Recombinant Human Collagen and Biomimetic Variants Using a De Novo Gene Optimized for Modular Assembly

Abstract

A collagen-mimetic polymer that can be easily engineered with specific cell-responsive and mechanical properties would be of significant interest for fundamental cell-matrix studies and applications in regenerative medicine. However, oligonucleotide-based synthesis of full-length collagen has been encumbered by the characteristic glycine-X-Y sequence repetition, which promotes mismatched oligonucleotide hybridizations during de novo gene assembly. In this work, we report a novel, modular synthesis strategy that yields full-length human collagen III and specifically defined variants. We used a computational algorithm that applies codon degeneracy to design oligonucleotides that favor correct hybridizations while disrupting incorrect ones for gene synthesis. The resulting recombinant polymers were expressed in Saccharomyces cereVisiae engineered with prolyl-4-hydroxylase. Our modular approach enabled mixing-and-matching domains to fabricate different combinations of collagen variants that contained different secretion signals at the N-terminus and cysteine residues imbedded within the triple-helical domain at precisely defined locations. This work shows the flexibility of our strategy for designing and assembling specifically tailored biomimetic collagen polymers with re-engineered properties.

Tuning cellular response by modular design of bioactive domains in collagen

“…Recombinant collagen has been previously expressed by other groups [12,13] and is commercially available; however, these recombinant systems generate collagen with native sequence and do not demonstrate the ability to design multiple, relatively significant changes at specified locations throughout the biopolymer. This is because a synthetic bottom-up approach has been elusive, due to the difficulties in generating synthetic genes encoding the (glycine-X-Y) N backbone of collagen and the stringent need to post-translationally hydroxylate prolines in the Y-position for stability [14]. Consequently, synthesis of designer collagen-mimetic material has primarily focused on peptides [15e19], unhydroxylated collagen-like polymers from bacteria [20,21], or tandem repeating collagen domains [22].…”

“…Our research group has developed a platform for creating recombinant human collagen III in which we can specifically tailor the identity, location, and frequency of functional sites within the biopolymer [14,23]. Collagen III was selected as the molecular scaffold because it is a homotrimer; experimentally, this requires only one gene to be synthesized with no need to separate the heterotrimeric populations of product.…”

Expanding Functionality of Recombinant Human Collagen Through Engineered Non-Native Cysteines

“…Fabrication of Genes Encoding Collagen Variants. Full length collagen-mimetic proteins were fabricated using the 15 The triple-helical region of collagen, for which the gene has been conventionally difficult to assemble, was divided into 12 modules to facilitate synthesis and sequence design. By utilizing the degeneracy of the genetic code, oligonucleotide sequences were optimized to yield higher melting temperatures for correct hybridizations than for incorrect hybridizations.…”

“…Primary fragments with and without Cys were mixed-andmatched to assemble secondary fragments with Cys in the desired locations following previously reported protocols; 15 details are provided in Supporting Information. Full length collagen variant genes were assembled from the secondary fragments and sequenced to confirm the correct product.…”

“…13,14 Collagen's glycine-X-Y tripeptide repeating sequence, together with the unique sequences embedded within the tripeptide repeats required for cell interaction sites, and the large overall number of amino acids, result in difficulties in generating an encoding synthetic gene due to oligonucleotide mishybridizations. 15 Furthermore, prolines in the Y position must be posttranslationally hydroxylated for sufficient thermostability of the collagen triple helix. 16 These difficulties have limited most collagen-mimetic materials to short peptides 13,17−21 or repeating collagen-like domains.…”