Molecular Detection and Genotyping of Male-Specific Coliphages by Reverse Transcription-PCR and Reverse Line Blot Hybridization

Abstract

In recent years, there has been increased interest in the use of male-specific or F+ coliphages as indicators of microbial inputs to source waters. Sero- or genotyping of these coliphages can also be used for microbial source tracking (MST). Among the male-specific coliphages, the F+ RNA (FRNA) viruses are well studied, while little is known about the F+ DNA (FDNA) viruses. We have developed a reverse line blot hybridization (RLB) assay which allows for the simultaneous detection and genotyping of both FRNA as well as FDNA coliphages. These assays included a novel generic duplex reverse transcription-PCR (RT-PCR) assay for FRNA viruses as well as a generic PCR for FDNA viruses. The RT-PCR assays were validated by using 190 field and prototype strains. Subsequent DNA sequencing and phylogenetic analyses of RT-PCR products revealed the classification of six different FRNA clusters, including the well-established subgroups I through IV, and three different FDNA clusters, including one (CH) not previously described. Within the leviviruses, a potentially new subgroup (called JS) including strains having more than 40% nucleotide sequence diversity with the known levivirus subgroups (MS2 and GA) was identified. We designed subgroup-specific oligonucleotides that were able to genotype all nine (six FRNA, three FDNA) different clusters. Application of the method to a panel of 351 enriched phage samples from animal feces and wastewater, including known prototype strains (MS2, GA, Qβ, M11, FI, and SP for FRNA and M13, f1, and fd for FDNA), resulted in successful genotyping of 348 (99%) of the samples. In summary, we developed a novel method for standardized genotyping of F+ coliphages as a useful tool for large-scale MST studies.

Traditional and alternative indicator microorganisms have been used for many years to assess and predict the presence of fecal pollution in water. Among them, male-specific (F+) coliphages are considered to be promising indicators of viral contamination (17, 24, 13). F+ coliphages are viruses that primarily infect the gram-negative bacterial genera Escherichia, Pseudomonas, Caulobacter, Salmonella, and Vibrio, which possess a plasmid coding for an F or sex pilus (10). These viruses include members of the families Inoviridae (F+ DNA [FDNA]) and Leviviridae (F+ RNA [FRNA]). Because FRNA viruses have been more fully studied while little data are available on the role of FDNA phages, most microbial source tracking (MST) studies focus on FRNA phages. FRNA phages have a cubic (icosahedral) capsid, and their genome consists of positive-sense linear single-stranded RNA. The family Leviviridae contains two genera (Levivirus and Allolevivirus) and three unclassified groups (a, b, and c) (29). The four coding regions of members of the alloleviviruses are oriented in one single reading frame, while those of the leviviruses are located in different reading frames that vary depending on the group (6). Based upon serological cross-reactivity (12), replicase template activity (27), and phylogenetic analysis (6), leviviruses and alloleviviruses each contain distinct subgroups. The genus Levivirus contains group I and group II phages, whereas the Allolevivirus genus contains subgroup III and IV phages (18, 29). Based on the presently available data, group II and III FRNA phages are mainly found in environments influenced by human waste, whereas group I and IV are mostly associated with animal pollution (9, 12, 34, 35). However, as exceptions have been documented, this distinction is not absolute (34). Because antisera for serotyping of FRNA phages are not readily available and may give inconclusive results (4, 18), genotyping using nucleic acid hybridization has been most widely used to differentiate subgroups (33, 34).

The FDNA phages, also called filamentous phages, belong to the family Inoviridae. Several members of the inoviruses, such as M13 viruses, are well characterized due to their value as cloning factors (23). Morphologically, they are nonenveloped flexible rod-shaped filaments containing a circular single-stranded DNA genome of ∼6,400 nucleotides. FDNA phages have usually been detected in environmental waters concurrently with FRNA phages (9). Reportedly, FDNA phages are more resistant than FRNA phages to sunlight exposure (37), and in some studies they have been detected at higher concentrations than the FRNA phages, especially during the summer and fall (9). These observations suggest that FDNA phages could be potentially useful for year-round source tracking studies (9, 25, 26). In contrast to FRNA phages, however, FDNA coliphages are a genetically homogeneous group of viruses and very little is known about their ecology.

To date, molecular detection of F+ coliphages is not routinely performed in the laboratory. To allow further study of the genetic variability of F+ coliphage strains and to identify differences that may be used to discriminate animal from human pollution sources, additional sequence information from field strains would be desirable. Hence, broadly reactive reverse transcription-PCR (RT-PCR) assays are needed. Furthermore, the probes presently used for genotyping FRNA coliphages into groups I to IV were not designed based on the same region of the genome and may not be reactive with certain subgroups of the group III phages (18).

Recently, a filter-based hybridization assay has been described that allow for simultaneous detection and typing of noroviruses without the use of ethidium bromide-stained agarose gels (40). In this assay, virus-specific amplicons hybridize to one of the multiple probes that are covalently linked to a nylon membrane in individual slots on a blot. In the present study, we developed and validated this method, named reverse line blot hybridization (RLB), for the simultaneous detection and genotyping of FRNA and FDNA coliphages. Because the labeled membranes can be reused, RLB is an ideal method for the standardization of coliphage genotyping methods between laboratories.

MATERIALS AND METHODS

Virus strains and host cells.

FDNA prototype strains (M13, fd, f1, and ZJ/2) were obtained from American Type Culture Collection; FRNA prototype strains MS2 (serogroup I), GA (serogroup II), Qβ (serogroup III), FI (serogroup IV), and SP (serogroup IV) were kindly provided by K. Furuse (Tokai University, Bohseidai, Japan), and FRNA strains M11 (serogroup III) and MX-1 (serogroup III) (5) were a gift from J. van Duin (Leiden University, Leiden, The Netherlands). A total of 557 field strains that had been isolated from wastewater, septic systems, surface water, swine lagoons, chicken litter, and cow manure from sites and farms in North Carolina, South Carolina, California, New Mexico, and Massachusetts were analyzed in this study. These field strains were isolated using the single- or double-agar layer method and subsequent RNase testing as described previously (38). Of these strains, 116 had been characterized previously as FRNA by serotyping or genotyping (18), and 230 strains had been characterized as FDNA by neutralization with antisera prepared against prototype FDNA phages (S. L. Long, S. S. El-Khoury, S. J. G. Oudejans, M. D. Sobsey, and J. Vinjé, submitted for publication). Escherichia coli HS(pFamp)R (11) was used as the host strain and was grown in tryptic soy broth (TSB; Difco, Sparks, Md.) supplemented with streptomycin-sulfate (15 mg/liter) and ampicillin (15 mg/liter). All coliphage and host strains were stored in TSB-20% glycerol at −80°C. For the validation of the FDNA and FRNA RT-PCR assays, five different somatic coliphage strains that had been enriched on host strain E. coli CN13 (38) were used.

Generation of sequence database for selection of generic primers.

Multiple alignments were constructed using Clustal W 1.4 software and were based on complete or partial genomic sequences of FDNA phages (M13, f1, and fd) and FRNA phages (MS2, fr, GA, KU1, M11, MX1, Qβ, SP, and NL95) (Table 1). For FDNA phages, gene IV, which encodes the outer membrane pore (secretin), pIV, through which FDNA phages exit from its host, was used as the target region for primer development, because this region showed the greatest sequence variation (range, 2.4 to 5.8%) (Table 2) and thus might be useful to differentiate between strains. For FRNA phages, the replicase gene was selected for the design of genus-specific primer pairs (JV80/JV81 for levivirus, JV40/JV41 for allolevivirus) (Table 3). All oligonucleotide primers were analyzed for the absence of possible hairpins, secondary structure, and melting temperature with NetPrimer (PREMIER Biosoft Int., Palo Alto, Calif.) and OligoAnalyzer 3.0 (IDT-DNA, Coralville, Iowa) primer evaluation software.

TABLE 1.

Sequences of prototype F+ coliphages used for primer development

Family	Genus	Strain	Genome length (nt)	Serogroup	GenBank accession no.	Reference
Inoviridae (FDNA)	Inovirus	fd	6,408		V00602	2
		M13	6,407		NC_003287	42
		f1	6,407		V00606	3
Leviviridae (FRNA)	Levivirus	MS2	3,569	I	NC_001417	41
		fr	3,575	I	NC_001333	1
		GA	3,466	II	NC_001426	19
		Kul	3,486	II	NC_002250	15
	Allolevivirus	Qβ	4,160	III	AY099114	1a
		M11	4,217	III	AF052431	4
		MX1	4,215	III	AF059242	4
		SP	4,276	IV	X07489	20
		NL95	4,248	IV	NC_001891	4

TABLE 2.

Nucleotide sequence variation among different genes of three FDNA coliphage strains (f1, fd, and M13)

Gene	Length (nt)	Range of nt differences (%)
I	1,047	3 (0.3)-31 (3.0)
II	1,233	3 (0.2)-33 (2.7)
III	1,275	13 (1.0)-16 (1.3)
IV	1,281	31 (2.4)-74 (5.8)
V	264	5 (1.9)-10 (3.8)
VI	339	2 (0.6)-4 (1.2)
VII	102	0 (0.0)-0 (0.0)
VIII	222	1 (0.5)-2 (0.9)
IX	99	0 (0.0)-0 (0.0)
Xa	336	0 (0.0)-10 (3.0)
XIb	327	0 (0.0)-4 (1.2)

^aRegion within gene II.

^bRegion within gene I.

TABLE 3.

Oligonucleotides for detection and genotyping of F+ coliphages used in this study

Oligonucleotide identitya	Sequence (5′-3′)b	Orien- tation	Targetc
Primer
JV40	AAG AAC AGT AAR ACW GAT CG	+	3
JV41	CCC ATD GAR GAW ATY TTC TC	−	3
JV80	GAG CCY GAT ATG AAT ATG	+	2
JV81	CCA RAA DAT CAT GGA CTC	−	2
MJV82	GTA TAG AYC TKA AYG AYC A	+	2 and 3
SL2	GTA ACT TGG TAT TCA AAG CA	+	1
SL3	AAG GAG CGG AAT TAT CAT CA	−	1
Probe
M13	TAA ACC TGA AAA TCT ACG C	+	1
fd	TTC CAT AAT TCA GAA ATA TAA C	+	1
CH	ATA ATT TGC GTA ATT TCT TC	+	1
con	TTC TTY ATH TCT GTT TTA CG	+	1
MS2	GAG ACG ATA CGA TGG GAA C	+	2
GA	AAT CAG GAR TTA GCC CGA	+	2
Qb	CAC GAA GGC TCC GTT AMT AAT	+	3
M11	GGC AGC CGT GAY GAT AA	+	3
SP	GAT GGG TCT YTG CTA AAT CAT	+	3
FI	GCA GCC AGC GAC TCT AT	+	3

^aProbes are 5′ hexylamine labeled.

^bIUPAC codes used to indicate degenerate positions.

^c1, inovirus; 2, levivirus; 3, allolevivirus.

Generic FRNA coliphage RT-PCR.

Based on published nucleotide sequences and FRNA coliphage sequences generated in this study, we developed a generic duplex RT-PCR assay with one forward primer (MJV82; Table 3) and two genus (Levivirus and Allolevivirus)-specific reverse primers (JV41 and JV81; Table 3). Different-sized RT-PCR products were generated for strains from each genus (Fig. 1). The assay was developed using a OneStep RT-PCR kit (QIAGEN, Valencia, Calif.) in a 25-μl reaction volume consisting of 0.8 μM MJV82, 0.8 μM JV41, and 0.6 μM JV81. Typically, enriched phage samples were heat released for 5 min at 99°C and then were iced for 2 min, followed by centrifugation for 10 s. For RT, 2.5 μl of a 1:50-diluted phage suspension was reverse transcribed at 45°C for 30 min, followed by heat activation of the Taq polymerase for 15 min at 95°C. PCR consisted of 40 cycles at 94°C for 30 s, 45°C for 30 s, and 72°C for 30 s. After a final extension of 10 min at 72°C, the RT-PCR products (expected size, 266 bp for leviviruses and 229 bp for alloleviviruses) were visualized on an ethidium bromide-stained 2% agarose gel.

FIG. 1.

Schematic representation of the genomic organization of leviviruses and alloleviviruses represented by MS2 (NC001417) and SP (NC004301) and the location of the broadly reactive primers (MJV82, JV41, and JV81). The shaded boxes represent the size of the amplicon generated for the different genera (266 bp for leviviruses and 229 bp for alloleviviruses). Boxed areas represent open reading frames carrying the specific genes.

Generic FDNA coliphage PCR.

Enriched phage suspensions were diluted 1:50 in water, and viral DNA was released by incubation at 94°C for 3 min. FDNA PCR was performed on 1 μl of this material in a 25-μl reaction volume using 0.3 μM each primer (SL2 and SL3; Table 3), 0.2 mM deoxynucleotide triphosphates (Roche, Indianapolis, Ind.), 10 mM Tris-HCl (pH 9.0), 1.5 mM MgCl₂, and 2.5 U of Taq DNA polymerase (NEN Biolabs). Samples were then subjected to 40 cycles at 94°C for 30 s, 50°C for 30 s, and 72°C for 30 s. PCR products (256 bp) were visualized on an ethidium bromide-stained 2% agarose gel.

DNA sequencing and phylogenetic analysis.

RT-PCR or PCR products were gel purified (Qiaquick PCR Purification kit; QIAGEN Inc.) and sequenced at the Lineberger Sequence facility (University of North Carolina, Chapel Hill) using the BigDye Terminator Cycle Sequencing Ready Reaction kit (PE Applied Biosystems). Multiple sequence alignments (based on 190 nucleotides [nt] of gene IV for inovirus, 189 nt of the replicase gene for levivirus, and 171 nt of the replicase gene for allolevivirus), including available sequences of known prototype strains, were generated using Clustal W and imported into TreeCon (V 1.3b) (39). Phylogenetic trees were drawn with Jukes and Cantor correction for evolutionary rate. The confidence values of the internal nodes were calculated by performing 100 bootstrap analyses. All sequences determined in this study are available from the authors upon request.

Design of serogroup- and cluster-specific probes for confirmation and genotyping.

In total, four probes (one consensus and three cluster specific; Table 3) for FDNA and six probes (MS2, GA, Qβ, M11, FI, and SP) for FRNA phages were selected (Table 3). Each probe was carefully designed using the following criteria (40): length between 17 to 22 nt, no hairpins, and no more than three mismatches between probe and strains of other clusters ideally evenly distributed over the probe sequence. All oligonucleotide probes were 5′ hexylamino labeled (IDT-DNA).

RLB.

The RLB protocol used in our study was described previously for genotyping of noroviruses (40) and was adapted for the detection and genotyping of F+ coliphages. A nylon membrane (Biodyne C; Pall Biosupport, Portsmouth, Mass.) was activated for 10 min with 16% (wt/vol) 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (Sigma, St. Louis, Mo.) at room temperature. After rinsing with water, the membrane was placed in a miniblotter (MN45; Immunetics, Cambridge, Mass.), where the oligonucleotide probes (155 μl of twofold dilutions in freshly prepared 0.5 M NaHCO₃) were covalently bound for 2 min at room temperature to the carboxyl groups of the activated membrane (22). The remaining activated esters on the membrane were hydrolyzed by incubation in 0.1 M NaOH for 8 min. Two washes at 60°C for 5 min in 2× SSPE (300 mM NaCl, 20 mM NaH₂PO₄, and 2 mM EDTA at pH 7.2) and 0.1% sodium dodecyl sulfate (SDS) was followed by a wash in 2× SSPE. The membrane then either was used directly for hybridization or was stored at 4°C for later use. The membrane was placed in the miniblotter with the slots perpendicular to the lines of the probes. Five microliters of RT-PCR product that had been generated with a 5′ biotin-labeled reverse primer(s) was diluted with 150 μl of 2- SSPE and 0.1% SDS, and the DNA strands were separated by heat denaturation for 10 min at 99°C. Each sample was then carefully loaded into the slots of the miniblotter. After hybridization for 1 h at 42°C, the membrane was washed twice in 100 ml of 2× SSPE with 0.5% SDS at 42°C for 10 min. The membrane was then incubated in 10 ml of 1.25 U of streptavidin peroxidase (Roche) diluted in 2× SSPE-0.5% SDS for 45 min at 42°C. After four 5-min washes with 2× SSPE-0.5% SDS and one wash with 2× SSPE, bound PCR product was detected by chemiluminescence using ECL detection liquid (Amersham Biosciences, Piscataway, N.J.) and visualized after 30 min of exposure to a Biomax Light X-ray film (Kodak, Rochester, N.Y.). The RT-PCR products were then removed from the membrane using three 10-min washes with 1% SDS at 70°C. After two final washes in 20 mM EDTA, the membrane was sealed in a plastic bag and stored moist at 4°C for reuse.

RESULTS

Selection of genomic regions for the development of generic primers and generation of sequence database.

Our first goal was to develop a generic PCR assay for FDNA and a generic RT-PCR assay for FRNA phages targeting genomic regions that would allow us to develop broadly reactive, genogroup-specific oligonucleotides. To identify such a region for FDNA phages, we constructed a multiple alignment of the complete genomic sequences of the three FDNA prototype strains (M13, f1, and fd) and calculated the differences among these strains for each gene. The sequence variation among the 11 genes was relatively small, with gene IV demonstrating the highest sequence variation (Table 2). Hence, conserved regions flanking regions of sequence variation in gene IV were chosen to design primers SL2 and SL3.

For FRNA phages, two separate multiple alignments of complete genomic sequences from four leviviruses (MS2, fr, GA, and KU1) and five alloleviviruses (Qβ, M11, MX1, SP, and NL95) were constructed. The replicase gene was chosen as the target region for design of generic diagnostic primers, because the capsid gene was too heterogeneous. Using primer pairs JV40/JV41 for alloleviviruses and JV80/JV81 for leviviruses, all FRNA prototype strains and 106 (99.1%) of the 107 serotyped or genotyped FRNA field strains yielded distinct RT-PCR products of the appropriate size that were confirmed by DNA sequencing. One strain tested negative and did not produce lysis of the E. coli Famp host upon reenrichment. Based on these novel sequences and on published sequences, a duplex primer pair MJV82/JV41/JV81 was designed to detect both levi- and alloleviviruses in a single RT-PCR.

Evaluation of generic RT-PCR or PCR assays. (i) FDNA PCR.

A panel (n = 83) of plaque-purified FDNA field strains that were neutralized by at least one of the antisera prepared against FDNA prototype strains (M13, f1, fd, or ZJ/2), five somatic phages, and six FRNA phages (MS2, GA, Qb, M11, SP, or Fi) were used for validation of the FDNA PCR assay. No PCR products were obtained with either somatic phages or the FRNA phages as template. All panel strains yielded PCR products of the expected size (256 bp) and could be confirmed by DNA sequencing.

(ii) FRNA RT-PCR.

In total, 107 FRNA strains that had been serotyped by neutralization or had been genotyped, five somatic phages, and four FDNA phages (M13, f1, fd, and ZJ/2) were used for validation of the duplex FRNA RT-PCR assay. No RT-PCR products were obtained for the somatic and FDNA phage isolates. All (100%) FRNA phage isolates tested positive in the duplex RT-PCR, with 47 samples showing an RT-PCR product (266 bp) for leviviruses and 59 samples giving a slightly smaller product (229 bp) for alloleviviruses (Fig. 2).

FIG. 2.

Results of ethidium bromide-stained agarose gel of products of generic FRNA RT-PCR (266 bp for leviviruses and 229 bp for alloleviviruses). Templates consist of RNA from field strains (lanes 1 to 15), MS2 and M11 as positive controls for levi- and alloleviviruses, respectively, and water as negative control. The RT-PCR products are flanked by a 100-bp DNA molecular size marker (Promega).

Genetic diversity of F+ coliphages. (i) DNA phages (inoviruses).

To investigate the genetic variability of FDNA phages, we determined the nucleotide sequence of a 190-nt region of gene IV from 101 field strains from different sources (wastewater, septic system, surface water, cow manure, and swine lagoon). All strains fell into three different clusters (M13, fd, and CH) (Fig. 3A), with a nucleotide sequence diversity ranging from 9 to 12% between clusters. A novel cluster, designated CH, including strains isolated from surface water, wastewater, and cow slurry, was identified that showed >10% sequence disparity with the known prototype strains (M13 and fd). No relationship between the genetic clusters and likely origin of the source (human and nonhuman) was observed.

FIG. 3.

Phylogenetic relationships among FDNA coliphages (inoviruses) (A), FRNA coliphages (leviviruses) (B), and FRNA coliphages (C). For FDNA coliphages the phylogenetic tree is based on a 190-nt region of gene IV of 47 field strains and the prototype strains f1, M13, and fd. Three genetic clusters designated M13, fd, and CH could be identified based on >5% nucleotide sequence diversity. Genetic relatedness among leviviruses is based on a 189-nt region of the replicase gene of 32 field strains and the prototype strains of genotype I (MS2 and Fr) and genotype 2 (GA) and for alloleviviruses is based on a 171-nt region of the replicase gene of 34 field strains and the prototype strains of genotype 3 (Qβ, M11, and MX1) and genotype 4 (NL95 and SP). Bootstrap values of internal nodes are indicated.

(ii) RNA phages (levi- and alloleviviruses).

To investigate the genetic diversity of the FRNA strains, we determined the nucleotide sequence of a region of the replicase gene of 32 levivirus and 34 allolevivirus field strains that were selected based on RLB hybridization results (see below). Leviviruses can be grouped into three main clusters (MS2-like, GA-like, and a potential novel group designated JS) (Fig. 3B). The two JS strains, both isolated from wastewater samples, had >40% sequence diversity with strains from MS2 or GA genetic groups. Within these genetic groups, sequence diversity between strains ranged from 10 to 18%. Among the alloleviviruses, strains fell into four different genetic groups represented by Qβ, M11, SP, and FI, with sequence diversity between groups ranging from 43 to 53% and from 0 to 15% within clusters (Fig. 3C). Of note, a single strain (G3&4_7, isolated from gull guano) had 25% sequence difference from the prototype strain FI, whereas all strains clustering with strain M11 were identical.

Development of serogroup- and cluster-specific probes for F+ coliphage genotyping by RLB.

The probes specific for the genetic clusters of FDNA and FRNA viruses (Table 3) were developed based on available sequences of prototype strains from GenBank and field strains generated in this study. In defining the distinct clusters, we used an arbitrary cutoff of >7.5% nucleotide sequence identity for FDNA phages and >30% nucleotide sequence identity for FRNA phages. For FDNA viruses, a generic probe was developed, whereas no conserved sequence regions for the design of consensus probes for both genera of the FRNA viruses could be found. For the differentiation of F+ coliphages (both DNA and RNA), a probe specific for each genetic group was developed. With the exception of a weak cross-reaction between the FI probe and occasionally another subgroup IV field sample (Fig. 4, lane 20), no cross-reactions with heterologous probes was observed. Optimal probe dilutions were determined empirically by testing four different probe concentrations (200, 100, 50, and 25 pmol).

FIG. 4.

Detection and simultaneous genotyping of 18 FRNA (lane 21 to 38) and 11 FDNA coliphage (lane 4 to 14) field strains from different animal- and human-affected (sewage) sources. For reference, hybridization patterns of strains representing each individual genetic cluster (FRNA, lanes 15 to 20; FDNA, lanes 1 to 3) are shown. Probes specific for the detection of six different FRNA genetic clusters, three different FDNA genetic clusters, and a conserved probe (con) for all FDNA viruses as described in Table 3 are indicated.

Evaluation of RLB assay for the detection and genotyping of coliphages.

In total, 216 FDNA and 135 FRNA RT-PCR-positive strains were used for validation of the F+ coliphage RLB hybridization assay. All 216 FDNA PCR-positive samples and 132 (98%) of the 135 RT-PCR-positive FRNA samples were confirmed by RLB. A selection of the strains (n = 29) is shown in Fig. 4. Overall, the hybridization patterns were unambiguous. Surprisingly, some FDNA field strains did react with the M13 probe while no signal was obtained with the consensus probe (Fig. 4, lane 11). Other strains reacted very weakly with their respective probes (GA-positive strain; Fig. 4, lane 27). Of the three RLB-negative FRNA strains, two strains (2GI13 and WWTP1_50) belonged to a potential novel levivirus cluster tentatively called JS (Fig. 3B), and one strain (G3&4_7; Fig. 4, lane 31) was FI-like by sequencing but had three mismatches with the FI probe.

DISCUSSION

Information on the human and nonhuman origin of fecal pollution may contribute to the assessment of public health risks and to the implementation of control measures (36). F+ coliphages may be a useful tool to fingerprint fecal pollution in surface waters, wastewaters, and oysters (9, 12, 18). With few exceptions, type II and type III FRNA coliphages have been associated with human waste-affected environmental waters, whereas type I and IV genotypes have been detected in animal feces (4, 13, 18, 34). The reported exceptions include the occasional detection of type II and type III coliphages in feces from pigs and poultry (16, 34). These viruses might, in fact, constitute distinct genetic subgroups within the type II and III viruses, but sequence information is lacking. Therefore, FRNA coliphage genotyping cannot presently be used for absolute distinction of human and animal fecal pollution, and additional genetic information about coliphage field strains from a wide variety of human- and animal-affected sources is needed to address this question. Present genetic classification of F+ coliphages involves distinguishing FRNA and FDNA coliphages by RNase testing (18) followed by preparing at least four replicates of nylon membranes that had been incubated on the agar surface of bacteriophage spot plates (4, 18, 34). This method has proved useful in several studies (7, 14, 34).

In this study, we have adapted an RLB method previously developed for the simultaneous detection and genotyping of Borrelia spp., Mycobacterium tuberculosis, and noroviruses (21, 30, 40) for the genotyping of F+ coliphages. Because this method uses membrane-bound genotype-specific probes which react with biotinylated RT-PCR products, we first developed an RT-PCR assay for the generic detection of FRNA viruses (leviviruses and alloleviviruses) and a generic PCR assay for FDNA viruses (inoviruses). We selected conserved regions of the replicase gene of FRNA phages for primer development, because the sequence variation among capsid genes as well as their flanking regions was too heterogeneous. This selection was supported by similar phylogenetic grouping of strains regardless of the gene (capsid or replicase) chosen for analysis (6). Using a one-step duplex RT-PCR assay yielding different amplicon sizes for leviviruses and alloleviviruses, we were able to detect all strains of a panel of previously serotyped or genotyped strains. A replicase gene-based RT-PCR has been described previously, but only a limited number of strains were evaluated (32).

In addition, a PCR assay for the detection on FDNA phages (inoviruses) was developed which was able to detect all samples from plaque-purified FDNA field strains. The generic primers were directed to gene IV, because this gene was shown to have the largest nucleotide sequence variation (5.8%) among a limited number of prototype strains and thus might be potentially useful for further subtyping of isolates to better discriminate between phages isolated from different sources.

Our study provides valuable sequence information of FRNA and FDNA coliphage field strains isolated from known sources, which allowed us to develop genotype-specific probes for a robust genotyping assay and gives us a better understanding of the genetic variation of these viruses in nature. For FRNA coliphages, the phylogenetic results obtained are consistent with estimates of relatedness based on serological cross-reactivity data (12). To date, genotype I viruses (MS2-like) and genotype II viruses (GA-like) have been recognized as members of the leviviruses (6). In this study, a potential novel genotype (JS) within the Levivirus genus was identified. The two JS strains were isolated from wastewater samples from two geographically different regions (Massachusetts and South Carolina), confirming that JS viruses may form a stable lineage within the leviviruses. Further genomic sequence and serological data are needed to confirm that they belong to a novel subgroup or genotype or whether these strains are a result of recombination or rearrangement events (8, 28).

FDNA phages are primarily known from the widespread use of M13 as a platform for phage display technology (31). However, little is known about the ecology of these viruses (12), and no information is available about whether specific subgroups of FDNA coliphages exist and are associated with specific sources. Our data are the first on the ecology of FDNA phages isolated from different sources. A more comprehensive study on these sources in relation to the diversity of FDNA phages will be described elsewhere (S. C. Long, S. S. El-Khoury, S. J. G. Oudejans, M. D. Sobsey, and J. Vinjé, submitted for publication). Based on the nucleotide diversity of a 190-nt region of gene IV, we were able to identify three different genetic clusters and one (CH) which has not been reported previously. This gene was arbitrarily chosen and does not reflect antigenic variation that may or may not exist among FDNA phages. Therefore, more studies are needed to confirm the results reported in this study before recommendations can be made regarding whether the presence of FDNA coliphages or certain subtypes can be used for source tracking of fecal contamination.

Genotype-specific probes for typing by RLB were developed based on multiple alignments of prototype and field strains targeting six different genetic FRNA clusters (MS2, GA, Qβ, M11, SP, and FI) and three different FDNA clusters (M13, fd, and CH). In addition, one generic FDNA probe was developed. The probes were immobilized onto a nylon membrane. Biotin-labeled amplicons were amplified using duplex RT-PCR for FRNA coliphages or PCR for FDNA coliphages and were hybridized to the membrane. Using this assay, we were able to detect and differentiate 98% of the FRNA strains and all FDNA strains in this study. Compared to the direct hybridization of viral RNA to a set of group specific probes (5, 18), the RLB method requires an RT-PCR step prior to hybridization and thus requires access to a thermocycler. However, recent advances in available ready-to-use one-step RT-PCR makes this technology a simple, routine tool in most laboratories. In addition, equivocal results with RNase testing and the finding in a recent study that a significant number of isolates cannot be propagated after isolation and thus cannot be tested for RNase inhibition (J. R. Stewart, J. Vinjé, S. J. G. Oudejans, G. I. Scott, and M. D. Sobsey, submitted for publication) also supports the advantages of the RLB method. Additionally, the same RLB membrane can be reused without substantial loss of activity (40). This offers the possibility of standardization of F+ coliphage genotyping among different laboratories and throughout a study. In our laboratory, we now use this coliphage RLB assay routinely to genetically classify F+ coliphage field strains.

In conclusion, we have developed novel laboratory assays for the genotyping of F+ coliphages. These assays include a duplex one-step RT-PCR assay for the generic detection of FRNA coliphages, a PCR assay for FDNA coliphages, and a novel RLB method for the genotyping of both RNA and DNA F+ coliphages. The RLB method is rapid, reproducible, cheap, and easy to perform with a high throughput of samples. This makes it an ideal candidate to become a standardized method for the detection of FRNA and DNA coliphages as source-specific indicators of fecal contamination in environmental waters and shellfish.