Molecular Epidemiology and Genetic Characteristics of Various bla_PER Genes in Shanghai, China

Abstract

We describe the genetic characteristics and possible transmission mechanism of bla_PER in 25 clinical Gram-negative bacilli in Shanghai. bla_PER, including bla_PER-1, bla_PER-3, and bla_PER-4, was located chromosomally or in different plasmids. Tn1213 harboring bla_PER-1 was first identified in two Proteus mirabilis isolates in China. The other bla_PER variants were preceded by an ISCR1 element inside the complex class 1 integron associated with IS26, Tn21, Tn1696, and a miniature inverted-repeat transposable element.

TEXT

PER-encoding genes have been predominantly detected in nonfermenting Gram-negative bacilli, such as Pseudomonas aeruginosa (1), Acinetobacter baumannii (2), and Alcaligenes faecalis (3), and in clinical Enterobacteriaceae isolates, including Salmonella enterica serotype Typhimurium (4), Klebsiella pneumonia (5), Enterobacter cloacae (6), Providencia stuartii (7), and Proteus mirabilis (8). PER-1, an extended-spectrum β-lactamase (ESBL) that was first identified in the P. aeruginosa strain RNL-1, hydrolyzes some oxyimino-cephalosporins and is inhibited by clavulanic acid and tazobactam (9). In total, eight PER variants have been reported worldwide and classified into two groups based on their similarity to PER-1. The first subgroup (PER-3, PER-4, PER-5, PER-7, and PER-8) is formed by replacing one or four amino acids in PER-1, and the other subgroup (PER-2 and PER-6) has only 86% similarity to PER-1 (6). Although PER-1 is prevalent in Europe and Asia (10–12), PER-2 has only been reported in South America, and PER-3, PER-6, and PER-7 have only been found sporadically (13–16). Diverse mechanisms are responsible for the dissemination and acquisition of bla_PER. Previous studies reported that bla_PER-1 was a part of the composite transposon Tn1213 located in a chromosome or carried by plasmids (17). Recently, the ISCR1 element inside a sul1-type integron structure was found to play a role in the dissemination of bla_PER-1, bla_PER-3, and bla_PER-7 (14, 16, 18).

In total, 458 consecutive nonduplicated Gram-negative bacilli resistant to ceftazidime and cefotaxime were isolated from Ruijin Hospital, Shanghai, China, between June 2011 and January 2014 and identified using a Vitek 2 Compact system (bioMérieux, Marcy L'Etoile, France). Among them, 25 (5.5%, 25/458) Gram-negative bacillus isolates (8 P. mirabilis, 6 P. aeruginosa, 6 A. baumannii, 1 Aeromonas hydrophila, 1 Acinetobacter lwoffii, 1 Providencia rettgeri, 1 Morganella morganii, and 1 E. cloacae) were identified as PER-type β-lactamase producers by PCR with specific primers (16). The sequencing analysis revealed that 20 (80%) harbored bla_PER-1, 1 (4%) harbored bla_PER-3, and 4 (16%) harbored bla_PER-4 (Table 1). bla_PER-1 was previously detected in China in A. baumannii, P. aeruginosa, Acinetobacter johnsonii, Vibrio cholerae, and Vibrio parahaemolyticus isolates (11, 18–20). bla_PER-3 was previously detected in Southern Taiwan in 2 Aeromonas caviae blood isolates (21). bla_PER-4 was first detected in China and was identical to that identified in Bulgaria in Proteus vulgaris (GenBank accession no. EU748544).

TABLE 1.

Genetic relatedness and genetic location of bla_PER variants identified in clinical Gram-negative bacilli isolated in Shanghai, China

Identification no.	Bacterium	Date of isolation (mo/yr)	Departmenta	PER identified by sequencing	PFGEb	Locationc	Plasmid size (kb)
RJ38	P. mirabilis	12/2012	Pneumology	PER-1	ApaI-A1	Chromosome
RJ 619	P. mirabilis	11/2011	Nephrology	PER-1	ApaI-A2	Chromosome
RJ 5	P. mirabilis	6/2012	Transplantation	PER-4	ApaI-A3	Chromosome
RJ 8	P. mirabilis	8/2012	Transplantation	PER-4	ApaI-A3	Chromosome
RJ 679	P. mirabilis	11/2011	Nephrology	PER-1	ApaI-A4	Chromosome
RJ 48	P. mirabilis	2/2012	Emergency	PER-1	ApaI-A5	Chromosome
RJ 727	P. mirabilis	7/2011	Emergency	PER-1	ApaI-A6	Chromosome
RJ 635	P. mirabilis	11/2011	Emergency	PER-1	ApaI-A7	Plasmid	∼175
RJ 242	P. aeruginosa	2/2012	Neurology	PER-4	SpeI-B1	Unknown
RJ 246	P. aeruginosa	6/2013	SICU	PER-1	SpeI-B2	Plasmid	∼400
RJ 248	P. aeruginosa	12/2012	RICU	PER-1	SpeI-B3a	Plasmid	∼80
RJ 249	P. aeruginosa	10/2012	RICU	PER-1	SpeI-B3b	Plasmid	∼80
RJ 250	P. aeruginosa	12/2012	RICU	PER-1	SpeI-B3b	Plasmid	∼80
RJ 252	P. aeruginosa	8/2013	SICU	PER-1	SpeI-B4	Plasmid	∼440
RJ 216	A. baumannii	12/2013	Burn surgery	PER-1	ApaI-C1	Unknown
RJ 232	A. baumannii	12/2013	Burn surgery	PER-1	ApaI-C1	Unknown
RJ 278	A. baumannii	1/2014	Burn surgery	PER-1	ApaI-C1	Unknown
RJ 305	A. baumannii	1/2014	Burn surgery	PER-1	ApaI-C1	Unknown
RJ 451	A. baumannii	8/2011	Nephrology	PER-1	ApaI-C2	Unknown
RJ 458	A. baumannii	6/2012	Burn surgery	PER-1	ApaI-C3	Unknown
RJ 604	A. hydrophila	10/2013	Emergency	PER-3	ND	Plasmid	∼173
RJ673	A. lwoffii	7/2012	Thoracic surgery	PER-1	ND	Plasmid	∼260
RJ 746	P. rettgeri	6/2011	Emergency	PER-1	ND	Plasmid	∼180
RJ 716	M. morganii	6/2012	Emergency	PER-1	ND	Unknown
RJ 443	E. cloacae	3/2013	SICU	PER-4	ND	Plasmid	∼305

^aSICU, surgical intensive care unit; RICU, respiratory intensive care unit.

^bPulsed-field gel electrophoresis (PFGE) profiles named by the corresponding restriction endonuclease, a letter, and a consecutive number. ND, not done.

^cUnknown, bla_PER gene neither located in a chromosome nor carried by a plasmid, but hybridized with ApaI-, SpeI-, and XbaI-restricted fragments.

Pulsed-field gel electrophoresis (PFGE) typing of isolates was then performed using CHEF Mapper XA system (Bio-Rad, Hercules, CA, USA) for 18 h at 14°C. P. mirabilis and A. baumannii were digested with ApaI, P. aeruginosa with SpeI (22, 23), and Salmonella enterica serotype Braenderup H9812 with XbaI, where the latter was used as a molecular size marker. As a result, eight bla_PER-harboring P. mirabilis isolates showed seven different PFGE profiles, six bla_PER-harboring P. aeruginosa showed four different PFGE profiles (three were considered subtypes of the same profile), and six bla_PER-harboring A. baumannii isolates showed three different PFGE profiles (Fig. 1; Table 1).

FIG 1.

Pulsed-field gel electrophoresis (PFGE) profiles of Proteus mirabilis isolates digested with ApaI (A), Pseudomonas aeruginosa isolates digested with SpeI (B), and Acinetobacter baumannii isolates digested with ApaI (C). M, Salmonella enterica serotype Braenderup H9812 digested with XbaI and used as a molecular size marker.

Southern hybridization of S1 digests with bla_PER-specific probes was positive for 10 isolates (1 P. mirabilis, 5 P. aeruginosa, 1 A. hydrophila, 1 A. lwoffii, 1 P. rettgeri, and 1 E. cloacae) and negative for the remaining 15 isolates, suggesting that bla_PER was located chromosomally. To confirm these results, I-CeuI PFGE profiles were hybridized with bla_PER-specific probes and subsequently with 16S rRNA gene probes. The hybridization results corroborated that bla_PER in 7 P. mirabilis isolates was located chromosomally. Eight isolates, including 1 P. aeruginosa, 6 A. baumannii, and 1 M. morganii, did not hybridize with S1 or I-CeuI-restricted fragments; they only hybridized with ApaI-, SpeI-, and XbaI-restricted fragments, respectively (Table 1). The conjugational transfer potential of ceftazidime resistance using Escherichia coli J53Az^r (sodium azide resistant) as a recipient demonstrated that only P. mirabilis RJ635 transferred the bla_PER-harboring plasmid to E. coli J53Az^r. In this study, bla_PER was either located in a chromosome or carried by different plasmids (80 to 440 kb) (Table 1). Only one bla_PER-carrying plasmid was transferred by conjugation, suggesting that conjugative plasmids might not be the main channel of gene transmission.

Differences in the genetic environment of bla_PER were also observed among the 25 isolates by overlapping PCR with sequence-specific primers (Table 2). The PCR products were purified and sequenced. The nucleotide sequences were compared using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). P. mirabilis RJ38 and RJ619 (Fig. 2, structure A) harbored a bla_PER-1-gst-like structure surrounded by ISPa12 and ISPa13 of Tn1213 similar to that in P. aeruginosa RNL-1 (17). This structure has been detected in different pathogens worldwide and appears to be a common mechanism mediating the mobilization of bla_PER-1 (5, 24); however, this is the first report on its detection in China.

TABLE 2.

Sequence-specific primers used in this study to determine the genetic context of bla_PER in clinical Gram-negative bacilli isolated in Shanghai, China

Target gene or region	Primer name	Primer sequence (5′ to 3′)	Reference or source
blaPER	PER-F	CCTGACGATCTGGAACCTTT	16
	PER-R	GCAACCTGCGCAATGATAGC	16
Intl1	Intl1	ACGAACCCAGTGGACATAA	This work
Upper blaPER	PER-UP	CGTCTCCCTGATACGCTTT	This work
qacEΔ1	qacEΔ1-R	TGCGGATGTTGCGATTACTT	This work
Lower blaPER	PER-LOW	CGTATCAGGGAGACGAGTTTAGT	This work
sul1	sul1-R	TCGCTGGACCCAGATCCTTTA	This work
tnpR of Tn1696	P4	GCCCTTCTTTGACGAACTCCA	18
IS26	IS26-F	CCTCCCGTCGTAACAGCAAA	This work
	IS26-R	GAAGGGTTACGCCAGTACCAG	This work
MITE	UPMITE-R1	TTGTTTGGGATTTGGTCCTC	19
catB3	catB3-R	GGGAAAGATGATGCCCAGTC	This work
dfrA17	dfrA17-R	TGGAAGAACACCCATAGAGT	This work
dfrA12	dfrA12-R	CAAAGCGATAGCGTGCGACA	This work
aacC1	aacC1-R	GGCTGATGTTGGGAGTAGGTG	This work
aacA4	aacA4-R	GGTTTCTTCTTCCCACCATCC	This work
blaPSE-1	PSE-1-R	TGCGTTCGGTCAAGGTTCTG	This work
blaIMP8	IMP8-R	TGGAAGGCGAGCATCGTTTG	This work
aadA6	aadA6-R	CCAAGGGACAACATCACCAT	This work
aadA13	aadA13-R	CCAAAGCCACGCTATGTTCT	This work
blaPER	P2	CTCGTCTCCCTGATACGCTTTC	18
abct	P7	CCACCACATACACCATCACATCC	18

FIG 2.

Genetic environment of bla_PER in clinical Gram-negative bacilli isolated in Shanghai, China. Structure A in Proteus mirabilis RJ38 and RJ619; structure B in Acinetobacter baumannii RJ216, RJ232, RJ278, and RJ305; structure C in P. mirabilis RJ48, RJ727, and RJ635; structure D in Aeromonas hydrophila RJ604; structure E in Morganella morganii RJ716 and Providencia rettgeri RJ746; structure F in Enterobacter cloacae RJ443; structure G in Pseudomonas aeruginosa RJ248, RJ249, and RJ250; structure H in P. mirabilis RJ5 and RJ8; structure I in P. mirabilis RJ679; structure J in Acinetobacter lwoffii RJ673; structure K in P. aeruginosa RJ242; structure L in P. aeruginosa RJ246 and RJ252; structure M in A. baumannii RJ451; structure N in A. baumannii RJ458; structure O in Vibrio cholerae RJ354 (GenBank accession no. KP076293); structure P in Vibrio parahaemolyticus V36 (GenBank accession no. KP688397); structure Q in Acinetobacter johnsonii XBB1 (GenBank accession no. KF017283); and structure R in Aeromonas punctata 159 (GenBank accession no. GQ891757).

Most bla_PER variants were preceded by an ISCR1 element and had genetic environments similar to those previously reported in China (18–20). The ISCR1 element is an unusual insertion sequence belonging to the IS91 family that probably mobilizes adjacent DNA sequences (25). Partridge and Hall confirmed that a small circular molecule containing ISCR1-dfrA10-sul1 could be transposed into a plasmid within the 3′-conserved sequence (CS) of a cloned class 1 integron in vitro (26). Li et al. first identified the circular molecule with ISCR1-bla_NDM-1-3′-CS in a clinical E. coli strain by Southern hybridization (27), and Li et al. and Wu et al. reported the existence of ISCR1-bla_PER-1-3′-CS in V. cholerae and V. parahaemolyticus using PCR with inverse primers (18, 20). In this study, a free circular molecule bearing the ISCR1-bla_PER structure was identified in all ISCR1-related isolates by PCR with a pair of inverse primers (Table 2). All amplicons were purified and sequenced to elucidate the structure of bla_PER-ISCR1-qacEΔ1/sul1-(3′-CS2)-abct, the same results as our previous report (18). Therefore, ISCR1-bla_PER-1-3′-CS might actually mediate formation of the complex class 1 integron. The universal existence of a circular intermediate suggests that it is a powerful genetic vehicle during the dissemination of bla_PER.

In 18 nonrepetitive isolates, we identified 13 complex class 1 integrons with different arrays of gene cassettes (Fig. 2, structures B to N). Previous studies in China also observed variations in gene cassettes upstream of ISCR1-bla_PER (Fig. 2, structures O to R). However, the region downstream of the ISCR1 element displayed significant similarity in all isolates. The differences on the right-hand side of ISCR1 can be partly attributed to the loss and acquisition of the ISCR1-bla_PER-1-3′-CS structure.

Recent studies have shown that the complex class 1 integron harboring ISCR1-bla_PER-1 is preceded by a miniature inverted-repeat transposable element (MITE) in A. johnsonii XBB1, Tn1696 in non-O1, non-O139 V. cholerae RJ354, and IS26 in V. parahaemolyticus V36 (18–20). These three mobile elements were also detected in our study; 17 isolates were associated with IS26, which can be freely inserted into different sites of Tn21 and intI1, 1 isolate with MITE, and 1 isolate with Tn1696. In particular, P. mirabilis RJ48, RJ727, and RJ635, in which bla_PER was either located in a chromosome or carried by plasmids of different sizes, had identical complex class 1 integrons preceded by IS26 inserted into an incomplete Tn21 (Fig. 2, structure C). Moreover, M. morganii RJ716 and P. rettgeri RJ746 had the same complex class 1 integrons mediated by IS26 and tnpM-Tn21 (Fig. 2, structure E), and P. aeruginosa RJ246 and RJ252, which had different PFGE profiles (bla_PER carried by different plasmids), had identical complex class 1 integrons with unknown mobile segments (Fig. 2, structure L). These results suggest that diverse mobile genetic elements (IS26, Tn21, MITE, and Tn1696) contribute to the horizontal dissemination of bla_PER through the transfer of the complete complex class 1 integrons.

In conclusion, bla_PER genes are widespread in Gram-negative bacilli in China. Our results indicate that ISCR1 plays a major role in the mobilization of bla_PER among clinical bacterial pathogens, forming a circular molecule. Moreover, diverse and complex mobile genetic elements, such as IS26, Tn21, MITE, and Tn1696, may be involved in the horizontal transfer of bla_PER. Additionally, this is the first report of bla_PER mediated by Tn1213 in China.

Nucleotide sequence accession numbers.

Partial nucleotide sequences of all isolates were deposited in the GenBank database under the accession numbers KU133335 to KU133347.