Persistence of multiple maternal genotypes of human immunodeficiency virus type I in infants infected by vertical transmission

Persistence of multiple maternal genotypes of

human immunodeficiency virus type I in infants

infected by vertical transmission.

S L Lamers, … , D J Barrett, M M Goodenow

J Clin Invest.

1994;

93(1)

:380-390.

https://doi.org/10.1172/JCI116970

.

The extent of nucleotide variation within the HIV-1 env hypervariable domains serves as a

marker of virus genotypes within infected individuals and as a means to track transmission

of the virus between individuals. We analyzed env V1 and V2 sequences in longitudinal

samples from two HIV-1-infected mothers, each with three children infected by maternal

transmission of the virus. Sequences in samples that were obtained from two infants at 2 d

and 4 wk after birth displayed more variation in V1 and V2 than maternal samples obtained

at the same times. Multiple HIV-1 genotypes were identified in each mother. In each family,

multiple maternal HIV-1 genotypes were transmitted to the infants. Specific amino acid

residues in the hypervariable domains were conserved within sequences from each family

producing a family-specific amino acid signature pattern in V1 and V2. Viruses that were

highly related to maternal viruses in signature pattern persisted for as long as 4 yr in the

older children. Results support a model of transmission involving multiple HIV-1 genotypes

with development of genetic variation from differential outgrowth and accumulation of

genetic changes within each individual.

Research Article

Find the latest version:

Persistence

of Multiple Maternal Genotypes of Human Immunodeficiency

Virus Type I in Infants Infected by Vertical

Transmission

SusannaL.Lamers,* John W.Sleasman,**_{Jin Xiong She,*}6_{KimberlyA.Barrie,*StevenM.Pomeroy,*} Douglas J. Barrett,** and Maureen M. Goodenow**

Departmentof *Pathology,tDepartmentofPediatrics,DivisionofImmunology, and the §Centerfor MammalianGenetics, UniversityofFlorida College ofMedicine, Gainesville, Florida32610

Abstract

The extentof nucleotidevariation within the HIV-1env hyper-variabledomains serves as a marker of virus genotypes within infected individuals and as a means to track transmission of the virus between individuals. We analyzed env VI and V2

se-quences in longitudinal samples from two HIV-1-infected mothers, each with three children infectedbymaternal trans-missionofthevirus.Sequencesinsamplesthat were obtained fromtwo infants at 2 d and 4 wk after birth displayed more variation in V1 and V2 than maternalsamplesobtainedatthe sametimes.MultipleHIV-1 genotypeswereidentified in each mother. In each family, multiple maternal HIV-1 genotypes weretransmittedtotheinfants.Specificamino acid residues in the hypervariable domains were conserved within sequences from eachfamily producingafamily-specificamino acid signa-turepattern in V1 and V2. Viruses thatwerehighlyrelatedto

maternalviruses insignaturepattern

persisted

foraslongas4 yr in the olderchildren.Results supportamodel of transmis-sioninvolving

multiple

HIV-1 genotypeswith

development

of genetic variation fromdifferentialoutgrowthand accumulation of genetic changes within each individual. (J. Clin. Invest. 1994.93:380-390.) Keywords: humanimmunodeficiencyvirus 1genetic variation*maternal transmission

Introduction

Maternaltransmissionof humanimmunodeficiencyvirus type

I

(HIV-1)

occurs in - ₁₅ to 35% of infants bornto infected women( 1-10 ).Whymorethan two thirds of the infants born to HIV-1-infectedwomen

apparently

escape infection

by

the virusraises important questionsabout virus-host interactions. Mother-to-child transmission ofHIV- 1 isundoubtedly

multi-factorial,

although the factors that contribute to, or inhibit, maternal transmission of HIV-1 are not completely defined ( 11 ). Becausepassive maternal

antibody

isthe

primary

immu-nity that a fetus acquires, it is likely that maternal humoral immunity isonefactor in vertical transmission ofHIV- 1 ( 12-15).However, thespecificities of maternal antibodiesthat can protectagainstHIV- 1 transmission areunclear ( 16-20). It is also

likely

that factors related to maternal virus titers, virus

Addresscorrespondence to Dr. Maureen M. Goodenow, Department

of Pathology, Box 100275,University of Florida College of Medicine,

Gainesville,FL32610.

Receivedforpublication21January 1993 and inrevisedform 13

July1993.

tropismfor different cell types, and virus variability contribute to verticaltransmission.

Genetic variation within HIV- 1 in vivo can have a signifi-cantimpactonthebiological characteristics and pathogenesis ofthe virus (21-25). Phenotypic changes of the virus can be related directly togenetic changes that result in a high fre-quency of amino acid substitutions, deletions, and insertions

clusteredinfivehypervariable domains, VI toV5, of the

env-encoded gp 120 molecule (26,27). Multiple functional determi-nantsfor virus cell tropism and host immune response map to the envhypervariable domains (28-36).

Nucleotide diversification in HIV-l envhypervariable do-mainsresults inviruses that can be distinguished by genetic analyses (37-41). Analyses of genetic variation primarily in

the V3 domain of env within infected mother-infantpairs indi-cated that HIV- 1 sequencediversity during the first 4moafter birthwasgreater insamplesfrom the mothers than from their infants (42,43). Because there was no evidence in either study for transmission ofmultiple maternal HIV- 1 genotypes to the infants, itwasproposedthat vertical HIV- 1 infection involves selectivetransmission ofaparticular maternal variantof the virus.

Our studies have focused on thegeneticvariability of

HIV-1 overtimein twodifferentfamilies, whereneonates and their older siblings were infectedbymaternal transmission of the virus. TheV1 and V2hypervariable domains ofenv were ana-lyzedbecause previous studiesdemonstrated thatextensive ge-netic variability in V1 and V2 could be used to distinguish among strains ofHIV-1 (37). The goals were to determine whether or notmultipleHIV- 1 genotypeswere presentinthe mothers,toassess thepossibilitythatmultiplegenotypes could betransmitted, andtoevaluate theextentofgenetic divergence overtime inepidemiologically related viruses infecting differ-entindividuals.

Methods

HIV-Iinfectedmothers andchildren.Samples ofperipheral blood were collectedoveraperiodof almost 2yrfrom members oftwofamilies, FamilyAandFamilyB(Fig.1). The mothers, MA and MB, each had

threeinfected children. Al and BI indicatethe eldestchild ineach

family;A2 andB2, the secondchild; A3 andB3, the youngestchild. Samplesarecodedwith numberstorepresent the dates collected. In all cases,theearliest available sampleswerestudied.Inseveralinstances, sampleswerecollectedfrom the mothersnear the time of conception

(Fig. 1). SampleswereobtainedthroughthePediatricImmunology

Clinic with maternal informedconsentunderaprotocol approved by

theInstitutional Review Board of the University ofFloridaCollegeof Medicine.

Peripheral blood T-cell subsetswereanalyzed by two-color flow

cytometry(FACSCANO; Becton Dickinson& Co.,Mountain View,

CA).Sera HIV-1 p24antigenwasdetermined byantigencapture

(Ab-bottLaboratories,NorthChicago, IL).Assays werecarriedoutin the J. Clin. Invest.

© TheAmericanSociety for Clinical Investigation,Inc.

0021-9738/94/01/0380/11 $2.00

Clinical Immunology Laboratory of Shands Hospital at the University ofFlorida.

MA was identified as HIV-1 infected when Al, her 21-mo-old child, was found to have lymphoid interstitial pneumonitis, lymphade-nopathy, and positiveHIV-l serology. MA had no history of intrave-nousdrug use by toxicology profile or evidence of other sexually trans-mitted diseases. Her risk factor was sexual contact with an intravenous druguser.MA's CD4 T-cell count fell from 587 cells/mm3 in October 1989, when she was enrolled in the study, to 38 cells/mm3 in January 1991, when zidovudine therapy was initiated. MA developed toxoplas-mosis as her AIDS-defining illness in June 1991.

The mother in FamilyBwasidentified as infected by HIV-Iwhen heroldest child, B 1, was evaluated at 9 mo ofage for lymphadenopathy andinterstitial pneumonitis. Similar to mother A, MB's risk factor was

asexual partner who was an intravenous drug abuser. MB was asymp-tomatic with 781 CD4 cells/mm3 in January 1988. Her CD4 T-cell

countdropped to 296 cells/mm3 in December 1990 and to 39 cells/

mm'in October 1991. During that time she developed HIV-associated wasting syndrome but had no opportunistic infections. Antiviral ther-apy wasinitiated after October 1991.

Eachof the six children were full-term infants born by spontaneous vaginal delivery with normal birth histories and neonatal physical ex-aminations. There was no clinical evidence for fetomaternal hemor-rhage.Clinical profiles of the children are summarized in Table I. Two infants in Family A were breast fed; Al, until 11 mo of age, and A2 until7wkof age. ChildAl developed generalized lymphadenopathy at

14 moand interstitial pneumonitis at 21 mo. Infant A2 developed pneumococcal meningitis at 18 wk of age. Infant A3 was evaluated at 9

wkof age for fever of unknown origin, oral candidiasis, and lymphade-nopathy, and developed Pneumocystis carinii and cytomegalovirus pneumonia at14wkof age. Infant B2 was positive for p24 antigen at 3 mo, but clinically asymptomatic until February 1991, when he devel-oped Pneumocystis carinii pneumonia. Infant B3 showed evidence of oralcandidiasis and lymphadenopathy at4 wkof age.

Nucleic acidextraction,amplification,andsequencing.Separation of peripheral blood mononuclear cells (PBMC), DNA purification, andamplificationprocedures were carried out in a laboratory free of cultured HIV-1 isolates orclonedHIV-1 sequences. DNA was

pre-pared from Ficoll-Hypaque separated PBMC accordingto standard procedures (45).

Sequences of HIV-IenvVI-V2-C2regionwereamplified usingthe PCRand nested oligonucleotide primers prepared inthe DNA synthe-sis core facility of the Interdisciplinary Center for Biotechnology

Re-search at theUniversity of Florida. The outer primerswere ENV5

(5'-GGTAGAACAGATGCATGAGGAT-3') and ENV6 (5'-CCA-TGTGTACATTGTACTGTGCT-3'), located at nucleotides 6102-6123 and 6544-6566 in the genome of

HIVLAI

(27).The innerprimers

were SK 122 and SK 123 (37). HIV- I envsequenceswereamplified

from 1 ggof PBMCDNAin 100MAlreactions with PCR buffer (1.75 mMMgCl2,50 mMKCl,20 mMTris-HCI,pH8.4,0.1%BSA), 100 pmol of each outer primer, 200 Mmolof each dNTP, and 2.5 units Amplitaq (Perkin Elmer Cetus Corp., Norwalk,CT) overlayedwith 25

,ulmineraloil. Reagent controls andDNAfrom uninfected cellswere

included in every experiment.Amplificationreactionswerecarriedout

inanautomated, 48-well thermalcycler(PerkinElmerCetusCorp.)

programmed for one cycle of denaturation(940Cfor 10 min), 25 cy-cles ofamplification (each cycleincluded 1 min at 940C, 1 min at

50'C, and 2 minat720C),followedbyonecycleat720Cfor 10 min. After passage throughaCentricon 100 filter(Amicon/W.R.Grace & Co., Beverly, MA)onetenthof theproductwasusedastemplateina

second round ofamplificationfor 20cycleswith innerprimersand the

sameprofileasabove.

Amplified products ofapproximately 400to 450basepairs (bp)

wereisolated from agarose gels and blunt-endligatedintopGEM-3Z

plasmid vectors(Promega Corp., Madison, WI). Multiple

recombi-nantplasmids from eachsampleweresequenced bythedideoxychain termination method with SP6 andT7primers,35S-labeleddATP,and Sequenase Version 2.0 (United States Biochemical Corp., Cleveland, OH).

Analyses of nucleotidesequences. Sequenceswerealignedtogive

minimumevolutionarydistancesusingthemultiplealignedsequence editor program(kindly provided byD.Faulkner andY.Jerka, Molecu-larBiologyComputer Research Resource, HarvardUniversity,

Cam-bridge, MA). Aftertranslationof nucleotide sequencesusing

Micro-genie(IntelliGenetics, Inc., Mountain View, CA), alignments were subsequentlyeditedbyhand duetoextensivelengthvariation in the

hypervariabledomains. Theamplified regionincluded 70to100bpof

TableI. ClinicalProfilesof HI V-i-Infected SiblingsinTwoFamilies

Dateof Birth Sample CD4 CD4/CD8 HIV

Patient birth weight code* Age CDC staget count ratio p24 Age

kig ceI)/.A!(//nIRIIit3

A3 9/16/90 2.86 0990 2d P1-A ND ND +

1190 9wk P2-A 1170 0.9 +

0191 14 wk P2-A, B,D1 1154 0.4 +

A2 8/30/89 2.75 1089 7wk P1-B ND ND +

0190 18wk P2-A,D2 1402 2.4 +

Al 1/05/88 3.43 1089 21 mo P2-A, C 1342 0.3

0590 29mo P2-A,C 925 0.5 ND

0191" 36mo P2-A, C 1012 0.8

B3 9/03/91 3.23 1091 4wk P2-A 2450 2.7

-B2 2/04/90 3.54 0291 12mo P2-A,DI 257 0.9 +

0491 14 mo P2-A,DI 1210 1.1 +

BI 2/21/87 3.10 0789 29mo P2-A, C 200 0.2 +

0989 31 mo P2-A, C 224 0.2 ND

0890 42 mo P2-A,C 287 0.3

1290"1 45 mo P2-A, C 724 0.4

* Sample code represents date(monthandyear)ofsamplecollection and isthesame asindicated inFig. 1. Inallcasestheearliestsample indi-catedwastheearliestoneobtained. *_{Centers for Disease Control classification system forstaging}ofpediatricHIV- I infection(44). §

Deter-mined by antigen capturetest(Abbott Laboratories, NorthChicago,IL).

1l

Zidovudine therapywasinitiatedbeforethesesamples.

VIsequences,100 to 150 bp of V2, and 140 bp ofconserved region C2. The V1 and V2domains weredefined by the cysteine residues pro-posed toformdisulfide linkages in the region (46).

The program SYNO (kindly provided byM. Nei,Pennsylvania

StateUniversity, University Park, PA)wasusedtocalculate nucleotide distances transformed by the Jukes-Cantor formula(47,48). The Stu-dent'st test wasusedtoevaluate thesignificanceof differences in

nu-cleotide distances withinanindividual and between individuals(47).

Phylogenetic trees resulted from analyses of selected sequencesusing

Fitch-Margolish version 3.2.

Nucleotide sequence accession numbers. Sequences have been sub-mitted to the GenBank database (accession numbers L15085 to LI5187), andalignments of sequences have beenprovidedtothe Hu-manRetroviruses and AIDS Database. Both dataresources are main-tainedatLos Alamos NationalLaboratory, Los Alamos, NM.

Results

HIV-I genetic variation within mothers and theirinfants. The

extent of nucleic acid variation in envV1-V2-C region was

analyzed within samples from each of thetwo mothers and three of theirinfants, A2, A3,andB3, shortlyafter birth(Fig. 1). IninfantA2,envsequenceswereessentially homogeneous in sample A2. 1089 obtainedat 7 wk of age andin asecond sample set, A2.0 190,obtained at 18 wk ofage(Table II). Yet, variation within maternal sample MA.1089, obtained at the

sametimeasinfantsampleA2.1089,was8.2%acrosstheVI

-V2-C region, reflecting variationof 22.4% in V1 and almost 12%inV2.Itis unlikelythat the sequencehomogeneityinthe infantwasduetononreplicatingvirus because A2waspositive forp24 antigenatboth timepoints (seeTableI).

In contrast,samplesfromtwoinfants,A3 andB3, displayed greaternucleotide sequence variation than samples obtained from theirmothersatthesametime(Table II). Variation in sequences from infant A3 at 2 d of age (A3.0990) was 1.3%

compared with variation of 0.2% in maternal sample MA.0990. Infant samplesA3. 1190 and A3.0191 increased in variation at 9 and 14 wkafter birth, relative to variation in

88p

90 91

MA onfl

I'V-A3 - ---I

Ur A2

§

_

- o

87 90 91

11S

MB

I-a

B3

I

v

§

s

V--8

B2 1

9

B1

l 1 _N

R

cm

Figure1.Samples from two families. Family A is on the left; Family B isonthe right. Line along the top indicates time in years. Lines

be-neathdate line represent individuals. MA and MB, mothers;

Al

and

B1, oldestchild; A2 and B2, second child; A3 and B3, youngest child.

Forinfant A2, the horizontal line begins at the time of birth. For infants A3, B2, and B3, dotted lines indicate gestation. Vertical lines

designatethedeathof infantsA2and A3. Black boxes represent sam-ples. Numbers underneath each box designate the sample by the month and yearof its collection. Samples with the same number were collectedatthe sametime.

Table II.NucleicAcid Variation in HIV-1envwithinSamples

from Mothers and Infants near Birth

Regionsof HIV-1env

sample

set* V1/V2/C VI V2

MA.1089 8.2(1.1)' 22.4(3.7) 11.9(2.2)

A2.1089 0 0 0

A2.0190 0 0 0

MA.0990 0.2(0.1) <0.2(0.2) <0.1 (0.6)

A3.0990 1.3(0.3) 1.0(0.6) 1.5(0.6)

MA.1190 1.6(.02) 2.3(1.1) 2.7(0.9)

A3.1190 1(0.4) 1.3(0.5) '(1.2)

A3.0191 (1.1) (2.3) (1.7)

MB.1091 4.5(0.7) 6.3(1.7) 6.8(1.7)

B3.1091 ((1.1) (3.5) 10.5(2.3)

*Each samplesetcontainsfive toeightsequences.

'Percent

mean

variation(±SEM). § Shaded values indicate infantsamplesthatare

significantlymorevariablethanmaternalsamples byStudent'st test

(53).

maternal env sequences (P<0.01).Ingeneral, both theV1and V2 domains in sequences from infantA3 were significantly morevariablethanmaternal MA.1190 sequences (P<0.001 ) (Table II). Likewise in Family B, the population ofenv se-quences from infantsample B3. 1091 wasmorevariable across the entire V1-V2-C region than maternal env sequences MB.1091 (P<0.01). Thesematernal-infantsamples were ob-tainedatthesametime when B3was4wkofage.Each hyper-variabledomainvariedindependently.V 1 sequences were sig-nificantlymorevariable in infantsample B3. 1091 than in ma-ternal sample MB.1091 (P < 0.001), whereas V2 sequences were not (P = 0.1 ). Results of these analyses indicated that HIV-1 sequencesin neonates canbe significantly more vari-able thanmaternal virussequencesatthe samepoint in time. Genetic variationin VI andV2sequences in

epidemiologi-cally

unrelated viruses. The extent of variation within virus populations that were epidemiologically related in a mother andherinfantsledus todefinethedistancebetween epidemiol-ogically unlinked virus populations.AllV1 and V2amino acid sequencesfromthetwomothers werealigned, guided by the 8 cysteine residues dispersedacrosstheregion, and the 12 con-served amino acid residuesbetween V1 and V2 (Table III). Because oflength variation in each of the hypervariable do-mains, spacing was introduced to maximize alignments

be-tweenthetwosetsofsequences.

In contrast tothe C2region, where nucleotide variability betweenMBand MA sequences was9.4±2.5%, distance in the

V I hypervariable domain was 30.5±4.2% and in V2, 15.9±2.7% (Table III). Variability in the HIV-1 sequences

be-tweenthetwomotherswasnotsignificantlydifferent in some

casesfromthevariation inHIV- 1 sequences betweeninfected membersofthe samefamily (data not shown). Yet, there was

noevidence fromtheclinical histories of the mothers to

fea-I

I NIN I .I W INM

P4. IT IT rI IJ I f1 IL fI rI rI 4 4 P

HI I I

COI

) II

C/) IZ 1:I1:-III I ~ ~ ~ ~ ~ ~ F-, H H

>1 IT4~~~~I

1)

)ZZ

CO Z Z

E-...

f-"W-_KW4_%IXs

HII I I ClH P4 P P4 P4 Z g/

>4 I I I

I l I

I

IIIII I~ ~ ~ ~ ~ ~~~I

C/)I~~~~()

) II I> P E-

-Ofl0~ EA E E4 Z Cl)

-,

I

III I ~

~~I

~

0) 0)

Cd

CD _Cd

O

H~ C H~

0Y 0)00

0)I C) H

dP

Cl) +1

C-'

+1

al)

Cd

C')

0\ -Hq

4J

Cd rij

Cd

0 0~

'A

0

E

4)

C,.

"0

4)

4)

C.

dO

4)

0.

.0 "0

4) 4) 1...) C. 4)

0 0

0

C-' 4) 4)

0"

4)

Cd 4)

4)

0

4) C-' 4) C.

0.

4)

"0

4)

"0 0

C-' 4) 4)

0

4)

4)

.0

0

10

4).

C)

cCd Q10

*Cd

Cd4

u>0

0 _CZ

.-W4 )

*

Hn

a)

Cdl

PQ (_)

134

turesoftheenvsequence, in additiontonucleotide

variation,

that characterizeepidemiologically related viruses. For

exam-ple, conserved amino acid signature patterns in env contrib-utedtoestablishingtheepidemiologicallink among HIV- 1 vari-ants in a groupof infected individuals (41 ).

Amino acid signature pattern. Based on the alignment of maternal sequences,itwaspossibletoidentifyapattern in the V 1-V2-C2 region where particular amino acids appeared in nearly all sequencesfromonemother butnotin the sequences from the second mother. This array of amino acidscomprises the signaturepattern foran individual'ssequences. A

unique

signaturepatternwasidentifiedin 16 amino-acid

positions

dis-persed alongthe V 1-V2-Cregion (Table III, shadedpositions). Only4signaturepositionswerelocated in the conserved

region

C2,whereas 12signature residueswerelocated in the hypervar-iable domains, 7inV1and 5 in V2. The patternof aminoacids in these positions did not predominate when V1-V2-C2

se-quences from molecular clones ofHIV- 1 (27) or in HIV- 1 isolates (37) wereanalyzed (data notshown), suggestingthat these residueswereconserved amongepidemiologicallyrelated viruses. If thiswerethe case, thenV1-V2-C2sequenceswithin the infants and children in each

family

should contain their respectivematernal signaturepatterns.

Toexamine this possibility, representativesequencesfrom each mother were used as reference for the alignment of all sequencesfromherchildren.

Significantly,

theaminoacid

sig-naturepatternidentifiedin the maternal sequenceswaspresent in most sequencesfromall membersofeachfamily (TablesIV

and

V,

shaded

residues).

Theaminoacid

_{signature positions}

in theV1domainwere

_designated

1

through

7 and those in

V2,

8

through

12. Positions in

C2, designated

13

through

16,

were alsoconservedina

_{family-specific}

_pattern

_(data

not

shown).

Analyses

of65 clones from

Family

A and 57 clonesfrom

Family

Brevealed in

general

two

_predominant

amino acidsat

any

signature

position (Table VI).

Therewas

essentially

no

overlap

inamino acidusageat_any

_{signature position}

between sequences from the two families. The

_{family-specific}

amino acid

signature

patternwas_{apparent within}

epidemiologically

related virusesin theface of considerable overallsequence vari-ation.

Moreover,

the

_signature

patterns

persisted

foras

long

as 3to4_yrnot

_only

_{in maternal HIV- 1 sequences, but in HIV- 1}

sequencesoftheirchildren aswell.

Transmission

of multiple

genotypes. It became clearfrom the

alignment

ofV1 and V2 sequences that a numberof dis-tinctHIV- 1 genotypes could beidentified withineach mother.

Likewise,

more than one genotype

appeared

in most ofthe children. The

relationship

between

_{representative}

maternal virus genotypes and sequences found in the infants ortheir older

siblings

wasassessed

by

phylogenetic

tree

_analyses.

In

Family

A,

three

_major

V1 genotypeswerefound in the maternal sequences,

represented by

MA.1089.

_1,

MA.

_1089.2,

and MA. 1089.3

_(Fig.

2

A,

upper

diagram).

Sequences

closely

related to two maternal genotypes were found in infant A3. Variant A3.0191.3 clustered with maternal genotype MA.

1089.2,

whereas several additional A3 variants clustered with maternal genotype MA.1089.1. In the V2

domain,

two

Table IV. HIV-1 Amino AcidSignature SequencesinMother-ChildSamples of FamilyA

Sequences* Clones

ofsamplest VI V2

MA.₁₁₉₀.1

_CVILTCT

PWG

_{NNTG"NSTGEKPEGEIKNCSFNITTS}

RRDK _REEYRIF"KLD VPI LDE GNKSFILIHC MA.1089.1* __3____

-[

W _

ENKL

a -T"----HF----

''N--NVG14 |- A--QnF--j- i MY-|

A3-.1190.2

- --- --- ____ID-AFi__---- 2 7 -~~.

I..1PaD

. G ~'QID AF A N

A3.0990.4* V-|---i---AF---AFA ---- _----AF

Y-A2.1089.1I SSR--SWDQ--..-GQIDAF A-S_..

A2.0190.1* -H- H aL SSR --SWDQ-f --- --G -QIDAF- A ---D

-S----Al.0191.1* D-SS --NSWGQ --QK--M -IG-RKKD AY -A-A - -S R-D-YR---.2 * --S0-K-WDQV-K---STS- GR-'L DRNDSRNSNYS-YR----F "T- - S *4 --0-N-- l- -D-L--NSWGQ-Q--- IGVR IKKD-AFAY-S-YR--A-4 DK

A1.0590.6 -NA--- -K-WDQV---M-IGGRKKAAF-A D S

RNDSR-D-YR-Y-B-89.

- - S-

---S-NNWGQQ---

IGGRLKKD-A--A -

SDLVPIDNDSRNDSR-S-YR---A1.10891.1* - N- T--S-- WG- -g--- IG-R QD-AF- A- -D0N

DSS-YR--:-2 N---SA-K-WDQ ---M---ST GR11KQLD-AF ADT S RNDSRNSNS-YR--'

5.2 ' N- ---SA-K-WDQ.KM---S-RIGGR...KKA.AF..A.--.---.-...RNDSR.D.YR....

--- T--G QD-AF- A- - G

R-D-YR---Signature

residues9 1

Table V.HIV-1 Amino Acid Signature Sequences in Mother-child Samples ofFamilyB

Sequences* Clones

ofsamplest V1 V2

MB.1091.5 C'

LNCS

SNQSRN

NEGJ

I

GEIKNCSFNITTSIRD

KEYALF

KL

VPI DNENNSHSRNNDSGSYRLI|C

MB.1290.1*

--

---I

T-V--

----NTS--

-

_I-R---

_---H--S

_-K---

-.3*

--

----

T---

---R-

--

---NQ---MB.O789.3

*

---T

_VTG

G-V--

_E

---

--K

----

SS

---

N----|

4*

_IMV-

_-V

_V--

S

_IDMK-

-.5-K::

- ----R ---_---D-NQ

---MB.0689.5*

--I-

--ST -V

D--

---

--- -- --K--IDM

----MWT----B3.1091.2* - ---K-- - KR

|-

K--NN-.3 -

-T-H---

----Y-R

---SSNY-M--N-R

---B 4*

TKV-

-T-

SSLG-

l-G--

1

1

I

-N--N--S-

S-B2.0491.1*

-V-

-

-VTG

E-

E

---|

--- KR- -RIIVNNK

---.7 -

---K-T

---

--- ---G--K-

---B2.0291.l* --

---KT---.4-- - --K - -R

B1.l290.1

TTV--- ---SKLE-

---A---

---K

-N -N--N---.5 -- --- _{TTV ---S--G- - G---} ---.6 I MVK -T-DSKLE--G-|- G-- --- --- ----

HN-NSS-N-S---|-.7 TTVS-- T-- GG -- A- N--S--N

-.9 T ----SKO- ---

N-N-Bl.0890.2 TTVS-S TT--SSLE- - - -S -N

Bl.0989.l TMVS- T--DSKLE- ---A---K-NSHSS-N-DN ---.3* T -I---SDGG-

---K-NNHSS-N--N---Bl.0789.4* T-V-T T -S LG- --- - --- - - --- --- S

-N--N--.8 -I -- TTVS--T- GG - - --- ---N--S- S

--Signature

residues67

* Alignment ofconstantregion isnotincluded because sequences in children are highly related to maternal sequences. tSequences of

representa-tiveclonesfromasampleareincluded. Stars indicatesequencesincluded in the phylogenetic tree analyses in Fig. 2.

major

genotypes were identified inboth the mother and her infant A3

(Fig.

2

A,

lower

diagram).

Infant

sequences

A3.0990.4 and A3.0191.1 clustered

closely

withonematernal genotype,

represented by

MA.1089.3orMA.1089.1.Asecond maternal

_variant,

MA.

1089.2,

wasrelatedtoinfantsequences

A3.1190.7

andA3.0191.3.

Multiple

HIV- 1 genotypes,

particularly

in

_Vl,

were also found inthe mother in

_Family

B

(Fig.

2 B,

upper

diagram).

Sequences

relatedto atleasttwomaternal_genotypeswere iden-tifiedin_{infant B3. One infant genotype,}

_B3.1091.2,

was

closely

related to a cluster of maternal variants

MB.0789.5,

MB.

_1290.3,

and MB.1091.1. Another infant genotype,

B3.1091.4,

was morerelatedto maternalvariantMB.1290.1. Therewasevidence for

multiple

maternalHIV-_{1 genotypes in}

the second

infant,

B2,

as well. One

_variant,

B2.029

1.1,

was

highly

related to the cluster that included maternal and B3 sequences, whileasecond

variant,

B2.049

_1.1,

was

_closely

re-latedtomaternalvariant MB.0789.3.

Persistence

of

genotypes.

Although

multiple

HIV-1 vari-antscould be

_transmitted,

allvariantswerenot

_necessarily

ap-parent in eithermaternal orinfant

samples

nearthetime of birth. For

_example,

infantvariant A3.0191.3wasfirst

identi-fied inasample obtained when A3was14 wkofage(TableI). The maternalprogenitorsequence,MA.1089.2, was notfound in the mother's samples obtained 2 d (MA.0990) or 9 wk (MA.1190) afterthebirth ofA3, but did appearin maternal sample MA.1089 obtained - 2 mobefore conception of A3

(Fig.

1).

Many of the individual Vl sequencesfound in the eldest child Al at21 to 36 moofage clustered togetherand,as a

group,were morerelatedtoMA.1089.2 than to other maternal variants

(Fig.

2

A, upper diagram). One variant,

A1.1089.1,

wasdistinct bothfrom

A1

sequences inthe cluster and from otherMAvariants. ThisA1 variantmay representavirus that evolvedindependentlyin thechildduringthetime after

trans-mission. Alternatively, variant A1.1089.1 may reflectyet

an-other maternalvariant thatwastransmitted and

persisted

in thechild.

Persistence ofHIV- 1 genotypes could bedetected in

Family

B, as well.Although therewas a

striking

similarity

in V2

se-quences among all membersof

family

B,

independent

ofthe timethat hadelapsedbetweenverticaltransmissionand subse-quent

samples,

V1 sequenceswere

clearly

morevariable

(Fig.

2 B,

upper

and

lower

diagrams).

Vl

sequences

in

samples

from

MB(MB.1091.l) andher infant B3 (B3.1091.2) were closely related toeach other and to maternal variant MB. 1290.3. The mother-infant

samples

designated

1091 were obtained 4 wk after the birth of B3, whereas maternal sample MB. 1290 was

obtained 10 mo

earlier, during

the firsttrimester of her preg- 8° nancy with B3 (Fig. 1). Even though MB's second infant, B2,

was 1 yrof age whensampleswere firstobtained,itwaspossible todistinguish twovariants in B2(B2.0291.1 and B2.049 1.1) that clustered with maternal variants MB.0789.5 and MB.

1290.3,

or

MB.0789.3,

respectively (Fig.

2B, upper

dia-gram).

Thesematernalvariantswereidentifiedin

samples

ob-tained - 18 moapart. One

sample

wasobtained

early

in

July

1989,duringthefirst trimester of herpregnancywith B2 and thesecond,in December1990,when B2was10moofage(Fig. 1).SimilarlyinMB's oldestchild,onevariant,B1.1290.1,was closelyrelatedtomaternal variant MB. 1290.1 (Fig.2B, upper

diagram).These maternal-childsampleswerecollected in De- z 00

cember of1990,whenB 1 was almost 4 yr ofage(Fig. 1).

z o

Discussion

The extent of nucleotide variation in env defined

multiple

>

HIV- 1 genotypesin each of the mothers inourstudies. A

num-berof distinct maternal HIV- 1genotypesweredetectable in the-infants ineachfamily.The mostplausibleexplanation forthe

presenceof

multiple

HIV- 1 genotypesin infants is that

multi-ple

virusescan be transmitted from the mother. Our

results,

LI which indicatethatpediatricHIV- I infection doesnotalways

resultfrom selective transmission ofmaternal HIV- 1

variants,

>

differ from otherstudiesinwhich infants infected

by

vertical C

transmissionwerefoundtoharbor

homogeneous

virus

popula-

.°.

tions relativetomaternal viruses (42, 43). XX

The families in our study were chosen because both _[ 0 0

mothers hadtransmittedHIV-1 to

multiple

infants. Vertical

transmission to more than one infant is not

unique

among 0

E HIV-I infectedwomenwithmorethanonechild

(1, 3, 6, 49,

50).The women in ourstudy didnottakezidovudineorother antiviral

therapies,

which

_{might impact}

the

frequency

of verti-cal transmission. Two of the six children whom we studied

were breast

fed,

but there was no evidence that

they

differed from the other HIV- 1 infected children whowere not. There

were noapparent differencesbetween otherstudies of vertical

x

transmission and ourswith regard tothe agesofthe

infants,

>

which

_ranged

from birth to - 4 mo, their

deliveries,

ortheir

clinical presentation (42, 43).

Because ourstudies werelongitudinal,oneexplanation for 0 °

our results could be thatdivergentHIV- 1 genotypes evolved 4 c

independently

in infantsand theirmothers froma common

progenitor

sequence.

Convergent evolution,

inwhich thesame aminoacid was fixed independently at the same position in the

V3

hypervariable

domain in more than one HIV-1

variant,

_e

_>°

apparently developed

invivoovera

period

of years

(51).

How- A

_Cn

ever,there isnoevidence thatdistinctHIV- 1 genotypes,such C.8 asthoseidentifiedinthe mothers and infants in our study, can

evolve

independently

within different individuals ina

period

of

weeks.Moreover, different HIV- 1 genotypes, which were iden-

_Mt8

tified in three oftheinfantsin our studies, were detected before

_4X

0

their births in earlier maternalsamples.

The mostlikelyexplanations for the difference between our

results and those from other studies of vertical transmissionare <!

00,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

~~~~~~~~~~~

0E

I~~~~~~~~~~~~~~~~~~~

LriW~~~~~~~~~~

i~~~~EhI!~~~~~

,.0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4

06l

a)

Ci~~ ~ ~ ~ ~ ~ ~ 0

0 a

I-t~ ~ ~ -Ca

MY U c~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a CYRCM ~ ~ ~aq

OJu

~~~~II

~~~~~~~~~

_~~~~~iI~~~~

and thatourstudieswere

longitudinal.

The V3 domainwasa

focus in other studies of maternal transmission of HIV- 1 (42,

43).

In contrast, both the VI and V2

hypervariable

domains of

envwereassessed inourstudies.V1 and V2 vary

independently

in HIV-1 genomes invivoandcan

display

more

genetic

vari-ability

than the V3 domain

(27, 52-55). Indeed,

development of

quasispecies

of HIV- 1 in vivo involvedmore

genetic

variabil-ity

ina

region

3'of the V3

loop,

than inV3itself(

56).

Because

ourstudies

analyzed

virus sequencesover

time,

rather thana

single

time

point,

thelikelihoodof

identifying multiple

HIV- 1 variants within individuals was increased. Different conclu-sions couldhave been drawn ifwehad limitedourstudiesto a

single

timepoint

or toa

single region

of theHIV- 1 genome;for

example,

the V2

domain,

which was

relatively homogeneous

among allindividuals in

Family

B.

Populations

ofHIV-1variantswerefoundtobemore heter-ogenous in

samples

from two infants

during

the first few months afterbirth than in

samples

from theirmothersatthe

sametime. Increasedsequence

heterogeneity

couldreflect ele-vatedlevelsof viremiainneonates

early

after infection by verti-cal

transmission,

similar to

early

viremia detected in adults after infection

by

other mechanisms

(57-59).

Oneinfant in

our

study

displayed

homogeneous

sequences inV1 and V2at

twoseparatetime

points,

suggesting

selective transmission of one maternal HIV- 1 variant.

Yet,

restricted variation in one

subgenomic region

ofHIV-1 doesnot

necessarily

indicate a

homogenous

population

of virus because different

regions

withinthesamevirusgenomecanvary

independently

(51, 53,

60, 61).

An alternative

explanation

is that

homogeneity

of virussequences inaninfant at7or 14 wkofage couldreflect selective

outgrowth

ofone of several maternal variants that

weretransmitted.

HIV-1 genotypes not found in

paired samples

from a

motherandone newborn close tothetime of birthwere de-tected

subsequently,

when all

samples

fromthe mother and her infantwere

analyzed.

A

representative

profile

of which

mater-nalvirusvariantsare

actually

transmittedmaynotbe

immedi-ately

evident,

inpartbecause oftechnical

factors,

suchasthe amountoftargetDNA

amplified

orthe numberofsequences

analyzed

in a

sample

set. In

addition, biological

factors

un-doubtedly

contributetothe

profile

of variants detectedbecause offluctuationsin

frequencies

ofHIV- 1 genotypesin peripheral blood cells

(61-64).

ParticularHIV-1 variantscanbe seques-tered inother

tissues,

suchasthe

lymph

nodes, before emerging in PBMCs

(65, 66). Compartmentalization

ofthevirusor acti-vation of virus

expression

would explain the appearance of additional maternal HIV- 1 genotypes in infants weeks after birth.

Timing

of vertical

transmission,

that

is,

whethertheinfants

wereinfected earlyingestationornearbirth,cannot be deter-mined fromourstudies primarilybecauseofthepersistence of maternal HIV-1 genetic variants, in some cases, throughout pregnancy.Nonetheless, studies ofgenetic variation in HIV- 1 do

provide insights

into mechanisms oftransmission.Because

multiple

maternal viruses were identified in theinfants, our studies indicatethatverticalHIV-1infectionis notalways lim-itedtoselective transmission ofasingle virionor a single virus genotype. Vertical transmissioncould occur by cell-free virus in maternal blood ifthe various maternal HIV-1 genotypes

were

expressed

(43). Mother-to-infanttransmission of HIV- 1 mayinvolve cell-associated virus,aswell(

11).

In ourstudies,

multiple

HIV- 1variantsdetected ininfantswere

closely

related

to many of the virus variants in maternal peripheral blood cells. Theextentof variation in HIV- 1 genotypes thatcanbe detected in maternal cells in small volumes ofperipheralblood would be sufficient to account for transmission of cell-asso-ciated virus. Maternallymphocyteshave beenidentified in the blood ofnewborns, for example, in infants with severe

com-bined immunodeficiency disease (67). The extent to which small numbers of maternal cells could be present in infants infected by vertical transmission of HIV- 1 remains to be exam-ined.

Genetically related virusespersistedforanumber of years withinmembersof each family. An unexpected result was the distinct pattern of amino acidsignatureresidues thatwas

con-served in V 1 and V2 sequences among populations of viruses in eachfamily. Persistence of the signature pattern indepen-dent of the overall extent ofgenetic variation in V 1 and V2 contributedto anunambiguous fingerprintthat linked epide-miologically relatedvirusesfor almost 4 yr after transmission. Resultsofourstudies indicatethat the VIand V2 domains can beimportant for evaluation of the molecular epidemiology of HIV- 1 transmission when the transmission link is in

question.

Acknowledgments

The authorsaregrateful forhelpful commentsanddiscussion from Livia Pedroza Martins, Nelson Michael, William Farmerie, and WaynePotts.

This research was funded in part by the National Institutes of Health under theauspices of the National Institute of Allergy and In-fectiousDiseases grantRO128210, and by the Pediatric AIDS Founda-tion grants 500028-7-PG and 500108-9-PGR.

References

1.Scott, G. B.,M. A.Fischl, N.Klimas,M. A.Fletcher,G. M.Dickinson, R.S.Levine, andW. P.Parks. 1985.Mothersof infants with theacquired immu-nodeficiency syndrome: evidence for both symptomatic and asymptomatic car-riers. J.Am.Med.Assoc.253:363-366.

2.Cowan,M.J.,D.Hellman,D.Chudwin,D. W.Wara,R.S.Chang,and A.J.Ammann. 1984. Maternal transmission ofacquired immune deficiency syndrome. Pediatrics. 73:382-386.

3.Kind, C.,B.Brandle, C.-A. Wyler,A.Calame, C.Rudin,U.B.Schaad,J. Schupbach, H.-P. Senn,L.Perrin,L.Matter, and the Swiss Neonatal HIVStudy Group. 1992.Epidemiologyof vertically transmittedHIV- Iinfection in Switzer-land:resultsofanationwideprospective study. Eur. J. Pediatr. 151:442-448.

4.Italian MulticenterStudy. 1988.Epidemiology,clinicalfeatures, and prog-nosticfactors ofpediatricHIVinfection. Lancet. 2:1043-1045.

5.European Collaborative Study. 1988. Mother-to-child transmission ofHIV infection. Lancet. 2:1039-1042.

6. Ryder,R.W.,W.Nsa, S.E.Hassig,F.Behets,M.Rayfield,B.Ekungola, A.M.Nelson, U. Mulenda, H. Francis,K.Mwandagalirwa,etal. 1989.Perinatal transmission of thehumanimmunodeficiency virustype 1 toinfants of seroposi-tivewomeninZaire.N.Engl. J. Med. 320:1637-1642.

7.Blanche, S., C.Rouzioux, M.-L. G.Moscato, F. Veber, M.-J. Mayaux, C. Jacomet, J.Tricoire,A.Deville,M.Vial, G. Firtion,etal. and theHIVInfection in NewbornsFrenchCollaborative StudyGroup. 1989. Aprospectivestudy of infants borntowomenseropositive forhumanimmunodeficiencyvirus type 1. N.Engi.J. Med. 320:1643-1648.

8.Andiman,W. A., J.Simpson, andB.Olson.1990. Rateoftransmission of HIV type 1infection frommother tochildand short-termoutcomeof neonatal infection.Am.J.Dis.Child. 144:758-766.

9. EuropeanCollaborative Study.1991. Children born towomenwith HIV- I infection:naturalhistoryandrisk oftransmission. Lancet.337:253-260.

10.Caldwell,M.B., and M. F.Rogers.1991. Epidemiology of pediatric HIV infection. Ped. Clin.NorthAm.38:1-16.

11. Report ofaConsensus Workshop. 1992. Maternalfactorsinvolved in mother-to-child transmission of HIV-1. J. Acquired Immune Defic. Syndr. 5:1019-1029.

12.Broliden,P.A., V.Moschese,K.Ljunggren,J. Rosen, C. Fundaro, A. Plebani,M.Jondal,P.Rossi, andB.Wahren. 1989. Diagnosticimplicationsof

13. Goedert, J.,H.Mendez, J.E.Drummond,M. Robert-Guroff, H. L. Min-koff, S. Holman,R.Stevens,A.Rubinstein, W.A.Blattner,A.Willoughby, and S.H.Landesman. 1989. Mother-to-infant transmission of human

immunodefi-ciency virustype 1: association with prematurityorlow anti-gp120. Lancet. 2:1351-1354.

14.Rossi,P., V.Moschese,P. A.Broliden,C.Fundaro, I. Quinti,A.Plebani, C.Giaquinto,P.Tovo,K.Ljunggren, and J.Rosen.1989.Presenceof maternal antibodiesto human immunodeficiency virus-I envelopeglycoprotein gpl20 epitopes correlates withtheuninfectedstatusof children born toseropositive mothers. Proc. Natl. Acad. Sci. USA. 86:8055-8056.

15.Devash, Y.,T.Cavelli,D.Wood,K. Reagan,andA.Rubenstein. 1990. Vertical transmission of human immunodeficiency virusiscorrelated withthe absence ofhigh-affinity/aviditymaternalantibodiestothegp120principal neu-tralizingdomain. Proc. Natl. Acad. Sci. USA. 87:3445-3449.

16.Rossi, P., V. Moschese,A.de Rossi,B.Wahren, M. Jansson,V.Lombardi, and H.Wigzell. 1991. Maternal antibody epitope mapping inmother-to-child transmissionof HIV. Immunobiol. Proteins Pept. 6:47-52.

17.Ljunggren, K.,V.Moschese, andP.Broliden. 1990. Antibodies mediating cellularcytotoxicityand neutralization correlate withabetterclinicalstagein childrenborn to humanimmunodeficiency virus infected mothers. J. Infect. Dis. 161:198-202.

18.Parkeh, B., N.Shaffer, C.Pau, E.Abrams,P.Thomas,H. Pollack,M. Bamji,A.Kaul, G. Schochetman,M.Rogers,J. R.George, and The NYC Perina-tal HIVTransmission Collaborative Group. 1991. Lack of correlation between maternalantibodiestoV3 looppeptides ofgp120 andperinataltransmission. AIDS(Phila.). 5:1179-1184.

19. Kollman,T. R., A.Rubinstein,W.D.Lyman,R.Soiero, andH. Gold-stein. 1991. Characterization of IgG and IgG subclass antibodies present in paired maternal andfetalserumwhicharedirected againstHIV- Iproteins. AIDSRes. Hum.Retroviruses. 7:847-854.

20. Ugen,K.E., J. J. Goedert, J. Boyer,Y.Refaeli, I. Frank,W. V.Williams, A.Willoughby, S. Landesman,A.Rubinstein, J.Berzofsky,T.Kieber-Emmons, andD.B.Weiner. 1992. Vertical transmission of HIV infection: maternal hu-moral immune responsestogpl20andgp4i peptides. J. Clin. Invest. 89:183-189.

21.Goudsmit, J.,K.Nicole,T.Back, andP.L. Nara.1991.Genomic diversity andantigenic variation ofHIV- 1:links betweenpathogenesis,epidemiology and vaccine development. FASEB (Fed.Am.Soc. Exp. Biol.). J. 5:2427-2436.

22. Fenyo,E.M., L.Morfeldt-Manson,F.Chiodi,B.Lind,A. vonGegerfelt, J. Albert, E.Olausson, andB.Asjo. 1988. Distinct replicativeandcytopathic characteristics of human immunodeficiency virus isolates. J. Virol. 62:4414-4419.

23.DeJong,J.-J.,J.Goudsmit,W.Keulen,B.Klaver,W. Krone, M. Ters-mette,andA.De Ronde. 1992.Humanimmunodeficiency virus type 1clones chimeric fortheenvelope V3 domain differ in syncytium formationand replica-tioncapacity.J. Virol. 66:757-765.

24.Schuitemaker, H.,M.Koot,N. A.Kootstra,M. W.Dercksen,R. E. Y.de Goede, R. P. vanSteenwijk,J. M. A.Lange,J. K. M.Eeftink Schattekerk, F. Miedema, andM. Tersmette.1992.Biological phenotype ofhuman immunodefi-ciency virustype Iclonesatdifferentstagesofinfection:progressionof disease is associated withashiftfrommoncytotropictoT-cell-tropicviruspopulations.J. Virol. 66:1354-1360.

25.DeRossi,A.,M.Pasti,F.Mammano,L.Ometto, C.Giaquinto,and L. Chieco-Bianchi. 1991. Perinatal infection byhuman immunodeficiency virus type I(HIV-I):relationshipbetweenproviralcopynumberinvivo,viral proper-tiesinvitro, and clinicaloutcome. J.Aled. Virol. 35:283-289.

26. Modrow, S.,B. H.Hahn, G.M.Shaw,R.C.Gallo,F.Wong-Staal,andH. de Wolf. 1987.Computer-assisted analysisofenvelopeproteinsequencesofseven

humanimmunodeficiency virus isolates:predictionsofantigenic epitopesin con-servedandvariableregions.J. Virol. 61:570-578.

27. Myers, G. M.,B.Korber, J.A.Berzofsky,R. F.Smith,andG.N.Pavlakis. 1992.HumanRetroviruses and AIDS. LosAlamos NationalLaboratory, Los Alamos,NM. 1050pp.

28.vanTijn, D., C. A.B. Boucher, M.Bakker, and J. Goudsmit. 1989.

Antigenicityof linearB-cellepitopes intheCl,VI,andV3region ofHIV-I gp 120.JAcquiredImmuneDefic. Syndr.2:303-306.

29. LaRosa, G. J., J.P.Davide,K.Weinhold,J. A.Waterbury,A.T.Profy, J. A.Lewis,A.J.Langlois, G.R.Dreesman, R. N.Boswell,P.Shadduck,etal. 1990. Conservedsequenceandstructuralelements in the HIV- Iprincipal neutral-izing determinant.Science(Wash.DC). 249:932-935.

30.Shioda, T.,J.Levy,and C.Cheng-Mayer. 1991. Macrophageand T cell-linetropisms ofHIV- I aredeterminedbyspecificregions oftheenvelopegp120 gene. Nature(Lond.).349:167-169.

31. Nara,P.L.,R.Garrity,A.Minassian,and J.Goudsmit. 1991.Envelope plasticity outsidetheV3 andgp 120-CD4binding epitopesresults inresistanceto neutralizationandfusion withoutlossofantibody binding: implicationsto HIV-1vaccinedesign.FASEB(Fed.Am.Soc.Exp. Biol.).J.5:175-181.

32. Westervelt, P.,D. P.Trowbridge,L.G.Epstein,B. M.Blumber,Y.Li, B.H.Hahn, G.M.Shaw,R. W.Price,and L. Ratner.1992. Macrophage tropic determinants ofhumanimmunodeficiencyvirus type1.J. Virol.66:2577-2582.

33.Sharpless, N. E., W.A.O'Brien,E.Verdin, C.V.Kufta,S.Irvin,Y.Chen, and M. Dubois-Dalco. 1992. Human immunodeficiency virus typeItropismfor brainmicroglialcellsis determined byaregion of theenvglycoproteinthat also controlsmacrophage tropism.J.Virol. 66:2588-2593.

34. Fung, M. S. C., C. Sun,W. L.Gordon, R.-S. Liou, T. W. Chang,W.Sun, E.S.Daar, andD. D.Ho. 1992. Identification and characterization ofa neutral-ization site withinthe secondvariable region ofhumanimmunodeficiency virus type 1gpl20. J. Virol. 66:848-856.

35.Sullivan, N., M. Thalix, C.Furman, D. D. Ho,and J.Sodroski. 1993. Effect of the amino and changes intheVl /V2 region of thehuman immunodefi-ciency virustype I gp120glycoproteinonsubunit association, syncytium forma-tion, andrecognitionbyaneutralizing antibody. J. Virol. 67:3674-3679.

36.Groenink, M.,R. A. M.Fouchier, S. Broersen, C.H. Baker,M.Koot, A.B. van'tWout, H. G. Huisman, F. Miedema, M. Tersmette, andH. Schuite-maker.1993. Relation of phenotype evolution ofHIV-I toenvelope V2 configu-ration. Science(Wash.DC).260:1513-1516.

37.Goodenow, M.,T.Huet, W. Saurin, S. Kwok, J. Sninsky, and S. Wain-Hobson. 1989.HIV- Iisolatesarerapidly evolving quasispecies: evidence for viral mixtures and preferred nucleotide substitutions. J. AcquiredImmuneDefic. Syndr.2:344-352.

38.Simmonds, P.,L.Q. Zhang,F.McOmish,P.Balfe,C.A.Ludlam, and A.J. LeighBrown.1991.Discontinuous sequence change of human immunodefi-ciency virus (HIV)type I envsequences in plasma viral and

lymphocyte-asso-ciatedproviral populations in vivo: implications for models ofHIVpathogenesis.

J.Virol. 65:6266-6276.

39. Wolfs,T. F. W., J.-J de Jong, H. vandenBerg, J. M.G.H. Tijnagel, W. J. A.Krone,and J.Goudsmit. 1990. Evolution of sequences encodingthe principal neutralization epitope of human immunodeficiency virus1is host

de-pendent, rapid,andcontinuous.Proc.Natl. Acad.Sci. USA.87:9938-9942.

40. Burger, H.,B.Weiser,K.Flaherty, J. Gulla,P.-NNguyen, andR.Gibbs. 1991. Evolution of human immunodeficiency virustype I nucleotide sequence

diversityamongclosecontacts. Proc.Natl.AcadSci. USA.88:11236-11240. 41.Ou, C.-Y., C.A.Ciesielski, G. Myers, C. I. Bandea, C-C. Luo,B. T. M. Korber, J.I.Mullins, G. Schochetman,R. L.Berkelman,A.N.Economous,etal. and theLaboratory Investigative Group, Epidemiologic InvestigationGroup. 1992.MolecularepidemiologyofHIVtransmission inadentalpractice.Science

(Wash.DC). 256:1165-1171.

42.Wolinsky, S., C.M.Wike,B.T. M.Korber, C. Hutto,W. P.Parks,L. L. Rosenblum,K.J. Kunstman,M.R.Furtado,and J. L.Munoz. 1992. Selective

transmissionof humanimmunodeficiencyvirustype Ivariantsfrom mothersto

infants. Science(Wash.DC).255:1134-1137.

43. Scarlatti, G., T. Leitner, E. Halap,J. Wahlberg, P. Marchisio, M. A. Clericii-Schoeller, H.Wigzel,E. M.Fenyo, J. Albert,M.Uhlen, andP.Rossi. 1993.Comparisonof variableregion 3 sequences of humanimmunodeficiency

virus typeIfrom infected children with theRNAandDNAof the virus popula-tions of their mothers.Proc.Natl. Acad.Sci. USA. 90:1721-1725.

44.Center for Disease Control. 1987. Classification system forhuman immu-nodeficiency virus (HIV) infection in children under 13 yearsofage. Morb. Mortal. Wkly Rep. 36:833-845.

45.Sambrook,J.,E. F.Fritsch,and T.Maniatis. 1989. MolecularCloning:A LaboratoryManual. ColdSpring Harbor Laboratory, Cold Spring Harbor,NY. 46.Leonard, C. K.,M. W.Spellman,L.Riddle,R. J.Harris,J.N.Thomas, andT. J.Gregory. 1990.Assignmentof intrachain disulfide bonds and character-izationofpotential glycosylationsitesofthe typeIrecombinanthuman

immuno-deficiency virusenvelopeglycoprotein (gp120)expressed in Chinesehamster ovarycells.J.Biol. Chem.265:10373-10382.

47.Nei, M., andL.Jin. 1989. Variances ofthe averagenumbers of nucleotide substitutions within and betweenpopulations.Mol.Biol. Evol. 6:290-300.

48.She,J.X.,S.Boehme,T. W.Wang,F.Bonhomme,andE. K.Wakeland. 1991.Amplificationofmajor histocompatibility complexclassII genediversity

by intra-exonic recombination.Proc.Natl.Acad.Sci. USA. 88:453-457. 49.Goedert,J.J.,A.-M.Duliege,C. I. Amos, S. Felton,R.J.Biggar,and The InternationalRegistryof HIV-exposed Twins. 1991.Highrisk ofHIV- Iinfection for first-born twins. Lancet. 338:1471-1475.

50.Scott, G. B., C. Hutto,R. W. Makuch,M.T.Mastrucci, T.O'Connor,

C.D.Mitchell,E.J.Trapido,and W. P. Parks. 1988. Survival in children with

perinatally acquiredhumanimmunodeficiencyvirustype Iinfection.N.Eng/.J. Med.321:1791-1796.

51. Holmes, E.C., L. Q. Zhang, P.Simmonds, C. A. Ludlam,and A. J. Leigh-Brown. 1992. Convergent anddivergentsequence evolutionin thesurface envelopeglycoproteinofhumanimmunodeficiencyvirustype Iwithinasingle

infectedpatient.Proc.Natl. Acad. Sci. USA. 89:4835-8439.

52. Lamers,S.L.,J. W.Sleasman,J. X.She,K. A.Barrie, S.M.Pomeroy, D. J.Barrett,and M. M.Goodenow. 1993.Independentvariationandpositive

selection inenvVI and V2 domains within maternal-infant strains ofhuman immunodeficiency virustype I in vivo.J. Virol. 67:3951-3960.

53.Kusumi, K.,B.Conway, S.Cunningham,A.Berson, C.Evans,A. K. N. Iversen,D.Colvin,M. V.Gallo,S.Coutre,E.G.Shpaer,etal. 1992.Human immunodeficiencyvirustype 1envelopegene structureanddiversityin vivo and after cocultivationin vitro.J. Viro/.66:875-885.

54.Pedroza Martins,L.,N.Chenciner,andS.Wain-Hobson.1992.Complex intrapatientsequencevariationin the VI and V2hypervariableregions of the HIV-1gp120 envelope sequence. Virology. 191:837-845.

55.Michael, N. L., G. Chang, P. K. Ehrenberg, M. T. Vahey, and R. R. Redfield. 1993. HIV-1 proviralgenotypes from theperipheralblood mononu-clear cells ofaninfected patientaredifferentially represented in expressed se-quences.J.Acquired ImmuneDefic.Syndr.6:1073-1085.

56.Cichutek, K.,H.Merget, S.Norley,R.Linde, W.Kreuz, M.Gahr, andR. Kurth. 1992. Development ofaquasispecies of human immunodeficiency virus type1invivo.Proc.Natl.Acad.Sci. USA. 89:7365-7369.

57.Saag,M.S.,M. J.Crain, W.D.Decker, S. Campbell-Hill, S. Robinson, W. E.Brown,M.Leuther,R.J.Whitley,B. H. Hahan, andG.M.Shaw. 1991. High-level viremia in adults and children infected with human immunodefi-ciency virus: relationtodisease and CD4+ lymphocytelevels. J. Infect.Dis.

164:72-80.

58.Piatak, M.,M.S.Saag,L.C. Yang, S.J.Clark, J. C. Kappes, K.-C. Luk, B. H.Hahn,G. M. Shaw, and J. D. Lifson. 1993. High levels of HIV-Iin plasma during allstagesof infection determined by competitive PCR. Science (Wash. DC). 259:1749-1754.

59. Pang, S.,Y.Shlesinger,E.S. Daar,T.Moudgil,D. D. Ho,andI. S.Y. Chen. 1992.Rapid generation ofsequencevariationduring primaryHIV- I infec-tion. AIDS. 6:453-460.

60.Delassus,S.,R.Cheynier, and S. Wain-Hobson. 1991. Evolution of hu-manimmunodeficiency virustype Inefand long terminalrepeat sequences over 4 yearsinvivoandin vitro.J. Virol. 65:225-231.

61. PedrozaMartins, L.,N.Chenciner,B.Asjo,A.Meyerhans, and S.

Wain-Hobson. 1991.Independent fluctuations ofHIV- 1 revand gp41 quasispeciesin vivo.J.Virol. 65:4502-4507.

62.Meyerhans,A., R.Cheynier, J. Albert,M.Seth, S. Kwok, J. Sninsky,L. Morfeldt-Manson,B.Ajso, and S. Wain-Hobson. 1989. Temporal fluctuations in HIVquasispecies in vivoare notreflected by sequential HIV isolations.Cell.

58:901-9 10.

63.Simmonds, P.,L.Q. Zhang,F.McOmish,P.Balfe, C.A.Ludlam, and A.J. LeighBrown.1991. Discontinuoussequencechanges of human immunode-ficiencyvirus (HIV)type 1 envsequences inplasma viral and lymphocyte-asso-ciatedproviral populationsin vivo:implications for models ofHIVpathogenesis.

J. Virol. 65:6266-6276.

64. Delassus,S., R. Cheynier, and S. Wain-Hobson. 1992. Inhomogeneous distribution of human immunodeficiency virus (HIV)type 1 genomeswithinan infected spleen. J. Virol. 66:5642-5645.

65. Pantaleo,G., C. Graziosi, J.F.Demarest,L.Butini,M.Montroni, C. H. Fox, J.M.Orenstein,D. P.Kolter, andA.S.Fauci.1993.HIVinfection is active andprogressive inlymphoidtissueduring theclinicallylatentstageof disease. Nature(Lond.). 362:355-358.

66.Embretson, J.,M.Zupanicic,J. L.Ribas,A.Burke,P.Racz,K. Tenner-Racz, andA.T. Haase. 1993. Massivecovertinfection of helperTlymphocytes andmacrophages byHIVduring theincubationperiodofAIDS. Nature(Lond.).

362:359-362.