Production of 2-Aminophenoxazin-3-one by Staphylococcus aureus Causes False-Positive Results in β-Galactosidase Assays

Abstract

Staphylococcus aureus can be distinguished from similar coagulase-positive staphylococci by its absence of β-galactosidase activity. This is commonly tested using o-nitrophenyl-β-d-galactopyranoside (ONPG) as the substrate. Unexpectedly, 111 and 58 of 123 isolates displayed apparent β-galactosidase activity in the ONPG assay and on the Vitek 2 system, respectively. Compositional analysis showed that the yellow coloration of the positive ONPG assay resulted from production of 2-aminophenoxazin-3-one. Alternative β-galactosidase substrates like X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) should be used for testing staphylococci.

TEXT

Staphylococcus aureus and several other members of the coagulase-positive staphylococci (CoPS) are important human and animal pathogens and are often encountered in clinical and veterinary laboratories. Definitive identification of S. aureus and its differentiation from other CoPS of veterinary origin can be difficult, as the diversity of phenotypic characteristics precludes a simple phenotypic classification to the species level. Even commercial phenotypic identification systems are limited in their capability to identify CoPS species such as S. pseudintermedius and S. delphini (19, 23). P agar supplemented with acriflavin can be used for positive screening of S. aureus isolates (8, 18), and numerous PCR assays, based on targets like 16S rRNA, nuc, hsp60, and kat genes, have been developed for identification of CoPS (3, 11, 20). However, these tests are not readily available in most clinical laboratories. The use of simple biochemical tests for rapid discrimination of S. aureus isolates from other CoPS would be desirable.

Among the common biochemical tests, the pyrrolidonyl arylamidase and β-galactosidase (β-Gal) tests have been reported to distinguish between S. aureus and members of the S. intermedius group (4, 17, 24). In particular, the absence of β-Gal activity can distinguish S. aureus from S. intermedius, S. pseudintermedius, and S. schleiferi subsp. schleiferi (2). However, we note that the Vitek 2 GP system (bioMérieux, France) reports a positive β-Gal result for S. aureus strain ATCC 29213, which is used as a control strain for the commercial phenotypic identification system. Furthermore, a variety of substrates are used for β-Gal assays, often without validation of their applicability to different bacteria. Thus, we aim to examine the utility of o-nitrophenyl-β-d-galactopyranoside (ONPG) and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) for β-Gal testing of S. aureus and other staphylococci.

A total of 123 S. aureus strains of human and animal origins were used in this study (10), including 2 clinical strains with atypical phenotypic features (22, 26). For comparison and control, 59 strains covering 8 other staphylococcal species were also tested (Table 1). Species identification of bacterial strains was performed using standard biochemical methods, the Vitek 2 GP system, and the Vitek MS system (bioMérieux). Additionally, species identity of S. aureus strains was confirmed genotypically with two different specific PCR assays (13, 27). For the ONPG assay, bacterial colonies were inoculated into 1 ml of ONPG broth containing 0.0133 M ONPG in 1 M monosodium phosphate solution to form a heavy suspension. This was incubated at 37°C under aerobic conditions and examined at 4 and 24 h. Development of a yellow color at either time indicates a positive reaction. Bacterial culture was inoculated into lactose broth as a control. Additionally, results were examined by measurement of absorbance at 420 nm, which corresponded to the presence of o-nitrophenol resulting from ONPG hydrolysis, on a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific). Measurements were performed after centrifugation to reduce light scattering from suspended bacterial cells. For β-Gal testing with X-Gal, tryptic soy agar was prepared and supplemented with 17.5 μg of X-Gal. Bacterial colonies were then streaked onto the agar, and a positive reaction was characterized by the appearance of blue color after 24 h incubation at 37°C under aerobic conditions.

Table 1.

Staphylococcal strains used in this study and their results in the ONPG assay

Species	Source	β−Gal activity according to reference 2a	No. of ONPG-positive isolates/total no. of isolates (%)b
S. aureus	Human	−	57/59 (96.6)
	Animal surveillancec	−	50/60 (83.3)
	Othersd	−	4/4 (100)
S. capitis	Human	−	1/8 (12.5)
S. caprae	Human	−	0/2 (0)
S. epidermidis	Human	−	14/14 (100)
S. haemolyticus	Human	−	2/3 (66.7)
S. hominis	Human	−	1/4 (25)
S. lugdunensis	Human	−	14/15 (93.3)
S. pseudintermedius	Canine	+	12/12 (100)
S. schleiferi	Human	ND	1/1 (100)

^a−, 90% or more strains negative; +, 90% or more strains positive; ND, not determined or insufficient data.

^bMajor discrepancies between previously described results and the assay results are in boldface.

^cOnly one isolate from each animal is included. Isolates were sampled from 21 chickens, 14 cats, 10 dogs, 13 pigs, and 2 rats.

^dThese include two atypical clinical strains as mentioned in the text, the reference stain ATCC 25923, and the laboratory strain RN4220.

For S. aureus, 111/123 (90.2%) strains were positive in the ONPG-based β-Gal test via development of a yellow color (Table 1). Various intensities of yellow coloration were observed in strains positive by the β-Gal test, and 46 strains developed a yellow color only after 24 h of incubation. For other staphylococcal species, all S. pseudintermedius strains gave a positive reaction as expected, while most S. lugdunensis and S. epidermidis strains were apparently positive in the ONPG assay (Table 1). The latter results did not agree with the known absence of β-Gal activities in these two species (2). For a more objective assessment, we also collected the Vitek 2 GP test results for the S. aureus strains, which showed that 58 (47.2%) strains were also β-Gal positive, as tested by this automated commercial system. It should be emphasized that the incubation time for the Vitek 2 GP system is typically up to 8 h (6), which is much shorter than the 24 h for our ONPG assay. No coloration was observed for any “β-Gal-positive” S. aureus, S. lugdunensis, and S. epidermidis strains in the X-Gal assay, suggesting that they are in fact negative for β-Gal (Fig. 1). All S. pseudintermedius strains were β-Gal positive in the X-Gal assay.

Fig 1.

Detection of β-Gal activity using assays based on two chromogenic substrates: ONPG (i) and X-Gal (ii). Organisms used include the positive control E. coli ATCC 25922 (A), S. aureus (B), and S. pseudintermedius (C). A false-positive result for S. aureus is observed using the ONPG-based assay; the same strain is negative for β-Gal activity on X-Gal. X-Gal plates consisted of tryptic soy agar supplemented with 17.5 μg of X-Gal.

To further investigate this discrepancy in β-Gal results, spectrophotometric measurements on the ONPG reaction mixtures were examined. The levels of absorbance at 420 nm for S. aureus strains were up to 20-fold lower than that observed in the E. coli positive controls (Fig. 2). Similar findings were obtained for S. lugdunensis and S. epidermidis strains (data not shown), while S. pseudintermedius displayed high levels of absorbance at 420 nm, similar to that of E. coli (Fig. 2). This finding suggests that a pigment other than o-nitrophenol is responsible for the yellow coloration in the ONPG assay when “β-Gal-positive” S. aureus is tested. To identify the responsible pigment, the ONPG assay mixture was analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) and nuclear magnetic resonance (NMR) spectroscopy (see the supplemental material). Results from these analyses identified 2-aminophenoxazin-3-one (APX) to be the yellow pigment in the assay mixture.

Fig 2.

Box plot of the measured absorbance at 420 nm of mixtures that were positive by the ONPG-based β-Gal test for S. aureus (n = 123), S. pseudintermedius (n = 12), and E. coli (n = 5) following incubation with ONPG for 24 h. The horizontal solid bar in each box represents the median, upper and lower boundaries of the boxes are the third and first quartiles of the data, and whisker bars above and below the box are the 90th and 10th percentiles, respectively.

Other investigators have previously obtained false-positive results when testing S. aureus in the ONPG assay and attributed the phenomenon to staphyloxanthin production (18, 25). However, the proposed explanation is not sound. First, staphyloxanthin has poor water solubility and would not have produced significant coloration in solution. Second, the ONPG broth could not support growth of S. aureus and the initial inoculum could not have resulted in a yellow suspension. Furthermore, the color is observed only in the presence of ONPG and not with S. aureus alone or in the presence of lactose. Lastly, the false-positive reaction is also observed with S. epidermidis, which does not produce staphyloxanthin. Hence, the false-positive ONPG assay is not a result of staphyloxanthin production.

The pathway for biosynthesis of APX from o-aminophenol in Streptomyces antibioticus has been studied previously (12), but no known pathway exists for the biosynthesis of APX from ONPG or o-nitrophenol. In addition, with the exception of the laccase gene, S. aureus genomes do not contain genes encoding homologs of known APX biosynthesis components, thus further suggesting that a novel pathway is involved in this instance. Given that APX is an antibiotic with an unusual antimicrobial spectrum as well as potential for anticancer treatment (1, 5, 7, 9, 14–16, 21), knowledge of additional biosynthetic pathways will facilitate synthesis of structural derivatives with potential biomedical or industrial applications.

For CoPS isolates from animals or at-risk patients, molecular assays may be the most discriminative method for species level identification. In the absence of such assays, the β-Gal test can be used with substrates such as 2-naphthyl-β-d-galactopyranoside (BNBG) and X-Gal to avoid the false-positive result described here. Several commercial biochemical phenotypic identification systems for Staphylococcus employ BNBG for β-Gal testing, which might explain why S. aureus was identified as β-Gal negative in previous literature (2). The present example is illustrative in showing how the results from a simple biochemical test can be subverted by the unknown metabolic capacity of the test organism. We speculate that a similar situation may occur for other bacteria, and the authenticity of β-Gal assay results when applied to a new organism should be verified by testing different substrates.

In summary, many S. aureus strains produce a false-positive result in ONPG assays due to production of APX from ONPG. An alternative β-Gal substrate, such as X-Gal, should be used for β-Gal testing of S. aureus and other staphylococci to avoid a false-positive reaction.