We compared the sequences of 43 strains to understand the genomic diversity of P. aeruginosa ST277. According to their BioSample records, strains from a 20-year period were included in this study (1997–2017). The vast majority obtained were from Brazil (33/43), and overall, they represent human-derived isolates (31/43). The other countries represented are United States (6), United Kingdom (1), Mexico (1) and China (1). Phylogenetic reconstruction based on SNPs showed that the ST277 genomes (disregarding PAO1) shared a total number of 16,649 SNPs (4,517 core and 12,132 non-core). The clade including all ST277 genomes share 510 unique SNPs, while the clade with genomes from Brazil, 4 from United States and the single from United Kingdom share 920 unique SNPs. Overall, the genomes’ phylogenetic relationships do not seem to be associated to the year of isolation (Figure 1).
Fig. 1: Whole-genome SNP-based parsimony tree of 43 ST277 P. aeruginosa isolates and the reference genome PAO1 generated by kSNP3.0. The branch lengths are expressed in terms of changes per number of SNPs. After strains identification, separated by an underline, there are the SNP counts. The tree was visualized using Dendroscope (40). The paint shows the presence (black), absence (gray) or duplication (blue) of the genetic determinants surveyed. The purple bars represent an additional mutation in the gyrA gene (D87N), while the green bars represent samples with spacers organization other than ST277 standard spacers content. BR: Brazil, CH: China, MX: Mexico, USA: United Stated of America, UK: United Kingdom, Hum: Human, Env: Environmental, -: not available.
We computationally searched for resistance genes in all ST277 strains. These data demonstrate that 90% (39/43) of the genomes are positive for In163, 70% (30/43) for blaSPM–1 (carbapenem resistance), 65% (28/43) for rmtD (aminoglycoside resistance), and 42% (18/43) for crpP (ciprofloxacin resistance). All ST277 strains carry the genes blaOXA–50 (beta-lactam resistance), PDC–5 (beta-lactam resistance, Pseudomonas-derived cephalosporinases), catB7 (chloramphenicol resistance), bcr1 (bicyclomycin resistance), fosA (fosfomycin resistance) and aph(3’)-IIb (aminoglycoside resistance). The prevalence of OXA–50, PDC–5, catB7, bcr1, fosA and aph(3’)-IIb, among 2,576 P. aeruginosa whole-genome shotgun assemblies available at NCBI, are higher than 97% [16]. Therefore, these genes are related to the P. aeruginosa specie, and not exclusively to the ST277 (Figure 1, Additional file 1).
Non-synonymous mutations were found in the following resistance related genes: oprD, mexT, gyrA, nalC, pmrA, and phoQ. All ST277 strains have the same amino acid alterations in OprD (T103S, K115T), NalC (S209R, G71E), and PmrA (L71R), besides 8 nucleotide deletion on mexT (240–247), which caused frameshift. Also, 38 strains (88%) have 2 nucleotide deletion on oprD (379–380) and 19 bp deletion on mexZ (437–456) genes which caused [the] frameshift. Forty strains have the T83I mutation on GyrA in addition to D87N mutation on the same protein observed for 9 strains (21%); and one strain has a mutation on PhoQ (V260G) (Figure 1, Additional file 1). Together, the mechanisms found in this work that may contribute to antimicrobial resistance in the P. aeruginosa samples of ST277 are shown in Table 1.
Table 1: Mechanisms contributing to antimicrobial resistance in the ST277 strains.
Mutations in the virulence genes were analyzed using the genes from P. aeruginosa PAO1 as a reference. Fourteen strains (32.5%) have different point mutations or one nucleotide deletion in lasR gene, 7 strains (16%) have the same non-synonymous mutation on rhlR (K222T), and a fragment with eleven nucleotides was absent in algB (382–393) for 3 strains (7%). All ST277 strains carrying the virulence genes exoS, exoT and exoY (Figure 1, Additional file 1).
Genomic comparisons using pangenome as a reference visually showed that some regions are not shared by all ST277 genomes. We verified that these regions were PAGIs, already described in previous works (PAGI–13, PAGI–14, PAGI–15, PAGI–19, PAGI–23, PAGI–25, PAGI–29, PAGI–32, PAGI–34, PAGI–36 and PAGI–39) [8, 17], or a plasmid with 49 Kb, present in 3 strains (Additional file 8). Furthermore, some of these PAGI harbor typical ST277 genetic determinants. In case of strains positive for the blaSPM–1 gene, it is inserted inside PAGI–15. For those strains that carrying class 1 integron In163 or rmtD gene, they are contained in PAGI–25.
Computational analysis identified intact type I-C CRISPR-Cas system in 30 genomes from ST277 (70%) (Figure 1, Additional file 1). Two P. aeruginosa isolates (AZPAE14819 and AZPAE14822) have the complete Cas operon, but the CRISPR associated was not detected. No other type of CRISPR-Cas system was found among the ST277 strains. When present, the type I-C CRISPR-Cas system was uniformly localized on PAGI–34 in the ST277 genomes analyzed here.
We analyzed the whole spacers content present in type I-C CRISPR-Cas system identified in P. aeruginosa genomes. Among ST277 strains with intact type I-C CRISPR-Cas system, 90% (27/30) have the same 39 spacers content, considered by us the standard spacers content of this ST (Additional file 2). The strain 213BR has a CRISPR array of 22 and another of 11 spacers, separated by 200bp from each other and near to Cas operon. These 33 spacers are included in the ST277 standard spacers content. We detected three CRISPR arrays in the strain PA19, with 12, 13 and 6 spacers, all of them included in the ST277 standard spacers content. The first CRISPR array (12 spacers) is associated with Cas operon. Strain LIM1680 has a set of 37 spacers, associated with Cas operon and included in the ST277 standard spacers content. We noticed that 64% (25/39) of the spacers are present in all genomes from this ST suggesting a strong link between phylogeny and spacer content. We need to consider that possible sequencing and assembly errors may have influenced the spacers identification, so the percentage can be higher than 64%. The ST277 spacers at both ends are conserved between the strains, and the deleted/unidentified spacers are positioned in CRISPR central region (Figure 2).
Fig. 2: Spacer content among P. aeruginosa carrying type I-C CRISPR-Cas system. The numbered spacers belong to the ST277 standard spacers content. Colored and numbered spacers are those shared between four different clones (ST277, 235, 17 and 853).
From 57 NCBI genomes non-ST277 which have type I-C CRISPR-Cas system, 6 (10%) are from ST235. The other 51 genomes are from at least 32 different STs (Additional file 1). All isolates with type I-C CRISPR-Cas system appear to harbor PAGI–34, but lack some regions. The clones with the PAGI–34 more like to that of the ST277 belong to STs 17, 235, 853, and 1232 (Additional file 9).
The CRISPR spacers associated with Cas subtype I-C were analyzed, and a total of 345 non-redundant spacers were obtained (Additional file 2). Most non-ST277 genomes belonging to ST235 (5/6) have 21 spacers, and 80% of them are corresponding to the spacers in the ST277 standard spacers content. Besides ST235, two other STs have spacers that match those of the ST277 standard spacers content (ST17 and ST853). Eleven spacers were identified in the genomes from these four clones (spacers 3, 4, 5, 28, 29, 30, 32, 33, 37, 38 and 39). Most of them belong to the CRISPR trailer end (Figure 2).
Besides type I-C CRISPR-Cas subtype, 29 genomes also have type I-F CRISPR-Cas subtype. These genomes are from 18 different STs (17, 160, 274, 439, 553, 557, 589, 645, 646, 874, 1074, 1184, 1527, 1605, 1920, 2329, 2651 and 2717) (Additional file 1).
BLAT between the ST277 standard spacers content and the NCBI plasmid and phage databases assigned targets for 8% (3/39) and 38% (15/39) of the unique spacers, respectively. In some cases, a unique spacer presented homology with more than one phage and plasmid. We observed that the spacers with the greatest number of matches are distributed throughout the CRISPR array (Figure 3, Additional file 3).
Fig. 3: Number of ST277 spacers’ matches against public sequences of plasmid and phage. X: Spacers; Y: Number of matches.
Among the available plasmids and phages in the NCBI, 6 plasmids and 40 phages were targeted by some spacer (Additional file 3). Two target plasmids harbor an integron encoding resistance to aminoglycosides, beta-lactam, and sulfonamides (pPB354_1 and pPB353_1). Among the phages, 15 perform a lysogenic cycle, 2 a lytic cycle, 2 lytic-lysogenic switch, and for 21 we could not have this information.
We searched for anti-CRISPR genes (acr) in ST277 genomes, and on phage and plasmids targeted by CRISPR spacers for subtype I-C. Proteins were identified as candidate anti-CRISPRs in 4 P. aeruginosa genomes (Additional file 5). Among these genomes, two have CRISPR-Cas system (2/30) but they are not present in the other two. In total, 3 plasmids (3/6) and 20 phages (20/40) have putative Acr proteins (Additional file 6–7, respectively). The two plasmids harboring resistance genes described above also code for putative Acr proteins. Between the 27 anti-CRISPRs candidates’ genes found in this study, 63% were identified as known groups of acr genes, all them characterized as anti-type I-F (Additional file 4).