key: cord-252604-1u14i4v1
authors: Smith, Alvin W.; Iversen, Patrick L.; Skilling, Douglas E.; Stein, David A.; Bok, Karin; Matson, David O.
title: Vesivirus viremia and seroprevalence in humans
date: 2006-03-22
journal: J Med Virol
DOI: 10.1002/jmv.20594
sha: 
doc_id: 252604
cord_uid: 1u14i4v1

Pathogenic caliciviruses of the genus Vesivirus circulate in oceanic ecosystems and spread to and among terrestrial mammals. Isolation of Vesivirus from natural and laboratory infections in humans led to this investigation of Vesivirus seroprevalence and viremia. Sera from four groups were tested for antibodies to Vesivirus as follows: blood donors whose units were cleared for donation, blood donors whose units were not accepted for donation solely because of elevated blood liver alanine aminotransferase (ALT) concentrations, patients with clinical hepatitis of unknown but suspected infectious cause, and patients with clinical hepatitis of unknown cause but associated with blood transfusion or dialysis. Additionally, sera were tested for Vesivirus genome by three methods: dot‐blot and two reverse transcription‐polymerase chain reaction (RT‐PCR) methods. The calculated seroprevalence against Vesivirus virions within these groups (N = 765) was 12%, 21%, 29%, and 47%, respectively (P < 0.001 for group differences). Additionally, 11 (9.8%) of 112 sera tested yielded RT‐PCR amplicons that by nucleotide sequence were distinct from each other and related to known Vesivirus. These data indicate that some blood donors in the population tested have serologic evidence of previous Vesivirus infection and some also have Vesivirus viremia. These results justify further investigation of an association between Vesivirus infection and illness in humans. J. Med. Virol. 78:693–701, 2006. © 2006 Wiley‐Liss, Inc.

Recent spread of newly recognized human pathogens from animal reservoirs, for example, Nipah virus, coronavirus associated with severe acute respiratory syndrome (SARS), and avian influenza strains, has focused renewed study of viral ecology [Enserink, 2000 [Enserink, , 2003 Smith, 2000; Trampuz et al., 2004] . A relatively under-explored potential reservoir of newly emergent human pathogens is the marine environment [Smith et al., 1998b; Culley et al., 2003] . We previously described laboratory and natural marine Vesivirus calicivirus infections in humans [Smith et al., 1998a] .

Vesivirus is one of four genera in the Caliciviridae, the others being Norovirus and Sapovirus, common causes of gastroenteritis in humans, and Lagovirus, widespread in Eurasia in rabbits and hares [Green et al., 2000] . Within the Vesivirus genus, serotypes of marine origin form a polyphyletic grouping distinct from those designated feline caliciviruses (FCVs) [Berke et al., 1997] . Vesivirus cycles naturally in oceanic reservoirs and a single Vesivirus serotype has infected species as diverse as fish, seals, shellfish, swine, cattle, and primates, including humans, and caused disease, sometimes severe, in marine and domesticated food-producing animals and humans [Smith et al., 1998a,b] . Classical virologic investigations established that epidemics of vesicular exanthema of swine virus (VESV) in North America from 1932 through 1959 were caused by serotypes of marine origin [Smith et al., , 1980b Smith, 2000] . The VESV epidemic serotypes (N ¼ 13) are closely related to serotypes (N > 25) of San Miguel sea lion viruses (SMSVs) also of ocean origin and the few other Vesivirus serotypes (N ¼ 5) isolated from terrestrial animals since 1970 [Smith et al., , 1980b Smith and Boyt, 1990; Neill et al., 1995 Neill et al., , 1998 Matson et al., 1996; Reid et al., 1999; Smith, 2000] . Viruses in the marine environment other than caliciviruses also are being discovered, a recent example being the characterization of four possibly new virus families related to caliciviruses at a high taxonomic level [Culley et al., 2003; Lang et al., 2004] . Of interest, the VESV epidemics were controlled, in part, by cooking of raw garbage containing fish and pork scraps fed to swine [Smith et al., , 1980b Smith and Boyt, 1990] .

The FCVs also are in the genus Vesivirus and neutralizing antibodies to them occur in marine mammals [Berke et al., 1997; Smith, 2000] . Other Vesivirus strains (N ¼ 5) as yet only isolated from terrestrial sources have antigenic and/or genetic properties similar to the Vesivirus serotypes often referred as the marine caliciviruses (e.g., Barlough et al., 1998; Smith, 2006] .

Vesivirus infection can have a multiplicity of clinical outcomes, dependent upon the host and strain, which include encephalitis, hepatitis, myocarditis, pneumonia, spontaneous abortion, and dermatitis (reviewed in [Smith et al., 1998b [Smith et al., , 2002a Smith, 2000] ). The human cases of marine Vesivirus disease which were described previously occurred in association with the North Pacific Basin, one in a laboratory scientist working with a marine Vesivirus who experienced a disseminated vesicular exanthem and the second in a field biologist working with marine mammals who had a vesicular exanthem on the face [Smith et al., 1998a] . One strain recovered from these human cases was by genome sequence related most closely to SMSV-5, the serotype known to infect more than 20 animal species and cause different clinical manifestations in them [Smith et al., 1977 [Smith et al., , 1980a [Smith et al., , 1998b Barlough et al., 1986] . The other strain from a human case was of a new Vesivirus serotype. To assess whether Vesivirus infections are more widespread among humans, the seroprevalence against Vesivirus and possible Vesivirus viremia were examined in adults whose donated blood was tested at a northwest-U.S. regional blood laboratory and, because of initial results, in patients with clinical hepatitis of unknown etiology.

Sequential, de-identified sera from blood donors were collected over a 4-year period (1996) (1997) (1998) (1999) , based upon availability, from the Regional Red Cross Blood Testing Laboratory in Portland, Oregon, which serves eight Western States and was processing samples from about 3,000 units of blood per day. One sample set included blood cleared for donation after clinical and laboratory screening, which included negative test results for hepatitis B surface and core antigens, HIV-1 and 2, HIV P24 antigen, human T-cell lymphotropic virus type 2, hepatitis C antigen, and serum alanine aminotransferase (ALT) levels below 120 International Units (IU). A second, concurrently acquired set of sequential blood specimens was collected from donors whose pre-donation laboratory values differed from the first set only in that the ALT level was 120 IU. Based upon testing of these two sets of samples, three additional sets of sera from cases of clinical hepatitis were acquired from an industry source. Two of these sets represented hepatitis cases of unknown but suspected infectious cause. Both were negative for known hepatitis viruses, but differed from each other in that some cases were associated with transfusion or dialysis. Other samples came from hepatitis cases caused by hepatitis B or C virus. The study was approved by the Institutional Review Boards of the authors' institutions.

Antigen for serum antibody detection was generated from four genus Vesivirus family Caliciviridae serotypes, three SMSVs and one a feline Vesivirus prototype, as follows: SMSV-5, originally recovered from blisters on the flippers of a Northern Fur Seal in the Bering Sea and subsequently from skin lesions on a researcher [Smith et al., 1998a] ; SMSV-13, recovered from large skin erosions on the flippers and mouth of a California Sea Lion (Zalophus californianus) at a marine mammal rescue facility and subsequently shown to cause similar disease in swine and cattle [Berry et al., 1990] ; SMSV-17, initially isolated from infected shellfish and an aborted California Sea Lion fetus on the Santa Barbara Channel Islands off the coast of California and subsequently demonstrated to establish persistent infections in experimentally exposed shellfish [Smith, 2000; Burkhardt et al., 2002] ; and prototype FCV F9 vaccine strain [Meanger et al., 1992] .

Each virus strain was grown in roller bottles to a titer of 10 7 plaque-forming units per ml in Vero cells. When cytopathology was scored 4þ (24 h post-inoculation), cells were lysed by freeze-thawing, cellular debris was removed by centrifugation and the supernatant applied to a continuous CsC1 gradient with a mean density of 1.35 g/ml. Virus bands were visible at a density of about 1.36 g/ml and were removed, cleared of CsC1 by dialysis and checked for purity and concentration by direct electron microscopic examination. Optimal dilutions of individual purified virus stocks were determined, then equal parts of each of each SMSV stock were pooled and both preparations (SMSV pool and F9) were retitrated for optimal dilution.

The EIA utilizing the SMSV pool antigen was described previously [Smith et al., 1998a] . SMSV pool antigen was applied directly to the plate and then incubated with a 1:100 dilution of human serum. Bound immunoglobulin was detected by an alkaline phosphatase-conjugated, anti-human immunoglobulin second antibody (1:40,000 dilution; Sigma, St. Louis, MO). Optical densities (ODs) of the color substrate Blue Phos (KPL, Gaithersburg, MD) were read at 650 nm on a Titertek Multiskan Plus EIA Plate Reader (Titertek, Huntsville, AL). For comparison, sera were tested against wells with no viral antigen. EIA data were analyzed as P-N values, in which the OD of the antigennegative well (N) was subtracted from the OD of the well containing viral antigen (P). The OD values of antigennegative (N) wells were 0.005 and most were 0.000. The cut-point for positivity was a P-N 0.043. Previous testing [Smith et al., 1998a] demonstrated antibody in EIA-positive samples by Western blot and by preabsorption with SMSV-5 antigen, and that paired sera from cases of Norwalk virus gastroenteritis showed no seroresponse to the SMSV pool. Vesivirus typing antisera show broad cross-reactivity of the SMSV-5 capsid antigen with that of other Vesivirus serotypes [Kurth et al., 2006] .

The FCV antigen also was utilized in the same EIA, in which all other reagents and steps were the same. The cut-point for the FCV antigen also was a P-N value 0.043, which was derived by utilizing antigen of assay potency comparable to that of the SMSV pool and a graphical method for estimating cut-point [Matson, 1999] .

To test the possibility that sera may contain Vesivirus genomes, as suggested by the clinical evidence of viremia in cases of Vesivirus illness, including humans [Smith et al., 1998a] , three complementary methods targeting three Vesivirus genomic regions were utilized to detect Vesivirus RNA in serum ( Fig. 1 ): dot blot for the ORF1 3C protease region, reverse transcriptionpolymerase chain reaction (RT-PCR) for the 3 0 terminal region of ORF1 encoding a portion of the viral RNA polymerase or for a portion of the viral capsid protein, and nucleotide sequencing of RT-PCR amplicons.

For the dot-blot method, a 280-bp biotin-labeled riboprobe (5R.3) specific for the Vesivirus 3C protease genomic region (designed from SMSV-5 and corresponding to nts 3,727-4,007 of primate Vesivirus prototype Pan-1, GenBank number AF091736) was synthesized by in vitro transcription from recombinant plasmid pCN5R.3 (not shown) and probed against 50 ng of total RNA extracted from blood donor serum using Trizol (Invitrogen, Inc., Carlsbad, CA) blotted onto a nylon membrane. Comparison controls also blotted included 20 ng of RNA extracted from CsCl-banded SMSV-5 and 50 ng of total RNA from mouse embryo (Clontech, Inc., Palo Alto, CA). The membrane was probed at 508C in Northern Max Hyb Buffer (Ambion, Inc., Austin, TX) and washed at high stringency (2X SSC, 1% SDS, 608C).

For one RT-PCR (RT-PCR 1), total RNA was extracted from 100 ml of serum using Trizol. A positive control (tissue culture of primate calicivirus) and a negative control (water) were included in each round of extraction in order to control for the quality of the extraction. Extracted RNA was diluted in water containing 0.4 U/ml of RNAsin (Promega, Madison, WI) and used directly in RT-PCR assays. The RNA was amplified with primers D3A 10F (5 0 -CCAAA GCCAA CAACC GTTGG TTCCA TG-3 0 , designed from prototype A48 strain, GenBank number U76884, and corresponding to nts 6,962-6,988 of Pan-1) and D3A 481R (5 0 -GTGTA GCAAT CCTGA CAACT TTGCT GG-3 0 , designed from prototype A48 strain, and corresponding to nts 7,415-7,441 of Pan-1) and yielding a product of approximately 480 nt. This RT-PCR utilized Enhanced Avian Reverse Transcriptase (Sigma) according to manufacturer's instructions and Platinum Pfx DNA Polymerase (Invitrogen) also following the manufacturer's recommended protocol. For the second RT-PCR (RT-PCR 2), total RNA from serum was amplified in a similar manner using primers 289rc, 5 0 -TATGG TGATG ACGGG GTCTA CA-3 0 (designed from nucleotides 5,123-5,144 of Pan-1), and primer ''precapsid,'' 5 0 -CACCT CACCA CTGAG CCC-3 0 (designed from nucleotides 5,650-5,633 of Pan-1), and yielding a product of approximately 510 base pairs. RT-PCR amplicons were resolved in a 1% agarose gel and visualized by ethidium bromide under UV light.

For nucleotide sequencing of RT-PCR amplicons, RT-PCR products were sequenced directly or after cloning. Cloning used the pGEM-T vector system I (Promega) following the manufacturer's protocol. Recombinants obtained from cloning were screened by PCR using forward and reverse primers targeting the SP6 and T7 polymerase promoters bracketing the multiple cloning cassette. Plasmid DNA from the positive clones was purified by QIA prep Spin Miniprep Kit (Qiagen, Valencia, CA) and sequenced with Sequi Therm EXCEL II Long-read DNA Sequencing Kit (Epicentre Technologies, Madison, WI) on an automated sequencer (ALF Express DNA Sequencer, Amersham Pharmacia, Piscataway, NJ).

Raw DNA sequences were edited in OMIGA 2.0 (Accelrys, San Diego, CA). Sequences were compared and aligned in CLUSTAL W with prototype sequences retrieved from the NCBI public databases using BLAST searches. Comparison Vesivirus sequences included human Hom-1 (GenBank number U623227), primate Pan-1 (AF091736), SMSV-2, SMSV-5, SMSV-6, SMSV-7, SMSV-13, and SMSV-14 (U18730, U18731, U18732, U18733, U18734, and U18735), VESV strains A48, C52, E54, I55, and 1934b (U76874, U18738, U18739, U18740, and U18736), bovine (U18741), feline strain F9 (M86379), rabbit (AJ866991), skunk strains 4-1L, 4-2S, and 7-2 (U14668, U14670, and U14672), and walrus (AF321298). Two sequences of SMSV-1 (U15301 and U14676), called here SMSV-1A and SMSV-1B, and of SMSV-4 (U15302 and U14674), called here SMSV-4A and SMSV-4B, were included.

GenBank numbers of sequences utilized or determined in this study include: for the 3C region riboprobe, DQ300285; for the capsid region amplicon of RT-PCR 1: sample 4, DQ300286; for the polymerase region amplicons of RT-PCR 2: sample N102, DQ300287; sample N104, DQ300288; sample 214, DQ300289; sample 298, DQ300290; sample 310, DQ300291; and sample N330, DQ300292.

Healthy blood donors (Normal) are adults volunteering to donate blood and whose donation cleared screening tests and was accepted for donation. Most are repeat donors. Persons with elevated ALT (high ALT) are other blood donors, like the healthy donors, except their blood donation was rejected solely because of the detection of elevated serum ALT values in the blood at the time of donation. Samples from persons with infectious hepatitis (B and C) and hepatitis of unknown etiology (non-A-G hepatitis) came from an industry repository established to discover new causes of human hepatitis.

Optical density (P-N) values were compared between groups by ANOVA and, for correlation, by regression. Difference in estimated prevalence of anti-Vesivirus antibodies among groups were assessed by the w 2 test. The trend of estimated prevalence among groups with increasingly greater evidence of hepatitis was assessed by the Mantel w 2 for the trend. P-values <0.05 were considered significant.

A total of 374 sera from Normal and 350 sera from high ALT blood donors were tested against both Vesivirus antigens. Results from this testing led to obtaining a set of sera from patients with hepatitis (see Materials and Methods) and retesting against the SMSV pool antigen because the estimated prevalence of antibody to both antigens in the high ALT donors was about twice that observed among the Normal blood donors (P < 0.05, Yates' corrected w 2 for each comparison) and mean P-N values were significantly higher among the high ALT donors than among the Normal donors when tested against the SMSV pool antigen but not when tested against the F9 antigen (P < 0.05, Student's t-test). The sera from all groups were recoded and tested together; the results of this combined testing are presented in Table I .

In the retesting, high ALT donors had a higher prevalence of antibody to both test antigens (21% for the SMSV pool and 14% for F9) than the Normal donors (P < 0.001 and P ¼ 0.026, respectively). The retesting also yielded a higher mean P-N value among positive samples in the high ALT donors for antibody to the SMSV pool (P < 0.001), but not for antibody to F9. Sera from 51 cases of clinical hepatitis had a yet higher estimated prevalence (25%) of anti-SMSV pool antibody [Dean et al., 1994] . and mean P-N value among positive samples. When these hepatitis cases were analyzed by subgroup, the highest prevalence (47%) of antibody to the SMSV pool was in the cases of Non-A-E hepatitis associated with transfusion or dialysis. Cases of Non-A-E hepatitis with an unknown exposure source had an estimated prevalence (19%) similar to that (21%) of the high ALT donors and cases of HBV or HCV hepatitis had an estimated prevalence (10%) similar to that of healthy donors. The increase of anti-SMSV pool prevalence along the axis ''Normal donor-high ALT donor-Non-A-E hepatitis case associated with transfusion or dialysis'' was highly significant (P < 0.001, w 2 for the trend). The mean P-N value of the positive samples from the Non-A-E hepatitis group was lower than that of the other groups studied (P > 0.05 for these comparisons).

The patterns of significant differences in estimated prevalence were similar and statistically significant for a range of antibody assay cut-points, up to 0.200, although estimated prevalence declined as the cut-point increased.

Total RNA was extracted from 30 donor sera and tested by RT-PCR 1. The laboratory performing this testing had no prior experience with Vesivirus genomes in the facility. RT-PCR 1 generated a Vesivirus amplicon in one of these samples that was confirmed by dotblotting (Fig. 2) and sequencing.

To reduce the possibility that a positive result for serum Vesivirus genome detection was generated by laboratory contamination at the first facility, a further 82 donor samples were tested in a second laboratory routinely performing RT-PCR and that had prior experience with Vesivirus RNA, but using newly designed primers to amplify a genomic region not previously amplified (RT-PCR 2). Ten of these 82 samples yielded amplicons of the expected size (Fig. 3) .

Ten (11%) amplicons came from 91 high ALT donor samples and 1 (4.8%) amplicon from 21 Normal donor samples (P ¼ 0.64, Yates' corrected w 2 ). Five (15%) of the amplicons were from 34 sera that scored EIApositive and 6 (7.7%) from 78 sera that scored EIA-negative (P ¼ 0.42, Yates' corrected w 2 ).

Six amplicons from the RT-PCR 2 (polymerase region) sample set and the one from the RT-PCR 1 (capsid region) sample set were successfully sequenced. Amplicons not successfully sequenced yielded a band too faint for successful direct sequencing and were not successfully cloned. Five of the six polymerase region amplicons were distinct from each other but closely related (1-6% divergence). The sixth amplicon, study number N104, was distinct from the other five polymerase region amplicons (24-38% divergence from N104). When the five similar polymerase region amplicons were compared with GenBank entries spanning the amplification region of RT-PCR 2, the two best matches were with primate Vesivirus Pan-1 (88-96% identity) and SMSV-6 (88-94%) and lower match was with SMSV-5 (84-86%).

J. Med. Virol. DOI 10.1002/jmv Fig. 2 . RT-PCR 1 product and confirmatory dot blot. RT-PCR 1 testing of blood donor sera revealed an amplicon of the expected size for Vesivirus RNA (A) and dot blot probing of total RNA from the same serum provided evidence of Vesivirus viremia (B). In (A), Lane 1: molecular weight ladder; lane 2: RT-PCR 1 amplicon. In (B), dot 1 is 20 ng of RNA extracted from CsCl-banded SMSV-5, dot 2 is 50 ng of total RNA from mouse embryo, and dot 3 is 50 ng of total RNA extracted from the donor serum sample. Fig. 3 . RT-PCR 2 products from blood donor sera. Several amplicons generated by RT-PCR 2 were of the expected size (lanes 5, 10, 19, and 22) . Lanes 1, 13, and 24: molecular weight ladder. Lane 2: positivecontrol RT-PCR 2. Lanes 3, 11, and 20 are negative controls containing all RT-PCR reagents without test specimen.

The sixth polymerase region amplicon (N104), was closest (68% identity) to SMSV-13 and 44% identical to F9. The capsid gene sequence from RT-PCR 1 best (97% identity) matched SMSV-5.

A smaller nucleotide (113 nt) region of the six polymerase region amplicons was shared with 22 SMSV-like Vesivirus entries in GenBank (Fig. 4) . The five similar polymerase region amplicons were distinct from each of the GenBank entries and diverged from published Vesivirus sequences for the most part at sites where other Vesivirus strains also differed from each other (Part A). The amplicon from donor N104 also had a Vesivirus sequence, but was distinct from other Vesivirus strains, including F9, differing from published Vesivirus most at the 5 0 end of the sequence, in a manner akin to how FCV differed from SMSV-like Vesivirus strains. Strain Hom-1 derived from a skin vesicle of a previously reported case of human Vesivirus infection [Smith et al., 1998a] had a short (47-64 nt) region of overlap with the serum-derived polymerase region amplicons. In this region of overlap, nucleotide sequence identity with the polymerase region amplicons ranged from 66% to 83%.

In this report, the detection of antibodies to Vesivirus and of SMSV-like Vesivirus viremia in blood donors is described at a U.S. regional blood bank serving eight Northwestern states. A higher prevalence of anti-Vesivirus antibody was observed, in comparison with normal blood donors, in donors with elevated serum ALT values, and the highest prevalence, among the groups tested, in cases of clinical hepatitis of unknown etiology associated with transfusion or dialysis. In addition, Vesivirus RNA was detected in healthy and high ALT blood donors, occurring numerically but not statistically significantly more frequently in the high ALT blood donors than among the healthy donors tested. The findings of this study and the attributes of Vesivirus calicivirus strains in mammals extend the potential for Vesivirus disease in humans from a few well-described cases to a broader population.

The complementary evidence for viremia included positive results from separate laboratories testing separate sample sets, utilization of RT-PCR assays that amplified non-overlapping genomic regions, direct detection of genomic RNA by hybridization, and amplicon sequencing that revealed non-identity of sequence compared with known Vesivirus, including strains characterized in one of the laboratories previously [Rinehart-Kim et al., 1999] and among the set of characterized amplicons themselves. These results together indicate that the positive laboratory results were not because of contamination in sample collection and handling.

The antigens utilized for detection of anti-Vesivirus antibody were purified by cesium chloride banding, which yields a homogenous population of viable particles. A pool of three SMSV strains was utilized, with each strain representing a different potential mechanism for exposure of humans to Vesivirus: SMSV-5 detected previously in a human case, SMSV-13 known to cause disease in two livestock species, and SMSV-17 recovered from edible shellfish. The separate phylogenetic cluster within the Vesivirus genus defined by FCV was represented by a live vaccine strain widely administered to cats. The noted differences among serum groups in patterns of estimated prevalence remained across a broad range of potential assay cutpoints. It would be unexpected for humans to have Vesivirus viremia, but not exhibit serologic evidence of infection.

Viremia was expected to be part of the natural history of Vesivirus infections in humans because one of two cases described had a disseminated vesicular exanthem of the hands and feet from which Vesivirus was cultivated [Smith et al., 1998a] . The Vesivirus strains causing Vesivirus viremia in this study were closest by genome sequence comparison to SMSV, marine Vesivirus. SMSV Vesivirus are widely distributed in marine and some terrestrial animals and may routinely ''traffic'' among these hosts, cycling from their large marine reservoirs onto land and perhaps back again. An ocean presence has been established by the isolation of virus, the presence of specific neutralizing antibodies, or by genome amplification and sequencing for 43 of the 46 known serotypes of the genus Vesivirus, including serotypes isolated initially from terrestrial hosts and named feline (FCV), primate (PCV Pan-1), bovine (BCV Bos-1), reptilian (RCV Cro-1), swine (VESV-A 48 -K 56 ), mink (MCV), and human (SMSV-5 Hom-1 and HuCV Hom-1) caliciviruses [Evermann et al., 1983; Smith et al., 1983 Smith et al., , 1998a Smith et al., ,b, 2002b Seal et al., 1995; Reid et al., 1999] . Another marine Vesivirus isolated from walrus (WCV) causes hepatitis in domestic animals [Smith, 2000; Ganova-Raeva et al., 2004] . Vesivirus caliciviruses are resistant to environmental degradation; stable in aquatic substrates; multiply to high titer, with an estimated 10 13 virions released into the ocean daily by a single California gray whale (Eschricitus gibbosus) [Smith et al., 1998b and, in the case of FCV, have a cosmopolitan distribution [Studdert, 1978; Smith et al., 1998b] . Such attributes indicate the potential for frequent contact between a diversity of Vesivirus biotypes and hosts.

Known examples of such interaction include mussels and oysters, aquatic filter feeders that concentrate particulates, including viruses, from the water column, preserve viral viability for 60 days or more and thereby can deliver large doses of viable Vesivirus to species ingesting contaminated shellfish [Smith, 2000; Burkhardt et al., 2002] . Another example is from the mid-Pacific (French Frigate Shoals), where Vesivirusinfected fingerlings of two fish genera (Aterinomorous spp. and Encrasicolina spp.) were eaten by white tern (Gygis alba) hatchlings that developed a Vesivirusassociated blistering disease of the feet Vesivirus strains represented in GenBank. Study strain sequences shared with these GenBank entries are aligned and placed next to the strains with which they had the highest sequence identity. GenBank sequences are grouped according to similarity to each other. Five of the study strains were similar to each other and closest to chimpanzee

Pan-1 and San Miguel sea lion serotype 6 strains. The other polymerase study strain also was closest to known

Vesivirus, but in the pattern of sequence homology similar to feline calicivirus.

gram of feces per day also migrate thousands of miles annually between the Sea of Cortez and the Arctic Ocean [Akers et al., 1974; Smith et al., 2004] . Vesivirus has been recovered at titers of 10 7 infectious virions per gram of spleen in naturally and experimentally infected opal-eye fish (Girella nigricans). These fish are resident along the Southern California coast [Smith et al., 1980a [Smith et al., ,b, 1981 and are a sports and commercial fish and a common food source for seals, some of which also have extensive migration cycle. Seals reproduce on land or ice, where Vesivirus-induced reproductive failure and death occur, where foraging scavengers can further redistribute the viruses into terrestrial ecosystems, and where exposure to Vesivirus likely occurred for one human case [Smith et al., 1998a] . The present findings indicate a broader potential for Vesivirus infection and, perhaps, illness in humans than previously recognized. The strains causing viremia and to which antibody was detected in this study are similar to the Vesivirus with an ocean reservoir. A finding of subclinical viremia and the detected highest seroprevalence in cases of clinical hepatitis associated with transfusion or dialysis suggest that blood exposure that may have led to hepatitis also could lead to higher exposure to Vesivirus. If Vesivirus causes hepatitis in humans, then an estimated rate for such causation can be derived from the study findings, as follows: the rate of 10 of 91 donors amplicon-positive in the serum who also had elevated serum liver transaminases (ALT) values, together with the rate of high ALT values occurring in about 1 in 1,000 blood donors, would correspond to a rate of 1 in 10,000 blood donors who might have active, subclinical Vesivirus hepatitis. The association of higher anti-Vesivirus antibody prevalence with clinical hepatitis of unknown etiology would require further study. The finding of Vesivirus viremia in otherwise normal blood donors indicates that blood exposure to caliciviruses of the genus Vesivirus could occur by multiple routes of exposure. The diversity of host species, mechanisms of exposure and tissue tropisms for Vesivirus with the findings of this study suggest additional Vesivirus disease manifestations might be found in humans with further investigation.

We thank Dr. Michael Houghton and Chiron Laboratories, Emeryville, California, for sharing industry-acquired sera with us. The authors thank Gayathri Devi for assistance in preparing the figures and Beth Chamblin for assistance in preparing the manuscript. 

Calicivirus antibodies in California gray whales (Eschrichtius robustus) and Steller sea lions (Eumetopias jupatus)

Antibodies to marine caliciviruses in the Pacific walrus (Odobenus rosmarus divergens Illiger)

Isolation of reptilian calicivirus Crotalus type 1 from feral pinnipeds

Phylogenetic analysis of the Caliciviruses

New marine calicivirus serotype infective for swine

Applied technique for increasing calicivirus detection in shellfish extracts

High diversity of unknown picorna-like viruses in the sea

Epi Info, Version 6: A word processing, database, and statistics program for epidemiology on microcomputers

Emerging diseases. Malaysian researchers trace Nipah virus outbreak to bats

SARS in China. China's missed chance

Ultrastructure of newly recognized caliciviruses of the dog and mink

New Calicivirus isolated from walrus

Taxonomy of the caliciviruses

Prevalence of vesivirus in a laboratory-based set of serum samples obtained from dairy and beef cattle

Genome sequence and characterization of a virus (HaRNAV) related to picorna-like viruses that infects the marine toxic bloom-forming alga Heterosigma akashiwo

Re-analysis of serologic data from the Australian study of human health risks of infection by rabbit hemorrhagic disease virus

Partial characterization of the genome of nine animal caliciviruses

Cloning and sequence determination of the feline calicivirus strain F9

Genetic relatedness of the caliciviruses: San Miguel sea lion and vesicular exanthema of swine viruses constitute a single genotype within the Caliciviridae

The capsid protein of vesicular exanthema of swine virus serotype A48: Relationship to the capsid protein of other animal caliciviruses

Detection of a non-cultivatable calicivirus from the white tern (Gygis alba rothschildi)

Development of a reverse transcription polymerase chain reaction procedure for the detection of marine caliciviruses with potential application for nucleotide sequencing

Complete nucleotide sequence and genomic organization of a primate calicivirus

Isolation of caliciviruses from skunks that are antigenically and genotypically related to San Miguel sea lion virus

Virus cycles in aquatic mammals, poikilotherms, and invertebrates

Caliciviruses of ocean origin: A review

Characterization of two new serotypes of San Miguel sea lion virus

Hazards of disease transfer from marine mammals to land mammals: Review and recent findings

Preliminary investigation of a possible lung worm (Parafilaroides decorus), fish (Girella nigricans), and marine mammal (Callorhinus ursinus) cycle for San Miguel sea lion virus type 5

Calicivirus pathogenic for swine: A new serotype isolated from opaleye Girella nigricans, an ocean fish

Calcivirus (SMSV-5) infection in experimentally inoculated Opaleye fish (Girella nigricans)

Isolation and partial characterization of a calicivirus from calves

In vitro isolation and characterization of a calicivirus causing a vesicular disease of the hands and feet

Calicivirus emergence from ocean reservoirs: Zoonotic and interspecies movements

Antisense treatment of caliciviridae: An emerging disease agent of animals and humans

Detection of vesicular exanthema of swine-like calicivirus in tissues from a naturally infected spontaneously aborted bovine fetus

Ice as a reservoir for pathogenic human viruses: Specifically, caliciviruses, influenza viruses, and enteroviruses

Avian influenza: A new pandemic threat?