Pichia pastoris as a cell factory for the secreted production of tunable collagen-inspired gel-forming proteins

It is the ability to establish triple helices and assemble into supramolecular structures, which makes collagen and its denature counterpart, gelatine, interesting for the food and biomedical industry. Collagen and gelatine array of applications is quite extensive, ranging from gelling agents in food, emulsifiers in photographic films, fillers in cosmetics and structural networks in drug delivery and tissue engineering systems. Despite their vast use, the animal origin of these protein materials, mainly bovine and porcine tissues, poses the risk of pathogen transmission or allergic reactions. However, mammalian tissues contain more than one type of collagen resulting in a lack of reproducibility and heterogeneity between batches. The variability in composition and structure of animal-derived collagen and gelatine presents a significant challenge for those using these proteins in medical applications, where reliable, predictable and traceable materials are essential. Additionally, the harvest of collagen and gelatine from animal tissues restrains us to the use of what is already available in nature and in commercial quantities. Thanks to the development of DNA recombinant technologies, the capability to design genes and obtain proteins with custom made conformations and functionalities has become a current day activity. In accordance, a myriad of recombinant systems have been developed to produce high yields of high quality heterologous proteins. However, there is no universal recombinant system that can be used by default. The choice between recombinant systems depends of the protein being expressed, the need for post translational modifications, the concentration of protein required and the costs associated to its production. In the present thesis, Pichia pastoris was the recombinant system of choice due to its many advantages. It is able to reach high cell densities on defined low-cost media, has a tight regulated AOX methanol induced promoter and can produce properly folded proteins with correct disulphide bond formation and other eukaryotic posttranslational modifications. The ability to secrete high amounts of heterologous proteins and low amounts of endogenous proteases is further considered as a plus by the biomedical industry, as downstream processing costs can be considerably reduced when compared to other industrial strains as Saccharomyces cerevisae or Escherichia coli. The possibility to custom design collagen and gelatine-inspired biomaterials and use a reliable production system like P. pastoris, is highly appealing for the safety demanding biomedical field. Our group has previously used this system for the secreted production of collagen-inspired proteins formed by two triple helix-forming (Pro-Gly-Pro)9 end blocks and one long random coil middle block. The triblock system allows the formation of gel networks with defined properties and porosities since i) triple helix is exclusively established by the (Pro-Gly-Pro) end blocks, connected by ii) a random coil middle block of known length. It was shown that the middle block can be independently tuned to originate gel networks with predictable rheological properties and drug delivery profiles and had no effect over the triple helices melting temperature. Only the collagen-inspired (Pro-Gly-Pro)9 end block domain seemed to contribute to the triple helix Tm which was ~ 41°C at a 1.1.mM protein concentration. Despite the possibility to establish triple helices (Tm ~ 41°C) at P. pastoris growth temperature (30°C), the secreted expression of these proteins was not impaired. Also, the elongation of the middle block proved to be no burden for P. pastoris secretory capability, since similar secretion yields were found for triblocks with middle blocks of ~37 or ~73 kDa. These positive results, paved the way to the design of a series of triblock polymers with different end block lengths and hence Tm. However, every system as a limit and P. pastoris is no exception. In Chapter 2, it was demonstrated that the elongation of the collagen-inspired (Pro- Gly-Pro)9 domain to (Pro-Gly-Pro)16 resulted in a considerable lower secretion yield and partial protein degradation by protease Yapsin 1. This result could have been explained by i) higher prolyl-tRNA turnover requirements, ii) higher hydrophobic content resulting in hydrophobic interactions or iii) incapacity to process highly stable triple helical conformations (Tm of ~74°C). The restitution of the secretion yields and intact product secretion by designing a polymer with noncollagen- like end blocks, revealed that the protein secretion impairment was due to intracellular triple helix formation. The non-collagen-like endblocks have the same amino acid composition but randomized sequence as to avoid triple helix formation. The intact gel-forming polymer could only be obtained by recurring to a P. pastoris strain deficient in the periplasmic protease Yapsin 1. Although the use of Yapsin 1 knockout strain allowed the secretion of intact polymer it did not restore the high secretion yields. To know where the secretion bottleneck was occurring, a closer look at P. pastoris cells was conducted. In Chapter 3, electron microscopy was performed to observe where triple helix formation was occurring. ER expansion and vesicle formation were observed only when gel-/triple helix-forming proteins with a high Tm were being expressed. Such observations were absent in cells expressing proteins with mutant end-blocks, endblocks with equal amino acid composition but unable to establish triple helices. The presence of extra organelles could have resulted from protein aggregation due to triple helix formation or ER volume enlargement as an attempt to minimize protein aggregation. This result suggests that triple helix formation should be avoided in order to achieve high secretion yields of proteins able to establish highly stable triple helices. In Chapter 4, it was shown that the thermomechanical stability of the triple helices could be tuned by varying the length of the (Pro-Gly-Pro) end blocks through genetic engineering. An increase of the triple helix Tm could be achieved by increasing the number of hydrogen bonds involved without the need to resort to chemical (covalent) crosslinking. All triblock copolymers studied formed stable hydrogels, and, an increase of the end block length resulted in higher stabilities under mechanical stress. The variation of the thermostability could be understood in terms of i) a simple linear relationship between the length of the end block and the free energy of helix formation, or ii) its constituting entropic and enthalpic components. The comparison with triple helices formed by free (Pro-Gly-Pro)n peptides seemed to indicate that this relationship was not influenced by the nature of the middle blocks. Regarding the use of P. pastoris as a cell factory, it was observed that there is an inverse relation between triple helix stability and secreted production yield. Proteins able to establish triple helices with high thermostablity were secreted at a lower yield. In Chapter 5, it was shown that denaturing SDS-PAGE gels can monitor the ability of collagen-inspired proteins to form supramolecular assemblies. During the destaining process proteins that cannot establish triple helices diffuse out of the acrylamide gel, while triple helix-forming collagen-inspired proteins are still visible after destaining over-night. Furthermore, it was observed that the diffusion speed from the gel is related to the triple helix Tm. Proteins that can establish triple helices with high thermostability diffuse slower from the gel. In addition, the migration of the middle blocks in a SDS-PAGE gel revealed that subtle differences in migration speed can expose changes in the amino acid sequence of random coil proteins. During the course of this project several attempts were made to increase the secretion yield of collagen-like proteins with high thermo stability. While not all experiments were successful, their results do yield additional hypothesis for further research that could lead to a further improvement of P. pastoris as a cell factory for the production of collagen-like proteins. These experiments are discussed in detail in Chapter 6.