Distribution and diversity of the enterococcal surface protein (esp) gene in animal hosts and the Pacific coast environment


Alexandria Boehm, Environmental & Water Studies, Stanford University, The Jerry Yang and Akiko Yamazaki Environment and Energy Building, Mail code 4020, Stanford, CA 94305, USA. E-mail: aboehm@stanford.edu


Aims:  This study sought to evaluate the distribution of the enterococcal surface protein (esp) gene in Enterococcus faecium in the Pacific coast environment as well as the distribution and diversity of the gene in Northern California animal hosts.

Methods and Results:  Over 150 environmental samples from the Pacific coast environment (sand, surf zone, fresh/estuarine, groundwater, and storm drain) were screened for the esp gene marker in E. faecium, and the marker was found in 37% of the environmental samples. We examined the host specificity of the gene by screening various avian and mammalian faecal samples, and found the esp gene to be widespread in nonhuman animal faeces. DNA sequence analysis performed on esp polymerase chain reaction amplicons revealed that esp gene sequences were not divergent between hosts.

Conclusions:  Our data confirm recent findings that the E. faecium variant of the esp gene is not human-specific.

Significance and Impact of the Study:  Our results suggest that the use of the esp gene for microbial source tracking applications may not be appropriate at all recreational beaches.


Bacterial contamination at beaches is a problem of growing concern in the United States and much of the developed world (Bartram and Rees 2000). In the United States, there were over 25 000 days of closures and advisories at recreational beaches in 2006, compared with 6200 in 1999, which were instigated by violations of faecal indicator bacteria (FIB) standards (Dorfman and Stoner 2007). Of the nearly 40 000 waters listed on the US States’ Clean Water Act 303(d) lists, almost a quarter (8964) are listed as a result of frequent FIB standard violations (http://iaspub.epa.gov/waters/national_rept.control, accessed 2 January 2008). These figures suggest that microbial pollution of US waters is widespread, threatening the health of visitors who collectively make 927·7 million trips to the beach each year (Weiher and Sen 2005), as well as jeopardizing the most rapidly growing sector of the ocean economy, tourism and recreation (Colgan 2003).

FIB standards for recreational waters, which include regulations for total coliform, Escherichia coli and enterococci levels, were promulgated on the basis of epidemiological studies (Boehm et al. 2009). These studies were conducted at beaches where water quality was known to be impacted by point sources of municipal wastewater, and showed a correlation between adverse health outcomes in swimmers and levels of E. coli and enterococci in fresh and marine waters, respectively (Cabelli et al. 1982; Dufour 1984; Freeman 1995). Yet it is understood that FIB may also originate from nonhuman animal faeces (Ashbolt et al. 2001), or persist as indigenous members of the coastal microbial community in some temperate and tropical areas (Byappanahalli and Fujioka 2004; Byappanahalli et al. 2006). A more recent epidemiological study conducted at a beach in southern California where poor water quality was attributed to wildlife and environmental sources indicated no correlative relationship between FIB and health risks (Colford et al. 2007). Thus, there is growing concern that the many beach closures and advisories caused by FIB from nonpoint sources may not be truly indicative of a health risk.

Microbial source tracking (MST) methods have emerged as a means to determine the source of FIB in environmental waters. With knowledge of FIB source in hand, remediative actions can be taken, and the relative risk of exposure might be determined (Field and Samadpour 2007; Boehm et al. 2009). There are numerous techniques in the MST ‘toolbox’, each with its own set of pitfalls and advantages (Simpson et al. 2002; Field and Samadpour 2007; Stoeckel and Harwood 2007). An international panel of experts convened by the US Environmental Protection Agency (USEPA 2007) identified the need for more research evaluating the efficacy of source tracking methods and their applicability to different geographical regions. The present study addresses this need by evaluating the potential use of a recently proposed MST target, the enterococcal surface protein-encoding gene (esp) in Enterococcus faecium (Scott et al. 2005), hereafter referred to as the esp HF marker, where HF stands for ‘human faecal’.

The esp gene is thought to be located on a mobile pathogenicity island (PAI; Leavis et al. 2004); its expression is associated with biofilm formation (Toledo-Arana et al. 2001; Heikens et al. 2007; Van Wamel et al. 2007) and increased virulence (Vergis et al. 2002; Lund and Edlund 2003). The esp gene has been described in both E. faecalis (Shankar et al. 1999) and E. faecium (Baldassarri et al. 2001; Willems et al. 2001). Various researchers have applied the esp HF marker as a source tracking tool in their studies (McDonald et al. 2006; McQuaig et al. 2006; Yamahara et al. 2007). However, little is known about the variability and diversity of the gene in animal hosts and in the environment. Whitman et al. (2007) investigated the presence of esp HF marker in faeces from the mid-western United States, adjacent to the Great Lakes, and found it was present in nonhuman faeces. Additional information about the prevalence of the esp HF marker in human and non-human enterococci, particularly from different geographical areas, would provide important evidence to support the use of this MST target. Further, no E. faecium esp DNA sequence data from nonhuman hosts has yet been presented; sequence data could reveal host-specific sequence variations that could be targeted for MST.

The objectives of the present study are: (i) to determine the distribution of the esp HF marker in enterococci from a variety of environments along the California coast and the north shore of Kaua’i, Hawai’i; (ii) to further evaluate the host specificity of the esp HF marker; and (iii) to examine the DNA sequence diversity of the esp gene among various animal hosts. The work presented here is the first to test the human specificity of the esp HF marker in the Pacific coast environment and to provide sequence data for the esp gene in animal hosts and the environment.

Materials and methods

Sample collection and processing

Environmental samples (water and sand) were obtained from the coasts of California and Kaua’i, Hawai’i during multiple field excursions spanning from July 2006 to March 2007 (Keymer et al. 2007; Santoro and Boehm 2007; Yamahara et al. 2007; De Sieyes et al. 2008; Knee et al. 2008). These samples (beach sands, storm drains, groundwater, estuaries, and surf zones) were chosen to assess the prevalence of the esp HF marker in a wide range of coastal environments. Mammalian and avian faecal samples from humans, dogs, horses, seagulls, seals, and sea lions were obtained from San Francisco Bay area farms, beaches, dog parks, and the Marine Mammal Center in Sausalito, California. These hosts were chosen for the study because they are common to beaches of the Pacific coast and could contribute enterococci to the coastal environment. Primary-treated sewage effluent was collected from the Palo Alto Water Quality Control Plant in Palo Alto, California.

Environmental water samples (groundwater, storm drain, estuarine, and surf zone) were screened for enterococci using either Enterolert implemented in a 97-well Quanti-tray (IDEXX, Westbrook, ME, USA) or membrane filtration onto mEI agar (Difco/BD, Sparks, MD, USA). Beach sand samples were eluted and the eluate was membrane filtered as previously described (Yamahara et al. 2007). Sewage samples were membrane-filtered onto mEnterococcus (Difco/BD) or mEI agar (EPA Method 1600). Animal faecal samples were combined with Dulbecco’s phosphate buffer saline (PBS; GIBCO/Invitrogen, Grand Island, NY, USA) to form slurries, which were then membrane-filtered in duplicate or triplicate onto mEnterococcus agar (Difco/BD) at dilution volumes producing 20–60 colonies per filter. Membrane filtration apparatuses were ultraviolet (UV)-disinfected for 5 min in a UV cross-linker (UVP, model CL1000, delivers c. 3000 μW cm−2) between samples, and filtration blanks were run.

Presumptive Enterococcus colonies from membrane filtration were enriched in trypic soy broth (TSB) as described by Scott et al. (2005). Enterococci from Enterolert Quanti-tray wells were enriched as follows: the back of the tray was sterilized with 100% ethanol, and enterococci-positive wells were aspirated using a 21½ gauge needle and syringe. Media from a single tray were combined and mixed, and then 1 ml was added to 15 ml of TSB. All TSB enrichments, regardless of the original culture media, were incubated at 41°C for 4–6 h.

Cultures of E. faecium strains E300, E470, and E734, carrying the esp gene, were graciously provided by Dr Rob Willems of University Medical Center Utrecht, The Netherlands. DNA obtained from these strains was used as template for positive controls for all esp polymerase chain reaction (PCR) assays. Staphylococcus aureus (ATCC #25923) and Streptococcus bovis (ATCC #33317) were used as negative controls from DNA extraction through PCR to ensure no cross-contamination occurred.

DNA was extracted from one millilitre of enrichment and control cultures with a QIAamp Min-Elute DNA Spin Kit (Qiagen Inc., Valencia, CA, USA), using the protocol for gram-positive bacteria. A blank experiment was performed to ensure there was no laboratory contamination at any stage of the work. Triplicate 100-μl aliquots of stationary-phase E. faecium E734 were membrane filtered, and after UV disinfection, the same filtration apparatuses were used to filter 10 ml of PBS following each aliquot of the culture. The same procedure was repeated using 1-ml aliquots of primary-treated sewage in place of the E. faecium E734. All 12 filters were placed on mENT and incubated for 48 h at 35°C. The samples were then subjected to the entire esp gene assay, from enrichment to PCR, regardless of whether there was growth on the culture media.

PCR amplification of the esp HF marker

We evaluated DNA extracts from samples described before for the presence of the esp HF marker using PCR primers described by Scott et al. (2005) (hereafter referred to as the Scott et al. primers). Each 25-μl reaction consisted of 1 ×  PCR buffer containing 1·5 mmol l−1 of MgCl2 (Qiagen), 0·4 μmol l−1 of each primer, 1 U of HotStar Taq (Qiagen), and 1 μl of template DNA (200–300 ng μl−1). DNase-/RNase-free water (GIBCO/Invitrogen) was used as a no-template control for each PCR run. Following an initial polymerase-activation step of 15 min at 95°C, PCR was carried out by 35 cycles of 1 min at 94°C, 1 min at 58°C, and 1 min at 72°C, with a final extension step of 5 min at 72°C. PCR products were analysed on a 1·5% agarose gel stained with ethidium bromide, and visualized and photographed using a GelDoc imaging system (Bio-Rad Laboratories, Hercules, CA, USA) and QuantityOne software (ver. 4.6.3; Bio-Rad Laboratories).

PCR amplification of 1-kb segment of esp gene

We used a separate primer pair (forward: BL9 5′-ATTTTGCTAATGCAAGTCCA-3′; reverse: esp5R 5′-TACTGCTAAATCGGTCGTG-3′) (hereafter referred to as the 1-kb primers) to explore sequence homology in esp from various hosts and environments by sequencing a 1-kb section of the gene. The reverse primer has been used previously to sequence the esp gene (Leavis et al. 2004); we designed the forward primer such that the pair would amplify all esp genes in GenBank, from both E. faecium and E. faecalis. PCR were carried out in 25 μl reactions containing 1 × PCR buffer (Invitrogen), 2·0 mmol l−1 of MgCl2, 0·4 μmol l−1 of each primer, 0·2 mmol l−1 of dNTP, 1 U of Platinum Taq (Invitrogen), and 1·0 μl of template DNA. All control strains, as well as a no-template control were included in each PCR run. Following an initial polymerase-activation step of 4 min at 95°C, PCR was carried out by 35 cycles of 1 min at 94°C, 1 min at 54°C, and 1 min at 72°C, with a final extension step of 5 min at 72°C. PCR products were visualized on a 1·5% agarose gel stained with ethidium bromide.

esp gene clone library construction

Enterococcus faecium esp PCR products were amplified from six DNA extracts (sewage, estuarine water from Elkhorn Slough, California, and dog, horse, seal, and sea lion faeces) using the Scott et al. assay. Amplicons were gel purified using the MinElute Gel Extraction kit (Qiagen) and cloned with a TOPO-TA cloning kit, using the pCR2·1-TOPO vector and TOP10 competent cells (Invitrogen).

Eight clones from each DNA sample were randomly chosen for sequencing to confirm amplification of the desired target, and to explore the diversity of the esp gene fragment among various animal hosts and Pacific coast environments. Plasmid preps and DNA sequencing were carried out by MCLab (South San Francicso, CA, USA), using an ABI 3730XL capillary sequencer (PE Applied Biosystems, Foster City, CA).

PCR was performed using the same six DNA samples as template with the 1-kb primers. In addition, 1-kb esp PCR products were also amplified from gull faeces and surf zone seawater from Lover’s Point, California. These amplicons were cloned as described before, and 12 clones from each DNA sample were randomly chosen for sequencing using the same procedure as before.

In summary, we obtained 48 esp sequences using the Scott et al. primers, and 96 sequences using the 1-kb primers. The collection included clones from sewage, seawater, estuarine water, and dog, horse, seal, sea lion, and gull faeces (see Table 1).

Table 1.   Number of esp sequences obtained from clones using two primer sets; DNA were extracted from enterococci enrichment cultures
DNA extractNo. of clones sequenced
Scott et al. primers1-kb primersTotal
Dog faeces81220
Horse manure81220
Seal faeces81220
Sea lion faeces81220
Gull faeces01212
Estuarine water81220

Phylogenetic analyses

Sequence data were processed (vector fragments and primers removed) and aligned into contigs using Sequencher (ver. 4·7; Gene Codes Corporation, Ann Arbor, MI, USA). Contigs were assembled into phylogenetic trees using Paup* (Swofford 2002). GenBank esp gene sequences that had complete data in the amplicon region were also included in the alignments. Enterococcus faecium E470 was chosen as an outgroup because it has <90% homology to the other reference sequences in this region of the gene. Clones from the same DNA sample whose sequences were 100% homologous to each other were condensed into a single position on the tree. Sequences obtained in this study have been deposited in GenBank under the following accession numbers: EU394205EU394213, EU621692, and EU815329EU815406.

Statistical analyses

Concentrations of enterococci in individual faecal samples were assigned the lower detection limit value (varied; on the order of 1–10 CFU g−1) if no colonies grew on the membrane filter, and the upper limit of detection (varied; range of 105–107 CFU g−1) if there were too many colonies to count. For those environmental samples that were processed using IDEXX Enterolert, an upper detection limit of 24 190 MPN per 100 ml was assigned if all wells were positive. When multiple DNA extracts from the same sample (faecal or environmental) were screened with PCR, the sample was assigned a positive result for the assay if at least one replicate produced an amplicon. Both parametric (n-way anova) and nonparametric methods (Kruskal–Wallis analysis of variance and chi-squared tests) were used to examine the variability of measured parameters between sample types. Rejection of the null hypothesis was deemed statistically significant if < 0·05. All analyses were completed using Matlab (ver. 7.3.0 R2006b; The Mathworks, Inc., Natick, MA, USA).


Quality control

All of our controls for the esp gene assays performed as expected: no amplicons were produced from no-template or negative controls. To confirm that we had no cross-contamination at any stage, we performed a blank experiment in which blank samples were processed alongside highly concentrated positive control samples from filtration to PCR. No growth (colonies or turbidity) was observed in the blank samples during the culturing steps, and no esp amplicons were produced from those DNA extracts (Fig. 1).

Figure 1.

 Agarose gel (containing ethidium bromide) showing esp human faecal marker assay blank method results. Lanes 1, 3, 5: 1 ml primary-treated sewage; lanes 7, 9, 11: 100 μl stationary-phase Enterococcus faecium E734; even-numbered lanes: blanks. Lanes 1 and 12 are flanked by 100-bp ladders.

The esp HF marker in the environment

We investigated the presence of the esp HF marker in enterococci cultured from the California and northern Kaua’i, Hawai’i coastal environments (Table 2). Over numerous sampling trips in 2006 and 2007, we collected 52 surf zone samples from Huntington Beach and Stinson Beach, California and Kaua’i, Hawai’i; 37 groundwater samples from Stinson Beach, California and Hanalei, Hawai’i; 34 sand samples from 12 locations along the California coast; 21 fresh/estuarine samples from the San Francisco Bay area (California) and Kaua’i, Hawai’i; and 9 storm drain samples from 5 Northern California beaches (see Table 3 for details). These samples were selected to characterize the distribution of the esp HF marker in a broad range of Pacific coast environments. The esp HF marker was frequently detected in each sample type; overall, 37% of 153 environmental samples were positive for the esp HF marker.

Table 2.   Incidence of the esp human faecal marker in environmental samples, as detected by the Scott et al. (2005) polymerase chain reaction assay
 nNo. of esp positive
Total environmental samples15457
 Stinson Beach, CA4517
 Lover’s Point, CA2612
 Huntington Beach, CA3418
 California coast (other)286
 Hawai’i coast214
Sample type
 Surf zone5222
 Storm drain94
Table 3.   Sampling locations of environmental enterococci, listed in order of decreasing latitude
SiteLatitude (decimal ºN)Longitude (decimal ºW)No. of samples*Sampling date(s)Avg. [ENT]†No. of esp HF marker positive
  1. *SZ, surf zone; GW, groundwater; E, estuarine; SD, storm drain.

  2. †Mean most probable number or colony-forming units of enterococci per 100 ml of water or 100 g of sand.

Walk On Beach, CA38·72924722123·48548062July 2006901
Stump Beach (Salt Point State Park), CA38·58156667123·33563611July 200619491
Cambell Cove (Bodega Bay), CA38·30471389123·056997221July 200640381
Hearts Desire Beach (Tomales Bay), CA38·13245556122·89333July 200619271
Millerton Point (Tomales Bay), CA38·10775122·84545563July 200620250
Stinson Beach, CA37·89641944122·641272273241July 2006, Mar. 2007242717
China Beach, CA37·788325122·49106671July 20062820
Ocean Beach, CA37·76494167122·51261671July 200612110
San Pedro Creek, CA37·5955122·50566673Oct., Nov., Dec. 200672691
Pescadero Creek, CA37·26083333122·40366671Dec. 2006520
Waddell Creek, CA37·09616667122·27816671Dec. 20063930
Soquel Creek, CA36·9725121·95283331Dec. 200634880
San Lorenzo River, CA36·96416667122·011Dec. 200620140
Cowell’s Beach, CA36·96166389122·023719411July 20067300
Kirby Park, CA36·84121·74351Dec. 2006100
Elkhorn Slough, CA36·80533333121·78683331Dec. 20067061
Salinas River, CA36·79066667121·79016672Nov., Dec. 20069580
Lover’s Point, CA36·62496667121·91638897273July, Aug., Nov. 200680012
Monterey Beach, CA36·60126111121·88746391July 2006610
Huntington State Beach, CA33·62984167117·9606861313July 2006197618
Ha’ena State Park, HI22·22052159·583781Feb. 2007300
Hono Piki Beach, HI22·21875159·4982211Aug. 2006, Feb. 20071291
Hanalei River, HI22·2148159·4971121Feb. 20071990
Hanalei Pier, HI22·21273159·4973512-Aug. 2006, Feb. 20071720
Waikoko, Hanalei, HI22·2073159·516822Aug. 2006, Feb. 200721070
Hanalei Pavilion, HI22·20728159·4984311Feb. 2007642
Waipa, Hanalei, HI22·2052167159·514551Feb. 20071450
Waioli Beach Park, Hanalei, HI22·20303159·504551111Feb. 20072071

A high culturable enterococci count in the sample was neither necessary nor sufficient for the esp HF marker to be detected in the environment (= 0·24; n-way anova). The occurrence of the esp HF marker was not significantly different between Hawaiian and Californian samples (= 0·07, chi-squared test), and the presence of esp HF marker did not vary across sample type (surf zone, sand, groundwater, estuarine, storm drain; = 0·36, chi-squared test). The high occurrence of the esp HF marker in environmental samples, particularly in samples with low concentrations of enterococci, motivated additional work into the host specificity of the marker.

The esp HF marker in animal faeces

The distribution of the esp HF marker in faeces of animal hosts was determined using faecal samples from 12 humans, 16 dogs, 16 horses, 22 seagulls, 4 seals, and 14 sea lions. All faeces contained culturable enterococci except four sea lions and five seagulls (lower limits of detection varied for each individual but were on the order of 1–10 CFU g−1 wet weight). Concentrations of enterococci in individual animals ranged from below the detection limit to >2·9 × 107 CFU g−1 wet weight; the majority contained 103–106 CFU g−1 wet weight, consistent with previously published data (Chenoweth and Schaberg 1990; Murray 1990). Concentrations were significantly different between hosts (< 0·05, Kruskal–Wallis anova), with horses having the highest concentrations of enterococci in their faeces, and seagulls the lowest (Fig. 2).

Figure 2.

 Concentrations of enterococci in human and animal faeces, in CFU per gram wet weight. The middle line in the box represents the median; the upper and lower bounds of the box represent the 75th and 25th percentiles, respectively; the whiskers indicate 1·5 times of the interquartile range; and markers show outliers. The width of the notches is computed so that boxes whose notches do not overlap have different medians at the 5% significance level.

DNA from faecal enterococci enrichment cultures were screened for the esp HF marker (Table 4); 83% and 64% of human and nonhuman faecal samples carried the esp HF marker. The frequency of esp HF marker detection varied significantly across the host groups (< 0·05, chi-squared test), with the esp HF marker most commonly detected in dogs. Our rates of esp HF marker detection in nonhuman samples differ from those reported in Whitman et al. (2007), who found the esp HF marker in 9 of 43 dog samples, whereas 100% of our dog samples were positive for the esp HF marker. On the other hand, our sewage results agree with Whitman et al. (2007), who reported that 93% of the sewage samples were positive for esp HF marker; 92% of our sewage samples were positive.

Table 4.   Incidence of the esp human faecal marker in sewage and faecal samples as detected by polymerase chain reaction
 nNo. of esp positive
Total faecal samples11080
Sewage enrichments2624
Total animals8456
 Sea lions149

esp gene sequence homology between enterococci sources

We obtained 144 sequences of the esp gene from various hosts and environments using the Scott et al. and 1-kb primers (Table 1). The goal of the sequencing was twofold: first, to confirm that the amplicons from our PCR reactions with the Scott et al. primers generate the correct amplicon; and second to determine if any host- or environment-specific variants of the gene exist, which could be targeted for MST applications. These sequences were assembled into neighbour-joining trees according to the primer set that generated them (Figs 3 and 4).

Figure 3.

 Neighbour-joining tree of esp human faecal marker amplicons produced by the Scott et al. primers from Enterococcus enrichment cultures of sewage, faecal, and estuarine water samples (ES indicates estuarine water from Elkhorn Slough, CA). Reference sequences, and their GenBank accession numbers, are shown in bold; 1000-replicate bootstrap values are shown for the major branches.

Figure 4.

 Neighbour-joining tree of a 1-kb section of the esp gene amplified from Enterococcus enrichment cultures of sewage, faecal, and water samples (ES indictaes estuarine water from Elkhorn Slough, CA; LP indicates surf zone seawater from Lover’s Point, CA). Reference sequences, and their GenBank accession numbers, are shown in bold; 1000-replicate bootstrap values are shown for the major branches.

The amplicons produced by the Scott et al. primers clustered into two clades and a singleton. The majority of the clones group with previously published E. faecium esp sequences (Fig. 3), confirming the presence of the intended PCR product. The second, smaller clade at the bottom of the tree contains two unique esp gene sequences from dog faeces, and six unique sequences from estuarine water at Elkhorn Slough, California. While these eight sequences are >99% similar to each other, they have <94% similarity to any of the GenBank reference esp sequences, suggesting this group represents a previously unknown variant of the esp gene. The singleton sequence (dog clone 19) is <99% similar to the reference sequences. As expected, no sequences were highly similar to E. faecalis esp reference sequences, as the assay is designed to be specific for E. faecium esp.

The 1-kb primers were used to amplify and sequence a 1-kb segment of the esp gene (which includes the Scott et al. priming regions) and a neighbour-joining tree was created (Fig. 4). All of the 1-kb esp gene sequences fell into two clades: an E. faecium-like clade and an E. faecalis-like clade, which clustered with their respective reference sequences. All clones in the E. faecium-like clade were from dog faeces, and all had 100% identity to the reference sequences in the Scott et al. priming regions. The E. faecalis-like group contains sequences from seals, horses, sewage, sea lions, gulls, Elkhorn Slough, and Lover’s Point, California. It appears that esp gene sequences are not divergent between human and nonhuman hosts, and that there is no host-specific variant of the gene captured by our clone libraries.


Eighty-three per cent of human faecal samples and 92% of sewage enrichments were amplified with the Scott et al. primers and thus were positive for the esp HF marker. These results indicate that the esp HF marker is not ubiquitous in enterococci from human hosts. Therefore, it is not surprising that Whitman et al. (2007) found that pit toilets in the mid-western United States, which receive limited human faecal sources, do not carry the esp HF marker, and that faeces from septic trucks contained the esp HF marker at low frequencies (30%). However, the methods used in this study do differ slightly from those used in Whitman et al. (2007); we used purified DNA extracts as the PCR template, while Whitman et al. (2007) used washed whole cells.

A majority (64%) of the nonhuman faecal samples contained the esp HF marker. It appears from this study and another (Whitman et al. 2007) that the esp HF marker is not limited to human hosts. In addition, it is known that the E. faecalis variant of the esp gene is also not human-specific (Hammerum and Jensen 2002; Harada et al. 2005; Poeta et al. 2006; Shankar et al. 2006). Frequent occurrence of the esp HF marker in animal faeces is not cause for alarm, as the esp gene alone is insufficient to cause pathogenicity in enterococci (Shankar et al. 2001). It is perhaps not surprising that we found the esp HF marker frequently in both human and nonhuman faecal samples, as all the animals sampled were in relatively close contact with the humans (domestic dogs and horses, gulls on an urban beach, and marine mammals under veterinary care in a rehabilitation centre).

Thirty-seven per cent of the 153 environmental samples were positive for the esp HF marker. This result suggests that either the environments we sampled contained enterococci from faeces (human or nonhuman), or that strains of enterococci indigenous to the extra-enteric environment contain the esp HF marker. The present study cannot distinguish between these two possibilities, because we cannot confirm that environmental enterococci were not of faecal origin. Use of additional source tracking tools (such as the host-specific Bacteroides and enterovirus markers) might provide more conclusive evidence of faecal pollution at these locations. Future work should examine if the esp HF marker can be found in enterococci that do not originate from faeces. Additionally, it remains unknown whether the relative abundance of esp-containing enterococci in faeces varies between hosts. Further work should be performed to determine the relative quantities of esp that each source contributes to the environment, as this data would inform the interpretation of esp presence/absence assays performed on environmental samples.

We obtained 48 esp HF marker sequences using six DNA samples from dog, horse, seal, and sea lion faeces and Elkhorn Slough (Table 1). The largest clade (Fig. 3) contains all of the reference esp HF marker sequences from GenBank, and none of the esp HF marker sequences amplified in this study clustered with the E. faecalis esp reference sequences, confirming that our esp HF assay amplified the intended target. The smaller clade and the singleton did not group with any reference sequences, indicating that these are previously unobserved variants of the esp gene; they are still the intended target, as they have >99% identity to the Scott primer regions. Further, the sequences in the small clade are >95% similar to the esp HF reference sequences at the amino acid level (data not shown). For comparison, the outgroup (E. faecium E470 esp) is 89–94% similar to the esp HF reference sequences at the amino acid level; thus, these unique sequences (small clade and singleton, Fig. 3) are most likely functional esp genes as well.

We cloned 1-kb esp amplicons from eight DNA extracts (Lover’s Point seawater, gull faeces, and the six samples that were used to obtain esp HF marker sequences) and sequenced these fragments to determine if a larger section of the gene would uncover sequence divergence that might be useful for MST. Enterococcus faecium-like 1-kb sequences were found only in dog clones, although the esp HF marker was amplifed and sequenced from other DNA extracts; this suggests that E. faecalis esp targets outnumbered E. faecium esp targets in those extracts. However, we cannot definitively determine that our sequences amplified from a particular species of Enterococcus, considering that the DNA templates were from enrichment cultures.

The sequences from both the Scott et al. primers and the 1-kb primers suggest that the esp HF marker lacks host specificity. Enterococcus faecium and E. faecalis have been shown to have 97·3% homology at the 16S rRNA level, which is less than the 16S rRNA homologies between E. faecium and other group II Enterococcus species (e.g. Enterococcus mundtii and Enterococcus casseliflavus; Patel et al. 1998). There is evidence that the esp gene can be transferred between isolates of E. faecium and E. faecalis via conjugation (Oancea et al. 2004; Lund et al. 2006), which suggests that the gene is not confined to a single species of Enterococcus or strain of E. faecium. Thus, it is perhaps not surprising that the esp HF marker does not exhibit host specificity. Ultimately, the esp gene is a potentially portable genetic element (Leavis et al. 2004), and given the nature of bacterial recombination and horizontal gene transfer in the environment, a mobile gene is a nonideal source tracking target. Future work should explore the possibility of a human-specific portion of the Enterococcus core genome. A new tool (modified transposon mutagenesis) for functional genomic analysis of E. faecalis has recently been developed (Kristich et al. 2008); this method and others could potentially be used to identify human-specific sequences in the core genome for MST applications.


This project was funded through the NOAA Oceans & Human Health initiative (Grant NA04OAR4600195) and the National Science Foundation (Grant OCE-0742048); B.A. Layton is also supported by a National Science Foundation Graduate Fellowship. The authors thank Dr Rob Willems for use of his clinical enterococci strains; Tracey Goldstein and Phil Bobel for their assistance in obtaining faecal and sewage samples; and Alyson Santoro, Daniel Keymer, Nick de Sieyes, Kevan Yamahara, Karen Knee, and Tim Julian for use of their environmental samples and/or suggestions for improving the manuscript. Two anonymous reviewers provided comments for improving the manuscript.