<p>In order to produce predictable and robust systems forprotein purification and detection, well characterized, small,folded domains descending from bacterial receptors have beenused. These bacterial receptors, staphylococcal protein A (SPA)and streptococcal protein G (SPG), possess high affinity to IgGand / or HSA. They are composed of repetitive units in whicheach one binds the ligand independently. The domains foldindependently and are very stable. Since the domains also havewellknown three-dimensional structures and do not containcysteine residues, they are very suitable as frameworks forfurther protein engineering.</p><p>Streptococcal protein G (SPG) is a multidomain proteinpresent on the cell surface of<i>Streptococcus</i>. X-ray crystallography has been used todetermine the binding site of the Ig-binding domain. In thisthesis the region responsible for the HSA affinity of ABD3 hasbeen determined by directed mutagenesis followed by functionaland structural analysis. The analysis shows that the HSAbindinginvolves residues mainly in the second α-helix.</p><p>Most protein-based affinity chromatography media are verysensitive towards alkaline treatment, which is the preferredmethod for regeneration and removal of contaminants from thepurification devices in industrial applications. Here, aprotein engineering strategy has been used to improve thetolerance to alkaline conditions of different domains fromprotein G, ABD3 and C2. Amino acids known to be susceptibletowards high pH were substituted for less alkali susceptibleresidues. The new, engineered variants of C2 and ABD shownhigher stability towards alkaline pH. Also, very important forthe potential use as affinity ligands, these mutated variantsretained the secondary structure and the affinity to HSA andIgG, respectively. Moreover, dimerization was performed toinvestigate whether a higher binding capacity could be obtainedby multivalency. For ABD, binding studies showed that divalentligands coupled using non-directed chemistry demonstrated anincreased molar binding capacity compared to monovalentligands. In contrast, equal molar binding capacities wereobserved for both types of ligands when using a directed ligandcoupling chemistry involving the introduction and recruitmentof a unique C-terminal cysteine residue.</p><p>The staphylococcal protein A-derived domain Z is also a wellknown and thoroughly characterized fusion partner widely usedin affinity chromatography systems. This domain is consideredto be relatively tolerant towards alkaline conditions.Nevertheless, it is desirable to further improve the stabilityin order to enable an SPA-based affinity medium to withstandeven longer exposure to the harsh conditions associated withcleaning in place (CIP) procedures. For this purpose adifferent protein engineering strategy was employed. Smallchanges in stability due to the mutations would be difficult toassess. Hence, in order to enable detection of improvementsregarding the alkaline resistance of the Z domain, a by-passmutagenesis strategy was utilized, where a mutated structurallydestabilized variant, Z(F30A) was used as a surrogateframework. All eight asparagines in the domain were exchangedone-by-one. The residues were all shown to have differentimpact on the alkaline tolerance of the domain. By exchangingasparagine 23 for a threonine we were able to remarkablyincrease the stability of the Z(F30A)-domain towards alkalineconditions. Also, when grafting the N23T mutation to the Zscaffold we were able to detect an increased tolerance towardsalkaline treatment compared to the native Z molecule. In allcases, the most sensitive asparagines were found to be locatedin the loops region.</p><p>In summary, the work presented in this thesis shows theusefulness of protein engineering strategies, both to explorethe importance of different amino acids regarding stability andfunctionality and to improve the characteristics of aprotein.</p><p><b>Keywords:</b>binding, affinity, human serum albumin (HSA),albumin-binding domain (ABD), affinity chromatography,deamidation, protein A, stabilization, Z-domain, capacity,protein G, cleaning-in-place (CIP), protein engineering, C2receptor.</p>

La régénération des solutions de lavage utilisées en Nettoyage En Place (NEP) présente un intérêt, tant économique qu’environnemental, pour les industriels laitiers. L’étape clé du NEP réside dans le lavage alcalin qui assure l’élimination des souillures organiques déposées sur les équipements. Ce lavage est généralement effectué par des lessives de soude, moins onéreuses que celles de potasse. Nos travaux ont visé à appliquer un procédé physico-chimique de régénération à des lessives de potasse en vue de rentabiliser leur intégration au NEP laitier. La régénération de solutions de potasse souillées par du lait a été examinée et comparée à celle de solutions de soude. L’efficacité du procédé à épurer les solutions de potasse en termes d’abattement de la DCO et de Ntot a été démontrée. L’analyse des solutions régénérées a par ailleurs mis en évidence l’amélioration de leurs propriétés interfaciales (tension superficielle, angle de contact). Ces propriétés, dont l’origine a pu être attribuée à l’accumulation de biotensioactifs, confèrent aux solutions de potasse régénérées un meilleur pouvoir nettoyant. Ce point a été validé lors d’essais de nettoyage de membranes organiques d’ultrafiltration colmatées par des protéines laitières. L’optimisation de la formulation de ces solutions a permis l’obtention de performances de nettoyage comparables à celles d’un détergent commercial classiquement utilisé pour ce type d’application. Enfin, les impacts économiques et environnementaux de l’intégration de lessives de potasse en substitution à celles de soude ont été évalués pour un NEP laitier industriel au travers d’une étude technico-économique et d’une Analyse de Cycle de Vie (ACV). / The regeneration of Cleaning In Place (CIP) solutions is interesting, both economically and environmentally, especially in dairy industry. The CIP’s key step lies on the alkaline washing which ensures the removal of organic contaminants deposited on the equipment. This washing is generally done by soda lyes, less expensive than potash ones. Our work aimed to apply a physicochemical regeneration process, based on adsorption-coagulation-flocculation phenomena, on potash lyes to make their integration in the dairy CIP affordable. The regeneration of potash solutions soiled with milk was examined and compared with the soda solutions regeneration. The process effectiveness to purify potash solutions in terms of COD and TN reduction has been shown. The solutions analysis has also highlighted the improvement of their interfacial properties (surface tension, contact angle). These properties, whose origin was attributed to the accumulation of biosurfactants, confer to potash regenerated solutions a best cleaning power. This point was validated during cleaning assays of organic ultrafiltration membranes fouled by milk proteins. The optimization of solutions formulation has also enabled obtaining a cleaning performance as efficient as those of a commercial detergent largely used for this type of application. Finally, economic and environmental impacts of the potash lyes integration in substitution to soda ones were evaluated for dairy industrial CIP through a techno-economic analysis and a Life Cycle Assessment (LCA).