This article is mentioned in:

Introduction

Shigella is a genus of pathogens responsible for acute bacterial diarrhea, chronic bacteritic dysentery and toxic dysentery. Certain Shigella strains can cause hemolytic uremic syndrome and Reiter's chronic arthritis syndrome, and with the extensive use of antibacterial agents, bacterial drug-resistance is becoming a serious concern. Mobile genetic elements, including plasmids, transposons and integrons are important in the spread of antibiotic resistance (1). Among these elements, class 1, 2 and 3 integrons are associated with drug resistance. A previous study showed that the involvement of the active efflux pump is significant in bacterial multi-drug resistance (2). Previous studies have mainly focused on the active efflux pump of Escherichia coli, AcrAB-TolC, which is also expressed in Shigella striains (3). The regulation of gene mutation may be important in the multi-drug resistance mediated by the gene (4). In the present study, the characteristics of integron genes were analyzed in clinical Shigella isolates from children, mutations of the active effiux pump gene (acrAB-tolC) and regulatory genes were examined, and the association between these genes and the drug-resistance of Shigella was investigated. This may provide evidence for the clinical treatment of dysentery and the control of multi-drug resistance of Shigella.

Materials and methods

Location of isolated strains

A total of 158 isolates of Shigella flexneri were isolated from the stool samples of 94 male cases and 64 female cases aged from 6 months to 14 years old between May 2012 and October 2015, in which there were 106 cases aged less than 5 years old. The isolates were identified using ID32E identificational strips and an ATB expression instrument, and were serotyped using diagnostic-serum of Shigella. Escherichia coli (cat. no. 25922; American Type Culture Collection) was preserved in the Clinical Microbiology Laboratory of The Fourth Hospital of Jinan City (Jinan, China) and it was the quality control strain for antimicrobial susceptibility assessment. The present study was approved by the Ethics Committee of Jinan Infectious Disease Hospital Affiliated To Shandong University (Shandong, China). Written informed consent was obtained from all patients or patient guardians for the present study.

Reagents and equipment

The following were obtained for use in the present study: Shigella diagnostic serum (Lanzhou Institute of Biological Products Co. Ltd., Lanzhou, China); antibotic discs of ampicillin, chloramphenicol, tetracycline, streptomycin, trimethoprim, sulfamethoxazole, ciprofloxacin, cefotaxime, ceftazidime, cefoperazone/sulbactam and imipenem, and M-H agar (Oxoid; Thermo Fisher Scientific, Inc., Waltham, MA, USA); 100 bp DNA Ladder marker and D15000+2000 DNA marker (Tiangen Biotech Co., Ltd., Beijing, China); Takara Taq kit, DL2000TMDNA marker and restriction enzyme (Takara Biotechnology Co., Ltd., Dalian, China), agarose (Invitrogen; Thermo Fisher Scientific, Inc.); bacterial identification instrument (ATB Expression; BioMérieux, Marcy l'Etoile, France); DNA amplifier (Biometra GmbH, Göttingen, Germany); electrophoresis apparatus (10C type; Beijing Liuyi Biotechnology Co., Ltd., Beijing, China); biosafety cabinets (1200IIA2; Shanghai Lishen Biotechnology Co., Ltd.). Casein hydrolysate acid (M-H) agar (Shanghai Lishen Biological Technology Co., Ltd.).

Strain identification and antimicrobial susceptibility assessment

The preserved strains were removed from the −86°C ultra-low temperature freezer, thawed at room temprature and re-identified. Subsequently, the strains underwent an agglutination test with Shigella diagnostic serum for the serotype identification. In total, 3–4 bacterial colonies of S. flexneri were detected following culture of the stains for 16–18 h at 35°C, followed by adjustment of the turbidity instrument to 0.5 McF with normal saline. Sterile cotton swabs were dipped into the bacteria solution and excess liquid was squeezed out against the tube wall, followed by application of the bacteria onto the M-H plate. The bacteria were applied ontp the M-H plate three times with a 60° rotation; the bacteria were also applied along the edge for entire 360° rotation. Following 5 min, the antimicrobial susceptibility disks were placed on the colonies and cultured for 16–18 h at 35°C. Following these procedures, the inhibition zone diameter was detected with vernier calipers and the data were input into WHONET 5.4 software (http://www.whonet.org/). According to the 2010 version of Performance Standards for Antimicrobial Susceptibility Testing of the National Committee for Clinical Standards Laboratory Institute) (5), the quality control strains was Escherichia coli (cat. no. 25922).

Assessment and sequencing for resistance gene cassettes of integrons and variable regions

The integron and plasmid DNA templates were prepared using a boiling method for 5 min, and they were preserved in a −20°C refrigerator. The primers and primer sequences were synthesized in the accordance with relative reports (6–9), as in Table I. The class 1, 2 and 3 integrons, integrase primers and variable region primers were synthesized by Takara Biotechnology Co., Ltd. The reaction system was as follows: 10X buffer containing Mg2+ (5 µl), dNTP (4 µl; 2.5 mmol/l), downstream and upstream primers (2.5 µl; 10 µmol/l of each), Taq eznzyme (0.25 µl; 5 U/µl) and DNA templates (5 µl). Deionized water was added to the reaction system to 50 µl. The amplification parameters for integrase were as follows: Pre-denaturation at 94°C for 5 min, followed by 94°C for 30 sec, annealing for 30 sec (annealing temperature shown in Table I) and 72°C for 40 sec, for 35 cycles, with final extension at 72°C for 5 min. The amplification parameters for variable regions were as follows: Pre-denaturation at 94°C for 5 min; touchdown polyermase chain reaction (PCR) for 10 cycles, annealing temperature reduction from 60 to 50°C (temperature decrease by 1.4°C each cycle), followed by 94°C for 30 sec, 52°C for 30 sec and 72°C for 180 sec, for 25 cycles with final extension at 72°C for 8 min. For product analysis, the amplification product (10 µl) was used for 1% agarose electrophoresis under 120 V for 20 min. Following EB staining, the products were observed under the UV absorption spectra analyzer. Images were captured and the results recorded. The amplification products were sent to Takara Biotechnology Co., Ltd. for gene sequencing and the sequences were compared with NCBI/BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Table I. Primers of genes associated with class 1 integron.

Table I.

Primers of genes associated with class 1 integron.

Primer	Sequence (5′-3′)	Primer sequence site
intI1 F	ACATGTGATGGCGACGCACGA	intI1
intI1 R	ATTTCTGTCCTGGCTGGCGA	intI1
In F	GGCATCCAAGCAGCAAGC	5′-conserved segment in the class 1 integrons
In R	AAGCAGACTTGACCTGAT	3′-conserved segment in the class 1 integrons
qacEΔ1	ATCGCAATAGTTGGCGAAGT	qacEΔ1
sul1	GCAAGGCGGAAACCCGCGCC	sul1
intI1ca	CGTAGAAGAACAGCAAGG	intI1
IS1ca	AGTGAGAGCAGAGATAGC	IS1
intI2 F	GTAGCAAACGAGTGACGAAATG	intI2
intI2 R	CACGGATATGCGACAAAAAGGT	intI2
intI2ca F	CGGGATCCCGGACGGCATGCACGATTTGTA	intI2
intI2ca R	GATGCCATCGCAAGTACGAG	3′-conserved segment in the class 2 integrons
intI3 F	GCCTCCGGCAGCGACTTTCAG	intI3
intI3 R	ACGGATCTGCCAAACCTGACT	intI3

[i] F, forward; R, reverse.

BamHI restriction enzyme digestion

Enzyme digestion was performed in the variable regions of the atypical class 1,2 integrons with the restriction enzyme HindIII and Hinf1. The specific steps were as follows: 10 µl enzyme buffer, 40 µl PCR products and 5 µl restriction enzyme were added to deionized water to 100 µl. The solution was placed at 37°C for the isothermal reaction for ~1 h. Subsequently, 10X loading buffer was added to terminate the reaction. The 10-µl samples were used for 2.5% agarose electrophoresis under 120 V for 20 min. Following EB staining, the products were observed under the UV absorption spectra analyzer. Images were captured and the results recorded.

Analysis of acrAB-tolC and its associated regulatory genes

The multi-drug resistant Shigella strains (resistant to five or more antibiotics) and sensitive strains were used for assessment. The PCR amplication and sequencing methods were applied. The specific steps were in accordance with the above methods. According to previous reports (10,11), the experimental primers were synthesized by Takara Biotechnology Co., Ltd. (Table II). The sequences were compared with NCBI/BLAST to investigate the numbers of mutations and mutational sites.

Table II. Active efflux pump acrAB-tolC and its regulatory genes.

Table II.

Active efflux pump acrAB-tolC and its regulatory genes.

Primer	Sequence (5′-3′)	Length (bp)	Annealing temperature (°C)
gyrA	TACACCGGTCAACATTGACG	648	50
	TTAATGTTGCCGCCGTCGG
parC	GCGTTGCCGTTTATTGGTGAT	469	52
	TGGACATCGTCATACCTCT
acrA	TGCGGCTTGCTGGTTATT	1,131	52
	GCGGTCGTTCTGATGCTC
acrB	GATTCCGACCATTGCCGTAC	510	52
	GCCAGAATACCGCCTACGC
marOR	CACTCTTTAGCTAGCCTTG	604	51
	TGGACATCGTCATACCTCT
acrR	AAACCCATTGCTGCGTTTAT	800	55
	AAACCGCAAGAATATCACGA

Results

Results of drug sensitivity tests

A total of 91.1% of the 158 strains were multi-resistant to ampicillin, chloramphenicol, tetracycline and streptomycin, and 69.6% of the strains were multi-resistant to sulfamethoxazole/trimethoprim. The resistance rate to ceftazidime, ciprofloxacin and levofloxacin was <32.9%. All (100%) of the strains were sensitive to cefoxitin, cefoperazone/sulbactam and imipenem. The frequent resistant pattern (ampicillin-tetracycline-chloromycetin-streptomycin) accounted for 91.1% of strains (144/158), and 77.2% (122/158) of the strains were resistant to more than five antibiotics. Multi-drug resistance was a serious problem. Only 14 strains were sensitive strains, which were not drug-resistant (Table III).

Table III. Results of class 1 and class 2 integron detection in 79 strains of Shigella flexneri.

Table III.

Results of class 1 and class 2 integron detection in 79 strains of Shigella flexneri.

Serotype	Typical class 1 integron (intI1, in, qacEΔ1-sul1-positive)	Atypical class 1 integron (intI1, intI1-IS1-positive)	Class 2 integron (intI2, intI2ca-positive)
S. flexneri 1a	0	16	14
S. flexneri 2a	20	70	74
S. flexneri 2b	2	10	10
S. flexneri 4a	4	34	34
S. flexneri × variant	0	4	4
Total	26	134	136

Integration assessment results

Among the 158 S. flexneri strains, 144 strains were intI1-positive and 136 strains were intI 2-positive. The positive rates were 91.1% (144/158) and 86.1% (136/158), respectively (Figs. 1 and 2). All strains were intI3-positive. In total, 10 strains were single ntI1-positive, accounting for 6.3% (10/158), and 134 strains were intI1-intI2 positive, accounting for 84.8% (134/158). Among the intI1-positive strains, the 3′-conserved segment of 16.5% of the strains (26/156) were positive. They were typical class 1 integrons [intI1, (conservative primer of class 1 integrons) in and qacEΔ1-sul1-positive; Fig. 3]. The 3′-conserved segment of 84.8% (134/158) of the strains were negative. They were atypical class 1 integrons (intI1 and intI1-IS1-positive; Fig. 4. In addition, 10.1% (16/158) strains (S. flexneri 2A) possessed typical and atypical class 1 integron class 1 integrons. The integron assessment results are shown in Table III.

Figure 1. Region of the class 1 integrase polymerase chain reaction electropherogram. M, DNA marker; 1–6 are the specimen numbers.

Figure 2. Region of the class 2 integrase polymerase chain reaction electropherogram. M, DNA marker; N, Negative control; 1–6 are the specimen numbers.

Figure 3. Region of the polymerase chain reaction amplification electrophoretogram of typical class 1 integron variable region. M, DNA marker; 1–8 are specimen numbers.

Figure 4. Region of polymerase chain reaction amplification electrophoretogram of atypical class 1 integron variable region. M, DNA marker. 1–6 are the specimen numbers.

Assessment and sequencing for resistance gene cassettes of integrons and variable regions

Among the 26 strains with typical class 1 integrons, PCR amplification of the variable regions in 22 strains yielded two types of fragment. The sizes were ~1.6 kb and 0.7 KB, respectively (Fig. 3). Confirmed by the gene sequencing, four S. flexneri 4a strains carried the dfrA17-aadA5 (1,624 bp) drug-resistant gene cassette, which was the dihydrofolate reductase gene, and aminoglycoside acyltransferase gene. A total of 18 S. flexneri 2a strains carried the dfrV (729 bp) drug-resistant gene cassette, which was the dihydrofolate reductase gene and aminoglycoside acyltransferase gene. The PCR amplification of four stains yielded nothing suggest an empty cassette. The assessment of antibiotic susceptibility showed that the drug resistance of pathogens to ampicillin and other drugs was severe, suggesting that the drug resistance was associated with the insertion of a 3′ conservative region into the C-terminal conservative region (9). The variable regions of 134 S. flexneri strains were amplified with the intI1ca-IS1ca primer. A total of 134 strains yielded 2.4 kb fragments, as shown in Fig. 4. The products had the same enzymatic maps following HindIII enzyme digestion. Four strains were randomly selected for sequencing. The carried drug resistant gene cassette was bla0XA-30-aadA1 (2,453 bp), which was the β-lactamase and aminoglycoside adenosine acyltransferase gene. A total of 134 strains with atypical class 1 integron were resistant to ampicillin, streptomycin, tetracycline and chloromycetin. Only 10 of the 26 strains without atypical class 1 integron were resistant to more than four antibiotics. There was a significant difference between two groups (χ2=35.96, P<0.01). The detailed results of the resistance gene in class 1 integron variable regions are shown in Table III.

Resistance gene cassette assessment and sequencing in variable regions of integrons

The variable region of integrons in the 136 strains was amplified by integron PCR. Following electrophoresis, all the strains yielded 2.2 kb fragments (Fig. 5). Following hinf1 digestion, the amplification products manifested the same characteristic bands in electrophoresis. Verified by gene sequencing, the drug-resistant gene carried by bacteria was the dfrA1-sat1-aadA1 gene (2,224 bp), which was responsible for the drug resistance to trimethoprim, streptomycin and aminoglycoside. The results of the resistance gene in class 2 integron variable regions are shown in Table IV.

Figure 5. Region of polymerase chain reaction amplification electrophoretogram of class 2 integron variable region. M, DNA marker; N, Negative control; 1–6 are the specimen numbers.

Table IV. Drug-resistant phenotype integrons and gene cassette of 56 Shigella flexneri strains.

Table IV.

Drug-resistant phenotype integrons and gene cassette of 56 Shigella flexneri strains.

	Class 1 integron and gene cassettes		Class 2 integron and gene cassettes
Drug-resistant phenotype (number of strains)	Class 1 integrase gene	Gene cassettes	Class 2 integrase gene	Gene cassettes
ATCR (2)	intI1 (n=2)	blaOXA-30-aadA1	intI2 (n=0)	–
	intI1 (n=4)	–	intI2 (n=2)	dfrA1-sat1-aadA1
	intI1 (n=16)	blaOXA-30-aadA1	intI2 (n=16)	dfrA1-sat1-aadA1
ATCRS (40)	intI1 (n=12)	blaOXA-30-aadA1	intI2 (n=12)	dfrA1-sat1-aadA1
		dfr V
	intI1 (n=6)	dfr V	intI2 (n=6)	dfrA1-sat1-aadA1
	intI1 (n=4)	blaOXA-30-aadA1	intI2 (n=0)	–
	intI1 (n=18)	blaOXA-30-aadA1	intI2 (n=18)	dfrA1-sat1-aadA1
ATCRP (2)	intI1 (n=2)	blaOXA-30-aadA1	intI2 (n=2)	dfrA1-sat1-aadA1
ATCRPL (10)	intI1 (n=10)	blaOXA-30-aadA1	intI2 (n=10)	dfrA1-sat1-aadA1
ATCRSPL (22)	intI1 (n=18)	blaOXA-30-aadA1	intI2 (n=18)	dfrA1-sat1-aadA1
	intI1 (n=4)	blaOXA-30-aadA1 dfrA17-aadA5	intI2 (n=4)	dfrA1-sat1-aadA1
ATCRSPLX (12)	intI1 (n=12)	blaOXA-30-aadA1	intI2 (n=12)	dfrA1-sat1-aadA1
ATCRSXZ (30)	intI1 (n=30)	blaOXA-30-aadA1	intI2 (n=30)	dfrA1-sat1-aadA1
ATCRSPLXZ (6)	intI1 (n=6)	blaOXA-30-aadA1	intI2 (n=6)	dfrA1-sat1-aadA1

[i] A, ampicillin; T, tetracycline; C, chloramphenicol; R, streptomycin; S, sulfamethoxazole/trimethoprim; P, ciprofloxacin; L, levofloxacin; X, cefotaxime; Z, trimethoprim.

Analysis of acrAB-tolC and its associated regulatory genes in multi-drug resistant Shigella

There were 12 strains with gene deletions among the 122 multi-drug resistant strains and two strains with gene deletions in 17 sensitive strains. Therefore 110 multi-drug resistant strains and 15 sensitive strains yeilded 131, 510, 604, 1,100 and 800 bp sequences (Fig. 6-D). Verified by gene sequencing, these fragments were the PCR amplification products of acrA, acrB, marOR, soxS and acrB. There were 110 cases in drug resistant group and four cases in the sensitive group. The gene analysis showed a low mutation rate in the acrA, acrB, acrR and soxS genes. In the multi-drug resistant strains, a base CATT deletion in the 36, 37, 38 and 39 sites of marOR was detected. However, no marOR mutation was found in sensitive strains.

Figure 6. Polymerase chain reaction amplification and electrophoretogram of multi-drug resistance gene acrAB-tolC and its regulatory genes (A) multi-drug resistance gene acrA, (B) multi-drug resistance gene acrB, (C) multi-drug resistance gene marOR and (D) multi-drug resistance gene acrR. 1, 368 strains; 2, 157 strains; 3, 3,171 strains; 4, 187 strains; 5, 4,536 strains; 6, 1,113 strains; 7, 22 strains; 8, N8; M, DNA marker (DL2000). Upper to lower progressively 2,000, 1,000, 750, 500, 250 and 100 bp. M, DNA marker; 1–8 are the specimen numbers.

Discussion

Bacillary dysentery caused by Shigella is a serious intestinal infectious disease and the incidence ranks third among the national statutory B infectious diseases. In addition to transient immunity following infection, the absence of cross-immunoreactivity between subtypes and problems in the immune system of children, the increased drug-resistance of Shigella poses a serious threat to childrens' health. A previous study showed that Shigella had the AcrAB-TolC active efflux pump, which caused multi-drug resistance to organic solvents, dyes, detergents and virious antibiotics drugs (12). In another study (13), the high expresion level of the active efflux pump can enhance the resistance of bacteria to organic solvents. It also been reported that the resistance of Escherichia coli to organic solvents was affected by certain inherited traits, including the imp/ostA gene, which coded a protein with a molecular mass of 87,000. The protein was attached to the outer membranes. The low expression of the gene increased the sensitivity to N-hexane (14). Certain cell elements that do not belong to the antivie efflux pump, for example the derivatives of the O-sidechain of lipopolysaccharides or the isomers of fatty acid membranes, may have an effect on the resistance of Pseudomonas to organic solvents (15).

The results of the present study showed that 91.1% S. flexneri strains were multi-drug resistant. According to the shigella treatment recommendations of the American Clinical and Laboratory Standards Institute, ampicillin, sulfamethoxazole/trimethoprim were not suitable for the empirical clinical medication. Although the drug resistance rate of ciprofloxacin and levofloxacin was 32.9%, the fluroquinolones were not suitable for pediatric patients due to the potential damage to cartilage in infants and children. The drug-resistance mechanism was associated with multiple-site mutations of gene gyrA and/or parC. The drug-resistance rate to cefotaxime and cefotaxime was <30.4%, which meant that these drugs were suitable for use in the treatment of bacillary dysentery in children. It is noteworthy that the drug resistance was significantly higher than that reported previously In China and elsewhere (16,17), sugesting that, in case of treatment failure with third generation cephalosporins or in case of pediatric dental patents, cefoxitin, cefoperazone/sulbactam and imipenem may be applied in the treatment.

The integron-mediated muhidrug-resistance has been verified previously. Class 1 integron was detected in 91.1% of 158 S. flexneri strains. Class 2 integron was detected in 86.1% of 158 S. flexneri strains. The propotions were significantly higher, compared with those reported in South Korea and Iran (18,19). The islolating regions and epidemic strains may accout for this. The typical class 1 integron gene was present in 15.2% of the isolates screened. The drug resistance gene cassette was frV and dfrA17-aadA5. The atypical class 1 integron accounted for 86.1% strains. The drug resistance gene cassette was blaOXA30-aadA1. The typical and atypical class 1 integrons were present in 10.1% of isolates screened, similar with the relative reports (20,21). The typical class 1 integron was present in 11.4% of the S. flexneri isolates screened. This result has not been reported previously, to the best of our knowledge. In individual studies, the dfrv gene was present in Escherichia coli, Salmonella bacteria and Gram-negative bacilli in waste water (20–22). Whether the resistance genes were transfered from these bacteria requires further investigation.

A total of 134 strains with atypical class 1 integron were resistant to ampicillin, streptomycin, tetracycline and chloromycetin. Only 10 of the 26 strains without the atypical class 1 integron were resistant to more than four antibiotics. There was a significant difference between them (χ2=35.96, P<0.01). This difference is associated with a situation that the blaOXA-30-aadA1 gene cassette in the atypical class 1 integron variable region was adjacent to the chloramphenicol- and tetracycline-resistant determinants (23). Those strains with no atypical class 1 integrons, which were resistant to the four antibiotics may have other drug resistance mechanisms. There were 10.1% strains (S. flexneri 2A) possessing typical and atypical class 1 integron class 1 integrons (24), which deteriorate the drug resistance of S. flexneri. The typical class 1 integron was found in conjugative plasmids and can mediate the horizontal gene transfer. In the present study, 86.1% S. flexneri isolates carried the class 2 integron and the gene cassette was dfrA1-sat1-aadA1, which was consistant with other reports (25,26). The gene cassettes reported currently comprise dfrAl-sat1, dfrAl-sat2-aadA1 and sat2-aadA1 (27–29). As the class 2 integron integrase is inadequate for integrating novel resistance genes from externally or to resect the resistance genes from its own variable region, the drug resistance gene cassettes of class 2 integron variable regions are relatively conservative.

The drug resistance rate to sulfamethoxazole/trimethoprim was 69.6%. The strains resistant to sulfamethoxazole/trimethoprim were compared with the sull gene carried in the atypical class 1 integron, the dfrAl gene carried in the class 2 integron, and the two together, suggesting that the two genes were not associated with the drug resistance. The two genes may be nonfunctional structural genes. The mechanism undferlying the drug resistance to sulfamethoxazole/trimethoprim requires further investigation.

The single-stranded conformation polymorphism suggested a low mutation rate of acrA, acrB and tolC in drug-resistant Shigella. The expression of AcrA and AcrB are regulated by several regulators. Of all the regulators the acrR suppressors and marOR operon are important. The results demonstrated a low mutation rate of marOR in drug-resistant Shigella, suggesting that the mutaion of the reglatory gene contributed more to the drug resistance of Shigella than the mutaion of a pump gene. Further investigations aim to examine the association between the mutation of regulatory gene marOR and the expression of acrA, acrB and tolC, which may provide a theoretical basis for the drug-resistant mechanism mediated by the active efflux pump of Shigella.

Acknowledgements

Not applicable.

Funding

No funding was received.

Availability of data and materials

The analyzed data sets generated during the study are available from the corresponding author on reasonable request.

Authors' contributions

QPM was responsible for specimen collection, bacterial identification, and experimental procedures. MXY was responsible for experimental operations and data collection. LS and JWL were responsible for data collection and statistical analysis. GYY was responsible for experimental design and data review.

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Jinan Infectious Disease Hospital Affiliated To Shandong University (Shandong, China).

Consent for publication

Written informed consent was obtained from all patients or patient guardians for the present study.

Competing interests

The authors declare they have no competing interests.

References