Although enzymes show great potential for use in industrial applications, their implementation from a practical perspective is still somewhat limited by various shortcomings in the area of enzyme immobilization. The use of silica sol-gels for protein entrapment has been studied extensively over the past 15 years or so. However, our understanding of the interactions between the immobilization matrix and the entrapped biomolecules is still relatively poor. Non-invasive in situ spectroscopic characterization is a promising approach to gain a better understanding of the fundamentals governing sol-gel immobilization. This thesis describes the application of Fourier transform infrared (FTIR) microscopy, two dimensional (2D) FTIR and fluorescence spectroscopy to characterize the immobilization matrix, entrapped model proteins and their interactions. Hydroperoxide lyase (HPL [E.C. 4.1.2.]) was chosen as a potential model protein for sol-gel entrapment. HPL activity was evaluated by use of factorial experimental design investigating the effects of KCl and Triton X-100 on HPL activity with 13-hydroperoxy-octadecadienoic acid (LA-OOH) and the novel water soluble 13-hydroperoxy-octadienoyl sulfate (LS-OOH) as substrates. The highest HPL activity was achieved under aqueous conditions with high salt and low surfactant concentrations and LA-OOH as the substrate. A significant interaction between salt and surfactant as well as salt and substrate was identified and a hypothesis to explain the basis of the interaction phenomena is presented. To analyze sol-gels with spectroscopic techniques, a sample format amenable to these techniques was needed. Therefore, a spin-coating technique for the preparation of aluminum or glass supported sol-gel thin films containing immobilized protein and a varying degree of the organically modified precursor propyltrimethoxysilane (PTMS) was developed. This approach produced samples that were suitable for chemical mapping using FTIR microscopy or fluorescence spectroscopic investigations. A data analysis method was developed to extract information on chemical speciation and distribution from FTIR data matrices obtained through FTIR microscopy. Results indicate that sol-gel thin films are not homogeneous on the microscopic level. Instead, they are heterogeneous with a clustering in the distribution of the model proteins studied (lysozyme [E.C. 3.2.1.17], lipase [E.C. 3.1.1.3] and bovine serum albumin (BSA)) at the scale investigated. The appearance of these clusters was found to depend on the type of protein entrapped, as well in some cases on the composition of the sol-gel. Moreover, the PTMS distribution was positively correlated with the protein distribution in the case of lipase and negatively correlated in the case of lysozyme and BSA. Additionally, sol-gels with a higher PTMS content appeared to conserve protein structure in areas where lipase clustered. Lysozyme and BSA, on the other hand, seemed to retain their structures in high concentration clusters better at lower PTMS content. A hypothesis taking into account the surface hydrophobicity of the proteins and the sol-gel composition as the basis for these phenomena is proposed. Fluorescence spectroscopy revealed that the PTMS content of the sol-gels had a direct effect on the physical properties of the immobilized proteins as evidenced by a blue shift of the intrinsic tryptophan (TRP) fluorescence. Temperature-dependent fluorescence spectroscopy revealed that the amount of TRP quenching was inversely proportional to the PTMS content of the sol-gel, suggesting that there were varying amounts of water available for quenching for the different immobilized enzyme systems. Analysis of the sol-gels by 2D FTIR spectroscopy with a focus on the amide A region using Gaussian peak deconvolution revealed two different species of water for the 50 % PTMS thin film sol-gels with BSA that could be described as fully and not fully H-bonded. It was also found that these species of water showed different removal profiles during thermal treatment. 2D FTIR of the amide I region followed by absorbance difference spectrum evaluation revealed that the temperature stability of the three model proteins was also sol-gel composition dependent. A hypothesis that the surface characteristics of the proteins determine the nature of the composition dependence is presented.

Although enzymes show great potential for use in industrial applications, their implementation from a practical perspective is still somewhat limited by various shortcomings in the area of enzyme immobilization. The use of silica sol-gels for protein entrapment has been studied extensively over the past 15 years or so. However, our understanding of the interactions between the immobilization matrix and the entrapped biomolecules is still relatively poor. Non-invasive in situ spectroscopic characterization is a promising approach to gain a better understanding of the fundamentals governing sol-gel immobilization. This thesis describes the application of Fourier transform infrared (FTIR) microscopy, two dimensional (2D) FTIR and fluorescence spectroscopy to characterize the immobilization matrix, entrapped model proteins and their interactions. Hydroperoxide lyase (HPL [E.C. 4.1.2.]) was chosen as a potential model protein for sol-gel entrapment. HPL activity was evaluated by use of factorial experimental design investigating the effects of KCl and Triton X-100 on HPL activity with 13-hydroperoxy-octadecadienoic acid (LA-OOH) and the novel water soluble 13-hydroperoxy-octadienoyl sulfate (LS-OOH) as substrates. The highest HPL activity was achieved under aqueous conditions with high salt and low surfactant concentrations and LA-OOH as the substrate. A significant interaction between salt and surfactant as well as salt and substrate was identified and a hypothesis to explain the basis of the interaction phenomena is presented. To analyze sol-gels with spectroscopic techniques, a sample format amenable to these techniques was needed. Therefore, a spin-coating technique for the preparation of aluminum or glass supported sol-gel thin films containing immobilized protein and a varying degree of the organically modified precursor propyltrimethoxysilane (PTMS) was developed. This approach produced samples that were suitable for chemical mapping using FTIR microscopy or fluorescence spectroscopic investigations. A data analysis method was developed to extract information on chemical speciation and distribution from FTIR data matrices obtained through FTIR microscopy. Results indicate that sol-gel thin films are not homogeneous on the microscopic level. Instead, they are heterogeneous with a clustering in the distribution of the model proteins studied (lysozyme [E.C. 3.2.1.17], lipase [E.C. 3.1.1.3] and bovine serum albumin (BSA)) at the scale investigated. The appearance of these clusters was found to depend on the type of protein entrapped, as well in some cases on the composition of the sol-gel. Moreover, the PTMS distribution was positively correlated with the protein distribution in the case of lipase and negatively correlated in the case of lysozyme and BSA. Additionally, sol-gels with a higher PTMS content appeared to conserve protein structure in areas where lipase clustered. Lysozyme and BSA, on the other hand, seemed to retain their structures in high concentration clusters better at lower PTMS content. A hypothesis taking into account the surface hydrophobicity of the proteins and the sol-gel composition as the basis for these phenomena is proposed. Fluorescence spectroscopy revealed that the PTMS content of the sol-gels had a direct effect on the physical properties of the immobilized proteins as evidenced by a blue shift of the intrinsic tryptophan (TRP) fluorescence. Temperature-dependent fluorescence spectroscopy revealed that the amount of TRP quenching was inversely proportional to the PTMS content of the sol-gel, suggesting that there were varying amounts of water available for quenching for the different immobilized enzyme systems. Analysis of the sol-gels by 2D FTIR spectroscopy with a focus on the amide A region using Gaussian peak deconvolution revealed two different species of water for the 50 % PTMS thin film sol-gels with BSA that could be described as fully and not fully H-bonded. It was also found that these species of water showed different removal profiles during thermal treatment. 2D FTIR of the amide I region followed by absorbance difference spectrum evaluation revealed that the temperature stability of the three model proteins was also sol-gel composition dependent. A hypothesis that the surface characteristics of the proteins determine the nature of the composition dependence is presented.

L’hydroperoxyde lyase (HPL) est une enzyme issue de la voie de la lipoxygénase, voie métabolique très représentée chez les végétaux, impliquée dans la production de composés aromatisants (l’hexanal, le 3Z-hexenal et le 2E-hexenal). Ces composés sont responsables de l’odeur fraîche de l’herbe coupée dite « note verte » et sont très utilisés par les industries cosmétiques et agroalimentaires. Leur biosynthèse résulte de l’oxydation des acides gras polyinsaturés en hydroperoxydes par la lipoxygénase, puis de leur clivage par l’hydroperoxyde lyase (HPL). Les procédés actuels de production de ces composés présentent certains inconvénients, ils sont notamment très polluants et peu performants, aussi l’utilisation d’enzymes recombinantes dans de tels procédés permettrait d’obtenir ces molécules de manière plus efficace tout en bénéficiant du label "naturel". L’ADNc codant pour l’hydroperoxyde lyase (HPLwt) a été isolé au laboratoire à partir d’olives noires. Afin d’améliorer la solubilité de l’enzyme, une HPL dépourvue de son peptide de transit chloroplastique (HPLdel) a également été produite. Les deux enzymes ont été exprimées chez E.coli, purifiées par chromatographie d’affinité puis caractérisées biochimiquement. Elles agissent exclusivement sur les 13-hydroperoxydes (13-HPL) à un pH et une température optimum de 7,5 et 25°C. De plus l’évaluation des paramètres cinétiques de l’enzyme montre qu’elles ont une meilleure efficacité catalytique (kcat/Km) sur les 13-hydroperoxydes d’acide linolénique (3,68 s-1.µM-1) que sur les 13-hydroperoxydes d’acide linoléique (0,54 s-1.µM-1). La bioconversion des 13-hydroperoxydes d’acide linoléique et linolénique en hexanal et 3Z-hexénal par l’action de l’HPLwt et l’HPLdel a été étudiée. Des taux de conversion maximum atteignant 93 % et 68 % pour la production d’hexanal et 73 % et 45% pour la production d’3Z-hexénal ont été obtenus quand l’HPLwt et l’HPLdel sont utilisées respectivement. La stabilité de l’enzyme a ensuite été étudiée. Des essais de conservation montrent que l’utilisation de glycérol à 10% (v/v) permet le maintien de la totalité de l’activité de l’HPLwt et de l’HPLdel durant cinq semaines de stockage à -80°C. De plus, l’ajout de composés chimiques tels que le KCl, le NaCl, le Na2SO4, la glycine et le glycérol permettent d’augmenter l’activité enzymatique des deux enzymes et d’améliorer les conditions de synthèse de l’hexanal et du 3Z-hexénal en diminuant la quantité d’enzyme nécessaire à leur production. / The hydroperoxide lyase (HPL) derives from a metabolic pathway named lipoxygenase pathway widely represented in plants and involved in the production of flavoring compounds (hexanal, 3Z-hexenal and 2E-hexenal). These volatile compounds are responsible for the fresh odor of cut grass known as "green note" and have a particularly interest for flavor and food industries. Their biosynthesis results from the oxygenation of linoleic and linolenic acids by lipoxygenase action to form fatty acid hydroperoxides, then of their cleavage by hydroperoxide lyase action. The processes of production currently used are highly polluting or lead to a low yield. To overcome these drawbacks, the use of recombinant enzymes in such processes constitutes an attractive alternative because they would allow producing these molecules in a more effective way, while benefiting from the "natural" label.A cDNA encoding for HPL (HPLwt) from black olive fruit was isolated, and in order to improve the enzyme solubility, the HPL deleted of its chloroplast transit peptide (HPLdel) was then produced. Both enzymes were expressed into E. coli (M15), purified by affinity chromatography, and characterized. They act exclusively on 13-hydroperoxide (13-HPL) and display an optimum pH at 7.5 and an optimum temperature at 25 °C. The bioconversion of 13-hydroperoxides of linoleic and linolenic acids in hexanal and 3Z-hexenal respectively, using HPLwt or HPLdel was studied. Conversion yields reach a maximum of 93 % and 68 % for hexanal production, and 73 % and 45 % for 3Z-hexenal when reactions were performed by HPLwt and HPLdel respectively.The enzyme stability was then studied. Conservations tests using 10 % glycerol (v/v) allows the maintenance of the entire activity of HPLwt and HPLdel during five weeks of storage at -80°C. Furthermore, the addition of chemical compounds such as KCl, NaCl, Na2SO4, glycine, and glycerol can increase the efficiency of both enzymes and improve the synthesis of hexanal and 3Z-hexenal by decreasing the amount of enzyme required to produce them.