Long-term analysis of Vibrio vulnificus and Vibrio parahaemolyticus virulence factors and their environmental associations in the Chesapeake Bay, Maryland, US

Abstract

Aims

The present study analyzed long-term trends in Vibrio spp. virulence factors associated with pathogenicity in Chesapeake Bay waters (in Maryland) across two 3-year sampling periods (2009–2012 and 2019–2022).

Methods and results

Vibrio parahaemolyticus (n = 1476) and V. vulnificus (n = 806) isolates were recovered from 13 sampling stations and tested for species-specific and virulence markers. Across both sampling periods, detection of V. parahaemolyticus tdh⁺/trh⁺ isolates was similar and more frequently observed in the spring and fall, and the upper sampling stations. However, significant differences were observed in the detection of V. vulnificus pilA⁺ and rtxA⁺ isolates, with a higher percentage of pilA⁺ in 2019–2022 and rtxA⁺ in 2009–2012. VcgC⁺ isolates were most frequently detected during the fall (n = 24%) in 2019–2022. Interestingly, a greater prevalence of V. vulnificus isolates harboring virulence markers was observed in the mid and lower stations.

Conclusions

Vibrio vulnificus isolates displayed greater changes in virulence marker frequency across sampling periods compared to V. parahaemolyticus and these markers were detectable throughout the year.

Impact Statement.

Incidence of potentially pathogenic Vibrio spp. remained relatively stable between 2009 and 2022. Nevertheless, exposure risk to pathogenic Vibrio strains in Maryland Bay waters is elevated during the summer, and may extend into winter, varying by location.

Introduction

Vibrio spp. are ecologically important Gram-negative bacteria that are ubiquitous in marine and estuarine environments, and flourish in warm water with moderate salinity (Singleton et al. 1982, Thompson et al. 2004). Vibrio spp. incidence is strongly influenced by environmental parameters (Colwell 1996, Vezzulli et al. 2013, Takemura et al. 2014, Brumfield et al. 2021, Sampaio et al. 2022), and they are known to be associated with aquatic invertebrates, especially zooplankton, crustaceans, and bivalves (Kaneko and Colwell 1973, Hlady 1997). Vibrio spp. are clearly beneficial to aquatic ecosystems, contributing to the degradation of polymeric substances, namely chitin, and aiding in carbon and nitrogen cycling (Zhang et al. 2018). Nonetheless, a few species are known to cause public health emergencies. In particular, Vibrio cholerae is the well-known etiologic agent of cholera that continues to plague many low- and middle-income countries (Colwell 1996). However, in the United States and other high-income countries, pathogenic non-cholera Vibrio species, including Vibrio parahaemolyticus and Vibrio vulnificus, are more likely to cause disease (Baker-Austin et al. 2017). Symptoms of non-cholera Vibrio illness (vibriosis) range from mild gastroenteritis to more severe cases that include wound infection and septicemia, leading to hospitalization and even death (Daniels and Shafaie 2000).

According to the Centers for Disease Control and Prevention (CDC), Vibrio bacteria cause an estimated 80 000 illnesses each year in the U.S., with approximately 40% of cases being attributed to V. parahaemolyticus, primarily transmitted via consumption of raw or undercooked seafood (notably oysters) (Iwamoto et al. 2010, Ndraha et al. 2020, CDC 2023). In addition, V. vulnificus is responsible for an estimated 150–200 cases each year but has a case-fatality rate among the highest of any waterborne pathogen, up to 50%, and infection is more commonly associated with exposure to brackish waters containing the bacterium (Oliver 2013, Baker-Austin and Oliver 2018). The pathogenicity of Vibrio spp. strains appears to be associated with their ability to produce multiple primary virulence factors that are often complex in nature and not yet entirely understood (Raghunath 2014, Choi and Choi 2022).

Thermostable direct hemolysin (tdh) is considered a major virulence factor of V. parahaemolyticus, capable of causing cytotoxicity and hemolytic activity, and is found in over 90% of all clinical strains (Nishibuchi et al. 1992, Hiyoshi et al. 2010). Thermostable direct-related hemolysin (trh) has also been identified as another significant virulence factor, which confers strains with the ability to produce urease (DePaola et al. 2000). Although the presence of these genes is typically used to identify pathogenic V. parahaemolyticus isolates, previous studies have found that not all disease-causing isolates carry tdh and/or trh, and other factors likely play a role in pathogenicity (Mahoney et al. 2010, Jones et al. 2012, Raghunath 2014). On the other hand, no single virulence gene that can identify all pathogenic V. vulnificus strains has been found, and the mechanisms involved appear to be more complex (Baker-Austin and Oliver 2018). Nevertheless, V. vulnificus isolates have been historically classified into biotype 1, responsible for most human infections, and further categorized into genotypes vcgC (clinical origin) and vcgE (environmental origin); biotype 2, a notable eel pathogen; and biotype 3, primarily found in Israel (Strom and Paranjpye 2000, Baker-Austin and Oliver 2018). Of the biotype 1 genotypes, the vcgC gene has been shown to strongly correlate (around 90%) with clinical type isolates and may provide a suitable indicator of virulence potential (Rosche et al. 2005, Çam et al. 2019). To date, the most significant virulence factors appear to be related to the ability of these bacteria to form biofilms and produce exotoxins, including hemolysin cytolysin (vvha) found in most strains, the repeats-in-toxin (RTX) (rtxA), and type IV pili (pilA) (Li and Wang 2020, Choi and Choi 2022). In particular, rtxA may be the strongest cytotoxic virulence factor of V. vulnificus that plays a significant role in sepsis as well as bacterial dissemination in host cells (Liu and Crosa 2012, Li and Wang 2020), whereas the presence of pilA in isolates may allow for greater transmission and dissemination in host organisms, such as oysters (Choi and Choi 2022).

Over the past two decades, there has been an increase in the annual incidence of vibriosis across the USA (Sims et al. 2011, Weis et al. 2011, Newton et al. 2012, Archer et al. 2023, Morgado et al. 2024). Multiple studies reported a significant geographic expansion of pathogenic Vibrio spp. and increased risk of infection, especially in more northern ocean and brackish waters, such as the eastern seaboard of the USA and Northern Europe (Baker-Austin et al. 2013, Vezzulli et al. 2013, Sterk et al. 2015, Ford et al. 2020, Fleischmann et al. 2022, Baker-Austin et al. 2024). There are also concerns about the impact of climate-related environmental changes (e.g. increased sea surface temperature) and the role of extreme weather events (e.g. hurricanes, extreme heat events) on prevalence and proliferation of Vibrio spp. worldwide (Baker-Austin et al. 2017, Deeb et al. 2018, Logar-Henderson et al. 2019, Froelich and Daines 2020, Brehm et al. 2021, Brumfield et al. 2023b). To this effect, a recent culture-dependent investigation of Vibrio spp. in water and oyster samples recovered from the Chesapeake Bay, the largest estuary in North America, reported a long-term increase and extended seasonality of V. parahaemolyticus and V. vulnificus across two time periods, namely 2009–2012 and 2019–2022 (Brumfield et al. 2023a). However, characterization of Vibrio spp. isolates, including carriage of clinically important virulence biomarkers, was not fully evaluated. Thus, the present study expanded on previous work by analyzing long-term trends in the occurrence of Vibrio spp. virulence factors in the Chesapeake Bay and potential implications for human health.

Materials and methods

Site description

Water and oyster samples were collected from the Chesapeake Bay, Maryland, USA, during two separate 3-year sampling events, namely 2009–2012 and 2019–2022, described in detail previously (Chen et al. 2017, Brumfield et al. 2023a). Briefly, between June 2009 and August 2012, samples were collected at two locations, Chester River and Tangier Sound, and between April 2019 and August 2022, samples were collected from 12 stations across the Chesapeake Bay (Fig. 1). The Chesapeake Bay is a commercial and recreationally important estuary characterized by a strong salinity gradient from the head of the Bay near fresh tidal rivers in the north to the opening of the Atlantic Ocean in the south, where salinities average 25–30 ppt (Roman et al. 2005). All sampling stations were located in the Maryland portion of the Bay and were divided into the following sections based on latitude and average salinity: upper (Chester River, Sandy Point, South River, Cox Creek), mid (Miles River, Broad Creek, Choptank River, Upper Patuxent River, Lower Patuxent River, Wicomico River), and lower (St. Mary’s River, Tangier Sound).

Figure 1.

Map of the Chesapeake Bay, Maryland, showing locations for both sampling periods (2009–2012 and 2019–2022) (adapted from Brumfield et al. 2023a).

Sample collection and processing

Chen et al. (2017) and Brumfield et al. (2023a, 2023b) have previously described sample collection and processing methods that are summarized in this study. Briefly, physical and chemical measurements were collected during each sampling event using a handheld water probe (Eureka, Austin, TX), including water temperature, pH, dissolved oxygen, salinity, and chlorophyll-a. Total chlorophyll-a concentration for the 2009–2012 sampling period was measured in methanol extracts on a Cary model 50 UV-visible-light spectrophotometer, whereas it was measured in acetone extracts on a Shimadzu UV 2401 PC spectrophotometer for the 2019–2022 sampling period. Between 2009 and 2012, all sample collection occurred twice a month during the summer (June through August), and once a month for the rest of the year, and between 2019 and 2022, samples were collected weekly during the summer and twice a month during the rest of the year. For the latter, restrictions imposed during the COVID-19 pandemic impacted the collection of samples from March through July of 2020.

Water samples (12 L) were collected just below the surface using a Van Dorn water sampler (WildCo, Buffalo, NY) and stored in a sterile Nalgene Carboy (Thermo Fisher Scientific, Waltham, MA). The carboy was rinsed three times with sample water from the site prior to collection. Up to 30 oysters were collected by dredging and stored in clean freezer bags. Samples were then transported to the laboratory in a cooler with ice packs, ensuring that samples did not come in direct contact with the ice packs, and kept refrigerated at 4°C until processing (<8 hrs). Sample processing followed methods described in the Bacteriological Analytical Manual for food sampling/preparation of sample homogenate (Andrews and Hammack 2022) and Vibrio detection (Kaysner et al. 2019), with modifications detailed in Brumfield et al. (2023b). Water samples were shaken vigorously 25 times in a 30 cm arc in 7 s. Oysters were rinsed and scrubbed under running deionized water to remove debris from shells and opened using a sterile shucking knife. Oyster tissue (measured in volume of up to 500 mL) in a 1:1 mixture with phosphate buffer saline (PBS) (pH 7.4), was homogenized in a sterile blender for 90 s. Following this, homogenized oyster and shaken water samples were added at various concentrations into alkaline peptone water (APW; 10% peptone, 1% NaCl [pH 8.5]). Briefly, three volumes of water (1 L, 100 mL, and 10 mL) and homogenized oyster tissue (10 mL, 1 mL, and 100 µL) were resuspended in APW (10X), which was diluted with sterile PBS (for oyster tissue) and seawater (for water samples) to ensure a final APW (1X) concentration, and incubated at 37°C for 16–18 hrs, with moderate aeration (orbit diameter 2.5 cm; 0.025 g).

Vibrio spp. isolation

Following initial incubation, a loopful from each APW-enriched water and oyster sample was subcultured onto selective media, including CHROMagar^TM  Vibrio (CHROMagar, Springfield, NJ), thiosulfate citrate bile salts sucrose agar (TCBS, Oxoid, Ontario, Canada), and M190 V. vulnificus agar [VVA; 2% peptone, 3% NaCl, 1% cellobiose, 0.06% bromothymol blue (pH 8.2)], and incubated at 37°C for 16–18 hrs. Up to 36 presumptive positive colonies were then selected for each Vibrio species per sample. These presumptive colonies were purified using Luria-Bertani (LB) agar (Becton, Dickinson and Co., Sparks, MD), and after PCR confirmation (see details below), an average of seven V. parahaemolyticus and five V. vulnificus colonies were confirmed per sampling event and archived in LB broth containing 25% (vol/vol) glycerol at −80°C.

Vibrio species confirmation

DNA was extracted from presumptive isolates of V. parahaemolyticus and V. vulnificus following methods previously described in Chen et al. (2017) and species were identified preliminarily using PCR targeting species-specific variations within the toxR biomarker, as described previously by Bauer and Rørvik (2007). Samples testing positive for toxR (V. parahaemolyticus, toxR_Vp, or V. vulnificus, toxR_Vv) were also screened for carriage of species-specific genes and virulence markers [V. parahaemolyticus: thermolabile hemolysin (tlh), thermostable direct hemolysin (tdh), thermostable direct-related hemolysin (trh); V. vulnificus: hemolysin cytolysin (vvha), virulence correlated gene environmental variant (vcgE), virulence correlated gene clinical variant (vcgC), RTX toxin (rtxA), type IV pili (pilA)]. It is worth noting that vcgC⁺ and vcgE⁺ data were not available for isolates recovered during the 2009–2012 sampling period.

PCR assays were performed using Promega GoTaq Green Master Mix 2x (Promega, Madison, WI, USA), and each reaction well contained a total of 25 µL, including 20 µL of mastermix solution (12.5 µL of GoTaq, 1 µL of each primer, and nuclease-free water to reaction volume) and 5 µL DNA template. The primer sequences, amplicon size, and thermal cycler conditions used for each PCR are detailed in Table 1.

Table 1.

List of primer sequences, amplicon size, and PCR conditions used for the detection of V. parahaemolyticus and V. vulnificus species-specific and virulence genes.

Primers	Primer sequence (5′-3′)	Amplicon (bp)	PCR conditions	Source
utox-F	GASTTTGTTTGGCGYGARCAAGGTT			Bauer and Rørvik (2007)
vptox-R	GGTTCAACGATTGCGTCAGAAG	297	95°C for 4 m; 34×: 95°C for 30 s, 55°C for 30 s, 72°C for 60 s; 72°C 5 m
vvtox-R	AACGGAACTTAGACTCCGAC	435	95°C for 4 m; 30×: 95°C for 30 s, 60°C for 30 s, 72°C for 60 s; 72°C 7 m
tlh-F	AAAGCGGATTATGCAGAAGCACTG	173	94°C for 3 m; 30×: 94°C for 60 s, 58°C for 60 s, 72°C for 60 s; 72°C 5 m	Bej et al. (1999), Rizvi and Bej (2010)
tlh-R	TGTGCCTTGATGAACTCGTTC
tdh-F	GTAAAGGTCTCTGACTTTTGGAC	270
tdh-R	TGGAATATGAACCTTCATCTTCACC
trh-F	TTGGCTTCGATATTTTCAGTATCT	500
trh-R	CATAACAAACATATGCCCATTTCCG
vvh-F	AGCGGTGATTTCAACG	411	94°C for 3 m; 34×: 94°C for 30 s, 55°C for 30 s, 72°C for 30 s; 72°C 5 m	Warner and Oliver (2008), Panicker and Bej (2005), Panicker et al. (2004), Rosche et al. (2005), Brasher et al. (1998)
vvh-R	GGCCGTCTTTGTTCACT
vcgC-F	AGCTGCCGATAGCGATCT	97	94°C for 3 m; 30×: 94°C for 40 s, 57°C for 40 s, 72°C for 40 s; 72°C 5 m
vcgC-R	TGAGCTAACGCGAGTAGTGAG
vcgE-F	CTCAGAAAGGCTCAATTGAC	199
vcgE-R	GATTAACGCTGTAAGGCCG
pilA-F	TGGCTGCTGTTGCTATTC	217	94°C for 3 m; 30×: 94°C for 60 s, 60°C for 60 s, 72°C for 60 s; 72°C 5 m	Natividad-Bonifacio et al. (2013), Paranjpye and Strom (2005)
pilA-R	GGTCCACCACTAGTACCAAC
rtxA-F	CGGGATCCTATGGCGTGAACGGCGAAG	1440	94°C for 3 m; 30×: 94°C for 30 s, 68°C for 30 s, 72°C for 60 s; 72°C 5 m	Natividad-Bonifacio et al. (2013), Lee et al. (2008), Kim et al. (2008)
rtxA-R	CGGGATCCAGCAGCCACAAGCGATTC

PCR products were kept at 4°C until visualization via gel electrophoresis. Positive controls included the following strains: V. parahaemolyticus ATCC 17803 (toxR⁺), NIHCB0757 (tlh⁺/tdh⁺), AQ 4037 (tlh⁺/trh⁺), and V. vulnificus ATCC 27562 (toxR⁺/vcgE⁺/vvha⁺), ATCC 29307 (vcgC⁺/pilA⁺/rtxA⁺). Nuclease-free water was used as a negative control in each reaction. PCR products were stained using BioLink Smart Glo Pre-stain and visualized using a 1.5% agarose gel at 110 V for 60–90 min and viewed under a UV transilluminator using a Gel Documentation System (GelDoc-IT^TM, UVP, LLC, CA, USA).

Data Analysis

The Wilcoxon rank sum and Kruskal–Wallis tests were used to evaluate differences between environmental water parameters (water temperature, salinity, dissolved oxygen, pH, chlorophyll-a) measured in each sampling period (2009–2012 and 2019–2022) and season. A paired t-test and the repeated measures ANOVA (season and season*year interaction as the fixed effects; year as the repeated measure) were used to assess differences in V. parahaemolyticus and V. vulnificus virulence factors across seasons for each sampling period.

The relative frequency (%) of Vibrio spp. virulence factors per sampling event was calculated based on their frequency of detection relative to the total number of isolates tested in samples from both sampling periods. The Kruskal-Wallis test was used to evaluate Vibrio spp. virulence factor differences between sections of Maryland Bay waters (upper, mid, lower) for both sampling periods combined. The non-parametric two-sample t-test (Wilcoxon Rank Sum test) was used to assess differences between the detection (absence/presence) of Vibrio spp. virulence factors and changes to environmental parameters. Correlation and regression analyses were also applied to evaluate the association between the relative frequency of Vibrio spp. virulence factors and environmental parameters. All statistical analyses were performed using SAS 9.4 (Cary, NC, USA).

Results

Water quality trends

In total, there were 196 sampling events: 110 sampling events between 2009 and 2012 and 86 between 2019 and 2022. All sampling events resulted in the successful recovery of Vibrio spp. isolates. However, in this study, we included only those sampling events positive for V. parahaemolyticus and/or V. vulnificus (n = 82 from 2009–2012 and n = 80 from 2019–2022) (Table 2).

Table 2.

Water sample characteristics by sampling period (2009–2012 and 2019–2022), season, and section of the Chesapeake Bay, including number of sampling events in the study, average temperature (T), salinity (S), pH, dissolved oxygen (DO), and chlorophyll-a concentrations (chla) per sampling event with ± standard deviation.

Study/season, section	# of sampling events included in study	T (°C)	S (ppt)	pH	DO (mg L−1)	Chla (µg L−1)
2009–2012
Season  a
Spring	12	16.3 ± 5.0	11.0 ± 4.4	7.3 ± 0.8	10.0 ± 4.1	14.0 ± 7.8
Summer	45	27.2 ± 1.6	11.0 ± 3.3	7.5 ± 0.6	6.6 ± 1.2	14.1 ± 5.4
Fall	15	17.6 ± 5.4	11.3 ± 4.4	7.7 ± 0.5	8.4 ± 1.6	12.0 ± 6.2
Winter	10	4.3 ± 3.2	11.0 ± 3.6	7.4 ± 0.2	11.8 ± 1.3	12.8 ± 6.7
Section  b
Upper	40	21.7 ± 7.7	8.2 ± 2.4	7.4 ± 0.6	7.6 ± 2.2	15.5 ± 6.9
Mid	—	—	—	—	—	—
Lower	42	20.4 ± 9.2	13.6 ± 2.6	7.6 ± 0.6	8.6 ± 3.0	11.8 ± 4.6
Overall (2009–2012)	82	21.0 ± 8.5	11.1 ± 3.7	7.5 ± 0.6	8.1 ± 2.7	13.6 ± 6.1
2019–2022
Season  a
Spring	11	13.3 ± 4.7	8.7 ± 2.6	8.3 ± 0.4	10.5 ± 1.6	39.6 ± 35.3
Summer	34	27.2 ± 1.8	10.3 ± 3.0	7.9 ± 0.4	6.4 ± 1.0	21.1 ± 12.2
Fall	22	20.1 ± 5.5	12.6 ± 2.1	7.8 ± 0.4	7.8 ± 1.4	15.6 ± 7.6
Winter	13	5.0 ± 2.1	12.4 ± 2.9	8.0 ± 0.6	11.9 ± 1.5	14.1 ± 9.4
Section  b
Upper	8	21.6 ± 8.8	9.4 ± 3.8	8.0 ± 0.5	8.4 ± 2.1	30.0 ± 25.0
Mid	56	19.1 ± 8.8	10.4 ± 2.2	7.9 ± 0.4	8.2 ± 2.7	21.8 ± 17.9
Lower	16	20.4 ± 9.6	13.7 ± 3.1	8.1 ± 0.4	8.4 ± 1.9	13.6 ± 8.2
Overall (2019–2022)	80	19.6 ± 8.9	11.0 ± 3	7.9 ± 0.4	8.3 ± 2.5	21.0 ± 17.7

^aSeasons defined as: spring (March, April, May); summer (June, July, August); fall (September, October, November); winter (December, January, February).

^bStations included in each section: upper (Chester River, Sandy Point, South River, Cox Creek), mid (Miles River, Broad Creek, Choptank River, Upper Patuxent River, Lower Patuxent River, Wicomico River), and lower (St. Mary’s River, Tangier Sound).

The overall average water temperature, salinity, and dissolved oxygen concentrations were not significantly different (P > 0.05) between the two sampling periods (Table 2). In contrast, pH concentrations were significantly different between sampling periods (P < 0.001), with higher average pH levels observed during the 2019–2022 time period.

Chlorophyll-a concentrations were also noticeable higher between 2019 and 2022, albeit with a higher degree of variability, but differences could not be directly compared due to the use of different analytical methods between sampling periods (Table 2) and are addressed elsewhere (Morgado et al. 2024). Average water temperature and dissolved oxygen concentrations varied significantly by season (P < 0.0001) for both sampling periods and followed a similar pattern (Table 2). As expected, water temperatures were highest during the summer and lowest during the winter, ranging between 4.3 and 27.2°C, while dissolved oxygen concentrations were lowest during the summer and highest during the winter, ranging from 6.4 to 11.9 mg L⁻¹. Average salinity, pH, and chlorophyll-a concentrations did not vary significantly between seasons for the 2009–2012 time period but were significantly different for 2019–2022 (P < 0.05). For the later study period, average salinity varied between 8.7 ppt during the spring and 12.6 ppt during the fall, while pH and chlorophyll-a concentrations were higher during the spring, and lower during the fall and winter, respectively (Table 2).

Average water temperature and pH did not vary significantly between the upper and lower sampling stations for the 2009–2012 sampling period nor between upper, mid, and lower sampling stations for the 2019–2022 time period (P > 0.05) (Table 2). On the other hand, average salinity and chlorophyll-a concentrations were significantly different between upper, mid, and lower sampling stations for both sampling periods (P < 0.01) and followed a similar trend. As expected, higher salinities close to 14 ppt were observed in the lower sampling stations, closer to the mouth of the Bay, and in the upper stations (near fresh tidal waters) the lowest salinity was observed, an average of 8.5 ppt. In contrast, average chlorophyll-a concentration declined moving from the upper stations to the lower stations, although it should be noted that there was a large degree of variability in chlorophyll-a measurements per sampling event, especially during the 2019–2022 sampling period. Lastly, average dissolved oxygen concentrations did not vary significantly by section during the 2019–2022 sampling period (P > 0.05) but were significantly different between sections for the 2009–2012 time period (P = 0.04). During this time, dissolved oxygen concentrations were on average 1 mg L⁻¹ higher in the lower section closest to the mouth of the Bay, compared to the upper section (Table 2).

Vibrio spp. abundance

There were a total of 1476 V. parahaemolyticus isolates (715 from 2009 to 2012 and 761 from 2019 to 2022) and 806 V. vulnificus isolates (565 from 2009 to 2012 and 241 from 2019 to 2022) recovered from water and oyster samples across both sampling periods that were included in the study, with most isolates collected during the summer and fall seasons. Most isolates (>90%) were recovered from water samples, and there were too few isolates recovered from oyster samples to observe any significant trends. The average relative frequency of species-specific markers per sampling event for V. parahaemolyticus (tlh) and V. vulnificus (vvha) was not significantly different between sampling periods or across seasons (Table 3).

Table 3.

Number of V. parahaemolyticus and V. vulnificus isolates and relative frequency (%) of species-specific and virulence factors (V. parahaemolyticus: tlh⁺, tdh⁺, trh⁺; V. vulnificus: vvha⁺, vcgE⁺, vcgC⁺, pilA⁺, rtxA⁺) by sampling period (2009–2012 and 2019–2022), season, and sample type (water or oyster).

Study/season	V. parahaemolyticus				V. vulnificus
	N	tlh + (%)	tdh + (%)	trh + (%)	N	vvha + (%)	vcgE+ (%)	vcgC+ (%)	pilA+ (%)	rtxA+ (%)
2009–2012
Water
Spring	61	100	11	11	71	85	*	*	39	45
Summer	329	100	2	2	309	97	*	*	17	40
Fall	126	100	0	2	182	99	*	*	17	48
Winter	84	100	7	7	1	100	*	*	0	0
Oyster
Spring	18	100	0	0	0	–	*	*	–	–
Summer	62	100	0	0	2	50	*	*	0	50
Fall	27	100	0	0	0	–	*	*	–	–
Winter	8	100	0	0	0	–	*	*	–	–
Overall	715	100	3	3	565	96	*	*	20	41
2019–2022 Water
Spring	37	88	0	9	9	87	87	0	11	0
Summer	291	97	0.3	0.7	110	96	92	12	33	0.9
Fall	182	100	0.6	0.6	84	96	81	24	21	6
Winter	88	100	0	0	4	67	67	17	33	0
Oyster
Spring	0	–	–	–	0	–	–	–	–	–
Summer	71	94	0	0	23	96	96	4	44	0
Fall	92	100	0	0	11	100	91	27	54	9
Winter	17	100	0	0	0	–	–	–	–	–
Overall	761	97	0.1	2	241	94	88	17	36	2

^*Data not available.

Spatial-temporal trends in isolate virulence

The relative frequency of V. parahaemolyticus virulence factors tdh and trh was not found to be significantly different (P > 0.05) between sampling periods for isolates collected from water samples (Table 3). On the other hand, the relative frequency of V. vulnificus virulence factors pilA and rtxA was significantly different between sampling periods, P = 0.02 and P < 0.0001, respectively, with higher overall percentage of pilA observed in isolates recovered between 2019 and 2022, and rtxA during the 2009–2012 time period (Table 3).

The relative frequency of V. parahaemolyticus virulence factors, tdh and trh, recovered from water samples varied by season, but only significantly for the 2009–2012 sampling period (P = 0.02) (Table 3). During the 2009–2012 sampling period, the relative frequency of tdh- and trh-positive isolates was highest in the spring (n = 11%) and lowest in the fall. For the 2019–2022 time period, tdh was infrequently observed in isolates across each season, and trh frequency was highest in the spring (n = 9%) and lowest in the winter (Table 3).

The relative frequency of V. vulnificus virulence factors, including vcgC, pilA, and rtxA, recovered from water samples varied by season but not significantly for either sampling period (P > 0.05) (Table 3). VcgC (recorded only in isolates collected during 2019–2022) was highest in the fall (n = 24%) and not detected during the spring. The frequency of pilA ranged between 39% in the spring to non-detectable in the winter for the 2009–2012 sampling period, while it was highest during the summer (n = 33%) and lowest in the spring (n = 11%) for isolates collected during 2019–2022. Lastly, rtxA frequency for both sampling periods was lowest in isolates collected during the winter (not detected) and highest during the fall (n = 48% in 2009–2012 and n = 6% in 2019–2022) (Table 3).

For combined sampling periods (2009–2012 and 2019–2022), the relative frequency of Vibrio spp. virulence factors in the isolates varied by section in the Chesapeake Bay. However, the differences were statistically significant only for tdh (P = 0.04), pilA (P = 0.0002), and rtxA (P < 0.0001) (Fig. 2). Overall, the relative frequency of tdh (n = 5%), trh (n = 5%), and rtxA (n = 33%) was highest in isolates collected from the upper section, which included Chester River, Sandy Point, South River, and Cox Creek stations. On the other hand, vcgC and pilA were more frequently detected in isolates, n = 18% and n = 48%, respectively, collected from the mid-section, which included Miles River, Broad Creek, Choptank River, Upper Patuxent River, Lower Patuxent River, and Wicomico River stations (Fig 2).

Figure 2.

Average relative frequency of Vibrio spp. virulence factors (V. parahaemolyticus: tdh, trh; V. vulnificus: vcgC, pilA, rtxA) by Chesapeake Bay section from both sampling periods combined. Sampling stations included in each section: upper (Chester River, Sandy Point, South River, Cox Creek), mid (Miles River, Broad Creek, Choptank River, Upper Patuxent River, Lower Patuxent River, Wicomico River), and lower (St. Mary’s River, Tangier Sound).

Environmental associations with virulence

Between both sampling periods, only 1% of V. parahaemolyticus isolates (n = 15), all cultured from water samples, tested positive for tdh (n = 9) and/or trh (n = 15) virulence factors. Detection of tdh and trh (absence/presence) was not significantly associated (P > 0.05) with average water temperature, dissolved oxygen, pH, and chlorophyll-a. Additionally, relative frequency of both of the virulence factors did not show a significant linear relationship or a strong correlation with average water temperature, pH, dissolved oxygen, and chlorophyll-a concentration. However, there was a significant, but weak, negative correlation between detection of trh and average salinity (ρ = −0.21, P = 0.01), but no such correlation observed for tdh.

Across both sampling periods, 45% of V. vulnificus isolates (n = 363) tested positive for vcgC (n = 37), pilA (n = 155), and/or rtxA (n = 217) virulence factors, and of these, 95% were cultured from water samples. Interestingly, 16 isolates also tested positive for both vcgC and vcgE virulence factors. Detection of vcgC, pilA, and rtxA was not significantly associated (P > 0.05) with changes in environmental water parameters. Moreover, the relative frequency of vcgC and pilA did not show a significant linear relationship or a strong correlation with any of the measured environmental water parameters. Conversely, relative frequency of rtxA showed a negative correlation with both average chlorophyll-a concentration (ρ = −0.23, P = 0.02) and pH levels (ρ = −0.28, P = 0.007). Notwithstanding, there was no observable significant linear relationship between rtxA and chlorophyll-a and pH levels, nor was there significant association with average water temperature, salinity, or dissolved oxygen concentration.

Although there was no significant or strong linear relationship between average water temperature, salinity, and relative frequency of Vibrio spp. virulence factors, an interesting pattern, denoting potential optimum environmental conditions per species, was observed, as shown in the bubble plot (Fig. 3). For V. parahaemolyticus isolates with virulence factors, tdh and/or trh (Fig. 3a), the highest observed average relative frequency occurred within water temperatures between 17 and 25°C and salinities of 1–7 ppt. In contrast, V. vulnificus isolates with virulence factors, vcgC (Fig. 3b), pilA (Fig. 3c), and rtxA (Fig. 3d), occurred more frequently in water temperatures ranging between 23 and 30°C and salinities of 7–14 ppt.

Figure 3.

Association between average temperature/salinity and the occurrence of V. parahaemolyticus and V. vulnificus virulence factors for both sampling periods combined. VcgC data only from 2019 to 2022. Disks in each bubble plot represent a single sampling event; disk size and color are indicative of detection and average relative frequency in percentage of tdh and/or trh (a), vcgC (b), pilA (c), and rtxA (d). The smallest white disks represent sampling events that did not detect the virulence factors but reported isolates positive for either tlh (V. parahaemolyticus) or vvha (V. vulnificus). Ellipses (dotted black circles) denote an area of greater virulence factor aggregation using an approximate least-squares fitting.

Discussion

Similarly to previous reports by Parveen et al. (2008, 2020), Banakar et al. (2011), Johnson et al. (2012), Jacobs et al. (2014), Chen et al. (2017), Davis et al. (2017), and Brumfield et al. (2023a), this study evaluated reported associations among pathogenic V. parahaemolyticus and V. vulnificus and environmental determinants in the Chesapeake Bay. The findings presented here confirm known associations but given the long-term nature of the sampling efforts reported here, also expand on changes and possible trends in the distribution and seasonality of potentially pathogenic Vibrio spp. in this important estuarine system.

Understanding how physico-chemical parameters change over time across coastal ecosystems is a critical aspect in the prevention of severe Vibrio spp. infections worldwide, as their occurrence and distribution in the environment have been shown to be strongly associated with certain environmental conditions (Davis et al. 2017, Williams et al. 2017, Tran et al. 2020, Brumfield et al. 2021). These conditions differ between Vibrio spp. and strain and may also change over time as the bacteria adapt to environmental stressors and increasingly adverse environments, including their ability to enter a viable but non-culturable state (VBNC) (Nowakowska and Oliver 2013, Davis et al. 2017, Colwell 2000). Not only have previous studies noted some variability in optimal growing conditions between Vibrio spp. but they also suggest that their abundance could shift depending on the environmental host (e.g. water, sediment, or oyster populations) (DePaola et al. 2003, Johnson et al. 2010, 2012, Chen et al. 2017, Ndraha and Hsiao 2021, Brumfield et al. 2023a). In the present study, environmental and potentially pathogenic V. parahaemolyticus and V. vulnificus were mostly found in water samples, compared to oysters, especially in the later time period (2019–2022). However, it is worth noting that this finding may be biased given the limitations of using culture-based methods to successfully recover pure isolates from complex sample types such as oysters.

According to a study by Najjar et al. (2010) that evaluated the potential impact of climate change in the Chesapeake Bay, the greatest changes by the end of the 21st century were predicted to be increased sea-level variability and water temperature, between 2 and 6°C, increased precipitation, especially in the winter and spring, and greater winter and spring streamflow. In the current study, overall average water temperatures were not significantly higher in the later sampling period of 2019–2022. However, it should be noted that the average water temperature was greater in the fall and winter compared to the earlier sampling period of 2009–2012. It is also important to recognize that there can be inter-decadal and intra-annual variability in temperature and precipitation patterns as well as the potential impact of extreme weather events, such as heat waves and hurricanes, that can lead to deviations from established trends (Banakar et al. 2011, Jacobs et al. 2014, Shaw et al. 2014, Brumfield et al. 2023b).

An interesting change between sampling periods was the observed decrease in both average water temperature and salinity during the spring, which might correlate to increased precipitation and streamflow. Other potential environmental changes forecasted by Najjar et al. (2010) included increased hypoxia, which was not observed between sampling periods, and altered phytoplankton dynamics, which would in turn impact the timing of zooplankton blooms, a known reservoir of Vibrio spp. (Kaneko and Colwell 1973). To this point, a noticeable increase and a greater variability in average chlorophyll-a levels were observed in the later sampling period across all seasons compared to the earlier time, and especially during the spring and summer. An additional concern includes the increase in acidification (lower pH) across the Chesapeake Bay (Najjar et al. 2010). A study by Waldbusser et al. (2011) found that pH has remained mostly unchanged in mesohaline waters across the Chesapeake Bay compared to more polyhaline areas further south, which may become increasingly corrosive to oyster shells. Moreover, the pH in tributaries that once supported large oyster populations has shown an increase over time, likely due to increased primary production from eutrophication, which may help explain the increased average pH levels observed during the later sampling period (2019–2022) compared to the earlier one (2009–2012).

Importantly, recent studies have found that V. vulnificus thrives in salinities between 5 and 25 ppt and at temperatures at or above 25°C (Williams et al. 2017, Baker-Austin and Oliver 2018). Our results appear to support these observations in terms of temperature preferences for V. vulnificus isolates that harbor virulence factors vcgC, pilA, and/or rtxA. Here, the salinity range for incidence of these biomarkers was between 7 and 14 ppt, which may suggest potential pathogen preferred conditions within the Maryland waters of the Chesapeake Bay and optimal seasonal conditions, in agreement with our previous work (Morgado et al. 2024). Notwithstanding, sampling efforts were not performed in more polyhaline areas of the Bay further south, which limited our ability to fully evaluate V. vulnificus virulence factor trends across the entire estuary. Interestingly, the significant, albeit weak, negative association observed between isolates rtxA positive and average pH and chlorophyll-a levels, may be indicative of optimal growing conditions in areas of lower pH and chlorophyll-a concentrations, i.e. more polyhaline areas closer to the mouth of the Bay, as well as the fall season when these conditions may be more likely. In fact, stations located in the mid and lower sections of the Maryland waters of the Bay appeared to harbor a greater percentage of V. vulnificus strains with virulence factors, compared to the upper section. These findings differ somewhat from a hindcast prediction for a 15-year period, 1991–2005, by Banakar et al. (2011), which found that the greatest hotspots of V. vulnificus were mostly located in the upper Bay, parts of western estuaries, and the mid Bay, with peaks in July and a likely decrease in presence until November. This particular study and findings by Jacobs et al. (2014), also noted a significant trend in which the mean probability of V. vulnificus occurrence in years with excessive precipitation (wet years) shifted southwards in the Bay towards more polyhaline areas, and the reverse happened during drier years, where the shift occurred northward. However, it should be noted that these previous studies analyzed trends for total V. vulnificus and not isolates carrying specific virulence factors. To this effect, it is important to note that while the chosen markers of vcgC, rtxA, and pilA may be suggestive of V. vulnificus pathogenicity, the multifactorial and complex nature of severe infection by this particular vibrio limits our ability to fully recognize virulent strains.

Earlier reports have noted that optimal growing conditions for V. parahaemolyticus can vary depending on the environmental host (water, oyster, sediment) as well as geographic location (Parveen et al. 2008, Johnson et al. 2012). For instance, Johnson et al. (2012) found that, with respect to tlh concentration, Gulf of Mexico strains had a greater affinity toward higher salinities and water temperatures. However, similarly to our study, they did not find that sea surface temperature was a strong predictor of the density for potentially pathogenic strains carrying tdh and/or trh. In fact, as suggested by others, it may be that total V. parahaemolyticus populations are not a suitable indicator of the presence of isolates containing these virulence genes, as they have shown to be variable and maybe even inversely related to temperature (DePaola et al. 2003, Williams et al. 2017, Davis et al. 2021). Although limited by the low percentage of tdh and/or trh-positive isolates cultured between sampling periods, our study appeared to indicate greater prevalence of potentially pathogenic subpopulations when temperatures were at or around 20°C and at lower salinities, between 1 and 7 ppt, and an inverse association was observed with average salinity for trh-positive isolates. The reduced isolation of strains containing virulence genes is in line with multiple studies that have found between 1% and 1.9% of pathogenic V. parahaemolyticus concentrations in environmental samples (Davis et al. 2017, Williams et al. 2017, Parveen et al. 2020).

In terms of location and seasonality, potentially pathogenic V. parahaemolyticus were more frequently observed in the upper Bay stations of Maryland and appeared in two peaks, one in the spring and another in the fall, observed in the later sampling period (2019–2022). It is important to note that our study relied on the ability to recover pure culture Vibrio spp. isolates from complex samples, and as reported previously, Vibrio can exist in the VBNC state, which can prevent detection using only culture-based methods (Parveen et al. 2008, Chen et al. 2017). Thus, our culture-based approach likely resulted in an underrepresentation of the true population of Vibrio spp., and subpopulations of pathogenic V. parahaemolyticus with tdh and trh genes may be too sparse to be properly enumerated with this technique (Johnson et al. 2012). On the other hand, solely relying on molecular methods, namely real-time PCR, may overestimate the abundance of potentially pathogenic strains in the environment, as they will not be able to differentiate between living and non-living populations. As such, a combination of culture-dependent and culture-independent methods may yield the best results, allowing a more complete description of pathogenic Vibrio spp. populations, including both culturable and VBNC, the latter dominant during colder months and possibly in other environmental hosts (e.g. sediment and oysters). Other limitations included a 10-year gap between the two sampling periods, as well as differences in study sites between time periods. Also, chlorophyll-a levels in each sampling period (2009–2012 and 2019–2022) were measured using different instruments and methodology, potentially influencing the results. Clearly, long-term continued monitoring is important, especially given intra- and inter-annual climate variability and rapidly changing physico-chemical conditions in the Chesapeake Bay. In addition, due to funding constraints, vcgC and vcgE data could not be obtained during the 2009–2012 sampling period.

In summary, potentially pathogenic V. parahaemolyticus and V. vulnificus can be detected in the Maryland waters of the Chesapeake Bay throughout the year, at frequencies posing greater risk of infection in the fall and early winter. Moreover, mid and lower sections of the Bay waters (within Maryland), which include commercial and recreationally important areas, may harbor greater risk of severe vibriosis from wound-associated water exposure, given the prevalence of potentially pathogenic V. vulnificus.

Author contributions

Michele E. Morgado (Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing), Kyle D. Brumfield (Formal analysis, Investigation, Methodology, Software, Writing – review & editing), Suhana Chattopadhyay (Methodology, Writing – review & editing), Leena Malayil (Methodology, Writing – review & editing), Anwar Huq (Resources, Resources, Writing – review & editing, Writing – review & editing), Rita R. Colwell (Resources, Writing – review & editing), and Amy R. Sapkota (Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing)

Conflict of interest

None declared.

Funding

MEM and ARS were supported by NRT-INFEWS: UMD Global STEWARDS (STEM Training at the Nexus of Energy, Water Reuse, and Food Systems) that was awarded to the University of Maryland School of Public Health by the National Science Foundation National Research Traineeship Program, Grant number 1828910. The 2009-2012 data collection and research were supported by NSF grant EF-0813285/EF-0813066 as part of the joint NSF-NIH Ecology of Infectious Diseases program. Further support was supplied by the National Institute of Environmental Health Sciences, National Institutes of Health (R01ES030317A), and the National Science Foundation (OCE1839171 and CCF1918749) awarded to AH and RRC, as well as the National Aeronautics and Space Administration (80NSSC20K0814 and 80NSSC22K1044) awarded to RRC.

Data availability

Available upon request.