You are on page 1of 7

Journal of Medical Virology 61:481487 (2000)

Human Cytomegalovirus Strains Associated With


Congenital and Perinatal Infections
Daniel E. Trincado,1,2 Gillian M. Scott,1 Peter A. White,1,2 Cheryl Hunt,3 Lucy Rasmussen,4 and
William D. Rawlinson1,2*
1

Virology Division, Department of Microbiology, South Eastern Area Laboratory Services, The Prince of Wales
Hospital, Randwick, Australia
2
School of Pathology, University of New South Wales, Kensington, Australia
3
School of Science, University of Western Sydney, Nepean, Kingswood, Australia
4
Stanford University School of Medicine, Center for AIDS Research and Division of Infectious Diseases and
Geographic Medicine, Stanford, California

The genotypes of human cytomegalovirus


(HCMV) isolates from pediatric patients differs
from those of infected adults in Australia. Genotypes were determined by PCR amplification of
glycoprotein B (gB) sequences, with subsequent
analysis by restriction fragment length polymorphism, single-stranded conformation polymorphism, heteroduplex mobility analysis and direct DNA sequencing. Restriction fragment
length polymorphism analysis of gB showed
genotypes gB1 (39%) and gB3 (30%) were more
prevalent in infected children and two new
genotypes (gB6 and gB7) were found. Singlestranded conformation polymorphism was used
to group isolates into 22 further subtypes and
suggested longitudinal co-infection or viral mutation was occurring over time. Heteroduplex
mobility analysis was found to be the most accurate and concise of the four methods used for
genotyping HCMV isolates. DNA sequencing
was used to confirm the results obtained from
heteroduplex mobility analysis, and identified
two isolates that were incorrectly genotyped by
restriction fragment length polymorphism
analysis. Heteroduplex mobility analysis efficiently genotyped all samples and allowed estimation of sequence variation between isolates.
These data suggest certain gB genotypes are associated more commonly with childhood infections, and these differ from strains associated
with invasive disease in HIV patients. J. Med.
Virol. 61:481487, 2000. 2000 Wiley-Liss, Inc.
KEY WORDS: glycoprotein B; polymerase
chain reaction; restriction fragment length polymorphism;
single-stranded conformation
polymorphism; heteroduplex
mobility analysis
2000 WILEY-LISS, INC.

INTRODUCTION
Human cytomegalovirus (HCMV) is responsible for a
wide range of diseases in newborn children, neonates
and immunocompromised patients. HCMV infection of
a child can occur in utero from transplacental transmission, during parturition, or perinatally [Alford,
1991]. Congenitally acquired HCMV causes illness in
510% of infected babies and can lead to a range of
disease up to the most severe form, cytomegalic inclusion disease [Britt and Alford, 1996]. Indeed, HCMV is
now the most common viral cause of congenital malformation in developed countries [Demmler, 1991].
Envelope glycoproteins of herpesviruses are involved
in the pathogenesis of infection, particularly in relation
to virus attachment, entry and egress of virion particles in host cells, and as targets for the host immune
response [Britt et al., 1990; Rasmussen et al., 1991].
Most genotyping techniques for HCMV are based upon
glycoprotein B (gB) sequence variation, and we studied
four different methods of typing gB-derived amplicons.
Using restriction fragment length polymorphism
analysis of amplicons from the variable region of gB
[Chou, 1990], HCMV isolates from children were identified and classified into one of four groups. Two new
genotypes (gB6 and gB7) based on restriction fragment
length polymorphism analysis classification, were

This work was performed at the Virology Division, Department


of Microbiology, South Eastern Area Laboratory Services, The
Prince of Wales Hospital, Randwick, Australia.
This work was presented at the 1997 Australian Society for
Microbiology Conference, September 28thOctober 3rd, 1997, Adelaide, Australia (abstract P11.6).
Grant sponsor: National Health and Medical Research Council;
Grant sponsor: University of Western Sydney-Seed Grant; Grant
sponsor: Sydney Childrens Hospital Foundation.
*Correspondence to: A/Prof. William D. Rawlinson, Virology
Division, Department of Microbiology, South Eastern Area Laboratory Services, The Prince of Wales Hospital, Randwick NSW
2031, Australia. E-mail: w.rawlinson@unsw.edu.au
Accepted 16 December 1999

482

Trincado et al.

identified and analyzed further using heteroduplex mobility analysis. Isolates from the four major groups
were also analyzed using single-stranded conformation
polymorphism, heteroduplex mobility analysis, and
DNA sequencing to detect further subgroups, and determine if a closer correlation between genotype and
disease could be ascertained. No single genotype was
found to be associated with congenital infection, although gB1 and gB3 genotypes were found more frequently in infected children than infected adults.
MATERIALS AND METHODS
Study Population
Clinical isolates were obtained from pediatric patients seen at the two childrens teaching hospitals in
Sydney, Australia. Urine samples, naso-pharyngeal aspirates (NPAs), and throat swabs were cultured in
MRC-5 cells and stored at 80C in minimum essential
medium containing 15% fetal calf serum and 10% dimethyl sulfoxide. Blood and urine specimens were obtained from patients with HIV/AIDS through routine
diagnostic testing. These individuals ranged in age
from 1868 years old and had invasive CMV disease at
time of diagnosis. Virus isolated from these individuals
was treated as above.
Sample Preparation
DNA was extracted using lysis buffer (NaCl 150 mM,
EDTA 1 mM, Tris Cl 15 mM and SDS 0.1%), and proteinase K (20 g/L). The solution was incubated at 55C
for 2 hr, the proteinase K denatured at 95C for 5 min,
the DNA ethanol precipitated, and resuspended in TE.
Glycoprotein B Genotyping by Restriction
Enzyme Digestion
gB genotypes were determined using previously described methods [Chou, 1990], involving a nested PCR
with 5 l of the outer product used as template in the
inner reaction. The gB PCR products were classified
into four existing (gB1, gB2, gB3 and gB4) and two new
(gB6 and gB7) genotypic groups, using restriction enzyme digestion with HinfI and RsaI [Chou, 1990].
Single-Stranded Conformation Polymorphism
Single-stranded conformation polymorphism was
performed by adding 10 l of the double-stranded PCR
product to 15 l of formamide buffer (formamide 95%,
bromophenol blue 0.05%, xylene cyanol 0.05% and
EDTA 20 mM), and 3 l of 3 M sodium hydroxide. The
DNA was denatured at 95C for 8 min then cooled rapidly on ice [Binder et al., 1995]. The denatured samples
were subjected to electrophoresis on a 10% polyacrylamide gel containing 0.5% polyethylene glycol at 300 V
for 7 hr at room temperature using a vertical water
jacketed electrophoresis tank [Glavac and Dean, 1993;
Markoff et al., 1997]. The DNA was visualized by silver
staining.

Heteroduplex Mobility Analysis


Heteroduplex mobility analysis was carried out by
mixing 5 l of each reference gB PCR product with 5 l
of a sample PCR product and 10 l of water. Reference
samples for each genotype (14) were chosen from the
most prevalent single-stranded conformation polymorphism subtypes within each gB genotypic group. The
DNA was denatured at 96C for 5 min then reannealed
at 60C for 1 hr. The resultant heteroduplexes and homoduplexes were subjected to electrophoresis on a 6%
polyacrylamide gel containing 0.5% polyethylene glycol
at 110 V for 75 min. DNA was stained with ethidium
bromide and visualized using ultraviolet light.
DNA Sequencing
PCR products were precipitated using a solution containing 27% polyethylene glycol 8000, 0.6 M sodium
acetate and 6.5 mM magnesium chloride, and sequenced on an ABI 377 automated DNA sequencer using Prism DyeDeoxy terminator chemistry (PerkinElmer).
DNA Sequence and Statistical Analysis
Pairwise and multiple sequence alignments of DNA
were carried out using the GCG programs GAP and
Clustal W respectively [Thompson et al., 1994]. Phylogenetic studies of the DNA sequences were carried out
using the programs DNADIST (Kimura two-parameter
method) and NEIGHBOR from the PHYLIP package
(version 3.5c) [Felsenstein, 1993]. Bootstrap resampling of 100 data sets obtained from the multiple sequence alignments was carried out using the programs
SEQBOOT, DNADIST and NEIGHBOR. The consensus tree was calculated using CONSENSE. Dendrograms were drawn using the RETREE program (with
midpoint rooting option) from the PHYLIP package
and the final graph produced with TREEVIEW (version 1.5) [Page, 1996]. Distribution differences were
calculated using the chi-square test when population
samples were greater than five, or Fishers exact test
when they were less than five.
RESULTS
Detection of New Viral Strains
Two new genotypes (gB6 and gB7), defined on the
basis of their restriction fragment length polymorphism pattern were identified. These were isolated
from two Australian pediatric patients infected perinatally for the first time. Only a single isolate of each
genotype was identified. Sequence analysis showed
these isolates to be variants of the gB3 and gB1 groups
both with 97% sequence homology respectively. Comparisons to a gB5 sequence published previously
[Shepp et al., 1998] showed 83% and 75% sequence
homology respectively, consistent with gB5, gB6 and
gB7 being different strains. Phylogenetic and bootstrap
analysis indicate that these isolates are variants of the
gB3 and gB1 groups but classified as new genotypes
based on restriction fragment length polymorphism
analysis.

HCMV gB Genotypes in Children

483

and produced identical results. Nine subtypes were


identified in the gB1 genotype group, seven in gB2, and
six in gB3. No single-stranded conformation polymorphism subtypes were found of genotypes gB4, gB6 and
gB7 (Fig. 2). The results were consistent with the gB
genotypes based on restriction fragment length polymorphism, that is, isolates of the same gB genotype
using restriction fragment length polymorphism produced similar single-stranded conformation polymorphism profiles (Fig. 2). The results also demonstrated
multiple HCMV isolates collected sequentially from the
one patient (identified as gB2 by restriction enzyme
digestion) were from different subtype groups, suggesting super-infection with different strains, or viral mutation over time (lanes 5 and 6).
Heteroduplex Mobility Analysis

Fig. 1. Prevalence of genotypes in a pediatric population with respect to congenital and perinatal infections.

Comparison of gB Genotype in Children With


Congenital and Perinatal Infections
Isolates were divided into congenital (virus isolated
within the first three weeks of life n 18) and perinatal (isolation of virus up to six months of age n 49)
infections. Isolates from congenitally infected children
were predominantly gB1, gB2 and gB3 and from perinatal infections were predominantly gB1 and gB3
(Fig. 1).
Comparison of the Distribution of gB
Genotypes in Children and HIV-AIDS Patients
in Australia
Children were infected most frequently with genotype gB1, although this did not differ significantly from
the adult rate of infection. A significant difference in
distribution was observed between infected pediatric
and HIV-AIDS patients for gB2 (20.9% vs. 33.9%, P
0.03), gB3 (29.9% vs. 12.9%, P 0.02) and gB4 (7.5%
vs. 24.2%, P 0.03) (Table I). Despite the differences
observed, the rate of infection was similar in pediatric
and adult patients.
Identification of Subtypes by Single-Stranded
Conformation Polymorphism
Based on single-stranded conformation polymorphism analysis of the gB-PCR product, 22 subtypes of
HCMV were identified from 66 patients. These experiments were confirmed by performing them in triplicate

Heteroduplex mobility analysis of PCR products


from isolates of the same genotype produced comigrating bands of 302bp. Heteroduplex mobility
analysis of genotypically heterologous products, however, revealed two or three bands; two co-migrating
homoduplexes and one (gB1/2) or two (all others) heteroduplexes with reduced mobility (Fig. 3). Two isolates (27EgB2 and 44EgB2), identified as genotype 2
strains by restriction fragment length polymorphism
analysis, were identified as genotypes gB3 and gB1 respectively by heteroduplex mobility analysis. Sequencing and subsequent phylogenetic analysis confirmed
the heteroduplex mobility analysis results indicating
the genotype was incorrectly assigned using restriction
fragment length polymorphism analysis (Fig. 4). Heteroduplex mobility analysis of the two new isolates
identified as gB6 and gB7 by restriction fragment
length polymorphism analysis produced a heteroduplex that could not be separated from the homoduplex
with the gB1 and gB3 reference samples respectively.
The migration of a heteroduplex correlated with the
genetic diversity between isolates as determined by sequencing, that is, the greater the diversity the greater
the reduction in mobility of the heteroduplexes (Fig. 3).
Heteroduplexes were separated from homoduplexes
when nucleotide sequences diverged by as little as 5%
(gB1 and gB2). Figure 3 shows pairwise heteroduplex
mobility analysis of the genotypes 14 as well as analysis of gB6 and gB7.
DNA Sequencing and Phylogenetic Analysis
Phylogenetic and bootstrap analysis of all sequenced
products indicated strong support for the existence of
three major groups, gB1/2, gB3 and gB4 (Fig. 4). The
gB sequence of amplicons from the heteroduplex mobility analysis reference samples were sequenced and
aligned to determine the extent of nucleotide variation
between them. Nucleotide variation was 5% (gB1/gB2),
18% (gB1/gB3), 6% (gB1/gB4), 18% (gB2/gB3), 10%
(gB2/gB4) and 20% (gB3/gB4). Those PCR products
exhibiting unique single-stranded conformation polymorphism patterns, or those with non corresponding
heteroduplex mobility analysis and restriction frag-

484

Trincado et al.
TABLE I. Distribution of gB Genotypes Using RFLP in Pediatric Patients and HIV
Infected Adults

Type of patient
Pediatric
HIV

No. genotypes/No. patients


67/66
62/53

gB1
(%)
38.8
25.8

gB2
(%)
20.9
33.9

gB3
(%)
29.9*
12.9

gB4
(%)
7.5
24.2

Other
(gB6 & gB7)
(%)
2.9
3.2

Statistical analysis was performed using the chi-square test or Fishers exact test (if n < 5). gB, glycoprotein B; HIV, human immunodeficiency virus.
*P-value 0.02.

P-value 0.03.

Fig. 2. Single-stranded conformation polymorphism analysis of gB


genotypes. Lane M contained molecular size markers (pGEM cut with
HinfI, RsaI and SinI). Lanes 14: gB1; lanes 58: gB2; lanes 912:
gB3; lane 13: gB4; and lane 14: mixed isolate gB1/4 (Demonstrated
gB1 and gB4 RFLP pattern). Additional bands detected in lanes 1, 6,
10, 11 and 14 (shown by arrows) indicate the presence of sequence
variation. The pattern of additional bands was used to classify each
genotype into further subtypes.

ment length polymorphism results were also sequenced (GenBank accession numbers AF210774 to
AF210791). Analysis of these products confirmed the
results obtained from heteroduplex mobility analysis
(Fig. 4).
DISCUSSION
Using restriction fragment length polymorphism
analysis, genotypes gB1 (38.8%) and gB3 (29.9%) were
the most prevalent in pediatric patients. No single genotype was highly associated with congenital infections. The majority of pediatric isolates studied were
from perinatal infections rather than congenital infections, consistent with studies showing 22% to 59% of
perinatally infected children have viruria whilst only

0.2% to 2.2% of all infants are infected in utero [Pass,


1991, 1985]. The distribution obtained for pediatric patients, gB1 (38.8%), gB2 (21%), gB3 (29.9%), gB4 (7%),
gB6 (1.5%) and gB7 (1.5%) is similar to results obtained in a previous study of 37 pediatric patients [gB1
(48%), gB2 (18%), gB3 (22%), and gB4 (3%)] [Bale et al.,
1996]. The remaining 9% of isolates in the previous
study were of mixed genotypes not commonly isolated
from children in this cohort. These results differ significantly from results obtained for HIV-infected patients where a predominance of gB2 occurred [Rasmussen et al., 1997]. There were significantly higher
numbers of genotype gB3 (P 0.02) and lower numbers of gB2 (P 0.03) and gB4 (P 0.03) from infected
children, than from patients with HIV-AIDS (Table I).
In contrast, the distribution of genotypes in transplant
recipients in the USA is similar to that of the HCMV
isolates in the pediatric population described here
[Fries et al, 1994; Rasmussen et al., 1997]. This similarity of distribution may be due to similar routes of
transmission, as isolates from transplant recipients often arise from reactivation of latent virus, that may
have been contracted during childhood [Demmler,
1991; Pass et al., 1982]. In comparison, HIV-AIDS patients in the USA have a different distribution of genotypes when compared to pediatric patients, with a
higher prevalence of gB2 isolates (51% of the total)
[Rasmussen et al., 1997]. The difference in the genotype distribution in the HIV-infected population may
be due to different routes of transmission or co-infection
with multiple isolates, where the co-infection may be
masked by the more dominant isolate. No evidence was
found for mixed infection with different strains of HCMV
on heteroduplex mobility analysis. If co-infecting strains
differed by less than 4%, however, they would not be detected using heteroduplex mobility analysis [White et al.,
1999], although they should have been detected using
single-stranded conformation polymorphism.
During the course of this study, two new genotypes
were identified using restriction fragment length polymorphism analysis, that were designated gB6 and gB7.
Only one of each genotype was isolated in children, and
one of each in adults (Table I). The gB6 strain exhibited
a HinfI restriction pattern similar to gB2 and gB3, but
with only one RsaI restriction site present. The gB7
isolate exhibited a HinfI restriction pattern similar to
gB1 and gB4 but with no RsaI site present. Compari-

HCMV gB Genotypes in Children

485

Fig. 3. Ethidium bromide stained 6% polyacrylamide gel showing


pair-wise heteroduplex mobility analysis of genotypes 14. Genotypes
used in the pair-wise analysis are shown along the top. Lanes containing reference samples are indicated by bold-face type and show
the homoduplex band. Subsequent lanes show heteroduplex mobility
analysis of the reference with other genotypes, as indicated. Hetero-

duplexes formed between two different HCMV genotypes can be seen


as bands with reduced mobilities in the gel. All combinations of genotypes gB1gB4 are shown. Genotypes gB6 and gB7 were excluded as
reference genotypes because they readily form homoduplexes with
genotypes gB3 and gB1 respectively (lanes 11 and 12). Lane M contains molecular size markers (pGEM cut with HinfI, RsaI and SinI).

son with gB sequences published previously [Shepp et


al., 1998] showed a high degree of sequence homology
between gB5 and both gB6 and gB7 (83% and 75%),
that would suggest there are two new genotypes. Phylogenetic and bootstrap analysis, however, indicate
that these isolates are variants of genotypes described
previously. The reasons for this may be due to viral
mutation or may be due to homologous recombination
between two isolates of different genotype as has been
suggested previously [Meyer-Konig et al., 1998]. The
model previously proposed is of two genotypically distinct strains that undergo homologous recombination
during viral replication in vivo. Such recombination
within the gB gene may account for the close relationship between the gB6 and gB7 isolates, and those previously classified. This finding was only possible using
heteroduplex mobility analysis and sequencing,
whereas restriction fragment length polymorphism
only detected changes occurring at the restriction sites.
This demonstrates the use of heteroduplex mobility
analysis for genotyping.
Using single-stranded conformation polymorphism
analysis, 22 variants (subtypes) were identified from
73 isolates. The results of the single-stranded conformation polymorphism analysis correlated with those
from the restriction enzyme digestion, that is, singlestranded conformation polymorphism analysis profiles
were unique to each genotypic group. Sequential isolates from the same patient identified as gB2 by restriction fragment length polymorphism analysis, displayed different additional bands by single-stranded
conformation polymorphism analysis. The differences
in nucleotide difference between these isolates suggested by single-stranded conformation polymorphism
analysis could be the result of adaptive mutation of the
virus with time. These isolates were 99% identical on
DNA sequencing. This indicates that single-stranded
conformation polymorphism analysis can detect even

minor sequence changes and may be useful for epidemiological studies and in cases of mother-to-child
transmission where the origin and development of a
virus could be tracked by comparing single-stranded
conformation polymorphism analysis profiles. The sensitivity and simplicity of single-stranded conformation
polymorphism analysis makes the assay well suited to
subtyping. The assays sensitivity can mean too many
variations (down to a point mutation) can be detected,
however, and the significance of such a change for alteration of protein function is uncertain.
Heteroduplex mobility analysis correlated well with
restriction fragment length polymorphism analysis for
all isolates, except for isolates 27EgB2 and 44EgB2,
that were identified as gB2 by restriction fragment
length polymorphism analysis, but gB1 and gB3 respectively by heteroduplex mobility analysis and sequencing. Using PCR and restriction fragment length
polymorphism analysis nucleotide sequence variation
directly around the restriction site can only be detected, whereas single-stranded conformation polymorphism analysis and heteroduplex mobility analysis can
detect variations throughout the entire sequence [Sheffield et al., 1993]. In some cases, restriction fragment
length polymorphism analysis leads to incorrect genotypic classification due to minor sequence variation
around the restriction site whereas heteroduplex mobility analysis classified these isolates into the correct
group. When heteroduplex mobility analysis was
carried out on two isolates from the same patient,
identified as gB2 by restriction fragment length polymorphism analysis, but that exhibited different singlestranded conformation polymorphism analysis patterns, all formed homoduplexes with the gB2 reference
sample. Sequencing showed that these isolates had
99% DNA homology. Overall, single-stranded conformation polymorphism analysis is too sensitive for genotyping purposes. Conversely, heteroduplex mobil-

486

Trincado et al.

Fig. 4. Phylogenetic tree generated by bootstrap analysis of the 300 bp region surrounding the cleavage site of gB. The internal node numbers
indicate the bootstrap values, obtained from 100 replicates. The use of this portion of gB shows that four groups are obtained, and these are
indicated by bold next to each branch. The branch lengths are proportional to the evolutionary distance between sequences and the distance
scale in nucleotide substitutions is shown.

ity analysis allows assignment to the correct genotype,


and has the advantage of allowing an estimation of the
percentage difference in nucleotide sequence between
two isolates, based on the shift of heteroduplexes from
the homoduplex. Heteroduplex mobility analysis and
sequencing results correlated completely but sequencing is more expensive and labour-intensive. Our conclusion is that heteroduplex mobility analysis is a better method for genotyping HCMV isolates, compared to
restriction fragment length polymorphism analysis,
single-stranded conformation polymorphism and DNA
sequencing.
ACKNOWLEDGMENTS
We would like to acknowledge the financial support
of the National Health and Medical Research Council,
the University of Western Sydney-Seed Grant, and the
Sydney Childrens Hospital Foundation. We would also
like to thank the Australian Pediatric Surveillance
Unit for their assistance in coordinating the response
of pediatricians.

REFERENCES
Alford C. 1991. Breast milk transmission of cytomegalovirus (CMV)
infection. Adv Exp Med Biol 310:293299.
Bale JF Jr, Petheram SJ, Souza IE, Murph JR. 1996. Cytomegalovirus reinfection in young children. J Pediatr 128:347352.
Binder T, Berg T, Siegert W, Schmidt CA. 1995. PCR-single-stranded
conformation polymorphism analysis: nonradioisotopic detection
with biotinylated primers and streptavidin-alkaline phosphatase
conjugate. Biotechniques 18:780781.
Britt WJ, Alford CA. 1996. Cytomegalovirus. In: Fields BN, Knipe
DM, Howley PM. editors. Fields virology, 3rd ed. Vol. 2. Philadelphia: Lippincott-Raven Publishers. pp 24932523.
Britt WJ, Vugler L, Butfiloski EJ, Stephens EB. 1990. Cell surface
expression of human cytomegalovirus (HCMV) gp55-116 (gB): use
of HCMV-recombinant vaccinia virus-infected cells in analysis of
the human neutralizing antibody response. J Virol 64:10791085.
Chou SW. 1990. Differentiation of cytomegalovirus strains by restriction analysis of DNA sequences amplified from clinical specimens.
J Infect Dis 162:738742.
Demmler GJ. 1991. Infectious Diseases Society of America and Centers for Disease Control. Summary of a workshop on surveillance
for congenital cytomegalovirus disease. Rev Infect Dis 13:315329.
Felsenstein J. 1993. PHYLIP inference package, version 3.5c. Department of Genetics, University of Washington, Seattle.
Fries BC, Chou S, Boeckh M, Torok-Storb B. 1994. Frequency distri-

HCMV gB Genotypes in Children


bution of cytomegalovirus envelope glycoprotein genotypes in bone
marrow transplant recipients. J Infect Dis 169:769774.
Glavac D, Dean M. 1993. Optimization of the single-strand conformation polymorphism (single-stranded conformation polymorphism
analysis) technique for detection of point mutations. Hum Mutat
2:404414.
Markoff A, Savov A, Vladimirov V, Bogdanova N, Kremensky I, Ganev
V. 1997. Optimization of single-strand conformation polymorphism analysis in the presence of polyethylene glycol. Clin Chem
43:3033.
Meyer-Konig U, Haberland M, von Laer D, Haller O, Hufert FT. 1998.
Intragenic variability of human cytomegalovirus glycoprotein B in
clinical strains. J Infect Dis 177:11621169.
Page RD. 1996. TreeView: an application to display phylogenetic trees
on personal computers. Comp App Biosci 12:357358.
Pass RF. 1985. Epidemiology and transmission of cytomegalovirus. J
Infect Dis 152:243248.
Pass RF. 1991. Day-care centers and the spread of cytomegalovirus
and parvovirus B19. Pediatr Ann 20:419426.
Pass RF, Stagno S, Dworsky ME, Smith RJ, Alford CA. 1982. Excre-

487
tion of cytomegalovirus in mothers: observations after delivery of
congenitally infected and normal infants. J Infect Dis 146:16.
Rasmussen L, Hong C, Zipeto D, Morris S, Sherman D, Chou S, Miner
R, Drew WL, Wolitz R, Dowling A, Warford A, Merigan TC. 1997.
Cytomegalovirus gB genotype distribution differs in human immunodeficiency virus-infected patients and immunocompromised
allograft recipients. J Infect Dis 175:179184.
Rasmussen L, Matkin C, Spaete R, Pachl C, Merigan TC. 1991. Antibody response to human cytomegalovirus glycoproteins gB and
gH after natural infection in humans. J Infect Dis 164:835842.
Sheffield VC, Beck JS, Kwitek AE, Sandstrom DW, Stone EM. 1993.
The sensitivity of single-strand conformation polymorphism
analysis for the detection of single base substitutions. Genomics
16:325332.
Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improving
the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties and
weight matrix choice. Nucleic Acids Res 22:46734680.
White PA, Zhengqian L, Rawlinson WD. 1999. Sequence diversity in
the 5-UTR region of GB virus C/hepatitis G virus assessed using
sequencing, heteroduplex mobility analysis and single-strand conformation polymorphism. J Virol Meth 83:91101.