Comparative genomic analyses of Polymyxin-resistant Enterobacteriaceae strains from China.

BACKGROUND: Mobile colistin resistance like gene (mcr-like gene) is a new type of polymyxin resistance gene that can be horizontally transferred in the Enterobacteriaceae. This has brought great challenges to the treatment of multidrug-resistant Escherichia coli and K. pneumoniae.
RESULTS: K. pneumoniae 16BU137 and E. coli 17MR471 were isolated from the bus and subway handrails in Guangzhou, China. K. pneumoniae 19PDR22 and KP20191015 were isolated from patients with urinary tract infection and severe pneumonia in Anhui, China. Sequence analysis indicated that the mcr-1.1 gene was present on the chromosome of E. coli 17MR471, and the gene was in the gene cassette containing pap2 and two copies of ISApl1.The mcr-1.1 was found in the putative IncX4 type plasmid p16BU137_mcr-1.1 of K. pneumoniae 16BU137, but ISApl1 was not found in its flanking sequence. Mcr-8 variants were found in the putative IncFIB/ IncFII plasmid pKP20191015_mcr-8 of K. pneumoniae KP20191015 and flanked by ISEcl1 and ISKpn26.
CONCLUSION: This study provides timely information on Enterobacteriaceae bacteria carrying mcr-like genes, and provides a reference for studying the spread of mcr-1 in China and globally.

Introduction

Polymyxin is a cyclic lipopeptide antibiotic discovered by Ainsworth et al. [1] in the 1940s. In 1959 [2], polymyxin B and colistin (polymyxin E) were introduced into clinical practice and used to treat infections caused by gram-negative bacteria. Due to the strong nephrotoxicity and neurotoxicity, and the popularity of more “safe” antibiotics such as beta-lactam antibiotics, polymyxins had not been used in clinical treatments in the following decades. In the past two decades, the outbreak of multidrug resistant (MDR) gram-negative bacteria and the lack of new antibiotics have caused polymyxins to return to clinical application as the last line of defense against gram-negative bacteria [3].

The resistance mechanisms of bacteria to polymyxins are mainly divided into two categories, two-component system [4, 5] and hyperproduction of CPS capsular polysaccharide (CPS) [6]. The two-component system mainly regulates polymyxin resistance by PhoPQ and PmrAB in Enterobacteriaceae, such as Pseudomonas aeruginosa and Salmonella enterica server Typhimurium. PhoQ can phosphorylate and activate PhoP in the presence of polymyxin. PhoP can increase the positive charge of the outer membrane of the bacteria and the resistance to polymyxins by activating the pmrHFIJKLM operon, causing lipid A to be modified by 4-amino-4-arabinose. Hyperproduction of CPS generally occurs in K. pneumoniae. Some K. pneumoniae strains can reduce the interaction between polymyxin and bacterial surface by synthesizing large amounts of CPS, which leading to the development of polymyxin resistance. Efflux pumps of some Gram-negative bacteria (such as AcrAB [7] and KpnEF [8] of K. pneumoniae) can participate in the resistance of bacteria to polymyxins, but the molecular mechanism is not yet clear. Although bacteria have evolved multiple polymyxin-resistance mechanisms, these mechanisms often require sacrificing their own development and are difficult to disseminate horizontally between strains. These factors limit the spread of these resistant genes among strains. However, in 2015, China reported a new colistin resistance gene, mcr-1, carried by E. coli in the intestine of edible pigs, can be transferred horizontally in Enterobacteriaceae [9]. According to statistics before 2016, mcr-1 positive strains have been reported in more than 40 countries [10], spreading across 7 continents, and may be further expanded. Several reports have shown that many drug-resistant genes, such as New Delhi β-lactamase (NDM) and other extended spectrum β-lactamase genes (ESBLs), were frequently found in the strains carrying mcr-1 [11, 12]. The emergence of mcr-1 not only subverted our understanding of polymyxin resistance genes, but also greatly increased the difficulty of treating MDR pathogenic microorganisms.

MCR-1 is a phosphoethanolamine (PEA) transferase with a 5-fold hydrophobic transmembrane helix located in the periplasmic domain and can reduce the net negative charge of the outer membrane of the bacteria by modifying PEA on the negatively charged lipid A on the lipopolysaccharide (LPS) of the bacteria [13]. The modification reduces the interaction of polymyxin on the outer membrane of bacteria, which in turn produces resistance to polymyxin [14]. Generally, mcr-1 forms a complex transposon Tn 6330 with the surrounding transposon sequence ISApl1 [15]. The complex transposon consists of a sequence of about 2600 bp containing mcr-1 (1626 bp), a PAP2 superfamily protein encoding gene (765 bp), and ISApl1 transposon insertions on both sides [16]. ISApl1 belongs to the IS30 family and therefore has similar functions and activities to IS30 members [17]. It is flanked by 27 bp inverted repeats (referred to as IRL and IRR) and contains a 927 bp open reading frame (orf). The ISApl1 transposon will self-cleave to form a circular sequence intermediate (ISApl1)2-mcr-1-pap2 [18, 19] if the ISApl1 transposon exists around the mcr-1 gene. The circular intermediate contains 2 bp of host flanking DNA between adjacent ISApl1 transposon ends and generates 2 bp of target site duplications (TSDs) after integration [20]. When the mcr-1 circular intermediate is integrated into the plasmid or genome of another strain, there is a probability that the ISApl1 transposon sequence will be lost. Loss of ISApl1 stabilizes mcr-1 in the plasmid or genome, which is conducive to the widespread spread of mcr-1.

Epidemiological studies have found that mcr-1 can be horizontally transferred in more than a dozen Enterobacteriaceae, mainly including E. coli, K. pneumoniae, Salmonella spp. [21], Enterobacter aerogenes [22], P. aeruginosa [23], Proteus putida [24], Enterobacter cloacae [22], Cronobacter sakazakii [25], Shigella sonnei [26], Kluyvera ascorbate [27], Raoultella ornithinolytica [28], Achromobacter spp [23] and Citrobacter spp [29]. These bacteria are mainly transmitted in nature through soil, water, food chains and animal migration [30, 31], and further lead to the global spread of mcr-1. The whole genome sequencing results of mcr-1 positive strains showed that the mcr-1-bearing plasmids were mainly IncI2, IncX4, IncHI2 [32], IncP [33], IncHI1 [34], IncFI, IncFII [35], IncFIB [36], IncK [37], IncY [38], IncN [31], F18:A–:B+ [39]. Among them, IncI2, IncX4 and IncHI2 are the main replicons, and are all conjugative transfer plasmids. These carried plasmids can be stably present in the recipient bacteria even in the absence of polymyxin.

Since mcr-1 was discovered, not only twenty-five genetic variants of the mcr-1 gene (such as mcr-1.1, mcr-1.2, etc.) were reported all over the world [40, 41], but also a variety of mcr-like genes were discovered, which were named mcr-1, mcr-2 [42], mcr-3 [43], mcr-4 [44], mcr-5 [45], mcr-6 [46], mcr-7 [47], mcr-8 [48], mcr-9 [49] and mcr-10 [50]. Among them, mcr-2 and mcr-3 were found in E. coli. Mcr-4, mcr-5 and mcr-9 were found in S. enterica subsp. The mcr-7 and mcr-8 were found in K. pneumoniae. The mcr-6 was found in Moraxella spp. These proteins encoded by these mcr-like genes have different amino acid sequence identity with MCR-1. MCR-6 has the highest amino acid sequence similarity to MCR-1 (82.7%), while MCR-4 has the lowest amino acid sequence similarity to MCR-1 (32.1%), so their sources are not the same. Among them, MCR-1 and MCR-2 are similar in structure, and there are PAP2 family protein coding genes downstream of the coding genes, and the transposition element located near mcr-2 is IS1595 instead of ISApl1. The structures of MCR-3, MCR-4 and MCR-9 are similar. In addition, Teo et al. [51], showed that the coexistence of some mcr-like genes did not significantly improve the polymyxin resistance of clinical Enterobacteriaceae strains.

In the current study, we performed a third-generation genome sequencing analysis of three strains of polymyxin B resistant K. pneumoniae (16BU137, KP20191015 and 19PDR22) and one strain of polymyxin B resistant E. coli (17MR471) from patients and environment. Then we combined with the phenotypes of related experiments to explain the resistance mechanism of mcr-like genes.

Results

Four multidrug-resistant strains all showed colistin resistance

We obtained four MDR strains resistant to polymyxin B, including three strains of K. pneumoniae (16BU137, KP20191015 and 19PDR22) and one strain of E. coli (17MR471). According to the whole-genome three-generation sequencing results, E. coli 17MR471 and K. pneumoniae 16BU137 carried the mcr-1.1 genes, and K. pneumoniae KP20191015 carried the mcr-8.2 gene (Fig. 1A). To determine the phenotypes of these four strains, we performed the determination of MIC value (Table 1). Based on polymyxin B resistance criteria (USCAST, MICs, ≥4 μg/ml) [53], these strains were identified as polymyxin B resistant strains. The MIC values of the four strains were 4 μg/ml (E. coli 17MR471), 8 μg/ml (K. pneumoniae 16BU137), 32 μg/ml (K. pneumoniae KP20191015) and 64 μg/ml (K. pneumoniae 19PDR22). Among these strains, K. pneumoniae 19PDR22 has the highest MIC value and E. coli 17MR471 has the lowest MIC value.

Fig. 1

The genetic environment of mcr-like genes and mgrB. A The genome of E. coli 17MR471 contains the mcr-1.1 gene. The IncX4 type plasmid p16BU137_mcr-1.1 of K. pneumoniae 16BU137 contains the mcr-1.1 gene. The IncFIB/ IncFII type plasmid pKP20191015_mcr-8 of K. pneumoniae KP20191015 contains the mcr-8 variant. No mcr-like gene was detected in K. pneumoniae 19PDR22. B The upstream sequence of mgrB in 19PDR22 was inserted by IS903. The arrow box indicate the target site for insertion of IS903

Table 1

Strains used in this study

Strain	Sourcea	MIC of polymyxin B (μg/ml)	MLST	Datab	Regionc
Strains
16BU137	Bus handrail	8	37	2016.12.17	Guangdong, China
17MR471	Subway handrail	4	1437	2017.10.28	Guangdong, China
KP20191015	Sputum	32	340	2019.08.02	Anhui, China
19PDR22	Urine	64	11	2019.09.16	Anhui, China
J53 [52]	Laboratory	0.5	10	2018.05.24	Kyoto, Japan
J53-16BU137	Laboratory	4	10	2021.01.02	Anhui, China

Source, source of isolates

Data, date of isolate collection

cRegion, geographic location of isolate collection

Transferability of mcr-1- and mcr-8-carring plasmids

Transconjugants conjugate of J53 and 16BU137 is called J53-16BU137, and transconjugants of J53 and KP20191015 is called J53-KP20191015. Through the drug susceptibility test, we found that neither 16BU137 nor KP20191015 can grow on MH plates containing 100 mg/L of sodium azide. PCR detection of mcr-like gene and K. pneumoniae-specific gene was performed on all transconjugants. Using wzi gene as a K. pneumoniae-specific gene (Table 2). PCR product of J53 showed no bands in agarose gel electrophoresis, showing negative, and other K. pneumoniae showed positive. The PCR result of mcr-1 of J53-16BU137 was positive, and the PCR result of wzi was negative, indicating that the mcr-1 plasmid carried by 16BU137 was successfully transferred to the J53, and the conjugation efficiency was 1 × 10− 4 per donor cell. The MIC of J53 for polymyxin B is 0.5 μg/ml, and the MIC of J53-16BU137 for polymyxin B is 4 μg/ml. It is proved that J53 is transformed from a polymyxin-sensitive strain into a polymyxin-resistant strain after receiving p16BU137_mcr-1.1. Although we also performed the same conjugation experiment on KP20191015, we did not observe the transfer of the plasmid carrying mcr-8 from KP20191015 to J53, which may indicate that the mcr-8 plasmid carried by KP20191015 is difficult to transfer between different strains.

Table 2

Sequences of primers used in this study

Primer	Oligonucleotide (5′-3′)a	Application
mcr-1 -F	GTCAGTCCGTTTGTTCTTG	Detection
mcr-1 -R	GGTGACATCAAACAGCTT	Detection
mcr-8 -F	CAACATAGCACTTTGGCA	Detection
mcr-8 -R	GGAAGACAGTGGTGTGTG	Detection
wzi-F	ATGATAAAAATTGCGCGCAT	Detection
wzi-R	GCGTGATCCGTTGCTGATCC	Detection

Genomic profiles of four colistin-resistant isolates

According to third-generation whole genome sequencing, the complete genome sequence of E. coli 17MR471 is 4,765,524 bp, containing 4433 CDS and 87 tRNA; the genome of K. pneumoniae 16BU137 is 5,269,011 bp, containing 4863 CDS and 86 tRNA; the genome of K. pneumoniae KP20191015 is 5,409,809 bp, containing 5110 CDS and 88 tRNA; the genome of K. pneumoniae 19PDR22 is 5,396,045 bp, containing 5077 CDS and 87 tRNA (Fig. 2). E. coli 17MR471 belongs to ST1437, harboring colistin resistance gene mcr-1.1 and other seven ARGs. K. pneumoniae 16BU137 belongs to ST37, harboring mcr-1.1 and other 25 ARGs K. pneumoniae KP20191015 belongs to ST340, harboring a mcr-8 variant (1698 bp, 99.71% nucleotide identity to mcr-8) and other 28 ARGs. K. pneumoniae 19PDR22 belongs to ST11, while lacked known plasmid-mediated colistin resistance gene.

Fig. 2

Circular chromosome map of K. pneumoniae 16BU137, K. pneumoniae KP20191015, and K. pneumoniae 19PDR22. 16BU137 (accession no. CP051161), KP20191015 (accession no. CP051160), 19PDR22 (accession no. CP051159). The map was drawn using BLAST Ring Image Generator (BRIG) (http://sourceforge.net/projects/brig/)

Molecular epidemiological features of mcr-positive E. coli and K. pneumoniae isolates

To better understand the genetic background of these colistin-resistant strains, we collected all E. coli isolates from Guangdong, China and K. pneumoniae isolates from Anhui and Guangdong, China in the NCBI database, and conducted a phylogenetic analysis on them (Fig. 3). The MLST type of E. coli in Guangdong shows diversified characteristics. The MLST type of K. pneumoniae in Anhui and Guangdong is more concentrated, most of which are ST11. Among the isolates we obtained, except for KP20191015, none of the other isolates formed an independent branch. 17MR471 formed a branch with a ST6335 E. coli isolate GDA49. 16BU137 formed a branch with K. pneumoniae P10 and P12 isolates of MLST type ST4298 from Guangdong. 19PDR22 was clustered with ST11 type K. pneumoniae isolates. It is worth noting that KP20191015 formed a branch on its own. The K. pneumoniae isolates distributed in Anhui and Guangdong were intertwined in the phylogenetic tree, which seems to indicate that the K. pneumoniae in China has spread and needs to be controlled immediately.

Fig. 3

Phylogenetic analysis of E. coli and K. pneumoniae isolates. A Phylogenetic analysis of E. coli isolates in Guangdong, China. Isolates obtained in this study are highlighted in red. B Phylogenetic analysis of K. pneumoniae isolates in Anhui and Guangdong, China. The analysis was performed using Parsnp [54] and iTOLv4 [55]

Virulence factors and colistin-related resistance genes of four isolates

A total of 48 virulence factors were predicted in E. coli 17MR471, 16 virulence factors were predicted in K. pneumoniae 16BU137, 10 virulence factors were predicted in K. pneumoniae KP20191015, 25 virulence factors were predicted in K. pneumoniae 19PDR22 (Table 3 and Table S1). In E. coli 17MR471, K. pneumoniae KP20191015 and K. pneumoniae 19PDR22, the identified virulence factors were all located on the chromosomes. In K. pneumoniae 16BU137, a total of 6 virulence factors located on the plasmid were identified. iucA, iucB, iucC, iutA and cseA are located on IncFIB(K)/IncFII type plasmids, while astA is located on IncQ1/IncFII type plasmid.

Table 3

Virulence factors predicted with VFDB database

Gene	16BU137	17MR471	19PDR22	KP20191015
aslA	–a	+ b	–	–
astA	+	+	–	–
cseA	+	+	–	–
csgB/F/G	–	+	–	–
ecpA/B/C/D/E/R	+	+	+	+
entA/B	+	+	+	+
entC/D/E/F/S	–	+	–	–
espL1	–	+	–	–
espR1	–	+	–	–
espX1	–	+	–	–
espX4	–	+	–	–
fepA	–	+	–	–
fepB	–	+	–	–
fepC	+	+	+	+
fepD	–	+	–	–
fepG	–	+	–	–
fes	–	+	–	–
fimA/B/C/D/E/F/G/H/I	–	+	–	–
fyuA	–	–	+	–
gspD/E/F/G/H/I	–	+	–	–
gspK/L/M	–	+	–	–
irp1/2	–	–	+	–
iucA/B/C	+	–	+	–
iutA	+	–	+	–
ompA	+	+	+	+
rmpA2	–	–	+	–
ybtA/E/P/Q/S/T/U/X	–	–	+	–

–, indicates virulence factor negative

+, indicates virulence factor positive

Colistin resistance related genes and other resistance genes in four isolates are shown in Tables 4 and 5 and Table S2. Colistin resistance related genes in K. pneumoniae KP20191015 are similar to K. pneumoniae 19PDR22. Compared with K. pneumoniae 16BU137, K. pneumoniae KP20191015 and K. pneumoniae 19PDR22 may contain more colistin resistance related genes. Among the three strains, the types of arnD, eptB, mgrB, opgE, pmrA, pmrB, pmrC, and pmrD are the same. Both K. pneumoniae KP20191015 and K. pneumoniae 19PDR22 contain two types of emrA, two types of emrB, and three types of phoP. K. pneumoniae 16BU137 contains one type of emrA, one type of emrB, and two types of phoP. In addition, IS903 inserted on the upstream sequence of mgrB in K. pneumoniae 19PDR22 (Fig. 1B), which may affect the normal expression of mgrB [56].

Table 4

Colistin resistance related gene in four isolates

Gene	16BU137	17MR471	19PDR22	KP20191015
arnD	Uniprot ID: P76472	Uniprot ID: P76472	Uniprot ID: P76472	Uniprot ID: P76472
emrA	Uniprot ID: P27303	Uniprot ID: P27303	Uniprot ID: P0DPR6*2; Uniprot ID: P27303	Uniprot ID: P0DPR6*2; Uniprot ID: P27303
emrB	Uniprot ID: P0AEJ0	Uniprot ID: P0AEJ0	Uniprot ID: P0DPR7; Uniprot ID: P0AEJ0	Uniprot ID: P0DPR7; Uniprot ID: P0AEJ0
eptB	Uniprot ID: P37661	Uniprot ID: P37661	Uniprot ID: P37661	Uniprot ID: P37661
mcr-like genes	mcr-1.1	mcr-1.1	None	mcr-8
mgrB	Uniprot ID: B5XQ45	Uniprot ID: P64512	Uniprot ID: B5XQ45, IS 903 is inserted upstream	Uniprot ID: B5XQ45
opgE	Uniprot ID: P75785	Uniprot ID: P75785*2	Uniprot ID: P75785	Uniprot ID: P75785
phoP	Uniprot ID: P13792; Uniprot ID: D0ZV90	Uniprot ID: P23836	Uniprot ID: P0DM78; Uniprot ID: P13792; Uniprot ID: D0ZV90	Uniprot ID: P0DM78; Uniprot ID: P13792; Uniprot ID: D0ZV90
phoQ	Uniprot ID: P23837	Uniprot ID: P23837	Uniprot ID: P23837	Uniprot ID: P23837
pmrA (basR)	Uniprot ID: P30843	Uniprot ID: P30843	Uniprot ID: P30843	Uniprot ID: P30843
pmrB (basS)	Uniprot ID: P30844	Uniprot ID: P30844	Uniprot ID: P30844	Uniprot ID: P30844
pmrC (eptA)	Uniprot ID: P36555	Uniprot ID: P30845	Uniprot ID: P36555	Uniprot ID: P36555
pmrD	Uniprot ID: P37589	Uniprot ID: P37590	Uniprot ID: P37589	Uniprot ID: P37589

Table 5

Antimicrobial resistance genes predicted with ResFinder-3.2

Gene	Identity(%)	Antibiotic_Resistance	Position
KP20191015
aac(3)-IV	100	Aminoglycoside	Plasmid
aadA1	100	Aminoglycoside	Plasmid
aadA2b	99.87	Aminoglycoside	Plasmid
aph(3″)-Ib	100	Aminoglycoside	Plasmid
aph(3′)-Ia	100	Aminoglycoside	Chromosome/Plasmid
aph(6)-Id	100	Aminoglycoside	Plasmid
armA	100	Aminoglycoside	Plasmid
blaCTX-M-15	100	Beta-lactam	Plasmid
blaDHA-1	100	Beta-lactam	Plasmid
blaSHV-182	99.88	Beta-lactam	Chromosome
blaTEM-1B	100	Beta-lactam	Plasmid
mcr-8	99.71	Colistin	Plasmid
fosA	99.27	Fosfomycin	Plasmid
mph(A)	100	Macrolide	Plasmid
mph(E)	100	Macrolide	Plasmid
msr(E)	100	Macrolide	Plasmid
catA2	96.11	Phenicol	Plasmid
cmlA1	99.92	Phenicol	Plasmid
oqxA	100	Quinolone	Chromosome
oqxB	100	Quinolone	Chromosome
qnrB4	100	Quinolone	Plasmid
sul1	100	Sulphonamide	Plasmid
sul3	100	Sulphonamide	Plasmid
tet(D)	100	Tetracycline	Plasmid
19PDR22
aac(3)-IId	99.88	Aminoglycoside	Chromosome
aadA2b	99.87	Aminoglycoside	Plasmid
aadA5	100	Aminoglycoside	Plasmid
aph(3″)-Ib	100	Aminoglycoside	Plasmid
aph(6)-Id	100	Aminoglycoside	Plasmid
armA	100	Aminoglycoside	Plasmid
rmtB	100	Aminoglycoside	Plasmid
blaCTX-M-65	100	Beta-lactam	Plasmid
blaKPC-2	100	Beta-lactam	Plasmid
blaSHV-12	100	Beta-lactam	Chromosome/Plasmid
blaSHV-182	99.77	Beta-lactam	Chromosome
blaTEM-1A	100	Beta-lactam	Plasmid
blaTEM-1B	100	Beta-lactam	Chromosome/Plasmid
blaTEM-1C	100	Beta-lactam	Plasmid
fosA	99.27	Fosfomycin	Chromosome
mph(A)	100	Macrolide	Chromosome
mph(E)	100	Macrolide	Plasmid
msr(E)	100	Macrolide	Plasmid
sul1	100	Sulphonamide	Plasmid
sul2	100	Sulphonamide	Plasmid
dfrA17	100	Trimethoprim	Plasmid
16BU137
aac(3)-IId	99.88	Aminoglycoside	Plasmid
aac(6′)-Ib-cr	100	Aminoglycoside	Plasmid
aadA16	99.65	Aminoglycoside	Plasmid
aph(3″)-Ib	100	Aminoglycoside	Plasmid
aph(3′)-Ia	99.88	Aminoglycoside	Plasmid
aph(6)-Id	100	Aminoglycoside	Plasmid
blaCTX-M-3	100	Beta-lactam	Plasmid
blaSHV-110	99.77	Beta-lactam	Chromosome
blaTEM-1B	100	Beta-lactam	Plasmid
mcr-1.1	100	Colistin	Plasmid
fosA	99.29	Fosfomycin	Chromosome
mph(A)	100	Macrolide	Plasmid
floR	98.27	Phenicol	Plasmid
aac(6′)-Ib-cr	100	Quinolone	Plasmid
oqxA	100	Quinolone	Chromosome
oqxB	100	Quinolone	Chromosome
qnrB2	99.84	Quinolone	Plasmid
qnrS1	100	Quinolone	Plasmid
ARR-3	100	Quinolone	Plasmid
sul1	100	Rifampicin	Plasmid
sul2	99.88	Sulphonamide	Plasmid
tet(A)	100	Tetracycline	Plasmid
dfrA27	100	Trimethoprim	Plasmid
17MR471
blaTEM-1B	100	Beta-lactam	Plasmid
mcr-1.1	100	Colistin	Chromosome
mdf(A)	99.92	Macrolide	Chromosome
floR	98.19	Phenicol	Plasmid
oqxA	100	Quinolone	Plasmid
oqxB	99.97	Quinolone	Plasmid
tet(B)	100	Tetracycline	Chromosome
tet(M)	96.15	Tetracycline	Plasmid

Location of mcr-1.1 on chromosome and plasmid

The mcr-1.1 gene was found to locate on the chromosome of E. coli 17MR471. Specifically, the mcr-1-pap2 gene cassette which encodes both MCR-1 and a hypothetical protein was flanked by two copies of ISApl1 (1070 bp, IS30 family) upstream and downstream in the same orientation. In K. pneumoniae 16BU137, mcr-1.1 located in an IncX4-type plasmid which named p16BU137_mcr-1.1 (Table S3). This plasmid is 33,309 bp in size and is predicted to encode 41 ORFs for which mcr-1.1 is the only resistance gene. No ISApl1 was found in the flanking sequences of mcr-1.1 in p16BU137_mcr-1.1(Fig. 4). IncX4 is the dominant plasmid type to harbor mcr-1.1 [57]. The mcr-1-bearing IncX4 plasmid was firstly identified in Germany in 2009. Since 2009, the majority of mcr-1 genes have been found on IncX4 plasmids. BLASTn revealed that the genetic context of mcr-1.1 in IncX4 plasmids are diverse. The examples included that mcr-1.1 without flanking ISApl1 (pAF48, KX032520). Also, mcr-1-pap2 could be flanked by ISApl1 upstream (pMCR-11EC-P293, KX555451), downstream (pPY1, KX711708) or both (pC214, KY120363). Plasmids like PN42 (MG557854) and pCDF8 (MF175191) have truncated IS elements in flanking regions of mcr-1. It has been hypothesized that after the loss of ISApl1, mcr-1 is immobilized in the plasmids [18].

Fig. 4

Circular chromosome map of p16BU137_mcr-1.1

A mcr-8 variant was found in an IncFIB/ IncFII plasmid

In K. pneumoniae KP20191015, a mcr-8 variant was found in an 107,661 bp IncFIB/ IncFII plasmid which named pKP20191015_mcr-8. mcr-8 was flanked by a reverse ISEcl1-like element (1336 bp, 99% identity to ISEcl1) upstream. Also, it was flanked by an ISKpn26-like element (1196 bp, 99% similarity to ISKpn26) at the same direction downstream. Consistent with the sequences in the mcr-8-carrying pKP91 (95,983 bp, MG736312) [48], both of which carried dgkA, baeS, and copR close to mcr-8 (Fig. 5). While mcr-8 in pKP91 was flanked by two intact IS903B sequences up- and downstream [48], and significant differences were observed in the remaining plasmid backbone (Fig. 5). BLASTn indicated that pKP20191015_mcr-8 carried novel components that showed limited identity to those known plasmid sequences (coverage < 75%). pKP20191015_mcr-8 is organized similar to that of plasmid pKP1814–2 (187,349 bp, KX839208) (69% coverage, 99.84% identity) identified in K. pneumoniae in China; p002SK2_A (159,714 bp, CP025516) (53% coverage, 99.80% identity) identified in K. pneumoniae in Switzerland; pKP121–2 (134,208 bp, CP031851) (53% coverage, 99.75% identity) identified in K. pneumoniae in China. They all carried plasmid transfer associated tra locus with different combination and the replicon encoding gene repB.

Fig. 5

Schematic presentation of major structural features of pKP20191015_mcr-8 in comparison with the reference plasmids pKP91, pKP1814–2, p002SK2_A, and pKP121–2. pKP20191015_mcr-8 (accession no. MT316509), pKP91 (accession no. MG736312), pKP1814–2 (accession no. KX839208), p002SK2_A (accession no. CP025516), pKP121–2 (accession no. CP031851). Annotation features represented the genes in pKP20191015_mcr-8

Discussion

Polymyxins are cyclic, positively charged peptides, which were first discovered to possess antibiotic properties in the 1940s [58]. Polymyxins can bind to lipid A of lipopolysaccharide (LPS) on the outer membrane of Gram-negative bacteria, and then displace Mg2+ and Ca2+ from cationic binding sites leading to disruption of bacterial membrane integrity [58, 59]. Polymyxins (polymyxin B and colistin) are a last resort treatment against human infections caused by multidrug-resistant (MDR) Gram-negative bacteria [60]. Colistin resistance is often associated with chromosomal point mutations that affect the expression of regulators, which modify lipid A and lead to alterations of LPS [61]. Bacteria can add phosphoethanolamine (PEtN) and 4-amino-4-deoxy-L-arabinose (L-Ara4N) to lipid A via biosynthesis, thereby decreasing the net negative charge of lipid A to reduce its binding affinity to polymyxins [62, 63]. The synthesis and transfer of PEtN and L-Ara4N are mediated by the expression of pmrCAB and arnBCADTEF (also called pmrHFIJKLM) [64] which were regulated by a two-component system (TCS) PmrA/PmrB [65, 66]. Mutations in the genes encoding PmrA/PmrB were shown to contribute to polymyxin resistance [67, 68]. Moreover, another TCS PhoP/PhoQ is known to develop polymyxin resistance via activation of its posttranscriptional activator PmrD to induce expression of the PmrA/PmrB system [69]. Mutations in the genes encoding the PhoP/PhoQ were also associated with colistin resistance [70]. Here, we report four polymyxin-resistant Enterobacteriaceae strains. Among them, K. pneumoniae 16BU137 and E. coli 17MR471 carries mcr-1, and K. pneumoniae KP20191015 carry mcr-8. The mcr-1 in 17MR471 is located on the chromosome, and the surrounding sequence is a typical Tn 6330 structure, (ISApl1)2-mcr-1-pap2. The mcr-1 in 16BU137 lacks upstream ISApl1 and downstream ISApl1, but still retains pap2. ISApl1 may be lost due to its involvement in mcr-1 transposition [18, 19]. However, pap2 always exists downstream of mcr-1, which seems to suggest that pap2 may play an indispensable function for mcr-1. Through further experimental verification, we identified K. pneumoniae 19PDR22 which conferred high MIC of colistin, while no known plasmid-mediated colistin genes was found. We found an IS903B-like element (97% similarity to IS903B) inserted into the upstream sequence of mgrB. This insertion appeared at position − 18 bp of the mgrB, which may lead to the inactivation of mgrB by interrupting its promoter region. The inactivation of mgrB conferred colistin resistance has been reported previously [71]. IS integration has also been reported to induce colistin resistance via transposition into the upstream putative promoter region of mgrB [72]. IS903, a member of IS5 family, is implicated in antibiotic resistance. Insertion sequences of the IS5 family have also been reported to truncate mgrB in Klebsiella oxytoca and yield elevated MICs for colistin [73]. More studies are needed to evaluate the mobilization of these elements from plasmids to the chromosome to disrupt the expression of potential resistance-associated genes. Our data show that the transformation efficiency of p16BU137_mcr-1.1 is higher than that of pKP20191015_mcr-8. This may indicate that the plasmid carrying mcr-1 has a higher transformation efficiency and stronger transmission ability than the plasmid carrying mcr-8. However, the experimental results still have limitations due to the small number of strains in this study. 16BU137 and KP20191015 carries the mcr-like genes, meanwhile carries a variety of ESBL genes, such as blaSHV, blaCTX and blaTEM. The existence of these resistance genes makes MDR enterobacteria a huge threat to public medical and health safety.

Conclusion

We present the complete genome of four polymyxin-resistant strains (including two clinically isolated strains and two environmentally isolated strains, both clinically isolated strains are K. pneumoniae). The high-quality complete genome sequence generated in this study will help to further study the mechanism of polymyxin resistance of K. pneumoniae and the horizontal transfer pathway of mcr-like genes. Although two strains are isolated from the environment, they still have high polymyxin resistance. And the types of virulence factors are basically the same as clinical strains, and still have the risk of infecting humans. These also warns us that the multi-drug resistant K. pneumoniae has spread seriously in China and needs to be controlled as soon as possible.

Methods

Bacterial isolation

The MIC of polymyxin B was tested on the MDR clinical isolates isolated from the inpatients in Affiliated Hospital of Anhui University of Traditional Chinese Medicine and the Anhui Provincial Hospital in 2019, and two polymyxin B resistant isolates were obtained (K. pneumoniae 19PDR22 and K. pneumoniae KP20191015). The environmental isolates of K. pneumoniae 16BU137 and E. coli 17MR471 were obtained from our previous studies [32]. They all carried mcr-1 and were resistant to polymyxin B. Briefly, the environmental samples were collected using sterilized swab with saline, and cultured by broth medium. Then, the cultured samples were plated on the MacConkey agar with colistin (2 μg/mL) and cultured under 37 °C overnight. Subsequently, we randomly selected 5 colonies for each plate which were subject to screen mcr-1 gene by PCR. Only one colony for each sample was included for the subsequent study. K. pneumoniae 19PDR22 was isolated from the urine of patient with urinary tract infection, and K. pneumoniae KP20191015 was isolated from the sputum of patient with severe pneumonia. Sputum and urine were plated on blood agar plates and cultured at 37 °C to isolate bacterial clones. VITEK 2 Compact System (bioMérieux, France) was used to identify positive culture strains.

Determination of minimum inhibitory concentration

K. pneumoniae and E. coli were cultured overnight in LB liquid medium at 37 °C for 220 rpm according to 1:100, and a small amount of liquid medium was streaked on LB plate and incubated overnight in 37 °C constant temperature incubator. Several monoclonal strains were selected to adjust the concentration of bacteria in MH (Mueller-Hinton Broth) medium so that the concentration of bacteria reached OD600 = 0.4 and then diluted 200 times in MH medium [74]. Mix 75 ml of MH medium with different concentrations of polymyxin B and 75 ml of MH medium with diluted bacterial solution and add them to each well of a 96-well plate according to the polymyxin concentration gradient. The final CFU of the well is 5 × 105. Each concentration gradient was divided into three parallel groups and grown at 37 °C and 220 rpm with shaking for 24 and 48 h. The experiment was repeated three times independently.

Plasmid conjugation experiments

E. coli J53 (LacZ–, AzrR, RifR) was used as the recipient, and the mcr-like gene-positive strain (16BU137, KP20191015) was used as the donor. Overnight culture (2 mL) of each donor and recipient bacteria was mixed together at a ratio of donor to recipient of 1:3. The mixture was added to a final volume of 5 mL LB liquid medium, and incubate at 37 °C for 12–18 h. Then spotted the mixture on Muller-Hinton agar plates containing 100 mg/L sodium azide and 2 mg/L polymyxin B as a selective medium for E. coli J53 transconjugants. Detection of mcr-like gene by PCR confirmed the putative transconjugants. Use wzi gene primers, mcr-1 gene primers and mcr-8 gene primers to distinguish the recipient strain (16BU137 and KP20191015) from the donor strain (J53).

Whole-genome sequencing and genotyping

K. pneumoniae and E. coli were cultured overnight in LB medium. Bacterial samples (5000 g 10 min at 4 °C) were collected and frozen at − 80 °C. The genomes of four isolates were performed using a PacBio RS II platform and Illumina HiSeq 4000 platform at the Beijing Genomics Institute (BGI, Shenzhen, China). Four SMRT cells Zero-Mode Waveguide arrays of sequencing, were used by the PacBio platform to generate the subreads set. PacBio subreads (length < 1 kb) were removed. The program Pbdagcon (https://github.com/PacificBiosciences/pbdagcon) was used for self-correction. Draft genomic unitigs, which are uncontested groups of fragments, were assembled using the Celera Assembler against a high quality corrected circular consensus sequence subreads set. To improve the accuracy of the genome sequences, GATK (https://www.broadinstitute.org/gatk/) and SOAP tool packages (SOAP2, SOAPsnp, SOAPindel) were used to make single-base corrections.

De novo hybrid assembly both of short Illumina reads and long PacBio reads was performed using Unicycler v0.4.3 [75]. Complete circular contigs were then corrected using Pilon v1.22 with Illumina reads. For each de novo assembled genome, coding sequences were predicted using Prodigal (v. 2.6) [76] and annotated using the rapid prokaryotic genome annotation tool Prokka [77]. Acquired antimicrobial resistance genes (ARGs) were identified using ABRicate version 0.5 (https://github.com/tseemann/abricate) by aligning genome sequences to the ResFinder database [78]. The virulence factors of the isolates were identified using VFDB database [79]. Insertion sequence (IS) elements were determined with ISFinder (https://www-isfinder.biotoul.fr). In silico multilocus sequence typing (MLST) was performed by MLST 1.8 (https://cge.cbs.dtu.dk/services/MLST/). Plasmid replicon types were detected using PlasmidFinder v1.3 [80].

Phylogenetic analysis

We collected all 87 E. coli strains from Guangdong, China and all 182 K. pneumoniae strains from Guangdong and Anhui, China (182 strains from Guangdong and 70 from Anhui) in the NCBI database (https://www.ncbi.nlm.nih.gov/pathogens/) as of December 2020. HarvestTools kit (Parsnp, Gingr and HarvestTools) was used to perform comparative genomics analysis and phylogenetic analysis of different isolates, Interactive tree of life (iTOL) v5 (http://itol.embl.de/) was used to construct a maximum likelihood phylogenetic tree [54, 55].

Additional file 1.

Additional file 2.

Additional file 3.