Genetic analysis of the human immunodeficiency virus type 1 integrase protein.

Abstract

Single-amino-acid changes in a highly conserved central region of the human immunodeficiency virus type 1 (HIV-1) integrase protein were analyzed for their effects on viral protein synthesis, virion morphogenesis, and viral replication. Alteration of two amino acids that are invariant among retroviral integrases, D116 and E152 of HIV-1, as well as a mutation of the highly conserved amino acid S147 blocked viral replication in two CD4+ human T-cell lines. Mutations of four other highly conserved amino acids in the region had no detectable effect on viral replication, whereas mutations at two positions, N117 and Y143, resulted in viruses with a delayed-replication phenotype. Defects in virion precursor polypeptide processing, virion morphology, or viral DNA synthesis were observed for all of the replication-defective mutants, indicating that changes in integrase can have pleiotropic effects on viral replication.

Full text

Page 1

JOURNAL OF VIROLOGY, Mar. 1994, P. 1633-1642
0022-538X/94/$04.00+0
Copyright © 1994, American Society for Microbiology
Genetic Analysis of the Human Immunodeficiency
Virus Type 1 Integrase Protein
CHA-GYUN SHIN,t BRUNELLA TADDEO,t WILLIAM A. HASELTINE,§ AND CHRIS M. FARNET*
Division ofHuman Retrovirology, Dana-Farber Cancer Institute, and Department ofPathology,
Harvard Medical School, Boston, Massachusetts 02115
Received 7 September 1993/Accepted 29 November 1993
Single-amino-acid changes in a highly conserved central region of the human immunodeficiency virus type
1 (HIV-1) integrase protein were analyzed for their effects on viral protein synthesis, virion morphogenesis, and
viral replication. Alteration of two amino acids that are invariant among retroviral integrases, D116 and E152
of HIV-1, as well as a mutation of the highly conserved amino acid S147 blocked viral replication in two CD4+
human T-cell lines. Mutations offour other highly conserved amino acids in the region had no detectable ellect
on viral replication, whereas mutations at two positions, N117 and Y143, resulted in viruses with a
delayed-replication phenotype. Defects in virion precursor polypeptide processing, virion morphology, or viral
DNA synthesis were observed for all of the replication-defective mutants, indicating that changes in integrase
can have pleiotropic effects on viral replication.
The integration of a DNA copy of the viral RNA genome
into the genome of the infected cell is an essential step in the
retroviral replication cycle (57). Genetic studies have shown
that two regions of the viral genome are essential for the
establishment of an integrated provirus: the ends of the linear
viral DNA molecule, including the sequences that are actually
joined to the host DNA target (7-9, 18, 42), and the 3' region
of the pol gene, encoding the integrase protein (6, 14, 15, 43,
46, 48, 50).
Recent in vitro studies of the enzymatic activities of purified
retroviral integrases have provided the following model of the
role of the integrase protein in the retroviral integration
reaction (3, 10, 19, 31, 51, 60). The long terminal repeat (LTR)
edges at the ends of the linear DNA molecule contain specific
cis-acting sequences recognized by the integrase protein. The
integrase protein catalyzes the endonucleolytic removal of the
two 3'-terminal nucleotides from each end of the linear viral
DNA molecule in a reaction referred to as the donor cleavage
reaction. The newly generated, recessed 3'-hydroxyl termini
are then joined to cellular DNA through an integrase-medi-
ated nucleophilic attack on phosphodiester bonds located on
the cellular DNA target (the DNA strand joining reaction).
Both the donor cleavage reaction and the DNA strand joining
reaction are catalyzed by purified retroviral integrases in vitro
by using short duplex oligonucleotides that match the ends of
linear viral DNA as substrates. In addition, purified integrases
can catalyze the reversal of the DNA strand joining reaction in
vitro in a process termed disintegration (5).
Mutational analysis of the in vitro enzymatic activities of
retroviral integrases has identified three potential functional
regions of the protein. The N terminus of the protein contains
an array of histidine and cysteine residues that is well con-
served among retroviral integrases (16, 29, 32) and is believed
to constitute a zinc finger-like domain (2, 4, 29). The N-
*Corresponding author. Mailing address: Dana-Farber Cancer In-
stitute, 44 Binney St., JFB824, Boston, MA 02115. Phone: (617)
632-4332. Fax: (617) 632-4338.
tPresent address: DepartmentofBiotechnology, Chung Ang Uni-
versity, Ansung Kyunggido, Korea 456-756.
tPermanent address: IstitutoSuperioredi Sanita, Rome, Italy.
§ Present address: Human Genome Sciences, Inc., Rockville, Md.
terminal domain is required for the donor oligonucleotide
cleavage and DNA strand joining reactions in vitro and may
play a role in the specific recognition of viral DNA sequences
present at the ends of the linear viral DNA molecule (19, 29a,
38, 56, 58). The C-terminal region of the protein is the least
conserved region among retroviral integrases and most likely
forms the major DNA-binding domain of the protein (49, 55,
59, 63, 64). The most highly conserved region of retroviral
integrases lies in the central region of the protein. The central
region is defined by three acidic amino acid residues in a highly
conserved spatial arrangement [the "D, D(35)E" motif] (32,
34). Mutations in any one of the three invariant aspartic or
glutamic acid residues coordinately block the donor cleavage,
DNA strand joining, and disintegration activities of integrases
in vitro (17, 20, 34, 36, 38, 56). These studies suggest that a
single active site is used to catalyze all of the polynucleotidyl
transfer reactions and that the conserved aspartic and glutamic
acid residues contribute to the active site of the enzyme (19,
60).
While the enzymatic activities of retroviral integrases are
becoming clearly defined through the use of in vitro integration
assays, other potential roles played by the integrase protein in
the establishment of a provirus in an infected cell may be more
readily identified by genetic analyses. For example, some
features of the integration process, such as the coordinated
joining of the two viral DNA ends to precisely spaced phos-
phodiester bonds in target DNA, the organization of viral
DNA and integrase protein into
preintegration complex, the entry of the preintegration com-
plex into the nucleus of the cell, and the features of the
preintegration complex that serve to prevent the autointegra-
tion of viral DNA, may involve functions of the integrase
protein that are not assayed by the current oligonucleotide-
based integration assays. Dissection of the roles of the inte-
grase protein in viral replication will be facilitated by the
identification of single-amino-acid substitutions throughout
the integrase protein that result in viruses whose only defect is
the conversion of the linear viral DNA molecule into an
integrated provirus. As a first step toward this goal, a number
of point mutations in the highly conserved central region of the
human immunodeficiency virus type 1 (HIV-1) integrase pro-
a stable and functional
1633
Vol. 68, No. 3

Page 2

1634
SHIN ET AL.
tein were constructed and analyzed for their effects on virus
replication in tissue culture cells.
MATERIALS AND METHODS
Construction of mutants. Plasmid targets for in vitro mu-
tagenesis were created by subcloning regions of thepol gene of
the infectious HXBc2 proviral DNA clone (25). To minimize
the likelihood of secondary mutations, small fragments of the
pol gene encoding portions of the integrase protein were used
for site-directed mutagenesis. For mutations in the region of
integrase amino acid D116, a 571-bp fragment spanning the
MscI and NdeI sites of HXBc2 (nucleotides 4551 to 5122,
according to the numbering scheme of Myers et al. [40]) was
subcloned into pGEMllZf(+) (Promega), and the resulting
plasmid was used as a target for site-directed mutagenesis as
described before (35). For mutations in the region of integrase
amino acid E152, a 474-bp fragment spanning the EcoRI and
NdeI sites of HXBc2 (nucleotides 4648 to 5122) was subcloned
into pGEM72f(+), and the resulting plasmid was used as the
target for mutagenesis. The mutagenic oligonucleotides used
are shown in Fig.
confirmed by DNA sequence analysis. To regenerate complete
proviral DNA clones, fragments carrying the desired mutation
were reinserted into the parental vector pHXB-SV (13), which
contains the full-length HXBc2 provirus and provides a simian
virus 40 origin of replication for episomal propagation in the
COS-7 monkey kidney cell line.
Cell culture, transfection, and infection. COS-7 cells were
grown in Dulbecco's modified Eagle's medium (DMEM) sup-
plemented with 10% fetal bovine serum. COS-7 cells (2 x 106)
were seeded into 100-mm culture dishes 24 h prior to trans-
1. The presence of the mutation was
A.
HIV-1
gag
IU3 IR IU5]
pol
a env
Integras
NH2
-
100
115
U3 |RU5
-COOH
288
200
160
TDNGSNFTSATVK AACW WAG IKQEFG IPYNPQ SQG VV ESMNK E LKK
A A KAN
A A
Q
fection. The cells were transfected with 10 ,ug of proviral DNA
by a DEAE-dextran procedure (39). Virus production was
evaluated by measurement of reverse transcriptase activity 48
h posttransfection as described previously (12). Alternatively,
virion release was measured by using a commercially available
HIV-1 p24 enzyme-linked immunosorbent-assay (ELISA) kit
(NEN/DuPont, Beverly, Mass.). Human SupTl and Jurkat
cells were maintained in RPMI 1640 medium supplemented
with 10% fetal bovine serum. SupTl or Jurkat cells (106) were
infected with cell-free virus equivalent to 50,000 cpm of reverse
transcriptase activity. Two-thirds of the culture medium was
replaced with an equal volume of fresh medium every third
day. Virus production was monitored by measuring the culture
supernatant reverse transcriptase activity every third day.
Metabolic labeling and immunoprecipitation. COS-7 cell
cultures were metabolically labeled for 14 h in medium con-
taining [35S]cysteine (50 ,iCi/ml) 48 h after transfection. Equiv-
alent numbers of labeled cells were lysed in RIPA buffer (140
mM NaCl, 8 mM Na2HPO4 2 mM NaH2PO4, 1% Nonidet
P-40, 0.5% sodium deoxycholate, 0.05% sodium dodecyl sul-
fate [SDS]), viral proteins were immunoprecipitated from the
cell lysates with an HIV-1-infected patient serum, and the
proteins were analyzed by electrophoresis on a 10 to 15%
(wt/vol) polyacrylamide gradient gel and visualized by fluoro-
graphy.
For pulse-labeling experiments, COS-7 cells were labeled for
20 min before the culture medium was replaced with fresh
medium lacking [35S]cysteine. For the analysis of virion pro-
teins, the labeling medium was collected and cleared of cell
debris by centrifugation. Virions were pelleted through a 20%
sucrose cushion by centrifugation at 25,000 x g for 1 h and
lysed in RIPA buffer, and the virion proteins were immuno-
precipitated and analyzed as described above.
Electron microscopy. Forty-eight hours after transfection,
COS-7 cells were fixed for 1 h at room temperature and then
for 12 h at 4°C in 4% (vol/vol) glutaraldehyde-3% (vol/vol)
paraformaldehyde in phosphate-buffered saline. Further pro-
cessing of the fixed COS-7 cells for thin-section electron
microscopy was done as described before (12).
Detection of viral DNA by PCR. The synthesis of viral DNA
in SupTl cells infected with the replication-defective integrase
mutants and wild-type virus was detected by a PCR technique
similar to ones described previously (36, 48). Equivalent doses
(100,000 cpm of reverse transcriptase activity) of virus from
COS-7 cell supernatants following transfection with wild-type
or S1471 or E1SA mutant proviruses were used to infect one
B.
C A A T A C
T A C A T A C
T C C T G A C
C T G A C A A
G A A T T C C
T C C
A G T A G T A G
T A G T A G A A
A A G A A
A
T
T G
A A
T G
C
g CT G A C
C
A A T
g G G C A
c
A G C
A
C A A T
T
C A A G
AT C
g C
T A T G
c
A G A A A
c
t
a
t
C A
A A
c
T A
T
TA
A A T G G C A GC
G G C A G C A AT
G C A A T T T AC
A A T T T C AC C
C C C C
A A A G T C
G A G T A G T A G A
T G A A T A A A G A
A A T A A A G A A T
A T T A T A G G A C
- 3'
- 3'
- 3'
- 3'
- 3'
- 3'
- 3'
- 3'
- 3'
FIG. 1. Mutation of the HIV-1 integrase protein. (A) Proviral DNA is shown at the top. The gag, pol, and env open reading frames are shown,
surrounded by the directly repeated LTRs. The 288-bp integrase protein encoded by the 3' region of the pol gene is enlarged below. The central
region of the integrase protein targeted for mutagenesis is shown in single-letter code from amino acids Ti 15 to K160. Amino acids that were
changed by mutagenesis are underlined, with the amino acid substitution indicated. (B) Mutagenic oligonucleotides used for the construction of
the mutations shown in panel A. Nucleotides corresponding to the wild-type HIV-1 integrase sequence are shown in uppercase letters, and the
nucleotide changes introduced to create the mutation are shown in lowercase letters.
Ti15A:
D116A: 5'- A A
N117K: 5'- C A
G118A: 5'- A T
Y143N
S1471:
E152A: 5'- G T
S153A:
K159Q
5' - G T
:5' - A C
5'- C C
5'- C A
5' - A T
A A A A A
A A C A A
A T A C A
A C A T A
T CG A G
T A C A A
C A A G G
A GG A G
G A A T A
J. VIROL.

Page 3

HIV-l INTEGRASE MUTANTS
1635
million fresh SupTI cells. Because cells transfected with mu-
tant D116A did not release reverse transcriptase activity into
cell supernatants, a direct comparison of this mutant with the
other replication-defective mutants and wild-type virus could
not be made. Twenty-four hours after infection, total DNA was
prepared from the infected cells by standard techniques (39).
Total DNA was treated with restriction endonuclease DpnI
prior to PCR to inactivate potentially contaminating plasmid
DNA carried over from the COS-7 cell transfections. One
microgram of endonuclease-treated total DNA was subjected
to 35 rounds of PCR amplification with primers envl (nucle-
otides 7950 to 7969 of HXBc2) and env2 (nucleotides 8545 to
8526 of HXBc2) in the envelope region with a commercially
available kit and conditions specified by the manufacturer
(Perkin-Elmer Cetus). The products of amplification were
separated on a 1.5% agarose gel, blotted onto a nylon filter,
and hybridized with an envelope-specific oligonucleotide probe
(env3; nucleotides 8285 to 8308 of HXBc2) labeled at its 5' end
with [y-32P]ATP by standard techniques (39). As a control for
plasmid contamination, 10 ng of plasmid pHXB-SV, treated
with DpnI or not treated, was used as a substrate for amplifi-
cation under the same conditions. The HXBc2 viral DNA
sequence contains five DpnI sites located between the primer
sequences used for amplification. The plasmid DNA used for
transfection of COS-7 cells is digested to completion by DpnI,
which requires a fully methylated recognition sequence for
cleavage; viral DNA synthesized in the cell following infection
is fully unmethylated and resistant to cleavage by DpnI (data
not shown). As a further control to eliminate the possibility of
plasmid contamination from the COS-7 cell transfections, PCR
experiments were also performed as described above after
infection of SupTl cells with supernatants prepared from
stable SupTI cell lines chronically producing viral mutant
S1471 or E152A (data not shown). These stable cell lines were
constructed by cotransfection of mutant proviral DNA clones
and a neomycin resistance plasmid into SupTl cells, followed
by selection of neomycin-resistant, virus-producing cell lines
(54a).
RESULTS
Effects of integrase mutations on viral protein synthesis and
processing. Single-amino-acid substitutions were introduced
into the highly conserved central region of the HIV-1 integrase
protein. The nature and location of each point mutation are
indicated in Fig. 1. Each of the mutations was introduced into
an HXBc2 provirus, and wild-type and mutant proviral DNAs
were transfected into COS-7 cells. A full complement of
35S-labeled viral proteins was found by immunoprecipitation of
COS-7 cell lysates with an HIV-1-infected patient serum at 48
h after transfection with wild-type and integrase mutant pro-
viruses (Fig. 2). Similar amounts of the processed capsid
proteins p24 and p17 were found after transfection of the
wild-type proviruses and all of the integrase mutants except for
mutant D116A. Cell lysates from COS-7 cells transfected with
mutant Dl 16A consistently showed reduced levels of the p24
and p17 proteins, while the levels of the other viral proteins,
including the p55gag precursor, were similar to those in the wild
type (Fig. 2, lane 4). Thus, the D116A mutant appeared to
exhibit a defect in processing of the p55gag precursor protein,
while none of the other single-amino-acid changes examined
had major effects on the synthesis or processing of viral
proteins.
The nature of the gag protein defect observed for mutant
D116A was further investigated by pulse-chase labeling anal-
ysis of viral proteins produced after transfection of COS-7
v
<< Y < Z - <
U O(N OC
- r-vt t
1_ 0 z- G >_
3
4
56
<
c)c)
Lo
C-n
10 11
cy
Lo
s-
o
0
E 3:
1
)N
cJ
Lo
F_
(O lJI
2
7 89
-
gp-6O
|t"t| |;|; -s
r
gpl120
VWOm
~~~~P55
- --p24
400,"W
map
~~~~~~~~..p17
FIG. 2. SDS-PAGE analysis of viral proteins synthesized in COS-7
cells after transfection with wild-type and mutant proviral DNAs.
COS-7 cells were labeled with [35S]cysteine at 48 h posttransfection,
and the viralproteinswere immunoprecipitatedfrom cellsupernatants
and resolved on a 10 to 15% (wt/vol) gradient polyacrylamide gel. The
cells were mock transfected (lane 1) or transfected withwild-type (WVT,
lane 2), TI I5A (lane 3), DI 16A (lane 4), NI 17K (lane 5), Gl 18A(lane
6), Y143N (lane 7), S1471 (lane 8), E152A (lane 9), S153A (lane 10),
or K159Q (lane 11) mutant proviral DNAs. The relative positions of
the HIV-I proteins are indicated.
cells. As shown in Fig. 3, lanes 2 and 3, similar levels of
355-labeled p559gag precursor polyprotein were formed in cells
transfected with wild-type and DI16A mutant proviruses after
a short labeling period with [355S]cysteine, indicating that viral
protein synthesis was not defective in the Dl 16A mutant. The
total amount of 35an-labeled gagproteinspresent inlysates of
cells transfected with wild-type provirus decreased during the
subsequent chase period (Fig. 3, lanes 5 and 8), presumably
reflecting the release of labeled precursor proteins from the
cells in the form of virions. However, a considerable fraction of
the labeled p55 precursorprotein remained cell associated in
the form of processed p24 and p17. In contrast, little or no
35c -labeledgag proteinwas found in lysatesof cells transfected
with the Dl 16A mutant during the chase period (Fig. 3, lanes
6 and 9). The disappearance of the 35 -labeled p553ag precur-
sor protein during the chase most likely resulted from the
release of the unprocessed precursor from the cell during
virion budding (see below). The greatly reduced levels of
processed p24 and p17 proteins in the experiment are consis-
tent with the low levels of mature gag products seen in the
experiment shown in Fig. 2 and suggest that the D116A
mutation interferes with theprocessingof the
polyprotein.
Effects of integrase mutations on virion formation. Virion
release from COS-7 cells transfected with wild-type and inte-
grasemutantproviruseswas measured 48 hposttransfection by
assaying culture supernatants for particle-associatedviral pro-
teins and reverse transcriptase activity. A wild-type pattern of
viral proteins was found in the supernatantsof cells transfected
with all of the integrasemutantproviruses except for mutant
DI16A (Fig. 4, lanes 7, 9, and 1o, and data not shown). All of
the mutants except DI 16A also directed the release of wild-
type levels of particle-associated supernatant reverse tran-
scriptase activity (data not shown), providing further evidence
p55D1A
precursor
VOL. 68, 1994

Page 4

1636
SHIN ET AL.
cell
20 min
pulse
3 hr
chase
6 hr
chase
1
2
3
45
6
7
8
9
>1
||l " ,,
1234567
* t
-~gp160
gp120
9
-p55
ss:
t3t:L
m1 < _ <
C
_-
2 3 4c
1 2 345
Y~
N-
,It
cu
m
o
E
virion
U)
C-
0-
.0
2
-r-
_t
E
§
a
mm
6
~~o
gp6
0F
_-_
-p24
mew
- p17
FIG. 3. Pulse-chase analysis of viral protein synthesis in COS-7
cells transfected with wild-type and mutant D116A proviruses. COS-7
cells were pulse-labeled with [35S]-cysteine at 48 h posttransfection
with wild-type (WT) or mutant D1 16A proviral DNAs, as described in
Materials and Methods. The cells were lysed immediately after the
pulse label (lanes 1 to 3) or after a 3-h (lanes 4 to 6) or 6-h (lanes 7 to
9) chase with unlabeled amino acid, and the viral proteins were
immunoprecipitated from the cell lysates and resolved on a 10 to 15%
(wt/vol) polyacrylamide gradient gel. Cells were mock transfected
(lanes 1, 4, and 7) or transfected with wild-type provirus (lanes 2, 5, and
8) or mutant D116A (lanes 3, 6, and 9) proviral DNAs. The relative
positions of HIV-1 proteins are indicated.
that their mutations had no detectable effect on the process of
virion formation. In addition, wild-type levels of particle-
associated gpl20 were found for all of the mutant viruses,
indicating that the mutations had no detectable effect on the
incorporation of envelope glycoprotein into the budding viri-
ons.
The appearance of particle-associated p55gag precursor pro-
tein in the supernatant of cells transfected with mutant D116A
indicated that virus particles were released from these cells
(Fig. 4, lane 8). Greatly reduced amounts of mature p24 and
p17 proteins were found in the virions produced by mutant
D116A, consistent with the earlier experiments that indicated
a defect in p55gag precursor processing for this mutant. Be-
cause of the generally high nonspecific background of radiola-
beled proteins found in the region of the gel corresponding to
the size of p16QWa"PoI, it was not possible to determine whether
unprocessed gag-pol precursor protein was incorporated into
virions of mutant D116A (Fig. 4, lane 8). Particle release from
cells transfected with the D116A mutant was also confirmed by
an HIV-1 p24 ELISA assay (Fig. 5B), which apparently can
detect p24 protein sequences in the context of the p55gag
precursor. Although viral particles were released from cells
transfected with mutant Dl 16A, particle-associated reverse
transcriptase activity was not detected in the supernatants of
transfected cells (Fig. 5A).
Effects of integrase mutations on virion morphology. Virion
~4W'qw 4sp
A~Afi
~1
- p24
- p17
FIG. 4. Viral proteins present in cell lysates and virions after
transfection of COS-7 cells with wild-type or replication-defective
integrase mutant proviral DNA. COS-7 cells were labeled with
[35S]cysteine at 48 h after transfection. Viral proteins were immuno-
precipitated from cell lysates (lanes
through sucrose cushions from culture supernatants (lanes 6 to 10) and
resolved on a 10 to 15% (wt/vol) gradient polyacrylamide gel. Cells
were mock transfected (lanes
proviral (lanes 2 and 7) or D116A (lanes 3 and 8), S147I (lanes 4 and
9), or E152A (lanes 5 and 10) mutant proviral DNAs. The relative
positions of HIV-1 proteins are indicated.
1 to 5) or from virions pelleted
I and 6) or transfected with wild-type
morphology was assessed by thin-section electron microscopy
of monolayer preparations of COS-7 cells transfected with
wild-type and integrase mutant proviruses. The virions pro-
duced by the integrase mutant proviruses fell into three
general morphological classes, as shown in Fig. 6. Mutants
T115A, N117K, G118A, Y143N, S1471, S153A, and K159Q
produced virions with a morphology indistinguishable from
that of wild-type virions (Fig. 6A). These virions typically
contained an eccentrically placed, cone-shaped, electron-dense
core surrounded by a thin envelope layer. In some sections, the
core can be seen to include a spherical nucleoid surrounded by
the cone-shaped p24 shell. Occasional virion particles having a
thick electron-dense shell at the inner circumference of the
envelope and an electron-lucent center without a condensed
core were produced by these integrase mutants at levels similar
to those in wild-type proviruses. These particles most likely
represent immature, newly released virions that have not yet
undergone precursor processing and core formation (27).
All of the virions produced by cells transfected with mutant
D116A had the appearance of immature virions (Fig. 6B). All
of the particles produced by mutant D116A contained thick
electron-dense outer shells and electron-lucent centers, with
no evidence of a condensed core, indicating that this mutation
resulted in a defect in virion core formation. Virions produced
by mutant E152A showed two general morphological types
(Fig. 6C). Some particles contained an eccentrically located,
electron-dense spherical core, characteristic of the spherical
-gpl6O
- gp520
-p55
J VlIROL.

Page 5

HIV-1 INTEGRASE MUTANTS
1637
A 80,000,
i-_
E
Cu
60,000 t
40,000 F
20,000[
0
B
3L
2
0~
't
cM
Q-
c]
o0
mock
wild-type
Dl 16A
mock
wild-type
Dl 16A
FIG. 5. Viral particle release from COS-7 cells transfected with
wild-type or mutant D116A proviruses. Forty-eight hours after trans-
fection, culture supernatants were assayed for viral particles by mea-
suring viral reverse transcriptase (RT) activity (A) or p24 antigen (B)
as described in Materials and Methods. O.D., optical density.
nucleoid found in wild-type viral particles. Other particles
contained, in addition to the spherical nucleoid structure, an
adjacent, irregularly electron-dense structure, giving the core
of the mutant virion a distended, bilobed appearance.
Replication of integrase mutants in tissue culture. The
effects of the integrase point mutations on viral replication
were assessed by infecting cells of the human SupTl T cell line
with supernatants prepared from COS-7 cells 48 h after
transfection with mutant
cultures of SupTl cells were infected with equivalent doses
(reverse transcriptase units) of wild-type or mutant virus. The
course of virus replication was monitored by supernatant
reverse transcriptase level every 3 days.
Three replication phenotypes were observed for the inte-
grase mutants examined after infection of SupTl cells. Similar
results were obtained after infection of Jurkat cells (data not
shown), demonstrating that the observed phenotypes were not
strictly cell type dependent. Mutants Ti1SA, G118A, S153A,
and K159Q had replication profiles indistinguishable from that
of the wild-type virus (Fig. 7A). Two of the mutants, N117K
and Y143N, exhibited delayed-replication phenotypes (Fig.
7C). For both of these mutants, peak virion production oc-
curred 20 to 30 days postinfection, compared with 10 to 15 days
postinfection for wild-type virus. Virus harvested from the
supernatants of cells infected with the N117K and Y143N
mutant viruses during the peak of replication (day 21 in Fig.
7C) showed similar delays in replication when used to infect
fresh SupTl cells (Fig. 7D), demonstrating that the delayed
appearance of peak reverse transcriptase levels displayed by
these two mutants was the result of a delayed course of
replication and was not due to the appearance of revertant
viruses. Two mutants, S1471 and E152A, failed to replicate to
detectable levels in SupTl cells, with no detectable virus
production for up to 30 days postinfection (Fig. 7B). Both of
these mutants also failed to replicate after direct transfection
into SupTi or Jurkat cells (data not shown).
Mutant DI116A did not release reverse transcriptase activity
after transfection into COS-7 cells, precluding a direct com-
or wild-type proviruses. Parallel
parison of its replication phenotype with those of the other
mutants. However, this mutant was shown to be completely
replication defective by the following criteria. No virus repli-
cation was detected in SupTl or Jurkat cells infected with
supernatants prepared from COS-7 cells 48 h after transfection
with D116A mutant proviruses or by cocultivation of the
transfected COS-7 cells with fresh SupTl or Jurkat cells (data
not shown). In addition, no virus replication was detected after
direct transfection of SupTl or Jurkat cells with D116A mutant
proviruses, even though viral protein production was readily
detected by metabolic labeling of the transfected cells with
35S-labeled amino acids (data not shown).
Detection of viral DNA in cells infected with thereplication-
defective integrase mutants. Mutant viruses having a specific
defect in the formation of integrated proviruses are expected
to be able to support the synthesisof viral DNAuponinfection
of a susceptible target cell. Each of the replication-defective
HIV-1 integrase mutants was tested for the abilitytosynthesize
viral DNA after infection of SupTl cells. Viral DNA was
readily detected by PCR 24 h after infection of SupTl cells
with wild-type virus or mutant E152A (Fig. 8, lanes 5 and 7).
No viral DNA was found in cells infected with mutant S1471
(Fig. 8, lane 6). Similar results were found whether virus was
prepared from the supernatants of transfected COS-7 cells or
from the supernatants of cells chronically producing wild-type
or mutant viruses (Fig. 8, lanes 8 and 9, and data not shown).
Because cells transfected with mutant D116A did not release
reverse transcriptase activity into the supernatant, a direct
comparison of this mutant with the others or with wild-type
virus was not possible. However, a dose of D116A mutant
virions could be estimated by electrophoretic analysis of virion
pellets and normalized towild-typevirusbytheintensityof the
gpl20 envelope glycoprotein (for example, see Fig. 4, lanes 7
and 8). By this standard, D116A mutant virions were unable to
direct the synthesis of viral DNA at levels detectable by the
PCR assay, while equivalent doses of wild-type virus synthe-
sized readily detectable levels of viral DNA (datanotshown).
In addition to the PCR assay described above, a Southern
blotting and hybridization procedure was used to detect viral
DNA synthesized in SupTl cells after cocultivation with cell
lines chronically producing the replication-defective viruses
(data not shown). These experiments confirmed that mutant
E152A synthesized nearly wild-type levels offull-lengthlinear
viral DNA, while viral DNA synthesis by mutants D116A and
S1471 was not detectable (54a).
DISCUSSION
The genetic analysis presented here demonstrates that a
variety of viral replication phenotypes can result from single-
amino-acid substitutions in the integrase protein.Three repli-
cation-defective viruses were generated by mutagenesisof the
highlyconserved centralregionof the HIV-1integrase protein,
each apparently blocked at a different stage of the viral
replication cycle. Previousgeneticstudies have established that
the integrase protein is required for only one stage of the
retroviral replication cycle, the formation of an integrated
provirus from the unintegratedlinear viral DNA intermediate
(14, 15, 43, 46, 50). However,thiswork, togetherwithprevious
work that demonstrated that someintegrasemutants can have
deleterious effects on viral protein processing and virion
release (47, 50, 53), demonstrates that mutantintegrase pro-
teins can interfere with avarietyofstepsin theearlyand late
stages of the viral replication cycle. This observation is not
surprising, since the integrase proteinissynthesizedaspartof
VOL. 68, 1994