Antimicrobial resistance patterns and genetic elements associated with the antibiotic resistance of Helicobacter pylori strains from Shanghai

Abstract

Background: Shanghai, in east China, has one of the world's highest burdens of Helicobacter pylori infection. While multidrug regimens can effectively eradicate H. pylori, the increasing prevalence of antibiotic resistance (AR) in H. pylori has been recognized by the WHO as 'high priority' for urgent need of new therapies. Moreover, the genetic characteristics of H. pylori AR in Shanghai is under-reported. The purpose of this study was to determine the resistance prevalence, re-substantiate resistance-conferring mutations, and investigate novel genetic elements associated with H. pylori AR.

Results: We performed whole genome sequencing and antimicrobial susceptibility testing of 112 H. pylori strains isolated from gastric biopsy specimens from Shanghai patients with different gastric diseases. No strains were resistant to amoxicillin. Levofloxacin, metronidazole and clarithromycin resistance was observed in 39 (34.8%), 73 (65.2%) and 18 (16.1%) strains, respectively. There was no association between gastroscopy diagnosis and resistance phenotypes. We reported the presence or absence of several subsystem protein coding genes including hopE, hofF, spaB, cagY and pflA, and a combination of CRISPRs, which were potentially correlated with resistance phenotypes. The H. pylori strains were also annotated for 80 genome-wide AR genes (ARGs). A genome-wide ARG analysis was performed for the three antibiotics by correlating the phenotypes with the genetic variants, which identified the well-known intrinsic mutations conferring resistance to levofloxacin (N87T/I and/or D91G/Y mutations in gyrA), metronidazole (I38V mutation in fdxB), and clarithromycin (A2143G and/or A2142G mutations in 23S rRNA), and added 174 novel variations, including 23 non-synonymous SNPs and 48 frameshift Indels that were significantly enriched in either the antibiotic-resistant or antibiotic-susceptible bacterial populations. The variant-level linkage disequilibrium analysis highlighted variations in a protease Lon with strong co-occurring correlation with a series of resistance-associated variants.

Conclusion: Our study revealed multidrug antibiotic resistance in H. pylori strains from Shanghai, which was characterized by high metronidazole and moderate levofloxacin resistance, and identified specific genomic characteristics in relation to H. pylori AR. Continued surveillance of H. pylori AR in Shanghai is warranted in order to establish appropriate eradication treatment regimens for this population.

Keywords: Antibiotic resistance; CRISPRs; Genetic characteristics; Helicobacter pylori; Outer membrane protein; Variation; Virulence.

Conflict of interest statement

The authors declare that they have no competing interests.

© 2022. The Author(s).

Figures

Fig. 1

Genome-wide gene family analysis and functional classification of unique genes of antibiotic resistant categories. Venn diagrams showing the number of common and unique genes in MTZ- (A), CLA- (B) and LEV- (C) resistant (yellow) and susceptible (green) categories. The number in the parentheses indicates the total number of genes identified in each category. Pie diagrams revealing the status of unique genes of MTZ-R (D), CLA-R (E) and LEV-R (F) categories in NCBI-NR, SwissProt and COG databases

Fig. 2

Analysis of OMP family genes in resistant and susceptible categories. Violin plots showing the distribution of the OMP genes in resistant and susceptible categories. The average presence levels of the total 62 OMP genes (A), the major OMP family genes along with two subfamilies genes (B), Hof family genes (C), Hom family genes (D), iron-regulated OMPs genes (F) and efflux pump OMP genes (F) in resistant and susceptible categories were compared, irrespective of the number of alleles of the genes. Solid lines indicate median levels and dotted lines indicate quartiles levels. G Proportion of strains with each OMP family gene in resistant and susceptible categories. Only genes with greater than 90% coverage were labelled as present. The number of strains containing the gene were listed on the right. A–G *P < 0.05, ***P < 0.001. Where there were no statistical analysis results labeled, there were no significant difference

Fig. 3

Analysis of virulence-associated genes in resistant and susceptible categories. Violin plots showing the distribution of the virulence-associated genes in resistant and susceptible categories. Comparisons of the average presence levels of the total virulence-associated genes (A), cagPAI genes (B) and flagellar genes (C) between resistant and susceptible categories were conducted, irrespective of the number of alleles of the genes. Solid lines indicate median levels and dotted lines indicate quartiles levels. Where there were no statistical analysis results labeled, there were no significant difference. D Two virulence-associated genes with significant different frequencies in MTZ-R and MTZ-S categories. *P < 0.05

Fig. 4

Analysis of efflux pump genes in resistant and susceptible categories. Proportion of strains with each efflux pump gene in resistant and susceptible categories. The number of strains containing the gene were listed on the right. *P < 0.05, ***P < 0.001. Where there were no statistical analysis results labeled, there were no significant difference

Fig. 5

Distribution and constitution of CRISPRs in resistant and susceptible categories and a previously unknown CRISPRs combination present exclusively in seven MTZ mono-R strains. The proportions of strains with 0, 1, 2, 3 and 4 CRISPRs in resistant and susceptible categories of MTZ (A), CLA (B) and LEV (C), respectively. **P < 0.01, n.s. not significant. The proportions of CRISPRs with 1, 2, 3, 4 and 6 spacers in resistant and susceptible categories of MTZ (D), CLA (E) and LEV (F), respectively. *P < 0.05, n.s. not significant. G The structural analysis of a previously unknown CRISPRs combination presented exclusively in seven MTZ mono-R strains, including DRs (purple, yellow and red diamonds in each CRISPR) and spacers (grey, blue and black rectangles in each CRISPR). The base sequences are shown below the CRISPR array. The DRs and spacers in the colored characters correspond with the colors of the respective diamonds and rectangles. The DR sequence in box, which exclusively presented in nine MTZ-R strains, indicated that there was a significant difference of the DR frequencies between MTZ-R and MTZ-S categories. *P < 0.05

Fig. 6

Phylogenetic analysis with respect to antibiotic resistance patterns and specific genetic loci profiles in H. pylori strains. Maximum likelihood phylogenetic tree based on whole-genome level single-copy genes of the 112 H. pylori strains, the resistance patterns, and the significantly different genomic loci observed above between susceptible and resistant categories for the corresponding antibiotic. The susceptible and resistant patterns are denoted by blue and red rectangles, respectively. The presence and absence of the specific loci are denoted by green and grey rectangles, respectively

Fig. 7

Synoptic representation of the genetic variations in the genome-wide resistance genes and linkage disequilibrium analysis. A The densities of total SNPs and Indels identified in the 112 H. pylori strains against H. pylori 26695 genome are overviewed. The numbers next to the rulers represent the number of SNPs and Indels per unit length of H. pylori 26695 genome. B The proportions of the strains with the nsSNPs and fsIndels present in the H. pylori genome-wide resistant genes in the resistant and susceptible categories of MTZ, CLA and LEV were analyzed, respectively. Only nsSNPs and fsIndels with significantly higher (red cells) or lower (blue cells) frequencies in the resistant categories of MTZ, CLA and LEV were displayed. *P < 0.05, **P < 0.01, ***P < 0.001. Conversely, grey cells indicate the variations not significantly associated with resistance or susceptibility to the antibiotics. Loci, genes, and variations are reported in rows and antibiotics in columns. The nature of the variations is color-coded on the left. The right panel represents the extent of linkage disequilibrium between variations

Fig. 8

Localization of the nsSNPs and fsIndels in the four proteins with the largest number of resistance-associated variations within the H. pylori genome-wide resistant genes. The available crystal structure of each protein was used to indicate the sites with amino acid changes. A Lon, B BabB, C XerD, and D TrpS. The color scale represents the corresponding antibiotics of the resistance or susceptibility association

Fig. 9

Cartoon model showing the putative genetic characteristics likely related to H. pylori resistance- or susceptibility- to MTZ, CLA and LEV at locus- and variant-levels. The intrinsic resistance or resistance-associated compensation related genetic features involves the variations in the genome-wide resistant gene-encoding proteins including antibiotic targets, translation-associated factors, transporters, OMPs, proteins probably related to biofilm formation, and other sporadic factors including lon protease with variations in linkage disequilibrium with many others, as well as the presence or absence of a porin, and an efflux pump transporter. In the underlying acquired resistance related genetic elements, the presence or absence of a virulence-associated factor belonging to ComB T4SS and a certain CRISPRs, as well as variations in several recombination required proteins, were demonstrated. Different colors of the genetic element names denote they are in relation to resistance or susceptibility of different antibiotics. Red (MTZ), gold (CLA), pink (LEV), orange (MTZ and CLA), blue (MTZ and LEV), brown (CLA and LEV)