Detection of Campylobacter jejuni diversity by clustered regularly interspaced short palindromic repeats (CRISPR) from an animal farm

Abstract Background Campylobacter jejuni is the leading bacterial pathogen that causes foodborne illness worldwide. Because of genetic diversity and sophisticated growth requirements of C. jejuni, several genotyping methods have been investigated to classify this bacterium during the outbreaks. One of such method is to use clustered regularly interspaced short palindromic repeats (CRISPR). Objectives The goal of this study was to explore the diversity of C. jejuni isolates with CRISPR from an animal farm. Methods Seventy‐seven C. jejuni isolates from an animal farm were used in this study. The day‐old broilers were reared with other poultry and farm animals, including layer hens, guinea hens, dairy goats and sheep. A small swine herd was also present on an adjacent, but separate plot of land. Isolation and identification of C. jejuni were performed according to the standard procedures. The CRISPR type 1 was PCR amplified from genomic DNA, and the amplicons were sequenced by the Sanger dideoxy method. The direct repeats (DRs) and spacers of the CRISPR sequences were identified using the CRISPRFinder. Results The CRISPR sequences were detected in all 77 isolates. One type of DRs was identified in these 77 isolates. The lengths of the CRISPR locus ranged from 100 to 560 nucleotides, whereas the number of spacers ranged from one to eight. The distributions of the numbers of CRISPR spacers from different sources seemed to be random. Overall, 17 out of 77 (22%) C. jejuni isolates had two and five spacers, whereas 14 out of 77 (18%) isolates had three spaces in their genomes. By further analysis of spacer sequences, a total of 266 spacer sequences were identified in 77 C. jejuni isolates. By comparison with known published spacer sequences, we observed that 49 sequences were unique in this study. The CRISPR sequence combination of Nos. 16, 19, 48 and 57 was found among a total of 15 C. jejuni isolates containing various multi‐locus sequence typing (MLST) types (ST‐50, ST‐607, ST‐2231 and ST‐5602). No. 57 spacer sequence was unique from this study, whereas the other three (Nos. 16, 19 and 48) sequences were found in previous reports. Combination of Nos. 5, 9, 15, 30 and 45 was associated with ST‐353. To compare the CRISPR genotyping with other methods, the MLST was selected due to its high discriminatory power to differentiate isolates. Based on calculation of the Simpson's index of diversity, a combination of both methods had higher Simpson's index value than those for CRISPR or MLST, respectively. Conclusions Our results suggest that the MLST from C. jejuni isolates can be discriminated based on the CRISPR unique spacer sequences and the numbers of spacers. In the future, investigation on the CRISPR resolution for C. jejuni identification in outbreaks is needed. A database that integrates both MLST sequences and CRISPR sequences and is searchable is greatly in demand for tracking outbreaks and evolution of this bacterium.

were found in previous reports. Combination of Nos. 5,9,15,30 and 45 was associated with ST-353. To compare the CRISPR genotyping with other methods, the MLST was selected due to its high discriminatory power to differentiate isolates. Based on calculation of the Simpson's index of diversity, a combination of both methods had higher Simpson's index value than those for CRISPR or MLST, respectively.

Conclusions:
Our results suggest that the MLST from C. jejuni isolates can be discriminated based on the CRISPR unique spacer sequences and the numbers of spacers. In the future, investigation on the CRISPR resolution for C. jejuni identification in outbreaks is needed. A database that integrates both MLST sequences and CRISPR sequences and is searchable is greatly in demand for tracking outbreaks and evolution of this bacterium.

K E Y W O R D S
Campylobacter jejuni, clustered regularly interspaced short palindromic repeats, CRISPR

INTRODUCTION
Clustered regularly interspaced short palindromic repeats (CRISPR) was first described by Ishino et al. (1987) that the highly unusual repetitive sequences of 29 nucleotides (as direct repeats [DRs]) were regularly spaced with 32 nucleotides (as spacers) during their study on the Escherichia coli iap gene. Since then, this similar pattern has been found in genomes of many archaea and prokaryotes (Ishino et al., 2018). The striking characteristics of the CRISPR pattern are as follows: (1) DRs interspaced with various numbers of unique, non-repetitive sequences (so-called spacers), (2) a leader sequence at the one side of the locus acting as a promoter and (3) various numbers of the cas family genes (CRISPR-associated genes) (Grissa et al., 2009;Ishino et al., 2018). The CRISPR-Cas systems in prokaryotes and archaea play important roles in defence of infecting bacteriophages and plasmids (Barrangou and Horvath, 2009;de Cardenas et al., 2015;Louwen et al., 2014). Also, these systems may act as virulence factors during bacterial pathogenesis (Ahmed et al., 2018;Hille et al., 2018). Recently, the CRISPR has been used for genotyping foodborne pathogen Salmonella to track outbreaks and evolution (e.g., Shariat and Dudley, 2014;Cox et al., 2019).
Campylobacter jejuni, a Gram-negative bacterium (Ryan et al., 2004;Ursing et al., 1994), is the leading foodborne pathogen worldwide (Kirk et al., 2015;Tack et al., 2019). It is estimated that this bacterium causes about 1.3 million cases of human campylobacteriosis in the U.S. annually (Crim et al., 2015;Scallan et al., 2011). The reservoirs of C. jejuni are found in guts of many animals where this bacterium is regarded as a member of gut microbiomes (Hermans et al., 2012;European Food Safety Authority, 2010). Therefore, control of this bacterium is extremely difficult (Lin, 2009;Sahin et al., 2015). In addition, because of the genetic diversity and sophisticated growth condition of C. jejuni, detection and identification of this bacterium with classic culture methods are problematic (On et al., 2003). Many genotyping methods for this bacterium have been investigated to solve these problems.
In this short communication, we applied the CRISPR typing to explore the diversity of C. jejuni isolates from an animal farm in 2016.

Bacterial cultures and genomic DNA isolation
Seventy-seven C. jejuni isolates from 2016 and an animal farm were used in this study and are listed in Table S1 (Rothrock et al., 2019).
Briefly, the farm was about 3 acres in size. The day-old broilers were transported to the farm, where other poultry and farm animals were also reared, including layer hens, guinea hens, dairy goats and sheep. A small swine herd was also present on an adjacent, but separate plot of land (Rothrock et al., 2019). Fresh fecal samples were collected from the pen area. At the same time, any fecal samples from other animal species surrounding the broiler area on the farm were also collected.
Cecal samples were collected after exsanguination. Rinsates were generated by rinsing the carcasses with 100 ml of 10 mM phosphatebuffered saline in sterile individual bags. All samples were placed on ice at the farm and transported to the laboratory. In the laboratory, isolation and identification of C. jejuni were performed according to the standard procedures. Bacterial cultures stored in 15% glycerol at -80 • C were revived in Müeller-Hinton agar plates at 42 • C for 48 h under microaerobic conditions as described previously (Hiett et al., 2008;Yeh et al., 2013).
Genomic DNA was isolated using a DNeasy Blood and Tissue kit (Qiagen Inc., Germantown, MD, USA) according to the manufacturer's instructions. The quality and quantity of genomic DNA were determined by agarose gel electrophoresis and a spectrophotometer (DS-11 FX spectrophotometer; DeNovix Inc., Wilmington, DE, USA), respectively. The DNA in 10 mM Tris-HCl (pH 8.0) was stored at -80 • C.

PCR amplification of C. jejuni CRISPR and sequencing
The PCR primers and conditions for amplification of the C. jejuni where Big Dye terminator chemistry on an ABI 3100 Genetic Analyzer (Thermo Scientific, Foster City, CA, USA) was used. Sequence chromatograms were edited for quality. The same primer pairs for amplification and sequencing were as follows (Price et al. 2007): CRISPR-F 5′-GCAACCTCCTTTTAGTGGAGTAATTAG-3′ and CRISPR-R 5′-AAGCGGTTTTAGGGGATTGTAAC-3′.

Analysis of CRISPR sequences
CRISPR sequences were submitted to the CRISPR Web Server and were identified using the CRISPRFinder program (Grissa et al., 2007; https://crispr.i2bc.paris-saclay.fr/). Simpson's index of diversity was calculated based on the Hunter and Gaston equation to determine the discriminatory power of genotyping methods (Carriço et al. 2006). The sequences were deposited in GenBank, and the accession numbers are MT199732-MT199808 (Table S1).

RESULTS AND DISCUSSION
Seventy-seven PCR amplified products with various sizes were obtained and subjected for sequencing by Big Dye terminator chemistry on an ABI 3100 Genetic Analyzer. The sequences were then submitted to the CRISPRs web server (Grissa et al., 2007) to locate DRs and spacer sequences from the isolates. Grissa et al. (2007) defined the sequences consisting of at least three motifs and at least two exactly identical DRs are regarded as 'confirmed' CRISPR, whereas the remaining is considered as 'questionable' CRISPR. Our results show that all 77 isolates had either confirmed or questionable CRISPRs with various lengths (Table S1). The DRs varied in lengths, but had a consensus sequence: 5′-ATTTTACCATAAAGAAATTTAAAAAGGGACTAAAA-3′. It seems to be common that one end at the DR was not totally conserved (de Cárdenas et al., 2015). A total of 266 spacer sequences were detected in 77 isolates. The numbers of spacers of each isolate range from one spacer to eight spacers (Table S1). By analysing these spacer sequences with the MUSCLE alignment program (Edgar, 2004), we observed 67 different spacer sequences in 77 C. jejuni isolates (Table 1).
Further comparing the current space sequences with those from previ-ous reports from poultry sources, such as cecal content and neck skin of broilers, laying hens from organic farms and sewage water (de Cárdenas et al., 2015;Kovanen et al., 2014;Louwen et al., 2013), shows that 18 out of 67 space sequences were identified previously, whereas 49 sequences were unique to isolates from this study. Among 67 space sequences, four sequences (Nos. 16,19,48 and 57) were found in the same 15 isolates (Table 1). The sequence Nos. 16, 19 and 48 (Table 1) were identified in the previous report ( The multi-locus sequence typing (MLST) (Dingle et al., 2001) has been the most used for genotyping foodborne bacterial pathogens.
In our unpublished observations, we found 13 MLST sequence types (ST) in 66 isolates (Table S1) Further, the CRISPR types of our current isolates were assigned based on the spacer numbers and sequences, and Simpson's index of diversity was calculated to measure the discriminatory power of these methods. As shown in Table 2, a combination of both methods had the Simpson's index value of 0.953 that is higher than those of 0.922 or 0.849 for CRISPR or MLST, respectively. These results suggest that the MLST can further be discriminated based on CRISPR spacer sequences and the numbers of spacers. Further investigation on the CRISPR molecular variations in and the numbers of spacers to increase the resolution is needed.
In summary, C. jejuni isolates from 2016 and an animal farm were subjected to CRISPR type 1 analysis. The CRISPR sequences were identified in all 77 isolates. One type of DR was detected in the CRISPR sequences. The lengths of the CRISPR sequences ranged from 100 to 560 nucleotides. The number of spacers ranged from one to eight.
By further analysis of spacer sequences, a total of 266 sequences     were identified from 77 C. jejuni isolates. Among them, 67 distinctive sequences were identified. Furthermore, by comparison with known spacer sequences, we observed that 18 from 67 sequences were known previously and 49 sequences were unique in this study. Further analysis shows that the MLST from C. jejuni isolates can be discriminated based on CRISPR spacer sequences and the numbers of spacers. In the future, investigation on the CRISPR resolution for C. jejuni identification in outbreaks is needed. A database that integrates both MLST sequences and CRISPR sequences and is searchable is greatly in demand. During the revision of this manuscript, a study demonstrates the CRISPR-Cas system is prevalent in the fluoroquinolone-resistant C. jejuni isolates (Adiguzel et al., 2021).