Download Consistent risk group-associated differences in human

JOurnal o f General Virology (1996), 77, 783-792.
783
Printed in Great Britain
Consistent risk group-associated differences in human immunodeficiency
virus type 1 vpr, vpu and V3 sequences despite independent evolution
Carla L. Kuiken, 1. Marion T. E. Cornelissen, ~ Fokla Zorgdrager, 1 Susan Hartman, 1
Adrian J. Gibbs 2 and Jaap Goudsmit I
1 Human Retrovirus Laboratory, Department o f Human Retrovirology, University o f Amsterdam, Meibergdreef 15,
1105 A Z Amsterdam, The Netherlands and ~Research School o f Biological Sciences, Australian National University,
Canberra, Australia
Human immunodeficiency virus type 1 vpr, vpu and V3
sequences from 15 homosexual men and 19 intravenous
drug users in the Amsterdam Cohort studies were
analysed. Previously, we reported that V3 domains of
viruses from drug users are distinguishable from those of
homosexual men on the basis of two silent mutations.
Phylogenetic analysis of vpr, vpu and V3 shows that
differences in all three regions correlate with risk group.
Two positions in both vpr and vpu were found to differ
significantly between the risk groups. The distinguishing
positions were confirmed for sequences from 11 Scottish
and four German samples. The three regions show
relatively independent evolution patterns; they resulted
in different phylogenies, the only stable clustering being
that based on the risk group distinction. Pairwise
differences between sequences of the genes were moderately correlated (around 0.30). Surprisingly, when only
silent changes were counted, the correlations dropped
almost to zero, indicating that the evolution towards
independence was more advanced in the silent than in
the non-silent positions. This suggests that selection at
the amino acid level is not the primary driving force for
the independent evolutionary behaviour of the genes.
Recombination, combined with restrictions on certain
amino acids because of epistatic interactions between
the genes, could be an alternative explanation of this
phenomenon.
Introduction
arisen from a duplication of the vpr gene. Large
quantities of Vpr protein are packaged in the virion,
suggesting that it may have a role in the early stages of
infection, before integration of the DNA (Yuan et al.,
1990). Several studies have shown that the protein is
involved in virus production in macrophages, but not (or
less pronounced) in T cells (Balliet et al., 1994; Ogawa et
al., 1989; Cohen et al., 1990; Yuan et aI., 1990; Connor
et al., 1995; Balotta et al., 1994). Vpr has the same
function in HIV-2 (Hattori et al., 1990). Recent evidence
indicates that it has a role in the regulation of latent
infection (Levy et al., 1994) through the control of
activation of both CD4 T cells and macrophages (Levy et
al., 1993).
The vpu gene has no known homologue in HIV-2 and
SIV. Vpu protein is synthesized in infected cells, but not
packaged in the virion (Strebel et at., 1989). Available
evidence indicates that Vpu is mainly associated with
particle maturation and budding (Strebel et al., 1989;
Klimkait et al., 1990; Terwilliger et al., 1989). It has also
been shown that independent of this mechanism, Vpu is
involved in CD4 degradation (Willey et at., 1994; Chen
et al., 1993; Raja et at., 1994; Buonocore et al., 1994),
which in turn affects HIV-1 particle production because
vpr and ~7~u are two of the accessory genes of human
imnmnodeficiency virus type 1 (HIV-1). They are located
just 5' of the envelope gene, with vpu partially overlapping it. The proteins are dispensable in vitro, since
laboratory strains can be defective for either gene
(Dedera et al., 1989; Cohen et al., 1988). However, in
vivo both are usually intact, and probably have a role in
the virus life cycle (Westervelt et aL, 1992; Klimkait et
al., 1990).
The vpr gene encodes a 96 amino acid protein. It is also
found in HIV-2 and some of the simian immunodeficiency virus (SIV) isolates known. Moreover,
similarities found between the vpr and vpx genes in the
HIV-SIV group suggest that Wx may actually have
* Author for correspondence. Fax +31 20 691 6531.
e-mail [email protected]
The sequence data reported in this paper have been deposited
in GenBank under accession numbers Z29280, Z29296, Z29298,
Z29299, Z29302-Z29304, Z29306, Z29307, Z29310-Z29312, Z29315,
Z29319-Z29324, Z68033, Z68047, Z68062, Z68070, Z68074, Z68086,
Z68508-Z68616 and Z68687-Z68693.
0001-3630 © 1996 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
784
C. L. Kuiken and others
CD4 molecules capture gp160 envelope precursors on
their way to the cell surface.
A study of V3 sequences from a group of HIV-1
infected people in the Amsterdam Cohorts of homosexuals and intravenous drug users has revealed that the
virus variants that were found in the homosexual men
and haemophiliacs could be distinguished by two
mutations from variants found in drug users (Kuiken et
al., 1993). These mutations were both silent, and one of
them was found with very high consistency. Subsequently, in a collection of over 100 homosexual seroconversion sequences the 'drug user variant' was found
only once (Kuiken et aI., 1996). In a second study, using
only published V3 sequences from homosexuals and drug
users (and their sexual partners) from other countries in
Europe, and judging by the most conserved distinctive
mutation drug users and homosexuals from Germany
(Hamburg) and the United Kingdom (Scotland:
Edinburgh and surrounding regions) could also be
reliably distinguished on the basis of the V3 sequences.
Sequences obtained from samples from Italian drug
users indicated that the epidemic there consists of a
mixture of the two virus variants (Kuiken & Goudsmit,
1994).
As the most consistent mutations were silent, we
assumed that the distinction was epidemiological in
origin, and represented a founder effect: the drug user
population appears to have become infected by one
accidental variant, which differs from the other variants
by a number of mutations. To further investigate this
hypothesis, we have sequenced other regions of the HIV1 genome.
In this paper we report the results of an analysis of the
vpr and vpu sequences fi'om 34 HIV-1 infected
participants in the Amsterdam Cohorts of homosexuals
and drug users, supplemented with those from 11
Scottish and four German samples. The data show that
the distinction between the risk groups is present in both
vpr and vpu.
Methods
Origins ofsera. For sequencing of vpr, vpu and V3, serum samples
were obtained from 15 homosexual men and 19 intravenous drug users,
all participants in the Amsterdam Cohorts (Van den Hock et al., 1988;
De Wolf et al., 1987). All drug user samples were obtained 3 months to
1 year after seroconversion, and form a random selection from all
participants who seroconverted between 1986 and 1991, Samples from
homosexuals were either seroconversion samples (seroconversion dates
between 1985 and 1988) or the first seropositive sample for those who
were already HIV infected when they entered the Cohort study
(sampling date 1984 or 1985).
In addition, four German samples [two from homosexuals, two from
drug users; a subset of those previously discussed in Kuiken &
Goudsmit (1994)] and 11 Scottish samples [five from homosexuals and
six drug users and sexual contacts of these; unrelated to the samples
discussed in Kuiken & Goudsmit (1994)] were used.
R N A isolation and sequencing. Genomic RNA was isolated from
100 gl of sermn or plasma according to Boom et aL (1990). The RNA
was converted to cDNA, which was subjected to a nested PCR. For V3,
the method was as described previously (De Wolf et al., 1994). Primers
used for vpr were 5'VPR-1 (5' GATCTCTACAITACTTGGCACT 3')
(I = inosine) and YVPR-4 (5' CTTCTTCCTGCCATAGGAGATGCC Y), final MgCI~ concentration 2-4 m~, for the first PCR, and
5'VPR-2-Sp6 (5" ATTTAGGTGACACTATAGCACCTTTGCCTAGTGTTAIGAA 3') and YVPR-3-T7 (51 TAATACGACTCACTATAGGGAAAGCAACACTTTTTACAATAGCA 3'), both at a final
MgClz concentration of 2.4 mM, for the second PCR. For vpu the
primer sets were 5'VPU-1 (5' GCATCTCCTATGGCAGGAAGAAG
3') and YVPU-4 (5' ATATGCTTTAGCATCTGATGCACAAAATA
3') for the first PCR (final MgCI~ concentration 3.0 mM) and 5'VPU-2Sp6 (5' ATTTAGGTGACACTATAGTCATCAAGITICTCTAICAAAGC Y) and YVPU-3-T7 (51 TAATACGACTCACTATAGGGCCATAAT1ACIGTGACCCACA Y) (final MgCI 2 concentration
3-5 mM) for the second PCR. In all cases, sequences were rejected if they
contained more than 1% heterogeneous or illegible positions after two
sequencing attempts. A minimum of six clones was then made by
cloning the PCR product in an AT cloning system. In three Scottish
samples, two subpopulations were clearly present; in these three cases,
one subpopulation was randomly selected after it was established that
the relevant nucleotide positions were the same in both populations.
Sequencing was done using an automated sequencer (.Applied Bitsystems).
Sequence analysis. Alignment was done by eye, keeping codons
intact. Phylogenetic analysis was done using the program MEGA
(Kumar et al., 1994). The neighbour-joining method (Saitou & Nei,
1987) was used to construct phylogenetic trees based on distances
calculated using the Kimura 2-parameter estimation method (Kimura,
1980). The variability within sets of sequences was calculated using
Hamming distances between all pairs of sequences (Hamming, 1986),
expressed as percentages of the total sequence length (including gaps).
These were averaged over all sequences in a set to calculate the
variation. Pearson correlation coefficients between the Hamming
distances on two genes/regions were used as a measure of co-variation.
Positions containing gaps or ambiguities were excluded from all
analyses.
Additional multivariate analysis of the sequences was done using the
program PCOORD (Higgins, 1992). Numbers of silent and non-silent
changes were calculated according to Nei & Gojobori (1986) using
MEGA. For these calculations, the vpr and vpu sequences were
truncated to span only the open reading frames of the genes.
Assessing the strength o f risk group association. To evaluate the
strength of the association between the positions and the risk groups,
the following method was used. For each position in the sequences, a
pairwise comparison matrix was constructed consisting of zeros and
ones, indicating for each pair of sequences whether they contained
identical or different nucleotides at that position (position matrix). A
second matrix contained a similar same/different index indicating
whether each pair of sequences came from the same or a different risk
group (risk group matrix). A Pearson correlation coefficient was then
calculated between corresponding cells of the position matrix and
the risk group matrix, yielding a position-wise indication of the
association between sequence differences and group membership.
The correlation coefficient, in the present case, was calculated on the
pairwise differences, which are not statistically independent. Although
this does not alter its interpretation, the associated significance test
loses its vMidity. Instead, a simulation method was employed to assess
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
Risk group differences in HIV-1 vpr and vpu
the correlation that one can expect to find accidentally, when no real
association is present. The sequences were randomly divided into two
groups 100 times, and for all positions in all three regions the
correlation was calculated; these never rose above 0.33. Therefore
actual correlations larger than 0.33 were considered significant. The
descriptive level for the resulting criterion is difficult to calculate due to
dependencies in the data, but the criterion is considerably stricter than
1%. Because the criterion is so conservative, difficulties due to multiple
testing are avoided.
the groups was seen (drug user: mean 6.4%, s o 1.7;
homosexuals: mean 6.0%, so 1.5). For the drug user
group the V3 variation was 8.6 % (SD 2"3) versus 6"5 %
(SD 1"9) in the homosexual group. In the drug user group,
there were two samples that contained very similar virus
variants (I3072 and I5038 : < 1% difference in all genes).
An epidemiological link between these persons could not
be established, but the samples may have been a few
transmissions apart. Ratios of silent to non-silent
changes (calculated per number of silent/nonsilent sites)
were 1.12 for V3, 3"3 for vpr and 1"8 for vpu. N o n e of the
sequences contained any mutations that generated stop
codons.
To investigate the evolutionary differences between
the genes, we looked at the extent to which differences
between viruses in one gene or region were associated
Results
Independent variation o f the genes
In both vpr and vpu, variation was lower than in V3. In
vpr, the variation in the drug user group (mean 3"5 %,
SD 1"3) was slightly lower than that in the homosexual
group (mean 4.3 %, SD 1"3). In vpu, no difference between
3<
(a)
17
16
.I0033
,I3077
.Ii164
,-I0142
,I0090
.I5032
Ii035
• Ii035
.Ii083
3
39=
-I3008
27
14
-" 15038
,I3007
.I7025
,I3064
--Jl
,I5037
I0221
H0617
92[
H0569
H0157
H0337
H0583
9911
.......13057-09
,I0070-07
H0172
H0230
H0186
H0211
H0537
36[
HI145
19
H0008
H0411
18~
'
785
H0140
471
H0709
Fig. 1 (a). For legend see page 787.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
786
C. L. Kuiken and others
(b)
.13072
38
15038
. . . . •I 0 2 2 1
17025
~
I
0
1
4
1
,I009(
3
I0070-07
2O
,'~I0142
.I3008
1
.
3]
.
.
.
15032
.
,
,
,I3007
IA
~ 1 5 0 3 7
34 ¸
',I0033
,11035
,II083
,Ii164
,I3064
]J
2vii
,':,I3077
HI145
.I3057-09
H0537
H0230
j
H0008
3,5
H0617
H0411
H0140
H0186
18
l[
...... H 0 5 8 3
....
3
,
H0569
H0157
52[
H0172
H0709
HO21i
581
H0337
F i g . 1 (b). F o r l e g e n d s e e f a c i n g p a g e ,
with differences in another. This co-variation was
calculated by correlating the numbers of pairwise
differences between the sequences of one gene with those
between the corresponding sequences of the other. The
genes do not co-vary strongly: the Pearson correlation
coefficient between the pairwise distances in vpr and vpu
was 0-36; between vpu and V3, 0'31 ; and between vpr and
V3, 0.32. Correlations between the numbers of silent
changes were 0.05 (vpr and vpu), 0.20 (vpr and V3) and
0-07 (vpu and V3). These numbers are much lower, which
is surprising because the numbers of silent changes
presumably reflect the evolutionary distances more
accurately, being less subject to various pressures.
[However, silent changes are not always selectively
neutral (e.g. Coffin, 1992).] Absolute numbers of silent
changes were large enough for meaningful comparison
(vpr: median 4, range 0-11 ; vpu: median 3, range 0-9;
V3: median4.5, range 0--13.5). The significance of the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
Risk group differences in HIV-1 vpr and vpu
(c)
787
.I3008
.I0142
,I5037
.I5032
~17025
-I3007
Ii035
,I0221
,I0141
05
]~'~I5038
I00]L---------~13072
I0033
....I3064
,I3077
10090
.Ii164
.Ii083
96].
~I
301' '
,I3057
....~I0070
.
H0211
H0569
2~
L~39
H0157
H0583
H0008
,,, H0617
02
H0186
,'
HI145
H0230
H0172
H0337
H0140
H0411
H0537
H0709
Fig. 1. Phylogenetic trees based on the vpr (a), ~7'u (b) and V3 (e) regions. Double lines indicate branches leading to HIV-1 sequences
from intravenous drug users, single lines those leading to HIV-1 sequences from homosexual men.
difference between the correlations of the numbers of
silent changes could not be directly tested, however,
because again they are based on pairwise comparisons. A
very conservative estimate (using Fisher's r-to-Z transformation with N set equal to the number of sequences)
indicates that the 95 % confidence interval around the
0-20 correlation has a width of 0-24, making the difference
between 0-20 and 0-05 non-significant.
This phenomenon of relatively independent variation
is also reflected in the phylogenetic analyses (Fig. 1).
Analyses based on the three genes yielded different
phylogenies. Only the I3072/I5038 pair consistently
clustered together. Several small clusters with moderate
to high bootstrap values (> 0.60) formed in phylogenetic
analyses based on all three genes [e.g. H0008/H0617
(0-66) in ~Tu, and H0569/H0617 (0.92) and 10090/15032
(0-96) in vpr]. Some of these clustered pairs appeared in
more than one of the trees, but none was present in all
(except the related viruses mentioned above). Bootstrap
values in the analyses based on individual genes were
generally low. This instability was reflected in an analysis
based on all three genes together, which yielded a tree
with three stable groupings (not shown): one with the
related drug user sequences (100%), one with the two
drug user sequences with homosexual characteristics (see
below, 100 %), and the separation between the two risk
groups (86 %). Two drug users who had sequences with
'homosexual' characteristics formed a separate subcluster within the homosexual cluster in all trees except
the one based on vpu sequences.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
788
C. L. Kuiken and others
relative
position
.........
i0
[ .........
20
[ .........
30
[ .........
40
[ .........
50
[ .........
60
[ .........
70
[ .........
80
~ .........
90
[
vpr
vpr
homo
IVDU
v p r ->
<- v l f
ACTGACAGAGGATAGATGGAACAAGCCCCAGAAGACCAAGGGCCACAGAGGGAGCCATACA%ATGAATGGACACTAGAG~Fz-i-i~AGAGGAG
...............................................
A ..................... G ....................
T/A
vpr
vpr
homo
IVDU
CTTAAGAATGAAG CTG~AGACA%~FF~CCTAGGATATGGCTCCATAGCTTAGGGCAACATATCTATGAAACTT~ATACTTGGGCA
..........................................................................................
IS0
vpr
vpr
homo
IVDU
GGAGTGGAAG C CATAATAAGAATTCTG CAACAACTG CTGTTTATTCA TTTCAGAATTGGGTGTCGACATAGCAG~TAGGCATTACTCAA
................................................................
A .........................
270
vpr
vpr
homo
IVDU
t a t ->
<- v p r
CAGAGGAGAG CAAGAAATGGAGCCAGTAGATCCTAGACTAGAGCCCTGGAAGCATCCAGGAAGTCAGCCTAAGACTGCTT~TACCAC
-G .....................................................................................
R/e
357
Q/R
relative
positio. . . . . . . . . .
i0
I .........
20
t .........
30
I .........
40
1 .........
50
I .........
60
r .........
70
t .........
80
L .........
90
I
vpu
vpu
homo
IVDU
vpu
- >
TAAGTAGTACATGTAATGCAAT~-i-i~i~%CAAATATTAGCAATAGTAGCATTAGTAGTAGCAACAATAATAGCAATAGTTGTGTGGTCCATA
..........................................................................................
vpu
vpu
homo
IVDU
env
G T A T T C A ~ A G A C A G G T T A A T T G A T A G A A T A A G A G A A A G A G C A G A A G A C A G T G G C A A T
..........................................................................................
vpu
vpu
homo
IVDU
<- V p U
GAGAGTGAAGGAGATCAGGAAGAATTATCAGCACTTGTGGAGA~CA~CATGCTCCTTGGGATATTGATGATCTGTAGTGCTGCAGAAAAATT
........ C ................................
A ........................... A .......... A ............
E/D
.........
i0
I .........
180
R/K
20
[ .........
30
~ .........
*
40
I .........
S/N*
50
{ .........
60
I .........
70
[ .........
80
[ .........
K/N*
90
]
V3 h o m o
V3 IVDU
GTAGTAATTAGATCTGACAATTTCA
CGGA CAATG CTAAAA CCATAATAGTACAG CTGAATCAATCTGTAGAAATTAATTGTACAAGACCC
.............. C .........................
T .................. C ..............................
r/l
K/N
V3 homo
V3 IVDU
A~CAACAATACAAGAAAAAGTATACATATAGGACCAGGGAGAGCA~F~-~~GGAGAAATAATAGGAGATATAAGACAAGCACAT
......................................
C ...................................................
V3
V3
TGTAAC CTTAGTAGAG CAAAATGGAATAACACTTTAAAACAGATAGTTAAAAAATTAAGAGAACAA~AATAAAACAATAGTCTTTAATCAA
...........................
G ..................... T ..............................................
N/D
K/I
homo
IVDU
276
T--
DIN
K/T*
relative
position
->
180
276
Fig. 2. vpr, ~Tu and V3 regions (based on the consensus sequence) depicting the mutations characteristic of homosexual and drug user
virus populations. Letters underneath the alignments indicate amino acid differences that result from the mutations: in X/Y, X
indicates the amino acid associatedwith the homosexualvariant, Y the one in the drug user group. If no translation is given, the change
is silent. In the vpu sequence, translations based on both Epuand env reading frames are shown. The numberingof the regions is relative
to the starting point of the sequences used in this study. * Amino acid changes in the Env reading frame.
Risk-group related differences between the sequences
Phylogenetic analysis clearly separated the two risk
groups in trees based on each of three regions (Fig. 1).
The drug user group contained two samples that yielded
virus which did not conform to the risk group classification (I3057 and I0070). These two samples mostly
clustered together and were grouped with the homosexual samples. The two groups could also be distinguished on the basis o f each region using principal
coordinate analysis (Higgins, 1992), again suggesting
that some positions differed consistently between the two
groups.
To evaluate the strength of the association for each
position, Pearson correlations were used (see Methods).
As noted previously for V3, associations varied from
almost perfect to very weak. In Table 1, all correlations
over 0-2 are included. In the Dutch group, correlations
that are significant according to the criterion defined
above are marked with an asterisk. In both vpr, vpu and
V3, two positions were found to be significantly
associated with risk group. In V3, both changes were
silent. In ~7~r, one of the two changes was non-silent
(Q/R). In vpu, both significant changes were silent in the
vpu reading frame (one was located in the stop codon).
They are both located in the section that also encodes
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
Risk group differences in HIV-1 vpr and vpu
Table 1. Pearson correlations between nucleotides
present at each position and risk group membership for
selected positions in vpr, vpu and V3
Analysis of Dutch vpr and vpu genes was based on a smaller
sample of 19 intravenous drug users and 15 homosexuals. Foreign
samples consist of four German (two homosexual, two drug users)
and 11 Scottish samples (fivehomosexuals and six drug users and
sexual contacts of these).
Position
Dutch samples
(n = 34)
Foreignsamples
(n = 15)
V3-15
V3-41
V3-60
V3-129
V3-208
V3-230
vpr-48
vpr-70
vpr-245
vpr-272
vpu-189
vpu-222
vpu-250
vpu-261
vpu-274
0.36*
0.33
0.33
0.73*
0.33
0.27
0.77*
0.20
0"33
0'77*
0"30
0"71*
0"33
0"48*
0"30
0.36
0.30
0.22
0.74
0"15
0.22
0.69
0"33
0-35
0.76
0"37
0"48
O"17
0"39
0"35
* Significant correlation (see Results).
Env; in the env reading frame, both changes were nonsilent (vpu-222, R / K ; vpu-261, S/N).
Other northern European samples
Previously, we reported that samples from G e r m a n y and
Scotland showed the same risk group-related distinction
as the Dutch ones. vpr and vpu sequences from 11
Scottish samples (five homosexuals and six drug users)
and four G e r m a n samples (two homosexuals, two drug
users) were analysed to assess whether the same pattern
was also found in these genes. One sample from a
Scottish homosexual was negative for vpr despite
repeated isolation attempts, so for vpr only four Scottish
homosexual samples were available.
Two Scottish samples from heterosexual women were
assigned to the drug user group, because the Scottish
epidemic is most severe a m o n g drug users (Brettle et al.,
1987; Peutherer et al., 1985), and because in northern
Europe the largest risk factor for women who do not use
intravenous drugs is sexual contact with a drug user.
Both samples yielded a V3 sequence that was unmistakably of the drug user type. Judging by the two
significant positions in the V3 region, one Scottish drug
user sample (IUK60), from a male drug user, yielded a
sequence with clear homosexual characteristics. All other
samples fell into the correct group based on V3 position
129 and vpr position 272. For vpr-48, one homosexual
sample deviated from the expected pattern; for V3-15,
two drug user samples had a different nucleotide than
789
expected. Consistency in vpu was lower: for both
positions (222 and 261), three of seven homosexual
samples adhered to the Dutch pattern. All drug user
samples except I U K 6 0 showed the expected pattern in
vpu-222, and five of these did so in vpu-261.
Discussion
The finding that the intravenous drug user population in
The Netherlands appeared to be infected with an H I V
variant that is distinct from the one found in homosexuals and haemophilia patients presumably reflects an
epidemiological founder effect: the ancestral virus
infecting the first drug user victim of the H I V epidemic
had several unique mutations which, being selectively
neutral or nearly so, remained present in the quasispecies
descending from it. The two mutations found in V3 were
both silent, and therefore presumably have little or no
effect on the fitness of the virus. Furthermore, haemophilia patients without exception carried the ' h o m o sexual' variant, which, for the Netherlands (where blood
donations are given voluntarily and without financial
compensation) is compatible with this epidemiological
model. Later on, the same distinction was found to be
present in other European countries as well, namely
G e r m a n y and Scotland (with high consistency) and Italy
(where four out of nine drug user samples contained the
' h o m o s e x u a l ' variant) (Kuiken & Goudsmit, 1994). It is
possible that all these drug user infections are epidemiologically related.
The V3-129 mutation is found in only four out of 47
sequences listed in the Los Alamos compendium alignment (Myers et al., 1994), all of these from the United
States: HIV-1 BAL, a macrophage-tropic isolate from a
New Y o r k patient, and three from intravenous drug
users from Baltimore (HIVUS712, -715 and -716). These
four exclusively also have vpu-261, and three have
V3-15. The mutation in vpu-222, however, is found in 13
of the sequences listed in the alignment, vpr sequences are
not available for these isolates. In a set of 30 V3
sequences from the United States (NIAID), three out of
four drug user sequences had the V3-129 mutation, as well
as two heterosexual transmission cases from Puerto Rico.
Other anecdotal evidence suggests that the Dutch drug
user variant is also associated with drug use in the United
States, though less consistently than in mid-Europe:
based on V3-129, it is present in three of 29 sequences
from control subjects in the Florida dentist case, one of
these from a drug user (Ciesielsky et al., 1992); it is found
in the only drug user in a set of three index persons from
New Y o r k City and their partners (Zhu et al., 1993), and
again in a drug user in a set of four United States patients
studied by Haigwood et al. (1993). In a set of 11 samples
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
790
C. L. Kuiken and others
from Baltimore drug users, five adhered to this pattern
(V. V. Lukashov and others, unpublished).
The present study was undertaken to investigate
whether a risk group-related distinction found in V3
could also be found in vpr and vpu. The results show that
this distinction is clearly present in both these genes. This
is consistent with a founder effect: accidental differences
present in the variant that was the founder o f the
epidemic in drug users would be scattered throughout
the genome. Most of the positions found to discriminate
between the risk groups in the Dutch dataset could be
confirmed independently using sequences from a Scottish
and a German sample set.
A distinction between viruses obtained from drug
users and homosexuals has been found in other studies as
well. A recent study by Albert et al. (1994) showed a clear
and consistent distinction in a set of reverse transcriptase
sequences. Our analysis of these sequences showed that
three positions in these reverse transcriptases are consistently different between the risk groups (not shown).
Analysis of a small set of 10 Swedish V3 sequences
suggests that, although again the sequences differ
between risk groups, this distinction is not the same one
that is found in the Dutch, Scottish and German viruses
(not shown). A Madrid study showed a distinction to be
present in gag as well as env (fusion domain) sequences
(Rojas et al., 1994), but again, based on V3 sequences,
this distinction seems to be different from the D u t c h /
German/Scottish distinction (C. Ldpez-Galindez, personal communication).
Combining all these results, we can conclude that it is
likely that several different founder effects are present in
the European HIV epidemic. The fact that the ' D u t c h
distinction' is also present in part of the United States
drug user population suggests that this particular variant
may have originated from the United States, and then
having crossed the Atlantic subsequently spread among
Dutch, German and Scottish, and some Italian drug
users. A sizeable proportion of Amsterdam drug users
are German or Italian; a study among Edinburgh drug
users showed that several of them had shared needles in
Amsterdam (Brettle et at., 1987). The source of the
homosexual infection could be a different one. At
present, data are too scarce to draw more inferences
about the origin of the other founder effects, or about the
ways in which the virus has spread through Europe. The
hypothesis that the northern European drug user variant
is derived from a United States variant is being
investigated at present.
The three regions investigated here appear to have
evolved more or less independently. The correlations
between pairwise differences were small, and the only
clustering that was reinforced when the sequences were
analysed together was the homosexual/drug user split.
This instability may be caused by the fact that the
sequences are rather short. Also, phylogenetic trees in
populations of closely related sequences typically have a
star shape, with short internal branches and long external
ones (Holmes et al., 1995); this results in very unstable
trees, where removal of one or two sequences often
results in a different structure.
Although the data have to be interpreted with care due
to the above considerations, alternative explanations can
be suggested. One is that the genes are evolving in
different directions because of different selection
pressures. However, this model is hard to reconcile with
the fact that the correlations between the pairwise
distances drop almost to zero when they are based on
only silent mutations. Although silent positions may also
be under some selective pressure in coding regions
(Coffin, 1992), it is hard to explain why they would be
m o r e diverged than non-silent ones. This finding may
reflect some form of epistatic interaction between the
genes, whereby certain amino acids encoded by one gene
necessitate or facilitate the presence of specific amino
acids encoded by the other.
Independent evolution in different genes has been
reported before in sequences from one patient (Pedroza
Martins et al., 1991 ; Delassus et al., 1991 ; Vartanian et
al., 1991; Howell et al., 1991), and frequent recombination events have been suggested as an explanation.
Increasingly, reports suggest that recombination occurs
frequently in HIV-1 (Monken et al., 1995; Diaz et al,,
1995; Zhu et al., 1995; Robertson et aI., 1995). It is
possible that the independent evolution found in the
present study is a reflection of recombination, where the
risk group distinction is preserved because the recombining viruses generally are from the same risk group and
share the same genotype. In this light, one might expect
that correlations between distances would be inversely
proportional to the distance between the genes (and thus,
to the probability of recombination). This does not
appear to be the case in the present dataset, but more
data are needed to draw this conclusion more firmly.
We would like to express our gratitude to Dr Michael Schreiberof
the BernhardNocht Institut ftir Tropenmedizinin Hamburg, Gemaany,
and to Dr Andrew Leigh Brown of the University of F~nburgh,
Scotland, for generouslydonating serum samples that were used for
this study; to Dr Edward Holmes of the University of Oxford, for
thoughtful comments; to Drs Jan Albert, Thomas Leitner and Cecilio
Ldpez-Galindez, for sharing unpublished results; and to Drs Tuofu
Zhu (Aaron Diamond Institute, New York) and Carol Ciesielsky
(Centers for Disease Control, Atlanta) for unearthing the risk groups
associated with published sequences.
References
ALBERT,J., WAHLBERG,J., LE1TNER,T., ESCANILLA,D. & UHLEN, M.
(1994), Analysis of a rape case by direct sequencing of the human
immunodeficiencyvirus type 1 pol and gag genes. Journal of Virology
68, 5918-5924,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
R i s k group differences in H I V - 1 vpr and W u
BALLIET,J. W., KOLSON,D. L., EIGER, G., KIM, F. M., MCGANN, K.
A., SRIN1VASAN,A. & COLLMAN,R. (1994). Distinct effect in primary
macrophages and lymphocytes of the human immunodeficiency
virus type l accessory genes vpr, vpu and nef: mutational analysis of
a primary HIV-I isolate. Virology 200, 623-631.
BALOTTA,C., LuSSO, P., CROWLEY,R., GALLO,R. C. & FRANCHINI,G.
(1994). Antisense phosphorothioate oligodeoxynucleotides targeted
to the vpr gene inhibit human immunodeficiency virus type 1
replication in primary human macrophages. Journal of Virology 67,
4409-4414.
BOOM, R., SOL, C.J.A., SALIMANS, M . M . M . , JANSEN, C. L.,
WERTHEIM-VANDILLEN, P. M. E. & VAN DER NOORDAA,J. (1990).
Rapid and simple method for purification of nucleic acids. Journal of
Clinical Microbiology" 28, 495-503.
BRETTLE, R. P., BtSSET,K., BURNS, S., DAVIDSON,J., DAVIDSON,S. J.,
GRAY, J. M. N., INCLIS,J. M., LEES,J. S. & MOK, J. (1987). Human
immunodeficiency virus and drug misuse: the Edinburgh experience.
British Medical Journal 295, 421-424.
BUONOCORE, L., TURI, T.G., CRISE, B. & ROSE, J.K. (1994).
Stimulation of heterologous protein degradation by the Vpu protein
of HIV-1 requires the transmembrane and cytoplasmic domains of
CD4. Virology 204, 482-486.
CHEN, M. Y., MALDARELLI,F., KARCZEWSKI,M. K., WILLEY,R. L. &
STREBEL, K. (1993). Human immunodeficiency virus type 1 Vpu
protein induces degradation of CD4 in vitro: the cytoplasmic
domain of CD4 contributes to Vpu sensitivity. Journal of Virology
67, 3877-3884.
CIESIELSKY, C . A , MARIANOS, D., OU, C . Y , DUMBAUGH, R.,
BERKELMAN,R., GOOCH,B., MYERS,G., LUO, C. C., SCHOCHETMAN,
G., HOWELL, J., LASCH, A., BELL, K., ECONOMOU,N., SCOTT, B.,
FURMAN,L , CtmRAN, J. & JArEE, H. (1992). Transmission of human
immunodeficiency virus in a dental practice. Annals of Internal
Medicine 166, 798-805.
COEEIN, J. M. (1992). Genetic diversity and evolution of retroviruses.
Current Topics in Microbiology and Immunology 176, 143-164.
COHEN, E. A., TERWILLIGER,E. F., SODROSKI,J, G. & HASELTINE,W.
A. (1988). Identification of a protein encoded by the vpu gene of
HIV-1. Nature 334, 532-534.
COHEN,E. A., TERWILLIGER,E. F., JALINOOS,Y., PROULX,J., SODROSKI,
J.G. 8¢ HASELTINq3, W.A. (1990). Identification of HIV-1 vpr
product and function. Journal of Acquired Immune Deficiency
Syndrome 3, 11-18.
CONNOR, R. 1., KUANCHEN,B., CHOE,S. t~ LANDAU,N. R. (1995). Vpr
is required for efficient regulation of human immunodeficiency virus
type 1 in mononuclear phagocytes. Virology 206, 935-944.
DEDERA,D., HU, W., VAN DER HEYDEN,N. & RATNER,L. (1989). Viral
protein R of human immunodeficiency virus types 1 and 2 is
dispensable for replication and cytopathogenicity in lymphoid cells.
Journal of Virology 63, 3205-3208.
DELASSUS,S., CHEYNIER,R. & WAIN-HoBsoN, S. (1991). Evolution of
human immunodeficiency virus type 1 nef and long terminal repeat
sequences over 4 years in vivo and in vitro. Journal of Virology 65,
1, 225-231.
DE WOLF, F., GOLq)SMIT,J., PAUL,D. A., LANGE,J. M. A., HOOYKAAS,
C., SCHELLEKENS,P. T. m., COUT1NHO,R. m. & VAN DER NOORDAA,
J. (1987). Risk of AIDS related complex and AIDS in homosexual
men with persistent HIV antigenaemia. British Medical Journal 295,
569-572.
DE WOLF, F., HOGERVORST, E., GOUDSMIT, J., FENY6, E.-M.,
R/.)BSAMEN-WAIGMANN, H., HOLMES, H., GALVAO-CASTRO, B.,
KARITA, E., WASI, C., SEMPALA,S. D. K., BAAN, E., ZORGDRAGER,
F., LUKASHOV,V. V., OSMANOV,S., KUIKEN,C. L., CORNELISSEN,M.
T, E. 8¢ THE WHO NETWORK ON H1V-1 ISOLATION& CHARACTERIZATION (1994). Syncytium-inducing (SI) and non-syncytiuminducing (NSI) capacity of human immunodeficiency virus type 1
(HIV-1) subtypes other than B: phenotypic and genotypic characteristics. AIDS Research and Human Retroviruses 10, 1387-1400.
DIAZ, R. S, SAmNO,E. C., MAYER,A., MOSL~Y,J. W. & BUSCH,M. P.
(1995). Dual human immunodeficiency virus type 1 infection and
recombination in a dually exposed transfusion recipient. Journal of
Virology 69, 3273-3281.
HAIGWOOD, N. L., MASON, O. B., YORK-HIGGINS,D , KORBER, B.T.
791
M., WE1SER, B. & BURGER, H. (1993). HIV-1 envelope sequence
diversity in long-term infected asymptomatic patients. Vaccines"
233-242.
HAMMING, R.W. (1986). Coding & h~brmation Theory, 2nd edn.
Englewood Cliffs: Prentice Hall.
HATTORI,N., MICHAELS,F., FARGNOLI,K., MARCON,L., GALLO,R. C.
& FRANCHINI,G. (1990). The human immunodeficiency virus type 2
vpr gene is essential for productive infection of human macrophages.
Proceedings of the National Academy of Sciences, USA 87,
8080-8084.
HI,GINS, D. G. (1992). Sequence ordinations: a multivariate analysis
approach to analysing large sequence data sets. CABIOS 8, 15-22.
HOLMES, E. C., NEE, S., RAMBAUT,A. R., GARNE1~, G. P. & HARX,q~Y,
P. H. (1995). Revealing the history of infectious disease epidemics
through phylogenetic trees. Philosophical Trtmsactions of the Royal
Society of London 349, 33~40.
HOWELL, R. M., FITZGIBBON,J. E., NOE, M., REN, Z., GOCKE, D. J.,
SCHWARTZER, T.A. t~; DUBIN, D.T. (1991). In vivo sequence
variation of the human immunodeficiency virus type 1 env gene:
evidence for recombination among variants found in a single
individual. AIDS Research and Human Retroviruses 7, 869-876.
KIMUgA,M. (1980). A simple method for estimating evolutionary rates
of base substitution through comparative studies of nucleotide
sequences. Journal of Molecular Evolution 16, 111-120.
KLIMKAIT, T., STREBEL, K., HOGGAN, M.D., MARTIN, M.A. &
ORENSTEIN,J. M. (1990). The human immunodeficiency virus type 1specific protein vpu is reqlaired for efficient virus maturation and
release. ,Journal of Virology 64, 621 629.
KU1KEN, C. L. • GOUDSMIT,J. (1994). Silent mutation pattern in V3
sequences distinguishes virus according to risk group in Europe.
AIDS Research and Human Retroviruses 10, 31%320.
KUIKEN, C.L., ZWART, G., BAAN, E., COUTINHO, R.A., VAN DEN
HOEH, J. A. R. & GOUDSMIT, J. (1993). Increasing antigenic and
genetic diversity of the HIV-1 V3 domain in the course of the AIDS
epidemic. Proceedingsof the National Academy of Sciences, USA 90,
9061-906L
KUIKEN, C. L., LUKASHOV,V. V., LEUNISSEN,J. A. M. & GOUDSMIT,J.
(1996). Restricted intra-patient evolution of genomic RNA encoding
the third variable (V3) domain of the HIV-1 envelope. AIDS 10,
31-37.
KUMAR, S., TAMLq/A, K. & NEI, M. (1994). MEGA: molecular
evolutionary genetics analysis software for microcomputers.
CABIOS 10, 189-192.
LEVY, D.N., FERNANDES,L.S., WILLIAMS,W.V. 8¢ WEINER, D.B.
(1993). Induction of cell differentiation by human immunodeficiency
type l vpr. Cell 72, 541-550.
LEVY,D. N., REFAELI,Y., MACGREGOR,R. R. & WEINER,D. B. (1994).
Serum vpr regulates productive infection and latency of human
immunodeficiency virus type 1. Proceedingsof the National Academy
of Sciences, USA 91, 10873-10877.
MONKEN, C.E., WU, B. & SRINIVASAN,A. (1995). High resolution
analysis of the HIV-1 quasispecies in the brain. AIDS 9, 345-349.
MYERS, G., WAIN-HOBSON,S., HENDERSON,L. E., KORBER,B., JEANG,
K.T. 8¢ PAVLAKIS, G. N. (1994). Human Retroviruses and AIDS
1994. Los Alamos, N. Mex.: Theoretical Biology and Biophysics
Group.
NEI, M. & GOJOBORI,T. (1986). Simple methods for estimating the
numbers of synonymous and nonsynonymous nucleotide substitutions. Molecular Biology and Evolution 3, 418-426.
OGAWA, K., SHIBATA,R., KIYOMASU,T., HIGUCHI, I., KISHIDA, Y.,
ISHIMOTO, A. & ADACHI, A. (1989). Mutational analysis of the
human immunodeficiency virus vpr open reading frame. Journal of
Virology 63, 4110-4114.
PEDROZA MARTINS,L., CHENCINER,N., fikSJO,B., MEYERHANS,A. &
WAIN-HOBSON,S. (1991). Independent fluctuation of human immunodeficiency virus type 1 rev and gp41 quasispecies in FIFO. Journal of
Virology 65, 4502-4507.
PEUTHERER,J. F., EDMOND,E., SIMMONDS,P., DICKSON,J. D. & BATH,
G. E. (1985). HTLV-III antibody in Edinburgh drug addicts. Lancet
ii, 1129-1130.
RmA, N. U., VINCENT,M. J. & JABBAR,M. A. (1994). Vpu-mediated
proteolysis of gpl60/CD4 chimeric envelope glycoproteins in the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11
792
C. L. Kuiken and others
endoplasmic reticulum: requirement of both the anchor and
cytoplasmic domain of CD4. Virology 204, 357-366.
ROBERTSON, D. L , HAHN, B.H. & SHARP, P.M. (1995). Recombination in AIDS viruses. Journal of Molecular Evolution 40,
24%259.
RO~AS, J.M., DOPAZO, J., NA3ERA, L, SANCHEZ-PALOMINO, S.,
OLIVARES,L, MARTIN,M. J., BERNAL,A., GARCIASAIZ, A., NAJERA,
R. & LOPEZ-GALINDEZ,C. (1994). Molecular epidemiology of HIV1 in Madrid. Virus Research 31, 331-.34_?.
SA:TOU, N. & Nt~I, M. (1987). The neighbor-joining method: a new
method for reconstructing phylogenetic trees. MolecularBiology and
Evolution 4, 406~25.
STREBEL,K., KLIMKAtT,T., MALDARELLI,F. & MARTIN,M. A. (1989).
Molecular and biochemical analyses of human immunodeficiency
virus type 1 vpu protein. Journal of Virology 63, 3784-3791.
TERWILrlGER, E.F., COHEN, E.A., LU, Y, SODROSKI, J.G. &
HASELTIN~0 W.A. (1989). Functional role of human immunodeficiency virus type 1 vpu. Proceedings of the National Academy of
Sciences, USA 86, 5163 5167.
VANDENHOEK,J. A. R., COUTINHO,R. A., VAN HAASTRECHT,H. J. A.,
VAN ZADELHOFF,A. W. ~¢ GOUDSMIT,J. (1988), Prevalence and risk
factors of HIV infections among drug users and drug-using
prostitutes in Amsterdam. AIDS 2, 55-60.
VARTANIAN, J.P., MEYERHANS, A., ]kSJ6, B. & WAIN-HOBSON, S.
(1991). Selection, recombination and G + A hypermutation of
human immunodeficiency virus type 1 genomes. Journal of Virology
65, 177%1788.
VqESTERVELT,P., HENKEL, T , TROWBRIDGE,D. B., ORENSTEIN,J. M.,
HEUSER,J., GENDEL~aN,H. E. & RATNER,L. (1992). Dual regulation
of silent and productive infection in monocytes by distinct hmnan
immunodeficiency virus type 1 determinants. Journal of Virology 66,
39257931.
WILLEY, R. L , BUCKLER-WroTE,A. & STREBEL,K. (1994). Sequences
present in the cytoplasmic domain of CD4 are necessary and
sufficient to confer sensitivity to the human immunodeficiency virus
type 1 Vpu protein. Journal of Virology 68, 1207-1212.
YUAN,X., MATSUDA,Z., MATSUDA,M., ESSEX,M. & LEE,T.-H. (1990).
Human immunodeficiency virus vpr gene encodes a virion-associated
protein. AIDS Research and Human Retroviruses 6, 1265-1271.
ZHU,T., Mo, H., WANG, N , NAM, D. S., CAO, Y., KouP, R. A. & Ho,
D. D. (1993). Genotypic and phenotypic characterization of HIV-1
in patients with primary infection. Science 261, 1179-1181.
ZHU, T., WANG, N., CARR, A., WOLINSKY, S. & Ho, D. D. (1995).
Evidence for coinfection by multiple strains of human immunodeficiency virus type 1 subtype B in an acute seroconvertor. Journal
of Virology 69, 1324-1327.
(Received 14 September 1995; Accepted 14 December 1995)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 11 May 2017 04:30:11