Quantification of Methicillin-Resistant Staphylococcus aureus Strains in Marine and Freshwater Samples by the Most-Probable-Number Method

Abstract

Recreational beach environments have been recently identified as a potential reservoir for methicillin-resistant Staphylococcus aureus (MRSA); however, accurate quantification methods are needed for the development of risk assessments. This novel most-probable-number approach for MRSA quantification offers improved sensitivity and specificity by combining broth enrichment with MRSA-specific chromogenic agar.

INTRODUCTION

Staphylococcus aureus strains are carried in the noses of 25% to 35% of humans, while carriage of methicillin-resistant S. aureus (MRSA) strains has increased from 0.8% (2000) to 1.5% (2004) (4, 8, 9). Identified risk factors associated with community-associated MRSA outbreaks include sharing of personal care products, frequent skin-to-skin contact, skin abrasions, and crowded living conditions. Recreational marine beaches are potential reservoirs for pathogenic bacteria (1), including S. aureus and MRSA (3, 5, 7, 11, 13, 14). Shedding of S. aureus and MRSA has been documented from bathers in marine waters (5, 11). Contact with seawater has been associated with a 4-fold increased risk of S. aureus skin infections in children (3). There are several potential point and non-point sources, such as recreational bathers (5, 11), combined sewer overflows (2), and urban runoff (12), possibly contributing to S. aureus and MRSA loads at recreational beaches. Currently, there is no risk assessment available to provide guidance on levels of MRSA that result in human infection or colonization from contact with contaminated sand or water at marine and freshwater beaches, and more research is needed (11). A key component needed for the development of a risk assessment model is accurate quantification of MRSA from these environments.

Previous studies have used one of the following two methods for isolation of MRSA from beach environments: membrane filtration onto MRSA-specific chromogenic media (7, 14) or selective broth enrichment followed by confirmatory biochemical tests (13). Membrane filtration onto chromogenic media, though simple, fast, and quantitative, has a high rate of false-positive and false-negative results with a positive predictive value of 25% (7, 14). In comparison to direct plating, broth enrichment followed by plating onto MRSA-specific chromogenic media has been shown to increase MRSA detection sensitivity by 16 to 24% (10).

The current study developed a most-probable-number (MPN) approach using broth enrichment in Quanti-Tray 2000 (IDEXX Laboratories, Westbrook, ME) for detection of Staphylococcus spp., followed by plating on MRSASelect media (Bio-Rad, Hercules, CA) to identify MRSA. The U.S. Environmental Protection Agency (EPA) has approved the Quanti-Tray 2000 MPN method for Escherichia coli and Enterococcus spp. for testing recreational water in accordance with the Clean Water Act (15). The Quanti-Tray 2000 methodology was adapted for MRSA detection by diluting 50-ml marine or freshwater samples 1:1 with 1.5× Bacto Staphylococcus Medium 110 broth (Difco Laboratories, Becton Dickinson & Co., Sparks, MD), a defined substrate-specific broth for Staphylococcus spp., supplemented with a final concentration of 75 μg/ml polymyxin B (Sigma, St. Louis, MO) and 0.01% potassium tellurite (Sigma) as previously described (13). The 100-ml final volume was then sealed into a Quanti-Tray 2000 and incubated at 36.5°C for 72 h. The majority of S. aureus strains oxidize potassium tellurite, generating a black precipitate that enables visual identification of positive wells (Fig. 1) (6). The number of wells with black precipitate was counted, and the presumptive Staphylococcus MPN/100 ml was calculated using the MPN chart provided by IDEXX adjusted for the dilution and sample volume. From each of the Staphylococcus-positive wells, 5 μl was spotted onto a MRSASelect plate and incubated at 36.5°C for ≥72 h. Each presumptive MRSA spot was then plated onto Brucella agar (Difco) supplemented with 5% sterile sheep blood to screen for beta-hemolysis. Positive isolates were coagulase tested using the Remel Staphaurex rapid latex kit (Thermo Fisher Scientific Remel Products, Lenexa, KS). Samples with beta-hemolytic coagulase-positive Gram-positive bacteria were considered MRSA positive and used to recalculate the MRSA-specific MPN/100 ml level.

Fig. 1.

Quanti-Tray 2000 with wells positive for Staphylococcus spp., identified by black precipitate from oxidation of potassium tellurite.

To evaluate the method, five MRSA environmental strains (strain 9-48 and 2 to 5) from local marine and freshwater beaches, two MRSA clinical strains (6 and 7), and one methicillin-susceptible S. aureus ATCC 25923 strain were serially diluted from a 0.5 McFarland stock using 0.85% saline solution. Prior to experiments, pooled marine and freshwater samples were filter sterilized through a 0.22-μm-pore-size filter (Pall Corporation, Port Washington, NY). To verify sterility, 100 μl of the sterilized marine and freshwater samples were plated in duplicate onto brain heart infusion (BHI) agar (Difco). The samples were added to marine water at approximately 100 CFU/50 ml, 10 CFU/50 ml, and 1 CFU/50 ml at a dilution of 1:1 with the Staphylococcus-defined substrate medium for a final volume of 100 ml. Each dilution was prepared in duplicate, and both replicates were sealed in a Quanti-Tray 2000 (Table 1). Fifty-milliliter replicates of freshwater were inoculated with the similar bacterial concentrations with three MRSA environmental strains (3 to 5) and one MRSA clinical strain (6) in Quanti-Tray 2000. The stock for each strain was plated onto BHI agar in triplicate to calculate the actual number of bacteria inoculated into the Quanti-Tray 2000. A negative control of 50 ml uninoculated filtered marine water or freshwater and 50 ml Staphylococcus defined substrate was prepared for each experiment.

Table 1.

Comparison of average CFU/100 ml and MPN/100 ml for eight environmental and clinical strains of MRSA in marine and freshwater samples

Type of isolate	Strain	Results for indicated sample type
		Marine water			Freshwater
		Avg CFU/ 100 ml	MPN/100 ml (95% CI)a	P valuee	Avg CFU/ 100 ml	MPN/100 ml (95% CI)a	P valuee
Environmental	9-48c	159	156.5 (111.6, 215.4)	0.47
		16b	14.6 (8.2, 24.6)
		2b	3.1 (0.7, 8.9)
	2d	143	115.3 (82.2, 158.1)	0.43
		14b	12.1 (6.5, 21.1)
		1b	3.0 (0.7, 7.4)
	3d	146	101.9 (72.7, 140.4)	0.36	183	81.6 (58.2, 110.3)	0.37
		15b	9.8 (4.7, 18.4)		18b	7.4 (3.2, 14.4)
		2b	2.0 (0.3, 7.1)		2b	2.0 (0.3, 7.1)
	4d	37	51.2 (36.5, 69.0)	0.38	37	55.6 (38.5, 77.2)	0.39
		4b	5.2 (1.8, 10.8)		4b	5.2 (1.8, 10.8)
		<0b	<1 (0.0. 3.7)		<0b	<1 (0.0, 3.7)
	5d	96	111.2 (79.3, 151.7)	0.43	96	111.9 (79.8, 154.0)	0.48
		10b	10.9 (5.6, 19.5)		10b	8.4 (3.7, 15.3)
		1b	3.1 (0.7, 8.9)		1b	1.0 (0.1, 5.5)
Clinical	6	131	151.5 (108.0, 207.8)	0.28	176	44.1 (30.6, 62.5)	0.48
		14b	20.3 (12.1, 32.2)		18b	4.1 (1.2, 9.1)
		1b	1 (0.1, 5.5)		2b	<1 (0.0, 3.7)
	7	52	95.9 (68.4, 130.5)	0.39
		5b	8.5 (3.9, 15.6)
		1b	<1 (0.0, 3.7)
	ATCC 25923	97	111.2 (79.3, 151.7)	0.49
		10b	8.5 (3.9, 15.6)
		1b	1 (0.1, 5.5)

^aMPN/100 ml and 95% confidence intervals (CI) calculated by IDEXX.

^bTheoretical value based on spread plating of serial dilutions.

^cPreviously characterized environmental strain (9).

^dUnpublished environmental strain obtained in summer 2010.

^eP values calculated using Student's paired t test.

The Quanti-Tray 2000 MPN/100 ml was highly correlated to the spiked concentrations, as determined by spread plating using Pearson's correlation coefficient for the eight isolates in marine water (R = 0.95), while the observed MPN/100 ml for fresh water had a lower correlation (R = 0.76). There was no statistically significant difference between the spiked concentrations and the MPN/100 ml by strain for both marine and freshwater samples with the paired Student's t test using Stata IC11.1 (P > 0.1) (Table 1).

To test the method with environmental samples, 27 freshwater and 8 marine samples from two different Seattle area beaches were concurrently processed with the Quanti-Tray 2000 MPN method and the previously described selective broth enrichment method (13). Two freshwater samples (1 and 2) were determined to be positive for MRSA by both methods, while the Quanti-Tray 2000 MPN method identified another nine freshwater samples and one marine water sample as MRSA positive (3 to 11 and 28) but were negative by the enrichment method. The remaining 23 samples (16 freshwater and 7 marine water) were determined to be negative by both methods (Table 2). A larger set of samples needs to be tested with both methods to verify that the difference in detection sensitivity is not due to the larger water sample volumes used in the MPN method (50 ml) versus the previously described enrichment method (25 ml). Limitations of this method include a minimum processing time of 72 h and technical support to confirm the presence of MRSA in the Staphylococcus-positive wells. However, the current method saves time and is more specific than previously described methods (1, 7, 11, 13, 14), and, until an adequate molecular method is found that can distinguish between MRSA and mecA-positive Staphylococcus spp., is an improvement over current methods. Nevertheless, the results of the current study suggest that the Quanti-Tray 2000 MPN approach offers MRSA quantification combined with the increased detection sensitivity of broth enrichment in recreational marine and freshwater samples. This quantification methodology improvement described will be useful as further research is done and for the development of a risk assessment model for MRSA colonization and infection of recreational beach visitors.

Table 2.

Comparison between Quanti-Tray MPN and broth enrichment on the same field samples

Sample(s)	MPN/100 ml	Broth enrichmenta	Water type
1	2.0	+	Fresh
2	2.0	+	Fresh
3	37.8	−	Fresh
4	66.2	−	Fresh
5	8.2	−	Fresh
6	4.0	−	Fresh
7	2.0	−	Fresh
8	2.0	−	Fresh
9	6.2	−	Fresh
10	6.2	−	Fresh
11	10.4	−	Fresh
12 to 27	<1.0b	−	Fresh
28	2.0	−	Marine
29 to 35	<1.0b	−	Marine

^a+, positive for growth and black precipitate; −, negative for growth and black precipitate.

^b0 positive wells.