Accurate species identification of a bacterial isolate is critical for managing infection in hospitalized patients. Staphylococcus aureus (S. aureus) is an important species to identify and is traditionally differentiated from coagulase-negative staphylococcal species through the coagulase tube test. S. aureus is frequently associated with complications and was responsible for over one million deaths globally in 2019 [1]. Conversely, most coagulase-negative staphylococcal species are rarely associated with disease; thus, coagulase testing of staphylococcal isolates is central to patient management.

S. aureus has multiple proteins that interact with and target platelets and host hemostasis pathways. Some examples include the clumping factors, which bind to fibrinogen or fibrin to create large cellular clumps [2]. Another significant protein is staphylocoagulase, which activates the clotting cascade by binding to complement C3, activating prothrombin to thrombin [3].

Diagnostically, this coagulating ability can be detected 4–24 hours after collection and well ahead of formal species identification using molecular methods. Developed in 1940, Fairbrother identified that combining an S. aureus colony with plasma in a tube would cause coagulation, later described as staphylocoagulase activity [4]. Subsequently, it was found that substituting rabbit plasma for human plasma would reduce the coagulase testing time [5], making it a staple phenotypic assay in the diagnostics lab for infections.

Advancements in understanding the underlying genetic makeup of S. aureus and other bacterial species have increased, allowing for high-throughput systems that can use unique genetic and biochemical markers to identify bacterial species quickly; however, most systems still require culturing first [6].

Some advanced diagnostic tests include molecular-based platforms such as the Xpert MRSA/SA BC assay and FilmArray systems, which use genetic probes and polymerase chain reactions [6]. One key advancement was the mass spectroscopy techniques, such as the gold-standard matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS), which increased the accuracy and reliability of the identification of species and drug resistance markers over phenotypic tests, with comparable times to phenotypic tests such as the coagulase tube test [6]. MALDI-TOF MS does have some diagnostic limitations, although not with S. aureus. Indeed, MALDI-TOF MS is known for its expense, meaning it is not a primary diagnostic tool for the laboratory [7]. Therefore, some laboratories still use the coagulase tube test as either a primary or presumptive identification; however, significantly, some S. aureus isolates have been observed to test negative [8, 9].

The reported rate of coagulase-negative S. aureus isolates ranges between 2% and 16% [9, 10, 11]. Meanwhile, the reasons for this range remain unclear, with one study suggesting that the high rate of negative isolates may be due to the high number of antibiotic-resistance genes present in an isolate [12]. Another possibility is that upstream and downstream regulations prevent the staphylocoagulase gene from being expressed. Prior studies have identified that the accessory gene regulator (Agr) regulates staphylocoagulase [13, 14], meaning Agr mutant strains could potentially affect coagulase expression. However, to our knowledge, a study has not focused on the genetic expression of this gene since the late 1990s, with more recent studies focusing on coagulation in coagulase-negative staphylococci species [15] or effects on production with novel chemical compounds [16, 17].

A potential reason for coagulase-negative isolates is the variation in specificity of the staphylocoagulase gene (coa). In the late 1990s, classification of staphylocoagulase types via serotyping experiments began [18]. Through the early 2000s, the serotyping scheme was correlated with the coa genotype [18]. The genotyping scheme has been extended from an initial 10 serotypes to 16 genotypes of Staphylococcus aureus complexes, including several subtypes, such as Staphylococcus argenteus [19, 20].

The staphylocoagulase protein contains six recognized regions: the C-terminal, D1, D2, central and tandem repeat regions, and N-terminal regions [18]. The genotype is determined from the nucleotide sequence encoding the D1, D2, and central regions. The C-terminal and D1 regions bind to fibrinogen, and the N-terminal regions play a role in prothrombin activation, suggesting that the D1 variability may play a role in coagulase binding specificity, which is important in the coagulase tube test [18]. Interestingly, to our knowledge, no expression studies have focused on staphylocoagulase expression at the nucleic level, with most studies using Western blot assays; no studies outside of functional characterization have assessed how these variations affect expression and, therefore, how it may impact diagnostics.

This study aimed to investigate the reasons for Staphylococcus aureusisolates testing negative in the coagulase tube test and its implications on healthcare. We aimed to identify whether there was an association between coa genotypes and negative results in the coagulase tube test. We also aimed to assess if there were potential factors outside of the coa genotype that could explain this negative phenotype. Finally, by investigating these factors, we aimed to determine the relevance of the coagulase tube test in a molecular testing era, which will improve the diagnostic and clinical decision-making ability of this deadly infection.

Of the 122 isolates used in this study, 81.1% (n = 99) of the cohort were positive for coagulase function via the tube test. All but one isolate was positive via the coagulase slide test. The negative coagulase phenotypes were not observed to be significantly associated with complications in bacteremia (Supplementary Table 3).

All isolates were confirmed as S. aureus via sequencing as evaluated using Kraken2 and on assembled genome size. A summary of the sequencing data can be found in Supplementary Table 1. All 122 isolates were found to have the staphylocoagulase gene. However, 20% of the isolates had the coa gene across two contigs. The staphylocoagulase genotype I-XI, but not genotype IX, was detected in the isolate collection, as shown in Table 2. Genotype II was most frequently identified (n = 33), followed by genotype IV (n = 26). Where multiple isolates had the multilocus sequence types (MLSTs), that subset of isolates had the same coa genotype.

By comparing coagulase tube test results with coa genotype isolates with genotypes VII, X, and XI, it was observed that these results were more likely to test negative for the tube test. These results are summarized in Supplementary Table 3. Due to the small sample size, an exploratory univariable regression model for types II, X, and VII was assessed statistically as these types had indicated negative isolates and had 10 or more representative isolates. Staphylocoagulase type X (OR: 5.22, 95% confidence interval (CI): [1.37, 19.91], p = 0.015) was statistically associated with a negative tube test. Staphylocoagulase type VII was not statistically significant (OR: 2.45, 95% CI: [0.86, 6.96], p = 0.093).

To assess for potential contributing factors outside of the genotype, we evaluated the structure of the gene and the surrounding genes upstream and downstream of the coa gene in each isolate. Only one isolate sequence had a large deletion at the 3^′ end of the gene, which might theoretically affect staphylocoagulase expression and function. The phenotypic results concurred with this inference from the gene synteny analysis. This isolate was classified as genotype XI. Within genotypes, there were slight SNP differences within the variant regions; however, no SNP differences were consistently observed with the negative tube test results.

The majority of isolates had a common gene order close to the coa gene, consisting of carobamoyl-phosphate synthase small chain (carA), long chain fatty acid CoA ligase (lcfb), crotonobetainyl-CoA reductase (caiA), 3-hydroxyacyl-CoA dehydrogenase (fadN), probable acetyl-Coa acyltransferase (fadA), upstream, and Staphylococcal complement inhibitor (scn), a hypothetical gene, pyruvate formate lyase activating enzyme (pflA), formate acetyltransferase (pflB), and a hypothetical protein (unknown function) downstream. Slight differences observed in the upstream and downstream genes did not correlate with the coagulase tube test result and did not differ by genotype.

As Agr is known to affect coagulase production indirectly, we assessed Agr function to determine if poor Agr function correlated with a negative coagulase tube test; however, no association was observed (OR: 0.50, 95% CI: [0.14, 1.82], p = 0.290; Supplementary Table 3).

With genetic structure and regulatory function not accounting for all the negative tube tests observed, we selected a subset of isolates representing different genotypes and coagulase tube test results to assess whether the coa gene is being expressed via reverse transcriptase real-time PCR. The isolates tested for coa gene expression are shown in Table 3. The tested isolates included positive coagulase tube test isolates from genotypes II, IV, and X and negative tube tests from genotypes II, VII, and X, with two isolates representing negative X isolates.

RNA was extracted from cultures grown under conditions for coa gene expression [13]. Expression of coa was detected in all tested isolates despite the negative coagulase tube test observed in some of them (see Supplementary Table 3 for individual results). One isolate (isolate 27, the type IV, coagulase tube test positive) had very low coa expression under these growth conditions (see Supplementary Table 2 for details).

The severity of disease associated with S. aureus infections highlights the importance of rapid and accurate identification of this organism in the laboratory. The coagulase tube test has been used since the 1940s to identify or presumptively identify S. aureus. The test is still used in resource-limited countries and presumptively even in countries with access to MALDI-TOF for isolate identification [36]. However, there are reports of S. aureus isolates that test negative in the coagulase tube test [37]. Among the 122 S. aureus isolates in this study, there was a tube test negative rate of 19.8%, similar to that seen in several prior studies [9, 10, 11, 38].

In this study, negative coagulase tube test results were not associated with complications in bacteremia. However, there was no significant negative association, suggesting that isolates that test negative in the coagulase tube test are no less virulent than those that do. Therefore, we investigated why some S. aureus isolates test negative in the coagulase tube test.

At the genetic level, all isolates possessed the staphylocoagulase gene. Only one isolate (isolate 116) had a major deletion in the conserved region of the coa gene that may affect expressed protein function (that could have theoretically affected coagulase tube test positivity) and was observed to be coagulase tube test negative. There were also no differences in the upstream and downstream genes from the staphylocoagulase gene that were exclusively found in tube test-negative isolates that may have affected coa gene expression in the isolates. In total, 20% of the isolates had the coa gene split across two contigs due to the tandem repeat structure of the gene, a known effect in short-read sequencing technology where repeat sequences are not adequately spanned by reads [39]. Despite the coa gene being split across two contigs in the draft genome sequences of those isolates, we could identify all coa gene regions and the contig breakpoint at the tandem repeat. Indeed, due to this limitation, repeat sequence elements in the genome often confound the reproduction of other typing methods, such as the pulse field electrogram fragment lengths [40]. In detecting the coa gene, genotyping and the analysis presented in this study were not affected by the poor assembly across the tandem repeat regions. Meanwhile, long-read sequencing would eliminate this assembly anomaly.

The Agr function indirectly affects coagulase production [13] and was not associated with negative coagulase tube test results. Other gene expression regulators may be related to the tube test result; however, given the association between genotype and tube test result, no additional regulators were investigated.

We next investigated the genetic variation in the staphylocoagulase gene. First identified molecularly by Watanabe et al. [33], the staphylocoagulase gene has different genotypes (identified initially by serotyping tests). The cohort isolate genes were mapped against references, and it was identified that all staphylocoagulase genotypes I–XII except type IX were present. This was not unexpected, as genotype IX is found in animal-associated isolates [41]. The staphylocoagulase genotypes are lineage-associated, and the gene is likely to be vertically inherited in S. aureus [33]. This is consistent with the observation from our isolates, whereby all isolates with the same ST have the same coa genotype.

Significantly, staphylocoagulase genotype X was negatively associated with a positive coagulase tube test result (OR: 5.22, 95% CI: [1.37, 19.10], p = 0.015). Genotype X was exclusively found in ST15 and ST3911, both of which are a part of clonal complex 15. This concurred with a prior study that found type X, sub-type Xa, in ST15 strains [33], although the preceding study only had one representative isolate.

In relation to the collection of isolates, variation in S. aureus ST prevalence across geographical areas has been observed previously. For example, the most common coa genotype observed was type II, usually ST5 and ST25; ST5 is the most common in Australia [42, 43]. The association between the coa genotype and ST types may explain the rates of coagulase-negative S. aureus isolates. ST15 is a common strain globally, but isolation rates differ depending on geographical region [44, 45]. The observation that ST15 may be more likely to be tube test negative could also explain the published variations in coagulase-negative S. aureus isolates.

We investigated whether there were other reasons preventing expression. Using real-time reverse transcriptase PCR, we observed that all isolates in the selected subset expressed staphylocoagulase. In cases where the isolates were coagulase-negative by tube test, this shows that staphylocoagulase is likely to be produced; however, this coagulase activity is not detected, perhaps due to possible variations in coagulase substrate specificity for particular genotypes.

S. aureus and mammals have co-existed and co-evolved for millennia [46]. Hence, coevolution with a particular animal host might narrow the staphylocoagulase substrate; for example, coagulase type IX is exclusively found in livestock-associated strains, so it may be that the rabbit plasma used in the coagulase test is not a suitable substrate for coa genotype X. Interestingly, a 2010 study investigated isolates collected from human, bovine, and dog plasma, and identified that dog plasma was superior to rabbit plasma. Although there were no coagulase-negative isolates, it should be noted that the collection of isolates was determined using a coagulate test [47].

An alternative driver for the postulated shift in specificity is that there may have been a functional shift in the affected genotypes. Staphylocoagulase type X has 98.5% nucleotide identity similarity with the fibronectin-binding protein A gene [11], so a change in function may impact the coagulation normally observed in the tube test.

With fast molecular techniques increasingly used in the clinical setting and the coagulase test used only as a preliminary test, whether it is time to move on from coagulase production as a marker for S. aureus should be posited. Numerous rapid molecular techniques have been developed recently, including the MALDI-TOF and BinaxNow with BinaxNow PBP2a assay [6]. While some methods, such as MALDI-TOF, have become routine in clinical laboratories, issues with these rapid diagnostics are limited to low throughput or high costs [6]. Therefore, avoiding the coagulase tube test may not be possible as it is cheap and has high throughput, making it an efficient preliminary test. However, this work highlights the importance of being aware of the limitations of the coagulase tube test.

This study has limitations, predominantly the small sample size, which could potentially make the regression association more prone to type II errors. Therefore, to minimize these errors, we only statistically assessed the samples with representative samples of 10 or more isolates and those with a distribution under 90%. Future studies investigating this ST would be beneficial in confirming this study’s findings.

Another limitation is the study setting: Study isolates were collected from a single-site collection point, meaning the genetic diversity represented in this study may not accurately reflect other clinical settings. The isolates were also all clinical, meaning the lineages were not evenly distributed, and the contribution of clonal complex 15 strains was small. However, this reflects the community-associated bloodstream lineages and incidence rates observed in a clinical laboratory setting for this site.

Despite these limitations, to our knowledge, this study is the first to identify associations between the staphylocoagulase genotype and coagulase tube test results, the first to use RT-PCR to determine whether coagulase-negative strains of S. aureus express the staphylocoagulase gene, and the first to see if Agr function would affect the coagulase tube test.

To conclude, this study does not recommend routinely screening for the staphylocoagulase gene for clinical or diagnostic purposes or removing the coagulase tube test from clinical practice. However, it does provide an important reminder that these bacterial lineages may have discrepancies in traditional laboratory tests. We observed that specific lineages, including those from MLST15 and MLST3911, may be more likely to test negative in a coagulase tube test, and scientists should be aware of false negatives.

The readsets for each isolate have been submitted to NCBI under the accession number PRJNA611667. All data not shown in the supplementary data and reported on in this paper will also be shared by the corresponding author upon reasonable request.

CB and EA designed the research study. CB performed the research. EA provided help and advice on the clinical significance of the findings. CB and DB analysed the data. CB and DB wrote the manuscript. All authors contributed to editorial changes in the manuscript. All authors have read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study was carried out in accordance with the guidelines of the Declaration of Helsinki. Ethical approval for this project was granted by the Barwon Health HREC (Reference number 16/33) and acknowledged by the Deakin University HREC (Reference number 2017-217). Due to the negligible to low risk and the retrospective data collection of this study, the Barwon Health HREC granted a waiver of consent for this study.

This research received no external funding.

The authors declare no conflict of interest.