Accessory genome of the multi-drug resistant ocular isolate of Pseudomonas aeruginosa PA34

Bacteria can acquire an accessory genome through the horizontal transfer of genetic elements from non-parental lineages. This leads to rapid genetic evolution allowing traits such as antibiotic resistance and virulence to spread through bacterial communities. The study of complete genomes of bacterial strains helps to understand the genomic traits associated with virulence and antibiotic resistance. We aimed to investigate the complete accessory genome of an ocular isolate of Pseudomonas aeruginosa strain PA34. We obtained the complete genome of PA34 utilising genome sequence reads from Illumina and Oxford Nanopore Technology followed by PCR to close any identified gaps. In-depth genomic analysis was performed using various bioinformatics tools. The susceptibility to heavy metals and cytotoxicity was determined to confirm expression of certain traits. The complete genome of PA34 includes a chromosome of 6.8 Mbp and two plasmids of 95.4 Kbp (pMKPA34-1) and 26.8 Kbp (pMKPA34-2). PA34 had a large accessory genome of 1,213 genes and had 543 unique genes not present in other strains. These exclusive genes encoded features related to metal and antibiotic resistance, phage integrase and transposons. At least 24 genomic islands (GIs) were predicated in the complete chromosome, of which two were integrated into novel sites. Eleven GIs carried virulence factors or replaced pathogenic genes. A bacteriophage carried the aminoglycoside resistance gene (AAC(3)-IId). The two plasmids carried other six antibiotic resistance genes. The large accessory genome of this ocular isolate plays a large role in shaping its virulence and antibiotic resistance.


Introduction
Pseudomonas aeruginosa is associated with many opportunistic and nosocomial human infections such as pneumonia, septicaemia, corneal ulcers (microbial keratitis) and chronic infections in cystic fibrotic lungs [1,2]. Antibiotic resistance in this bacterium is alarmingly on the a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 rise and P. aeruginosa has been included in one of the top three priority pathogens urgently requiring new antimicrobial therapies for treatment by the World Health Organization [3]. Resistance to different antimicrobials in P. aeruginosa is due to its inherent capacity to oppose the action of antibiotics and its capacity to acquire genetic elements that often carry antibiotic resistance genes [4].
The study of the mechanisms used by different strains of P. aeruginosa to become resistant to antibiotics or acquire virulence have benefited from genome sequencing and these have identified a number of mobile genetic elements (MGEs) [5][6][7]. However, many of the genomes investigated are not completely closed and, in this context, accessory genes associated with resistance or virulence in addition to mutations may be missed. This may limit the understanding whether resistance and virulence genes are present on MGEs or "standard" chromosomal regions as well as their prevalence in general [8]. For example, less than 5% of P. aeruginosa draft genomes are complete; furthermore, only 37 plasmids of the total (3003 as of 15/10/2018) in the NCBI P. aeruginosa database are complete. Additionally, many of the isolates sequenced are derived from cystic fibrosis infections, while relatively few ocular and other isolates are represented. This is an important source of genetic information for P. aeruginosa as specific subpopulations are thought to be associated with microbial keratitis [9]. Notably, those subpopulations are predominantly characterised by possession of genes associated with type IV pili twitching motility, cytotoxicity (exoU), and certain genomic islands [9]. Indeed, to date, the complete genome of only one P. aeruginosa keratitis isolate has been published [10]. Very little information is available about the accessory genomes of ocular isolates.
P. aeruginosa strain PA34 (referred to as PA34 hereafter) is a multi-drug resistant microbial keratitis isolate which is resistant to gentamicin, imipenem, ciprofloxacin and moxifloxacin. Analysis of its draft genome revealed that PA34 belongs to sequence type 1284 based on multilocus sequence typing and carried at least 12 acquired resistance genes and possessed the exoU gene [11], an important effector gene in the pathogenesis of microbial keratitis [12]. A class I integron (In1427) that carries two antibiotic resistance genes has been shown to be integrated into a Tn3-like transposon carried by PA34 [13]. However, the genetic context of these resistance and pathogenic genes has not been completely elucidated. Hence, in this study we sought to analyse the complete genome of PA34, to examine the genetic structure of accessory genome and to compare the complete genome of PA34 with other genomes of P. aeruginosa from the public database.

Bacterial strain and microbiology
The strain Pseudomonas aeruginosa PA34 was isolated in 1997 from the cornea of a microbial keratitis patient in a tertiary eye care centre L.V. Prasad Eye Institute, Hyderabad, India. The cause of microbial keratitis was recorded as trauma induced by a stone in the right eye of a 21-year old male. Three complete genomes of P. aeruginosa were used for comparison: I) P. aeruginosa PAO1, which has the most complete and curated annotations [14,15], II) P. aeruginosa PA14, an exoU positive strain [16] and III) P. aeruginosa VRFPA04, an Indian microbial keratitis isolate [10].
The minimum inhibitory concentration (MIC) to three heavy metals (mercury, copper and cobalt) was determined by broth dilution methods as described elsewhere with minor modification [17,18]. PA34 and PAO1 (as the control strain) were examined against Hg ++ , Cu ++ and Co ++ by using twofold serial dilutions from 16mM-0.31mM. The metal salt solutions (HgCl 2 , CuCl 2 and CoCl 2 ) were prepared in deionised water, sterilized by membrane filtration and added to Mueller-Hinton broth (Oxoid Ltd, Hampshire, UK) to obtain the required concentrations. The experiments were performed in triplicate and repeated three times. Wilcoxon matched-pairs signed rank test was used to confirm significant difference between metal tolerance and the strain. A p-value of <0.05 was taken as significant.
Cytotoxicity was examined by the trypan blue dye exclusion assay [19]. Briefly, human corneal epithelial cells (HCEC) were grown in 96 well plates in the presence of SHEM (DMEM (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco), 1.05mM CaCl 2 , 0.5% DMSO, 2ng/mL epidermal growth factor (Gibco), 1% ITS-X (Gibco)). 200μL of 5x10 5 CFU/mL P. aeruginosa in SHEM was used to expose confluent HCEC. Following 3 h incubation at 37˚C, the cells were stained with 0.4% trypan blue. The stained cells were observed by microscopy and photographs were taken for the quantitative determination of cytotoxicity. P. aeruginosa PAO1 (a non-cytotoxic strain) were used as a control. The experiment was performed in triplicate and repeated three times.

MinION sequencing and complete genome assembly
Genomic DNA was extracted using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA) as per the manufacturer's protocol. The amount of extracted DNA was determined using a Qubit fluorometer (Life Technologies, Carlsbad, CA, USA). The library was prepared using the Ligation Sequencing Kit 1D R9 Version as per the manufacturer's instructions. The sequencing library was loaded into the Flow Cell Mk 1 Spot-ON of the Min-ION system using Library Loading Bead Kit R9 Version according to the manufacturer's instructions. The raw reads were obtained using the MinKNOW v1.7.14 in a 24-h run experiment and base calling was performed using Albacore v2.0.2. A total of 50,831 reads ranging from 174bp to 58,887bp (mean = 3,292bp) were obtained. The sequence reads from Illumina (which was obtained using MiSeq (Illumina, San Diego, CA) generating 300bp paired-end reads [11]) and MinION data were assembled using three hybrid assembly pipe-lines; Unicycler v0.4.3 [20], Japsa v1.8.0 [21] and Spades v3.12.0 [22]. Spades resulted in the best assembly in terms of contig numbers ( Table 1). The genome coverage was calculated using BBmap v35.82 [23].
All contigs were reordered with reference to the three complete genomes of P. aeruginosa. (PAO1, PA14, and VRFPA04) using MUMmer v3.0 [24]. Gaps between the contigs were determined by ABACAS v1.3.1 [25] using the "circular reference genomes" flag. Pseudogenomes of different sizes were observed after comparison with different reference strains and in all comparisons, contigs 6 and 7 were observed as unused contigs indicating they were not part of the chromosome. This was further verified by performing a BLAST of unused contigs against reference genomes. The fewest gaps between contigs in the PA34 genome (99 bp) were observed when PA14 was used as the scaffold compared to the other strains. Further, the size of the PA34 pseudogenome was slightly less than the size of the draft genome (~6.8 Mbp) [11]  . Therefore, this comparison was considered as the best and was further used to complete the genome of PA34. Primers were designed for each gap (S1 and S2 Tables) using Primer3web v4.1.0 [26] and the gap size was verified by PCR. Using this approach, we were able to close the gaps between contigs to generate a single contiguous chromosome based on contigs 1-5. Contigs 6 and 7 had a lower G+C content than rest of the contigs and showed similarity with different plasmids based BLASTn searches against NCBI database. Contigs 6 and 7 were therefore estimated to represent two different plasmids and were both circularised, referred to here as pMKPA34-1 and pMKPA34-2 (Plasmid 1 and 2 of Microbial Keratitis strain PA34).

Bioinformatics analysis
The complete chromosomal genome was annotated using NCBI Prokaryotic Annotation

General features of P. aeruginosa PA34 genome
The statistics of the complete genome of P. aeruginosa PA34 ( Table 2) were broadly matched with other published complete genomes of P. aeruginosa [39]. A total of 6,462 coding sequences (CDS) were predicted in the 6.8 Mbp genome of PA34 using NCBI Prokaryotic Genome Annotation Pipeline (PGAP) [40]. Of the predicted CDS, 6,314 were predicted to form functional proteins and 148 were pseudogenes. In the genomic comparison with three reference genomes (PAO1, PA14 and VRFPA04) and the genome of PA34, it was estimated there were 7,643 orthologs in the pangenome. Of those orthologs, 5078 were common amongst all four strains and are suggested to represent the size of the core genome. PA34 had the largest accessory genome compared to these reference strains (having 1,213 genes [~20% of total genes]) that had 543 genes unique to PA34. These exclusive genes were determined to encode features related to metal and antibiotic resistance, phage integrase and transposons. Furthermore, in the context of individual reference genomes, PA34 has 886, 737, 946 genes with no orthologs in PAO1, PA14, and VRFPA04, respectively (Fig 1) (See S1 File for full annotations of exclusive orthologs). A high proportion of strain-specific accessory genomic elements are common in P. aeruginosa strains and this represents the diversity of its genome [41,42]. However, in this study, we observed a lower range (3-10%) of exclusive orthologues than previously reported range (10-20%) [41]. The difference observed may be related to the use of the complete genome for the analysis over the use of draft genomes in previous studies. Since draft genomes are likely to be mis-assembled due to the presence of a high content of repetitive sequences in bacteria [43], analysis based on the complete genome is expected to be more relevant. PA34 shared 124 orthologs with VRFPA04, an eye isolate, and those genes were associated with chloramphenicol resistance, pathogenesis (type IV secretion system) and phages.
The diversity of core and accessory genomes observed in this study contradicts findings on environmental species P. fluorescens, where comparative genomics of three isolates demonstrated that they shared a core of 61% genes (vs 66% in this study) and 22-27% of unique genes to each isolate [44]. Furthermore, comparative genomics of three strains of the plant pathogen P. syringae showed a smaller core genome of 56% [45] than that of P. aeruginosa observed in this and another study [46]. The overall low diversity amongst P. aeruginosa strains compared to other Pseudomonas spp. may result from the selection of closely related strains (i.e. those containing exoU) in the present study.
No significant CRISPR and associated Cas genes were observed in PA34 based on comparison with the CRISPRCasFinder database [33]. CRISPR-Cas system may be negatively correlated with the size of the accessory genome and acquired antibiotic resistance genes because this system restricts the invasion and incorporation of MGEs. Thus, strains without a CRISPR-Cas system are expected to have a larger genome size [47].

Genomic islands
Genomic islands (GIs) refer to blocks of horizontally acquired DNA that integrate into certain loci in the core genome [48]. Such loci are known as Regions of Genomic Plasticity (RGPs). GIs should have a minimum of four contiguous open reading frames, which are not present in any of the set of genomes to which they are compared [5]. More than 80 RGPs have been identified in the genome of P. aeruginosa [42]. The position of RGPs in the genome are indicated by homologous flanking loci in PAO1 [5]. RGPs constitute a major portion of the accessory genome and are essential for the adaption of P. aeruginosa in diverse habitats. In this study, 24 RGPs were observed in the complete genome of PA34 (Table 3 and Fig 2) when compared with genomes of P. aeruginosa strains PAO1, PA14 and VRFPA04. A study has reported 27 to 37 RGPs in individual genomes of five P. aeruginosa strains [5]. The difference in number of observed RGPs may be due to a difference in the number of genomes used for comparison. Nevertheless, out of 24 GIs in the current study, two GIs were integrated into loci which have not been reported previously [42] and were deemed as GIs MKPA34-GI1 and MKPA34-GI2 in this study ( Table 3). The GI MKPA34-GI1 was 68.9 Kbp, was integrated between PA2858 and PA2859 homologues of PAO1 and carried genes for mercury and chromate resistance. The GI MKPA34-GI2 had a size of 39.9 Kbp, encodes genes associated with phages and was integrated into loci flanking 4856/4857 of PAO1. In addition, three GIs (RGP23, RGP56 and RGP84) carried phage related genes; phages can be associated with virulence [49]. Our analysis showed that at least five GIs were associated with replacement or insertion of pathogenic genes and one GI (RGP23) carried an acquired aminoglycoside resistance gene (AAC(3)-IId). Three GIs (RGP2, RGP50, and RGP57) were associated with the type IV secretion system.
Integrative conjugative elements (ICEs) are chromosomally integrated self-transmissible MGEs that can exist as circular extrachromosomal elements and are antecedents of various GIs [48]. Our analysis revealed three ICEs in the complete genome of PA34. The ICEs observed in RGP5 and RGP29 were related to clc-like ICE elements (Genbank Accession number AJ617740). The parental clc element contains genes for chlorocatechol (clc) degradation [50]. However, these genes were lost from both RGP5 and RGP29 of PA34. Nevertheless, genes required for integration and conjugation were identified in both of the ICEs in RGP5 and RGP29 (Fig 3A) indicating that these elements may have the capacity to transfer within and between species. This may be the reason for observing clc-like GIs in two different loci in this strain. In addition, RGP5 carried mercury resistance genes, which are not present in the parental clc elements. This finding suggests that the clc-like elements may incorporate antibiotic resistance genes during genetic recombination.
The third ICE observed in RGP41 of PA34 was related to the pKLC102 family (GenBank Accession number AY257538) and was similar to P. aeruginosa pathogenicity island PAPI-1 (GenBank Accession number AY273869) first identified in P. aeruginosa clone C (Fig 3B). PAPI-1 is self-transmissible and carries virulence factors such as genes for type IV pilus biogenesis and the cupD gene clusters The cupD genes are essential for the formation of cell surface fimbriae that are involved in biofilm formation [51]. Interestingly, the cupD gene cluster appears to have been lost from the RGP41 of PA34, although the genes for type IV pilus biogenesis are still present.
In addition, various members of pKCL102 ICE family are found to be associated with the carriage of exoU/spcU genes [48]. These GIs have frequently been referred as exoU islands such as exoU island A, exoU island B, exoU island C and PAPI-2 [52,53]. Possession of the exoU gene, that encodes the type III secretion system effector cytotoxin ExoU, markedly enhances the virulence of P. aeruginosa [54]. An exoU-island of 7.5 Kbp carrying five open reading frames was observed in PA34 (RGP7). In contrast to other exoU-islands that contain genes associated with integration, transposition, conjugations and sometimes antibiotic resistance [55], the exoU island observed here contains only three genes that have homology to a site-specific integrase (Fig 4A). The presence of both PAPI-1 and PAPI-2 like elements that act synergistically in virulence [56] may indicate an enhanced virulence of PA34.
To determine whether the exoU gene was functional, cytotoxicity of PA34 was examined in a human corneal epithelium cell line and compared with P. aeruginosa PAO1 which is a noncytotoxic invasive strain [57]. Microscopic examination after staining with trypan blue indicated that PA34 was highly cytotoxic (Fig 5) showing a large number of dead cells. Based on Accessory genome of ocular P. aeruginosa the percentage of dead cells, maximum cytotoxicity (of score 4) was observed in PA34 [19]. Previous studies have shown that possession of exoU is associated with in vitro and in vivo cytotoxicty, for example by using knockout mutants of P. aeruginosa strains such as PA103 and PAO1 [58,59]. The results in the current study suggest that exoU in RGP7 of PA34 was transcribed and translated into a functional protein that retained cytotoxicity. The finding also demonstrates that the size of the exoU island was not be associated with activity. However, direct confirmation of this result by producing exoU knockout of P. aeruginosa PA34 is required to confirm this result. A mercury resistance operon was also observed in the MKPA34-GI1 RGP that also carried chromate resistance genes and features related to transposition and conjugation. A BLASTn search against the plasmid database in NCBI did not show significant sequence similarity with MKPA34-GI1. However, 51% of the MKPA34-GI1 sequence was loosely identical (205 matches with an average identity of 97%) to a the complete genome of Pseudomonas stutzeri strain 273 (GenBank accession number CP015641.1). Pseudomonas stutzeri is an environmental organism and has the capacity to degrade pollutants such as toxic metals [60]. These observations indicate that the MKPA34-GI1 may have been acquired by horizontal transfer from organisms in mercury polluted environments. The ability of MKPA34-GI1 to transfer between strains needs to be confirmed by further experiments that will also help to confirm the MKPA34-GI1 RGP loci is mobile. In order to compare the phenotypic susceptibility to heavy metals, MIC to Hg ++ , Cu ++ and Co ++ were examined. High mercury tolerance was observed in PA34 (p<0.05) in comparison to PAO1 (Fig 6), possibly due to the presence of mercury resistance operon in two different GIs (RGP5 and MKPA34-GI1) in PA34. Copper tolerance was slightly higher in PA34 than PAO1 (p>0.05). Although bacteria require copper as a cofactor for many metabolic processes and can tolerate low concentration of Cu ++ , decreasing sensitivity to copper can be associated with acquisition of copper resistance genes [61]. The presence of a copper resistance operon in RGP23 in PA34 may be associated with higher MIC to Cu ++ . On the other hand, cobalt tolerance, for which no acquired genes were observed in this study was low in both PA34 and PAO1 (p<0.05). P. aeruginosa can tolerate Cu ++ and other heavy metals including Co ++ mainly due to P-type ATPase transporter and resistance nodulation cell division efflux pump [62]. In PAO1, the P-type ATPase encoded by PA3920 was found to be important for copper tolerance [63]. In addition, bacterial plasmids also carry metal resistance determinants [64,65]. However, plasmid associated metal resistance determinates were not detected in PA34.
The Pseudomonas phage MP38 was identified in the MKPA34-GI2 (flanking loci homologous to 4856/4857 of PAO1) (Fig 4B). MP38 is D3112-like phage which is a transposable and has been isolated from many clinical isolates of P. aeruginosa [66]. We examined the presence of MP38 in genomes of different P. aeruginosa and found that a similar phage was integrated into VRFPA04 near the PA4204 homologue of PAO1. This confirms the result of another study [67] that has shown that transposable phages may have variable integration sites, indicating that the phage MP38 does not have any specific integration site in the genome of P. aeruginosa. Therefore, the MKPA34-GI2 integration site may not be a constant RGP for P. aeruginosa.
Four RGPs (RGP31, RGP73, RGP9 and RGP60) of PA34 were associated with replacement of pathogenic genes and were related to lipo-polysaccharide O-antigen, pyoverdine (pvdE), flagella (fliC) and pilus (pilA) synthesis. Despite the replacement islands contains the same genes and being integrated into the same loci in the core genome, they highly diverse between strains [68]. These components are important for interaction of a bacterium with external environments including other species or hosts and hence are under continuous selection pressure [68][69][70][71]. The pvdE and fliC orthologs have been shown to vary greatly between different strains of P. aeruginosa [72]. The replacement islands may contribute to higher virulence in PA34 although this needs to be tested for swimming or twitching motility and for expression of pvdE under low iron condition.
The largest GI (RGP23) observed in PA34 was 125.1 Kbp and carried resistance genes for aminoglycosides (AAC(3)-IId) and tunicamycin. An insertion sequence of the family IS1182 was observed in this GI, which may be responsible for carriage of these resistance genes. A BLASTn search revealed that the antibiotic resistance genes were best matched with those of Acinetobacter sp. WCHA45 plasmid pNDM1_010045 (GenBank accession number NZ_CP028560.1) and were similar to many from other members of the Enterobacteriaceae where IS1182 is also present. In addition, a phage tail protein gp37, which was first identified in Enterobacter phage T4 [73] was found inserted into this island. This suggests that this resistance island may be derived from phages.

Plasmid features
The Illumina and ONT reads of the whole genome sequence of PA34 were assembled into seven contigs (> 500 bp) using hybrid strategies in SPAdes. Out of seven contigs, contigs 6 and 7 had a G+C content that less than that of other contigs and showed a significant match with plasmids in the NCBI database. This suggests that contigs 6 and 7 are likely to represent two different plasmids carried by PA34 and were named here as pMKPA34-1 and pMKPA34-2. pMKPA34-1 is 95.4 kbp with 57% G+C content and pMKPA34-2 is 26.8 kbp with 61% G+C content. Automatic annotation followed by manual confirmation with BLASTn revealed 98 CDS in pMKPA34-1 and 33 CDS in pMKPA34-2. Out of 98 predicted genes in pMKPA34-1, 46 were predicted to encode proteins with unknown functions whilst all 33 CDS in pMKPA34-2 were predicted to encode known proteins.
General features of pMKPA34-1. The putative plasmid pMKPA34-1 contains a replication gene (repE), a chromosome partition gene (smc), a plasmid stabilization gene (parB) and a plasmid conjugal transfer mating pair stabilization protein (traN). These genes may help replication, maintenance and transfer of the plasmid [74,75] (Fig 7). Based on BLASTn searches against the NCBI plasmid database, pMKPA34-1 had the best match (31% query cover with 94% identity) with plasmid pLIB119 of P. stutzeri strain 19SMN4 (GenBank accession number CP007510), followed by Citrobacter freundii strain 18-1 plasmid pBKPC18-1 (23% query cover with 99% identity) (Fig 7). The parA, smc and repE genes display high similarity to the corresponding genes of pLIB119. The presence of the traN gene, which encodes an outer membrane protein and is essential for F-mediated bacterial conjugation [75], indicates the potential exchangeability of pMKPA34-1. This transfer system is common amongst bacterial plasmids notably in Enterobacteriaceae [76] and its presence in pMKPA34-1 may indicate that the plasmid could be exchanged with members of the Enterobacteriaceae.
Transposase, integrase and recombinase genes formed a major portion of the mobile genetic elements in pMKPA34-1. ISsage analysis identified two putative prophage integrases, and site-specific recombinases (xerC and xerD) of the tyrosine family. The XerC-XerD system, which is found on both the chromosome and on plasmids helps separation and segregation of newly replicated bacterial dimeric chromosome by resolving it to monomers [77]. This system is essential in the segregation of multicopy plasmids and contributes to the stable inheritance of multicopy plasmids [78,79].
General features of pMKPA34-2. The plasmid pMKPA34-2 lacks identifiable replication genes. However, the genome coverage statistics (109x vs. 50x) ( Table 1) suggests it may have a higher copy number than the pMKPA34-1. pMKPA34-2 may use alternative mechanisms for replication. In the BLASTn search against the NCBI plasmid database, pMKPA34-2 was best matched (29% query cover with 88% identity) with plasmid pSSE-ATCC-43845 of Salmonella enterica subsp. enterica serovar Senftenberg strain 775WP (GenBank accession number CP016838), followed by plasmid tig00000727 of Klebsiella pneumoniae strain AR_0158 (Gen-Bank accession number CP021699) (14% query cover with 99% identity) (Fig 8). Moreover, pMKPA34-2 carried the series of genes (tnsA, tnsB, tnsC, tnsD and tnsE) that are associated with transposition and are related to Tn7 transposons, a phage integrase and a putative resolvase. Upstream of these mobile genetic elements, a multi drug export protein gene (mepA) was identified. MepA is multi drug efflux transporter of MATE (Multidrug And Toxic Compound Extrusion) family whose role in P. aeruginosa is not well understood. [81]. However, MepA has been associated with resistance to many antibiotics and microbicidal dyes (crystal violet and ethidium bromide) in Staphylococcus aureus [82]. Indeed, the presence of the mepA gene in this plasmid and its association with mobile genetic elements suggest that this gene may have been acquired through horizontal gene transfer.

Conclusions
The large accessory genome of P. aeruginosa strain PA34 indicates that this strain has a diverse genomic structure. The strain harbours twenty-four genomic islands and two plasmids that carry various metal and antibiotic resistance genes as well as several genes associated with virulence. This may be associated with the observance of higher antibiotic resistance, mercury tolerance and in-vitro cytotoxicity of PA34. The in-silico analysis showed that six antibiotic resistance genes were present in two different plasmids and one antibiotic resistance gene plus various mercury resistance genes were integrated into different GIs and these resistance genes have sequence similarities with that of either other environmental or enteric bacteria. Furthermore, the genome of the PA34 has been integrated by phage element (gp37) that has its origin in enteric bacteria and carried aminoglycoside resistance gene (AAC(3)-IId). These findings suggest that resistance and virulence in PA34 may have evolved due to environmental selection pressure where organisms acquire traits to survive predation by other inhabitants. Thus acquired traits enhance the pathogenesis process in human. Given the eye is susceptible to being infected by environmental strains of P. aeruginosa, examination of a larger number of eye isolates may be necessary to uncover any additional acquired genetic features associated with microbial keratitis. This will help our understanding of different aspects of Pseudomonas keratitis.
Supporting information S1