Transposition mechanism of ISApl1—the determinant of colistin resistance dissemination

ABSTRACT

Multidrug-resistant Enterobacteriaceae, a prominent family of gram-negative pathogenic bacteria, causes a wide range of severe diseases. Strains carrying the mobile colistin resistance (mcr-1) gene show resistance to polymyxin, the last line of defense against multidrug-resistant gram-negative bacteria. However, the transmission of mcr-1 is not well understood. In this study, genomes of mcr-1-positive strains were obtained from the NCBI database, revealing their widespread distribution in China. We also showed that ISApl1, a crucial factor in mcr-1 transmission, is capable of self-transposition. Moreover, the self-cyclization of ISApl1 is mediated by its own encoded transposase. The electrophoretic mobility shift assay experiment validated that the transposase can bind to the inverted repeats (IRs) on both ends, facilitating the cyclization of ISApl1. Through knockout or shortening of IRs at both ends of ISApl1, we demonstrated that the cyclization of ISApl1 is dependent on the sequences of the IRs at both ends. Simultaneously, altering the ATCG content of the bases at both ends of ISApl1 can impact the excision rate by modifying the binding ability between IRs and ISAPL1. Finally, we showed that heat-unstable nucleoid protein (HU) can inhibit ISApl1 transposition by binding to the IRs and preventing ISAPL1 binding and expression. In conclusion, the regulation of ISApl1-self-circling is predominantly controlled by the inverted repeat (IR) sequence and the HU protein. This molecular mechanism deepens our comprehension of mcr-1 dissemination.

INTRODUCTION

Recently, the misuse of antibiotics has increased multidrug-resistant bacteria, posing significant challenges for clinical treatment and prevention (1). The World Health Organization has prioritized multidrug-resistant Enterobacteriaceae, including Klebsiella pneumoniae and Escherichia coli, as a critical issue (2, 3). Polymyxin, especially against carbapenem-resistant Enterobacteriaceae, is considered the last-line defense for treating multidrug-resistant gram-negative bacteria, Despite its severe nephrotoxicity (4–6). However, bacteria have developed resistance to polymyxin, especially via the mgrB or two-component systems PmrAB, PhoPQ, and CcrAB, which target lipid A (7). In addition, in 2015, the mcr-1 gene was reported, introducing a novel polymyxin resistance mechanism (8). Subsequently, strains positive for mcr-1 and its variants (mcr-2 to mcr-10), including Acinetobacter baumannii, Salmonella enterica, K. pneumoniae, and Escherichia fergusonii, among others, have been detected in more than 30 countries and regions worldwide (9).

Previous studies have shown that ISApl1 assists in the cyclization of mcr-1 from Tn6330 (10, 11). The ISApl1 gene is initially identified in Actinobacillus pleuropneumoniae and consists of a coding region (927 bp) flanked by an inverted repeat left (IRL), inverted repeat right (IRR) sequence (27 bp), and a direct repeat (DR) sequence (2 bp) (12). ISApl1 belongs to the IS30 family and possesses the characteristic DDE domain (Asp, Asp, and Glu) found in the DD (E/D) superfamily of transposable enzymes (13, 14).

E. coli has the most abundant DNA binding protein heat-unstable nucleoid protein (HU), composed of two highly homologous subunits, HUα and HUβ (15). The two subunits can form as homo- or heterodimers because of the differential expression and stability of the two subunits during the growth cycle (16). It has been reported that it can regulate the transposition of Tn10 and bacteriophage Mu (17, 18). Meanwhile, HU induces changes in gene expression by modulating the 3D arrangement of DNA. This includes altering DNA looping in the promoter region, trapping free supercoils, indirectly affecting supercoiling through DNA topoisomerases, and contacting long-range DNA-DNA interactions (19–22).

This study demonstrated the widespread distribution of mcr-1-positive strains in China. The cyclization of ISApl1, a critical factor in regulating the propagation of mcr-1, is mediated by the IRs at both ends and the proteins of ISAPL1 and HU. These results offer insights into the molecular mechanisms of ISApl1 self-excision and the dissemination of mcr-1.

RESULTS

mcr-1-positive strains are distributed throughout China

To study the distribution of mcr-1 in China. We obtained all the E. coli genome sequences submitted to the NCBI database from China. The research data revealed that mcr-1-positive strains were prevalent in various regions of China, particularly in Sichuan, Guangdong, and Shandong (Fig. 1A; supplemental Excel 1). Moreover, there was a greater occurrence of positive strains from environmental sources compared to clinical strains (Fig. 1A; supplemental Excel 1).

Fig 1.

The distribution of multidrug-resistant mcr-1-positive strains in China. (A) Distribution of mcr-1-positive strains in China. E. coli genomes were downloaded from the NCBI database and statistical analysis of the regional distribution of mcr-1-positive strains was performed. A darker red color represents a higher number of strains. Red indicates that the strain was isolated from the environment, and blue indicates that the strain was isolated from a clinical source. (B) Resistance genes carried by mcr-1-positive strains. We used Abricate for the analysis of resistance genes. Genes resistant to the same antibiotic are marked in the same color. The proportion of resistant genes greater than 0.2 was counted. (C) Statistical information on the location of mcr-1 within the genome. There are 477 samples with mcr-1 on chromosomes (marked in blue), 594 samples with mcr-1 on plasmids (marked in red), and 124 samples of mcr-1 with two copies on chromosomes and plasmids (marked in green). (D) Types of plasmids carrying mcr-1. Plasmid types were categorized using a software tool named SankeyMATIC. Plasmids carrying mcr-1 are indicated in blue, while those lacking ISApl1 are labeled in green, and plasmids containing ISApl1 are marked in orange. The thickness of the line corresponds to the number of plasmids.

Previous reports have indicated that mcr-1-positive strains also confer resistance to other types of antibiotics (23). Consequently, we analyzed the resistance spectrum in mcr-1-positive strains and identified multiple antibiotic-resistant genes, including aminoglycoside, tetracycline, chloramphenicol, sulfonamide, quinolone, and β-lactam resistance-related genes (Fig. 1B; supplemental Excel 1). These findings revealed that mcr-1-positive isolates contained multiple drug resistance genes, implying significant challenges in clinical treatment. To gain insight into the widespread dissemination of mcr-1, we analyzed its location sites in the genome. The contigs of mcr-1 were analyzed using the PlasmidFinder to ascertain its plasmid location. The findings revealed that mcr-1 can be present in both the chromosome and plasmid (Fig. 1C), indicating its genomic transferability. The data revealed the presence of mcr-1 in 12 distinct plasmid backgrounds, with IncI2 (47.1%) and IncX4 (30.8%) being the predominant types. Subsequently, we compared the distribution of 12 plasmid types carrying ISApl1. While IncI2 and IncX4 plasmids were less prevalent, the remaining plasmid types (IncHI2A, IncP1, IncHI2, p0111, and IncY) exhibited contrasting proportions (Fig. 1D; supplemental Excel 1). The absence of ISApl1 may account for the stable integration of mcr-1 in the plasmid, leading to a low carrier rate of ISApl1 in these plasmids. In conclusion, the study reveals that mcr-1-positive bacteria primarily exhibit multidrug resistance and are widely distributed across China.

The determinant of mcr-1 transposition element-ISApl1 can circle by itself

To investigate the widespread distribution of mcr-1 in China, we conducted sequence alignments surrounding the mcr-1 gene. The findings revealed that ISApl1 was predominantly present upstream or downstream of the mcr-1 gene on both plasmids and chromosomes, indicating its potential involvement in the transposition of mcr-1 (Fig. 2A). To explore the relationship between ISApl1 and mcr-1 further, we examined the environmentally isolated strain E. coli 17MR471, which is known to harbor Tn6330 (24). The products generated by P1/P2 and P3/P4 primers had lengths of approximately 3 kb and 1.4 kb, respectively, suggesting that Tn6330 was excised from E. coli 17MR471 in the form of ISApl1-mcr-1-pap2 structure (Fig. 2B). Similar findings demonstrated that the formation of a 1.4 kb product was exclusive to Tn6330, observed in both constructed model strain (Top10, recA^-), as well as in the clinical strains (Fig. S1). The findings corroborated previous reports that suggested the ability of Tn6330 to generate ISApl1-mcr-1-pap2 intermediates, thereby facilitating the dissemination of mcr-1 (11). These results confirmed that the mobility of mcr-1 relies on the intact Tn6330 element. Through sequence alignment, we offered an initial demonstration of the evolutionary process of Tn6330 and validated the involvement of ISApl1 excision and cyclization in the degradation of Tn6330 (Fig. 2C). However, examination of the Tn6330 cyclizing product using P1/P2 primer pair revealed the presence of an additional 400 bp product, which was subsequently sequenced and identified as ISApl1 self-cyclization (Fig. 2B). Whole-genome sequencing revealed 12 copies of the ISApl1 genes in the 17MR471. To investigate the property of ISApl1 self-cyclization, we inserted the ISApl1 at various positions in 17MR471 into pUC19 plasmids. The plasmids were subsequently extracted and analyzed using agarose gel electrophoresis. The results indicated the presence of an additional band of approximately 1 kb (mini-plasmid), known as ISApl1 (Fig. 2D). Following PCR amplification with P1/P2 primers, a 2 bp junction spacer, almost identical to the DR, was detected between the IRL and IRR (Fig. 2D). Subsequently, the majority of ISApl1 insertion sites in 17MR471 were found within the AT-rich regions (Fig. 2E). The results demonstrated that ISApl1 can undergo independent self-cyclization during mcr-1 movement, thereby regulating the transmission of mcr-1.

Fig 2.

The transposition of mcr-1 depends on ISApl1. (A) The location of ISApl1 is found to be surrounding mcr-1 in both the chromosome and plasmid. The sequences around the mcr-1 on the plasmid and chromosome were aligned, respectively. (B) The excision pattern of the composite transposon Tn6330 in 17MR471.The ISApl1, mcr-1, and pap2 genes are marked in orange, red, and yellow, respectively. The P1-P4 primer sequences are shown in Table S2. M: Marker. (C) Sequence alignment around mcr-1 in different isolates. mcr-1-positive strains were compared, XH988, 659, and G3 × 16–2. The three strains sequence information from the NCBI database. (D) ISApl1 can transposition by itself. (I) We cloned ISApl1, from different genome positions of 17MR471, into pUC19. Plasmids were extracted to be subjected to agarose gel electrophoresis. L₁-L₁₂ represents pUC19 containing ISApl1. L₁₃ represents the pUC19 wild type, serving as the negative control. (II) PCR was performed using primers P1 and P2, followed by sequencing of the PCR products. L₁-L₁₂ represents pUC19 containing ISApl1, which was formed at different positions of 17MR471. Sequences in red boxes indicate DR, and bases in red font indicate junction spacer. DR: direct repeat; IRL: inverted repeat left; IRR: inverted repeat right. Red font: junction spacer. (E) Insertion site of ISApl1. Alignment of the ISApl1 insertion sequence information in 17MR471 was performed using the online tool WebLogo.

ISAPL1 transposase participates in ISApl1 cyclization

The results presented above indicate that the excision of ISApl1 played a role in the abortion of Tn6330, leading to the transposition of mcr-1. However, the mechanisms underlying the self-cyclization of ISApl1 remain unclear. Our findings demonstrated that ISApl1 could not transpose when disrupted by the kanamycin resistance gene (Fig. 3A). Conversely, when functional ISApl1 was complemented, the transposition of ISApl1 was detected (Fig. 3A). We conducted a preliminary exploration of the evolution of ISAPL1 through phylogenetic tree analysis, which indicated its close relation to the evolution of ISEnfa364 and ISSlu1, both of which have received limited study (Fig. S2A). In addition, sequence alignment of the entire IS30 family demonstrated high homology of ISAPL1 with other family members (Fig. S2B). To predict the function of ISApl1, sequence alignment was conducted between ISApl1, ISEnfa364, ISSlu1, and IS30. The results show that four protein sequences in the DDE (Asp, Asp, Glu)-domain were highly conserved (Fig. 3B) (12). Next, we made a single mutation of putative key sites (D163A, D217A, and E251A) on the complement ISApl1. No cyclic products were observed when the mutation (D163A, D217A, and E251A) was present (Fig. 3A). These data show that ISApl1 cyclization was dependent on its transposase activity, especially the DDE domain.

Fig 3.

ISApl1 cyclization depends on the function itself. (A) ISApl1 transposition depends on the function itself. The kan gene was inserted into the ORF of ISApl1 to destroy the CDS. pSTV28 +ISApl1 and its point mutations D163A, D217A, and E251A plasmids were used for complement. The occurrence of transposition was verified by PCR using P1/P2 primers. The ISApl1 and kan genes are marked in orange and yellow, respectively. (B) ISApl1 transposition depends on the DDE domain. ISApl1 was aligned with the IS30, ISEnfa364, and ISSlu1 sequences. The amino acid sequences around the DDE domains of the three transposases were compared. The key amino acids in the DDE domain are marked in the yellow box.

To further investigate the significance of the transposition enzyme encoded by ISApl1 during transposition, we expressed and purified ISAPL1 protein. Unfortunately, our attempts at purification were unsuccessful. According to sequence alignment (Fig. 3B), we expressed and purified the ISAPL1-HTH (Helix Angle Helix) domain protein (Fig. 4A). An electrophoretic mobility shift assay (EMSA) experiment was performed to confirm the interaction between the ISAPL1-HTH protein and IRs of the ISApl1 gene. The results demonstrated the specific binding of the ISAPL1-HTH protein to IRL and IRR (Fig. 4B and D). These results uncover the function of transposase ISAPL1 which can bind the terminal sequence of ISApl1 and help its cyclization.

Fig 4.

ISAPL1 binds to the sequence flanking end of ISApl1. (A) Purification of ISAPL1 DNA-binding domain. The sequence of the ISApl1-HTH domain was amplified from the 17MR471 genomic DNA. pET28a-His₆-ISAPL1-HTH-His₆ was used to purify the ISAPL1-HTH protein. The purified ISAPL1-HTH protein was verified by 15% SDS-PAGE followed by Coomassie blue staining. (B) Electrophoretic mobility shift assay—the interaction between ISAPL1 and IRL-FAM-WT. FAM-labeled probe (IRL-FAM-WT) was added to each well and incubated with the concentration gradient protein. The IRL-no-FAM probe, consistent with the IRL-FAM-WT sequence but without FAM modification, was used for a specific competition. 16SrRNA is a random sequence of DNA amplified from the genome for non-specific competition. IRL: inverted repeat left. (C) Electrophoretic mobility shift assay—the interaction between ISAPL1 and IRL-FAM-NO-IRL. FAM-labeled probe (IRL-FAM-NO-IRL) was added to each well and incubated with the concentration gradient protein. The probe sequence of IRL-FAM-NO-IRL means without IRL compared with IRL-FAM-WT. (D) Electrophoretic mobility shift assay—the interaction between ISAPL1 and IRR-FAM-WT. FAM-labeled probe (IRR-FAM-WT) was added to each well and incubated with the concentration gradient protein. The IRR-no-FAM probe without FAM modification, compared with the IRR-FAM-WT sequence, was used for a specific competition. 16SrRNA is the same to (B). IRR: inverted repeat right. (E) Electrophoretic mobility shift assay—the interaction between ISAPL1 and IRR-FAM-NO-IRR. FAM-labeled probe (IRR-FAM-NO-IRR) was added to each well and incubated with the concentration gradient protein. The probe sequence of IRR-FAM-NO-FAM means without IRR compared with IRL-FAM-WT.

Flanking DNA sequences at the left and right sides regulate the excision of ISApl1

Despite extensive research on the functionality of the ISAPL1 transposase (Fig. 3 and 4), its excision process remains unexplored. The ISApl1 gene, with or without its IRL and IRR sequences, was cloned into pUC19 (Fig. 5A) and subsequently transferred into the Top10 strain (recA-) (25). Cyclization products were observed in the IRs and IRs-out-10 bp groups, whereas the no-IRs and no-IRR groups did not exhibit cyclization products (Fig. 5B). In addition, partial cyclization products of 252 bp were observed when there was no IRL (Fig. 5B). Sequencing analysis revealed that the product was attributed to homologous recombination between IRR and another CDS sequence (5′-ctcgcacagggcaaaaaacaagcagaa-3′) of ISApl1 (Fig. S3A). Subsequently, we conducted EMSA experiments to further validate the significance of the IR sequences. The results demonstrated that a shift was produced when the probe contained IRL (Fig. 4B), whereas no shift product was observed in the absence of IRL (Fig. 4C). The combination of IRR sequence of ISApl1 and ISAPL1-HTH is similar to the result of IRL (Fig. 4D and E).

Fig 5.

Inverted repeat (IR) sequences affect self-cyclization of ISApl1. (A) Diagram of ISApl1 transposition. ISApl1 was cloned into Puc19 and the ISApl1 self-cyclization was detected using the P1/P2 primers. The product of ISApl1 transposition was approximately 400 bp. Yellow and orange represent IRL and IRR, respectively. (B) ISApl1 transposition is dependent on the IR at both flanking ends. Five plasmid variants—pUC19 ISApl1-IRs (IRL and IRR), pUC19 ISApl1-no IRs (no-IRL and no-IRR), pUC19 ISApl1-no-IRR, pUC19 ISApl1-no-IRL, and pUC19 ISApl1-IRs-out-10 bp (contain IRL, IRR and extend of 10 base pairs outside of the IRs)—were constructed. The out-10 bp was the original sequence located in both flanking outside ends of the IR. The original sequences were “agtttaatcg” and “ggtaatattt.” The P1/P2 primer pair was used for detection. (C) Detection of the excision of ISApl1 by PCR. IRL and IRR were shortened respectively and then verified using PCR. The P1/P2 primer pair was used for detection. (D) RT-qPCR detection of the excision frequency of ISApl1 by RT-qPCR. ns, not significant; Student’s two-tailed unpaired t-test was utilized to calculate significant differences. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; *****P < 0.00001. (E) The IRL-FAM and IRR-FAM probes were truncated and incubated with or without ISAPL1 protein. The number indicates the length of IR in the probe. “-” means no protein, “+” means with protein. (F) Detection of ISApl1 self-cyclization when the flanking bases were S (C and G). Primers of ISApl1-circle-F and ISApl1-circle-R were used to detect the frequency of ISApl1 self-cyclization by RT-qPCR; WT-WT (5′-agtttaatcg-ISApl1-ggtaatattt-3′), WT-S (5′-agtttaatcg-ISApl1-ssssssssss-3′), S-WT (5′-ssssssssss-ISApl1- ggtaatattt-3′) and S-S (5′-ssssssssss-ISApl1-ssssssssss-3′) (W: A, T; S: G, C). Student’s two-tailed unpaired t-test was utilized to calculate significant differences. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; *****P < 0.00001. (G) Detection of ISApl1-self-cyclization when the flanking bases were W (A, T). Primers of ISApl1-circle-F and ISApl1-circle-R were used to detect the frequency of ISApl1-self-circle by RT-qPCR; WT-WT (5′-agtttaatcg-ISApl1-ggtaatattt-3′), WT-W (5′-agtttaatcg-ISApl1-wwwwwwwwww-3′), w-wt (5′-wwwwwwwwww-ISApl1-ggtaatattt-3′), and w-w (5′-wwwwwwwwww-ISApl1-wwwwwwwwww-3′). Student’s two-tailed unpaired t-test was utilized to calculate significant differences. NS, not significant (P > 0.05). (H) The ability of ISAPL1 bound to different bases bias flanking of IR. The binding ability of HTH protein with IRL-FAM-WT (6 fmol), IRR-FAM-WT (6 fmol), IRL-FAM-10S (6 fmol), IRR-FAM-10S (6 fmol) was observed by increasing the amount of protein. IRL-FAM-WT (IRL-FAM-agtttaatcg), IRR-FAM-WT (IRR-FAM-agtttaatcg), IRL-FAM-10S (IRL-FAM-SSSSSSSSSS), and IRR-FAM-10S (IRR-FAM-SSSSSSSSSS). “−” means no probe, “+” means with probe.

To investigate the impact of IR length on ISApl1 excision, we shortened the original IRL and IRR sequences. The findings revealed that the minimum length of ISApl1 transposition-dependent IRR was 12 bp (Fig. 5C), and a mismatch occurred when the IRR was truncated to 18–15 bp (Fig. S3C). Surprisingly, ISApl1 could transpose in a mismatched manner when there was NO-IRL (Fig. S3B). To achieve transposition without mismatches, the minimum length of IRL was 14 bp (Fig. 5C). In addition, the RT-qPCR revealed that the excision rate of ISApl1 corresponded to the length of the IRs (Fig. 5D). Furthermore, the EMSA experiment showed that the binding ability of the probes with the protein decreased with the shortening of IRL and IRR length (Fig. 5E). These findings provided further confirmation that the IRs were essential for the transposition of ISApl1.

In addition, the excision frequency was found to be significantly lower in the pUC19-ISApl1-IRs group (“5′-aagcttgtga-ISApl1-gtcgactcta-3′”) compared to the pUC19-ISApl1-IRs-out-10-bp group (5′-agtttaatcg-ISApl1-ggtaatattt-3′) (Fig. S4). It is interesting to note that the pUC19-ISApl1-IRs-out-10-bp group had a higher AT content in the flanking end of IRs. Subsequently, 10 bp random sequences S (C, G) or W (A, T) were added to the outside region of IRL and IRR to construct pUC19 variants. The excision rate of ISApl1 significantly decreased when it was flanked by CG-rich sequences (Fig. 5F). However, the rate was no significantly different when it was flanked by AT-rich sequences (Fig. 5G). Subsequently, we conducted EMSA experiments to determine whether the excision rate decreased due to the binding ability of ISAPL1 and the IRs at both ends. The results revealed that IRL-FAM-WT and IRR-FAM-WT had stronger binding abilities to ISAPL1 compared to IRL-FAM-10S and IRR-FAM-10S (Fig. 5H). As illustrated in Fig. 2E, the ISApl1 insertion site was primarily located in regions with high AT regions, which conferred its excision activity. In conclusion, the excision of ISApl1 was affected by the length of the IRs and the base bias at both flanking ends, which primarily affected the binding ability with ISAPL1.

hupA or hupB can regulate ISApl1-self excision

To further explore the novel mechanism of ISApl1 excision regulation, a pull-down assay was performed, and the results were analyzed by SDS-PAGE and mass spectrum (Fig. S5; Excel 2 and 3). HUα and HUβ, which are encoded by hupA and hupB respectively, were chosen as our target proteins (Table 1). Heat-unstable nucleoid protein (HU) has been reported to be involved in the global regulation and mediation of transposition (26, 27). To explore the role of hupA and hupB in ISApl1 transposition, we overexpressed the two genes and detected the excision rate in 17MR471. The results showed that the excision rate of ISApl1 was suppressed by hupA or hupB (Fig. 6A). In addition, the expression of ISApl1 was evaluated, revealing that hupA or hupB could suppress the expression of ISApl1, thereby regulating its excision (Fig. 6B). We failed to knock out hupA or hupB in 17MR471 due to its multi-resistance. Consequently, we knocked out hupA or hupB in E. coli Top10 and transformed the mutational strain with pUC19 + WT-ISApl1-WT. Unfortunately, the deletion of hupA or hupB did not affect the excision of ISApl1 (Fig. S6). To further explore the mechanism of hupA or hupB regulated the expression of ISApl1, HUα, and HUβ proteins were purified (Fig. 6C). EMSA assay confirmed their binding to the ISApl1 promoter (Fig. 6D). Therefore, HUα or HUβ suppressed the expression of ISApl1 by directly binding to its promoter. To investigate the specific binding sites of HUα and HUβ to the ISApl1 promoter, the IRL-FAM-long probe was shortened. The results showed that HUα or HUβ could bind to IRL-FAM-long or IRL-FAM-WT probes (with IRL sequences) but no shift was observed when IRL-FAM-NO-IRL was employed (without IRL sequences) (Fig. 6E). This suggests the binding site of HUα and HUβ is the IRL sequence. However, the HTH domain of ISAPL1 also binds to the IR sequence (Fig. 4). We hypothesized that HUα and HUβ regulated the transposition of ISApl1 by competing with ISAPL1 transposase for the same DNA sequence. Subsequently, we conducted competition experiments. The results showed that HUα or HUβ can displace ISAPL1 from the IRL-FAM-WT sequence (Fig. 6F). HUα and HUβ bound to the IRL sequence and regulated the excision of ISApl1 suggests that they may also bind to the IRR sequence, impacting the regulation of ISAPL1 transposition. To investigate this, an EMSA was performed using HUα or HUβ proteins incubated with the IRR-FAM-long probe. The experiment results revealed that both HUα and HUβ could bind to the probe (Fig. 6G). In addition, we observed that truncating IRR-FAM-long resulted in a notable decrease in the binding ability of IRR-FAM-NO-IRR (lacking the IRR sequence) when compared to IRRL-FAM-Long (with IRR) and IRR-FAM-WT (with IRR) (Fig. 6H). Furthermore, competition experiments demonstrated that both HUα or HUβ could bind to the IRR sequence, inhibiting ISAPL1 binding to IRR (Fig. 6I). The above results show that HUα or HUβ inhibit ISAPL1 transposition through two aspects: suppressing ISApl1 expression and competing for the same DNA sequence as ISAPL1.

TABLE 1.

Proteins from Pull-down

Protein	Functiona
HUα Uniprot: A0A0E0VE47	Histone-like DNA-binding protein which is capable of wrapping DNA to stabilize it
HUβ Uniprot: A0A0A5SD62	Histone-like DNA-binding protein which is capable of wrapping DNA to stabilize it

^a The database from Uniport.

Fig 6.

HU can inhibit the excision of ISApl1 (A) Overexpression of hupA or hupB in 17MR471 was followed by whole-genome extraction to examine the excision of ISApl1. 16SrRNA was used as an internal control. Student’s two-tailed unpaired t-test was utilized to calculate significant differences. *P < 0.05; **P < 0.01; ***P < 0.001. (B) Overexpression of hupA or hupB in 17MR471 was followed by RNA extraction and detection of ISApl1 expression. 16SrRNA was used as an internal control. Student’s two-tailed unpaired t-test was utilized to calculate significant differences. *P < 0.05; **P < 0.01; ***P < 0.001. (C) Expression and purification of HUα and HUβ proteins. (D) The promoter of ISApl1 (containing IRL) was incubated with HUα or HUβ proteins, respectively. IRL-FAM-Long was amplified using primers ISApl1-R-FAM, ISApl1-F-probe. (E) IRL-FAM-long probes were truncated and incubated with HUα or HUβ proteins. Probe 1: IRL-FAM-Long (ISApl1-R-FAM, ISApl1-F-Probe); Probe 2: IRL-FAM-WT (ISApl1-R-FAM, ISApl1-F-IRL-27); Probe 3 (ISApl1-R-FAM, ISApl1-F-NO-IRL). Probe 1 and Probe 2 with the IRL, Probe 3 without IRL. IRL was presented with a yellow color. (F) HUα or HUβ competed with ISAPL1 for IRL. The HUα: ISAPL1, HUΒ: ISAPL1 hybrid proteins were incubated with the probe (IRL-FAM-WT), and the HUα or HUβ content was gradually increased. (G) The IRR sequence of ISApl1 was incubated with HUα and HUβ proteins, respectively. IRR-FAM-Long was amplified using primers ISApl1-F-FAM, mcr-1-R-probe. (H) IRR-FAM-long probes were truncated and incubated with HUα or HUβ proteins. Probe 4: IRR-FAM-Long (ISApl1-F-FAM, mcr-1-R-probe); Probe 5: IRR-FAM-WT (ISApl1-F-FAM, ISApl1-R-IRR-27); and Probe 6: IRR-FAM-NO-IRL (ISApl1-F-FAM, ISApl1-R-NO-IRR). IRR sequences are shown in orange. (I) HUα or HUβ competed with ISAPL1 for IRR. The HUα: ISAPL1, HUβ: ISAPL1 hybrid proteins were incubated with the probe (IRR-FAM-WT), and the HUα or HUβ content was gradually increased.

ISApl1 can relieve the inhibition of HU

The interaction between HU and IRL or IRR sequences inhibits the transposition of ISApl1 by competitively binding to the same DNA sequence as ISAPL1 (Fig. 6). To determine whether the inhibitory effect of HU on ISApl1 transposition could be reversed by increasing the amount of ISAPL1, a competition experiment was conducted. HUα, HUβ, or ISAPL1 proteins were separately incubated with IRL-FAM-WT probe. Subsequently, ISAPL1 was incubated with HUα or HUβ, and the ISAPL1:HUα and ISAPL1:HUβ ratios were gradually increased. The results showed that ISAPL1: HUα generated the same shift as ISAPL1 instead of HUα or HUβ (Fig. 7A). Similar outcomes were observed with IRR-FAM-WT probe (Fig. 7B), indicating that ISAPL1 can displace HU from the IRL or IRR sequences, thereby restoring ISApl1 transposition. To validate this, ISApl1 was overexpressed and examined for transposition. The results demonstrated a significantly higher transposition frequency of ISApl1 compared to the WT, hupA, and hupB groups (Fig. 6A and 7C). Overall, these findings suggest that overexpression of ISApl1 relieves the inhibitory effect of HU on ISApl1 transposition.

Fig 7.

ISAPL1 can restore the excision of ISApl1. (A) ISAPL1 competed with HUα or HUβ for IRL, respectively. The ISAPL1: HUα, ISAPL1: HUβ mixture proteins were incubated with a probe (IRL-FAM-WT), and the content of ISAPL1 was gradually increased. (B) ISAPL1 competed with HUα or HUβ for IRR, respectively. ISAPL1: HUα, ISAPL1: HUβ mixture proteins were incubated with a probe (IRR-FAM-WT) and the content of ISAPL1 was gradually increased. (C) ISApl1 can promote the excision of ISApl1. We overexpressed ISApl1 in 17MR471 and detected the excision of ISApl1.16SrRNA as an internal reference. Student’s two-tailed unpaired t-test was utilized to calculate significant differences. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. (D) Schematic of HU-regulated ISApl1 transposition. (I) ISAPL1 transposase can bind to IRs, and its binding ability is affected by the length of IRs and the base bias on the flanking of IRs. (II) HU could bind to IRs to inhibit the expression of ISApl1 and compete with ISAPL1 transposase for the same DNA sequence.

DISCUSSION

Molecular epidemiology and retrospective studies have shown that food animals may serve as the origin of mcr-1-positive isolates (28). The plasmid pHNSHP45 (contained mcr-1) is detected from SHP45 isolate, which is separated from animal, was detectable in various bacteria, and can be transmitted (8). Subsequently, a variant of pHNSHP45, known as mcr-1-carrying conjugative plasmid, was found to have spread among Salmonella isolates obtained from humans, pigs, and chickens (29). The transmission of the mcr-1 gene on the plasmid may be the key factor in its transmission from animals to humans. This could explain why the number of environmental isolates exceeds that of clinical isolates (Fig. 1A). Moreover, the variation in the regional distribution of strains in China may also be influenced by several factors, including economic development, the number of hospitals and scientific research institutions, and research capacity.

To date, a total of 15 Inc-type plasmids carrying mcr-1 have been documented (30). Among these, IncX4, IncHI2, and IncI2 have emerged as the predominant plasmids carrying mcr-1 (9, 31). Our findings reveal the presence of 12 types of plasmids carrying mcr-1, with IncI2 and IncX4 being the primary types responsible for the transmission of mcr-1. In addition, we observed a higher frequency of plasmids carrying ISApl1 when compared to the predominant plasmids (IncI2, IncX4). These other plasmids (IncHI2A, IncP1, IncHI2, P0111, IncY, IncHI1A, IncN, IncHI1B, IncFII, RepA) exhibited a high frequency of carrying ISApl1. The presence of both ISApl1 and mcr-1 in a plasmid may confer the ability of mcr-1 transposition, which could explain the higher frequency of mcr-1 in IncI2- and IncX4-type plasmids.

The occurrence of transposition can be influenced by various factors, including transcription factors (26), the sequence at the ends of transposons (32–34), and the regulation of transposases (25, 35). Our results show that multiple factors affect ISApl1 transposition. These factors include the IRs and the base bias at both the flanking ends of ISApl1, as well as the transposase encoded by ISApl1 and the HU protein. Previous reports have highlighted the significance of transcription factors, such as IHF, FIS, and H-NS, in regulating transposition by binding to DNA sequences of transposons (32, 36–38). For instance, H-NS has been shown to bind to the terminal sequence of Tn5 transposons, thus modulating their transposition (32). IHF and H-NS have also been implicated in mediating the dissociation of a refractory protein-DNA composite during Tn10/IS10 transposition (39). Our results indicate a significant decrease in the excision rate of ISApl1 when hupA or hupB is overexpressed. Further research is needed to determine whether the effect is exerted by the homodimer (HUαα, HUββ) or the heterodimer (HUαβ). In addition, when we knocked out hupA or hupB in the TOP10 strain, there was no change in ISApl1 transposition. This suggests that the regulation of ISApl1 transposition by HU may be strain-specific.

The transfer of mcr-1 relies on the Tn6330 complex transposon but the molecular mechanism of how ISApl1 facilitates mcr-1 transmission requires further investigation. Erik Snesrud et al. have demonstrated that the birth and demise of the ISApl1-mcr-1-ISApl1 composite transposon hinged on the specific recognition and insertion of the mcr-1 surrounding sequence by ISApl1, as determined through comparative genomics (40). Furthermore, the stability of mcr-1 in the genome is contingent upon the deletion of both ends of the ISApl1 sequence (9). More experiments may be needed to verify the relationship between ISApl1 and mcr-1. The presence of ISApl1 surrounding mcr-1 in mcr-1-positive strains has been frequently observed (9, 41, 42) but the proportion of ISApl1 in relation to mcr-1 has yet to receive much attention. We evaluated the probability of ISApl1 quantitatively surrounding mcr-1. Out of the 947 mcr-1-positive strains, 66 were identified through third-generation sequencing, while 881 were identified through second-generation sequencing. It is important to note that the limitations of second-generation sequencing technology may influence this approximate result. Through sequence comparison, we discovered that the presence of ISApl1 upstream and downstream of mcr-1 accounted for 17.6% and 5.2%, respectively. Whenever ISApl1 is detected downstream of mcr-1, it is always accompanied by ISApl1 upstream; in other words, the structure of mcr-1-pap2-ISApl1 has yet to be found to exist. The neighboring sequences of ISApl1 potentially influence this observation. Previous studies have demonstrated that the DRs on both sides of ISApl1 remain consistent when inserted into a novel location (43). However, when comparing the sequence at both ends of ISApl1 around mcr-1, we found that the DRs on both sides upstream and downstream of ISApl1 are inconsistent but in the structure of Tn6330, the DRs of its were consistent. Wang et al. showed that the Tn6330 element synchronizes the DRs at both flanking ends by transposing them to a new position (9).

In this study, we discovered the widespread distribution of mcr-1-positive bacteria across China. We also found that ISApl1, a crucial factor in mcr-1 transmission, is capable of self-transposition. In addition, the self-cyclization of ISApl1 relied on the involvement of its encoded transposase. Moreover, the excision of ISApl1 was influenced by the length of IRs and the nucleotide bias at both flanking ends. Importantly, the excision of ISApl1 was influenced by HU, which suppressed the expression of ISApl1 and competed for the same DNA with ISAPL1. These results give us a better understanding of the molecular mechanism of mcr-1 dissemination.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions

The strains and plasmids are listed in Table S1. E. coli cultures were grown in a Luria broth medium (Oxoid). Plasmid maintenance involved antibiotic concentrations of chloramphenicol (15 µg/mL), ampicillin (150 µg/mL), and kanamycin (50 µg/mL). In addition, we obtained 2,463 whole-genome sequences of E. coli from the NCBI database (https://www.ncbi.nlm.nih.gov/pathogens/isolates/). These strains were extensively distributed across 33 regions of China, encompassing both environmental and clinical isolates.

Analysis of genome profiling and comparative genomics

ABRicate (44) was employed to identify resistance genes and plasmids in mcr-1-positive isolates. The results, aligned using Clustal Omega (45–47), were manually edited and corrected using Jalview. WebLogo (48) was used to align the surrounding sequence of ISApl1 insertion sites. The IS30 family protein sequence was downloaded from the ISfinder database. MAGE was used to generate a phylogenetic tree, which was edited and enhanced using iTOL (https://itol.embl.de/upload.cgi).

Plasmid construction

A series of plasmids derived from pUC19, including IR truncation, base preference at both flanking ends of IR, were constructed. The primers used for constructing are listed in Table S2. To generate ISApl1 mutant strain, we inserted the kanamycin resistance gene sequence into ISApl1. In addition, we created plasmids pSTV28 + ISApl1 and its point mutations (D163A, D217A, E251A) for complementation. To examine the impact of hupA and hupB on excision, we individually ligated them to pUCK19 plasmid (49).

Expression and purification of ISAPL1-HTH domain, HUα, and HUβ proteins

E. coli BL21 cells that carried the respective plasmids were incubated in LB at 37°C to an OD₆₀₀ of 0.5 and induced with 1 mM IPTG at 16°C for 20 h. The His-tagged proteins were purified using Purification Kit (P2226, Beyotime) and confirmed by 15% SDS-PAGE followed by Coomassie blue staining, and their concentrations were determined using Bradford Protein Assay Kit (P0006C, Beyotime).

Electrophoretic mobility shift assay

Fluorescein amidite (FAM)-labeled probes were incubated with proteins in binding buffer (50 mM NaH₂PO₄, 300 mM NaCl, 50 mM imidazole) at 25°C for 30 min. After incubation, a 5% native polyacrylamide gel in 1× Tris-borate-EDTA (TBE) buffer was used, applied at 140 V for 30 min at 4°C. The bind shifts were detected by Amersham Typhoon NIR (GE) and analyzed by ImageQuant TL 8.1.

Knockout hupA and hupB

The hupA or hupB knockout strains were generated using red recombination. Briefly, the pKD46 plasmid was transformed into the wild-type (WT) strain. Then, the chloramphenicol resistance (cat) gene, flanking by 40 bp homology arms located upstream and downstream of the target gene, was amplified via PCR using pKD3 as a template. Then, the cells were induced by L-arabinose. The positive strain was cultured in LB without antibiotics at 42°C for 16 hours to remove pKD46. The pCP20 plasmid was used to remove the cat gene. To eliminate the pCP20, the overnight culture was incubated at 30°C and then subjected to a 48-hour incubation at 42°C.

RT-qPCR detection of excision frequency

Plasmids or genomes were extracted and the excision rate of ISApl1 was determined using RT-qPCR. 16SrRNA served as the internal reference.

Total RNA extraction and RT-qPCR

An overnight culture of 17MR471 was prepared. The bacterial cells were collected by centrifugation at 12,000 × g for 1 minute. Next, 1 mL of RNAiso plus (108–95-2, TaKaRa) was added. The cells were then disrupted using 0.1 mm silica beads and a fast PRE24 automated system (MP Biomedicals). RNA was reverse-transcribed into cDNA using the PrimeScript RT reagent kit (RR420A, TaKaRa), and 16SrRNA served as the internal reference.

DNA Pull-down assay

DNA pull-down assays were conducted as previously described (50). The biotin-labeled ISApl1 promoter was amplified from the genomic DNA of 17MR471. 16SrRNA-biotin was used as a negative control. Cell cultures were inoculated into 20 mL of fresh LB medium and incubated overnight. The cells were then collected and lysed. The lysate was then centrifuged at 12,000 × g at 4°C for 40 minutes to remove insoluble debris. The supernatant containing 10 µg/mL of poly (dI-dC) was added to the DNA-coated beads and incubated at 4°C for 1 hour. Then, the beads were supplemented with ddH₂O (70 µL) and incubated at 70°C for 10 minutes. Samples were separated by SDS-PAGE. The entire lanes containing ISApl1 and 16SrRNA were excised and subjected to in-gel digestion with trypsin (0.6 mg). The resulting tryptic peptides were analyzed using liquid chromatography-tandem mass spectrometry (LC-MS/MS) with an LTQ mass spectrometer (ProteomeX-LTQ; ThermoFisher Scientific). Sequence and peptide fingerprint data were then analyzed using the NCBI database.

Statistical analysis

All experiments were performed in biological triplicates. GraphPad Prism (version 8.3d) was employed for all statistical analyses. Student’s two-tailed unpaired t-test was utilized to calculate significant differences. NS, not significant (P > 0.05); *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.01231-23.

Supplemental Excel 1. aac.01231-23-s0001.xlsx.

List of mcr-1-positive isolates.

Supplemental Excel 2. aac.01231-23-s0002.xlsx.

Pull-down for ISApl1.

Supplemental Excel 3. aac.01231-23-s0003.xlsx.

Pull-down for 16S rRNA.

Supplemental Fig S1. aac.01231-23-s0004.tif.

the cyclization of Tn6330.

Supplemental Fig S2. aac.01231-23-s0005.tif.

IS30 family evolutionary tree and IS30-family-align.

Supplemental Fig S3. aac.01231-23-s0006.tif.

Sanger sequence of cyclization product.

Supplemental Fig S4. aac.01231-23-s0007.tif.

Detection of cyclization frequency.

Supplemental Fig S5. aac.01231-23-s0008.tif.

Pull-down.

Supplemental Fig S6. aac.01231-23-s0009.tif.

Detection of cyclization frequency.

Supplemental Fig. S1-S7 and Tables S1-S2. aac.01231-23-s0010.docx.

Fig. S1-S7 and Excel S1-S3 descriptions and lists of isolates, plasmids, and primers.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.