Integrative conjugative elements mediate the high prevalence of tmexCD3-toprJ1b in Proteus spp. of animal source

ABSTRACT Integrative conjugative elements (ICEs) are important mobile elements that are associated with the dissemination of antibiotic resistance genes (ARGs) in Proteus. Recent studies demonstrated that the tigecycline resistance gene cluster tmexCD-toprJ has emerged in ICEs of Proteus. However, the prevalence of tmexCD-toprJ positive Proteus from animal sources is unclear. To cover the gap, a total of 762 Proteus spp. were isolated from animal source from six provinces of China to identify the tmexCD-toprJ positive isolates. Eight tmexCD-toprJ positive isolates were identified, with the tmexCD-toprJ positive rate of 1.05% against all Proteus spp. and 2.79% against ICE-bearing Proteus spp. The tmexCD-toprJ gene cluster in these Proteus spp. were tmexCD3-toprJ1b and all of them were carried by ICEs. Genetic structure analysis showed that tmexCD3-toprJ1b-bearing ICEs were complicated and plastic, but the tmexCD3-toprJ1b was specifically integrated into variable region III (VRIII) of ICEs with the help of integrases. Furthermore, we found that the umuC gene, which is presented in VRIII of ICEs and other genetic structures of many other bacterial genomes, was a hotspot for the integration of tmexCD-toprJ-bearing potential mobile elements. In summary, our results suggested that Proteus were important reservoirs for tmexCD-toprJ due to the high prevalence of ICEs. Therefore, continuous surveillance of ICEs-associated ARGs in Proteus is necessary and significant for controlling their future transmission. IMPORTANCE The emergence and spread of tmexCD-toprJ have greatly weakened the function of tigecycline. Although studies have demonstrated the significance of Proteus as carriers for tmexCD-toprJ, the epidemic mechanism and characteristics of tmexCD-toprJ in Proteus remain unclear. Herein, we deciphered that the umuC gene in VRIII of SXT/R391 ICEs was a hotspot for the integration of tmexCD3-toprJ1b-bearing mobile genetic elements by genomic analysis. The mobilization and dissemination of tmexCD3-toprJ1b in Proteus were mediated by highly prevalent ICEs. Furthermore, the co-occurrence of tmexCD3-toprJ1b-bearing ICEs with other chromosomally encoded multidrug resistance gene islands warned that the chromosomes of Proteus are significant reservoirs of ARGs. Overall, our results provide significant insights for the prevention and control of tmexCD3-toprJ1b in Proteus.

trimethoprim resistant Vibrio cholerae isolate O139 (4).R391 ICE was initially thought to be IncJ plasmid, which was discovered in a clinical Providencia rettgeri isolate in 1972 (5).SXT and R391 ICEs have a similar genomic skeleton structure and a fixed chromosomal insertion and recombination site.Hence, the two types of elements are classified as the same family, SXT/R391 ICE (6).
Currently, SXT/R391 ICEs have been found in many different genera of bacteria with low GC content, including Vibrio, Proteus, Shewanella, and Providencia, from humans, animals, and environments (7)(8)(9)(10)(11)(12).A conservative genetic context of roughly 47 kbp and many variable regions make up the backbone of SXT/R391 ICE.The conservative areas encode a total of 52 genes involved in activities such as integration, excision, and conjugative transfer (13).Five insertion hotspots (HS1, HS2, HS3, HS4, and HS5) and five variable regions (VRI, VRII, VRIII, VRIV, and VRV) in ICEs are typically reported to have varying DNA sequences (14).In addition to regulating bacterial motility, antibiotic resistance, and heavy metal resistance in the host, it has been discovered that the common SXT/R391 variable region genes also encode for a toxin-antitoxin system that prevents the loss of ICEs from the host (8).Therefore, as self-transmissible elements, ICEs played an important role in host adaptation and horizontal genetic structure exchanges in bacteria.
TMexCD1-TOprJ1 is a plasmid-mediated resistance-nodulation-division (RND) family efflux pump conferring resistance to tigecycline, which is first identified in Kleb siella pneumoniae (32).Subsequently, a tmexCD1-toprJ1 variant tmexCD3-toprJ1b was identified in Proteus mirabilis (33).In addition, studies provided that tmexCD-toprJ tended to integrate into the VRIII region of ICE in Proteus (30,34).To investigate the preva lence of tmexCD-toprJ in ICEs of Proteus, we isolated 762 Proteus spp.from 729 animal source samples from six provinces of China in 2020 and 2022.A total of 281 SXT/R391 ICEs positive isolates were identified, and among them, eight tmexCD-toprJ positive Proteus spp.were characterized with genomic analysis to decipher the current molecular epidemiology of tmexCD-toprJ positive ICEs.

The prevalence and characteristics of ICEs in Proteus spp. from animal sources
In 2020, 175 fecal and non-fecal samples (including blood, wastewater, and environmen tal samples) were collected from a swine slaughterhouse in Jiangsu Province.A total of 210 Proteus spp.were isolated from these samples according to the inherent resistance characteristics of tigecycline and colistin.Among them, 76 isolates were positive for SXT/R391 ICEs, and the positive rate was 36.19% (Fig. 1; Table S1).In 2022, a total of 552 Proteus spp.were isolated from 554 fecal samples collected in nine different chicken farms from six provinces of China, with 211 isolates testing positive for SXT/R391 ICEs.The prevalence of ICE-bearing Proteus spp.from different chicken farms varied from 20.83% to 69.77% (Fig. 1; Table S2).The overall positive rate of Proteus spp.harboring SXT/R391 ICEs from chicken farms was 38.22%.In general, the prevalence of ICEs in Proteus spp.from swine source and chicken source was similar and at high level.Species analysis showed that the 287 ICEs positive Proteus spp.consisted of four different species, including Proteus mirabilis, Proteus vulgaris, Proteus terrae, and Proteus penneri, with Proteus mirabilis dominated (89.20%, 256/287) (Fig. S1).Antimicrobial susceptibility testing showed that these ICEs positive Proteus spp.were 100% resistance to trime thoprim-sulfomethoxazole, tigecycline, amoxicillin, chloramphenicol, and colistin.The resistance rate of ICE-positive Proteus spp. was higher in chicken farms than that in the swine slaughterhouses (Fig. S1).

The prevalence of tmexCD-toprJ in ICE-bearing Proteus spp. and the resist ance phenotype of tmexCD-toprJ positive Proteus spp.
A total of eight tmexCD-toprJ positive isolates were identified from 762 Proteus spp., consisting of seven P. mirabilis isolates from chicken farms, and one P. terrae isolate from a swine slaughterhouse (Table 1).Among the seven tmexCD-toprJ positive isolates of chicken source, three were isolated from a chicken farm in Changzhou, Jiangsu.Another four were isolated from a chicken farm in Nanyang, Henan (Table S2).Of note, the eight isolates were all found in ICEs positive Proteus spp.The positive rate of tmexCD-toprJ against all Proteus spp.and ICE-bearing Proteus spp.were 1.05% (8/762) and 2.79% (8/287), respectively.Antimicrobial susceptibility testing showed that the eight isolates were multidrug resistant bacteria.All of them were resistant to florfenicol, streptomycin,  tetracycline, tigecycline, trimethoprim/sulfamethoxazole, amoxicillin, and colistin but were susceptible to meropenem.In addition, conjugation assay showed that ARGs carried by the eight isolates could not be transferred to recipient strain C600 under laboratory conditions.
The genomic characteristics of tmexCD3-toptJ1 positive Proteus spp.
To investigate the genomic features of the eight tmexCD-toprJ positive Proteus spp., whole genome sequencing was performed.Of which, the complete closed genomes of seven isolates except NYP69 were generated using both long-read and short-read sequencing methods.The genome sizes of the eight isolates were ranged from 4,059,985 bp to 4,149,995 bp.Genomic analysis showed that tmexCD-toprJ gene cluster in the isolates were all tmexCD3-toprJ1b variants.Apart from tmexCD3-toprJ1b, many other ARGs, including four extended-spectrum beta-lactamase (ESBL) genes bla CTX-M-14 , bla HMS-1 , bla CTX-M-3 and bla CTX-M-65 , were also detected in these isolates, implying that Proteus spp. is an important reservoir for ARGs (Table 1).However, only two Col3M-type qnrD1-bearing plasmids were found in the eight isolates, indicating that the ARGs in these isolates were almost located on the chromosome.Seven tmexCD3-toprJ1b-bearing ICEs were found in the complete genomes of the seven Proteus spp.Although isolate NYP68 was only sequenced with short-read method, the backbone of tmexCD3-toprJ1b-bearing ICE was found in short-read assembly.According to the genetic alignments, tmexCD3-toprJ1b-bearing ICE in isolate NYP68 showed highly similar to ICEPmiChnNYP73 in isolate NYP73.Hence, we confirmed that the tmexCD3-toprJ1b gene cluster in the eight isolates was located on chromoso mal ICEs.The complete ICEs in the seven isolates ranged in length from 75,940 bp (ICEPmiChnNYP73) to 123,747 bp (ICEPciChnTP22), with an average length of 104,721 bp.In addition, these ICEs contained two to nine ARGs, indicating that many ARGs in these isolates were located on other genetic structures of their chromosomes (Table 1).
Subsequently, we analyzed the genetic structures of other chromosomal ARGs of the seven tmexCD3-toprJ1b positive isolates.We found that majority of them were carried by genetic islands such as integrons and composite transposons (Fig. S2).The sizes of the seven genetic islands ranged from 11 kbp to 84 kbp.The largest genetic island found in the seven isolates was a multidrug resistance transposon in isolate CZP17.It contained 20 different ARGs, including ESBL gene bla CTX-M-14 .Complex class one integrons were detected in three isolates from Nanyang chicken farm, and a more complex structure of class one integrons was found in isolate NYP6, highlighting the variable structures of complex class one integrons.

Genetic structure analysis of tmexCD3-toprJ1b-bearing SXT/R391 ICEs
We identified seven tmexCD3-toprJ1b-bearing complete SXT/R391 ICEs in this study.Meanwhile, six previously reported tmexCD3-toprJ1b-bearing ICEs were retrieved from NCBI nr database and analyzed (Table 1).The comparative analysis found that ICEPmiChnNYP73, ICEPmiChnNYP69, and ICEPmiChnNYP6 detected in isolates from a chicken farm in Nanyang from this study share almost identical structures (Fig. S3).In addition, ICEs in isolates SDQ8C180-2T, SDQ8C176-1T and SDQ8C113RT from a chicken farm in Shandong also have almost identical gene arrangements (27).The phenom enon highlights that the tmexCD3-toprJ1b-bearing ICEs could widely spread in different niches via horizontally transfer or clonal spread.Notably, we found that tmexCD3-toprJ1b-bearing ICEs with different host species and geographical locations have a high degree of similarity, such that ICEPmiChnRGF134-1 and ICEPmiChn4280 with similar structures were isolated from a slaughterhouse and a patient, and ICEPmiChnCZP17 and ICEPmiChnTP22 with similar structures were carried by a P. mirabilis from chicken feces and a P. terrae from a slaughterhouse, respectively (Fig. 2).
Further analysis found that tmexCD3-toprJ1b gene cluster in these ICEs were all located on VRIII, and the core genetic environment of tmexCD3-toprJ1b was typical of a potential mobile structure tmexCD3-toprJ1b-hp-hp-int-int (Fig. 2; Fig. S4).The phenomenon also provided that the integration of tmexCD3-toprJ1b in VRIII of ICEs was likely assisted with integrase.Apart from tmexCD3-toprJ1b, many other ARGs were also located on VRIII.Due to the integration of foreign genes in insertion hotspots and variable regions, the sizes of these tmexCD3-toprJ1b-bearing ICEs were greater than the initial SXT/R391 ICEs identified in V. cholerae O139 and P. rettgeri R391 (Fig. 2).Hence, the SXT/R391 ICEs were mobile elements with consistent evolution ability.What is notable is that ICEPmiChnNYP73 had a significantly different backbone from other SXT/R391 ICEs.A total of 20 conserved genes involved in conjugation and other functions were lost in ICEPmiChnNYP73.Blastn analysis showed that ICEPmiChn3 was most similar to ICEPmiChnNYP73 (Fig. 2).Compared with ICEPmiChn3, a tmexCD3-toprJ1b-bearing region, with gene arrangement tmexCD3-toprJ1b-hp-hp-int-int, was integrated into the umuC gene of VRIII of ICEPmiChnNYP73.This further illuminated that the genetic structure of tmexCD3-toprJ1b-hp-hp-int-int was self-transmissible.

Phylogenetic analysis of ICE-bearing Proteus spp.
Many tmexCD3-toprJ1b-bearing ICEs have similar genetic structures.To investigate whether the structure of ICEs is related to the evolution of host Proteus spp., a phylo genetic tree consisting of all tmexCD3-toprJ1b positive Proteus spp.and some other Proteus spp. was constructed (Fig. 3).The four P. mirabilis isolates NYP6, NYP68, NYP69, and NTP73 from Nanyang chicken farm were clustered into one clade, implying that they originated from a common ancestor.Of note, the structures of ICEs in the four isolates were also almost identical, indicating that the spread of tmexCD3-toprJ1b-bear ing ICEs in the farm was caused by the clonal transmission of these isolates.Isolates CZP44, CZP17, and CZP26 from a chicken farm in Changzhou were distributed in three different clades in phylogenetic tree and they carried different ICEs.This indicated that tmexCD3-toprJ1b-bearing ICEs in the farm were diversified.We found that some other ICE-harboring isolates were phylogenetically close to CZP17 and CZP26, showing that some P. mirabilis isolates were widely distributed in a variety of settings and served as important ICEs reservoirs.For isolate TP22, although it was closely related to isolates SDQ8C180-2T, SDQ8C113-RT, and SDQ8C176-1T in the phylogenetic tree, they harbored different ICEs.According to the results, the prevalence of ICEs in Proteus was complex and diverse.The tmexCD3-toprJ1b gene cluster was initially integrated into ICEs of different Proteus spp.and spread horizontally or vertically to different niches.

DISCUSSIONS
In addition to the generally recognized plasmid-mediated transmission of most ARGs, SXT/R391 ICEs were also considered to be a significant factor in the spread of ARGs (35,36).Our results showed that the prevalence of ICEs in Proteus from animal source is over 30%, which is higher than that in humans and food (10).Such highly prevalent ICEs in Proteus may accelerate the aggregation and horizontal transfer of ARGs within Proteus.Furthermore, many recent researches have demonstrated that Proteus were important hosts for tigecycline resistance gene cluster tmexCD3-toprJ1b (27,29,33).Meanwhile, tmexCD3-toprJ1b in Proteus is commonly carried by ICEs.Hence, ICEs were regarded as important carriers for the horizontal transfer of tmexCD3-toprJ1b.In this study, eight tmexCD3-toprJ1b positive isolates were identified from 762 Proteus spp., the positive rate was 1.05%, which was higher than that in clinical Proteus spp (29).Of note, the tmexCD3-toprJ1b gene cluster in the eight isolates was all located on ICEs.The positive rate of tmexCD3-toprJ1b in ICE-bearing isolates was 2.79%.It is amazing that nearly 3 out of every 100 ICE-bearing Proteus isolates of animal origin carried tmexCD3-toprJ1b.Our results warned that ICE-mediated tmexCD3-toprJ1b had high prevalence in Proteus.Apart from the surveillance of plasmid-mediated tmexCD-toprJ in other genera of bacteria (29,37,38), we also should take notice to monitor and prevent the prevalence of tmexCD3-toprJ1b in Proteus.
Source analysis of all available tmexCD3-toprJ1b positive Proteus isolates showed that the majority of them were isolated from animals.Only two isolates were isolated from humans.According to previous surveillance, the prevalence of tmexCD3-toprJ1b in clinical patients was at an extremely low level (0.2%, 1/437) (29).Comparatively, our investigation showed that the positive rate of tmexCD3-toprJ1b in Proteus isolates from animal sources was considerably higher than from human sources.The different prevalence of tmexCD3-toprJ1b in Proteus from human and animal sources is likely associated with the survival condition of bacteria.Of note, tigecycline was not allowed to be used in the animal industry.Whereas tetracycline was found in feed for animals in extremely high amounts (71,800 ± 8,860 µg/kg) (39), which could result in serious antibiotic residues in the intestinal or fecal environment of livestock and might lead to the selection and enrichment of ARGs, including tmexCD3-toprJ1b.We noticed that two human-source tmexCD3-toprJ1b positive isolates were closely related in the phylogenetic tree, and they were close to a porcine tmexCD3-toprJ1b positive isolate.More impor tantly, almost identical tmexCD3-toprJ1b-bearing ICEs were found in Proteus isolates from both animal and human sources.These phenomena implied that tmexCD3-toprJ1b positive isolates or tmexCD3-toprJ1b-bearing ICEs might have been transmitted between animals and humans.Hence, the rational use of antibiotics should be conducted in animals to curb the emergence of ARGs and prevent the transmission of ARGs of animal origin to humans.
Genetic structure analysis found that tmexCD3-toprJ1b has been integrated into many kinds of ICEs in Proteus.Meanwhile, we found that the core genetic environments of tmexCD3-toprJ1b in ICEs, tmexCD3-toprJ1b-hp-hp-int-int, were in line with the genetic environments of other tmexCD-toprJ variants in other genera of bacteria (37,38,40).The integration sites of these tmexCD-toprJ-bearing structures were mostly the umuC gene, suggesting that umuC was likely an integration hotspot for the two integrases.Mean while, the insertion site of VRIII of ICEs was also the gene umuC.This might explained that the high prevalence of tmexCD3-toprJ1b in ICEs.Although the phenomenon has been noticed in many studies, the transfer of tmexCD3-toprJ1b-hp-hp-int-int in different bacteria has not been proved with experiments.
We noticed that the majority of ARGs in the eight tmexCD3-toprJ1b positive Proteus spp.were located on chromosomes.However, only partly of them were carried by ICEs (Table 1).Hence, other ARGs-associated genetic structures in the chromosome of Proteus were also important to the dissemination of ARGs.Previous studies have demonstrated that many chromosomal genetic islands, including large transposons and some mobile genetic elements, were usually found in the chromosome of Proteus, and they could mediate the horizontal transfer of many ARGs (22,41).In this study, we identified many different genetic islands harboring ARGs in the tmexCD3-toprJ1b positive Proteus spp., indicating that Proteus was important host for ARGs-bearing genetic islands.In addition, many Proteus were found to co-harbor chromosomal mobile elements and plasmids (23,27).In consideration of the wide distribution of Proteus in natural environments, we should pay more attention to their role in the transmission and shelter of ARGs, especially significantly novel ARGs.

Conclusion
In conclusion, a comprehensive investigation about the prevalence of tmexCD-toprJ in Proteus from animal sources was conducted.We discovered that the prevalent tmexCD-toprJ variant in Proteus was tmexCD3-toprJ1b.The high prevalence of tmexCD3-toprJ1b in Proteus was mediated by ICEs.The VRIII of ICEs was a hotspot for the integration of tmexCD3-toprJ1b-bearing genetic structures.

Bacterial isolation
In 2020 and 2022, 729 samples were collected from a swine slaughterhouses and nine chicken farms from six provinces of China (Fig. 1; Table S1 and S2).The pretreatment of samples was performed as previously described methods (21).Briefly, the samples were incubated in 5 mL tryptic soy broth (TSB) for 6 hours.Then, Proteus was screened on tryptic soy agar (TSA) and MacConkey agar plates supplemented with 2 mg/L tigecycline and 4 mg/L colistin due to that they were inherently resistant to tigecycline and colistin.Suspected Proteus spp.were purified on TSA plats, and then confirmed by PCR targeting the genus-specific gene tuf and 16S rRNA gene with Sanger sequencing (42).The SXT/ R391 ICE positive isolates were screened by PCR targeting conserved int genes (19).The tigecycline resistance gene cluster tmexCD-toprJ in Proteus spp.were screened by PCR using previously reported primers and positive PCR products were subjected to Sanger sequencing (28).

Antimicrobial susceptibility testing and conjugation experiment
Antimicrobial susceptibility testing for commonly used antibiotics was performed by broth microdilution according to the Clinical and Laboratory Standards Institute guidelines (43).Escherichia coli ATCC25922 was used for quality control.To verify the transferability of tmexCD-toprJ in Proteus, conjugation assay was performed using rifampin resistant E. coli C600 (Rif R ) as the recipient.The donor strain and the recipient strain in the log phase of growth were mixed at a ratio of 1:1 and then cultured overnight in 15-fold volume fresh LB broth or on LB agar plates.The transconjugants were screened on LB agar plates containing rifampin (300 mg/L), trimethoprim/sulfa methoxazole (4/76 mg/L), and tigecycline (2 mg/L).At last, the transconjugants were confirmed by PCR for tmexCD-toprJ and 16S rRNA.

DNA extraction, whole genome sequencing and bioinformatics analysis
Genomic DNA of the SXT/R391 ICEs positive isolates were extracted using the TIANamp Bacteria DNA Kit (Tiangen, China) following the manufacturer's instructions.Then, we used Qubit 4 Fluorometer and NanoDrop (Thermo Scientific) to evaluate the quality and purity of genomic DNA, respectively.Following that, the genomic DNA was carried out short-read sequencing at Illumina Hiseq 2,500 platform and long-read sequencing at Oxford Nanopore Technologies MinION platform, respectively.Short-read genome assembly was performed with SPAdes (44).The closed complete genomes were obtained using a hybrid assembly strategy combining of short-read and long-read sequences with Unicycler (45).Plasmid replicon genes, insertion sequences and ARGs were investigated using abricate tool based on Plasmidfinder (46), ISfinder (47) and AMRFinder (48) databases.Subsequently, the genomes were annotated using Prokka and the resulting GFF3 files were pipelined into Roary to create core genome comparisons (49,50).The phylogenetic trees of ICE-bearing Proteus spp.were constructed using Roary and FastTree based on SNPs of core genomes (50,51).Easyfig was used to visualize the genetic comparisons (52).

FIG 1
FIG 1 The distribution of collected samples from chicken farms and swine slaughterhouse in different provinces of China and the prevalence of ICEs in Proteus spp. of animal source.The fecal samples of chicken were collected from nine cities of six different provinces in China.The positive rate of ICEs in Proteus spp. of chicken farms and swine slaughterhouse from different cities were noted after the city's name.

FIG 2
FIG 2 Genetic structures analysis of tmexCD3-toprJ1b-bearing SXT/R391 ICEs in this study and other similar ICEs.The gray-shaded region indicates the homologous region of the two ICE structures.The red arrows indicate the ARGs.The genes pfrC, umuC, and umuD are shown as cyan arrows.Pink arrows represent mobile elements.The gene prfC represents the insertion site of ICEs.The ∆ symbol indicates that the gene is truncated.

FIG 3
FIG 3 Phylogenetic analysis of tmexCD3-toprJ1b positive Proteus spp.and other ICE-harboring Proteus spp.The isolate names in this study were highlighted with blue.The ARGs numbers carried by isolates were represented by red stars of different sizes.

TABLE 1
Genome characteristics of the tmexCD-toprJ-positive isolates from this study and the NCBI database.