JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1993, p. 2433-24380095-1137/93/092433-06$02.00/0Copyright © 1993, American Society for Microbiology

Analysis of Acquired Human Cytomegalovirus Infections byPolymerase Chain Reaction

JAMES F. BALE, JR.,12* MARSHA E. O'NEIL,1 SARAH S. FOWLER,1 AND JODY R. MURPH'Departments of Pediatrics1 and Neurology, 2 The University ofIowa

College ofMedicine, Iowa City, Iowa 52242

Received 12 March 1993/Returned for modification 3 May 1993/Accepted 7 June 1993

We used the polymerase chain reaction and primers corresponding to three regions of the humancytomegalovirus (HCMV) genome to study HCMVs isolated from 16 children attending a single day-care centerand the father of two children in the same center. When we analyzed isolates with primers for the pp65 andmajor immediate-early genes, we observed nearly uniform amplification yielding products of predicted sizes.By contrast, primers for the a sequence demonstrated variability among HCMV strains, supporting the use ofthese primers as an epidemiologic tool. Analysis of a-sequence products from two isolates demonstrated 50 to70% nucleotide homology with the a sequence of HCMV Towne strain DNA. We observed 95% nucleotidehomology for the two a-sequence products derived from the father-child pair. Analysis of day-care centerisolates indicated that two children excreted two distinct HCMV strains during the study interval.

Children in group day-care centers in the United States,the parents or caretakers of such children, and adults withmultiple sexual partners acquire human cytomegalovirus(HCMV) at rates that greatly exceed those of the generalpopulation (3). Rates of HCMV infection among day-carecenter workers, for example, range from 8 to 20% annually(4) versus 1 to 5% among persons lacking such exposures.Although the majority of these acquired infections do notproduce disease, HCMV can cause an infectious mononu-cleosis syndrome in previously healthy individuals and mayinduce serious, life-threatening disorders in persons withimpaired cell-mediated immunity (8).Recent reports (6, 7, 12) indicate that the polymerase

chain reaction (PCR) can be used to characterize the geneticvariability of the virus, suggesting that PCR has potential asa method to investigate HCMV transmission. We used PCRwith primers corresponding to three regions of the HCMVgenome to study HCMV isolates derived from one largegroup day-care center. Our results indicate that PCR with ana-sequence primer can be used to characterize the epidemi-ology of HCMV transmission and provide additional evi-dence that children in group day-care centers can be rein-fected with distinct HCMV strains.

MATERIALS AND METHODSHCMV isolates. The virus strains analyzed in this study

consisted of HCMVs isolated between May 1986 and March1991 from (i) 16 children in a single large day-care center (19total isolates distinct from those in prior reports [14]) and (ii)the father of two children in the above-mentioned center(one isolate). HCMV was isolated from this parent after theonset of hepatitis attributed to HCMV. The HCMV Townestrain was propagated separately as a positive control forPCR and restriction enzyme analysis.Primary HCMV isolations were achieved by incubating

urine or saliva on confluent monolayers of human foreskinfibroblasts (HFFs) grown in 24-well plates in 5% CO2 at37°C. Monolayers were fed with Eagle's minimum essentialmedium containing 10% fetal calf serum, 50 U of penicillin

* Corresponding author.

per ml, 50 ,ug of streptomycin per ml, and 50 ,ug of genta-micin per ml and observed weekly. All isolates exhibited thecharacteristic HCMV-induced cytopathic effect, and 15 iso-lates were further confirmed by immunofluorescent stainingwith an HCMV-specific monoclonal antibody (MicroTrak;Syva Company, Palo Alto, Calif.).PCR analysis. DNAs extracted from HCMV-infected HFF

monolayers were amplified by PCR with five separate sets ofHCMV primers (Table 1). Reaction mixtures for typical PCRamplifications consisted of 1.25 U of Taq polymerase, 200,uM each deoxynucleoside triphosphate, 50mM KCl, 10 mMTris-HCl (pH 8.4), 1.5 mM MgCl2, 100 ,ug of gelatin per ml,20 pM each primer, and approximately 1 ,g of target DNA ina final volume of 50 ,ul. Typical PCR parameters consisted ofan initial denaturation for 2 min at 94°C, which was followedby 25 to 40 cycles of denaturation for 30 s at 94°C, annealingfor 30 s at 55°C, extension for 1.5 min at 72°C, and then afinal extension for 10 min at 72°C.

Controls for each run included amplification of (i) purifiedviral DNA extracted from HFF monolayers infected with theTowne strain of HCMV (courtesy of Mark Stinski, Depart-ment of Microbiology, University of Iowa), (ii) DNA fromuninfected HFFs, and (iii) buffer with sterile, distilled H20replacing target DNA. In several experiments, DNAs werecoamplified with primers for beta-globin DNA (10) as aninternal PCR control (primer 1, 5'-GGTTGGCCAATCTACTCCCAGG-3'; primer 2, 5'-GCTCACTCAGTGTGGCAAAG-3'; product size, 536 bp). Appropriate precautions weretaken to prevent contamination.For initial size analysis, 8 ,u1 of each PCR product was

studied by electrophoresis in 1.8% agarose gels (SeaKem;FMC Corp., Rock Island, Ill.) run at 30 V for 4 h. Molecularsize markers (123-bp ladder; Bethesda Research Laborato-ries, Gaithersburg, Md.) were incorporated in the initial andfinal lanes of each run. Gels were recorded by Polaroidphotography, and bands were scored by two readers (J.F.B.and M.E.O.).The specificities of PCR products were determined by

Southern blot analysis with HCMV probes specific for thetarget region. Amplification products were separated on0.9% agarose gels run at 60 V for 2 h and were transferred toZetabind membranes (CUNO, Inc., Meriden, Conn.) by

2433

Vol. 31, No. 9

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 26

Dec

embe

r 20

21 b

y 12

1.16

9.98

.33.

2434 BALE ET AL.

TABLE 1. Primers for detection of HCMV genes

Size (bp)Primer Sequencea of pre-dicted

product

a sequencebA

Primer 1 TTC CCC GGG GAA TCA CAC AG 130-220Primer 2 TTT TTA GCG GGG GGG TGA AAProbe GCG CCOG CAC GTC GCT TTT AT

IE-scPrimer 1 CCA CCC GTG GTG CCA GCT CC 159Primer 2 CCC GOCT CCT CCT GAG CAC CCProbe CTG GTG TCA CCC CCA GAG TCC

CCT GTA CCC GCG ACT ATC CIE-m

Primer 1 TTG GOCC GAA GAA TCC CTC AA 404Primer 2 TCT GOCA AAC ATC CTC CCA TCProbe GAT TAA GGT TCG AGT GGA C

pp65Primer 1 TAG GTG ACC AGT ACOG TCA AG 223Primer 2 TCC AGC ATG ATG TGOC GAG ATC TProbe AAG CTC TTT ATG CAC GTC A

pp65bPrimer 1 AAA GAG CCC GAC GTC TAC TAC ACOG T 159Primer 2 CCA GGT ACA CCT TGA COGT ACT GGT CProbe GTT CTC CAT GGA GOCA AAC CAG CTC

GTG COGCa Oligonucleotides 5' to 3'.b From Zaia et al. (22). The probe for a-sequence products was derived

from the nucleotide sequence of isolate 311u.c From Shibata (15).

using a vacuum transfer apparatus (Red-Evac; Hoefer Sci-entific Instruments, San Francisco, Calif.). Membranes wereprehybridized according to the manufacturer's instructions.Bound DNAs were hybridized with the appropriate probesend-labeled with [_y-32P]ATP and were visualized by autora-diography.

Restriction enzyme analysis. Several virus strains werestudied by modifications of methods described previously(14). After three or four passages in HFFs, viral genomic andcellular DNA was harvested by proteinase K digestion,chloroform-phenol extraction, and ammonium acetate-etha-nol precipitation. DNAs were resuspended in sterile H2O,and the concentration was adjusted to 1 ,ug/,lI by the additionof H20. DNAs were stored in H20 at 4°C until used.For restriction enzyme analysis, 1 to 3 ,ug of DNA was

incubated with EcoRI or HindIlI overnight at 37°C in theappropriate buffer. Digested DNAs were separated by elec-trophoresis on 0.8% agarose gels for 2 h at 100 mA andtransferred to Zetabind membranes (CUNO, Inc.), and themembranes were prehybridized for 3 h by procedures de-scribed by the manufacturer. Bound DNA was then hybrid-ized overnight with an HCMV probe, theXbaI D fragment ofHCMV Towne (courtesy of Mark Stinski) labeled with[32P]dCTP by random priming, and visualized by autoradiog-raphy. This probe spans part of the long repeat, all of theshort repeat, and part of the short unique regions of theHCMV genome (20). Our methods correspond to the junc-tional hybridization approach described by Spector et al.(18).DNA sequencing. Selected PCR products were sequenced

by the University of Iowa DNA Facility by using fluores-cence automated DNA sequencing and an Applied Biosys-tems (Foster City, Calif.) model 373A automated DNA

bp

492-

246-123-

-536

-223

ml 3 5 7 9 11 13m

FIG. 1. Amplification of selected HCMV isolates with pp65primers. Shown are an ethidium bromide-stained gel (top) and acorresponding autoradiogram with a 32P-labeled probe (bottom).Lanes: 1, no DNA (control); 2, HFF DNA; 3, HCMV Towne DNA;4, 368s; 5, 310u; 6, 365s; 7, 380u; 8, 356u; 9, 356s; 10 and 11, 369s;12, 311s; 13, 311u. m, molecular size marker lanes. The 536-bpproduct in this and Fig. 2 to 4 corresponds to the amplificationproduct of the beta-globin gene primers.

sequencer according to instructions provided by the manu-facturer. The volumes of PCR reagents were doubled toprovide 50 to 100 ng of PCR product in a final volume of 100,ul. Primers were provided at a concentration of 0.8 pm/ul.Hardcopy sequence data were reviewed and compared withrespect to HCMV strains.

RESULTSPCR analysis. Amplifications with one or both pairs of

primers for the pp65 gene yielded predicted products for allbut one HCMV isolate (344u). Amplifications with the IE-sprimer, corresponding to exon 4 of the major immediate-early gene, resulted in uniform products for all but fourisolates (329u, 344u, 356u, and 361u). However, a productwas observed when 329u and 361u were amplified with theIE-m primer (corresponding to exons 3 and 4 of the majorimmediate-early gene), suggesting that these isolates con-tained nucleotide variations in the region of the IE-s primer.The nucleotide sequences of the primers used in theseexperiments are indicated in Table 1, and representativeamplification products are shown in Fig. 1 and 2.

In contrast to the above-described results, considerablepolymorphism was observed among HCMV isolates (Fig. 3)when analyzed with the a-sequence primer described byZaia et al. (22). The isolates yielded a-sequence products ofthree sizes: approximately 220 bp (group A), approximately180 bp (group B), and approximately 130 bp (group C). Theproduct derived by amplification of purified HCMV TowneDNA was intermediate in size between group B and Cproducts.By using this schema, seven day-care isolates could be

classified as group A, three could be classified as group B,and three could be classified as group C. Two additionalisolates, shown in Fig. 3A, lane 11, and Fig. 3B, lane 5, hadfaint bands that appeared to categorize them as group Astrains. Four isolates, shown in Fig. 3A, lanes 9 and 12, andFig. 3B, lanes 9 and 13, had no a-sequence products. One ofthese isolates (364s) had a faint product on a duplicate run(Fig. 3B, lane 5), suggesting that it was a group A strain,whereas a group C product was observed in one run for 369s.

Beta-globin coamplification products were identified in the

J. CLIN. MICROBIOL.

A'000 :-rA6.-ll-A"-',Z. -.-%.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 26

Dec

embe

r 20

21 b

y 12

1.16

9.98

.33.

PCR ANALYSIS OF ACQUIRED HCMV INFECTIONS 2435

Abp

Z5

m 1 2 3 4 17 l9 1 1112S114 mm l 3 5 7 9 11 13 m

FIG. 2. Amplification of selected HCMV isolates with IE-mprimers. Shown are an ethidium bromide-stained gel of PCR prod-ucts (top) and an autoradiogram with 32P-labeled probe for the IE-mamplification product (bottom). Lanes: 1, no DNA (control); 2, HFFDNA; 3, HCMV Towne DNA; 4, 319u; 5, 318u; 6, 375u; 7, 312u; 8,364s; 9, 364s; 10, 378u; 11, 361u; 12 and 13, 329u; 14, 368s. m,molecular size marker lanes.

runs for three of the negative isolates (361u, 364s [Fig. 3B,lane 9], and 369s), and elimination of the beta-globin primersgave similar results (data not shown). The three isolates haddetectable products when amplified with an IE or pp65primer. Thus, we believe that the results for these strainsindicated nucleotide variations in the a sequence sufficient toinhibit PCR amplification.When the groups of isolates were compared according to

the dates of sample acquisition, it was found that group Aisolates were recovered between November 1988 and March1991, group B isolates were recovered between April 1988and November 1988, and group C isolates were recoveredbetween May 1986 and March 1988 (Table 2). The twoisolates that had faint bands suggesting that they were groupA strains were isolated in October and December 1987.Isolates without a-sequence products were recovered be-tween December 1987 and June 1989. Thus, categorizationby a-sequence product size generally correlated with tem-poral clusterings of isolates, a feature compatible with hor-izontal transmission ofHCMV strains in the day-care center.Three sets of paired isolates obtained 12 or more months

apart from each of three children were among the isolatesanalyzed in this study (Fig. 3 and 4). Two children (311 and356) appeared to excrete two distinct strains during the studyinterval. Child 356 appeared to excrete a group A strain insaliva in 1987 (Fig. 3A, lane 11) and a group B strain in urinein 1988 (Fig. 3B, lane 11). Child 311 excreted a group C strain(Fig. 3A, lane 7, and Fig. 4, lane 2) in saliva in 1986 and agroup A strain (Fig. 3A, lane 14, and Fig. 4, lane 1) in urinein 1989. A group A strain isolated in 1989 was also excretedin urine by the father of child 311.Comparison of PCR and restriction enzyme analysis. We

next compared the restriction endonuclease fragment pat-terns of several group A isolates. These results, shown inFig. 5, suggested that at least two strains (isolates 368s and378u) were closely related. The pattern shown by theseisolates was similar to the pattern of the group A strainsexcreted by child 311 and his father. One isolate (356s) hada distinct restriction enzyme pattern differing from that ofother group A strains by at least four bands in the combineddigests.

Bbp

492369246

123

ml 2 3 4 5 6 7 8 9 1011 1213M

FIG. 3. Amplification of HCMV isolates with a-sequence prim-'ers. (A) Lanes: 1, no DNA (control); 2, HFF DNA; 3, HCMVTowne DNA; 4, 329u; 5, 368s; 6, 310u; 7, 311s; 8, 365s; 9, 344u; 10,380u; 11, 356s; 12, 369s; 13, isolate from a congenitally infectednewborn; 14, 311u. m, molecular size marker lanes. (B) Lanes: 1, noDNA (control); 2, HFF DNA; 3, HCMV Towne DNA; 4, 319u; 5,364s; 6, 318u; 7, 375u; 8, 312u; 9, 364s; 10, 378; 11, 356u; 12, 368s;13, 361u. m, molecular size marker lanes. A, B, and C in the rightmargins correspond to the sizes of a-sequence products, as de-scribed in Results.

The a-sequence products from several group A isolateswere further characterized by digestion with restrictionenzymes BssHII and Mnll (data not shown). We observedidentical digestion patterns for isolates 368s and 378u (cut byBssHII but not by MnlI), supporting the conclusion thatthese strains were closely related. The a-sequence amplifi-cation product of isolate 365s had a similar digestion pattern,whereas 369s did not.

Sequence analysis. We performed sequence analysis of twoa-sequence products. This analysis demonstrated approxi-mately 95% nucleotide sequence homology between isolatesfrom child 311 and his father (Fig. 6A), group A strainsknown to have similar restriction enzyme profiles. Com-pared with the published nucleotide sequence for the HCMVTowne DNA a sequence (11), the clinical isolates exhibited50 to 70% homology. The isolates demonstrated approxi-mately 90% homology in the a-sequence pac-2 motif, con-

serving the six-nucleotide sequence (5'-TlTlTAT-3') (Fig.6B).

DISCUSSIONOur results indicate that a-sequence primers have utility in

investigations of HCMV transmission. We observed that

bp

492-369-246-

123-

-536-404

492369246

123'Wz.0, .IW-

VOL. 31, 1993

.0-B

.0-C

-A-o- B.0-C

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 26

Dec

embe

r 20

21 b

y 12

1.16

9.98

.33.

2436 BALE ET AL.

TABLE 2. Summary of HCMV isolates and PCR results

Isolate' Date (mo/day/yr) of PRgoPsample collection PCR group"310u 11/16/88 A311u 6/14/89 A356s 10/02/87 A (faint)364s 12/08/87 A (faint)365s 11/16/88 A368s 11/16/88 A375u 2/20/91 A378u 2/22/91 A380u 1/04/91 A318u 4/21/88 B329u 5/19/88, 6/15/89 BC356u 11/07/88 B311s 5/13/86 C312u 5/20/86 C319u 3/30/88 C344u 6/14/89 NPd361u 1/08/88 NP364s 12/08/87 NP369s 11/16/88 NPIsolate from father 4/07/89 A

of child 311

a Numbers correspond to designations for children in the day-care center.s, saliva; u, urine.

b Determined according to the product size (in base pairs) when the isolatewas amplified with a-sequence primers. A, approximately 220 bp; B, approx-imately 180 bp; C, approximately 130 bp.

c On initial size analysis bya sequence. On a subsequent run (Fig. 4), a faintgroup A band was seen in the autoradiograph.

d NP, no product. In one run, a faint group A product was seen for 364s anda faint group C product was observed for 369s.

HCMV isolates can be grouped according to the sizes of thea-sequence products and that strains with similar-size a-se-quence products tended to cluster temporally in the day-carecenter, suggesting that these strains were horizontally trans-mitted. Transmission of certain group A strains was further

bp,:

56...5369246

123-ml1 2 34 56 7 8910m

B 0 : -

1 23 678910

FIG. 4. Amplification of HCMV isolates with a-sequence prim-ers. (A) Ethidium bromide-stained gel of PCR products. (B) Auto-radiogram with a 32P-labeled probe for a-sequence products. Lanes:1, 311u; 2, 311s; 3, 356s; 4, 310u; 5, 329u; 6, HCMV Towne DNA;7 and 8, blank; 9, 11FF DNA; 10, no DNA (control). m, molecularsize marker lanes.

A

1 2 3 4 5 6 7 8

B

Kb-23.1

-9.4

-6.6

-4.4

-2.3-2.0

Kb-23.1

-9.4

-6.6

-4.4

1 2 3 4 5 6 7 8 9

FIG. 5. Restriction enzyme analysis of selected group A strains.(A) Autoradiogram of EcoRI digests. Lanes: 1, 356s; 2, 365s; 3,378u; 4, 368s; 5, 364s; 6, 369s; 7, HCMV Towne DNA; 8, HFFDNA. (B) Autoradiograph of HindIII digests. Lanes: 1, 312u; 2,356s; 3, 365s; 4, 378u; 5, 368s; 6, 364s; 7, 369s; 8, HCMV TowneDNA; 9, HFF DNA.

supported by detection of similar restriction fragment pat-terns. Horizontal transmission ofHCMVs among children ingroup day-care centers has been documented in prior UnitedStates studies (1).However, a-sequence amplification alone may not com-

pletely distinguish HCMV strains. One isolate (356s), cate-gorized as a group A strain, had a restriction fragmentpattern substantially different from that of other group Astrains, an observation that indicates that additional strate-gies, such as restriction enzyme digestion of a-sequenceproducts (22) or amplification with other PCR primers (16),may be necessary to establish strain relatedness. We alsoobserved that a-sequence amplification of some isolates didnot consistently yield products, and we suspect that suchisolates possess nucleotide variations within the a sequence.In vitro selection or alteration of the a sequence duringpassage could potentially affect strain characterization. Ob-servations regarding the Towne strain indicate that the asequence remains stable through repeated passage (13),whereas the AD169 strain may be unstable (21).

J. CLIN. MICROBIOL.

IRWI"

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 26

Dec

embe

r 20

21 b

y 12

1.16

9.98

.33.

PCR ANALYSIS OF ACQUIRED HCMV INFECTIONS 2437

A10 20 30 40 50

311U aacagacACGCGTCTTCgxTTTCGCCGTGCGCGCCGCACGtCGCTTTTA

FATHER ttccttcaacagaaACGCGTCTTCgcTTTCGCCGTGtGCGCCGaACGtCGCTTTTA

TOWNE ACGCGTCTTC-TTTTCGCCGTGCGCGCCGCACGCCGCTTTTA

60 70 80 90 100 110

311U TtCGCCGtCGCCGgctTCCgCg-ACCGCcgtCcCtAcCGCaccACaCCgcaccgcAGc

FATHER TtCGCCGtCGCCGgctTCCgCg-ACCGCcgtCcCtAcCGCatcACaCCgcaccgcAGcTOWNE TGCGCCGCCGCCG---TCC-C-AACCGCAC-CGC-AACGCG--ACTCC------AAG-

120 130 140 150 160

311U ACaCgCAA-ctattcgccgtcgccgtccacacacgcaactccaat

FATHER ACaCgCAA-ctattcgccgtctccgtacacacacgcaactccatt

TOWNE ACTC-CAAA

BISOLATESANTT)Mr-'F

311U

FATHER

CGCCGCACGtCGCTTTTATtCGCCGtCGCC

CGCCGaACGtCGCTTTTATtCGCCGtCGCC

TOWNE CGCCGCACGCCGCTTTTATGCGCCGCCGCCFIG. 6. Sequence data from two HCMV isolates. (A) Comparison of a-sequence products from HCMVs isolated from child 311 and his

father. (B) Comparison of pac-2 motifs. Towne sequence data were derived from Kemble and Mocarski (11). Lowercase letters indicatedeviations from the HCMV Towne sequence. Nucleotide sequences are 5' to 3'. x, unknown.

Our results indicate that certain gene regions are con-served among clinical HCMV isolates. Primers forpp65, agene that encodes a tegument protein, and for the majorimmediate-early gene (IE-s primer) amplified most of theseHCMV strains. This result is similar to that of Zaia andcolleagues, who studied HCMV isolates from several clinicalsettings (22). However, recent data suggest that nucleotidevariations occur within exon 4 of the major immediate-earlygene, indicating that primers such as IE-s may not uniformlyamplify all HCMV isolates (6).Our data regarding the group A strains support the con-

cept that certain HCMV strains may predominate in day-care environments. In a prospective study of 104 children inday-care centers, Adler detected 14 different strains among55 HCMV-excreting children (1). However, three strainsinfected 44 of the children and were apparently transmittedwithin the day-care environment. Several factors couldaccount for this observation. Excretion of similar HCMVstrains could reflect person-to-person transmission duringclose contact (i.e., several children grouped together in oneroom) or transmission via fomites or caretakers (9). In thepresent study, group A strains were excreted by severaldifferent children in the day-care center over a period of atleast 2 years.The predominance of certain HCMV strains within a

day-care center could also reflect the intrinsic characteristicsof viral strains. The PCR data suggest that there are genomicdifferences in the a-sequence region among groups ofHCMV strains. Although analysis with restriction enzymes

has not identified biological differences among strains (19),variations in the HCMV genome may be important. Certainstrains could differ in their replicative efficiencies, a featurethat may influence viral titers in urine or saliva. Much of thepac-2 motif was conserved in the two isolates studied here,but the significance of the nucleotide variations within theremainder of the amplified a-sequence products is uncertain.Our results indicate that children in group day-care cen-

ters can be reinfected with newHCMV strains. Two childrenin the present study excreted two strains with distincta-sequence groupings. One child (311) first excreted HCMVin saliva in 1986. He subsequently had several negativesaliva cultures and one positive urine culture in 1987. He wasfound to excrete HCMV in urine in 1989, coincident withexcretion of HCMV by his younger brother and his father,who had viral hepatitis. We do not know whether child 311acquired the new strain from his sibling, other children in thecenter, or, conceivably, his father, but we are intrigued bythe observation that his second isolate was a group A strain,the predominant HCMV pattern.

Several studies of diverse human populations suggest thatreinfection with HCMV occurs. Chandler and colleaguesidentified women who either shed multiple strains ofHCMVsimultaneously from different sites or had different HCMVstrains detected during serial sampling (5). A similar phe-nomenon has been observed with homosexual men withAIDS (17). Our data and the results of Adler (2) regardingHCMV reinfection of children in day-care centers have

VOL. 31, 1993

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 26

Dec

embe

r 20

21 b

y 12

1.16

9.98

.33.

2438 BALE ET AL.

substantial relevance to the biology of HCMV infections inhosts without high-risk sexual behaviors or AIDS.

ACKNOWLEDGMENTS

This work was supported by NIH grants HD22136 and HD19937and by grants from the Deafness Research Foundation and TheChildren's Miracle Network Telethon.We thank Mark Stinski for his advice and review of the manu-

script and Inara Souza for assistance with the restriction enzymedigestion of PCR products.

REFERENCES1. Adler, S. P. 1988. Molecular epidemiology of cytomegalovirus:

viral transmission among children attending a day care center,their parents, and caretakers. J. Pediatr. 112:366-372.

2. Adler, S. P. 1991. Molecular epidemiology of cytomegalovirus:a study of factors affecting transmission among children at threeday care centers. Pediatr. Infect. Dis. J. 10:584-590.

3. Alford, C. A., and W. J. Britt. 1990. Cytomegalovirus, p.1981-2010. In B. N. Fields, D. M. Knipe, et al. (ed.), Virology,2nd ed. Raven Press, Ltd., New York.

4. Bale, J. F., and J. R. Murph. 1991. The risk of cytomegalovirusinfection in hospital personnel and child care providers: areview. Pediatr. Rev. Commun. 4:233-246.

5. Chandler, S. H., H. H. Handsfield, and J. K. McDougall. 1987.Isolation of multiple strains of cytomegalovirus from womenattending a clinic for sexually transmitted diseases. J. Infect.Dis. 155:655-660.

6. Chou, S. 1992. Effect of interstrain variation on diagnostic DNAamplification of the cytomegalovirus major immediate-earlygene region. J. Clin. Microbiol. 30:2307-2310.

7. Chou, S., and K. M. Dennison. 1991. Analysis of interstrainvariation in cytomegalovirus glycoprotein B sequences encod-ing neutralization-related epitopes. J. Infect. Dis. 163:1229-1234.

8. Ho, M. 1991. Cytomegalovirus: biology and infection, 2nd ed.,p. 229-300. Plenum Medical Book Co., New York.

9. Hutto, C., E. A. Little, R. Ricks, J. D. Lee, and R. F. Pass. 1986.Isolation of cytomegalovirus from toys and hands in a day carecenter. J. Infect. Dis. 154:527-530.

10. Innis, M. A., and D. H. Gelfand. 1990. Optimization of PCRs, p.16. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White(ed.), PCR protocols: a guide to methods and applications.

Academic Press, Inc., San Diego, Calif.11. Kemble, G. W., and E. S. Mocarski. 1989. A host cell protein

binds to a highly conserved sequence element (pac-2) within thecytomegalovirus a sequence. J. Virol. 63:4715-4728.

12. Lehner, R., T. Stamminger, and M. Mach. 1991. Comparativesequence analysis of human cytomegalovirus strains. J. Clin.Microbiol. 29:2494-2502.

13. Mocarski, E. S., A. C. Liu, and R. R. Spaete. 1987. Structureand variability of the a sequence in the genome of the humancytomegalovirus (Towne strain). J. Gen. Virol. 68:2223-2230.

14. Murph, J. R., J. F. Bale, J. C. Murray, M. F. Stinski, and S.Perlman. 1986. Cytomegalovirus transmission in a midwest daycare center: possible relationship to child care practices. J.Pediatr. 109:35-39.

15. Shibata, D. 1990. Detection of human cytomegalovirus, p. 369.In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White(ed.), PCR protocols: a guide to methods and applications.Academic Press, Inc., San Diego, Calif.

16. Sokol, D. M., G. J. Demmler, and G. J. Buffone. 1992. Rapidepidemiologic analysis of cytomegalovirus by using polymerasechain reaction amplification of the L-S junction region. J. Clin.Microbiol. 30:839-844.

17. Spector, S. A., K. K. Hirata, and T. R. Neuman. 1984. Identifi-cation of multiple cytomegalovirus strains in homosexual menwith acquired immunodeficiency syndrome. J. Infect. Dis. 150:953-956.

18. Spector, S. A., T. R. Neuman, and K. K. Hirata. 1985. Rapiddetermination of molecular relatedness of isolates of humancytomegalovirus. J. Infect. Dis. 152:755-759.

19. Stinski, M. F. 1990. Cytomegalovirus and its replication, p.1961. In B. N. Fields, D. M. Knipe, et al. (ed.), Virology, 2nded. Raven Press, Ltd., New York.

20. Stinski, M. F. (University of Iowa). 1992. Personal communica-tion.

21. Tamashiro, J. C., and D. H. Spector. 1986. Terminal structureand heterogeneity in human cytomegalovirus strain AD169. J.Virol. 59:591-604.

22. Zaia, J. A., G. Gallez-Hawkins, M. A. Churchill, A. Morton-Blackshere, H. Pande, S. P. Adler, G. M. Schmidt, and S. J.Forman. 1990. Comparative analysis of human cytomegalovirusa-sequence in multiple clinical isolates by using polymerasechain reaction and restriction fragment length polymorphismassays. J. Clin. Microbiol. 28:2602-2607.

J. CLIN. MICROBIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

cm o

n 26

Dec

embe

r 20

21 b

y 12

1.16

9.98

.33.