Molecular cloning, sequence analysis and homology modeling of galE encoding UDP-galactose 4-epimerase of Aeromonas hydrophila.

A. hydrophila, a ubiquitous gram-negative bacterium present in aquatic environments, has been implicated in illness in humans, fish and amphibians. Lipopolysaccharides (LPS), a surface component of the outer membrane, are one of the main virulent factors of gram-negative bacteria. UDP-galactose 4-epimerase (GalE) catalyses the last step in the Leloir pathway of galactose metabolism and provides precursor for the biosynthesis of extracellular LPS and capsule. Due to its key role in LPS biosynthesis, it is a potential drug target. The present study describes cloning, sequence analysis and prediction of three dimensional structure of the deduced amino acid sequence of the galE of A. hydrophila AH17. The cloned galE consists of the putative promoter-operator region, and an open reading frame of 338 amino acid residues. Sequence alignment and predicted 3Dstructure revealed that the GalE of A. hydrophila consists of the signature sequences of the epimerase super family. The present study reports the molecular modeling / 3D-structure prediction of GalE of A. hydrophila. Further, the potential regions of the enzyme that can be targeted for drug design are identified.

Abbreviations

UDP-GlcNAc, UDP -UDP-N-acetylglucosamine ,GalE, UDP - galactose 4-epimerase, LPS -Lipopolysaccharides ,

Background

Aeromonas hydrophila is a member of the family Aeromonadaceae, associated with disease conditions mainly in fish, amphibians and humans [1]. Identification of strains of A. hydrophila capable of causing illness in apparently healthy individuals, by infecting open wounds and possibly by ingestion of the microorganism in food or water, has generated immense interest in this organism [2]. Some of the known virulence factors responsible for pathogenesis of A. hydrophila are O-antigen lipopolysaccharide, capsules, exotoxins, enterotoxins, and certain exoenzymes [3,4]. LPS has been reported to be involved in adherence and may play a role in antigenic variation [5-7]. Importance of the enzymes involved in galactose metabolism in bacterial virulence has been demonstrated [8-10]. GalE is one of the enzymes involved in galactose metabolism that mediates the incorporation of galactose in extra cellular polysaccharide material such as the O-side chain of lipopolysaccharide. The essential role of UDP-galactose 4-epimerase in virulence of many other gram-negative bacteria is well documented [11-15]. The fact that the epimerase mutants have altered LPS core biosynthesis with significant reduction in their ability to adhere and invade the host cells makes epimerase a potential drug target. GalE from different species exhibits a significant degree of interspecies variation at their gene and quaternary structure. In the present study, we report cloning, characterization of the galE, including its putative promoter, and structure modeling of the deduced amino acid sequence of the GalE of A. hydrophila

Methodology

Bacterial strains and vector

A. hydrophila (AH17) isolated from pond water was obtained from Dr. I. Karunasagar, College of Fisheries, Mangalore, India. Escherichia coli DH5α and BL21 (DE3) strains were from GIBCO BRL, USA and Novagen, USA, respectively. Plasmid pBCKS+ was procured from Stratagene (USA).

Cloning and sequencing of galE of A. hydrophila

Genomic DNA from A. hydrophila (AH17) was isolated essentially as described earlier [16]. The galE of A. hydrophila was PCR amplified using the genomic DNA as a template and the forward and reverse primers (5'-AGTCTGAGAAAAAGCGCGTGTG -3', 5'-TTAATCGGGATATCCCTGTGGATGG-3', respectively), designed on the basis of available sequence information of galE of E. coli (Acc. No. NC_000913) obtained from Microsynth, Switzerland. The PCR amplified product was purified and the ends of purified PCR product were phosphorylated using T4 polynucleotide kinase (NEB, USA), followed by ligation to Sma I digested dephosphorylated pBCKS (+) vector. Competent E. coli DH5α cells (Novagen, USA) were then transformed with the ligation mix and the transformants were analyzed by colony PCR and were further confirmed by restriction enzyme digestion for the release of the insert. The construct thus made was designated as pAHGalE. The integrity of the galE insert was verified by automated DNA sequencing (Applied Biosystem Model 393A).

Phylogenetic analysis

Sequence analysis tools of the ExPASy Molecular Biology Server of Swiss Institute of Bioinformatics were used to process nucleic acid sequence for the deduced amino acid sequence. The deduced amino acid sequence of the GalE of A. hydrophila was aligned with the GalE of other species and A. hydrophila AH3 strain was carried out using ClustalW (Version 1.83) [17]. The phylogenetic tree was inferred using Phylip's inference package, Version 3.5c.

Structure Modeling and Visualization of Model

The most appropriate template for Homology modeling of A. hydrophila GalE (Accession No. AJ785765) was identified using BlastP analysis. The available structure of GalE from Escherichia coli in the Protein Database (PDB) (PDB entry 1udcA, resolution =1.65, R value =0.177) was referred [18]. The target and the template sequences were aligned using ClustalW. Homology modeling program Swiss-Model was employed to generate a comparative 3D- structure model of A. hydrophila GalE [19]. Swiss-Model [20] is a server for automated comparative modeling of three-dimensional (3D) protein structures. No other refinements were applied. Swiss PDB viewer software [21] was employed as a tool to envisage the generated structural model.

Validation of the generated model

The generated 3D-model was assessed/reviewed evaluated at various structure verification servers viz. PROCHECK [22] that relies on Ramachandran plot [23], WHAT_CHECK, a subset of WHATIF programme [24,25], and VERIFY3D [26,27].

Discussion

Sequence analysis of the galE of A. hydrophila

Sequencing of the PCR amplified cloned galE fragment revealed the insert to be of 1140 bp, representing full length galE and its promoter-operator sequences (Figure 1). Sequence analysis revealed the presence of putative RNA polymerase binding site (-35 region) at 57-65 bp and a pribnow box (-10 region) at 76-84 bp. Putative binding site for catabolite repressor protein or cyclic AMP receptor protein (CRP) is also present at 49-69 bp, overlapping -35 region. The binding site for the GalR overlaps the translation start site and is present at 131-139 bp. Presence of all the regulatory sequences and components of a promoter upstream of the GalE encoding region of the cloned fragment indicate that the organization of galETK operon of A.hydrophila is similar to that of other gram-negative bacteria, which is organized in the order of galE, galT and galK. Open reading frame of the cloned galE contains a single protein translation start site ATG at 124-126 bp and a termination codon, TAA, is present at the 1138-1140 bp (Figure 1). Putative ribosome binding site is located 7 bp upstream of the ATG at 112-117 bp position. The encoded protein is of 338 amino acid residues with a theoretical pI of 5.64 and molecular weight of 36501.36. Genome database search (Blast N) showed varying degrees of similarity to nucleotide sequence of the galEs of other species. Blast P of the deduced amino acid sequence showed that GalE of A. hydrophila shares the percentage identities ranging from ~60-95% (95% with Shigella boydii and 63% with Photobacterium profundum) with different species of bacteria. It is of interest to note that while the GalE of the A. hydrophila AH17 shows significant identity with other bacteria, it showed only 59% identity and 85% similarity with the GalE of another strain of A. hydrophila, AH3 [15], though the active site and the catalytic sites/residues have remained conserved between the two. Bacteria of Aeromonas spp. are highly heterogeneous group of bacteria, and the differences in the GalEs of the two Aeromonas hydrophila strains may only be an indication of heterogeneity.

Figure 1

Nucleotide and deduced amino acid sequence of cloned galE of A. hydrophila. The open reading frame encodes for a protein of 338 amino acid residues. Initiation and termination codons are shown in bold. Putative -35 and -10 regions are shown as bold and underlined nucleotides. The shaded and the boxed bases represent putative ribosome binding site and the GalR binding site, respectively. Nucleotides pertaining to the primers used for amplification of the gene are underlined.

Analysis of amino acid sequence alignment of the GalE (Figure 1 in supplementary material) revealed that the A. hydrophila GalE consisted of the characteristic Tyr-X-X-X-Lys couple (position 128 to 133) that plays a key role in catalysis and a complete N-terminal NAD binding GXXGXXG motif (position 7 to 13), popularly known as ’Rossman fold‘. Both these motifs are conserved in all the family members and among all the species (Figure 2, boxed sequences) [28,29]. The signature sequences of the epimerase super family ‐ FSSSATVYG, ALLRYFNPVGAHP, NNLMPXXAQVAXGRR-XXXX-IFGNDYPTEDGTGVRDYIHV, YNLGAGXXXSVLDVVN that have remained conserved across the species are also present in the GalE of A. hydrophila (Supplementary figure 1, shaded sequences). Thus, epimerases from all the species appear to have the same evolutionary origin and employ similar catalytic mechanisms though they differ significantly in their subunits, quaternary structure and requirement for NAD. It is also of interest to note that though the signature sequences and catalytic couple have remained conserved in GalE of A. hydrophila, it differs significantly from GalE of other species outside these domains. As evident from the Phylogram (Figure 2) generated from the sequence alignment, the closeness of A. hydrophila GalE with that of E. coli, S. typhi, S. boydii and many others is not surprising as these are all enteropathogenic. What is of interest to note that the encoded GalE of A. hydrophila exhibited only 51% and 53% identities with that of H. sapiens and D.rerio, respectively, in which it is an important disease causing bacteria, thus making it a potential drug target. The distance between these species is also evident from the inferred phylogenetic tree.

Figure 2

Rooted phylogenetic tree of the deduced amino acid sequence of the GalE of A. hydrophila and other organisms. Amino acid sequences for different organisms were obtained from NCBI database and aligned using Clustal W program. The distances from the nodes, i.e. the branch lengths denoted in the tree, correspond to sequence divergence.

Structural model and Overall Architecture

X-ray resolved crystal structure of GalE from Escherichia coli (PDB entry 1udc) is available from Protein Data Bank (PDB). Based on the sequence alignment, GalE from Escherichia coli was found to be the best template structure for homology modeling of the target sequence. The comparative 3D- structure model of A. hydrophila GalE was generated by homology-modeling program Swiss-Model. The predicted model of A. hydrophila GalE (Figure 3) depicted in the form of ribbons is composed of twelve α-helices and eleven β-strands.

Figure 3

Homology model of the GalE of A. hydrophila. This model is produced by Swiss-Model program. Visualization of the structure was done by SWISS PDB VIEWER and is represented in the form of ribbons.

The assessment of the predicted model using the Ramachandran plot showed that the modeled structure has 89.2% residues in the most favorable regions, 10.8% residues occurring in the allowed regions and none of the residues in the disallowed regions. Such figures assigned by Ramachandran plot represent a good quality of the predicted model (Figure 4). All Ramachandrans show 6 labelled residues out of 336, whereas chi1-chi2 plots show 0 labelled residues out of 192. The main chain and side chain parameters for all of them were found to be concentrated/convoluted in the ’better‘ region. No bad contact was detected in the modeled structure. To define a model reliable, the score for G-factor (a log odds score based on the observed distribution of stereochemical parameters such as main chain bond angles, bond length and phi-psi torsion angles) should be above -0.50. The observed G-factor score for the present model was 0.04 for dihedrals bonds, 0.38 for covalent bonds and 0.18 overall. The distribution of the main chain bond lengths and bond angles were 99.9% and 99.1% within the limits, respectively. The modeled A. hydrophila GalE structure was also validated by other structure verification servers as such WHAT_CHECK and Verify-3D. For the modeled structure of GalE of A. hydrophila, 96.76% of the residues had an averaged 3D-1D score > 0.2 indicating a good quality of modeled structure. The modeled structure of A. hydrophila GalE is comparable to the structurally resolved GalE from Escherichia coli, wherein structural motifs have been identified to remain conserved. Since A. hydrophila has also been reported to infect humans, it is important to compare the depicted model with that of human UDP-Galactose 4-Epimerase, with which it shares only 51% identity. A superimposition of the A. hydrophila GalE onto the human epimerase monomer along with UDP-GlcNAc and NADH is shown in Figure 5.

Figure 4

Ramachandran plot of the predicted model of A. hydrophila GalE: This figure is generated by PROCHECK. The red regions in the graph indicate the most allowed regions whereas the yellow regions represent allowed regions. Glycine is represented by triangles and other residues are represented by squares.

Figure 5

(A). Superimposition of homology modeled structure of A. hydrophila GalE onto a Homo sapiens GalE monomer. A. hydrophila GalE is shown in red and Homo sapiens GalE in blue. α2 and β12 correspond to human GalE. Structural differences between the two structures are encircled. (B). Superimposition of the catalytic site of human epimerase/UDPGlcNAc/NADH structure onto the catalytic site of the predicted model of the A. hydrophila GalE. The human protein is shown in blue with UDPGlcNAc (in CPK color scheme) and NADH (in green) while the A. hydrophila protein is depicted in red. Figure shows the substitution of Tyr 299 in the A. hydrophila epimerase with a Cys307 in the human enzyme.

Superimposition of modeled structure of A. hydrophila GalE onto a Homo sapiens GalE monomer (subunit A of PDB entry 1HZJ) matches 335 Cα atoms with rms distance of 1.35Å and there is high conservation of sequence and structure between the two (Figure 5A). The core of the GalE subunit is highly conserved in both the structures, with differences confined to active site and some areas distant from the active site. In Homo sapiens GalE, three residues Ala305, Ala306 and Cys307 form a beta strand (β12) in C-terminal domain, whereas corresponding residues Pro297, Ala298 and Tyr299 of A. hydrophila GalE form coiled structure. When comparing A. hydrophila GalE to Homo sapiens GalE, and ignoring single amino acid differences, a stretch of six amino acids GGSLPE make a loop between β2 and α2 in Homo sapiens GalE, which is absent in A. hydrophila GalE, . Moreover, α2 of A. hydrophila GalE adopts a slightly different orientation compared with Homo sapiens GalE. A. hydrophila GalE also differs from that of fish species (D. rerio) by about 47%, therefore, the regions of differences between the two, can be targeted for drug design against the pathogen enzyme.

Superimposition of the catalytic site of the Homo sapiens and A. hydrophila GalE along with the UDP-GalNAc and NADH is shown in Figure 5B. It is well known, that in addition to catalyzing the interconversion of UDP-galactose and UDP-glucose, the human epimerase is also capable of interconverting UDP-GalNAc and UDP-GlcNAc [30]. Markedly, E. coli epimerase has not been reported for this activity. It is clear from the superimposition of A. hydrophila GalE onto Homo sapiens GalE (Figure 5B) Tyr299, a conserved residue in A. hydrophila GalE as well as in E. coli GalE has been is replaced with a Cys307 in Homo sapiens GalE. It can be suggested that the substitution of more bulky Tyr299 in the A. epimerase with a Cys307 in the human epimerase most likely prohibit UDPGalNAc from binding in the A. hydrophila GalE active site as has been reported for the E. coli GalE. These points can be taken into consideration for designing suitable inhibitors against A. hydrophila GalE.

Conclusion

The GalE activity is crucial for lipopolysaccharide biosynthesis, one of the virulent factors of A. hydrophila, and GalE mutants exhibit altered core LPS biosynthesis and reduced ability to infect the host cell. Therefore, inhibition of this enzyme can result in controlling Aeromonas infection. In the present study, cloning and sequence analysis of GalE of one of the Indian isolate of A. hydrophila revealed it to be different from other strains of the bacterium. The GalE of A.hydrophila exhibited greater degree of differences between the hosts, fish and human. Structure modeling of the A. hydrophila GalE resulted in identification of the structural differences between the GalE of the host and the pathogen. These differences can be targeted for drug design against the pathogen.