BCH br Materials and methods br Results br Discussion

Materials and methods
Results
Discussion Prolyl 4-hydroxylases are oxygenases with key roles in a variety of biological processes including oxygen sensing, siRNA regulation and collagen folding (Gorres and Raines, 2010). Hydroxyproline is particularly abundant in collagenous proteins which present a characteristic helical region composed by the repetition of Xaa-Yaa-Gly sequences, where X and Y frequently are proline and hydroxyproline (Vanderrest and Garrone, 1991). Collagen is widely used in a plethora of industrial and medical applications in the form of animal tissue extracts (Lee et al., 2001). Since these products are not well characterized at the molecular level and may be contaminated with potential immunogenic and infective agents (Lee et al., 2001, Moon et al., 2014), the recombinant technology has been suggested and developed at various degrees of success, to obtain high quality and animal derived contaminant-free collagens (Olsen et al., 2003, Ritala et al., 2008, Vuorela et al., 1997, Xu et al., 2011). In the last years, an increasing interest in sponge collagen production for biomedical and cosmetic purposes has been developed (Nicklas et al., 2009a, Nicklas et al., 2009b, Swatschek et al., 2002), but until now, its recombinant production was hampered by lack of biomolecular information on the collagenous genes and the BCH involved in their biosynthesis. However, the recent identification and cloning of the first C. reniformis collagen gene (Pozzolini et al., 2012), as well as of an enzyme involved in its post-translational modifications (Pozzolini et al., 2015), has definitely laid the foundations for the setup of sponge collagen large scale production by recombinant technology. The present paper demonstrates that with the optimal combination of expression vectors it\'s possible to integrate the three C. reniformis genes responsible for sponge collagen synthesis (ColCH) and post-translational modifications (α and β P4H) into the P. pastoris genome using a commercial strain without resorting to particular genetic manipulations. Usually the available commercial yeast strains are designed to produce a single recombinant polypeptide at once, typically such yeast strains are defective in a gene coding for enzymes involved in fundamental metabolic pathways and are transformed with expression vectors that can restore this defect. In order to obtain a P. pastoris strain able to produce a stable sponge P4H tetramer, the cloning strategy used was a co-transformation of the commercial Pichia Pink strain with ade+ genotype with the pPink expression vector containing the sponge αP4H gene together with an αpPIC6 vector containing the sponge PDI gene (Scheme 1). The co-transformed strains were then easily selected on blasticidin agar plates lacking adenine. In the αP4H cDNA we maintained the original ER retention signal peptide sequence while in the PDI cDNA the same was replaced with the S. cerevisiae alpha mating factor pre-pro sequence as it had been already reported that this signal peptide gave the highest amount of P4H active tetramer in a previous work contemplating recombinant expression of human α and βP4H in P. pastoris (Vuorela et al., 1997). Using four different Pichia Pink strains and two different pPink vectors (Table 1), eight different yeast strains were obtained. The level of recombinant proteins produced was compared in the eight strains by P4H radioactive enzymatic assay. The highest enzymatic activity was obtained in the αβH4 strain cell lysate obtained by transformation of a P. pink strain defective for two different vacuolar proteases with the High-copy pPink vector (Fig. 2A). The analysis of the inserted gene copy number revealed that in this case there was no correlation between the P4H enzymatic activity and αP4H inserted gene copy number (Fig. 1A). In contrast, an inverse correlation between yeast intracellular protease activity and the recombinant P4H enzyme stability could be observed by comparing the enzymatic activity of P4H in the various protease-defective yeast strains (Fig. 2A) especially in the High-copy transfected strains (H strains). In fact, strain H1, with the lowest recombinant activity, has a completely efficient protease apparatus, strain H2 and H3, both measuring half recombinant enzymatic activity respect to strain H4, are defective for just one protease while strain H4, the best performing, is the one lacking most proteases (proteinases A and B, that consequently can neither activate carboxypeptidase Y), a condition clearly increasing the survival of the recombinant proteins inside the yeast cells (Cregg et al., 2000). Surprisingly, an enzymatic activity in the P. pastoris strain transformed only with the α subunit was also measured, indicating that sponge αP4H, differently from its human counterpart (Vuorela et al., 1997), could at least partially assemble with the yeast PDI, or alternatively the sponge enzyme could partially hydroxylate without the support of its β subunit, with a measured enzymatic activity 46.2±3.6% lower than the activity of the whole α and β P4H sponge tetramer (data not shown).