Introduction
Mussels can rapidly and strongly adhere to a variety of wet surfaces by secreting specialized adhesives, mussel adhesive proteins (MAPs), which are considered potential water-resistant bioglues (Cha et al., 2008; Lim et al., 2011; Silverman and Roberto, 2007; Waite et al., 2005). MAPs are characterized by the presence of specific post-translational modifications (Lim et al., 2011). In particular, hydroxylation of tyrosine residues results in a high content of 3,4-dihydroxy-L-phenylalanine (Dopa) with a reactive catechol group (Danner et al., 2012; Nicklisch et al., 2012). Dopa has been suggested as a key molecule for underwater mussel adhesion because Dopa mediates various interactions, including bidentate-hydrogen-bond formation, metal coordination, cation-π interaction, and oxidative cross-linking (Anderson et al., 2010; Harrington et al., 2010; Lee et al., 2006; Lu et al., 2012; Sever et al., 2004; Yu et al., 2013; Zeng et al., 2010). These interactions result in very strong underwater surface interaction and cohesion between proteins for integrity, water-resistance, self-healing properties, and so on (Harrington et al., 2010; Holten-Andersen et al., 2007; Holten-Andersen et al., 2011; Kim et al., 2014; Nicklisch and Waite, 2012).
To mimic mussel underwater adhesion in various fields, including medical, environmental, and industrial applications, biotechnology has been developed to produce bioengineered MAPs (bMAPs) via E. coliexpression due to its superior productivity compared to direct extraction from mussels, and because it has less toxicity, less immunogenicity, and better biocompatibility than synthetic polymers (Cha et al., 2008; Choi et al., 2011; Cha et al., 2008; Yang et al., 2013). However, E. coli –derived bMAPs lack of Dopa due to its intrinsic inability for post-translational modification, and this critically limits underwater adhesion. Because of this, there have been many trials to incorporate Dopa efficiently into bMAPs (Choi et al., 2012; Lim et al., 2011). To date, in vivo residue-specific Dopa incorporation is known as the most efficient method of incorporateing Dopa into bMAPs via mis-aminoacylation of Dopa by endogenous tyrosyl-tRNA synthetase (TyrRS) in a tyrosine-auxotrophic strain under minimal conditions (Yang et al., 2014). While the conventional method, in vitro mushroom-tyrosinase modification, showed a low modification yield of less than 15% (Yang et al., 2013, 2014), and in vivo modification via tyrosinase coexpression showed high instability between reduced and oxidized forms of Dopa, as tyrosinase can also catalyze oxidization of Dopa to Dopa-quinone (Choi et al., 2012), in vivo residue-specific Dopa incorporation allowed a high incorporation yield, around 90%, close to the yields of natural MAPs, and allowed Dopa to stay in its reduced form (Yang et al., 2014). Also, Dopa-incorporated bMAPs (Dopa-bMAP) produced by this method showed superior underwater adhesion ability, similar to natural MAPs (Yang et al., 2014). However, potential use of these high-quality bMAPs as practical bioglues was hindered by their low production yield, 3 ~ 5 mg/L (Yang et al., 2014). Therefore, there is a pressing need for increasing productivity toward commercialization. Such low production yield of protein obtained from residue-specific incorporation of a noncanonical amino acid is usually attributed to lower translational activity of a noncanonical amino acid compared to natural amino acids (Calendar and Berg, 1966; Hogenauer et al., 1978; Johnson, et al, 2010) (Fig. 1A). Therefore, in order to increase the productivity of Dopa-bMAPs, changes in translational machinery are required. Considering that the activation rates of noncanonical amino acids by aminoacyl-tRNA synthetase (AARS) were usually lower than those of natural amino acids, overexpression of AARS recognizing noncanonical amino acids is one feasible option for overcoming low activation rates.
In the present work, we engineered the translational machinery inE. coli to improve production yield of Dopa-bMAP with a high DOPA incorporation yield using two strategies. First, overexpression of Dopa-specific AARS and its cognate tRNA was investigated. Mutations at the binding pocket of AARS might be required for recognition of a specific noncanonical amino acid (Alfonta et al., 2003) (Fig. 1B). Previously, AARS with high selectivity for Dopa has been developed by screening the mutant libraries of M. jannaschii tyrosyl tRNA-synthetase (MjTyrRS) for site-specific Dopa incorporation into myoglobin with its cognate mutant tyrosine amber suppressor tRNA (MjtRNATyrCUA) in response to an amber codon (Alfonta et al., 2003). After mutation of the anticodon of MjtRNATyrCUA into AUA to recognize a Tyr codon (UAU), we designed a coexpression vector for Dopa-specific aminoacyltRNA synthetase (MjTyrRS-Dopa) and its cognate tRNA (MjtRNATyrAUA) and examined the improvement of production yield. Second, overexpression of E. coli tyrosyl-tRNA synthetase (EcTyrRS) was investigated. Overexpression of endogenous AARS in E. coli cells was shown to increase the production yield of recombinant proteins containing phenylalanine or methionine analogs (Kiick et al., 2000) (Fig. 1C). Although Tyr is a better substrate than Dopa for EcTyrRS, in minimal media deficient of tyrosine, EcTyrRS still activates Dopa, resulting in Dopa incorporation. Therefore, overexpression of EcTyrRS was expected to compensate low affinity of EcTyrRS for Dopa, which finally results in improved production yield, with plenty Dopa-charged E. coli tyrosyl tRNAs.