Environmental Surveillance and Characterization of Antibiotic Resistant Staphylococcus aureus at Coastal Beaches and Rivers on the Island of Hawaiʻi

Staphylococcus aureus are human facultative pathogenic bacteria and can be found as contaminants in the environment. The aim of our study was to determine whether methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-susceptible S. aureus (MSSA) isolated from coastal beach and river waters, anchialine pools, sand, and wastewater on the island of Hawaiʻi, Hawaiʻi, are a potential health risk. Samples were collected from three regions on Hawaiʻi Island from July to December 2020 during the COVID-19 pandemic and were characterized using whole-genome sequencing (WGS). From WGS data, multilocus sequence typing (MLST), SCCmec type, antimicrobial resistance genes, virulence factors, and plasmids were identified. Of the 361 samples, 98.1% were positive for Staphylococcus spp. and 7.2% were S. aureus positive (n = 26); nine MRSA and 27 MSSA strains were characterized; multiple isolates were chosen from the same sample in two sand and seven coastal beach water samples. The nine MRSA isolates were multi-drug resistant (6–9 genes) sequence type (ST) 8, clonal complex (CC) 8, SCCmec type IVa (USA300 clone), and were clonally related (0–16 SNP differences), and carried 16–19 virulence factors. The 27 MSSA isolates were grouped into eight CCs and 12 STs. Seventy-eight percent of the MSSA isolates carried 1–5 different antibiotic resistance genes and carried 5–19 virulence factors. We found S. aureus in coastal beach and river waters, anchialine pools, and sand at locations with limited human activity on the island of Hawaiʻi. This may be a public health hazard.


Introduction
Staphylococcus aureus is a facultative pathogen that can be part of the normal flora of humans, as well as some animal species. They have been found in domestic and wild animals, as well as in a variety of environments [1][2][3][4][5][6]. In the United States, S. aureus has been isolated from coastal beach waters, freshwaters, and sand in California [7], Florida [5,8,9], Washington State [3,10,11], Ohio [4], the Great Lakes [12], and Hawai'i [13][14][15][16]. Methicillinresistant Staphylococcus aureus (MRSA)-and methicillin-susceptible S. aureus (MSSA)-related mortality rates have been estimated at 14 and 11 deaths per 100,000 hospitalizations, respectively, in the United States [17]. Previous studies have shown a higher prevalence of MRSA infections in the State of Hawai'i compared to the national average [18,19]. Native Hawaiians, Pacific Islanders, and children are disproportionately affected and have a higher infection rate of community-acquired MRSA infections [13,20,21]. A steady increase in the proportion of MRSA, compared to MSSA infections, has been reported in inpatient and outpatient clinics, jails, and nursing homes in Hawai'i [13,20,22,23].
Marine waters and beaches have been proposed as possible sources for S. aureus infections, but only a few studies have examined the risk to beachgoers in Hawai'i and abroad. A correlation between swimmer density and staphylococci isolated from seawater has previously been demonstrated [24]. This resulted in four times higher odds of developing staphylococcal skin infections, as well as increases in respiratory and total illnesses in humans exposed to seawater compared to those who were not; however, the association could not always be confirmed [5,[25][26][27].
The aim of the current study was to assess the public health hazard of S. aureus found in coastal beach and river waters, anchialine pools, sand, and wastewater on Hawai'i Island, Hawai'i, by characterizing the isolates using whole-genome sequencing (WGS). This is the first study in the State of Hawai'i to characterize MRSA and MSSA isolates using WGS from environmental sources. As this project started, the COVID-19 pandemic began closing tourism to the State of Hawai'i which severely reduced beach activities of residents and visitors through lockdowns and limitations on public gatherings. Consequently, this study may underrepresent the full extent of S. aureus contamination in coastal and aquatic Hawaiian environments.

Genetic Characterization of MRSA and MSSA Isolates
Thirty-six S. aureus isolates included 12 different STs (Figures 2 and 3; Table 2). All nine MRSA isolates [coastal beach water (n = 5), anchialine pool (n = 1), sand (n = 3)] were ST8 (CC8). These isolates were highly clonal based on the core genome SNP analysis (0-16 differences) (Figure 3), despite being isolated from two different districts on the island. The MRSA isolates obtained from the environment in this study were compared with the sequences of clinical MRSA isolates from hospital patients in the State of Hawai'i dating back from 2011 [2]. These isolates differed by 54-144 SNPs (Figure 4), and can thus be regarded as clonally related.

Genetic Characterization of MRSA and MSSA Isolates
Thirty-six S. aureus isolates included 12 different STs (Figures 2 and 3; Table 2). All nine MRSA isolates [coastal beach water (n = 5), anchialine pool (n = 1), sand (n = 3)] were ST8 (CC8). These isolates were highly clonal based on the core genome SNP analysis (0-16 differences) (Figure 3), despite being isolated from two different districts on the island. The MRSA isolates obtained from the environment in this study were compared with the sequences of clinical MRSA isolates from hospital patients in the State of Hawaiʻi dating back from 2011 [2]. These isolates differed by 54-144 SNPs (Figure 4), and can thus be regarded as clonally related.

Plasmids
Among the 36 S. aureus strains, nine replicon families (rep family) were detected. The rep genes were on the same contig as the AMR and/or virulence genes ( Table 3). The MRSA isolates contained 2-4 rep genes. Among the MSSA isolates, 12 contained no rep family genes, 14 contained a single rep family, and 1 contained 2 rep families ( Table 3 The tet(K) gene was associated with rep 7a for ST1181 (isolate #79) and ST5 (isolate #111). While two MSSA isolates carried the mph(C) gene, only one had rep 19 . Of the six MSSA isolates that carried the blaZ gene, (3/6) were associated with rep 16 , (2/6) associated with rep 20 , and (1/6) had an unknown rep sequence. The ACME cluster on a single MRSA isolate was associated with rep 7c . The aur virulence gene from three MRSA and two ST72 were associated with rep 7c . Two ST5 carried the sed and sej toxin genes and were associated with rep 20 and found on the SAP074A plasmid.

Discussion
In this study, we isolated both MRSA and MSSA from known clonal complexes. The number of positive water samples for S. aureus in our study was lower (7.2%) than expected compared to study sites in Florida, which had a 37% positive rate for MSSA and 1% positive rate for MRSA for coastal waters [5]. Similarly other Flordia studies have found more MSSA vs. MRSA in their samples [8,9]. The lower prevalence rate of S. aureus compared to the other studies [4,14,15] are most likely a result of the COVID-19 pandemic, whereby the beaches were much less populated with beachgoers and toruists (Hawai'i Tourism Authority 2020 https://www.hawaiitourismauthority.org/media/6395 /december-2020-visitor-statistics-press-release-final.pdf, accessed on 8 August 2021). This is reflected by studies in Hawai'i that had a prevalence up to 92% for S. aureus in aquatic environments [14][15][16]. Nevertheless, the strains isolated are of interest as they are most likely contaminated by the local population rather than tourists.
The MRSA, despite being isolated from two different districts on the island (Figure 2), were clonally related and thus, probably indigenous to the region, as they were also clonally related to strains previously isolated in Hawai'i State 10 years prior and from a different island [2]. It is clear that this MRSA clone had been circulating on the Hawaiian Islands for a long time. However, limited data are available on MRSA strains circulating in Hawai'i. A study conducted in 2003-2004 on the island of O'ahu, Hawai'i typed 12 environmental MRSA isolates via PFGE; seven of the MRSA contained recognizable pulse field types, including: USA100 (4/12), USA300 (1/12), USA1000 (1/12), and USA1100 (1/12) [16]. Another study in 2004 collected 40 MRSA isolates from skin and soft tissue infections from hospital patients; 65.0% (26/40) were classified as USA300 and of those, 84.6% (22/26) were classified as USA300-0114 [13]. Thus, while a diversity of MRSA was found in the mid-2000 s, the dominant MRSA strain in the past decade appears to be the USA300 clone. Nevertheless, given the high infection rate with MRSA in Hawai'i [18,19], this indicates that the strain is also widespread amongst the population, as beachgoers during our sampling period were primarily local residents.
The MRSA in this study was the typical USA300 ST8 SCCmec IVa community-acquired MRSA, carrying the PVL toxin, which is a very prevalent strain in North America and the strains, though an outgroup, cluster together with the USA300 "North American Epidemic" (USA300-NAE) clone [28]. Moreover, an episode of acute increase of MRSA infections in a coastal community in Hawai'i has been described in the past [29]. Suggestive of a clear public health hazard associated with this strain circulating on coastal beaches. The USA300 strains have also been detected by Pulse Field Gel Electrophoresis (PFGE) analysis in the past, but only in low numbers [16]. This indicates that this clone has been spreading successfully on the island of Hawai'i. Nevertheless, future studies should include human S. aureus isolates, in connection with epidemiological data, to examine whether there is a link between recreating in S. aureus-contaminated coastal waters and the onset of skin infections. Aside from two older publications, there has been limited indication of an increased risk of developing staphylococcal skin infections as a result of recreating in staphylococci-contaminated waters and beach sand in Hawai'i [24,27]. Recreational swimmers in Hawai'i may be disproportionately affected by S. aureus infections, particularly on Hawai'i Island, due to the presence of rocky beaches and coral reefs where skin abrasions/wounds are common. These have previously been identified as an important risk factor for acquiring staphylococcal infections [30]. S. aureus has also been found in rivers during storms [15], which attracts surfers who may be disproportionately at risk for acquiring these infections [31].
Overall, this indicates that colonization of S. aureus strains may occur at coastal beaches and may, subsequently, spread in the community and cause infections. Additional genomic information of currently circulating strains infecting humans on the island of Hawai'i are necessary to bring more clarity to the situation.
The only former WGS data on AMR genes in MRSA from Hawai'i indicated that the eight USA300 strains contained 2-9 different antibiotic resistant genes [2]. The aac(6')-aph(2"), and mupA genes were not detected in the MRSA isolates in the current study. The AMR genes [erm(T), lnu(A), qacD, qacG, and tet(K)] found in the current study were not found in the clinical isolates. This suggests that during the period of 10 years, several additional resistance genes were acquired. However, additional studies are necessary to better assess the resistance genes present.
The strains in this study and the formerly isolated eight MRSA [2] carried the same virulence genes, except for the ACME cluster, which was carried in only 50% (4/8) of the clinical MRSA compared to 100% (9/9) of the environmental MRSA isolates in the current study. Indeed, the current USA300-NAE strains are characterized by the presence of the ACME cluster, while the South American clones are ACME negative [28].
The MSSA strains in this study were more diverse, less resistant, and carried fewer virulence genes. The MSSA isolates included closely related ST72 strains from different locations. Similarly, some of the ST5 strains were related, and others were less related to other ST5 strains. The ST518 strains were very related. Moreover, the ST398 strains were nearly the same despite being isolated from all regions sampled. This suggests that these strains may also be endemic in Hawai'i, but due to the lack of data, it cannot be stated for certain. The MSSA all belong to well-known clones, such as ST398 MSSA, which is human-associated unlike the MRSA ST398 [32] and carries the erm(T) gene, which are not typically found in livestock-associated strains [33]. Furthermore, future in-depth studies should also focus on the epidemiology of MSSA, as these strains account for the majority of invasive S. aureus infections in the United States resulting in considerable morbidity and mortality, [34] and may also pose a health risk for beachgoers.
The AMR genes have been shown to be located on different mobile genetic elements in MRSA and MSSA for blaZ. Indeed, blaZ is located, in general, on a transposon and may jump between different locations [35]. Similarly, the erm(T) gene, originally described in an MSSA ST398 [36], has been found on different plasmids, though to our knowledge never on USA300 strains. Unfortunately, it is unknown what the prevalence of ST398 is in Hawai'i. However, we found it rather abundantly in this study, and it is a human-associated strain. It appears to be quite prevalent, and the erm(T) gene might have been transferred from ST398 to ST8, ST5, and ST518 (Table 2).
Few plasmids are common between the MRSA and MSSA, though it is also clear that the MRSA population is quite clonal and may represent a bias in the diversity. Nevertheless, it seems that the USA300 clone is also the predominant clone on the island. However, there are also clear indications of the exchanges of rep 7c , rep 7a and rep 19 plasmids. While these plasmids have the same replicon type, they do not always carry the same resistance genes. It is clear that the MRSA are able to more easily acquire plasmids as they are overwhelmingly present in the MRSA opposed to MSSA strains.

Study Locations and Sampling
Water and sand samples were collected multiple times at 36 stations across the three most populous districts in Hawai'i County, including the most populated district, Hilo, and the less populated ones of South Kohala, North Kona, and Puna ( Figure 5). Hilo stations were chosen adjacent to and within the Hilo Bay watershed, spanning~12 km of coastline, including the Wailuku and Wailoa Rivers and Honoli'i Stream. A majority of households in Hilo rely on decentralized wastewater management, where~8700 cesspools leak~21.2 million liters/day into groundwater, which contaminates nearshore coastal waters [37]. In Hilo, there is a wastewater treatment plant that utilizes a series of gravityfed systems with pump stations and provides primary and secondary treatment, where effluent is discharged offshore via an ocean outfall. As a result of heavy rainfall, cesspool seepage and sewage overflows periodically occur representing a potential health hazard and thus, sewage was sampled. Twenty-one stations were selected, which included 14 State and county beaches (stations #1-12, #16, #20; 188 samples) seven river/stream stations (76 samples) along the Wailoa River (stations #13-15), Wailuku River (stations #17- 19), and Honoli'i Stream (station #21). Sand was collected at five stations in Hilo (stations #1-2, #11, #16, #20) each on five occasions (25 samples). Wastewater influent, effluent, and treated effluent collected from a manhole was collected monthly over five months (15 samples) from the Hilo Wastewater Treatment Plant ( Figure 5A). and thus, sewage was sampled. Twenty-one stations were selected, which included 14 State and county beaches (stations #1-12, #16, #20; 188 samples) seven river/stream stations (76 samples) along the Wailoa River (stations #13-15), Wailuku River (stations #17- 19), and Honoliʻi Stream (station #21). Sand was collected at five stations in Hilo (stations #1-2, #11, #16, #20) each on five occasions (25 samples). Wastewater influent, effluent, and treated effluent collected from a manhole was collected monthly over five months (15 samples) from the Hilo Wastewater Treatment Plant ( Figure 5A).  Figure 5B).
Puna is a rapidly developing district located southeast of Hilo, and in 2018, a black sand beach and numerous anchialine pools formed at a popular surfing spot (Pohoiki) following a volcanic eruption [38]. Two county beaches (stations #31-32; 10 samples) and  Figure 5B).
Puna is a rapidly developing district located southeast of Hilo, and in 2018, a black sand beach and numerous anchialine pools formed at a popular surfing spot (Pohoiki) following a volcanic eruption [38]. Two county beaches (stations #31-32; 10 samples) and four anchialine pools (stations #33-36; 10 samples) were sampled. Sand was collected at two stations in Puna, each on three occasions (stations #31-32) (6 samples) ( Figure 5C).
A total of 361 samples were collected and processed from coastal beach waters (n = 225), rivers/stream (n = 76), anchialine pools (n = 10), sand (n = 35), and wastewater (n = 15) between July-December 2020 ( Figure 5). Liquid samples were collected in sterile, acid-washed polypropylene plastic bottles between 6 AM and noon, stored on ice, and processed within 6 h of collection. Sand samples were collected in sterile, polypropylene bottles using a metal spade 1-2" deep, within 1 m of the high tide line, transported on ice, and processed within 6 h.

Isolation and Quantification of Staphylococci
Water samples were cultured for Staphylococcus spp. using broth enrichment (1.5 X m Staphylococcus broth (Becton, Dickinson, and Company, Sparks, MD, USA), 75 µg/mL polymyxin B, and 0.01% potassium tellurite) in Quanti-Tray 2000 ® (IDEXX Laboratories, Westbrook, ME, USA) and incubated for 72 h at 37 • C, as previously described [3]. Positive wells contained a black precipitate [3,10]. The number of positive wells were counted and the presumptive Staphylococcus spp. MPN/100 mL was calculated using the MPN IDEXX chart and adjusted for dilution. From each Quanti-Tray 2000 ® sample, ≤1 mL of broth from positive wells were placed into a 1.5 mL microcentrifuge tube and vortexed for 15 s. An aliquot was inoculated onto both mannitol salt agar (Becton, Dickinson, and Company) and Oxacillin Resistance Selective Agar Base (ORSAB) with supplement (Thermo Fisher Scientific TM Oxoid TM Products, Lenexa, KS, USA), and incubated at 37 • C to screen for MRSA and MSSA, respectively. Presumptive MRSA/MSSA colonies were inoculated onto blood agar (Tryptone Soya agar with sheep blood) (Thermo Fisher Scientific TM Remel TM Products), incubated at 37 • C for 18-24 h in a candle extinction jar and screened for βhemolysis. Presumptive MRSA and MSSA isolates were identified using the Staphaurex TM latex agglutination test kit (Thermo Fisher Scientific TM Remel TM Products). Presumptive MRSA isolates were biochemically identified using the penicillin-binding protein (PBP2) latex agglutination test kit (Thermo Fisher Scientific TM Remel TM Products). A single MRSA/MSSA isolate was stored at 4 • C until further characterized. In seven (30%) of the 23 positive water (beach/river/anchialine pool) samples, two different wells contained S. aureus, which were characterized to assess diversity in the STs (Table 2).
Sand samples were processed using modified methods, as previously described [39]. Briefly, 10 g of wet sand from each station were shaken with 200 mL of 0.15 M NaCl for two min at 100 rpm where 25 mL of the supernatant was transferred into a sterile bottle, and sterile water added for a final volume of 50 mL. This volume was then combined with 50 mL of the enrichment broth and processed, as described above in IDEXX trays [3]. Up to 20 mL of wastewater was transferred; sterile water was added for a final volume of 50 mL. Samples were then combined with 50 mL of enrichment broth and placed in Quanti-Tray 2000 ® and processed, as previously described [3]. In two (67%) of the three positive sand samples, 2-3 different wells contained S. aureus, which were characterized to assess diversity in the STs (Table 2).

Conclusions
The beaches in Hawai'i Island contaminated both MRSA and MSSA. The MRSA in this study were the typical USA300 strains, which are likely circulating in the community. The MRSA strains were very clonal. The MSSA strains isolated in this study were less resistant and have less virulence traits; however, there is also a lot of clonality amongst the strains. This indicates that the beaches are a potential health risk for both MRSA and MSSA infections in humans.