Molecular characterization and homology modeling of short-chain dehydrogenase (GcSDR) and NDP-sugar epimerases (GeNSEs) cDNAs from Gracilaria changii

Gracilaria changii is a source of agar in Malaysia. Despite its economic importance,many of its genes are unknown. The objective of this study was to characterize the sequence and functions of three novel proteins from G. changii using an approach that combines molecular biology and bioinformatics. Amino acid sequence analysis of Gracilaria changii short-chain dehydrogenase (GcSDR) showed that the GxxxGxG co-enzyme binding motif and the YxxxK active-site motif are well conserved. GcSDR transcript encoded 282 amino acids with a predicted molecular mass of 48 kDa. The open reading frame (ORF) of GcSDR was successfully cloned into pET32a(+) and expressed as a soluble protein in Escherichia coli strain BL21 (DE3) pLysS. Recombinant GcSDR showed the highest activities when pyruvaldehyde dimethyl acetal was used as a substrate compared to alcohols and steroid, with a Km value of 174.1 mM and a Vmax value of 0.87 μmol of product formed min-1 mg-1. GcSDR favoured NADPH rather than NADH, NADP+ and NAD+ as its primary co-factor. The three-dimensional structure of GcSDR was predicted using the crystal structure of a short-chain dehydrogenase (SDR) from Streptomyces avermitilis which shares 34.5 % identities. Virtual screening showed that CMP-Nacetyl-beta-neuraminate(2-) has the highest possibility to be the substrate of GcSDR with a total binding energy of -157.336 kJ/mol. The Gracilaria changii NDP-sugar epimerase 1 (GcNSE1) transcript encoded 576 amino acids (82 kDa) while Gracilaria changii NDP-sugar epimerase 2 (GcNSE2) encoded 401 amino acids (62 kDa). GcNSE1 shared high identities with the NADH: ubiquinone oxidoreductase complex I from Fischerella sp. JSC-11 while GcNSE2 has high identities with several NAD dependent epimerases /dehydratases. GcNSE1 possesses conserved co-enzyme binding pattern GxxxGxG and active site motif YxxxK for SDRs, while GcNSE2 has an atypical SDR co-enzyme binding site (GGxxxxG) and active site motif (YxxxN). The ORFs of GcNSE1 and GcNSE2 were cloned into pET32a(+) and transformed into E. coli strain BL21 (DE3) pLysS. Both recombinant GcNSE1 and GcNSE2 were produced as insoluble proteins. The insoluble fractions of both GcNSE1 and GcNSE2 were solubilized and refolded to produce soluble proteins. Threading using i-TASSER was performed for GcNSE1 using the X-ray structure of a tyrosine-phosphorylated protein from Arabidopsis thaliana (At5g02240) as template. The 3D structure of GcNSE2 was predicted using the crystal structure of Thermus thermophilus Hb8 UDP-glucose 4-epimerase complex with NAD. Virtual screening showed that UDP-D-glucose with the lowest total binding energy (-180.755 kJ/mol) could be the possible substrate of GcNSE2. The 5’ flanking regions of GcSDR, GcNSE1 and GcNSE2 were analyzed and putative cisregulatory elements (CREs) were predicted. The light responsive elements were frequently found in the 5’-flanking regions of GcSDR, GcNSE1 and GcNSE2 indicating that these three proteins might be regulated by light. The full length gene sequences encoding these three cDNAs have no introns. In conclusion, the approach used in this study has shed lights on the amino acid sequences and putative functions of three unknown seaweed proteins