consommateurs<br>The increasing demand for high-quality chocolate creates the need to diversify the range of flavors offered by chocolate producers. Cocoa quality derives strongly from its flavor, which may in itself be seen as the result of the biochemical reactions that take place during the cultivation, harvest and post-harvest processing of the cocoa beans. Both, volatile and non-volatile compounds contribute to the final flavor perception of cocoa. During processing, the characteristic flavors of chocolate are developed mainly during fermentation, drying, roasting and, to a lower extent, conching. Although cocoa fermentation has an important impact on the homogeneity of cocoa quality and has been studied for several decades, it is still an empirical and not mastered process. Roasting and conching conditions also have a significant impact on the subsequent transformation of the flavor precursors obtained during fermentation, mainly driven by the non-enzymatic Maillard reactions that take place during these steps. By carrying out fermenting, roasting and conching trials, all under specific conditions, and by later proceeding to the analysis of the physical, chemical and sensory characteristics of cocoa beans throughout their transformation all the way into chocolate, it is hoped to better understand the mechanism of aroma formation and its link with flavor perception. In this study, an important focus has been placed on the impact of fermentation on the development of aroma compounds and their precursors. The use of yeast starter cultures during fermentation has been of special interest in this study, as well as their impact and that of fermentation time on the compositional differences in the taxonomy of fungal and bacterial microbiota present in the fermentation mass throughout the entire process. This, because the microbial composition has the potential of greatly influencing the resulting volatile and non-volatile composition of the beans, which translates into differences in the perceived flavor profiles of the liquors and chocolates obtained thereof. Lastly, predictive models have been developed in an attempt predict the flavor of the cocoa products based on their chemical composition, taking into account the processing conditions to which they had been submitted. Globally, the study aims to gain a deeper understanding of flavor formation in cocoa and provides important tools for the production of chocolates capable of displaying unique and highly sought-after flavors in an attempt to better meet the increasing demand for fine flavor cocoa and chocolate products

A series of studies was conducted using SEM in conjunction with chemical analysis, in situ and in vitro digestion techniques, to characterize the anatomical components from barley, corn, sorghum and wheat grains which constitute "fiber" and investigate their susceptibility to rumen microbial digestion. Fractured grains were used to identify anatomical features and cell types prior to and after extraction or digestion. Certain anatomical features, including pericarp tissue, aleurone cells, endosperm cell walls, corneous and floury endosperm tissue and lemma and palea from barley, were easily identifiable in fractured and ground grains, and in neutral detergent extracted or digested residues. In situ and in vitro incubation conditions were varied to assess the effect of concentrate and/or reduction of pH on the disappearance of identifiable grain fractions. In situ incubations were conducted using steers adapted to 0-, 30- and 90% concentrate diets. In vitro inoculum buffered at pH 7 or 6 was provided by a steer fed 0- or 90% concentrate. Tissues resistant to rumen microbial digestion during extended (144-h) in situ incubations and shorter term (12- to 48-h) in vitro incubations were primarily those identified in NDF, and included pericarp, lemma and palea, and small amounts of corneous endosperm. Remaining tissues identified included barley lemma, palea and pericarp; corn pericarp, tip cap and small amounts of corneous endosperm; sorghum pericarp and corneous endosperm with matrix and protein bodies; and wheat pericarp. In vitro disappearance of isolated NDF after 48-h ranged from 43% for barley to 89% for corn. Labile structures included embryonic tissue and portions of endosperm cell walls, protein matrix and residual starch. Resistant tissues included pericarp, aleurone cell walls, tip cap and portions of the corneous endosperm. Relative rankings of NDF digestibility under all conditions studied were similar (corn > sorghum > wheat > barley) whether determined using isolated NDF or calculated from TIVDMD residues. Neither concentrate level fed to the host animal nor pH of the in vitro incubation flask affected rankings among grains, although increasing concentrate level and/or reducing pH appeared to reduce in vitro NDF disappearance. Evaluation of electron micrographs of fractured grains suggested that similar anatomical structures in the various grains differed in their resistance to microbial digestion. For example, pericarp from barley and wheat appeared to be more resistant than that from corn or sorghum. Endosperm of barley was less resistant than that of sorghum.

Made available in DSpace on 2014-06-11T19:28:22Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-02-17Bitstream added on 2014-06-13T18:34:30Z : No. of bitstreams: 1 ribeiro_af_me_jabo.pdf: 572955 bytes, checksum: f64dd01902ef75161ee1881df3afa8ed (MD5)<br>Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)<br>Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)<br>A avaliação nutricional de alimentos para ruminantes tem sido de grande importância para adequar os bancos de dados dos sistemas de formulação de dietas em regiões de clima tropical. Devido à grande quantidade de opções de forrageiras, várias alternativas de avaliação dos alimentos foram desenvolvidas nas últimas décadas, entre elas, as técnicas in vitro. O objetivo desse trabalho foi avaliar a composição química, digestibilidade e a produção de gases in vitro de três cultivares de Brachiaria: Brachiaria brizantha cv. Marandu, Brachiaria brizantha cv. Xaraés e Brachiaria híbrida cv. Mulato, submetidas a níveis de oferta (4, 7, 10 e 13% do peso vivo animal), sob ciclos de pastejo rotativo. A cultivar Xaraés apresentou os maiores teores de FDN, FDNcp, FDA, LIG e menor produção de gases em 96 horas no componente lâmina foliar . Houve diminuição dos CNF e fração B2 e o aumento da fração C dos carboidratos nos componentes colmo e lâmina foliar entre ciclos de pastejo em todos as cultivares. A cultivar Marandu apresentou maiores valores de digestibilidade (DIVMO e DIVMS) nos dois componentes. A cultivar Marandu apresentou a melhor composição química e maiores valores de digestibilidade<br>Nutritional evaluation of ruminant feed has been of great importance to adapt the databases of the systems of feed formulation in tropical regions. Due to the large amount of various alternative options for forage food assessment were developed in recent decades, among them the in vitro techniques. The aim of this study was to evaluate the chemical composition, digestibility and gas production in vitro of three cultivars of Brachiaria, Brachiaria brizantha cv. Marandu, Brachiaria brizantha cv. Xaraés and Brachiaria hybrid cv. Mulato, subjected to offer levels (4, 7, 10 and 13% of live weight) under rotational grazing. Xaraés cultivar showed the highest contents of NDF, NDFcp, ADF, lignin, and lower gas production in 96 hours in the stem and leaf fractions. There was a decrease in NFC and B2 fractions and increased the fraction C of carbohydrates in the stem and leaf fractions between grazing cycles in all the cultivars. Marandu cultivar showed the highest digestibility (IVDMD and IVDOM) in both fractions. The leaf fraction had lower levels of fiber and higher protein content in relation to the stem fraction. Cultivar Marandu had the best chemistry and higher digestibility values

Thesis (MSc)--Stellenbosch University, 2014.<br>ENGLISH ABSTRACT: Producers of South African (SA) dry and semi-dry Chenin blanc table wines are currently experimenting with winemaking techniques to modulate the flavours of the predominantly fruity styles of this genre. An important stage during wine style development is sensory profiling paired with consumer acceptance testing, before wine is produced on industrial scale. With those aforementioned goals in mind, this study was conducted in partnership with two commercial SA wine cellars. The main focus of the study was an investigation into the treatment effects of two winemaking techniques, respectively grape skin contact and natural fermentation, on the sensory profiles of experimentally produced Chenin blanc wines. Results obtained with descriptive sensory analysis (DA) of the wines were compared to those obtained by two rapid sensory profiling methods, namely projective mapping (PM) and frequency of attribute citation (FC). A consumer preference study was also done on the wines. In order to understand the treatment effects better, the dominant non-Saccharomyces yeasts that were present during the natural fermentations were identified using polymerase chain reaction (PCR), while the major volatile chemical compounds were identified with chromatography and mass spectrometry. The sensory and chemical profiles of the naturally fermented wines were significantly different from those of the inoculated wines. PCR analysis identified some of the yeasts present during alcoholic fermentation. In comparison to the inoculated fermented wines, the naturally fermented wines were generally perceived to have more intense and riper tropical fruity aromas, with enhanced sweetness and reduced intensities of sourness, bitterness and astringency. The wines fermented on the skins (FoS) had lower ester and volatile thiol concentrations than the control wines (with no skin contact) and limited skin contact wines. Sensory attributes linked to the FoS wines included aromas of dried fruit, dried grass and a vegetative character, with an increased sour and bitter taste and astringent mouthfeel. In contrast, the wines that were produced with limited skin contact (12 hours) retained their tropical fruity aromas better than the FoS wines. Limited skin contact seemed to have had a less harsh effect on the taste and mouthfeel than the FoS wines. A consumer study was done to establish a Generation Y consumer group’s (18-35 years) preference for the different treated wines. Overall, the naturally fermented wines, which were described as having a strong tropical fruit character, were preferred. The FoS wines were generally disliked by the consumers. Results obtained from the three sensory analysis methods, respectively DA, PM and FC, were similar, thereby confirming the suitability of the rapid methods PM and FC, to extract qualitative information from the sensory profiling of white wine. The results of this study made a significant contribution towards validation of rapid sensory methods for wine evaluation, which are particularly valuable in the context of sustainability and technology transfer to research and industry alike. The knowledge gained on the chemical profiles of SA Chenin blanc is novel and this is one of the first reports on the volatile thiol content of SA Chenin blanc wine.<br>AFRIKAANSE OPSOMMING: Produsente van Suid-Afrikaanse (SA) droë en semi-droë Chenin blanc tafelwyne eksperimenteer tans met wynmaaktegnieke om die oorwegend vrugtige style van hierdie genre te varieer. ‘n Belangrike fase van wynstylontwikkeling is sensoriese profilering van die produkte met gepaardgaande verbruikersaanvaardingstoetse, voordat die wyne op industriële skaal geproduseer word. Met dié bogenoemde doelwitte in gedagte, was hierdie studie in vennootskap met twee kommersiële SA wynkelders gedoen. Die hooffokus van die studie was om die invloed van twee wynmaaktegnieke, onderskeidelik druifdopkontak en natuurlike fermentasie, op die sensoriese profiele van eksperimenteel geproduseerde Chenin blanc wyne te ondersoek. Resultate wat met beskrywende sensoriese analise (DA) van die wyne verkry is, is vergelyk met dié wat deur twee vinniger sensoriese profileringsmetodes metodes, naamlik projektiewe kartering (PM) en frekwensie van kenmerkaanhaling (FC) verkry is. 'n Verbruikervoorkeurstudie is ook op die wyne gedoen. Ten einde die behandelingseffekte beter te verstaan, is die dominante nie-Saccharomyces giste wat teenwoordig was tydens die natuurlike fermentasie geïdentifiseer met behulp van die polimerasekettingreaksie (PCR), terwyl die vernaamste vlugtige chemiese verbindings geïdentifiseer is met chromatografie en massaspektrometrie. Die sensoriese en chemiese profiele van die natuurlike gefermenteerde wyne was beduidend anders as dié van die geiinokuleerde wyne. PCR analise het sommige giste wat teenwoordig was tydens alkoholies fermentasie geïdentifiseer. In vergelyking met die geiinokuleerde wyne, het die natuurlik gefermenteerde wyne intenser en ryper tropiese vrugtige aromas getoon, met ‘n verhoogde perspesie van soetheid en laer intensiteite van suurheid, bitterheid en vrankheid. Die wyne wat berei is deur fermentasie op die druifdoppe (FoS) het laer konsentrasies van esters en vlugtige swawelverbinding gehad, as die kontrole wyne (geen dopkontak) en beperkte dopkontak wyne. Sensoriese eienskappe gekoppel aan die FoS wyne was geasssosieer met gedroogde vrugte, gedroogde gras en 'n vegetatiewe aroma, met gepaardgaande verhoogde suur en bitter smaakpersepsies en ‘n frank mondgevoel. In teenstelling, het die wyne wat geproduseer is met 'n beperkte dopkontak (12 uur) hul tropiese vrugtige aroma beter behou as die FoS wyne. Beperkte dopkontak het ook ‘n minder negatiewe uitwerking op die smaakpersepies van bitterheid en vrankheid gehad. ‘n Verbruikerstudie is gedoen om 'n groep jong Generasie Y verbruikers (18-35 jaar) se voorkeure vir die verskillende behandelings te bepaal. Oor die algemeen is die natuurlike gefermenteerde wyne, wat beskryf is met 'n sterk tropiese vrugtekarakter verkies, terwyl die groep nie van die FoS wyne gehou het nie. Resultate wat verkry is met drie sensoriese analise metodes, onderskeidelik DA, PM en FC, was soortgelyk en die geskiktheid van die vinniger metodes, PM en FC, om kwalitatiewe inligting van witwyn se sensoriese eienskappe te verkry, is bevestig. Die resultate verkry met hierdie studie maak ‘n beduidende bydrae tot die validering van vinniger sensoriese metodes vir wynevaluering, wat veral waardevol is in die konteks van volhoubaarheid en tegnologieoordrag na navorsing en die industrie. Die nuwe kennis wat in hierdie studie gegenereer is met die chemiese profilering van SA Chenin blanc wyn is baie waardevol en hierdie is een van die eerste narvorsingsprojekte oor die inhoud van vlugtige swawelverbings in SA Chenin blanc wyn.

Thesis (PhD (Chemical Engineering))--Cape Peninsula University of Technology, 2020<br>The use of bioproducts in different commercial sectors such as medicine, agriculture, cosmetics, food, and chemical industries motivates the need for easy production and recovery techniques of bioproducts at laboratory and pilot scale. This study aims at the production and recovery of some selected bioproducts using supercritical fluid extraction techniques. Three products are chosen as case studies: these are ethanol, acetoin, and vanillin, since the ease of separation is influenced by the concentration of the product in the broth, these compounds were selected based on their concentration in the fermentation broth, according to literature sources. A standard method was developed in a spectrophotometer for quantifying the targeted product in the broth, while the product recovery studies was carried out using a supercritical fluid extraction pilot plant. Saccharomyces and Bacillus species were chosen for the bioproduction of the selected bioproducts. Experimental design and statistical analysis of results were carried out using response surface methodology (RSM) and artificial neural network (ANN). Studies on each of the selected bioproducts are as justified in the paragraphs below. Bioethanol production has recently become an increasing trend in research, with a focus on increasing its economic viability. Hence, the need to develop a low-cost fermentation medium with minimum redundant nutritional supplements, thereby minimizing the costs associated with nutritional supplements whereby inoculum preparation becomes necessary for ethanol production. Corn steep liquor (CSL) in glucose fermentation by Saccharomyces Type 1 (ST1) strain and Anchor Instant Yeast (AIY), which are low-cost media, are used as replacements for yeast extract (YE). The fermentation process parameters were optimized using artificial neural networks (ANN) and the response surface methodology (RSM). The study shows that for CSL, a maximum average ethanol concentration of 41.92 and 45.16 g/L representing 82% and 88% of the theoretical yield were obtained after 36 h of fermentation in a shake flask for ST1 and AIY respectively. For YE, ethanol concentration equivalent to 86% and 88% of theoretical yield were obtained with ST1 and AIY respectively after 48 h. Although, ANN predicted the responses of ethanol yield better than RSM, optimum conditions for ethanol production were better predicted by RSM. The consumers’ preference for ‘naturally’ produced aromas drives the development of bioproduction of acetoin from glucose with a view to optimize its production. The results revealed that by using a cheap nitrogen source, corn steep liquor, the yield of acetoin was similar to those of yeast and beef extracts. Furthermore, it was shown that by using Box-Behnken design, the optimum parameters such as glucose concentration, corn steep liquor, and inoculum size to maximize the concentration of acetoin produced were 78.40 g/L, 15.00% w/v and 2.70% v/v respectively. The validated concentration of acetoin produced in a triplicate analysis, 10.7 g/L, was 0.06% less than the predicted value. Increasing awareness of consumers of healthy, eco-friendly flavors and fragrances motivates the bioproduction of vanillin. The interactive effects of three variables on vanillin yield were evaluated by response surface methodology (RSM) with Box-Behnken design (BBD) model. The results showed the optimum conditions for the biotransformation of ferulic acid into vanillin can be achieved with maximum overall desirability (D) of 1.0 and a significant (p<0.05) quadratic model with regression coefficient (R2) of 0.995. Corn steep liquor, initial ferulic acid concentration and pH significantly influence the concentration of vanillin in the broth. The results in triplicate experiments confirmed vanillin yield of 386 mg/L after validation, which was in agreement with the prediction of the model. The maximum vanillin yield of 384.40 mg/L was predicted when corn steep liquor, ferulic acid concentration and pH were 7.72 g/L, 2.33 g/L, and 9.34 respectively. Fermentation system in a bioreactor has been proven to be an efficient system for the study of controlled fermentation variables when compared to a shake flask study. The influence of agitation, aeration, time and pH were analysed by Taguchi orthogonal array design for the upscale of acetoin in a bioreactor. The optimized parameters in 1.3L of fermentation vessel were as follows: 300 rpm agitation, 1.5 slpm aeration; 2 days’ fermentation time and 6.5 pH value. Agitation with above 70% was the most contributing factor and other variables were less than 30% in the percentage analysis of variance of each fermentation variables in the batch study of acetoin. A fourfold gain in acetoin titre (42.30 g/L) was obtained with the same substrate concentration in a lab-scale bioreactor on scaling up when compared with the shake flask batch study. The validated acetoin concentration of 41.72 g/L was obtained after a triplicate experiment to confirm the possibility of reproducing acetoin using the optimized conditions. Many separation techniques have been proven to recover value-added products from fermentation broth with a preference for several methods above other and new techniques that are emerging. Supercritical fluids separation using CO2 is one such technique. The feasibility of acetoin concentration and recovery was studied in supercritical CO2 pilot plant with pressure ranges of 100 to 300 bar, CO2 feed rate of 5 to 15 kg/h, at a process temperature of 37 and 80 °C in simulated and fermentation broth, respectively. The validated conditions for the fractionation of acetoin by supercritical fluid extraction (SFE) were determined as follows: extraction pressure, 300 bar; CO2 feed rate, 15 kg/h; extraction temperature 37 °C; and fractionation time of 30 minutes. At these operating conditions, the percentage recovery of acetoin with respect to the feed solution at the raffinate for the simulated and actual ermentation broth was 77.8% (0.20 g/L) and 77% (0.15 g/L) respectively. A two-fold extract increase was obtained after 30 minutes of fractionation. The study provides the technical feasibility and the base case data which are critical to the development and design of processes for production and recovery of acetoin. The lesson gleaned from this study may be extended to develop processes for the production and recovery of other bioproducts (ethanol and vanillin).

Ce travail a pour but d'etudier sur des pailles le traitement a l'uree (hydrolyse enzymatique) et ses effets sur la valeur alimentaire (realisation en milieu pauvre en eau et de faible cout). La source d'urease a ete la graine de soja depelliculee broyee. Au laboratoire les effets de plusieurs facteurs ont ete etudies: presence de paille, humidite, dose ou concentration en uree initiale, en soja, temperature. L'amelioration de la paille a ete caracterisee par l'azote additionnel total, fixe et la digestibilite de la matiere organique estimee par densitometrie et in vivo. L'activite ureasique native des pailles, mesuree par une methode originale a ete effectuee. Des traitements ont ete realises a moyenne puis a grande echelle, en suivant 2 techniques: arrosage d'homogeneisat et solide+eau asperge a part. La degradabilite du residu parietal traite ou non a l'ammoniac anhydre, marque ou non au carbone 14 sur les glucides a ete mesuree en minifermenteurs simulant le rumen

Resumo: A avaliação nutricional de alimentos para ruminantes tem sido de grande importância para adequar os bancos de dados dos sistemas de formulação de dietas em regiões de clima tropical. Devido à grande quantidade de opções de forrageiras, várias alternativas de avaliação dos alimentos foram desenvolvidas nas últimas décadas, entre elas, as técnicas in vitro. O objetivo desse trabalho foi avaliar a composição química, digestibilidade e a produção de gases in vitro de três cultivares de Brachiaria: Brachiaria brizantha cv. Marandu, Brachiaria brizantha cv. Xaraés e Brachiaria híbrida cv. Mulato, submetidas a níveis de oferta (4, 7, 10 e 13% do peso vivo animal), sob ciclos de pastejo rotativo. A cultivar Xaraés apresentou os maiores teores de FDN, FDNcp, FDA, LIG e menor produção de gases em 96 horas no componente lâmina foliar . Houve diminuição dos CNF e fração B2 e o aumento da fração C dos carboidratos nos componentes colmo e lâmina foliar entre ciclos de pastejo em todos as cultivares. A cultivar Marandu apresentou maiores valores de digestibilidade (DIVMO e DIVMS) nos dois componentes. A cultivar Marandu apresentou a melhor composição química e maiores valores de digestibilidade<br>Abstract: Nutritional evaluation of ruminant feed has been of great importance to adapt the databases of the systems of feed formulation in tropical regions. Due to the large amount of various alternative options for forage food assessment were developed in recent decades, among them the in vitro techniques. The aim of this study was to evaluate the chemical composition, digestibility and gas production in vitro of three cultivars of Brachiaria, Brachiaria brizantha cv. Marandu, Brachiaria brizantha cv. Xaraés and Brachiaria hybrid cv. Mulato, subjected to offer levels (4, 7, 10 and 13% of live weight) under rotational grazing. Xaraés cultivar showed the highest contents of NDF, NDFcp, ADF, lignin, and lower gas production in 96 hours in the stem and leaf fractions. There was a decrease in NFC and B2 fractions and increased the fraction C of carbohydrates in the stem and leaf fractions between grazing cycles in all the cultivars. Marandu cultivar showed the highest digestibility (IVDMD and IVDOM) in both fractions. The leaf fraction had lower levels of fiber and higher protein content in relation to the stem fraction. Cultivar Marandu had the best chemistry and higher digestibility values<br>Orientador: Euclides Braga Malheiros<br>Coorientador: Telma Teresinha Berchielli<br>Banca: Ana Cláudia Ruggieri<br>Banca: Paulo Henrique Moura Dian<br>Mestre

Le procédé de production d’éthanol à partir de biomasse lignocellulosique nécessite l’hydrolyse de cette dernière en sucres simples. Cette hydrolyse est le plus souvent réalisée par voie biologique grâce à des enzymes appelées cellulases. La production de ces enzymes représente cependant un verrou économique majeur au développement du procédé à grande échelle. Les cellulases sont généralement produites industriellement par le champignon filamenteux aérobie Trichoderma reesei, doté d’une forte capacité de sécrétion d’enzymes. Les cultures sont réalisées en bioréacteurs aérés et agités mécaniquement. Elles nécessitent de contrôler la concentration des substrats, ce qui requiert la maitrise de conditions hydrodynamiques et physicochimiques. En effet, le milieu de culture de T. reesei devient une suspension de cellules de champignons associées en filaments, de structure complexe, dont la viscosité augmente avec la concentration microbienne selon un comportement rhéofluidifiant. La viscosité est fonction de la morphologie du microorganisme qui peut, elle-même, varier avec les conditions de cultures. Cet accroissement de viscosité est un critère clef de l’extrapolation du procédé, car il affecte le transfert d’oxygène. Afin de maintenir une concentration en oxygène dissous suffisante, l’agitation et l’aération sont en général augmentées, entraînant un accroissement du cisaillement. Cet accroissement impacte en retour la morphologie du champignon, ralentit sa croissance puis diminue la production de cellulases. Ainsi, les conditions hydrodynamiques et rhéologiques engendrées au sein du bioréacteur sont complexes et variables dans le temps. L’interrelation entre conditions opératoires, morphologie, croissance du champignon et viscosité du moût de fermentation impose l’intégration de tous ces phénomènes pour l’optimisation du procédé, notamment à grande échelle. L’objectif de la thèse est de mettre en place une approche, visant à étudier au laboratoire la croissance de T. reesei et sa production d’enzymes, en reproduisant les contraintes hydrodynamiques associées aux conditions de fonctionnement des fermenteurs industriels. Pour ce faire, deux méthodologies originales ont été développées : une méthode de mesure de la viscosité du milieu, optimisée pour les champignons filamenteux, représentative des conditions rencontrées à grande échelle et qui s’appuie sur l’utilisation d’un rhéomètre rotatif équipé d’un rotor hélicoïdal ; une méthode d’analyse d’images associant un microscope motorisé et des algorithmes d’analyse d’images innovants, qui permet de générer des données sur la morphologie du champignon et d’identifier un critère morphologique pertinent basé sur le nombre de « trous » au sein d’un filament. Parallèlement à ces méthodes, différentes contraintes de cisaillement ont été mises en oeuvre en fermentation, afin de reproduire, à l’échelle du laboratoire, les conditions rencontrées à l’échelle industrielle. Ces outils ont été utilisés conjointement et validés lors de cultures non conventionnelles mimant les conditions industrielles en termes de cisaillement. Ils ont permis d’identifier un critère représentatif du cisaillement (EDCF) et d’établir, à partir de ce critère, des corrélations capables de prédire la viscosité du moût de fermentation, le taux de croissance maximum du microorganisme ainsi que certains paramètres morphologiques de la souche. De façon originale, ces corrélations déterminées à l’échelle du laboratoire ont été validées par des mesures effectuées à l’échelle industrielle. Au final, l’approche développée permet d’identifier au plus tôt les contraintes d’extrapolation à ne pas dépasser, afin d’orienter les choix technologiques des fermenteurs industriels impliquant des champignons filamenteux<br>Ethanol production from lignocellulosic biomass requires its transformation into fermentable sugars before the alcoholic fermentation. This step called hydrolysis is catalyzed by cellulases and is often considered as the major technical and economic challenge for the process development. Cellulases are industrially produced by the filamentous fungus Trichoderma reesei, thanks to its high secretion capacity. This fungus is strictly aerobic and is thus cultivated in aerated and stirred bioreactors. The fermentation optimization requires control of physicochemical conditions. Actually the growth of fungi induces an increase of the broth viscosity with shear thinning behavior because of the formation of three-dimensional mycelial structures (from micrometer to millimeter). This viscosity increase has a negative impact on the oxygen transfer. In order to keep the dissolved oxygen concentration higher than a critical limit, it is necessary to increase the power input thereby increasing the shear stress, which may affect the morphology of the fungus as well as its growth and cellulose production. Actually, physico-chemical conditions generated inside the bioreactor are complex and vary with time. These interrelations, between process conditions, morphology, growth and viscosity, require the integration of all these parameters to optimize the full-scale process. The goal of the thesis work was to develop a scale-down approach at lab-scale to mimic hydrodynamic conditions of industrial bioreactors and to study their impact on T. reesei growth and cellulase production. For that purpose, two new tools were developed. The first one consists in a new rheological measurement set-up using a helical rotor dedicated to filamentous fungi preventing mycelium degradation during the measurements. The second one is an original image analyses method that uses specific algorithms. It was then possible to record various morphological data on fungi and to select the most relevant ones (like the number of holes). Meanwhile, a wide range of shear stress conditions were explored in the laboratory bioreactor to reproduce industrial conditions. The new tools we had developed, coupled to these unconventional cultures lead to identifying a shear stress relevant

Cette fermentation complexe met en jeu plusieurs substrats carbones et azotés transformés par des levures floculantes en éthanol, CO2, glycérol et de nombreuses autres molécules peu concentrées qui caractérisent le goût et l'arôme de la bière. Le manque de connaissance de base en microbiologie et la quasi inexistance de capteurs directs représentent un handicap pour la surveillance, l'optimisation et l'automation de la fermentation en vue d'obtenir une bière aux qualités gustatives constantes. L’objectif de ce travail est d'apporter une contribution pour améliorer les connaissances sur le comportement de la levure en fermenteur discontinu, développer des nouveaux capteurs directs, proposer un modèle mathématique précis qui, jumelé à une technique de filtrage numérique élaborée, permette d'estimer les variables d'état non mesurées. La première partie est consacrée à l'étude cinétique de croissance et de floculation des levures, d'assimilation des sucres et de production des principaux métabolites avec des conditions opératoires variables (température, pH, agitation, nature du sucre et âge de culture des levures). L’étude de nouveaux capteurs directs pour la mesure d'éthanol et de CO2 dissous est abordée. Les connaissances acquises servent alors à définir et à mettre au point un modèle physiologique et un capteur logiciel par la méthode de Kalman. Dans la dernière partie, les capteurs directs et indirects sont utilisés pour la surveillance de la fermentation

The self-cycling fermentation (SCF) technique was used to grow three species of the genus Streptomyces which have been described as producing antibiotics. Application of this technique led to reproducible growth patterns for filamentous organisms. Control of the system using the CO$ sb2$ concentration in the off-gas was achieved under conditions which prevented monitoring of the dissolved oxygen. It was also possible to use the SCF system with either defined or complex media.<br>Production of tetracycline by S. aureofaciens was obtained on an iron-deficient medium. The production rate was as high as 0.272 g tetracycline/g sucrose/h. This was comparable with that reported for tetracycline in the patent literature. Growth patterns of the SCF when tetracycline was present in the fermenter were erratic, demonstrating the negative effect of the antibiotic on the growth of the producer organism. Further growth in a secondary fermenter system resulted in additional production over a period of about fifteen hours. In these experiments, tetracycline exhibited characteristics of a siderophore.

Self-cycling fermentation is a new technique used to cultivate synchronous populations of cells. This work used the technique in the production of citric acid from a strain of Candida lipolytica. Application of this technique resulted in extremely stable, reproducible patterns of growth and induced a high level of cell synchrony. Synchrony was maintained even throughout long periods of extended nutrient starvation. Thus, self-cycling fermentations can be operated with cycle times significantly longer than the doubling time of the microorganism. Use of this technique led to a full order of magnitude increase in the specific biomass production rate compared to literature values. This biomass was found to be suitable for the production of citric acid in a second stage reactor. Although this preliminary work with the second stage reactor was not optimized, the specific citric acid production rate obtained was comparable to rates found in the literature.

Solid state fermentation (SSF) refers to the microbial fermentation, which takes place in the absence or near absence of free water, thus being close to the natural environment to which the selected microorganisms, especially fungi, are naturally adapted. The current status of SSF research globally was discussed in terms of articles publication. This was followed by discussion of the advantages of SSF and the reason for interest in SSF as a notable bioprocessing technology to be investigated and compared to submerged fermentation (SmF) for the production of various added-value products. SSF also proved to be a potential technology to treat solid waste produced from food and agricultural industry and to provide environmental benefits with solid waste treatment. A summary was made of the attempts at using modern SSF technology for future biorefineries for the production of chemicals. Many works were carried out in the Satake Centre for Grain Process Engineering (SCGPE), University of Manchester, to prove the strategy of using SSF for the production of hydrolysate rich in nutrients for sequel microbial fermentation with or without adding any commercial nutrients. The research findings presented in this thesis are based on a series of SSF experiments carried out on systems varying in complexity from simple petri dishes to our own design of bioreactor systems. They were conducted to assess a solution for biomass estimation, enzymes production, and successful mass and heat transfer. A proper technique for inoculum transfer prior to the start of the fermentation process was developed. In SSF, estimation of biomass presents difficulties as generally the fungal mycelium penetrates deep and remains attached with the solid substrate particles. Although many promising methods are available, the evaluation of microbial growth in SSF may sometimes become laborious, impractical and inaccurate. Essentially, this remains another critical issue for monitoring growth. In these studies, measurement of colour changes during SSF are presented as one of the potential techniques that can be used to describe growth, complementary to monitoring metabolic activity measurement, such as CER, OUR and heat evolution, which is directly related to growth. For the growth of Aspergillus awamori and Aspergillus oryzae on wheat bran, soybean hulls and rapeseed meal, it was confirmed that colour production was directly proportional to fungal growth. This colourimetric technique was also proved to be a feasible approach for fungal biomass estimation in SmF. This new approach is an important complementation to the existing techniques especially for basic studies. The key finding is that the colourimetric technique demonstrated and provided information of higher quality than that obtained by visual observation or spores counting. The effect of aeration arrangements on moisture content, oxygen (O2), mass and heat transfer during SSF was investigated. A. awamori and A. oryzae were cultivated on wheat bran in newly designed four tray solid state bioreactor (SSB) systems. The new tray SSB systems were: (1) single circular tray SSB, (2) multi-stacked circular tray SSB, (3) Single rectangular tray SSB and (4) multi-square tray SSB. The purpose was to study the effect, on heat and water transfer, of operating variables, fermentation on the perforated base tray and internal moist air circulation under natural and forced aeration. Temperature, O2 and carbon dioxide were measured continuously on-line. Enzyme activity, moisture content and biomass were also measured. The results suggest that the air arrangements examined have a remarkable effect on the quantity of biomass produced using measurement of spores and enzymes production. The strategy presented in these studies allowed quantitative evaluation of the effect of forced internal moist air circulation on the removal of metabolic heat. With the proposed strategy, it was possible to maintain the bed temperatures at the optimum level for growth. However, the effect on moisture content was very different for the two fungi. It was found that the ability of A. oryzae to retain moisture was much higher than that of A. awamori. This is possibly due to the higher levels of chitin in A. oryzae. Greater spores and enzymes (glucoamylase, xylanase and cellulase) production was observed for A. awamori in multi-stacked circular tray and multi-square tray SSB systems compared to the conventional petri dishes and the other two systems. A. oryzae was excellent in producing protease in the same bioreactors. A direct technique of establishing a correlation between fungal growth and CER, OUR, heat evolved was proven successful in this work. The information obtained from CER and OUR led to the estimation of respiratory quotient (RQ). RQ describes the state of the fungal population in the tray SSB and gives an indication of fungal metabolic behaviour. RQ values < 1 were obtained from 38 experiments using four tray SSB systems for the two fungi. A kinetic model based on CO2 evolution instead of biomass concentration was examined in order to simplify the required experiments for kinetic model development. A Gompertz model was used to fit the integrated CO2 data and predict the quantity of CO2 evolution in all experiments. A correlation was found between the heat evolution and CER. The performances of tray SSB systems can be improved by constructing them as multi-trays. The multi-tray systems improved the mass transfer considerably compared with single tray systems. In addition, the multi-tray systems allowed precise measurement of the gradients of CO2, enzymes, spores and fungal biomass. In addition, the air arrangements using moistened air were successful in maintaining moisture content, adequate O2 supply and control of temperature, and hence, increased the productivity of both fungi. Overall A. awamori and A. oryzae have their own ability and performance to degrade and utilise the complex compositions contained in the solid substrate and fermentation conditions may lead to possible comparisons. In addition, multi-stacked circular tray and multi-square tray SSB systems demonstrated an excellent system for further investigations of mass transfer and possibly for large scale operation, though considerable optimisation work remains to be done, for example the height/diameter ratio and total number of trays should be optimised.

La premiere partie concerne la mise au point de biocapteurs amperometriques specifiques de l'ethanol, du l-malate, du d-lactate, du pyruvate et de la l-alanine et leur adaptation a des milieux biologiques. Ils sont tous bases sur le meme principe et sont compatibles avec un appareillage commercialise. Ces outils analytiques sont fiables, peu onereux et permettent des dosages precis (erreur de 2%) et rapides (2 minutes). Les capteurs specifiques de l'ethanol, du l-malate et du d-lactate ont ete implantes pendant une campagne de vinification dans le chai d'un viticulteur et a la chambre d'agriculture du tarn et garonne et ont permis la realisation de 80 dosages journaliers de chaque substrat. Les deux derniers biocapteurs proposes pour l'analyse clinique ne necessitent pas de traitement du sang dose, contrairement aux dosages en cours actuellement. La deuxieme partie de l'etude porte sur l'analyse globale d'un milieu de fermentation au moyen de la mesure du potentiel a courant nul pris par une electrode de platine par rapport a une electrode de reference. La valeur du potentiel est bien definie dans chaque phase d'une fermentation et renseigne sur la presence d'oxygene dissous dans le milieu fermentaire ou sur la mise en uvre d'une activite cellulaire, comme par exemple le debut et de la fin de la fermentation malolactique dans le cas de l'elaboration d'un vin. La proposition d'un fermenteur de type filtre-presse fait l'objet de la troisieme partie, dans le but de moduler l'activite du microorganisme en regulant le potentiel du milieu par application d'un champ electrique. Mais la modification du potentiel s'est averee impossible, meme en presence d'un mediateur d'oxydoreduction

Ce travail s'attache a presenter les principales caracteristiques de la physique qualitative. Nous proposons ensuite les differents aspects methodologiques developpes a savoir: une modelisation par graphe d'influences causales entre les diverses variables du procede; un raisonnement sur l'etat standard dont la finalite consiste a retablir la chaine des causalites des phenomenes observes a partir de l'observation des tendances d'evolution des variables accessibles. Ces concepts, generalisables a un grand nombre de problemes poses par les systemes physiques, sont ensuite adaptes aux procedes biotechnologiques. La finalite est ici de reconstituer l'etat qualitatif de certaines variables internes au procede (niveau microscopique) grace a une analyse et une interpretation des observations externes (variables macroscopiques). Dix operateurs qualitatifs ont ete definis pour traduire les interactions majeures entre les variables du procede. Ces operateurs sont par la suite exprimes selon le formaliste de kheops, un systeme de programmation et d'interpretation de bases de connaissances en logique propositionnelle etendue developpe au sein de notre laboratoire. Durant cette etude, les travaux ont ete testes sur des procedes fermentaires de nature differentes et ont permis d'elaborer deux systemes experts fonctionnant en ligne sur site reel: au stade du laboratoire (650 regles), les essais se sont deroules au centre de transfert en biotechnologie et microbiologie de l'ups/insat sur la croissance de bacteries. Un test a egalement ete effectue sur des levures afin de prouver une certaine robustesse du systeme expert en fonction des microorganismes utilises; a l'echelle industrielle (550 regles), le systeme expert assure une aide a la conduite d'un procede fermentaire de la societe sanofi-chimie (usine d'aramon) utilisant des champignons filamenteux

Les bases necessaires a la construction de modeles stoechio-cinetiques sont les connaissances experimentales acquises dans le laboratoire ou sur les pilotes industriels, et les connaissances theoriques obtenues par une etude bibliographique. La demarche classique consiste a construire les lois cinetiques a partir des connaissances theoriques, et a identifier les parametres qui s'y rattachent sur les differents jeux de donnees experimentales disponibles. Lorsque l'adequation modele/experience n'est pas correcte, il est difficile de repondre aux questions suivantes : les donnees experimentales sont-elles correctes ? les parametres cinetiques peuvent-ils etre estimes ? a cette approche globale, qui ne permet pas de retour direct sur les questions posees, nous preferons une approche sequentielle et progressive qui permet de juger tres tot de la qualite des mesures et ensuite de mettre en place les lois simplement a partir d'une analyse du resultat experimental. Ces deux premieres etapes reposent sur des contraintes imposees par les bilans de matiere. La premiere etape correspond au jugement de la qualite des donnees experimentales, les mesures erronees sont detectees, puis eliminees par filtrage sur des bilans atomiques. La deuxieme phase, une analyse factorielle des donnees, permet d'assister la conception du modele cinetique. Le schema reactionnel propose ainsi que les differents mecanismes cinetiques, sont testes pour savoir s'ils sont en adequation avec la connaissance experimentale. L'estimation des parametres cinetiques intervient dans la troisieme phase. Un modele tres general de reacteur biologique integrant les aspects thermodynamiques et les resistances aux transferts, a ete developpe et valide. Cette approche a necessite d'abord de travailler les deux premieres etapes sur le plan theorique. Les methodes sont presentees et critiquees, avec dans chaque chapitre, une illustration sur deux exemples de reacteurs biologiques utilises en brasserie.

Abstract The general aim of this thesis was to study rumen fermentation processes using in situ and in vitro digestibility and in vitro gas production techniques. The thesis includes the results of five experiments. The first experiment investigated the validity to reduce the amount of feed sample incubated in filter bags with DaisyII (from 0.50 to 0.25 g feed/bag). This hypothesis was verified by considering the results of two separate trials (conducted on 4 and 7 feeds, respectively) in which the digestibility values provided by DaisyII were compared with those achieved using a conventional batch culture. Results evidenced that, by using the lower sample size (0.25 g feed/bag), digestibility values provided by DaisyII resulted more similar, less variable and better correlated with those achieved with the batch culture. The second experiment was conducted on 11 feeds and compared the digestibility values achieved: i) in situ and in vitro using two different kinds of bag (nylon and filter) with a different pore size (40 and 25 µm); ii) in situ, placing the bags in the rumen environment, and in vitro using rumen fluid collected from intact cows by a suction technique. Results of the experiment showed that the use of filter bags induced a systematic, but predictable, underestimation of digestibility values compared to the nylon bags. Rumen fluid collected from intact cows provided digestibility values directly proportional to those obtained in situ. This latter result is relevant, as the validity of using rumen fluid collected from intact cows can be of interest for many research centers, in order to overcome ethical and public concerns related to the use of surgically treated animals. The third experiment was conducted on the same feeds analysed in the second experiment and it was aimed to compare the energy value of feeds estimated using three different equations: i) one equation based on feed chemical composition and in situ digestible NDF at 48 h of incubation (MENRC); ii and iii) two equations based on feed chemical composition and in vitro gas production at 24 h of incubation (MEMenke and MEUCD). A further aim was to evaluate the validity of reducing the in situ incubation time from 48 to 24 h. Results of the trial evidenced that, when MENRC estimates were taken as reference, MEUCD equation provided feed energy estimates more accurate and precise compared to MEMenke. However, the accuracy and precision of MEMenke estimates increased when four feeds rich in protein (CP > 16% DM) were excluded from the analysis. As the in situ digestibility values achieved at 24 and 48 h were highly correlated and showed a comparable reproducibility, the reduction of in situ incubation time from 48 to 24 h seems can be proposed for reducing the cost of the tests. The fourth experiment was aimed to compare the effects of two different systems of gas release (venting at fixed times or at fixed pressure) on gas production (GP) kinetics provided by two feeds with a different degradability (meadow hay and corn meal), using an innovative GP system. Results of the experiment showed that the venting system exerted significant effects on GP kinetics. The effects were particularly accentuated for corn meal and for venting at fixed times. Therefore, when venting at fixed times is applied, the headspace volume of GP system, the venting frequency and the amount of fermentable matter incubated must be carefully balanced to avoid the generation of high pressures, which could alter GP kinetics. However, venting at fixed pressure is preferable to that at fixed times, as it can strongly reduce these shortcomings. The fifth experiment aimed to evaluate the effects of increasing dosages of two antioxidants (BHT and a red chicory extract) on some in vitro rumen parameters, when they are incubated with meadow hay and corn meal. Results of the experiment showed that the two antioxidants exerted only small effects on rumen degradability, GP kinetics and volatile fatty acids profile. At the highest dosage, the red chicory extract showed to improve the efficiency of nitrogen utilization in the rumen, by increasing the amount of nitrogen captured by rumen microbes and utilized for microbial protein synthesis<br>Riassunto Lo scopo generale del presente lavoro di tesi è stato quello di studiare le fermentazioni ruminali mediante la tecnica in situ, e le più moderne tecniche in vitro che valutano la digeribilità e la produzione di gas degli alimenti zootecnici. La tesi comprende i risultati di 5 diverse prove sperimentali. L’obiettivo della prima prova è stato quello di valutare la possibilità di ridurre la quantità di campione alimentare incubato all’interno dei filter bags previsti dal sistema DaisyII (da 0,50 a 0,25 g di alimento/filter bag). La suddetta ipotesi è stata verificata analizzando i risultati ottenuti in due differenti prove sperimentali, condotte rispettivamente su 4 e 7 alimenti, il cui comune obiettivo era il confronto dei valori di digeribilità forniti dal DaisyII e da un sistema in vitro tradizionale. I risultati ottenuti hanno dimostrato che, con l’incubazione di 0,25 g di alimento/filter bag, i valori di digeribilità forniti dal DaisyII risultavano complessivamente meno variabili, meglio correlati e comparabili con quelli forniti dal sistema in vitro tradizionale. La seconda prova sperimentale è stata condotta su 11 alimenti e si poneva l’obiettivo di confrontare i valori di digeribilità ottenuti: i) in situ, attraverso l’impiego di due diversi tipi di sacchetto (nylon e filter bags) caratterizzati da una differente porosità (40 e 25 µm); in situ, incubando i campioni alimentari direttamente nel rumine dell’animale, ed in vitro, incubando i campioni alimentari con liquido ruminale prelevato mediante sonda esofagea da bovine non fistolate. I risultati della prova hanno evidenziato che l’utilizzo dei filter bags, in sostituzione dei nylon bags, induceva una sottostima sistematica, ma comunque stimabile, dei valori di digeribilità degli alimenti. I campioni alimentari incubati in vitro con liquido ruminale prelevato da bovine non fistolate hanno fornito valori di digeribilità direttamente proporzionali a quelli incubati in situ. Un tale risultato è sicuramente soddisfacente: la possibilità di utilizzare liquido ruminale prelevato da bovine non fistolate può rappresentare infatti un’importante opportunità per molti centri di ricerca, consentendo di superare le problematiche di natura etica legate all’impiego di bovine sottoposte ad operazione chirurgica per l’applicazione di fistole ruminali permanenti. La terza prova sperimentale è stata condotta sugli stessi alimenti impiegati nel secondo contributo e si poneva l’obiettivo di determinare il valore energetico degli alimenti applicando tre differenti equazioni: i) un’equazione basata sulla composizione chimica dell’alimento e sul valore di digeribilità della fibra misurato in situ dopo 48 h di incubazione (MENRC); ii e iii) due equazioni basate sulla composizione chimica dell’alimento e sulla produzione di gas misurata in vitro dopo 24 h di incubazione (MEMenke e MEUCD). Un secondo scopo era quello di valutare la possibilità di ridurre il tempo di incubazione in situ da 48 a 24 h. I risultati della prova hanno dimostrato che, prendendo come riferimento le stime di valore energetico ottenute applicando l’equazione MENRC, l’equazione MEUCD forniva stime più accurate e precise rispetto all’equazione MEMenke. Tuttavia, il livello di accuratezza e di precisione dell’equazione MEMenke è migliorato notevolmente escludendo i quattro alimenti proteici (PG > 16% SS) dall’analisi. Poiché i valori di digeribilità ottenuti in situ a 24 e a 48 h sono risultati fortemente correlati e hanno evidenziato una riproducibilità comparabile, la riduzione del tempo di incubazione in situ da 48 a 24 h può rappresentare una valida soluzione al fine di contenere i costi di analisi. La quarta prova sperimentale ha valutato gli effetti di due diverse modalità di rilascio del gas (a tempi fissi oppure a pressione fissa) sulle cinetiche di produzione di gas fornite da due alimenti caratterizzati da un differente livello di degradabilità ruminale (un fieno di graminacee e una farina di mais). Ai fini della prova è stato utilizzato un sistema innovativo per la misurazione del gas prodotto dalla fermentazione di matrici alimentari. I risultati della prova hanno evidenziato che la modalità di rilascio del gas ha condizionato significativamente le produzioni di gas fornite dai due alimenti. Tali effetti sono risultati particolarmente rilevanti nel caso della farina di mais e nel caso del rilascio a tempi fissi. Alla luce di questi risultati, quando si adotta la modalità di rilascio a tempi fissi, lo spazio di testa del sistema, la frequenza di rilascio del gas e la quantità di campione alimentare incubato dovrebbero essere opportunamente bilanciati per evitare lo sviluppo di pressioni molto elevate all’interno del sistema, le quali potrebbero alterare le normali cinetiche di produzione del gas. Comunque, al fine di ridurre tali rischi, l’impiego di sistemi che rilasciano il gas a pressione fissa è da ritenersi preferibile. La quinta prova sperimentale ha valutato gli effetti di due sostanze antiossidanti (il BHT ed un estratto di radicchio) dosate a due livelli differenti ed incubate con farina di mais e fieno di graminacee, sui principali parametri ruminali. I risultati della prova hanno evidenziato che i due antiossidanti hanno esercitato effetti poco rilevanti sulla degradabilità ruminale, sulle produzioni di gas e sui profili acidici forniti dai due alimenti. Al dosaggio più elevato, l’estratto di radicchio ha tuttavia migliorato il bilancio azotato ruminale, favorendo un aumento significativo della quota di azoto impiegato dai microorganismi ruminali per la sintesi di proteina batterica

Schwanniomyces occidentalis presente des caracteristiques qui pourraient faire de cette levure un nouveau systeme d'expression efficace pour la production a grande echelle de proteines heterologues. En effet contrairement a saccharomyces cerevisiae cette levure secrete, des proteines de taille importante dans le milieu extra-cellulaire, n'hyperglycosyle pas les proteines secretees, ne produit pas de quantite mesurable de proteases et se developpe sur des milieux peu onereux. Enfin, schwanniomyces occidentalis possede un metabolisme non fermentaire en aerobiose et presente une croissance rapide. Elle atteint une densite cellulaire maximale importante permettant une production de proteines importante. A l'heure actuelle, le developpement de schwanniomyces occidentalis, en tant que systeme hote, necessite au prealable la construction d'un vecteur d'expression efficace (marqueurs de selection, origines de replication, promoteurs et terminateurs de transcription). Nous disposions d'un vecteur comportant un marqueur d'auxotrophie (gene ade2) ainsi qu'une origine de replication autonome. Nous avons dans un premier temps, mieux caracterise ce vecteur et mis au point une methode de transformation efficace dans s. Occidentalis. Afin d'evaluer la possibilite d'utiliser s. Occidentalis comme systeme hote, nous avons isole la region 3' non codante ainsi que differents fragments (0,6 kb, 1,2 kb et 3,4 kb) de la region 5' non codante, du gene codant pour l'invertase. Une fois les conditions de derepression de ce gene etablies, nous avons utilise le gene gus, facilement dosable, comme systeme rapporteur. Nos etudes indiquent que ce gene n'est pas approprie dans s. Occidentalis. Neanmoins, nous demontrons que le fragment de 1,2 kb est necessaire et suffisant pour l'expression regulee du gene rapporteur dans s. Cerevisiae, dans les conditions de derepression precedemment definies

Na tentativa de reproduzir os benefícios ruminais dos ionóforos, pesquisadores exploram as propriedades antimicrobianas dos compostos secundários dos vegetais. Técnicas in vitro de produção de gás são amplamente utilizadas nas etapas iniciais de pesquisa. Óleos essenciais são compostos hidrofóbicos, sendo comum sua diluição com etanol em experimentos in vitro. Etanol é metabolizado no ambiente ruminal, havendo principalmente produção de acetato. No primeiro experimento, o objetivo foi avaliar o efeito de 10, 100 e 1000 µL de etanol em 75 mL de fluido ruminal tamponado, correspondendo a 0,13; 1,3 e 13,3 µL/mL, respectivamente. As inclusões de 100 e 1000 µL de etanol alteraram a fermentação ruminal in vitro. A dose de 10 µL não afetou a fermentação, exceto pela tendência (P < 0,10) de aumento na produção de gás ou na concentração de valerato ao se incubar feno ou dieta de alto concentrado, respectivamente. Sugere-se a utilização da menor dose possível de etanol (0,13 µL/mL). O segundo experimento trata do uso de brancos (frascos sem substrato, contendo somente inóculo ruminal e meio de incubação) em experimentos in vitro para se estimar as produções líquidas de gás e metano (CH4), assim como a degradação líquida da matéria orgânica incubada. Foi demonstrado que aditivos ruminais (monensina, carvarol, eugenol) afetaram a fermentação de frascos contendo substrato assim como dos brancos. Dessa forma, brancos específicos (brancos contendo aditivo) são necessários ao se avaliar aditivos ruminais sob condições in vitro. No terceiro experimento, incubações in vitro foram conduzidas para triar os efeitos de óleos essenciais sobre a fermentação ruminal. Foram selecionados os óleos essenciais de erva-baleeira (Cordia verbenacea), aroeira-vermelha (Schinus terebinthifolius; óleo extraído das folhas ou frutos), macela (Achyrocline satureoides), guaco (Mikania glomerata), carqueja (Baccharis cylindrica), arnica (Lychnophora pinaster), capim cidreira (Cymbopogon citratus), capim limão (Cymbopogon flexuosus) e citronela (Cymbopogon winterianum). Foram também incluídos os óleos resinóides de copaíba mari-mari (Copaifera reticulata), copaíba angelim (Copaifera multijuga), copaíba zoró (Copaifera langsdorfii) e copaíba vermelha (Copaifera langsdorfii). Os óleos de ervabaleeira, macela e as quatro óleoresinas de copaíba pouco alteraram a fermentação ruminal. Os outros óleos apresentaram claro efeito antimicrobiano, evidenciado pela queda na degradação de substrato. Os resultados mais promissores foram observados ao se incubar dieta de alto concentrado com inóculo adaptado a esta dieta. Os óleos essenciais que apresentaram os melhores resultados foram aroeira vermelha (folhas e frutos) e arnica. Sob a condição de alto concentrado, esses óleos aumentaram a concentração de propionato, reduziram a relação acetato:propionato e/ou diminuíram a produção de CH4. Os óleos essenciais de aroeira vermelha (extraídos das folhas e dos frutos) e arnica foram selecionados para subsequente avaliação in vivo.<br>In an attempt to reproduce the benefits of ionophores on rumen fermentation, researchers have been exploiting the antimicrobial properties of plant secondary metabolites. In vitro gas production techniques are widely used during the screening phase. Essential oils are hydrophobic compounds, being usual its dilution in ethanol for in vitro experiments. In the rumen environment, ethanol is metabolized mostly to acetate. In the first experiment, the objective was to evaluate the effects of 10, 100 and 1000 µL of ethanol in 75 mL of buffered rumen fluid, corresponding to 0.13, 1.3, and 13.3 µL/mL, respectively. Ethanol inclusions of 100 and 1000 µL affected in vitro rumen fermentation. The inclusion of 10 µL of ethanol had no effects on fermentation, except for the tendency (P < 0.10) of gas production increase or valerate increase when incubating hay or a high-concentrate diet, respectively. It is suggested that ethanol should be included at the lowest dose as possible, which corresponded to 10 µL (0.13 µL/mL) in our conditions. The second experiment dealt with the use of blanks (flasks without substrate, containing only inoculum and incubation medium) in in vitro experiments to estimate net production of gas and methane (CH4), as well as net degradation of organic matter incubated. It was demonstrated that rumen additives (monensin, carvacrol, eugenol) affected fermentation of flasks containing substrate and blanks. Thus, specific blanks (blanks containing additive) are necessary when rumen additives are evaluated in vitro. In the third experiment, in vitro incubations were conducted to screen the effects of essential oils on rumen fermentation. The selected essential oils were: cordia (Cordia verbenacea), Brazilian peppertree (Schinus terebinthifolius; extracted from leaves or fruits), macela (Achyrocline satureoides), guaco (Mikania glomerata), carqueja (Baccharis cylindrica), arnica (Lychnophora pinaster), West Indian lemongrass (Cymbopogon citratus), East Indian lemongrass (Cymbopogon flexuosus), and citronella (Cymbopogon winterianum). Oleoresins from copaiba mari-mari (Copaifera reticulata), copaiba angelim (Copaifera multijuga), copaiba zoro (Copaifera langsdorfii), and copaiba vermelha (Copaifera langsdorfii) were also included. The essential oils from cordia, macela and all copaiba oleoresins had little effect on in vitro rumen fermentation. The remaining essential oils showed a clear antimicrobial effect, mainly on truly degraded organic matter. The most promising results were obtained when using high-concentrate diet and inoculum from animals fed this same diet. The best results were observed for Brazilian peppertree (leaves and fruits) and arnica essential oils. In the condition of high-concentrate diet, these oils increased propionate concentration, had lower acetate to propionate ratio, and/or reduced CH4 production. The essential oils from Brazilian peppertree (leaves and fruits) and arnica were selected for further in vivo evaluation.

L'objectiu d'aquesta tesi va ser l'adaptació i validació de diferents tècniques independents de cultiu per a la detecció i quantificació de la microbiota present a la fermentació vínica. Es van assajar la QPCR (PCR quantitativa) per a la quantificació i seguiment d'espècies clau i la diversitat ecològica es va analitzar per DGGE (electroforesi en gel desnaturalitzant) i la clonació d'un fragment ribosomal.<br/>Tanmateix es va estudiar l'aplicació de la Hibridació in Situ (FISH) i la QPCR amb colorants específics per tal de diferenciar cèl·lules vives i mortes. <br/>Aquestes tècniques es varen aplicar a fermentacions industrials, essent notable la detecció de bacteris acètics i llevats No-Saccharomyces en concentracions superiors a les detectades com a poblacions cultivables. Es van estudiar interaccions entre llevats Saccharomyces i No-Saccharomyces al laboratori, observant-ne la supervivència d'aquestes en estats viables però no cultivables. <br/>Aquestes tècniques independents de cultiu indiquen una població i dinàmiques microbianes prèviament desapercebudes.<br>The objective of this thesis was the adaptation and validation of different culture-independent techniques for detection and quantification of the microbiota present in wine fermentation. We tested the QPCR (PCR quantitative) for the quantification and monitoring of key species. The ecological diversity was analyzed by DGGE (denaturing gradient gel electrophoresis) and by cloning of a ribosomal fragment. <br/>However, we studied the application of in Situ Hybridization (FISH) and QPCR with specific dyes to differentiate between live and dead cells. <br/>These techniques were applied in industrial fermentations, being significant the detection of acetic acid bacteria and Non-Saccharomyces yeast in concentrations higher than those identified as culturable populations. We studied the interactions between Saccharomyces and Non-Saccharomyces yeast in the laboratory, observing their survival in these viable but non-culturable state. <br/>These culture-independent techniques indicate a population and microbial dynamics previously unnoticed.

Un fermenteur aere, agite par un ruban helicoidal a ete concu afin de repondre aux contraintes specifiques des fermentations mettant en jeu des milieux de culture presentant des viscosites apparentes tres elevees, superieures a 1 pa. S. La production par voie fermentaire de scleroglucane par sclerotium rolfsii a ete retenue comme support de cette etude. Une etude du comportement rheologique des solutions de scleroglucane a permis en utilisant le modele de herschel-bulkley de correler les parametres rheologiques des solutions a leur concentration en scleroglucane dans une gamme de concentration allant de 5 a 30 g/l. L'etude des capacites de melange de notre installation par une technique conductimetrique en regime laminaire et en regime de transition a confirme les excellentes capacites de melange des agitateurs a rubans helicoidaux en milieu fortement visqueux. L'etude du transfert d'oxygene dans notre installation en utilisant une sonde a oxygene a montre des capacites de transfert en milieu faiblement visqueux nettement inferieures a celles observees dans des systemes agites par des turbines. En milieu fortement visqueux, on observe un transfert d'oxygene, meme faible, dans tout le volume de la cuve. L'etude de la puissance consommee par notre installation a montre que celle-ci etait 2 a 3 fois inferieure a celle demandee par une installation agitee par turbine. L'etude de fermentations dans notre appareillage mettant en uvre sclerotium rolfsii a montre qu'il etait possible d'obtenir des concentrations en scleroglucane plus elevees que dans un fermenteur classique

Due to the warming of the climate system, public opinion has increased its interest in greenhouse gas, with particular predilection for methane (CH4 ) and their relationship with livestock sector. For this reason, several researches have proposed different nutritional strategies to reduce methane production in ruminants. The two main aims of this thesis were: i) to study methodological factors affecting in vitro CH4 production and to develop a new approach to measure in vitro CH4 production; ii) to evaluate effectiveness of some nutritional strategies on in vitro CH4 reduction. The objectives have been achieved through the research activities and reported in five experimental contributions. The first contribution aimed to analyse specific factors affecting in vitro gas production (GP), CH4 production and their ratio by a meta-analysis approach using 274 control treatments from 39 scientific papers published over 12 years. The common factors included in the analysis were: the pressure in the GP equipment (constant or increasing), the incubation time (24 or ≥48 h), the timing of rumen fluid collection (before or after feeding of donor animals), the presence of N in the buffer solution (presence or absence), and the ratio between amount of buffered rumen fluid and feed sample (<130 or 130-140 or >140 ml/g DM). The NDF content of feed sample incubated (392 ± 175.3 g/kg) was considered as a continuous variable. Results showed that GP measures were influenced mainly by the incubation time; values of CH4 production were affected mainly by the ratio between amount of buffered rumen fluid and feed sample; values of CH4 proportion (CH4 /GP) were mainly affected by timing of rumen fluid collection. Results suggest that factors considered in the meta-analysis explained a significant part (about 65%) of variability of GP and CH4 measures obtained in vitro. Thus, a higher standardization of specific methodological protocols would be desirable in order to compare results of different trials. The second contribution compared two in vitro techniques (closed or vented bottles) of sampling and measuring CH4 production from ruminant feeds. Five feeds (meadow hay, ryegrass hay, corn grain, dry sugar beet pulp, and expeller flaxseed) were tested. The experimental design was: 2 incubations×5 feeds×3 replications×2 gas sampling techniques, plus 4 blanks (bottles without feed sample), for a total of 64 bottles incubated. Half of the bottles were not vented during the incubation, whereas the remaining were vented at a fixed pressure and gas was collected into a tight bag connected to each bottle. At the end of each incubation, gas samples were sampled from the headspace of closed bottles or from headspace and bags of vented bottles and analyzed for CH4 concentration. GP values were adjusted or not for the amount of CO2 solubilized in the fermentation fluid. Values of CH4 concentration (ml CH4 /100 ml gas) and production (ml CH4 /g DM) were computed using corrected or uncorrected GP values. Closed bottles showed lower uncorrected GP (−18%) compared with vented bottles, whereas the correction for dissolved gas reduced but did not remove differences between techniques. Closed bottles showed unadjusted CH4 concentrations 23% greater than that of vented bottles, but the adjustment of measurements for solubilized CO2 reduced but did not remove this difference. Adjusted CH4 production was not influenced by technique. Closed bottles provide good measurements of CH4 production but not of GP. Venting of bottles at low pressure permits a reliable evaluation of total GP and CH4 production. The third contribution evaluated the influence that changes in main dietary constituents (starch:ADF ratio, protein and lipid contents) could have on in vitro GP, CH4 concentration and production. All changes have been made within ranges of diets used in intensive farms of the North Italy. A reference diet was used (273, 361, 158, and 33 g/kg DM of starch, aNDF, CP, and lipids, respectively). Other 6 diets had a smaller or a greater starch:ADF ratio (0.40 or 1.77, respectively), CP content (115 or 194 g/kg DM, respectively), or lipid content (26 or 61 g/kg DM, respectively), compared to the reference diet. The experimental design was: 4 incubations × 7 diets × 5 replications, plus 20 blanks (5 blanks/run), for a total of 160 bottles. Bottles were vented at fixed pressure and at the end of each incubation, gas samples were collected from the bottle headspace and analyzed for CH4 . The proportion of CH4 lost with venting was estimated. An increasing starch:ADF ratio reduced GP per g DM and per g TDMd, increased the amount of CH4 produced per g DM, whereas CH4 produced per g TDMd was unchanged. The increase of CP content decreased GP, whereas CH4 production was reduced in low protein diet only when expressed as proportion of GP. The increase of lipid content reduced GP per g DM, but no influences were observed on CH4 values. Results of this study showed that the magnitude of effects exerted by the dietary changes on in vitro fermentation was small. The fourth contribution aimed to explore effects of 4 pure plant extracts (allyl-sulfyde; cinnamaldehyde; eugenol; limonene) and 1 synthetic compound (monensin), all with antimicrobial properties, added to a commercial diet for dairy cows, on GP, CH4 concentration and production. Two dosages were tested: 3 or 30 mg/g of diet for pure plant extracts, 0.015 or 0.030 mg/g of diet for monensin. Incubation procedures were the same as the previous experiment. The experimental design was: 4 incubations×5 additives×2 dosages×3 replications, plus 12 bottles as control (without additive; 3/run) and 12 bottles as blanks (3/run), for a total of 144 bottles. Results showed that low dosages of all compounds did not exert any effect on in vitro GP and CH4 production. Compared to the control, high dosage of allyl-sulfyde, cinnamaldehyde, eugenol, limonene, and monensin significantly reduced in vitro GP (ml/g DM; -16%, -12%, -9%, -38%, -12%, respectively). In vitro CH 4 production was significantly reduced only by high dosage of allyl-sulfyde, cinnamaldehyde, limonene, and monensin (-32%, -12%, -43%, -18%, respectively, compared to the control). Only high dosage of allyl-sulfyde and limonene significantly reduce CH4 proportion (-18%, -12% respectively, compared to the control). The most promising results were observed for cinnamaldehyde, as the depression of CH4 production was not accompanied by a reduction of in vitro degradability. The fifth contribution investigated the possibility to reduce in vitro gas and CH4 production by combining a direct inhibitor of methanogenisis (chloroform) with two indirect inhibitors: i) nitrate, acting as a H2 sink, and ii) saponins, acting as antiprotozoal agent. The experiment was conducted using a semi-continuous flow system (RUSITEC ®). A diet, based on grass hay, was incubated alone (diet 1) or supplemented with 31.5 g/kg of nitrate (diet 2) or with 50 g/kg of saponins extracted from ivy extract (diet 3). Three diets were incubated without or with chloroform (2 µL/L), obtaining a total of 6 dietary treatments. Each of the 6 dietary treatments was incubated in four vessels, for a total of 24 vessels. Each vessel was inoculated with 800 ml of buffered rumen fluid. Whole incubation lasted 21 d. Total GP (on average 2.56 L/d) was not affected by diet or chloroform addition. When incubated alone, chloroform, nitrate and saponins reduced daily CH4 production by 96, 66 and 22%, respectively. However, when chloroform was combined with indirect inhibitors, neither nitrate or saponins had any additional effect in chloroform treated vessels on in vitro CH4 production. Results suggest that no synergistic effects emerged between the direct inhibitor and the two indirect inhibitors.<br>Il recente scenario del riscaldamento climatico globale, ha portato ad un maggiore interesse dell’opinione pubblica riguardo ai gas che causano l’effetto serra, ponendo particolare attenzione al gas metano (CH4 ) e alla sua relazione con il settore zootecnico. Per questa ragione, numerosi gruppi di ricerca hanno proposto delle strategie nutrizionali per ridurre le emissioni di CH4 da parte dei ruminanti. Nella presente tesi sono stati sviluppati due obiettivi principali. Il primo riguardava lo studio dei fattori metodologici che influenzano la produzione in vitro del CH4 e lo sviluppo di un nuovo approccio per la misurazione dello stesso gas prodotto durante delle fermentazioni in vitro. In secondo luogo, si è inteso valutare l’efficacia di alcune strategie nutrizionali sulla riduzione della produzione di CH4 . Tali obiettivi sono stati raggiunti attraverso cinque diverse attività di ricerca. Nel primo lavoro è stato valutato l’effetto che specifici fattori metodologici possono avere sulla produzione in vitro di gas e di CH4 (espresso sia in termini di produzione che di proporzione). Questa analisi è stata effettuata attraverso un approccio di meta-analisi utilizzando 274 osservazioni, che rappresentavano i trattamenti di controllo di 39 articoli scientifici pubblicati negli ultimi 12 anni. I fattori considerati erano: la pressione che si forma nello spazio di testa dello strumento utilizzato per le fermentazioni in vitro (costante o incrementale), il tempo di incubazione (24 o ≥48 ore), il momento di raccolta del liquido ruminale (prima o dopo la somministrazione del pasto agli animali donatori), la presenza di azoto nella composizione della saliva artificiale (presenza o assenza), e il rapporto tra la miscela di liquido ruminale e saliva artificiale sul campione alimentare incubato (130 or 130-140 or >140 ml/g DM). Questi cinque fattori sono stati considerati come variabili discrete, invece il contenuto di NDF del campione alimentare incubato (392 ± 175.3 g/kg) è stato considerato come variabile continua. I risultati hanno mostrato che la produzione totale di gas è principalmente influenzata dal tempo di incubazione, mentre la produzione di CH4 è influenzata soprattutto dal rapporto tra la miscela di liquido ruminale e saliva artificiale sul campione alimentare incubato. Quando invece il CH4 viene espresso in termini di proporzione sul gas totale prodotto, i valori sono principalmente influenzati dal momento di raccolta del liquido ruminale. Inoltre, questa meta-analisi dimostra che i fattori considerati spiegavano una parte considerevole (circa il 65 %) della variabilità dei dati di gas e CH4 prodotti durante le fermentazioni in vitro. Quindi, sarebbe desiderabile una maggiore standardizzazione dei protocolli metodologici internazionali, in modo da facilitare il confronto di dati ottenuti in diversi sperimentazioni. Il secondo contributo sperimentale ha inteso mettere a punto una procedura di raccolta e di misurazione del CH4 prodotto, utilizzando due diverse tipologie di fermentazione in vitro: i) un “sistema aperto”, con sfiato regolare del gas all’interno di un sacchetto connesso alle bottiglie di fermentazione, dal quale viene prelevato il campione di gas per l’analisi del CH4 ; ii) un “sistema chiuso”, con accumulo progressivo dei gas di fermentazione nello spazio di testa delle bottiglie, dal quale viene prelevato il campione per l’analisi del CH4 . Per le fermentazioni sono stati usati cinque alimenti singoli utilizzati nell’alimentazione dei ruminanti (fieno polifita, loietto, farina di mais, panello di lino e polpe di bietola). Il disegno sperimentale prevedeva: 2 incubazioni × 5 alimenti × 3 replicazioni per alimento × 2 tecniche di campionamento del gas, più 4 bianchi (bottiglie incubate senza campione alimentare), per un totale di 64 bottiglie incubate. Metà delle bottiglie non venivano sfiatate, mentre le rimanenti venivano sfiatate a pressione fissa e il gas era raccolto in un sacchetto a tenuta connesso ad ogni bottiglia. Alla fine di ogni incubazione, il gas veniva campionato dallo spazio di testa delle bottiglie utilizzate per il sistema chiuso o dallo spazio di testa e dal sacchetto delle bottiglie utilizzate per il sistema aperto. Tutti i campioni sono stati poi analizzati per quantificare la concentrazione di CH4 . I valori di gas prodotto venivano o meno corretti per la quantità di CO2 che si era disciolta nel liquido di fermentazione. I valori di produzione (ml CH4 /g DM) e di proporzione (ml CH4 /100 ml gas) di CH4 sono stati calcolati utilizzando i valori corretti o non corretti di produzione di gas totale. Il sistema chiuso ha mostrato una produzione di gas totale non corretto inferiore (-18%) rispetto al sistema aperto, mentre la correzione del gas ha ridotto ma non rimosso le differenze tra le due tecniche. Le bottiglie chiuse hanno mostrato una proporzione di CH4 non corretto superiore (+23%) rispetto al sistema aperto, mentre la correzione ha ridotto ma non rimosso le differenze tra le due tecniche. La produzione di CH4 corretto non è stata influenzata dalla tecnica utilizzata. Concludendo, il sistema chiuso non offre buone misurazioni della produzione del gas, mentre il sistema aperto consente una valutazione attendibile sia del gas che del CH4 prodotto. Nel terzo contributo sperimentale è stato valutato l’effetto che le variazioni quantitative dei principali componenti chimici (rapporto amido:ADF, contenuto proteico e contenuto lipidico) di diete per vacche da latte, posso avere sulla produzione di gas e CH4 . Le variazioni sono state fatte tenendo conto degli intervalli di fibra, amido, proteina e lipidi effettivamente utilizzate negli allevamenti intensivi del Nord Italia. La dieta di riferimento utilizzata aveva la seguente composizione chimica: 273, 361, 158, e 33 g/kg SS di amido, NDF, CP, e lipidi, rispettivamente. Le altre 6 diete avevano un minore o maggiore rapporto di amido:ADF (0.40 or 1.77, rispettivamente), o di contenuto proteico (115 or 194 g/kg DM, rispettivamente), o di contenuto lipidico (26 or 61 g/kg DM, rispettivamente), rispetto alla dieta di riferimento. Il disegno sperimentale prevedeva: 4 incubazioni × 7 diete × 5 replicazioni per dieta, più 20 bianchi (bottiglie incubate senza campione alimentare), per un totale di 160 bottiglie incubate. Il gas prodotto veniva regolarmente sfiatato a pressione fissa e alla fine di ogni incubazione un campione di gas veniva raccolto dallo spazio di testa delle bottiglie e veniva analizzato per misurare la concentrazione di CH4 . La quantità di CH4 perso durante lo sfiato del gas di fermentazione è stata stimata. I risultati mostrano che all’aumentare del rapporto amido:ADF si riduce la produzione di gas (per g di SS e per g di degradabilità "vera" della SS), aumenta la produzione di CH4 (per g di SS), mentre la produzione di CH4 espressa come g di degradabilità "vera" della SS non varia. All’aumentare del contenuto proteico si riduce la produzione di gas, mentre la proporzione di CH4 è stata ridotta solo nella dieta ipoproteica. L’aumento del contenuto lipidico ha ridotto la produzione di gas (per g di SS), ma non ha influenzato i valori di CH4 . Si può quindi concludere che l'entità degli effetti, esercitati dalle variazioni quantitative dei principali componenti chimici delle diete sulla fermentazione in vitro, è stata quasi inesistente. Il quarto lavoro ha inteso valutare l’effetto di quattro estratti puri di piante (allil-sulfide, cinnamaldeide, eugenolo e limonene) e di un composto sintetico (monensin), tutti con proprietà antimicrobiche, utilizzati come additivi di una dieta per vacche da latte, sulla produzione di gas e CH4 . Sono stati utilizzati due diversi dosaggi: 3 or 30 mg/g di per gli estratti puri di piante, 0.015 or 0.030 mg/g di dieta per il monensin. Le procedure di incubazioni utilizzate erano le stesse dell’esperimento precedente. Il disegno sperimentale prevedeva: 4 incubazioni × 5 additivi × 2 dosaggi × 3 replicazioni, più 12 bianchi (bottiglie incubate senza campione alimentare, 3 per incubazione), e 12 controlli (dieta incubata senza additivi, 3 per incubazione) per un totale di 144 bottiglie incubate. I risultati hanno mostrato che tutti i composti testati con il basso dosaggio non hanno mai influenzato la produzione in vitro di gas e CH4 . Rispetto al controllo, gli alti dosaggi di allil- sulfide, cinnamaldeide, eugenolo, limonene, e monensin hanno ridotto significativamente la produzione in vitro di gas (ml/g DM; -16%, -12%, -9%, -38%, -12%, rispettivamente). La produzione in vitro di CH4 è stata significativamente ridotta solo dell’alto dosaggio di allil-sulfide, cinnamaldeide, limonene, e monensin (-32%, -12%, -43%, -18%, rispetto al controllo). Solo gli alti dosaggi di allil- sulfide e limonene hanno ridotto significativamente anche la proporzione di CH4 (-18% e -12% rispetto al controllo). I risultati più promettenti sono stati osservati per la cinnamaldeide, che ha depresso la produzione CH4 senza influenzare negativamente gli altri parametri fermentativi. Il quinto e ultimo contributo sperimentale ha valutato la possibilità di ridurre la produzione in vitro di gas e CH4 attraverso la combinazione di un diretto inibitore del CH4 (cloroformio) con due inibitori indiretti: i) i nitrati, che agiscono come accettori di idrogeno e ii) le saponine, dei noti agenti antiprotozoari. L’esperimento è stato condotto utilizzando un sistema di simulazione ruminale a flusso semi-continuo (RUSITEC®). Una dieta base è stata incubata singolarmente (dieta 1) o addizionata con 31.5 g/ kg di nitrati (dieta 2) o con 50 g/kg saponine ottenute dall’estratto d’edera (dieta 3). Queste tre diete sono state incubate senza o con l’aggiunta di cloroformio (2 µL/L), ottenendo un totale di 6 trattamenti testati. Ogni trattamento è stato incubato in quattro bottiglioni, per un totale di 24 bottiglioni. L’intera incubazione è durata 21 giorni. I risultati mostrano che la produzione totale di gas (in media 2.56 l/d) non è stata influenzata dalla dieta o dalla presenza del cloroformio. Quando utilizzati singolarmente, cloroformio, nitrati e saponine riducono la produzione giornaliera di CH4 del 96, 66 e 22%, rispettivamente. Nonostante ciò, quando il cloroformio era combinato con inibitori indiretti, non si è evidenziato nessun effetto addizionale sulla riduzione di CH4 prodotto. Concludendo, si può quindi affermare che nessun effetto sinergico è emerso tra l’inibitore diretto di CH4 e i due inibitori indiretti.

Foam separation is a simple and economic method for separation of surface-active molecules such as proteins and enzymes from aqueous solutions. In this study, lactic acid broth, spent brewer&rsquo<br>s yeast extract and residual beer was used to investigate the applicability and efficiency of foam separation technique in partial purification of fermentation products and recovery of valuable components from industrial waste streams. The effects of the process variables initial feed concentration, air flow rate, foaming time, liquid pool height and temperature on separation performance were studied and optimum conditions for removal of proteins from lactic acid broth was determined. Highest enrichment (172.2) and separation ratio (314) with a high protein recovery (45.2 %) were obtained by foaming 200 ml of lactic acid broth with an initial feed concentration of 0.018 mg/ml at an air flow rate of 38.5 cm3/min. Selectivity of foam separation in protein purification, and its effect on protein structure was investigated in brewery wastes using SDS-PAGE and native PAGE, respectively.

Industrial beets may compete against corn grain as an important source of sugars for non-food industrial fermentations. However, dependable and energy-efficient systems for beet sugar storage and processing are necessary to help establish industrial beets as a viable sugar feedstock. Therefore, technical and economic aspects of beet sugar storage and processing were evaluated. First, sugar retention was evaluated in whole beets treated externally with either one of two antimicrobials or a senescence inhibitor and stored for 36 wk at different temperature and atmosphere combinations. Although surface treatment did not improve sugar retention, full retention was enabled by beet dehydration caused by ambient air at 25 ?C and with a relative humidity of 37%. This insight led to the evaluation of sugar retention in ground-beet tissue ensiled for 8 wk at different combinations of acidic pH, moisture content (MC), and sugar:solids. Some combinations of pH ? 4.0 and MC ? 67.5% enabled retentions of at least 90%. Yeast fermentability was also evaluated in non-purified beet juice acidified to enable long-term storage and partially neutralized before fermentation. None of the salts synthesized through juice acidification and partial neutralization inhibited yeast fermentation at the levels evaluated in that work. Conversely, yeast fermentation rates significantly improved in the presence of ammonium salts, which appeared to compensate for nitrogen deficiencies. Capital and operating costs for production and storage of concentrated beet juice for an ethanol plant with a production capacity of 76 ? 106 L y-1 were estimated on a dry-sugar basis as U.S. ?34.0 kg-1 and ?2.2 kg-1, respectively. Storage and processing techniques evaluated thus far prove that industrial beets are a technically-feasible sugar feedstock for ethanol production.<br>North Dakota Renewable Energy Commission<br>North Dakota Agricultural Experiment Station<br>Green Vision Group<br>U.S. Department of Transportation (USDOT) through the SunGrant Initiative

La modelisation, l'estimation et la commande d'une fermentation alcoolique font l'objet de ce memoire. Un modele mathematique simple du procede de fermentation alcoolique est etabli pour rendre compte des resultats experimentaux obtenus sur l'unite pilote. Ce modele est utilise pour tester les algorithmes d'estimation et de commande developpes par la suite et pour determiner leurs parametres de synthese. Le probleme d'estimation non-lineaire de l'etat et des parametres pour les bioprocedes est etudie. Deux techniques d'estimations sont developpees et proposees pour pallier au manque de capteurs pour la mesure de certaines variables d'interet. Leurs performances sont etudiees par des simulations sur le modele etabli du procede. L'un des estimateurs est valide experimentalement sur une fermentation alcoolique fonctionnant en mode discontinu, semi-continu et continu. Une loi de commande adaptative est proposee pour reguler la concentration en substrat a l'interieur du bioreacteur en agissant sur le taux de dilution. C'est une strategie de commande non-lineaire couplee a l'estimateur valide auparavant. Les performances de cette loi de commande sont mises en evidence par une experimentation sur l'unite pilote de fermentation

L’objectif de ce travail est d’étudier la valorisation des produits connexes de l’industrie gabonaise du bois d’okoumé dans un contexte de bioraffinerie. La production de bioéthanol cellulosique et de lignine en se basant sur un procédé de prétraitement transposable à l’échelle industrielle a été étudiée. Nos travaux, réalisés à l’échelle du laboratoire, portent sur (1) l’optimisation du prétraitement par explosion à la vapeur (EV), (2) l’extraction de la lignine de bois explosé par percolation avec de l’éthanol comme solvant, (3) la caractérisation des biopolymères (cellulose et lignine) et l’étude de l’impact des paramètres du prétraitement sur leur structure chimique et (4) la conversion du bois prétraité en éthanol par hydrolyse enzymatique et fermentation. Dans un premier temps nous nous sommes intéressés à optimiser les différents paramètres du procédé EV (granulométrie, pré-imprégnation, temps de séjour et température) concernant les teneurs en glucose, xylose et en lignine du bois explosé. L’impact du procédé sur la production de produits de dégradation des sucres (furfural et le HMF ) potentiellement inhibiteurs sur les enzymes et microorganismes a également été étudié. Il a été constaté que la taille des particules est un paramètre important à considérer dans l’amélioration du procédé, les particules fines conduisant à une dégradation importante de la cellulose durant le prétraitement. L’influence de la concentration en acide, utilisé pour imprégner la biomasse, ainsi que des paramètres du procédé (temps de séjour et température) sur l’hydrolyse des biopolymères (lignine et hémicelluloses) ont ensuite été étudiés. Des conditions permettant d’optimiser des teneurs en cellulose (&gt; 59%) dans la matière prétraitée ont été retenues. Afin de mieux comprendre l’impact du prétraitement sur la lignine extraite à partir du bois explosé et anticiper une valorisation future, des analyses spectroscopiques (RMN HSQC et 31P) et chromatographiques (SEC) ont été réalisées. Une diminution de la teneur en β-O-4 est observée avec l’augmentation de la sévérité du traitement d’EV avec une dégradation quasi-totale des unités β-O-4 dans des conditions de plus grande sévérité (T = 210°C, t = 2,5 min et H2SO4 = 0,5%). Des phénomènes de recondensation par voie radicalaire de la lignine ont également été observés. La cellulose des différentes fractions solides récupérées après traitement EV a également été caractérisée par SEC, DRX pour étudier l’effet du traitement et identifier les paramètres clés dans l’amélioration de l’efficacition de la conversion de la cellulose en glucose. Une augmentation de la cristallinité avec l’augmentation de sévérité a été observée avec une chute importante du DP de la cellulose pour les EV réalisées avec imprégnation acide. La digestibilité enzymatique des résidus cellulosiques après EV par T. Reesei et l’extraction des lignines ont été étudiées. Les principaux paramètres affectant positivement la réactivité aux enzymes sont l’élimination d’une partie des polymères non cellulosiques et la chute de DP de la cellulose. A forte sévérité d’EV, une baisse de réactivité a été attribuée à la formation de pseudolignines. Les conditions optimales d’EV en vu d’une hydrolyse enzymatique sont les suivantes : T = 210°C, t = 5 min et H2SO4 = 0,25%. Une bonne fermentabilité des hydrolysats par Saccharomyces Cerevisiae a été obtenue. A partir de 100 Kg d’aubier d’okoumé, il est possible dans les conditions de notre étude de produire 11,9 Kg d’éthanol<br>The objective of this work is to study the valorization of co-products from the Gabonese okoumé wood industry in a biorefinery context. The production of cellulosic bioethanol and lignin based on a pre-treatment process that can be transposed to an industrial scale was studied. Our work, carried out on a laboratory scale, focused on (1) optimization of pre-treatment by steam explosion (SE), (2) extraction of lignin from exploded wood by percolation with ethanol as solvent, (3) characterization of biopolymers (cellulose and lignin) and study of the impact of pre-treatment parameters on their chemical structure, and (3) conversion of pre-treated wood to ethanol by enzymatic hydrolysis and fermentation. We were first interested in optimizing the different parameters of the SE process (particle size, preimpregnation, SE residence time and temperature) concerning the glucose, xylose and lignin contents of the exploded wood. The impact of the process on the production of sugar degradation products (furfural and HMF) potentially inhibiting enzymes and microorganisms was also studied. It was found that particle size is an important parameter to consider in process improvement, as fine particles lead to significant degradation of cellulose during pre-treatment. The influence of the concentration of acid, used to impregnate the biomass prior to SE, as well as process parameters (residence time and temperature) on the hydrolysis of biopolymers (lignin and hemicelluloses) was then studied. Conditions for optimizing cellulose contents (&gt; 59%) in the pre-treated material were selected. In order to better understand the impact of pre-treatment on the lignin extracted from exploded wood and to anticipate future SE lignin utilizations, spectroscopic (HSQC and 31P NMR) and chromatographic (SEC) analyses were carried out. A decrease in β-O-4 content is observed with the increase in the severity of the treatment of SE with an almost total degradation of the β-O-4 units under conditions of highest severity (T = 210°C, t = 2.5 min and H2SO4 = 0.5%). Recondensation of lignin through radical coupling processes was also observed. The celluloses of the different solid fractions recovered after EV treatment were also characterized by SEC, DRX to study the impact of the treatment and to identify the key parameters in the improvement of the efficiency of cellulose to glucose conversion. An increase in crystallinity with increasing severity was observed with a significant drop in cellulose DP for acid impregnated SE. The enzymatic digestibility of cellulosic residues after SE and lignin extraction by T. Reesei was studied. The main parameters positively affecting the enzymic reactivity are the removal of part of the non-cellulosic polymers and the drop of cellulose DP. At high SE severity, a decrease in reactivity was attributed to the formation of pseudolignins. The optimal conditions for SE before enzymatic hydrolysis are as follows: T = 210°C, t = 5 min and H2SO4 = 0.25%. Good fermentability of the hydrolysates by Saccharomyces Cerevisiae was obtained. From 100 Kg of okoumé sapwood, it is possible to produce 11.9 Kg of ethanol

L'objet de ce travail a ete la mise au point d'un capteur de suivi de fermentation en ligne, utilisant la technique de resonance magnetique nucleaire bas champ, basse resolution. Pour cela il a ete necessaire d'effectuer des developpements methodologiques dans deux domaines: sequences d'impulsions permettant la selection de l'information provenant de certains protons, traitement des donnees par decomposition multiexponentielle utilisant la methode de deconvolution de laplace-pade, ou par analyse multidimensionnelle, afin de detecter les differents composants de l'echantillon. Ces techniques ont ensuite ete utilisees pour etudier les melanges eau-sucre, et pour mettre au point des dosages: eau dans les sirops et les melasses, sucres et ethanol separement, ou simultanement dans un milieu de fermentation. Un automate: le c. A. F. E. (controle automatique de fermentation ethanolique), a enfin ete construit pour effectuer toutes les operations permettant un suivi en continu des fermentations alcooliques

ROQUETTE transforme et valorise le pois (Pisum sativum) pour produire des protéines, des fibres et des amidons. Au cours du procédé, diffé-rentes fractions secondaires sont générées dont le soluble de pois, désigné LAB 4960 après atomisation. Ce coproduit riche en fibres est inapte pour la nutrition humaine car il peut causer des désordres diges-tifs, engendrés par la forte teneur en α-GOS, des α-galacto-oligosaccharides formés de 1 à 3 unités de galactose liées par des liaisons α-(1-6). Parmi les α-GOS, on retrouve le raffinose, le stachyose et le verbascose qui sont non digérés par l’homme mais fermentés par le microbiote intestinal. L’objectif de ce projet de thèse est donc de ré-duire la teneur en α-GOS du LAB 4960 par voie microbienne afin d’en améliorer sa digestibilité. Pour atteindre cet objectif, une stratégie a été mise en place en deux temps. Dans une première partie, une collec-tion microbienne très diverse en termes d’espèce et d’origine (végétale vs animale) a été constituée à partir de graines de pois avec la caracté-risation de la diversité microbienne du pois de différents terroirs et à partir des collections internes de l’INRAE et de la société ROQUETTE. Dans une deuxième partie, la collection a été testée pour son aptitude à hydrolyser les α-GOS du LAB 4960. Le criblage de souches a été réparti sur trois phases, impliquant différents critères de sélection. La phase 1 a permis de sélectionner les souches capables de croître sur le LAB 4960 gélosé dans deux conditions d’oxygénation (aérobie et anaérobie) et de pH (acide et neutre). La phase 2 a permis d’identifier les sucres par Chromatographie sur Couche Mince après 72 h de culture sur le LAB 4960 liquide. Les souches ayant réduit les α-GOS ont été retenues en phase 3 pour le dosage des sucres par Chromatographie liquide haute performance couplée à la spectrométrie de masse. Dans la première partie, l’étude par métagénétique de la diversité de la surface de pois de différents terroirs après trempage, a montré une forte dominance d’espèces bactériennes appartenant aux Proteobacteria (57%) et Firmicutes (28%) et d’espèces fongiques appartenant aux Ascomyco-ta (89%) et Basidiomycota (11%). La structure de la communauté épiphyte associée à la graine de pois a été fortement impactée par son origine (coopératives et pays). A partir du jus de trempage des graines de pois, 102 souches ont été isolées et assignées à 52 es-pèces. Les 52 souches du pois représentatives de chaque espèce identifiée ont été ajoutées aux 157 souches représentatives de 82 espèces microbiennes des collections internes. Dans la seconde partie, le criblage de la collection a montré que 89% des souches testées ont pu croître sur le LAB 4960 gélosé. A peu près 20% des souches ont dégradé uniquement le saccharose. L’apparition des sucres mélibiose, manninotriose et manninotétraose traduisant une défructosylation a suggéré que 19% de souches ont hydrolysé les α-GOS via une β-fructosyltransférase dont 4% provenaient du pois. Enfin, 4% des souches ont hydrolysé les α-GOS via une α-galactosidase dont 1% provenait de pois. Sur les 49 (23%) souches hydrolysant les α-GOS, deux souches se sont démarquées par leur forte activité hydrolytique : Candida pseudoglaebosa CBS 6715T et Serratia liquefaciens GBM09. Une étude sur milieu minimum, milieu LAB 4960 et en bioréacteur sur LAB 4960 de concentrations diffé-rentes a montré que, dans des conditions optimales de croissance, la bactérie GBM09 est capable d’hydrolyser les α-GOS par ordre de degré de polymérisation croissant à pH neutre et à 20°C alors que la levure CBS 6715T hydrolyse l’ensemble des α-GOS simultanément à pH acide et à 28°C. Ces essais préliminaires ont permis de valider une première preuve de concept d’un aliment fonctionnel fermenté et laissent espérer favorablement leur développement à l’échelle industrielle en ouvrant la voie à de nombreuses innovations<br>ROQUETTE transforms and valorizes peas (Pisum sativum) to produce proteins, fibers and starches. During this process, various secondary fractions are generated, including the pea soluble, designated LAB 4960 after atomization. This fiber-rich co-product is unfit for human consumption in its current form as it can cause digestive disorders, caused by the high content of α-GOS, α-galactooligosaccharides formed from 1 to 3 galactose units linked by α-(1-6) bonds. Among the α-GOS are raffinose, stachyose and verbascose which are not digested by humans, but fermented by the intestinal microbiota. The aim of this thesis project is therefore to reduce the α-GOS content of LAB 4960 by microbial fermentation in order to improve its digestibility. To achieve this objective, a twofold strategy has been implemented. In a first part, a microbial collection that is very diverse in terms of species and ori-gins (plant vs. animal) was built up from pea seeds with the characteri-zation of the microbial diversity of peas from different terroirs and from the private collections of the INRAE Laboratory and the ROQUETTE Company. In a second part, the constituted collection was tested for its ability to hydrolyze the α-GOS from LAB 4960. The screening of the strains was split into three steps, involving different selection criteria. Step 1 allowed the selection of strains capable of growing on LAB 4960 agar under two conditions of oxygenation (aerobic and anaerobic) and pH (acidic and neutral). Step 2 allowed the identification of sugars by Thin-layer chromatography after 72 hours of culture on the liquid LAB 4960. The strains that reduced the α-GOS were selected in step 3 for the quantification of sugars by High performance liquid chromatog-raphy coupled to mass spectrometry. In the first part, the metagenetic study of pea surface diversity according to different terroirs after soak-ing, showed a strong dominance of bacterial species belonging to Proteobacteria (57%) and Firmicutes (28%) and fungal species be-longing to Ascomycota (89%) and Basidiomycota (11%). The structure of the epiphytic community associated with the pea seed was strong-ly influenced by its origin (storage cooperatives and countries). From the pea seed soaking juice, 102 strains were isolated and assigned to 52 species. The 52 pea strains representative of each identified spe-cies were added to the 157 strains representative of 82 microbial species in the internal collections. Screening of the collection showed that 89% of the strains tested were capable of growing on LAB 4960 agar. About 20% of the strains degraded only sucrose. The occurrence of sugars as melibiose, manninotriose and manninotetraose, known to be the product of defructosylation, suggested that 19% of the strains hydrolyzed α-GOS by a β-fructosyltransferase of which 4% came from peas. Finally, 4% of the strains hydrolyzed α-GOS by an α-galactosidase, of which 1% came from peas. Among the 23% strains hydrolyzing α-GOS, two strains stood out for their strong hydrolytic activity: Candida pseudoglaebosa CBS 6715T and Serratia liquefa-ciens GBM09. A study on minimum medium, LAB 4960 medium and in a bioreactor on LAB 4960 of different concentrations showed that, under optimal growth conditions, the GBM09 bacterium is capable of hydrolyzing the α-GOS in increasing order of degree of polymeriza-tion at neutral pH and at 20°C whereas the yeast CBS 6715T hydro-lyzes all the α-GOS simultaneously at acid pH and at 28°C.These preliminary trials have made it possible to validate a proof of con-cept for a fermented functional food and hold out promise of their development on an industrial scale, paving the way for many innova-tions

Le yogourt grec, pouvant être obtenu par concentration du yogourt traditionnel par ultrafiltration (UF), connaît une croissance exceptionnelle en Amérique du Nord (+100% depuis 2012), et représente le premier segment de marché des produits laitiers fermentés en 2014. Cependant, d'un point de vue environnemental, la production du yogourt grec fait face à plusieurs enjeux et défis. Son élaboration nécessite trois fois plus de lait que le yogourt traditionnel de par l'étape de concentration nécessaire à l'atteinte de la concentration protéique cible. De plus, l'étape d'UF du yogourt génère un perméat acide (coproduit du yogourt) difficilement valorisable. Néanmoins, une alternative consistant à effectuer l'étape d'UF sur le lait avant sa fermentation permet d'éliminer la production du perméat acide, et génère un perméat de lactosérum doux déprotéiné dont les voies de valorisation sont davantage connues. Cette stratégie pourrait donc potentiellement réduire l'impact environnemental du procédé et générer des coproduits plus facilement valorisables, améliorant ainsi l'écoefficience du procédé de fabrication de yogourt grec. Dans cette optique, ce projet de recherche visait à comparer l'impact environnemental sur l'ensemble du cycle de vie de la production de yogourt grec selon deux procédés : en effectuant l'étape de concentration par UF avant l'étape de fermentation (UF LAIT), ou après (UF YOG) comme utilisé dans l'industrie. Ainsi, des expérimentations à échelle pilote ont été réalisées dans le but de comparer ces deux procédés. Le nouveau procédé (UF LAIT) permettrait une réduction des consommations d'énergie à l'étape de transformation étant donné que l'UF du lait avant fermentation permet de réduire la quantité de matière première à transformer d'environ un tiers. Cependant l'Analyse du Cycle de Vie (ACV) des deux procédés donne comme résultat un bilan environnemental défavorable à (UF LAIT) comparativement au procédé traditionnel (UF YOG) à cause d'une plus grande consommation de lait, qui est responsable d'environ 80% des impacts sur le cycle de vie du yogourt grec. Cet impact majeur pour UF LAIT l'est encore même lorsque dans une analyse de sensibilité le perméat doux de UF LAIT est alloué à l'étape d'UF contrairement au perméat acide de UF YOG qui est considéré comme un déchet non valorisable.