• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. May 2007; 73(10): 3431–3436.
Published online Mar 30, 2007. doi:  10.1128/AEM.02702-06
PMCID: PMC1907115

High-Resolution DNA Melt Curve Analysis of the Clustered, Regularly Interspaced Short-Palindromic-Repeat Locus of Campylobacter jejuni[down-pointing small open triangle]


A novel method for genotyping the clustered, regularly interspaced short-palindromic-repeat (CRISPR) locus of Campylobacter jejuni is described. Following real-time PCR, CRISPR products were subjected to high-resolution melt (HRM) analysis, a new technology that allows precise melt profile determination of amplicons. This investigation shows that the CRISPR HRM assay provides a powerful addition to existing C. jejuni genotyping methods and emphasizes the potential of HRM for genotyping short sequence repeats in other species.

Clustered, regularly interspaced short palindromic repeats (CRISPRs) are a class of short sequence repeats that have been found in many bacterial genomes (7, 10, 11). CRISPRs comprise near-perfect direct repeats (DRs) interspersed with similarly sized nonrepetitive spacer sequences. The Campylobacter jejuni CRISPR locus harbors few DRs but extensive spacer variation (18). The aim of this study was to develop a high-resolution melt (HRM)-based assay for interrogating the hypervariable CRISPR locus of C. jejuni. It was hypothesized that HRM would be effective for discriminating CRISPR variants and that this assay would efficiently add resolution to existing single nucleotide polymorphism (SNP) and binary gene-based typing methods for C. jejuni (5, 6, 15-17, 19).

CRISPR detection in C. jejuni.

Schouls and coworkers (18) have previously demonstrated the absence of the CRISPR locus in 10% of C. jejuni and Campylobacter coli strains, with a further 15% harboring a single repeat without a spacer. However, due to polymorphisms between the original CRISPR primers and the genome-sequenced C. jejuni strain RM1221, the distribution of CRISPRs in C. jejuni and C. coli may have been underestimated. Therefore, new CRISPR primers were designed: CRISPR-For (5′-GCAACCTCCTTTTAGTGGAGTAATTAG-3′) and CRISPR-Rev (5′-AAGCGGTTTTAGGGGATTGTAAC-3′). A total of 210 Australian C. jejuni and C. coli isolates, including 181 that have been previously described (13, 15, 16), were tested for the presence of the CRISPR locus using the redesigned primers (Sigma-Proligo, Lismore, Australia) and the manufacturer's PCR reagent and thermocycling schedule (Invitrogen).

Twelve (6%) isolates did not yield a CRISPR PCR product, including five C. jejuni (2%) and seven C. coli (100%) isolates. Seventy-four (35%) strains contained a single DR lacking a spacer unit. The remaining 138 isolates contained between 2 and 11 DRs, consistent with the data set of Schouls et al. (18), in which between 1 and 13 DRs were identified. All four sequence type 42 (ST-42) isolates were CRISPR negative, in agreement with the study by Schouls and colleagues in which six of seven ST-42 isolates were absent for the CRISPR locus (18). The C. jejuni isolate F079, an ST-536 isolate belonging to the ST-21 multilocus sequence typing (MLST) clonal complex (CC), was also CRISPR negative. The SNP genotype consistent with ST-48 (SNP group 10) was numerically dominant in all isolate collections used in this study, represented by 43 (20%) strains. The dominance of this genotype is not seen in other countries, such as the United Kingdom and United States, in which the ST-21 and ST-828 CCs, respectively, are most commonly isolated (8). These results suggest that the ST-48 genotype is common in the Australian C. jejuni population and contributes substantially to human gastroenteritis in this country. All 43 isolates with this SNP profile possess a single DR. This finding is consistent with a previous study which found that all ST-48 isolates contained a single DR (18).

CRISPR sequencing.

Thirty-two PCR-positive CRISPRs of various sizes were subjected to DNA sequencing using the CRISPR-For and CRISPR-Rev primers. Sequencing of CRISPR loci enabled (i) correct size determination of all CRISPR products, (ii) comparison of spacer sequences between the present study and that of Schouls et al. (18), and (iii) assessment of the performance and resolving power of HRM analyses. DNA sequencing revealed 22 different CRISPR types (CTs) within the 32 isolates, of which 8 isolates contained a single DR (Table (Table1).1). Schouls et al. (18) identified 170 unique spacers among 137 Campylobacter isolates, and 55 unique spacers from 32 C. jejuni strains were found in the present study. While the DR sequence was strongly conserved, the spacer sequences are all novel. This confirms the highly polymorphic nature of the CRISPR spacers.

CRISPR spacer sequences of 32 C. jejuni isolates from the present study

HRM analysis of the CRISPR locus.

Real-time PCR is an attractive platform for high-throughput bacterial genotyping, as it is single step, closed tube, and cost-effective, and real-time PCR devices are used extensively in both research and analytical laboratories. Recently, real-time PCR devices that contain HRM capabilities have emerged (9). HRM differs from conventional PCR product melting temperature (Tm) measurement in two ways. First, the accuracy of the melt curve is maximized by acquiring fluorescence data over small temperature increments (as low as 0.01°C). Secondly, the precise shape of the HRM curve is a function of the DNA sequence being melted, allowing amplicons containing different sequence to be discriminated on the basis of melt curve shape, irrespective of whether amplicons share the same Tm. HRM analysis makes use of melt curve normalization and comparison software that allows the user to determine whether two similar melt curves differ from one another.

HRM analysis of the C. jejuni CRISPR was achieved by transferring the conventional PCR procedure to the Rotor-Gene 6000 (Corbett Research, Sydney, Australia) platform. First, the levels of effectiveness of different double-stranded nucleic acid-specific fluorescent dyes in performing HRM were compared. SYTO9 (Invitrogen-Molecular Probes) has recently been reported as a suitable chemistry for HRM analysis, and due to its lack of PCR inhibition activity, it can be used at higher concentrations than SYBR green I (3, 12). However, there is currently limited information on the comparative performance of intercalating DNA dyes for HRM analysis, particularly for examination of complex amplicons. SYTO9 reaction mixtures contained 5 pmol (0.5 μM) of each primer, 1.5 mM MgCl2, 0.2 mM deoxynucleoside triphosphates (dNTPs), 1 U Platinum Taq DNA polymerase and the relevant PCR buffer, 2.5 μM SYTO9, and 1 μl genomic DNA, made to a volume of 10 μl with double-distilled H2O. SYBR green I reactions were performed as previously detailed (15). Despite altering several parameters, including dye and genomic DNA concentration, efficient and reproducible amplification using SYBR greenER (Invitrogen) could not be achieved, and this dye was not used further.

SYTO9 and SYBR green I cycling parameters were as follows: 95°C for 2 min, 40 cycles of 95°C for 3 s and 58°C for 45 s, and final extension at 72°C for 2 min. HRM analysis of all samples was undertaken postrun by incubation at 50°C for 20 s followed by ramping from 70 to 85°C, with fluorescence data acquisition at 0.05°C increments. HRM analyses were carried out using Rotor-Gene 6000 1.7.34 software. Both conventional and normalized dissociation plots were generated. The normalized dissociation plot was used to construct difference graphs, a feature of the Rotor-Gene 6000 software that allows the user to quantitatively determine the sample deviation relative to a control sample. Four controls were included in each run to facilitate inter-run comparisons, and all samples were tested in duplicate to ensure reproducibility of the melt curves and to determine appropriate cutoffs for “same” and “different” genotypes. Classification of “different” CRISPR HRM genotypes was determined by a cutoff of ≥±5 U in the difference graph relative to a control sample of interest. “Different” genotypes inside the cutoff were further distinguished based on whether the curves traversed the midpoint of the difference graph two or more times (with a deviation of ≥±2 U) or whether distinct peaks were evident in the difference graph (Fig. (Fig.1).1). Where appropriate, the normalized and conventional dissociation plots were also used to ascribe genotypes as the “same” or “different” by visual inspection of the curves relative to the control sample of interest.

FIG. 1.
Classification of “same” and “different” genotypes using HRM analysis of the Campylobacter jejuni CRISPR locus. Data that illustrate the limits of the technology were deliberately chosen: i.e., the alleles shown here were ...

The levels of performance of SYTO9 and SYBR green I were compared using CRISPR amplicons from a subset of 29 C. jejuni isolates. For all 29 isolates, the SYBR green I melt profiles were highly reproducible and easy to interpret, whereas the SYTO9 melt profiles were much less reproducible (Fig. (Fig.2).2). This was an unforeseen finding, as SYBR green I has been shown to translocate between amplicons during melt curve analysis to higher-Tm products (2, 4, 20), suggesting instability of this dye during HRM analysis. A previous Tm study comparing the two chemistries demonstrated that SYBR green I melt profiles were more greatly influenced by dye, MgCl2, and starting DNA concentrations than melt profiles with its counterpart (12). In addition, SYTO9 has been shown to generate more reproducible DNA melting curves over a larger range of concentrations than SYBR green I, as the dye is less inhibitory to PCR at higher concentrations and does not selectively detect amplicons in multiplex reactions (2, 12).

FIG. 2.
Comparison of SYBR green I and SYTO9 by HRM analysis of the CRISPR locus of Campylobacter jejuni. Two genotypes (gray and black lines) are shown. (Left panels) SYTO 9. (Right panels) SYBR green I. (Top) Normalized high-resolution melt curve. (Middle) ...

To confirm the improved performance of SYBR green I over SYTO9, a further 125 C. jejuni isolates were compared. Inter-run variability was monitored by including four control strains per run. As previously observed, the SYBR green I amplicons generated highly reproducible and sensitive melt curves, whereas SYTO9 amplicons were difficult to interpret due to large differences between replicates, with many replicates called as “different” using a relaxed cutoff value of ≤±7 U. As a result of the superior performance of the SYBR green I chemistry over SYTO9, the remaining strains were examined using SYBR green I only.

Comparison of the CRISPR sequence data with the HRM data revealed that SYBR green I enabled discrimination of CRISPR amplicons containing small sequence differences, independent of amplicon length. For example, CRISPR HRM discriminated between the single-DR (237 bp) amplicons of F448 and F509, which differ by an SNP at their 5′ end. Such sensitivity was not apparent in all cases, with F448 and F486 identical by HRM but harboring an SNP at the 3′ end of the amplicon. Overall, eight CRISPR HRM genotypes were identified within isolates containing a single DR. The larger (~900 bp) amplicons of F050 and F119, which differ at 1 base as well as having a length difference of 4 bp, were efficiently discriminated by the SYBR green I HRM method. Importantly, the HRM profiles of CRISPRs with identical sequences, such as F509, F168, 01M27530, F421, and F492, were indistinguishable. These results demonstrate the power of the HRM method for discriminating “same” or “different” CRISPR genotypes in C. jejuni.

Comparison of genotyping methods with CRISPR HRM.

Previously, we have developed SNP and binary gene typing methods for C. jejuni and C. coli using the real-time PCR platform and have generated considerable data on the performance of these methods using the 181 Australian C. jejuni and C. coli isolates (15, 16). The present study addressed whether the HRM procedure provided comparable resolution to the current “gold standard” pulsed-field gel electrophoresis (PFGE) procedure when combined with SNP and/or binary gene interrogation of C. jejuni isolates.

The SNP-binary gene approach has previously been shown to provide comparable resolution to MLST-flagellin A short variable region (flaA SVR) sequencing. However, these combinatorial methods were unable to reach the discriminatory power of PFGE (16). One of the shortcomings of the study by Price et al. (16) was that the Campylobacter strains examined were obtained from sporadic gastroenteritis cases collected over a 3-year period, and therefore few definitive epidemiological data were available. To overcome this, 29 previously uncharacterized C. jejuni isolates from Queensland, Australia, were obtained from Queensland Health Scientific Services (QHSS). These isolates comprised six distinct outbreak clusters with confirmed epidemiological data, as well as seven sporadic gastroenteric isolates. The QHSS isolates were tested blind by SNP typing, binary gene typing, and flaA-restriction fragment length polymorphism analysis using DdeI (Roche). PFGE was carried out on the QHSS isolates using SmaI (Roche) digestion. Digested DNA was electrophoresed using a CHEF DR III system (Bio-Rad) for 10 to 25 s at 6 V cm−1 in 0.5% Tris-borate-EDTA (pH 8.3) for 22 h at 14°C.

In the sporadic isolate collections (designated 84, 154, and PA isolates in Table S1 in the supplemental material), the binary-CRISPR HRM and binary-flaA SVR approaches provided the highest discriminatory power of two methods in combination. Significantly, the use of three methods (SNP-binary-flaA SVR/CRISPR HRM or MLST-binary-flaA SVR) enabled further delineation of the sporadic isolates and provided resolution comparable to or surpassing that of PFGE (see Table S1 in the supplemental material). In contrast, addition of CRISPR HRM to the SNP-binary profiles did not increase resolution of outbreak QHSS genotypes, most likely attributable to the high genetic similarity of these isolates; however, the discriminatory power remained comparable to that of PFGE. The PFGE and SNP-binary (with or without CRISPR HRM) genotypes for the 22 outbreak isolates corroborated in almost all instances, and both methods were effective in differentiating between outbreaks (Table (Table2).2). These results demonstrate that the SNP-binary-CRISPR HRM approach is a feasible alternative to PFGE in both sporadic and outbreak isolate investigations of C. jejuni.

QHSS isolate genotyping results

To conclude, we have demonstrated the application of HRM as an alternative to DNA sequencing, using the CRISPR locus of C. jejuni and C. coli as a model. The CRISPR HRM assay, in combination with other established real-time PCR methods for C. jejuni and C. coli, provides a novel approach to bacterial genotyping that equals or surpasses the resolving power of PFGE. The value of HRM for characterizing complex DNA sequence is not limited to CRISPRs or C. jejuni and can potentially be applied to any polymorphic region, ranging from SNPs to entire genes.

Nucleotide sequence accession numbers.

The nucleotide sequences of the CRISPRs from 32 C. jejuni isolates have been deposited in GenBank under accession no. EF017316 to EF017347.

Supplementary Material

[Supplemental material]


E.P.P. is in receipt of a postgraduate studentship from IHBI, QUT. This work was supported by the Cooperative Research Centres Program of the Australian Federal Government.


[down-pointing small open triangle]Published ahead of print on 30 March 2007.

Supplemental material for this article may be found at http://aem.asm.org/.


1. Fouts, D. E., E. F. Mongodin, R. E. Mandrell, W. G. Miller, D. A. Rasko, J. Ravel, L. M. Brinkac, R. T. DeBoy, C. T. Parker, S. C. Daugherty, R. J. Dodson, A. S. Durkin, R. Madupu, S. A. Sullivan, J. U. Shetty, M. A. Ayodeji, A. Shvartsbeyn, M. C. Schatz, J. H. Badger, C. M. Fraser, and K. E. Nelson. 2005. Major structural differences and novel potential virulence mechanisms from the genomes of multiple Campylobacter species. PLoS Biol. 3:e15. [PMC free article] [PubMed]
2. Giglio, S., P. T. Monis, and C. P. Saint. 2003. Demonstration of preferential binding of SYBR Green I to specific DNA fragments in real-time multiplex PCR. Nucleic Acids Res. 15:e136. [PMC free article] [PubMed]
3. Giglio, S., P. T. Monis, and C. P. Saint. 2005. Legionella confirmation using real-time PCR and SYTO9 is an alternative to current methodology. Appl. Environ. Microbiol. 71:8944-8948. [PMC free article] [PubMed]
4. Herrmann, M. G., J. D. Durtschi, L. K. Bromley, C. T. Wittwer, and K. V. Voelkerding. 2006. Amplicon DNA melting analysis for mutation scanning and genotyping: cross-platform comparison of instruments and dyes. Clin. Chem. 52:494-503. [PubMed]
5. Huygens, F., A. J. Stephens, G. R. Nimmo, and P. M. Giffard. 2004. mecA locus diversity in methicillin-resistant Staphylococcus aureus isolates in Brisbane, Australia, and the development of a novel diagnostic procedure for the Western Samoan phage pattern clone. J. Clin. Microbiol. 42:1947-1955. [PMC free article] [PubMed]
6. Huygens, F., J. Inman-Bamber, G. Nimmo, W. Munckhof, J. Schooneveldt, B. Harrison, J. A. McMahon, and P. M. Giffard. 2006. Staphylococcus aureus genotyping using novel real-time PCR formats. J. Clin. Microbiol. 44:3712-3719. [PMC free article] [PubMed]
7. Jansen, R., J. D. Embden, W. Gaastra, and L. M. Schouls. 2002. Identification of genes that are associated with DNA repeats in prokaryotes. Mol. Microbiol. 43:1565-1575. [PubMed]
8. Jolley, K. A., M.-S. Chan, and M. C. Maiden. 2004. mlstdbNet—distributed multi-locus sequence typing (MLST) databases. BMC Bioinformatics 5:86. http://pubmlst.org/campylobacter/. [PMC free article] [PubMed]
9. Krypuy, M., G. M. Newnham, D. M. Thomas, M. Conron, and A. Dobrovic. 2006. High resolution melting analysis for the rapid and sensitive detection of mutations in clinical samples: KRAS codon 12 and 13 mutations in non-small cell lung cancer. BMC Cancer 6:295. [PMC free article] [PubMed]
10. Mojica, F. J., C. Diez-Villasenor, E. Soria, and G. Juez. 2000. Biological significance of a family of regularly spaced repeats in the genomes of Archaea, Bacteria and mitochondria. Mol. Microbiol. 36:344-346. [PubMed]
11. Mojica, F. J., C. Diez-Villasenor, J. Garcia-Martinez, and E. Soria. 2005. Intervening sequences of regularly spaced prokaryotic repeats derive from foreign genetic elements. J. Mol. Evol. 60:174-182. [PubMed]
12. Monis, P. T., S. Giglio, and C. P. Saint. 2005. Comparison of SYTO9 and SYBR Green I for real-time polymerase chain reaction and investigation of the effect of dye concentration on amplification and DNA melting curve analysis. Anal. Biochem. 340:24-34. [PubMed]
13. O'Reilly, L. C., T. J. J. Inglis, L. Unicomb, et al. 2006. Australian multicentre comparison of subtyping methods for the investigation of Campylobacter infection. Epidemiol. Infect. 134:768-779. [PMC free article] [PubMed]
14. Parkhill, J., B. W. Wren, K. Mungall, J. M. Ketley, C. Churcher, D. Basham, T. Chillingworth, R. M. Davies, T. Feltwell, S. Holroyd, K. Jagels, A. V. Karlyshev, S. Moule, M. J. Pallen, C. W. Penn, M. A. Quail, M.-A. Rajandream, K. M. Rutherford, A. H. M. van Vliet, S. Whitehead, and B. G. Barrell. 2000. The genome sequence of the food-borne pathogen Campylobacter jejuni reveals hypervariable sequences. Nature 403:665-668. [PubMed]
15. Price, E. P., V. Thiruvenkataswamy, L. Mickan, L. Unicomb, R. E. Rios, F. Huygens, and P. M. Giffard. 2006. Genotyping of Campylobacter jejuni using seven single-nucleotide polymorphisms in combination with flaA short variable region sequencing. J. Med. Microbiol. 55:1061-1070. [PubMed]
16. Price, E. P., F. Huygens, and P. M. Giffard. 2006. Fingerprinting of Campylobacter jejuni by using resolution-optimized binary gene targets derived from comparative genome hybridization studies. Appl. Environ. Microbiol. 72:7793-7803. [PMC free article] [PubMed]
17. Robertson, G. A., V. Thiruvenkataswamy, H. Shilling, E. P. Price, F. Huygens, F. A. Henskens, and P. M. Giffard. 2004. Identification and interrogation of highly informative single nucleotide polymorphism sets defined by bacterial multilocus sequence typing databases. J. Med. Microbiol. 53:35-45. [PubMed]
18. Schouls, L. M., S. Reulen, B. Duim, J. A. Wagenaar, R. J. L. Willems, K. E. Dingle, F. M. Colles, and J. D. Van Embden. 2003. Comparative genotyping of Campylobacter jejuni by amplified fragment length polymorphism, multilocus sequence typing, and short repeat sequencing: strain diversity, host range, and recombination. J. Clin. Microbiol. 41:15-26. [PMC free article] [PubMed]
19. Stephens, A. J., F. Huygens, J. Inman-Bamber, E. P. Price, G. R. Nimmo, J. Schooneveldt, W. Munckhof, and P. M. Giffard. 2006. Methicillin-resistant Staphylococcus aureus genotyping using a small set of polymorphisms. J. Med. Microbiol. 55:43-51. [PubMed]
20. Varga, A., and D. James. 2006. Real-time RT-PCR and SYBR Green I melting curve analysis for the identification of Plum pox virus strains C, EA, and W: effect of amplicon size, melt rate and dye translocation. J. Virol. Methods 132:146-153. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...