Upon further evaluation of the

Upon further evaluation of the phage obtained from phage display, we compared the capture of each of the 4th round phage to a phage possessing the wild-type arylsulfatase A sequence (LCTPSR) using ELISA. Each phage was equivalently exposed to FGE and to the beads for capture; however, we found that the wild-type sequence (LCTPSR) exhibited a much higher ELISA signal than that of the 4(2) phage (Fig. 5). As such, the 4(2) sequence can be converted by FGE, but contrary to our expectation it was is in fact not the optimal substrate for FGE conversion. We believe that the 4(2) substrate sequence arose as the dominant phage display sequence perhaps by an additional selection mechanism other than FGE conversion/capture. Specifically, the 4(2) sequence could serve not only as a substrate for FGE conversion but also perhaps as a substrate for cleavage by trypsin. With this in mind, the 4(2) sequence may have arisen from preferential elution over other sequences that did not possess the auxiliary trypsin cut site. As shown in Fig. 6, proteolytic digest of the 4(2) peptide by trypsin could be observed, occurring at the 5th position of the HCTPRRP sequence. The effect of this on the extent of phage liberation would certainly be a key factor in the evolutionary selection process. From these results, we believe the selection process may have enriched for the 4(2) sequence as an FGE substrate that could also exhibited preferential elution rather than enriching for an optimal FGE substrate. The results of our new screening approach provided an effective means for enriching an FGE substrate sequence given the use of covalent coupling as the mode of selection. Computer simulations were used to model the FGE from M. tuberculosis as no crystal structure has been solved for the FGE from this organism, though as mentioned earlier prokaryotic forms of FGE exhibit impressively similar structures to eukaryotic and even human FGE [19]. Looking to existing literature finds that researchers have examined the human FGE substrate sequence, arylsulfatase A, to reveal the minimum motif C(X/T)P(X/S)R necessary for oxidation by the FGE product of human gene SUMF1 [7], [20]. The active site of both the prokaryotic and human FGE variants are capable of accepting 6 amino ephrin receptor residues inclusive of the CXPXR motif [19]. For the human analog, it was proposed that mutations of the P and R site will greatly reduce the efficiency of the enzyme and that mutations of the T or S residue are acceptable as long as one of those two residues remain [20]. In contrast, the FGE used in our study from M. tuberculosis was previously revealed to be tolerant to mutations in the CXPXR motif and flanking regions, as mutants of the native AtsG sulfatase motif of M. tuberculosis all remained active FGE substrates upon alanine replacement with the exception of mutations in the key cysteine residue to be transformed [11]. We performed docking simulations of a modeled FGE active site with the 4(2) substrate. Computational modeling of the FGE from M. tuberculosis using MODELLER [21] was based on its high homology with the FGE of Streptomyces coelicolor. In looking at existing crystal structures of human FGE bound with the CTPSR motif of arylsulfatase A, the presence of a key cysteine pair separated by four residues, namely catalytic C336 and the nearby C341 serving to form a disulfide bond with the substrate cysteine, are seen to exist adjacent to the purported oxygen binding pocket [8]. Comparing these structures with our simulated docking of 4(2) with the homology modeled FGE of M. tuberculosis (performed using AutoDock Vina) [22], several interesting similarities were found in our docking simulations including the arrangement of the substrate which appears to support disulfide bonding of the substrate cysteine (C2) with the exterior Cys268 (analogous to Cys341in humans) thus leaving the interior Cys263 (analogous to Cys336 in human) free for reaction with molecular oxygen (Fig. S4b,c). Several side-chain and main-chain peptide backbone interactions are found to occur with the FGE active site [23] as seen in Fig. S4c,d. In addition, the position of the guanidinium group in the substrate arginine (R6) lies in close proximity to the carboxylate of the FGE Asp282 which may suggest complementary electrostatic interactions (Fig. S4b). It has been suggested that for human FGE substrate sulfatases, substitutions of the substrate arginine residue for lysine are tolerated while substitution to alanine causes a dramatic reduction in substrate modification by FGE [7]. Interestingly, in this docking configuration the substrate proline (P4) does not appear to be tucked against the aromatic tryptophan residues which may support existing results showing tolerance of FGE from M. tuberculosis to substrates lacking this residue. The analogous placement of substrate C2 and R6 seen in our simulation as compared to known substrate bound FGE structures may be expected given the highly conserved nature of this enzyme and the fact that M. tuberculosis natively possesses CXPXR-type substrate sulfatases [24]. Although FGE functions in aerobic organisms, there exists related machinery in anaerobic organisms (i.e., Clostridium perfringens) for formylglycine formation that acts upon sulfatases having the CXAXR type substrate [25].