| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Brief Communications |
Cooperative Research Centre for Diagnostics, Institute of Health and Biomedical Innovation, Queensland University of Technology, Queensland, Australia;
aAddress correspondence to this author at: the Cooperative Research Centre for Diagnostics, Institute of Health and Biomedical Innovation, Queensland University of Technology, 60 Musk Ave, Kelvin Grove QLD 4059, Australia. e-mail f.huygens{at}qut.edu.au.
Abstract
Background: The staphylococcal protein A (spa) locus of Staphylococcus aureus contains a complex repeat structure and is commonly used for single-locus sequence-based genotyping. The real-time PCR platform supports genotyping methods that are single step and closed tube and potentially can be carried out simultaneously with diagnosis. We describe here a method for genotyping S. aureus using high-resolution melting (HRM) analysis of the spa polymorphic region X.
Methods: The conventional PCR spa assay was modified and optimized for the Rotor-Gene 6000 instrument (Corbett Life Science). HRM analysis on the Corbett Rotor-Gene 6000 instrument was used to test 22 known spa sequences obtained from 44 diverse methicillin-resistant S. aureus (MRSA) isolates. Criteria for calling pairs of melting curves "same" or "different" were developed empirically by converting the data to difference graph format with one curve defined as the control. HRM curve comparison between runs was done to determine the portability of the method. The assay performance was assessed by genotyping uncharacterized isolates, carrying out blind trials, and comparing HRM profiles from different runs.
Results: HRM analysis of 44 diverse MRSA isolates generated 20 profiles from 22 spa sequence types. The 2 unresolved HRM spa types differed by only 1 bp. Two blind trials demonstrated complete reproducibility with respect to calling the different spa types. Interrun comparisons of HRM curves were successfully developed, indicating the robustness of the method.
Conclusion: Analysis of the spa locus by HRM resolves spa sequence variants. This single- and closed-tube single-step method for S. aureus genotyping can be easily combined with the interrogation of other genetic markers.
The hypervariable region X of staphylococcal protein A (spa) is used for single-locus sequence-based genotyping of Staphylococcus aureus (1)(2)(3)(4). This locus contains a complex repeat structure that is thought to rapidly evolve through slipped-strand mispairing and recombination (5)(6)(7). Additionally, the spa extracellular domains are subject to immune surveillance, which is expected to increase its speed of evolution. Consequently, there are currently 2366 spa types in the Ridom SpaServer database (http://spaserver.ridom.de).
Although sequencing technology is now extremely effective, non–sequence-based genotyping methods are also advantageous. In particular, the real-time PCR platform supports genotyping methods that are single step and closed tube, can potentially be carried out simultaneously with diagnosis, and can interrogate different classes of genetic polymorphisms. These features provide real advantages for the clinical microbiology laboratory. A recent development in real-time PCR technology is high-resolution DNA melting (HRM) analysis (8)(9). Although melt curves are predominantly used to determine the melting temperature (Tm) of amplified double-stranded DNA, it is recognized that the precise shape of a melting curve is a function of the DNA sequence, and this characteristic forms the basis of HRM analysis (10). Accurate melting curves are derived using very small temperature increments, and normalization and comparison of melting curves allows sensitive determination of whether different amplicons have the same or different sequence (9). The potential resolving power of this approach is much greater than conventional melting curve analysis because in HRM melting curves from different amplicons can be differentiated on the basis of shape, even when they define the same Tm values.
HRM has previously been applied to human mutation screening (11)(12)(13) and to differentiating the hypervariable CRISPR (clustered regularly interspaced short-palindromic-repeat) locus of Campylobacter jejuni (14). We tested the hypothesis that HRM analysis on the Corbett Rotor-Gene 6000 instrument (Corbett Life Science) can be used to differentiate spa alleles. We also hypothesized that HRM curves can be compared between different runs of the Rotor-Gene 6000, a characteristic that would indicate that HRM analysis can be used as a library-based genotyping method instead of being strictly comparative. Initially, the HRM analysis was tested against 22 known spa sequences that were derived during the course of this study from 44 diverse MRSA isolates (Table 1
) (15)(16)(17). To rigorously test the discriminatory power of HRM, the collection included multiple isolates with identical multilocus sequence types (MLST), or MLST-derived single-nucleotide polymorphism profiles (16) thought to possess very similar spa sequences. The 22 spa sequences included 5 novel types; 3 with novel combinations of known repeat units, and 2 including novel repeat units. As anticipated, these included sequences that differed from each other at a single base, thus allowing rigorous testing of the HRM resolving power.
|
The conventional PCR spa assay using primers 1095F (5'-AGACGATCCTTCGGTGAGC-3') and 1517R (5'-GCTTTTGCAATGTCATTTACTG-3') (4) was modified and optimized for the Rotor-Gene 6000 instrument (Corbett Life Science). Each 10-µL reaction contained 5 µL Platinum® SYBR® Green qPCR SuperMix-UDG (2x, Invitrogen Life Technologies), 0.25 µL of each primer (20 µM stock, final concentration 0.5 µM), 3.5 µL ddH20, and 1 µL DNA template (20 fg/L). The real-time PCR thermocycling parameters were: 50 °C for 2 min, 95 °C for 2 min, 40 cycles of (95 °C for 5s, 60 °C for 30s), 72 °C for 2 min, and 50 °C for 20 s, followed by HRM ramping from 75–87 °C with fluorescence data acquisition at 0.05 °C increments. Reactions were routinely carried out in duplicate.
The Rotor-Gene 6000 proprietary software (version 1.7.34) enables the user to visualize HRM data in multiple ways. The negative derivative of fluorescence (F) over temperature (T) (df/dt) curve primarily displays the Tm, the normalized raw curve depicts the decreasing fluorescence vs increasing temperature, and difference curves (9), which display a user-defined curve as the baseline (i.e., the x-axis), and depicts other normalized curves in relation to that baseline. Criteria for calling melting curves as "same" or "different" using difference graphs were developed empirically. Melting curves are called the same as the defined control if the difference graph lies within ±4 U relative to the x-axis, and does not display reproducible differences such as double peaks or crossing the x-axis more than twice in both replicates (Fig. 1
). In addition, a 2-step procedure was followed to determine if an unknown HRM curve was the same as a known curve. First, the normalized HRM curve for the unknown type was compared to known normalized HRM profiles. These profiles are either generated together with the unknown samples, or more practically, have been previously produced. Second, the closest known HRM profile was selected as the difference graph control, and comparison of the difference curves was use to determine whether the unknown isolate was the same as or different from the known. For each data analysis, the digital filter was set to "heavy" and the replicate grouping option was selected. When conflicts between replicates occurred, repeat or sequence analysis was performed. According to the above criteria, 22 known spa sequences generated 20 reproducibly different HRM curves. As expected, this result was a higher degree of resolution than could be obtained from Tm determination. For example, HRM curves 1 and 5 were derived from different spa sequences but defined identical Tm values, as did HRM curves 11, 14, and 15. Two pairs of spa sequences were not resolved by HRM analysis. These were t002 and the NEW2, and t037 and t1155. Both pairs of sequences differed at a single base. We concluded that spa sequences can be differentiated by HRM even when they are closely related, but the resolving power using SYBR Green dye does not reliably extend to sequences that differ by a single base pair.
|
The software supplied with the Rotor-Gene 6000 does not allow the generation of difference graphs using curves from different runs of the device. To determine the practicality of HRM curve comparison between runs (and by extension, the development of a standard library of HRM curves), the normalized data were exported and converted to "difference" format in Microsoft Excel, and difference curves assembled using the freeware chart drawing program Teechart Office (18). We found that the HRM curves were completely portable, and that the practice of comparing data from different real-time PCR runs caused no loss in the ability to differentiate different spa alleles (Fig. 1
). To further confirm the robustness of this method, we carried out 2 blinded experiments. To increase the rigor of these tests, different batches of reagents were used. The first experiment entailed the analysis of 10 MRSA isolates of unknown spa sequence, as indicated in Table 1
. Subsequent spa sequence determination revealed that all the spa sequences were identical to spa types found in isolates in Table 1
, and all were correctly identified by HRM analysis. The second blinded experiment was the reanalysis of the original set of 34 isolates. Once again, the spa types were 100% consistent with the first 2 times these isolates were analyzed. Interestingly, because the melt curves were clearly visually distinct, we successfully differentiated spa sequences that differed by a single base and could not be separated in the first 2 analyses of these isolates owing to the very conservative sequence discrimination criteria. It is likely that these criteria will be refined as more data become available, thus increasing the resolving power of HRM analysis. In summary, HRM analysis was performed 3 times on the 34 isolates, and the results proved to be completely reproducible with respect to calling the spa types. There was one very minor reproducibility issue. When the Corbett 6000 runs were separated by several weeks, HRM curves from these runs could be offset by up to 0.2 °C, although there was complete consistency of curve shape. These results were likely attributable to very slight changes to the calibration of the instrument’s thermometer and can easily be overcome by including a known control in all runs.
We concluded that HRM analysis of the spa locus is an effective method to easily and rapidly identify spa alleles. The major advantage of this method is that it is a single-step closed-tube procedure performed on a moderately priced and generic piece of laboratory equipment. This technique offsets the occasional inability to differentiate spa alleles, because spa interrogation could simply be combined with the real-time PCR interrogation of, for example, clonal-complex–specific single-nucleotide polymorphisms (16), toxin-encoding genes (16), or binary markers that subtype SCCmec (19). In this context, HRM-mediated spa interrogation may be regarded as a facile means of adding resolving power to other genotyping methods for interrogating less polymorphic markers that more reliably define population structures. The portability of the HRM curves is very significant, because this characteristic makes the method potentially library based rather than comparative. A resource that would be very useful is a library of spa HRM curves for the major MRSA epidemic- and community-acquired clones, and this library has been provided in part by this study. Similar genotyping approaches could be developed for other bacteria.
Acknowledgments
Grant/funding Support: This work was supported by the Cooperative Research Centres Program of the Australian Federal Government.
Financial Disclosures: None declared.
Acknowledgment: The authors thank Jan Bell for supplying several of the MRSA isolates.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
