Journal of Medical Virology 41:179-184 (1993)

Efficacy of a Recombinant Norwalk Virus Protein Enzyme Immunoassay for the Diagnosis of Infections With Norwalk Virus and Other Human “Candidate” C aliciviruses

Simon Parker, David Cubitt, Jason Xi Jiang, and Mary Estes Department Virology, Institute of Child Health, London, UK (S.P., D.C.) and Division of Molecular Virology, Baylor College of Medicine, Texas Medical Center, Houston, Texas (J.X.J., M.E.)

A recombinant Norwalk virus antigen enzyme immunoassay (rNV EIA) for the detection of Nor- walk virus (NV) antibodies was evaluated using sera from volunteers and patients infected with Norwalk, Hawaii, Snow Mountain Agent, small round structured viruses (SRSV), and human calicivirus (HuCV) strains. The results for NV an- tibodies correlated well with previous results ob- tained using a NV radioimmunoassay. The test rNV EIA was found to be highly specific for NV. The results of tests on sera obtained from pa- tients involved in confirmed outbreaks of SRSV in the UK showed that only 1/12 was due to Nor- walk virus. Further evidence is presented that SRSV strain UK2 is antigenically related to NV. 0 1993 Wiley-Liss, Inc.

KEY WORDS: recombinant Norwalk virus, EIA, RIA, calicivirus, small round structured virus

~~~ ~

INTRODUCTION Since the first description of Norwalk virus (NV) by

Kapikian et al. [1972] Norwalk and morphologically similar viruses (small round structured viruses, SRSV) have been increasingly recognised as a major cause of community outbreaks of food and waterborne infection in the United Kingdom: Public Health Laboratory Ser- vice [19881, Cubitt [19911; the United States, Kaplan et al. [1982], Kapikian and Chanock [1989]; and Japan: Sekine et al. L1989J. A knowledge of the true extent of SRSV infections worldwide has been hampered by a lack of antigen to produce widely available serological assays and the virtual dependence on electronmicros- copy for the detection of virus particles. The success of Jiang et al. [19901 in cloning and characterising the ge- nome of NV and the subsequent expression of recombi- nant norwalk virus protein (rNV) in baculovirus provides the potential for producing an enzyme immunoassay (EIA) [Jiang et al., 19921 for large seroepidemiological 0 1993 WILEY-LISS, INC.

studies and diagnosis of infections. The genomic orga- nization, biophysical, and biochemical characteristics of NV have led recently to it being classified as a mem- ber of the Caliciviridae by the International Committee for Taxonomy of Viruses [Cubitt et al., 19931. At present, none of the other human “candidate” calicivi- ruses have been sequenced and the true extent of their antigenic relationships has yet to be established.

This paper presents studies designed to determine the specificity and the sensitivity of the rNV EIA using well-characterised sera from patients involved in out- breaks of Norwalk (NV), small round structured vi- ruses (SRSV), or human calicivirus (HuCV) infection and paired sera from experimentally infected adult vol- unteers and to establish whether known outbreaks of SRSV in the United Kingdom were due to NV.

MATERIALS Sera

Norwalk virus. Acute and convalescent phase Norwalk sera from adult human volunteers were ob- tained from M. Estes, Baylor College of Medicine, N. Blacklow, University of Massachusetts, P. Madore, University of Rochester, and R. Glass, Centers for Dis- ease Control, Atlanta.

Acute and convalescent phase sera from patients in- volved in an epidemic of Norwalk infection associated with the consumption of oysters which occurred in Aus- tralia in 1978 [Murphy et al., 19791 were obtained from A.M. Murphy, Institute of clinical Pathology and Medi- cal Research, Sydney and R. Schnagl, La Trobe Univer- sity.

Data on the Norwalk antibody titres determined by immunoelectron microscopy (IEM), radioimmunoassay (RIA), or biotinylated enzyme immunoassays (BIA) in some of the sera were provided with the samples.

Accepted for publication October 5, 1992. Address reprint requests to David Cubitt, Department of Virol-

ogy, Institute of Child Health, 30 Guilford St., London WC 1N lEH, UK.

180 Parker et al.

TABLE Ia. Norwalk IgG Responses in Adult Volunteers Hawaii and Snow Mountain. Acute and conva- lescent phase sera from adult volunteers were obtained from P. Madore, University of Rochester.

Small round structured viruses, strains UKl , UK2, UK3. Acute and convalescent phase sera from patients involved in SRSV associated outbreaks in the United Kingdom were obtained from D. Lewis, Public Health Laboratory, Leeds [Lewis et al., 1990; Lewis, 19911.

Human caliciviruses. Sera from patients in- volved in the HuCV UK1 [Cubitt et al., 1979, 19801, UK2 [Cubitt et al., 19811, UK3 [Humphrey et al., 19841, and UK4 [Gill et al., 19831 outbreaks were obtained. Paired sera to a Japanese strain were obtained from I. Oishi, Osaka Prefectural Institute of Public Health.

Acute and convalescent phase sera were obtained from pa- tients involved in 18 outbreaks of gastroenteritis in the United Kingdom, Finland, United States, and South Africa in which SRSVs had been shown by electron microscopy to be the causative agent.

Untyped small round structured viruses.

METHOD rNV EIA

Ninety-six-well polyvinyl chloride, flat bottomed (Dynatech Immulon-2) microplates were coated with a 1:10,000 dilution of baculovirus expressed Norwalk protein (rNV) (5 mg/ml) in 0.01 M phosphate-buffered saline, pH 7.2 a t room temperature for 4 hr. Each well was washed once in 200 p1 of 0.05% Tween 20-PBS (PBST) using an automated wellwasher (Denley In- struments, Billinghurst, Sussex, UK). Wells were blocked by adding 200 pl of 0.1 M Tris/O.l% Tween 2013% bovine serum albumen (Fraction V) in PBS and storing a t 4°C overnight. Wells were washed twice in PBST and 100 pl of sera was added in a series of dou- bling dilutions in 1.0% skimmed milk (Marvel)/PBS from 1:1,000 to 1:64,000 in each column. The eighth well, which was uncoated, acted a s a control for nonspe- cific binding of each serum at a dilution of 1:1,000. Plates were incubated for 2 h r at 37°C in a humidified chamber and then washed three times in PBST. A 1:2,000 dilution of horseradish peroxidase-conjugated goat anti-human IgG (Dako Ltd) (100 p1) was added to each well and incubated for 2 h r a t 37°C. Plates were washed three times in PBST before adding 100 pl of substrate TMB. The reaction was stopped after 5 min by the addition of 50 pl of 2 N sulphuric acid. Optical densities were read a t A,,, in a spectrophotometer. ODs with a P-N value 20.2 and a P/N value 32.0 were regarded as positive. Sera were coded prior to test- ing and many sera were tested in both London and Houston.

RESULTS Tests on paired sera from 16 adult volunteers experi-

mentally infected with NV showed that 15 of them de- veloped significant antibody responses (Table Ia,b). The other volunteer had a high titre (264,000) of NV IgG in both acute and convalescent phase sera.

Source" BCM

BCM

Mass

Mass

Mass

Mass

Mass

Mass

Rochester

Rochester

Rochester

No

1 A C

2 A C

3 A C

4 A C

5 A C

6 A C

7 A C

8 A C

9 A C

10 A C

11 A C

Code

505-1 505-4 555-1 555-4 MC MC BE BE DO DO PE PE FR

(11) (11)

rNV EIA <1,000 64,000 2,000

128,000 (1,000

264,000 4,000

364,000 1,000

264,000 2,000

264,000 < 1,000

264,000 <1,000 32,000 <1,000 64,000 <1,000

8,000 4,000

256.000 "BCM, Baylor College of Medicine; Rochester, University of Roches- ter; Mass, University of Massachusetts.

TABLE Ib. Norwalk IgG Responses in Adult Volunteers Code

12

13

14

15

16

GVS 1731 A GVS 1731 C GVS 1973 A GVS 1973 C VGL 4349 A VGL 4349 C VGL 4361 A VGL 4361 C VGL 4367 A VGL 4367 C

Norwalk RIA 400

3,200 6,400

12,800 400

33,200 <loo

800 200

1.600

rNV EIA 8,000

364,000 264,000 364,000

4,000 364,000

2,000 264,000

4,000 364.000

The results on sera obtained from adults involved in four outbreaks of NV infection which occurred in Aus- tralia, the UK, and the United States are shown in Table 11. NV IgG was detected in the acute phase serum of 719 individuals and all of them showed a significant seroresponse (24-fold rise in antibody titre) following infection. The results of previous tests using RIA, [Cu- bitt et al., 19871 were in close agreement with those obtained using the rNV EIA, but titres were generally a t least 10-fold higher in the rNV EIA (Fig. 1).

The results of a one way cross to establish whether or not patients infected with Hawaii agent or Snow Moun- tain agent develop anamnestic responses to NV indi- cate that the rNV assay is specific for NV. There were no significant seroresponses in the adults infected with Hawaii or Snow Mountain agents.

Tests on paired sera from symptomatic patients in- volved in outbreaks in England showed that patients shown to be excreting SRSV UK1, UK3, and UK4 did not serorespond to NV. However, the person infected with SRSV UK2 seroconverted (Table 111).

Recombinant Norwalk Virus EIA, Calicivirus 181

TABLE 11. Norwalk IgG Titres in Patients Involved in Outbreaks of Norwalk Infection”

Source Case Code IEM RIA rNV EIA

Melbourne, Australia 1 A 9177178 2+ <200 8000 C 10577178 4+ ,3200 256000

2 A 9176178 1+ 200 200 128000 c 10584178 4+ 200

3 C MU 800 4000 4 C NO <loo 4000 5 C QU >6400 32000 6 C BR 3200 128000 7 C J O <200 <loo0

Manchester , U . K . 1 A 4721179 NT 4000 C 11023179 NT 64000

2 A 10425179 NT 4000 C 11024179 NT 32000

Source Case Code White Pine. USA 1 A 3934

c 3934 2 A 3935

c 3935 3 A 3998

C 3998 Philadelphia, USA 1 A 87-602

C 87-602 2 A 87-621

C 87-621

BIA rNV EIA <50

> 1600 <50

21600 100 800 100

26400 200

a6400

<loo0 216000 <loo0

216000 2000

216000 1000

a16000 8000

a16000 *RIA, Radioimmunoassay; BIA, biotinylated enzyme immunoassay.

Tests on sera from children known to be excreting HuCV strain UK1 or “Japan” were either negative (NV titres < 1,000) or showed no variation in titre between the acute and convalescent phase samples (data not presented in tables).

Antibody responses in individuals involved in out- breaks associated with infection with human calicivi- rus strain HuCV “UK2” are shown in Table IV. One patient showed a small (2-fold) rise in antibody levels but none of them developed a significant seroresponse (34-fold rise) to NV following infection. However, in a previous study, using a Norwalk RIA [Cubitt et al., 19871, 2/3 patients in the Harefield HuCV “UK2” out- break showed a significant seroresponse to NV. Only one of these pairs of sera were still available for testing [Norwalk RIA titre, acute (A) = 200; convalescent (C) = 800; rNV EIA titre, A = 4,000, C = 4,0001.

The results of tests on individuals involved in two oyster associated outbreaks caused by HuCV strain “UK4” are presented in Tables Va and b. Some of the data have been published previously [Cubitt et al., 19871. In the Colchester outbreak, 114 symptomatic pa- tients and 116 asymptomatic individuals showed signif- icant seroresponses to NV using the rNV EIA. How- ever, the one patient shown to be excreting HuCV UK4 did not serorespond to Norwalk. In the Tower Hamlets outbreak 618 of the symptomatic group showed signifi- cant seroresponses to Norwalk when tested by RIA and by the rNV EIA. In contrast, none of the asymptomatic group seroconverted.

Two elderly patients involved in an outbreak in a geriatric ward caused by HuCV UK4 did not serore- spond to NV.

Tests on paired sera from patients shown to be excret- ing SRSVs involved in 18 outbreaks (UK, 12; Finland, 1; United States, 3; South Africa, 2) showed that only one of the UK outbreaks, “Manchesterll979,” was due to NV (Table 11). The only other Norwalk outbreak identified occurred in South Africa in 1991.

DISCUSSION The rNV EIA was found to be a simple and effective

means for screening large numbers of sera for the pres- ence of Norwalk virus specific IgG. The assay showed good correlation with results previously obtained by RIA and was found to be a t least 10 times more sensi- tive. Tests using paired sera from individuals infected with other caliciviruses or SRSVs showed that the as- say was highly specific for NV. There was no evidence of significant anamnestic responses in any individual infected with HuCV strains UK1, UK2, Hawaii, Snow Mountain agent or SRSV strains UK1, UK3, or UK4. However, i t should be noted that the scarcity of refer- ence material enabled us to test sera from only a small number of individuals, The results of tests on the SRSV UK2 sera confirmed the previous findings of Lewis 119911 that this UK strain of SRSV is probably NV, but definitive proof of their identity will not be available until SRSV UK2 has been sequenced.

Cubitt et al. 119871 previously found that patients involved in the two oyster-associated outbreaks of HuCV UK4 infection developed significant serore- sponses to both HuCV UK4 and NV using immunoelec- tron microscopy and RIA, although HuCV UK4 failed to react in a radioimmunoassay for the detection of NV virus. The rNV antibody EIA confirmed our previous

182 Parker et al.

$J =5 1 2 4 8 16 32 E

Rlfl x 100

Fig. 1. Comparison of Norwalk virus antibody titres determined by rNV EIA and RIA. rNV EIA, recombinant Norwalk virus EIA; RIA, Cubitt e t al. [19871; N, number of samples tested.

TABLE 111. Norwalk IgG Responses in Patients Infected With Small Round Structured Viruses,

Strains UK1. 2. 3. and 4* Norwalk IgG titre

Strain Location Code rNV EIA SRSV UK1 Taunton, UK A <1,000

C <1,000 SRSV UK2 12238 A <1,000

13089 C SRSV UK3 Blake. UK 9887 A

11078 C SRSV UK3 Barnet, UK 1164186 A

1714186 C Barnet, UK 1163186 A

1717186 C Barnet, UK 1713186 C Barnet, UK 1715186 C Barnet,UK 1716186 C

SRSV UK4 Bo A

s;ooo 1,000 2,000

32,000 32,000 4,000 4,000

32,000 32,000 8,000

< L O O 0 Bo c <1;000

*A, acute phase serum; C, Convalescent phase serum [Lewis et al., 1990, 19911.

findings that sera from 6/8 symptomatic individuals from the Tower Hamlets outbreak developed signifi- cant antibody responses to NV. In view of the lack of detection of cross reactions with serum from patients from out,breaks associated with any other strain of

1

1

3

4

HuCV or SRSV it seems probable that the patients may have been coinfected with both agents. Support for this view comes from the following findings: the one asymp- tomatic individual who seroresponded to NV did not respond to HuCV UK4 (Table V); two elderly patients infected with HuCV UK4 during an outbreak of diar- rhoea and vomiting in a geriatric ward seroconverted to HuCV UK4 but not to NV (data not shown in tables). On the other hand the arguments for an anamnestic response are that the probability that oysters from two distinct geographical locations both being simulta- neously infected with HuCV UK4 and NV seems im- probable and stool samples from two patients who sero- responded to NV were shown to contain HuCV UK4 but were found to be negative when tested in an RIA for NV antigen. Therefore it is not possible to draw any firm conclusion as to whether HuCV UK4 and NV are anti- genically related or whether the patients were coin- fected.

Norwalk virus has been shown by RIA and EIA to be a major cause of community outbreaks of diarrhoea and vomiting in the United States [Kaplan et al., 19821. In contrast the results of the present study using the highly specific rNV EIA indicate that although SRSV outbreaks commonly occur in the United Kingdom, very few (1/12) were caused by NV. Additional support for this conclusion comes from the finding that when

Recombinant Norwalk Virus EIA, Calicivirus 183

TABLE IV. Antibody Responses to Human Caliciviruses and Norwalk Virus in Patients Infected With Human Calicivirus UK2

UKla UK2’ Japanb Norwalk‘ Norwalk Lab No. HuCV IEM HuCV IEM HuCV RIA RIA rNV EIA

Harefield 1979, HuCV UK2 24391Ad <10 <10 NA 24660C <10 1280 NA 24395Ad NT 20 400 24669C NT >640 400 24392Ad NT 40 400 24664C NT 640 400

1678Ad 40 <20 50 85436 40 160 <50 1681Ad <20 <20 400 1682C <20 320 400 8823C <20 320 400 1680Ad <20 <20 400 2720C <20 320 800 8824C <20 160 NA 1683Ad <20 20 200 1684C <20 160 200 1685A 40 <20 3,200 1686C 40 320 6,400

“Cubitt 119851. bCubitt et al. [19871. ‘Cubitt et al. 119811. ’Calicivirus particles detected by electronmicroscopy in patient’s stool. “A, not available; NT, not tested.

Portsmouth 1980, HuCV UK2‘

TABLE Va. Antibody Responses to Human Calicivirus and Norwalk Virus in Patients Involved in an Outbreak Due to

Human Calicivirus Strain UK4: Colchester HuCV UK4”

HuCV UK4” HuCV” Norwalk” rNV Lab No. IEM JananRIA RIA EIA

20879A 21721C 20878A 21720C 21102A’ 22171C 22155A 23141C

21101 1 22158 2 21005 1 22151 2 21103 1 23215 2 21639 1 22152 2 21096 1 23461 2 20998 1 22159 2 23910 2 23462 2

Symptomatic Group <20 200 <50

>320 400 400 20 200 200

320 NT NT <20 200 400 320 100 NT <20 50 800 320 50 800 Asymptomatic Individuals

20 NT 1600 20 NT 1600 20 NT 100 40 NT 50

160 NT 50 160 NT 400 20 NT NT 20 NT NT

<20 NT NT <20 NT NT <20 NT NT <20 NT NT <20 NT NT <20 NT NT

4,000 8,000 2,000

32,000 4,000 8,000

64,000 64,000

8,000 8,000 NT 2,000 4,000 8,000

16,000 32,000 2,000

64,000 64,000 64,000

2,000 8,000

“Details of outbreak in Cubitt 119851. aData published in Cubitt et al. (19871. bPatient shown to be excreting calicivirus.

400 1600

800 400 200 800

400 400

1600 800 800

50 50

100 400 400

<loo <loo

NA‘ NA 8,000 NA 4,000 4,000

4,000 4,000

16,000 8,000

16,000 1,000 2,000 2,000 2,000 4,000 2,000 2,000

TABLE Vb. Tower Hamlets, HuCV UK4, 1983

HuCV UK4 HuCV Japan Norwalk Norwalk Lab No. IEM RIA” RIA” protein

Symptomatic Group’ 1206AC 20 400 <50 2459C 80 400 <50 2458C‘ 80 400 400 1316A“ 40 200 400 1914C 640 NA NA 1204A 20 400 1,600 1908C 160 400 1,600 1506A 20 400 200 2198C 320 400 800 1504A 80 400 NT 2197C 640 800 800 1209A 40 400 1,600 2018C 2640 400 6,400 1213A 40 200 1,600 1915C 160 200 6,400 1262A 40 400 800 1920C 160 400 3,200 1503A 40 200 200 2196C 320 100 800

1499 1 20 200 (50 2193 2 20 200 <50 1265 1 40 <50 < 50 1927 2 40 <50 <50 1260 1 40 200 400 1930 2 40 200 200 1259 1 40 50 <50 1918 2 40 200 <50 1269 1 80 200 200 1925 2 160 200 200 1313 1 (20 200 <50 1913 2 20 100 100

“Cubitt et al. L19871. bGill et al. 119831. ‘HuCV UK4 detected.

Asymptomatic Individuals

<1,000 1,000 8,000 2,000

32,000 4,000 8,000 1,000 8,000 1,000 4,000 4,000

128,000 32,000

256,000 NA 2,000 4,000

16,000

1,000 <1,000

1,000 <1,000

8,000 8,000

<1,000 <1,000

8,000 16,000 2,000 4,000

184 Parker et al.

102 faecal samples from confirmed outbreaks and spo- radic of S R S V ~ and H ~ C V infection in ~ ~ ~ l ~ ~ d were tested in an rNV EIA for NV antigen none was found to be positive (Cubitt, Jiang, and Estes, in prepa- ration). These findings highlight the need for continued international collaboration to sequence other calicivi-

epidemiological studies to be carried out prior to developing strategies for produc- ing vaccines.

Cubitt WD, Blacklow NR, Herrmann JE, Now& NA, Nakata S, Chiba s (1987): Antigenic relationships between human caliciviruses and Norwalk virus. Journal of Infectious Diseases 156:806-813.

Cubitt WD, McSwiggan DA, Arstall S (1980): An outbreak of calicivi- rus in a mother and baby unit. Journal of Clinical Pathology 33:1095-1098.

Cubitt WD, McSwiggan DA, Moore W (1979): Winter vomiting disease caused by calicivirus. Journal of Clinical Pathology 32:386-393.

Cubitt WD, Pead PJ, Saeed AA (1981): A new serotype of calicivirus associated with an outbreak of gastroenteritis in a residential home for the elderly. Journal of Clinical Pathology 34:924-926.

Gill ON, Cubitt WD. McSwiaaan DA. Watnev BM. Bartlett CLR

and SRSVs to

(1983): Epidemic of gastroenteritis caused by “oysters contaminated with small round structured viruses. British Medical Journal

ACKNOWLEDGMENTS This work was sumorted bv a Wellcome Trust Trav- 287:1532-1534.

I I

elling Fellowship granted to“&-. D, Cub& and grant, Public Health Service Cooperative Research Agree- merit A130448, to Dr. M. Estes from the National Insti-

Humphrey TJ, Cruickshank JG, Cubitt WD (1984): An outbreak of calicivirus associated gastroenteritis in an elderly persons’ home. A possible zoonosis? Journal of Hygiene (Cambridge) 92:293-299.

Jiana X. Graham DY. Wane K. Estes MK (1990): Norwalk virus

Clinical Res:arch Center, supported by Grant 00350 from the Division of Research Resources of the NIH. We would like to thank Drs. J. Herrmann, N. Blacklow, P. Madore, R. Glass, 0. Meurmann, M. Taylor, W. Grabow, H. Appleton, A. Curry, D. Lewis, and U. Des- selburger, A. Murphy, R. Schnagl, and I. Oishi for pro- viding us with sera. Data shown in Tables Va and b were taken from Cubitt et al. 119871, Journal of Znfec- tious Diseases 156:806-813, and are reprinted with per- mission from the publisher, The University of Chicago Press, Copyright 0022-198918715605-0018.

REFERENCES Appleton H, Caul EO, Clewley JP , Cruickshank JG, Cubitt WD, Kurtz

JB, McSwiggan DA (1988): Foodborne viral gastroenteritis. PHLS Working Party Overview. Public Health Laboratory Service Di- gest 5:69-76.

Cubitt WD (1991): Waterborne viral gastroenteritis. A review. Water Science Technology 24:197-203.

Cubitt WD, Bradley D, Carter M, Chiba S, Estes M, SaifL, Schaffer F, Smith A, Studdert M, Thiel H J (1993): Caliciviridae, In “Classifi- cation and Nomenclature of Viruses.” Sixth Report of the Interna- tional Committee for Taxonomy of Viruses. Berlin: Springer-Ver- lag (in press).

Jiang X, Wang M, Graham DY, Estes MK (1992): Expression, self- assembly, and antigenicity of the Norwalk virus capsid protein. Journal of Virology 66:6527-6532.

Kapikian AZ, Wyatt RG, Dolin R, Thornhill TS, Kalica AR, Chanock RM (1972): Visualization by immune electronmicroscopy of a 27nm particle associated with acute infectious nonbacterial gas- troenteritis. Journal of Virology 10:1075-1081.

Kapikian AZ, Chanock RM (1989): Viral gastroenteritis. In Evans AS (ed): “Viral Infections of Humans.” New York: Plenum, pp 293- 340.

Kaplan JE , Gary W, Baron RC, Singh N, Schonberger LB, Fieldman R, Greenberg HB (1982): Epidemiology of Norwalk virus in out- breaks of acute non bacterial gastroenteritis. Annals of Internal Medicine 96:756-761.

Lewis DC (1990): Three serotypes of Norwalk-like virus demonstrated by solid-phase immune electronmicroscopy. Journal of Medical Vi- rology 30:77-81.

Lewis DC (1991): Norwalk and other small round structured viruses in the UK. Journal of Infection 23:220-222.

Murphy AM, Grohmann GS, Christopher PJ, Lopez WA, Davey G, Millsom R J (1979): An Australia-wide outbreak of gastroenteritis from oysters caused by Norwalk virus. Medical Journal of Austra- lia 2:329-333.

Sekine S, Okada S, Hayashi Y, Ando T, Terayama T, Yabuchi K, Miki T, Ohashi M (1989): Prevalence of small round structured virus infections in acute gastroenteritis outbreaks in Tokyo. Microbiol- ogy and Immunity 33:207-217.