The study was approved by the Florida Department of Health Internal Review Board (IRB 1491; DOH IRB Number, H07164) and the University of Miami Internal Review Board (IRB 20070306). Consent forms were signed by each study participant (or parent/guardian), and participant identity was kept confidential. The field experimental design followed that of Elmir et al. [17, 18], including the use of the same study site (a sub-tropical non-point source recreational marine beach).
Pool field studies
The "Large Pool" field study was used to determine the total amount of S. aureus and the distribution of S. aureus relative to MSSA and MRSA released from the bodies of adult bathers [17, 18]. Briefly, after filling an inflatable pool with 1400 L of local off-shore marine water (where no humans were observed swimming at the time of collection), two groups of 10 adult participants were subjected to a series of four continuous 15-minute bathing cycles conducted on a single day in July 2008 beginning at 9:00 AM and 12:00 PM, respectively. In between bathing cycles, the pool was cleaned and refilled from the same source water. Participants had no sand exposure during the first two cycles, but were exposed to beach sand during the last two cycles. Samples of the source water, pool water before participant contact (in triplicate) and pool water after participant contact (in triplicate) were collected after each cycle. Source water, pool water and residual sand samples were analyzed as described below. The demographic characteristics of the 20 adult "Large Pool" participants (10 males and 10 females) included an age range from 19 to 51 years old, and body weights ranging from 50 to 100 kg [18].
The "Small Pool" field study was used to determine the total amounts of S. aureus and the distribution of S. aureus among MSSA and MRSA released from the bodies of a pediatric population, including an estimate of the contribution from the sand adhered to the pediatric participant [18]. Briefly, in the same area of the beach as the adult studies during two days in July and August of 2008, 14 individual toddlers wearing bathing suits over diapers spent 15 to 30 minutes on the beach sand (e.g. playing, sitting, lying, walking, etc). Following sand exposure, toddlers were placed in a 190-liter tub, while local off-shore marine water (14 L) was poured from sanitized watering cans gently over their heads and bodies. When necessary the toddlers were held upright in pool by an adult with either gloved hands or hands sanitized with alcohol. Sanitation of the pool and sample collections (in triplicate) were performed as described [18]. Source water, pool water and residual sand samples were analyzed as described below. The demographic characteristics of the 14 "Small Pool" toddlers (2 males and 12 females) included ages ranging from 5 to 47 months, and weights ranging from 6.8 to 16.3 kg [18].
Prior to study initiation, nasal cultures were obtained from the anterior nares from all participants using rayon swabs (BBL culture swab: Becton, Dickinson and Company) and S. aureus were cultured as described below.
Bacterial isolation and identification
S. aureus was isolated from the water samples using a standard membrane filtration (MF) method [19], followed by growth on selective media, Baird Parker agar (Becton, Dickinson and Company, Sparks, MD) with Egg Yolk (EY) Tellurite Enrichment (Becton, Dickinson and Company), BP, and CHROMagar, CHR (Becton, Dickinson and Company) (see Figure 1 for process flow). MSSA and MRSA isolated from BP plates were subjected to genetic tests and compared to organisms isolated from nasal cultures. MSSA isolated from CHR plates was used for direct comparison of colony counts obtained during previous studies. Facilities used for processing samples were located within minutes from the study site, allowing for the processing of samples within one hour after collection. Volumes of the source water used for filtration were 10 ml and 100 ml; volumes of the pool water samples used for filtration prior to and after adult participant contact were 10 ml and 50 ml respectively; volumes of the water used for filtration after contact with the pediatric participants were 5 ml, 10 ml, and 50 ml. Multiple volumes were filtered in order to obtain quantifiable colony counts as the levels of bacteria in both the source water and the experimental pool water samples were unknown.
The analysis of S. aureus in sand was similar to that for water with the exception of two pre-processing steps. The first step measured the water content of sand (weight difference of sand before and after drying at 110°C for 24 h). The second step extracted bacteria from the sand particles to a predefined volume of sterile water. To accomplish this, pre-weighed un-dried sand was aseptically removed from the corresponding sample container and placed into a sterile pre-weighed jar. One hundred and ten milliliters of sterile phosphate buffer saline (PBS) were added to each jar, and the jars were shaken vigorously for 30 seconds. The samples were permitted to settle for 30 seconds, and the supernatant was subsequently used for membrane filtration. One hundred milliliters of the sand eluate samples were used for the filtration and bacterial quantification.
Following standard MF, filter membranes were placed on BP and CHR, and incubated aerobically at 37°C for a minimum of 24 h. After incubation, colonies found to be black, shiny, convex, 2-5 mm in diameter, and surrounded by clear zones (BP) or mauve (CHR), were considered presumptive S. aureus, and subjected to confirmatory tests. All presumptive positive isolates were transferred to Mannitol Salt agar (Becton, Dickinson and Company), for the determination of mannitol fermentation, and incubated aerobically at 37°C for 16-24 h. All mannitol-fermenting isolates were enriched [20] on Trypticase Soy Agar with 5% Sheep Blood (TSA II, Becton, Dickinson and Company) for determination of colony morphology and gross pigmentation, the ability to lyse red blood cells and to provide bacterial cells for latex agglutination tests for clumping factor and protein A using the Remel BactiStaph Latex Agglutination Test (Thermo Fisher Scientific, Lenexa, KS).
The analysis of the nasal swab cultures focused on detection and genetic characterization, rather than quantification. The method used was the same as that used for the water samples, except that the membrane filtration step was omitted. Utilizing standard aseptic techniques swabs were placed in 0.5 ml tripticase soy broth TSB (Becton, Dickinson and Company) supplemented with 6.5% sodium chloride, and incubated overnight at 37°C for enrichment. One hundred micro-liters of the overnight broth were transferred to Mannitol Salt agar (Becton, Dickinson and Company), and the organisms were identified and confirmed as detailed above.
Chromosomal DNA was extracted from colonies isolated from water, sand, and nasal cultures. Whole cell extracts were prepared from latex agglutination positive bacterial isolates using the Amplicor MTB Sputum Specimen Preparation Kit (Roche Molecular Systems, Inc., Indianapolis, IN) according to the manufacture's recommendations, and used as template for confirming and characterizing polymerase chain reactions (PCR) as outlined below. These DNA extracts (up to a maximum of 22 per filter) were subjected to PCR analysis of the S. aureus specific gyrA gene for S. aureus confirmation and the mecA gene for genetic MRSA confirmation. Oligonucleotide primers and thermal cycling conditions were used as described previously [21], with the minor modification that 5-µl of whole cell extract was used as template in initial PCR reactions instead of purified chromosomal DNA.
All organisms determined to be genotypic MRSA (testing positive for mecA) were re-isolated from agar plates, and grown on oxacillin resistance screening agar base media ORSAB (Remel; Thermo Fisher Scientific), a selective media for confirmation of phenotypic MRSA. All genotypic MRSA isolates from this study showed the phenotypic characteristics of MRSA. All confirmed MRSA (n = 17) and MSSA (n = 162) collected from water and sand samples and all nasal cultures were stored as stock strains at -80°C.
The number of colonies testing positive for gyrA gene (for S. aureus counts) and mecA gene (for MRSA counts) were reported. Counts were then adjusted to colony forming units per 100 ml water (CFU/100 ml) or per 100 g sand (CFU/100 g) using the volume of water applied to the filters or the weight of the sand collected from the pool. The numbers of microbes shed per person were determined by multiplying the difference in microbial concentrations measured before and after bathing in the pools by the water volumes corresponding to each person.
Genetic characterization
Bacterial isolates determined to be positive for S. aureus specific gyrA and MRSA specific mecA were subjected to additional PCR to test for the toxin genes for Panton-Valentine leukocidin, pvl, to evaluate the pathogenic potential of isolated organisms as previously described [21]. Staphylococcus cassette chromosome methicillin, SCCmec, type was determined for all MRSA as described [22]; and Staphylococcus protein A, spa, type was determined for all MRSA and a representative subset of MSSA as described [23] and using RIDOM spa type server to analyze sequences. SCCmec and spa types were determined in order to provide specific genetic characteristics for comparison between the nasal isolates from the colonized participants and the water samples from the pools. Genetic characteristics were evaluated for all MRSA isolates and all MSSA isolates from the nasal swabs, and from the water and sand samples from the small pool. Due to the large number of S. aureus isolates from the large pool, genetic characterization was conducted on a representative set of the MSSA isolates from each large pool water collection and choosen to include a subset of all colony morphology, gross pigmentation and RBC hemolysis type present in each set. All MSSA collected from the small pool water samples from the single colonized pediatric participant were analyzed.
Statistical analyses
Data analyses (including Pearson Correlations, Student T-Tests, and Sum Rank Tests) were performed using Microsoft Excel 2003 and Sigmaplot 11.