Hintergrund: Die Synthese von Chemikalien aus fossilen Rohstoffen wird wegen ihrer begrenzten Verfügbarkeit und ihren negativen Auswirkungen auf die Umwelt zunehmend kritisch bewertet. Eine Alternative bietet die „Weiße Biotechnologie“, insbesondere die Fermentation nachwachsender Rohstoffe mithilfe von Hefen. Die oleophilen Hefen Pseudozyma (P.) tsukubaensis und Yarrowia (Y.) lipolytica sind natürliche Säureproduzenten. Ihre Hauptprodukte sind Metabolite des Tricarbonsäurezyklus: Citrat (CA), α-Ketoglutarat und Malat. In kleineren Mengen werden auch andere Stoffe wie Isocitrat (ICA) oder Itaconat (ITA, nur von P. tsukubaensis) sekretiert. Das Interesse an den beiden Letztgenannten hat in den vergangenen Jahrzehnten stetig zugenommen. Bis heute gibt es allerdings keinen etablierten Wirtsorganismus für die ICA-Produktion. ITA hingegen wird mithilfe von Aspergillus terreus synthetisiert. Jedoch stößt die ITA-Produktivität dieses Hyphenpilzes auch mit großem wissenschaftlichem Aufwand an ihre Grenzen. Daher wird ein neuer Wirtsorganismus benötigt. Ergebnisse: In dieser Studie wurden ein vielversprechender P. tsukubaensis-Stamm für die Produktion von ITA und ein Y. lipolytica-Stamm für ICA konstruiert. Zunächst wurde das Genom von P. tsukubaensis sequenziert. Infolgedessen wurde ein Gencluster für die Synthese und den Export von ITA identifiziert, das homolog zu dem von Ustilago maydis ist. Die Überexpression von vier der fünf Clustergene erhöhte die ITA-Sekretion nicht deutlich. Das fünfte Gen kodiert den vermeintlichen Transkriptionsfaktor Ria1p, der vermutlich das Gencluster steuert. Die Überexpression des PtRIA1 Gens führte zu einer signifikant erhöhten ITA-Produktion von bis zu 31,4 g/l in Mikrotiterplatten. Durch die Optimierung der Wachstumsbedingungen wurden im Bioreaktor innerhalb von 7 d 113,6 g/l ITA ohne die Notwendigkeit eines Triggers produziert. Für die ICA-Produktion wurden zwei mutmaßliche mitochondriale Citrat-Transportproteine in Y. lipolytica identifiziert, welche von den Genen YlCTP1 sowie YlYHM2 kodiert werden. Die Funktionsweise der beiden Proteine scheint sich stark voneinander zu unterscheiden. Die Deletion von YlCTP1 führte zu einer leichten Verschiebung des ICA:CA-Verhältnisses, aber die Gesamtmenge beider Säuren nahm stark ab. Durch die Deletion von YlYHM2 stieg die ICA:CA-Produktrate von 12 % auf 95 % im Vergleich zum Wildtyp. Innerhalb von 5 d wurden bis zu 131,9 g/l ICA mit Sonnenblumenöl, bzw. 22,0 g/l ICA mit Glukose als einzige Kohlenstoffquelle in einem Bioreaktor unter kontrollierten Produktionsbedingungen erreicht. Durch die zusätzliche Hemmung des Isocitratlyase-Proteins mit ITA stieg das ICA:CA-Verhältnis bis 98 %. Fazit: Mittels Metabolic Engineering wurden im Rahmen dieser Arbeit die beiden Hefestämme P. tsukubaensis HR12 und Y. lipolytica ΔYHM2 erzeugt. Mit ihrer Hilfe ist es möglich, die hochwertigen Chemikalien ITA oder ICA in hohen Mengen (> 100 g/l) aus nachwachsenden Rohstoffen wie Glukose oder sogar Pflanzenölen herzustellen. / Background: The synthesis of chemicals from fossil fuels is being evaluated increasingly critically, mainly due to its expected exhaustion and negative impact on the environment. An alternative offers ‘white biotechnology’, especially the fermentation of renewable resources with the help of yeasts. The oleophilic yeast species Pseudozyma (P.) tsukubaensis and Yarrowia (Y.) lipolytica are both natural organic acid producers. Their main products are metabolites of the tricarboxylic acid cycle, namely citrate, α-ketoglutarate and malate. In smaller amounts, other compounds like isocitrate (ICA) or itaconate (ITA, solely with P. tsukubaensis) are also secreted. The interest for the latter two has been rising steadily during the last decades. However, to this date, there is no established host organism for the ICA production. ITA, on the other hand, is being synthesised with Aspergillus terreus. Even with great scientific effort, the ITA productivity of this hyphal fungus appears to reach its limits. Therefore, a different host organism is needed. Results: In this study, a promising P. tsukubaensis strain has been constructed for the production of ITA and a Y. lipolytica strain for ICA. First, the genome of the ITA producer P. tsukubaensis has been sequenced. As a result, a gene cluster for the synthesis and export of ITA, homologous to that of Ustilago maydis, has been identified. By overexpressing four of the five cluster genes, respectively, none to low increases in ITA secretion were observed. The fifth gene is encoding the putative transcription factor Ria1p which probably controls the gene cluster. The overexpression of the gene PtRIA1 led to a significantly increased ITA production of up to 31.4 g/l in micro-wells. By optimizing the growth conditions 113.6 g/l ITA could be produced within 7 d under controlled conditions in a bioreactor without the need of a trigger like phosphate limitation. For the production of ICA, two putative mitochondrial citric acid transporter proteins were identified in Y. lipolytica. One carrier protein is encoded by the novel gene YlYHM2, the other one by YlCTP1. The mode of function for the two deduced proteins appears to be very distinct from one another. The deletion of YlCTP1 led to a minor shift in the ICA:CA ratio but the total amount of acids decreased greatly. By deleting YlYHM2, the ICA:CA product ratio could be increased from 12 % to 95 % compared to the wild type strain. Within 5 d up to 131.9 g/l ICA with sunflower oil and 22.0 g/l with glucose as the sole carbon source could be achieved under controlled production conditions in a bioreactor. Further inhibition of the isocitrate lyase protein with ITA increased the ICA:CA ratio to 98 %. Conclusion: Within this work, the two yeast strains P. tsukubaensis (HR12) and Y. lipolytica (ΔYHM2) have been created via metabolic engineering. With their help, it is possible to produce the value-added chemicals ITA or ICA on a high scale (> 100 g/l) from renewable resources like glucose or even vegetable oils.

Cyanobakterier är intressanta mikroorganismer för produktion av biobränslen från solljus, vatten och atmosfärisk koldioxid och anses därför vara potentiella mikrobiella cellfabriker. Men på grund av långsam tillväxt och låg produktion är genteknologi processen intensiv och tidskrävande för cyanobakterier. En alternativ metod till prototypteknik för metabola vägar är att använda cellfri metabolisk teknik där cellysat av överuttryckta enzymer används. I detta projekt försökte vi utveckla en metod för cellfri metabolisk ingenjörsteknik för cyanobakterien Synechocystis PCC 6803 med hjälp av den övre mevalonatvägen som exempelreaktionsväg. Vi började med att utveckla tre fluorescensbaserade metoder för att detektera proteinöveruttryck med hjälp av de tre enzymerna från mevalonatreaktionsvägen. Dessa metoder använde fusering av YFP-proteinet till målproteinet, en delad GFP-reporterprotein eller translationskoppling. Ett av de överuttryckta enzymerna verkade vara giftigt för Synechocystis-celler så flera inducerbara promotorer användes för att försöka uttrycka enzymet. Den högst uttryckande konstruktionen för varje gen valdes ut och proteiner extraherades och blandades i en cellfri metabolisk ingenjörsreaktion. Även om inget mevalonat kunde detekteras med hjälp av gaskromatografi i detta projekt, berodde detta sannolikt på otillräckligt högt proteinöveruttryck av mevalonatgenerna. / Cyanobacteria are desirable microorganisms for the production of biofuels from sunlight, water and atmospheric carbon dioxide, and are therefore considered potential microbial cell factories. But due to slow growth rate and low production rates, the engineering processes for bioproduction is labour intensive and time consuming. An alternative method to prototype metabolic pathway engineering is to use cell-free metabolic engineering, where cell lysates of enriched enzymes are used. In this project, we attempted to develop a method for cell-free metabolic engineering for the cyanobacterium Synechocystis PCC 6803 using the upper mevalonate pathway as an example pathway. We started by developing three fluorescence-based methods for detecting protein overexpression using the three enzymes from the mevalonate pathway. These methods used YFP fusion to target proteins, a split GFP reporter tag or translation coupling. One of the overexpressed enzymes appeared to be toxic to Synechocystis cells so several inducible promoters were used to try and express the enzyme. The highest expressing construct for each gene was selected and proteins were extracted and mixed in a cell free metabolic engineering reaction. Although no mevalonate could be detected using gas chromatography in this project, this was likely due to insufficiently high protein overexpression of the mevalonate pathway genes.

Dihydroxyacetone phosphate (DHAP)-dependent aldolases have been intensively studied and widely used in the synthesis of carbohydrates and complex polyhydroxylated molecules. However, the strict specificity toward donor substrate DHAP greatly hampers their synthetic utility. We transformed DHAP dependent aldolases mediated in vitro reactions into bioengineered Escherichia coli (E. coli). Such flask-to-cell transformation addressed several key issues plaguing in vitro enzymatic synthesis: 1) it solves the problem of DHAP availability by in vivo hijacking DHAP from glycolysis pathway of bacterial system, 2) it circumvents purification of recombinant aldolases and phosphatase, and 3) it dephosphorylates resultant aldol adducts in vivo, thus eliminating the additional step for phosphate removal and achieving in vivo phosphate recycling. The engineered E. coli strains tolerate a wide variety of aldehydes as acceptor, and provide a set of biologically relevant polyhydroxylated molecules in gram scale. Sialic acids exist in abundance in glycan chains of glycoproteins and glycolipids on the surface of all eukaryotic cells and some prokaryotic cells. Their presence affects the molecular properties and structure of glycoconjugates, modifies their functions and interactions with other molecules. The sialylation status, referring to the expression levels and linkages of sialic acids on the cell surface, is determined by the dynamic balance between sialylation and desialylation (removal of sialic acids). Sialylation is mainly regulated through expression and activity of sialyltransferases. And the mainstream idea attributes desialylation to the sialidases. However, more and more emerging evidences support the existence of ROS/RNS mediated chemical desialylation process under some pathological conditions. We used electrochemical oxidation of sialic acid conjugates to mimic ROS mediated chemical desialylation. Such electrochemical desialylation mimicry reveals that 1) β-linked sialic acid is much more difficult to de desialylated than α-linked sialic acid, 2) electron withdrawing residue and bulky underlying residue can facilitate the desialylation, 3) α- 2,3-linked sialic acid is easier to be desialylated than α-2,6- and α-2,8-linked sialic acid. This information is highly valuable for identifying the ROS species participated in ROS mediated desialylation and unveiling corresponding mechanisms. The mechanism of ROS mediated desialylation was proposed to go through radical decarboxylation.

The chemical diversity and biological activities of terpene and terpenoids have served in the development of new flavors, fragrances, medicines and pesticides. While terpenes are made predominantly by plants and microbes in small amounts and as components of complex mixtures, chemical synthesis of terpenes remains technically challenging, costly and inefficient. In this dissertation, methods to create new yeast lines possessing a dispensable mevalonate biosynthetic pathway wherein carbon flux can be diverted to build any chemical class of terpene product are described. The ability of this line to generate diterpenes was next investigated. Using a 5.5 L fed bath fermentation system, about 569 mg/L kaurene and approximately 207 mg/L abietadiene plus 136 mg/L additional isomers were achieved. To engineer more highly modified diterpenes might have greater industrial, agricultural or medicinal applications, kaurenoic acid production reached 514 mg/L with byproduct kaurene and kaurenal at 71.7mg/L and 20.1mg/L, respectively, in fed batch fermentation conditions. Furthermore, ZXM lines for engineer monoterpene and ZXB lines for engineer triterpene were generated by additional specific genomic modification, 84.76 ±13.2 mg/L linalool, 20.54±3.8 mg/L nerolidol and 297.7mg/L squalene were accumulate in ZXM144 line ana ZXB line, respectively, in shake flask conditions.

Metabolic engineering of microorganisms for chemical production involves the coordination of regulatory, kinetic, and thermodynamic parameters within the context of the entire network, as well as the careful allocation of energetic and structural resources such as ATP, redox potential, and amino acids. The exponential progression of “omics” technologies over the past few decades has transformed our ability to understand these network interactions by generating enormous amounts of data about cell behavior. The great challenge of the new biological era is in processing, integrating, and rationally interpreting all of this information, leading to testable hypotheses. In silico metabolic reconstructions are versatile computational tools for integrating multiple levels of bioinformatics data, facilitating interpretation of that data, and making functional predictions related to the metabolic behavior of the cell. To explore the use of this modeling paradigm as a tool for enabling metabolic engineering in a poorly understood microorganism, an in silico constraint-based metabolic reconstruction for the anaerobic, cellulolytic bacterium Clostridium thermocellum was constructed based on available genome annotations, published phenotypic information, and specific biochemical assays. This dissertation describes the analysis and experimental validation of this model, the integration of transcriptomic data from an RNAseq experiment, and the use of the resulting model for generating novel strain designs for significantly improved production of ethanol from cellulosic biomass. The genome-scale metabolic reconstruction is shown to be a powerful framework for understanding and predicting various metabolic phenotypes, and contributions described here enhance the utility of these models for interpretation of experimental datasets for successful metabolic engineering.

Au cours des dernières décennies, la fermentation IBE (isopropanol, butanol and éthanol) a connu un regain d'intérêt en vue de la production de carburants ou de composés chimiques à partir de matériaux renouvelables. Dans cette étude, nous avons étudié la production d'IBE avec le producteur naturel Clostridium beijerinckii NRRL B593 et avec des souches modifiées de Clostridium. acetobutylicum ATCC 824. En culture discontinue, la souche C. beijerinckii NRRL B593 excrétait 13.2 g/L d'IBE (dont 4,5 g/L d'isopropanol). Afin d'améliorer la production d'IBE, le gène codant pour l'alcool déshydrogénase secondaire (s-Adh) de NRRL B593, ainsi que différentes combinaisons des gènes des enzymes actives de la conversion de l'acétoacétyl-CoA en acétone, c.-à-d. l'acétoacétyl-CoA acétate/butyrate transférase (ctfA et ctfB) et l'acétoacétate décarboxylase (adc), ont été exprimées dans la souche productrice d'ABE (acétone, butanol éthanol), C. acetobutylicum ATCC 824. Les résultats montrent que la sur-expression des gènes ctfA et ctfB augmentait significativement la productivité et les concentrations finales en IBE tandis que la surexpression du gène adc n'avait qu'un effet limité. Cultivée en discontinu, la meilleure souche, exprimant les gènes adh, ctfA, ctfB et adc a produit 24.4 g/L d'IBE dont 8.8 g/L d'isopropanol avec une productivité de 0.7 g/L h. Cultivée en mode continu à un taux de dilution de 0.1 h-1, la productivité en IBE a été portée à 1.7 g/L h. Puisque le mélange IBE est considéré comme un additif carburant de qualité, les transformants obtenus constituent une avancée réelle vers le développement d'un procédé IBE industriel de production de biocarburants. / Over the past decades, the IBE fermentation (isopropanol, butanol and ethanol) has received a renew interest for the production of fuels or biochemicals from renewable materials. In the present study, we have investigated the IBE fermentation using the natural producer C. beijerinckii NRRL B593 and genetically-modified strains of Clostridium acetobutylicum ATCC 824. In batch culture, C. beijerinckii NRRL B593 was found to excrete 13.2 g/L IBE of which 4.5 g/L was isopropanol. To increase IBE production, the gene coding the secondary alcohol dehydrogenase (s-Adh) of C beijerinckii NRRL B593 and different combinations of genes coding for enzymes active in acetoacetyl-CoA to acetone conversion i.e. acetoacetate decarboxylase (adc) and acetoacetyl-CoA: acetate/butyrate: CoA transferase subunits A and B (ctfA and ctfB) were expressed in the ABE (acetone, butanol ethanol) producer C. acetobutylicum ATCC 824. Results showed that the overexpression of the ctfA and ctfB genes significantly increased both speed and extent of the IBE production while the overexpression of the adc gene had only a little effect. In batch culture, the best mutant (expressing adh, ctfA, ctfB and adc) produced 24.4 g/L IBE (of which 8.8 g/L was isopropanol) and displayed an IBE productivity of 0.7 g/L h. Cultivated in continuous mode at the dilution rate of 0.1 h-1, IBE productivity was increased to 1.7 g/L h IBE. As the IBE mix has been considered as a valuable fuel additive, the transformants obtained are a real step forward towards the development of an industrial IBE process for biofuel production.

Developing microbial systems capable of converting low cost lipids into value added products depends on the ability to acquire substrates from the growth media. Saccharomyces cerevisiae can acquire free fatty acids from the growth media and a portion of these lipids can be converted into new lipid products. However, they cannot acquire complex lipids from the growth media unless a nonspecific lipase is included. To circumvent lipase addition, we are genetically engineering S. cerevisiae to secrete a lipase into the growth media. We selected the LIP2 gene from Yarrowia lipolytica, which encodes a nonspecific lipase. Several modifications were made to the LIP2 gene to improve processing. Results identified strains secreting the most lipase. From these results, high producing strains were inserted into an oil inducible vector. Halo assays confirmed lipase secretion, while measuring the fatty acid composition confirmed triacylglycerol breakdown, and yeast uptake of the free fatty acids released.

Une stratégie d’ingénierie du métabolisme de C. acetobutylicum a été développée afin de construire une souche capable de produire de l’isopropanol à partir de sucres en C5, en C6 ou de substrats plus complexes. Dans un premier temps, une souche de C. acetobutylicum a été ingénieriée pour la production d’un mélange isopropanol/butanol/éthanol (IBE), ce microorganisme n’étant pas capable de produire naturellement de l’isopropanol. Différents opérons, exprimant une voie synthétique de production d’isopropanol, ont été construits et introduits à partir d’un plasmide dans une souche chez laquelle la voie de synthèse du butyrate a été supprimée (C. acetobutylicum ATCC 824 Δcac15ΔuppΔbuk). La souche la plus performante a été sélectionnée à partir de cultures réalisées en fermenteur, en mode discontinu à pH 5,0 et s’est avérée être celle exprimant la voie de l’isopropanol sous la dépendance du promoteur thl. Une optimisation des paramètres de culture a conduit à la production d’un mélange IBE, à partir de glucose, à une concentration de 21 g.l-1, un rendement de 0,34 g.g-1 et une productivité de 0,8 g.l-1.h-1. La production du mélange IBE à partir de xylose ou de xylane comme unique source de carbone a également été démontrée et permet une production IBE de 10,4 g.l-1 avec un rendement de 0,31 g.g-1 sur xylose et une production IBE de 4,28 g.l-1 avec un rendement de 0,28 g.g-1 sur xylane. Enfin, l’analyse des flux passant par la voie de l’isopropanol a permis d’identifier l’étape limitant la production de ce composé. Cette dernière semble être liée à la concentration en acétate intracellulaire et aux propriétés catalytiques la CoA-transférase, qui possède une faible affinité pour l’acétate. Ainsi, une CoA-transférase synthétique basée sur les caractéristiques de la CoA-transférase AtoAD d’E. coli, qui est décrite comme ayant un Km pour l’acétate plus faible, a été conçue et exprimée dans la souche précédement construite afin de tenter de lever la limitation de la voie de synthèse de l’isopropanol. Dans un deuxième temps, des modifications supplémentaires du métabolisme de C. acetobutylicum ont été effectuée afin de produire de l’isopropanol comme unique produit de fermentation à partir de glucose ou de xylose. Différentes stratégies ont alors été évaluées dans le but de contourner le déséquilibre rédox causé par la délétion des voies parasites consommatrices de carbone. Ainsi, des outils permettant la mesure d’activité hydrogénase, in-vivo et in-vitro, ont été développés pour tester la fonctionnalité de 3 hydrogénases, utilisant la bifurcation d’électrons pour la production d’H2 à partir de NADH et de ferrédoxine. Une deuxième stratégie utilisant les potentialités de la voie des phosphocétolases pour la métabolisation du xylose en acétyl-CoA a été étudiée et des résultats prometteurs ont été obtenus malgré les limitations actuellement rencontrées / First, C. acetobutylicum was metabolically engineered to produce a biofuel consisting of an isopropanol/butanol/ethanol (IBE) mixture. Different synthetic isopropanol operons were constructed and introduced on plasmids in a butyrate minus mutant strain (C. acetobutylicum ATCC 824 Δcac15ΔuppΔbuk) in which the butyrate pathway was deleted. The best strain expressing the isopropanol operon from the thl promoter was selected from batch experiments at pH 5.0. By further optimizing the pH of the culture, an IBE mixture with almost no by-products was produced at a titer of 21 g.l-1, a yield of 0.34 g.g-1 and productivity of 0.8 g.l-1.h-1, values never reached before. IBE production was also shown to be efficient using xylose or xylan as the sole carbon source with 10.4 g.l-1 IBE produce at a yield of 0.31 g.g-1 from xylose and 4.28 g.l-1 IBE produce at a yield of 0.28 g.g-1 from xylan. Furthermore, by performing in vivo and in vitro flux analysis of the synthetic isopropanol pathway, this flux was identified to be limited by acetate intracellular concentration and the high Km of CoA-transferase for acetate. A synthetic CoA-transferase based on the AtoAD E. coli characteristics was designed, synthesized and evaluated in vivo. This enzyme, that displays a lower Km for acetate, was found to be a good candidate to alleviate the bottleneck of the isopropanol pathway. Secondly, several strategies were evaluated to redraw C. acetobutylicum metabolism and finally construct a strain able to produce isopropanol as the only fermentation product from glucose or xylose. To overcome the severe redox imbalance caused by homo-isopropanolic fermentation, several strategies were investigated. On the one hand, a new class of electron bifurcating enzyme, the NADH hydrogenases, that can use NADH and ferredoxin to produce H2, were evaluated in C. acetobutylicum. This strategy opens the alternative to produce isopropanol and H2 from glucose without any carbon lost. On the other hand, the use of an alternative catabolic pathway, the phosphocetolase pathway, for xylose utilization and acetyl-CoA production was evaluated. These results allow the identification of the metabolic bottlenecks to overcome to obtain a C. acetobutylicum strain able to produce only isopropanol from xylose at high yield

Les plantes halophiles sont des plantes résistantes au stress salin, qui subissent une grande variété de stress dans leur environnement naturel. Ces conditions les ont menées à synthétiser des molécules de défense, qui peuvent présenter des activités biologiques intéressantes de par leur structure et diversité. Dans le cadre d’une collaboration avec la PME Salipouss, trois espèces ont été choisies sur la base de tests antioxydants préliminaires, avec pour objectif d’optimiser (i) la multiplication de plants in vitro pour des cultures industrielles en serre et (ii) d’améliorer le niveau d’activité de leurs extraits. La diversité des composés potentiellement actifs présents dans ces extraits est ensuite analysée par fractionnement bioguidé, afin d’isoler des molécules valorisables. Ce fractionnement est appuyé par des analyses de composés par RMN, permettant d’obtenir des informations sur la structure des composés bio-actifs. Les résultats obtenus montrent le fort potentiel de valorisation de ces trois espèces dans l’industrie, et plus particulièrement dans la nutrition et la cosmétique. / Halophytes are salt tolerant or salt-resistant plants which undergo high stress in their natural habitat. As a consequence of environmental stresses, they produce a number of active defense molecules which display interesting biological activities because of their diverse actions or structures. For the present study, three halophytic species were selected from preliminary antioxidant screening. In collaboration with Salipouss SME, objectives of the work were (i) to optimize in vitro halophyte multiplication in order to produce biomass under greenhouse and (ii) to elicit particular metabolic pathways in order to improve extract activities. To attempt to isolate molecules with potentially valuable activities, the variety of compounds from these extracts is reduced by successive fractionations. In addition, NMR analyzes allow to obtain indications on the nature and on the structure of the active compounds. First results highlight the strong activities of the selected halophytes, making them promising candidates for industrial uses, especially in nutrition and cosmetics.

Ces dernières années ont vu un grand développement des biotechnologies blanches et de l'ingénierie métabolique avec l'objectif de remplacer les procédés de synthèse de molécules d’intérêt de l’industrie chimique classique par des voies de synthèse biologique. Dans ce contexte, les réactions anaplérotiques, qui produisent les acides dicarboxyliques, sont particulièrement intéressantes puisqu'au delà de la production de ces molécules d’intérêt elles permettent une fixation nette de carbone, réduisant ainsi l’impact environnemental des procédés. Ce travail de thèse a donc porté sur l'élaboration d'une stratégie d'ingénierie métabolique faisant appel à des réactions de fixation de carbone inorganique chez la levure pour la production d'acide malique, une molécule plateforme ayant de nombreuses applications industrielles. La levure Saccharomyces cerevisiae a été choisie comme hôte pour sa commodité d’utilisation dans les procédés industriels et ses nombreux outils génétiques. L'approche développée repose sur la mise en place d'une voie de production d'acide malique par surexpression de la phosphonéolpyruvate carboxylase d'Escherichia coli (PEPC), de la malate déshydrogénase peroxysomale de S. cerevisiae relocalisée dans le cytosol (MDH) et du transporteur d'acides dicarboxyliques de Schizosaccharomyces pombe. La souche de levure recombinante obtenue a été caractérisée lors d'essais en fioles, en présence notamment de carbonate de calcium pour assurer un apport de carbone inorganique. Ces essais ont permis de mettre en évidence un effet stimulant de l'apport de carbone inorganique sur la production de malate et d'obtenir des concentrations de malate de l'ordre de 2,5 g/L à partir de 50 g/L de glucose, pour un rendement maximal de 0,046 gramme de malate par gramme de glucose. Des essais en bioréacteur de 5 L en présence d'air ou d'air enrichi à 5% de CO2 ont montré un effet positif de l'apport de carbone inorganique sous forme de dioxyde de carbone sur la production de malate. La concentration maximale de malate obtenue est de 1,46 g/L à partir de 50 g/L de glucose, soit un rendement de 0,029 gramme de malate par gramme de glucose. Des souches intermédiaires exprimant la PEPC et la MDH obtenues pour la production de malate ont également été caractérisées pour la production d'éthanol, car elles semblaient présenter une augmentation du rendement de production d'éthanol par effet transhydrogénase par rapport à la souche sauvage. Les essais n'ont cependant pas permis de confirmer cette augmentation de rendement. / White biotechnologies have been developing quickly during the last decades, aiming at replacing chemical syntheses by biological processes for the industrial production of target compounds. In this context, the implementation of anaplerotic reactions in the metabolism is of great interest, since those reactions allow both production of dicarboxylic acids and direct fixation of inorganic carbon. This work is about the development of a metabolic engineering strategy using inorganic carbon fixation reactions to produce malic acid, a compound with various industrial applications. The yeast Saccharomyces cerevisiae was chosen as a host for its convenient use in industrial processes and the availability of genetic tools. The approach developed to produce malic acid is based on the overexpression of Escherichia coli phosphoenolpyruvate carboxylase (PEPC), S. cerevisiae peroxysomale malate dehydrogenase relocated in the cytosol (MDH) and Schizosaccharomyces pombe dicarboxylic acid carrier. A recombinant yeast strain expressing those three genes was obtained and characterised in shake-flasks experiments, involving more specifically calcium carbonate as an inorganic carbon source. Those experiments showed an enhancement of the malate production in the presence of calcium carbonate and allowed to obtain a concentration of 2.5 g/L from 50 g/L glucose, for a maximal yield of 0.046 gram malate per gram glucose. Fermentation experiments were performed in a 5 L bioreactor in the presence of air or 5% CO2 enriched air; they confirmed the positive effect of inorganic carbon addition as CO2 on malate production. A malate concentration of 1.46 g/L from 50 g/L glucose was obtained, giving a yield of 0.029 gram malate per gram glucose. Intermediate recombinant strains expressing PEPC and MDH were also characterised, for ethanol production, as they seemed to give increased ethanol yields, probably due to a transhydrogenase effect. Shake flasks and bioreactors experiments did unfortunately not confirm the yield improvement.