Redefining the Origin and Evolution of Chromosomally Encoded bla_CTX-M/KLU in the Context of a Revised Taxonomy of Genus Kluyvera

ABSTRACT

Changes in Kluyvera taxonomy may clarify each species contribution for recruitment and dissemination of their relevant β-lactamases. The CTX-M-2 subgroup is linked to Kluyvera ascorbata, KLUC to Kluyvera cryocrescens, and CTX-M-25 to Kluyvera georgiana. The CTX-M-8 subgroup can be linked to Kluyvera genomospecies 3 and CTX-M-9 to Kluyvera genomospecies 2. Kluyvera sichuanensis and Kluyvera genomospecies 1 harbor new subgroups. The CTX-M-1 subgroup has a direct counterpart in an isolate proposed as a new genomospecies 5.

INTRODUCTION

CTX-Ms are the most prevalent group of extended-spectrum β-lactamases (ESBLs) among pathogens, representing a global pandemic (1–4). There are ∼233 CTX-M variants (http://bldb.eu/) (5), distributed in at least five gene clusters (6).

Because the first chromosomic variants reported in Kluyvera ascorbata displayed 95% to 100% amino acid identity with CTX-M-2, they were suggested to be the progenitors of the plasmid-encoded CTX-M-2 subgroup (7). Later, other chromosomally encoded enzymes, such as CTX-M-76 and CTX-M-95, were found already circulating among pathogens (8).

Other chromosomal counterparts of acquired CTX-M β-lactamases were identified in different species of Kluyvera. KLUC-1 (from Kluyvera cryocrescens) was the proposed ancestor of the CTX-M-1 cluster (9), but other plasmid-encoded enzymes were found to have direct chromosomal counterparts, i.e., CTX-M-3 and CTX-M-37, in an environmental isolate identified as K. ascorbata (6), and a K. cryocrescens isolate from a urinary tract infection (10). KLUG-1 (now CTX-M-152), the chromosomal β-lactamase from Kluyvera georgiana, was proposed as the putative origin of the CTX-M-8 subgroup (11) and the chromosomally encoded KLUY-1-4 (K. georgiana) for the CTX-M-9 subgroup (12), whereas the most recent member of this subgroup is the chromosomally encoded CTX-M-213 in Kluyvera ascorbata strain KA2 (13). CTX-M-78, a chromosomally encoded β-lactamase also reported in K. georgiana, has been suggested as the putative progenitor of the CTX-M-25 subgroup (14).

Even though participation of Kluyvera as the origin of antimicrobial resistance (AMR) genes is accepted, each species contribution is unclear because the assigned species from which chromosomal β-lactamases were recovered need to be revisited (15). A previous version of our manuscript included a deep reexamination of this genus taxonomy; however, a publication by Liu et al. (16) proposed a new species, Kluyvera sichuanensis sp. nov., and that other isolates whose sequences were available and used in our analysis should correspond to new genomospecies: K. cryocrescens L2, as the novel genomospecies 1; K. georgiana KA2 as a novel genomospecies 2; K. georgiana PO2S7 in the novel genomospecies 3; and K. intermedia D51-sc-1712206 as novel genomospecies 4 (ERR2221162). The cited reference sequences corresponding to Kluyvera genomospecies 4 could not be found.

(Part of this work was presented at ASM Microbe 2019, San Francisco, CA [15]).

In this study, we performed a genomic analysis comparing bla_CTX-M/KLU genes present in the Kluyvera chromosome to revise their association with the new proposed taxons and correspondence to subgroups of CTX-M.

We included 15 isolates of Kluyvera sp. recovered from clinical sources (see Table S1 in the supplemental material), preliminarily identified with biochemical tests, 16S rRNA gene sequencing, and matrix-assisted laser desorption ionization–time of flight mass spectrometry (data not shown). Whole-genome sequencing (WGS) was performed in a HiSeq X10 sequencer (Illumina, San Diego, CA, USA); paired forward and reverse reads were used as inputs for de novo assembly using the Velvet package (https://www.ebi.ac.uk/∼zerbino/velvet/) and 27 genomes from type, reference, and/or representative strains (including WGS assemblies related to Kluyvera obtained from NCBI genomes between 20 November 2016 and 25 March 2021) (see Table S2 in the supplemental material).

Sequences were aligned with ClustalX (http://clustalx.software.informer.com/2.1/). The molecular evolution model was estimated with JModelTest2 (http://gihub.com//ddarriba/jmodeltest2/releases), and unrooted phylogenetic trees (1,000 bootstraps) were obtained with PhyML (http://www.atgc-montpellier.fr/phyml/versions.php) and MEGA 7 (17), visualized and edited by FigTree (http://tree.bio.ed.ac.uk/software/figtree/).

WGS data from Kluyvera isolates were used to perform digital DNA-DNA hybridization (dDDH) using the genome-to-genome distance calculator (version 2.1; http://ggdc.dsmz.de) (18, 19) and average nucleotide identity (ANI) using the OrthoANIu (https://www.ezbiocloud.net/tools/ani) (20) (Table 1), confronting the isolates with the correspondent type strain (Table S2). Minimal cutoff points of 70% dDDH and 95% OrthoANI values were considered to represent species delineation.

TABLE 1.

Mean ANI and dDDH values for all strains

Straina	ANI (%)b	Estimated dDDH (%)c
Kluyvera ascorbata ATCC 33433	100	100
K. ascorbata 3162	98.65	89.10
K. ascorbata LFC	97.15	86.90
K. ascorbata 4663	98.46	87.90
K. ascorbata 58	98.58	88.10
K. ascorbata 68	98.73	89.60
K. ascorbata 8633	98.37	86.40
K. ascorbata 220	98.44	87.40
K. ascorbata 280	98.51	87.80
K. ascorbata 711	98.52	87.40
K. ascorbata TP1631	99.60	99.92
K. ascorbata Colony 413	94.74	56.40
K. ascorbata 13608	90.99	42.90
K. cryocrescens 169	85.12	29.50
Kluyvera sichuanensis 090646	100	100
K. ascorbata 13608	97.50	78.60
K. cryocrescens 169	85.42	29.70
Kluyvera georgiana ATCC 51603	100	100
K. georgiana 14751	97.89	82.30
K. cryocrescens 169	85.08	29.20
Kluyvera cryocrescens NBRC102467	100	100
K. cryocrescens 1919	99.21	93.50
K. cryocrescens 4701	99.26	94.40
K. cryocrescens SCW13	99.23	94.20
K. cryocrescens NCTC10483	99.18	92.70
K. cryocrescens NCTC12993	99.71	98.70
K. cryocrescens 169	89.89	40.00
Kluyvera genomosp. 1_L2	100	100
K. cryocrescens 169	89.50	38.20
Kluyvera intermedia ATCC 33110	100	100
K. intermedia N2-1	99.01	91.70
K. intermedia NCTC12125	99.95	99.90
K. intermedia HR2	99.01	91.70
K. cryocrescens 169	85.85	30.40
Kluyvera genomosp. 2_KA2	100	100
K. cryocrescens 169	84.71	28.90
Kluyvera genomosp. 3_PO2S7	100	100
K. cryocrescens 169	85.26	29.20
Phytobacter diazotrophicus DSM17806	100	100
P. diazotrophicus UAEU22	99.00	91.30
Kluyvera strain 4105	97.75	80.50
Kluyvera sp. Nf5	98.88	90.20

^aANI and dDDH values are relative to reference strain, in boldface.

^bRelative to reference genome (in boldface) and other average of genome nucleotide identity analyzed.

^cIdentity/high-scoring pair length.

Sequences that were found to be unrelated to the appropriate type or reference strains were submitted to the online TYGS platform (https://tygs.dsmz.de/user_requests/new) to assess their identification at the genus or species level (21). Kluyvera species Nf5 and 4105 (from our collection) matched with Phytobacter diazotrophicus and therefore were not included in further analyses.

Most isolates could be assigned to the accepted species K. intermedia, K. ascorbata, K. cryocrescens, and K. georgiana; the proposed species K. sichuanensis sp. nov.; or the proposed genomospecies 1, 2, and 3. The isolate identified preliminarily as K. cryocrescens 169 did not match with any of the already established species within the genus. With ANI values and dDDH still compatible within the genus, we propose that this isolate is representative of a new genomospecies 5. This correlates with differences in its chromosomal bla_CTX-M genes and with a preliminary analysis using MLST (data not shown) (15).

A phylogenetic tree of all mature CTX-M enzymes showed that the chromosomally encoded enzymes from Kluyvera can be included in at least six clusters (Fig. 1): (i) CTX-M-1 subgroup containing only two variants (CTX-M-3 and CTX-M-37); (ii) CTX-M-2 cluster, including 16 enzymes from K. ascorbata isolates (KLUA-1 to ‐12, CTX-M-76, CTX-M-77, CTX-M-95, and CTX-M-115); (iii) CTX-M-8 subgroup comprising KLUG-1 from K. georgiana; (iv) CTX-M-25 subgroup with CTX-M-78 from K. georgiana 14751, suggested as the origin of the group (14); (v) the “compact” CTX-M-9 subgroup with the KLUY enzymes from not-well-defined Kluyvera species isolates (and CTX-M-213 from the current Kluyvera genomospecies 2 strain KA2) (13); (vi) KLUC cluster, including KLUC-1 from K. cryocrescens NBRC102467 and four other named KLUC β-lactamases, although from plasmid.

FIG 1.

Phylogenetic tree of mature CTX-M enzymes. CTX-M enzymes from Kluyvera isolates are shown in red. The widths of the branches are related to bootstrap values.

The proposed hybrid enzymes derived from CTX-M-1 and CTX-M-9, their point-mutant derivatives, and other CTX-Ms with extensive amino acid changes, such as CTX-M-45 (formerly, Toho-2, having several frameshifts throughout the original sequence) and CTX-M-151 from a Salmonella choleraesuis, could not be assigned to a single group and distort the picture (22–26).

According to the phylogenetic tree of bla_CTX-M/KLU from Kluyvera species isolates, most genes are distributed in two clades (Fig. 2).

FIG 2.

Phylogenetic tree of bla_CTX-M genes from Kluyvera isolates.

The first clade (node A) groups K. cryocrescens isolates with strains NBRC102467 (encoding KLUC-1), 4701, 1919 (encoding KLUC-5), NCTC10483 (also encoding KLUC-1), NCTC12993 (containing a single deletion at 133 bp), and SCW13 as a monophyletic group and the remaining isolates (corresponding to Kluyvera genomospecies 1 strain L2 and Kluyvera genomospecies 5 strain 169, see below) as external branches. True K. cryocrescens isolates in this clade carry a chromosomal bla_KLUC and share high nucleotide identity. New KLUC-1-derived variants (WP_061284984.1) were found in K. cryocrescens 4701 (Ala219Thr mutant) and K. cryocrescens SCW13 (a variant with 93% amino acid identity).

The second clade (node B) includes two main subclades: (i) node C, containing all K. georgiana isolates and Kluyvera genomospecies 3 strains (YDC799 and PO2S7), in two separate subclades, and Kluyvera genomospecies 2 strain KA2 as an external branch; and (ii) node D, including one subclade containing all K. ascorbata isolates as a monophyletic group and a second subclade containing K. sichuanensis 090646 and the assigned 13608 strain (deposited as K. ascorbata), each encoding new CTX-Ms close to CTX-M-95 (94% and 91% amino acid identity, respectively). Within the K. ascorbata group, isolates with new variants (all related to the CTX-M-2 cluster) include K. ascorbata 711, encoding a new enzyme having 99% amino acid identity with CTX-M-95 (Leu80Val and Val249Ile; WP_063860092.1); K. ascorbata LFC, encoding a new variant having 99% amino acid identity with CTX-M-5 (Asn254Ser; WP_032072602.1); and K. ascorbata OT2 and colony 413, encoding a variant having 99% and 97% amino acid identity with KLUA-10 and CTX-M-95, respectively.

Summarizing, the novel bla_CTX-M/KLU genes in the Kluyvera taxons include (i) a novel bla_CTX-M (86% nucleotide identity and 91% amino acid identity with the closer relative CTX-M-95) from K. sichuanensis 090646 and K. sichuanensis 13608 with no plasmidic counterpart; (ii) a new CTX-M-9-like allele, bla_CTX-M-213, from Kluyvera genomospecies 2 strain KA2 (16); (iii) bla_CTX-M-37 (bla_CTX-M-1 subgroup) from Kluyvera genomospecies 5 strain 169 (deposited as K. cryocrescens) (10); (iv) Kluyvera genomospecies 1 strain L2, encoding a new CTX-M closer to KLUC and having ≤80% acid identity with the described KLUC enzymes and the CTX-M-1 cluster; and (v) Kluyvera genomospecies 3 strain PO2S7, encoding a single-mutant variant (Thr109Ala) related to CTX-M-8. YDC799 sequence could not be analyzed due to the presence of multiple stop codons that suggest sequencing errors.

If we accept that plasmid-borne CTX-M enzymes derive from recruitment into mobile elements of previously native (chromosomal) gene counterparts present in different isolates of Kluyvera, it would be expected to have heterogeneity based on the long evolution of these environmental bacteria in different isolated geographic locations. Because different Kluyvera species have been able to at least transiently colonize the human/animal gut, selection under antibiotic pressure of promiscuous bla_CTX-M/KLU genes is probably the driving force for recruitment. Once existing as plasmidic genes, further diversification may be facilitated by the higher copy number, granted by their location in proper and promiscuous platforms, plasmids, and strains.

Incorporating the recent taxonomic proposals by Liu et al. (16) with our findings that depict even more diversity within the genus with a phylogenetic approach to the different bla_CTX-M gene origins, including at least sequences from all the new or better-defined taxons within Kluyvera, we provide some clarification about their origin in Table 2.

TABLE 2.

Proposed CTX-M subfamilies and their corresponding chromosomally encoded progenitors in Kluyvera species

Group denomination	Representative plasmid-borne gene	Best-fitting chromosomal gene (ancestor)	Closest plasmid gene (to ancestor)	Species	GenBank accession no.
CTX-M-1	bla CTX-M-3	bla CTX-M-3	bla CTX-M-3	Kluyvera sp.a	AJ632119
	bla CTX-M-15	bla CTX-M-37	bla CTX-M-37	Kluyvera genomosp. 5	FN813246
CTX-M-2	bla CTX-M-2	blaCTX-M-5/blaKLUA-2	bla CTX-M-5	K. ascorbata	AJ251722
		bla KLUA-1, 3, 4, 12	bla CTX-M-124	K. ascorbata	AJ272538, AJ427461, AJ427462, AJ427469
		bla CTX-M-115	bla CTX-M-115	K. ascorbata	NZ_JMPL00000000.1
		bla CTX-M-76		K. ascorbata	NG_049028
		bla CTX-M-77		K. ascorbata	NG_049029
		bla CTX-M-95		K. ascorbata	NG_049049
CTX-M-8	blaCTX-M-8, blaCTX-M-40, blaCTX-M-41, blaCTX-M-63	Not yet assigned	bla CTX-M-8	Kluyvera genomosp. 3	NZ_CP050321.1
				Kluyvera genomosp. 3	CP022114.1
CTX-M-25	blaCTX-M-25, blaCTX-M-26, blaCTX-M-160, blaCTX-M-100, blaCTX-M-89, blaCTX-M-217, blaCTX-M-39, blaCTX-M-94, blaCTX-M-91, blaCTX-M-205	bla CTX-M-78	bla CTX-M-205	K. georgiana	AM982522
		bla CTX-M-152		K. georgiana	AF501233
		bla CTX-M-185		K. georgiana	KX266838
CTX-M-9	bla CTX-M-9	bla CTX-M-213		Kluyvera genomosp. 2	MH094805
	bla CTX-M-14	bla KLUY-1		Kluyvera sp.a	AY623932
KLUC	bla KLUC-1-5	bla KLUC-1	bla KLUC-1	K. cryocrescens	AY026417
		blaKLUC-5 allele	bla KLUC-5	K. cryocrescens 1919	(SAMN16845439)b
No namec	Unknown		New blaCTX-M gene group	K. sichuanensis	NZ_JABBJF010000000
				K. sichuanensis 13608	(SAMN16845436)b
No namec	Unknown		New blaCTX-M gene group	Kluyvera genomosp. 1	NZ_LGHZ01000000
Outliers
CTX-M-151				S. choleraesuis	NG_048927
CTX-M-64	blaCTX-M-64, blaCTX-M-199, blaCTX-M-123, blaCTX-M-132, blaCTX-M-153, blaCTX-M-174, blaCTX-M-234
CTX-M-137	bla CTX-M-221

^aNo WGS data available.

^bBioSample.

^cWill depend on the number after the first formal deposit in β-lactamase databases.

The CTX-M-1 subfamily may be the most prevalent group present in epidemic plasmids and clones and has been a driver or coselected for multiresistance. The most “popular” enzymes are CTX-M-3 and its derivative CTX-M-15, whose carriers also exhibit resistance to ceftazidime. A 100% identical counterpart to CTX-M-3 was originally proposed as the chromosome origin from an isolate classified (at that time) as K. ascorbata. Unfortunately, this strain could not be recovered for full reexamination. The other chromosomal counterpart described so far is CTX-M-37, well embedded in the cloud represented by this subfamily. In this case, the corresponding isolate was identified as K. cryocrescens 169, but it should be reclassified as a new genomospecies 5 (Kluyvera genomospecies 5 strain 169) (15).

The CTX-M-2 subfamily is well populated on both sides, plasmidic and chromosomal enzymes. In fact, the proposal of KLUA enzymes as ancestors of this family was initially supported within this subfamily (7), and with direct counterparts 100% identical, more ancestors (e.g., bla_CTX-M-76, bla_CTX-M-77, and bla_CTX-M-115) should be considered directly recruited. Presence of a new bla_CTX-M from K. ascorbata colony 413, showing 94% nucleotide identity and borderline ANI values compared with the monophyletic K. ascorbata clade, should be considered with caution because the deposited genome has only 2,832,874 bp, compared to the ∼5,000,000 bp for other Kluyvera genomes.

In the case of the very compact CTX-M-9 subfamily, the origin is most likely a new Kluyvera genomospecies 2, represented by strain KA2 as a close relative of the indeterminate origin of KLUY enzymes.

Within the CTX-M-8 subfamily, only four plasmid-borne variants are included, with maximum 2.40% nucleotide divergence. Kluyvera genomospecies 3 strain PO2S7 is likely to be related to the CTX-M-8 group ancestor.

The CTX-M-25 group includes nine plasmid-borne members; three chromosomal genes from K. georgiana (bla_CTX-M-78, bla_CTX-M-152, and bla_CTX-M-185) group better with plasmidic bla_CTX-M-205 than with other bla_CTX-M-25 alleles, so perhaps with further sequencing, they may consolidate as a new group, especially if a closer chromosomic counterpart is found for this last group.

Kluyvera genomospecies 1 strain L2 harbors a gene for a new CTX-M subfamily cluster, so far not recruited as clinically relevant plasmidic counterparts (16).

Because some of the previously known “plasmidic” enzymes have a strict identity with chromosomal encoded enzymes (and vice versa), the limits between the nomenclature as KLU or CTX-M (chromosomal or plasmidic, respectively) also produce obscure inconsistencies that should be examined to clarify our understanding of this family volution by providing a univocal single (consensus) name for a single (mature) enzyme sequence.

Data availability.

Short reads for all sequenced isolates have been submitted to NCBI under BioProject number PRJNA679803.