Chapter III. Metagenomic analysis of isolation methods of a targeted
3.5. Conclusions
128
129
effectively, enabling further analyses, which include epidemiological investigation and molecular characterization of strains.
This study provides a new perspective and possibilities for applying metagenomics in microbiological research, such as improving diagnostic methods for fastidious bacteria that are difficult to isolate.
130 Table 7. Information of samples and isolation results of C. jejuni. Enrichment type No enrichment Bolton broth Preston broth Ratio - 1:101:1001:1000 1:101:1001:1000 Selective media mCCDA Preston agar mCCDA Preston agar mCCDAPreston AgarmCCDAPreston agar mCCDA Preston agar mCCDAPreston agar mCCDA Sample
1 CJ CJ CJ 2 CJ 3 CJ 4 CJ 5 CJ CJ CJ 6 CJ CJ CJ CJ 7 CJ CJ 8 CJ CJ CJ CJ 9 CJ CJ CJ CJ 10CJ CJ CJ 11CJ CJ CJ 12CJ CJ CJ 13CJ CJ CJ CJ 14CJ CJ CJ CJ 15CJ 16CJ CJ 17CJ CJ CJ 18CJ 19CJ CJ CJ CJ 20CJ CJ 21CJ CJ 22CJ CJ 23CJ CJ CJ 24CJ CJ CJ CJ 25CJ CJ 26 CJ 27 CJ
131
28 CJ CJ 29CJ 30CJ CJ CJ 31 CJ 32 CJ 33 CJ 34CJ CJ 35CJ CJ Total 17/35 (48.6%)19/35 (54.3%)0/35 (0%)0/35 (0%)0/35 (0%)0/35 (0%)0/35 (0%)0/35 (0%)0/35 (0%)0/35 (0%)11/35 (31.4%)0/35 (0%)34/35 (97.1%)2/35 (5.7%) CJ: positive for C. jejuni in culture-based results, Blank: negative for C. jejuni in culture-based results. Gray color: samples used for microbia community analysis. A total of 54 samples (at least seven samples per process) were used for microbial community analysis.
132 Table 8. Primer list for polymerase chain reaction (PCR), quantitative PCR, and bacterial DNA amplification in this study Target microbe TargetSizePrimer Sequence (5’-3’) PCR C. jejuni hipO323 bpCJ-F ACTTCTTTATTGCTTGCTGC CJ-R GCCACAACAAGTAAAGAAGC C. coli glyA126 bpCC-F GTAAAACCAAAGCTTATCGTG CC-RTCCAGCAATGTGTGCAATG Campylobacter spp.23S rRNA 650 bp23s-F TATACCGGTAAGGAGTGCTGGAG 23s-R ATCAATTAACCTTCGAGCACCG E. coliMalB promoter585 bpEco-F GACCTCGGTTTAGTTCACAGA Eco-RCACACGCTGACGCTGACCA E. faecium- 658 bpmFM-F TTGAGGCAGACCAGATTGACG mFM-R TATGACAGCGACTCCGATTCC E. faecalis - 941 bpmFL-FATCAAGTACAGTTAGTCTTTATTAG mFL-RACGATTCAAAGCTAACTGAATCAGT Enterococcus spp. 16S rRNA 320bp mENT-FGGATTAGATACCCTGGTAGTCC mENT-R TCGTTGCGGGACTTAACCCAAC Bacterial DNA amplification 16S V3-V4 region V3-V4 region MiSeq 341F TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG Miseq 805RGTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAAT Quatitative PCRC. jejuni hipO123bp qCJ-F AATGCACAAATTTGCCTTATAAAAGC qCJ-R TNCCATTAAAATTCTGACTTGCTAAATA probe qCJ-probe FAM-ACATACTACTTCTTTATTGCTTG-BHQ1
133 Table 9. Culture-dependent isolation results of Campylobacter jejuni and competing microbes in each procedures. Enrichment type No enrichment Bolton brothPreston broth Ratio - 1:101 1:102 1:103 1:101 1:102 1:103 Selective mediamCCDA Preston agarmCCDAPreston agarmCCDAPreston agarmCCDAPreston agarmCCDA Preston agarmCCDAPreston agarmCCDAPreston agar Target & competing microbes
C. jejuni48.6% (17/35) 54.3% (19/35) - - - - - - - -
ac 31.4% (11/35) -
abc 97.1% (34/35) 5.7% (2/35) ESBLs producing E. coli
a 48.6% (17/35) 8.6% (3/35)
ac 80.0% (28/35) -
ac 97.1% (34/35) 2.9% (1/35)
ac 82.9% (29/35)
bc 54.3% (19/35)
a 48.6% (17/35) -
a 42.9% (15/35) -
b 11.4% (4/35) - P. mirabilis42.9% (15/35)
a 88.6% (31/35) 45.7% (16/35)
a 94.3% (33/35)71.4% (24/35) 91.4% (32/35)
c 62.9% (22/35) 82.9% (29/35) 65.7% (23/35) 94.3% (33/35)54.3% (19/35)
a 80.0% (28/35)
b 28.6% (10/35)
a 97.1% (34/35) Enterococcus spp.
a 22.9% (8/35) 2.9% (1/35) - - - - - -
ac 20.0% (7/35) 2.9% (1/35)
ac 31.4% (11/35) 2.9% (1/35) 17.1% (6/35) 8.6% (3/35) A total of 14 procedures were applied for all 35 fecal samples, which were combinations of different enrichment type (without enrichment process or enriched in Bolton broth or Preston broth), ratio of sample-to-enrichment broth (1:101 , 1:102 , or 1:103 ), and selective media (mCCDA or Preston agar). a : Significant differences according to selective media (enriched in the same enrichment broth and the same ratio of sample-to- broth),b : Significant differences according to the ratio of sample-to-broth (enriched in the same enrichment broth and same selective media), Significant differences according to the type of enrichment broth (enriched in the same ratio of sample-to-broth and the same type of selective media), -: negative result (0%).
134
Figure 15. A schematic diagram of the isolation method of C. jejuni.
A schematic diagram of the current study. All chicken fecal samples were analyzed to compare isolation methods of C. jejuni using culture independent
135
sequence-based metagenomics and culture-dependent tools. Metagenomic analysis was performed to investigate the effects of seven enrichment procedures. Six procedures involved enrichment processes (Bolton and Preston broth at 1:101, 1:102, and 1:103, respectively), and one procedure did not. In culture-dependent analysis, a total of 14 procedures were applied for all samples, which were combinations of different enrichment broths (without enrichment process or enriched in Bolton broth or Preston broth), ratio of sample-to-enrichment broth (1:101, 1:102, or 1:103), and selective media (mCCDA or Preston agar).
a The ratio of fecal sample-to-enrichment broth
136
137
Figure 16. Proportion of competing microbes according to isolation procedures of C. jejuni based on culture-dependent tools.
Proportion of A) ESBL-producing E. coli and B) P. mirabilis among competing microbes according to the isolation procedure of C. jejuni. The proportion of ESBL-producing E. coli was significantly different according to the isolation procedure including type of enrichment broth, selective media, and combination of different enrichment broths, ratio of sample-to-enrichment broth, and selective media. The proportion of P. mirabilis was significantly different according to the isolation procedure including type of selective media and combination of different enrichment broths, ratio of sample to enrichment broth, and selective agars.
*: p <0.05, **: p <0.01
138
139
Figure 17. Evaluating the effect of competing flora on the isolation of C jejuni using culture-dependent tools.
The proportion of A) ESBL-producing E. coli and B) P. mirabilis according to the isolation results of C jejuni. The proportion of ESBL-producing E. coli and P. mirabilis was significantly higher in C. jejuni-negative fecal samples.
*: p <0.05, **: p <0.01
140
141
Figure 18. Alpha diversity of each seven enrichment procedures using metagenomics tools.
A) The number of observed operational taxonomic units (OTUs), B) Shannon diversity index, and C) Faith’s phylogenetic diversity in each seven procedures.
The number of OTUs, microbial abundance and evenness, and microbial richness of fecal samples significantly decreased in the enrichment process as the ratio of sample-to- enrichment broth decreased.
*: p <0.05, **: p <0.01
142
143
Figure 19. Similarity of the microbial community of fecal samples in each procedure.
A) Principal coordinates analysis (PCoA) plot of fecal samples in seven procedures based on weighted UniFrac distance metrics and B) Heatmap of
144
fecal samples in seven procedures. Fecal samples enriched in Preston broth at 1:103 ratio were clustered together.
145
146
147
Figure 20. Microbial community of fecal samples in each procedure at the genus level.
A) Taxonomic composition and B) relative abundance of Campylobacter, Escherichia, and Proteus of fecal samples in each procedure. The relative abundance of Campylobacter in fecal samples in 103–Preston broth was significantly higher than in other procedures. In the case of Escherichia- Shigella, the relative abundance was reduced, compared to feces, but was still present in large proportions in the process of 101-/102-Bolton broth enrichment.
In the case of Proteus, the relative abundance increased during the enrichment process in both enrichment broths, at all ratios of sample-to- broth, with the exception of 103-Bolton broth.
148
149
150
Figure 21. Bacterial taxa that are differentially abundant in microbial community of fecal samples in each procedure.
A) Linear discriminant analysis effect size (LEfSe) and B) taxonomic cladogram between fecal samples enriched in Preston broth at the 1:103 ratio and not undergoing the enrichment process. C) LEfSe and D) taxonomic cladogram between fecal samples enriched in Preston broth and Bolton broth at the 1:103 ratio. The logarithmic linear discriminant analysis score cut-off was set to 2.0. The relative abundance of Campylobacter in fecal samples in 103- Preston broth was significantly higher than in other procedures, while the relative abundance of Escherichia-Shigella was significantly lower than in other procedures.
151
152
Figure 22. Relationship between microorganisms in microbial community of fecal samples.
Correlation plot in A) Bolton broth and B) Preston broth regardless of the ratio of sample- to-enrichment broth. Campylobacter was negatively correlated with Proteus in Bolton broth, while Campylobacter was negatively correlated with Escherichia-Shigella in Preston broth.
153
154
Figure 23. The relative abundance of each microbe according to the isolation results of C. jejuni.
The relative abundance of A) Campylobacter, B) Escherichia, and C) Proteus according to the isolation results of C. jejuni based on culture-dependent tools.
The relative abundance of Campylobacter was significantly higher in the fecal sample from which C. jejuni was isolated, while, the relative abundance of Escherichia was significantly lower. There was no difference in the relative abundance of Proteus in the fecal sample according to isolation result of C.
jejuni.
*: p <0.05, **: p <0.01
155
Figure 24. The correlation between colony forming units and cycle threshold (Ct) values of C. jejuni standard strains (NCTC 11168 and ATCC 33560).
156
157
Figure 25. Quantitative PCR of C. jejuni applied to all 54 samples.
A) The amount of C. jejuni in each procedure inferred from standard curve and B) correlation of metagenomics data and quantitative polymerase chain reaction (qPCR) results. The amount of C. jejuni was significantly higher in fecal samples in 103-Preston broth, 1.0×109.4–1.0×1011.4. In addition, the cycle threshold (Ct) values from qPCR results and the number of Campylobacter reads from metagenomics results had a high correlation.
158
Chapter IV.
The wild mouse (Micromys minutus): reservoir of
a novel Campylobacter jejuni strain
159
Abstract
Background: Campylobacter jejuni is one of the most common zoonotic pathogens worldwide. Although the main sources of human C. jejuni infection are livestock, wildlife can also affect C. jejuni transmission in humans.
However, it remains unclear whether wild mice harbor C. jejuni and are involved in the ‘environment–wildlife–livestock–human’ transmission cycle of C. jejuni in humans. Here, we characterized C. jejuni from wild mice and identified genetic traces of wild mouse-derived C. jejuni in other hosts using a traditional approach, comparative genomics, and sequence-based approach.
Methods and Results: We captured 115 wild mice (49 Mus musculus and 66 Micromys minutus) without any clinical symptoms from 22 sesame fields in Korea over 2 years. Among them, Micromys minutus were typically caught in remote areas of human houses and C. jejuni was solely isolated from Micromys minutus (42/66, 63.6%). We identified a single clone (MLST ST- 8388) in all 42 C. jejuni isolates, which had not been previously reported, and all isolates had the same virulence/survival-factor profile, except for the plasmid-mediated virB11 gene. No isolates exhibited antibiotic resistance, either in phenotypic and genetic terms. Comparative-genomic analysis and MST revealed that C. jejuni derived from Micromys minutus (strain SCJK2) was not genetically related to those derived from other sources (registered in the NCBI genome database and PubMLST database). Furthermore, Micromys
160
minutus-derived C. jejuni had different ancestral lineages from those derived from other sources, and there was a low chance of C. jejuni transmission from Micromys minutus to humans/livestock because of their habitat. In gut microbiota analysis, there was no Campylobacter in the gut of all Mus musculus, while in the gut of all Micromys minutus, Campylobacter was present in high proportions, regardless of the culture-based isolation results.
Conclusions: In conclusion, Micromys minutus may be a potential reservoir for a novel C. jejuni, which is genetically different from those of other sources, but may not be involved in the transmission of C. jejuni to other hosts, including humans and livestock. This study could form the basis for further studies focused on understanding the transmission cycle of C. jejuni, as well as other zoonotic pathogens originating from wild mice.
Keywords: Campylobacter jejuni, Wild mouse, Transmission cycle, Whole genome sequencing, Comparative genomic analysis, Microbial community analysis
161
4.1. Introduction
Campylobacter jejuni is one of the most common zoonotic pathogens worldwide, and human infections with this pathogen have increased in both developing and developed countries (Kaakoush et al., 2015). In particular, C.
jejuni directly causes gastrointestinal diseases such as diarrhea and abdominal pain, and it can also lead to late complications such as neurological diseases, including Guillain–Barre and Miller–Fisher syndromes in humans (Nachamkin et al., 1998). The main sources of C. jejuni transmission in humans are livestock, including poultry and cattle, and humans in particular are easily exposed to C.
jejuni during the handling or ingestion of poultry (Bronowski et al., 2014).
Environmental sources are also reservoirs for C. jejuni infections in humans, and transmission through contaminated water and soil is prevalent (Gras et al., 2012). In addition, wildlife species can serve as potential reservoirs for C. jejuni infection in humans (Waldenström et al., 2007). Therefore, humans, livestock, environmental sources, and wildlife form complex interactions that contribute to C. jejuni infection and constitute a transmission cycle. Studies of C. jejuni in various reservoirs are needed to better understand the potential for C. jejuni transmission and the impact that such transmission could have on human health.
Studies on C. jejuni in wildlife have been conducted mainly with wild birds, since they can contribute to C. jejuni infections in humans and livestock (French et al., 2009; Weis et al., 2016). In addition to research on wild birds, C.
162
jejuni infections in wild mice are also important to study, since wild mice are prevalent, come in close contact with humans, and are known carriers of various pathogens (Kruse et al., 2004; Han et al., 2015). However, few studies have examined C. jejuni infection in wild mice, and no studies evaluating the effects of wild mice on human C. jejuni infection have been reported. Only one study provided an estimate of C. jejuni-transmission events between rodents and pigs on the same farm (Meerburg et al., 2006). Thus, the transmission of C. jejuni between wild mice and other hosts remains unclear, and no studies have shown that wild mice (similarly to wild birds) contribute to human C. jejuni infection (Meerburg and Kijlstra, 2007).
However, it is difficult to completely detect C. jejuni using common culture methods because of its specific growth conditions (i.e. requires a microaerophilic environment) and the presence of viable but nonculturable (VBNC) (Ugarte‐Ruiz et al., 2012). Currently, advances in next-generation sequencing (NGS) technologies have made bacterial identification and detection more accurate (Kim et al., 2019). These NGS technologies, such as Illumina Genome Analyzer (HiSeq, MiSeq), Applied Biosystems SOLiD System, Life Technologies Ion Torrent, and the PacBio RX system, can be used to understand the microbial community in the gastro-intestinal tract of host and to determine the distribution or proportion of each microbe, ultimately determining whether it is a commensal or transient flora (Huang et al., 2016).
163
The objective of this study was to determine whether wild mice are involved in the ‘environment–wildlife–livestock–human’ transmission cycle of C. jejuni in human infections. To characterize C. jejuni from wild mice, C.
jejuni was isolated from wild mice (Mus musculus and Micromys minutus) in 22 sesame fields over the course of two years. The clonal distribution of C.
jejuni isolates from these wild mice was identified, and the virulence/survival- factor profiles and antibiotic-resistance patterns of all isolates were determined.
To investigate a genetic trace of C. jejuni transmission between wild mice and other hosts, the genetic relatedness of C. jejuni between wild mice and other sources was compared using comparative-genomic analysis. Furthermore, gut microbiota of wild mice (Mus musculus and Micromys minutus) was analyzed to investigate the distribution and ratio of Campylobacter in the gut of wild mouse and to determine the actual presence of Campylobacter in wild mice whose culture method cannot confirm the presence of Campylobacter.
164
4.2. Materials and Methods
Isolation of C. jejuni
This study was reviewed and approved by the Institutional Animal Care and Use Committee of Hallym University (approval number Hallym2017-5).
Wild mice, including Mus musculus and Micromys minutus, were captured under piles of dry sesame plants in 22 sesame fields that were separated by rivers, roads, and mountains in Hongcheon and Chuncheon (Gangwon Province, Korea) over the course of two years. The global positioning system (GPS) coordinates for the regions where the sesame fields were located are shown in Figure 26. All captured wild mice were immediately transported to the laboratory, and each wild mouse was transferred to a single disinfected cage.
Fresh fecal samples were collected within 10 min and stored at 4℃.
Subsequently, pathological signs or lesions were investigated for each wild mouse.
Within 3 h after fecal sampling, we began isolating C. jejuni from the feces of the wild mice. Fecal samples collected from each wild mouse were homogenized in phosphate-buffered saline. The homogenized contents were directly spread onto modified charcoal–cefoperazone–deoxycholate agar plate (mCCDA; Oxoid, Ltd., Hampshire, UK) with a CCDA-selective supplement (Oxoid, Ltd.). Next, all plates were incubated at 42℃ for 2 days under 83% N2, 7% CO2, 4% H2, and 6% O2 (Mace et al., 2015) in micro-aerobic jars using
165
Anoxomat Mark II (MART microbiology B.V., Lichtenvoorde, Netherland).
Next, colonies suspected to be C. jejuni were transferred to Müller–Hinton agar (Oxoid Ltd.), and genomic DNA was extracted from each colony using the boiling method. Briefly, bacterial cells were suspended in 200 μL sterile distilled water, boiled for 10 min, and centrifuged at 13,000 × g for 3 min. The resulting supernatants were used as a template for polymerase chain reaction (PCR) experiments. C. jejuni was confirmed by running polymerase chain reaction (PCR) experiments (Table 10) (Wang et al., 2002). Then, to identify false negatives for C. jejuni, an isolation method with an enrichment process was applied. Briefly, the fecal samples that were negative for C. jejuni were enriched in Bolton broth (Oxoid, Ltd.) with a Bolton broth-selective supplement (Oxoid, Ltd.), and the enrichment broths were incubated at 42°C for 48 h under 83% N2, 7% CO2, 4% H2, and 6% O2 in micro-aerobic jars using Anoxomat Mark II (MART microbiology B.V., Lichtenvoorde, Netherland).
Subsequently, the presence of C. jejuni in these samples was confirmed following the same procedure described above. Finally, the chi-square test was performed to compare the isolation rates of C. jejuni by the gender/age of the mice and the region/year in which mice were captured.
Multilocus sequence typing (MLST) of Micromys minutus-
derived C. jejuni
166
MLST was performed on all C. jejuni isolates from wild mice, according to the PubMLST protocol (https://pubmlst.org/). Briefly, seven housekeeping genes (aspartase A, glutamine synthetase, citrate synthase, serine hydroxymethyl transferase, phosphoglucomutase, transketolase, and ATP synthase alpha subunit) were amplified and sequenced with primer sets described in the PubMLST protocol. Amplification and sequencing were performed using Lamp Taq DNA Polymerase (BIOFACT, Korea) and an ABI 3730XL DNA analyzer (Applied Biosystems, USA), respectively. Furthermore, a minimum-spanning tree based on the allelic MLST profiles of the current study (ST-8388) and other sequence types (STs) registered in PubMLST (accessed on 20 November 2019, about 6,800 STs from C. jejuni) was generated using the PHYLOViZ software (http://www.phyloviz.net/) (Francisco et al., 2012).
Profiling virulence/survival factors and antibiotic- resistance patterns of Micromys minutus-derived C. jejuni
For all C. jejuni isolates, the presence and absence of virulence/survival-related genes were confirmed by PCR (Table 10) (Konkel et al., 1999; Datta et al., 2003; Müller et al., 2006; Bui et al., 2012; González-Hein et al., 2013; Harrison et al., 2014; Koolman et al., 2015; An et al., 2018). These genes corresponded to three secretion systems (flhB, virB11, and hcp), one
167
adherence/colonization-related protein (cadF), two cell invasion-related proteins (pldA and iamA), one cytotoxin related protein (cdtB), and three survival-related factors (crsA, perR, and htrA). The minimum inhibitory concentration (MIC; for erythromycin, chloramphenicol, ciprofloxacin, tetracycline, telithromycin, gentamicin, azithromycin, streptomycin, and nalidixic acid), disk diffusion (for ciprofloxacin, tetracycline, and nalidixic acid), and sequence-based tests (for gyrA and tetO genes) were conducted for all C. jejuni isolates to evaluate antibiotic resistance (Table 10) (Zirnstein et al., 1999; Gibreel et al., 2004).
Whole-genome sequencing of representative Micromys minutus-derived C. jejuni strains
A representative strain of wild mouse-derived C. jejuni (strain SCJK2) was selected, and genomic DNA was extracted using the MGTM Genomic DNA Purification Kit (Macrogen, Korea). Whole-genome sequencing was performed using a Pacific Biosciences RS Ⅱ sequencer (Pacific Biosciences, Menlo Park, CA, USA). De novo microbial genome assemblies were carried out using Hierarchical Genome Assembly Process, version 3.0 (Chin et al., 2013), and all contigs were circularized using Circlator 1.4.0 (Sanger institute, UK). The
NCBI prokaryotic genome annotation pipeline
(https://www.ncbi.nlm.nih.gov/genome/annotation_prok/) and the EzBioCloud
168
genome database (https://www.ezbiocloud.net/) were used for gene annotation.
Functional classification was performed using the EggNOG (Powell et al., 2014), SwissProt (Pundir et al., 2015), KEGG (Kanehisa et al., 2015), and SEED (Overbeek et al., 2005) databases. Antibiotic resistance- and virulence- related genes of wild mouse-derived C. jejuni (strain SCJK2) sequence were analyzed using the comprehensive antibiotic resistance database (CARD) (Jia et al., 2016) and virulence factors of pathogenic bacteria database (VFDB) (Chen et al., 2015), respectively.
Comparative-genomic analysis of Micromys minutus- derived C. jejuni with previously reported isolates
For comparative genomic analysis, 174 complete genome sequences of C. jejuni from various sources, including humans, poultry, bovines, and sheep, as well as environmental isolates, were downloaded from the NCBI Genome database (https://www.ncbi.nlm.nih.gov/genome/; Table 11). For phylogenetic characterization of the wild mouse-derived C. jejuni (strain SCJK2) sequence with other sequences obtained from the NCBI Genome database, functional annotation of all 175 genomes was performed using Prokka (Seemann, 2014), and then core genes (genes possessed by more than 95% genome) and accessory genes (genes possessed by less than 95% strains) were identified using Roary (Page et al., 2015). Multiple alignments were performed on core genes and