Modulation of TRIM5alpha activity in human cells by alternatively spliced TRIM5 isoforms.
ABSTRACT TRIM5α is a restriction factor that can block an early step in the retroviral life cycle by recognizing and causing the disassembly of incoming viral capsids, thereby preventing the completion of reverse transcription. Numerous other isoforms of human TRIM5 exist, and isoforms lacking a C-terminal SPRY domain can inhibit the activity of TRIM5α. Thus, TRIM5α activity in a given cell type could be dependent on the relative proportions of TRIM5 isoforms expressed, but little information concerning the relative expression of TRIM5 isoforms in human cells is available. In this study, we demonstrate that mRNAs coding for TRIM5α represent only 50% of total TRIM5 transcripts in human cell lines, CD4(+) T cells, and macrophages. Transcripts coding for, in order of abundance, TRIM5ι (TRIM5-iota), a previously uncharacterized isoform, TRIM5γ, TRIM5δ, and TRIM5κ are also present. Like TRIM5γ and TRIM5δ, TRIM5ι and TRIM5κ do not inhibit HIV-1 replication, but both have dominant-negative activity against TRIM5α. Specific knockdown of TRIM5ι increases TRIM5α activity in human U373-X4 cells, indicating that physiological levels of expression of truncated TRIM5 isoforms in human cells can reduce the activity of TRIM5α.
- SourceAvailable from: ncbi.nlm.nih.gov[show abstract] [hide abstract]
ABSTRACT: TRIM5 is a determinant of species-specific differences in susceptibility to infection by retroviruses bearing particular capsids. Human immunodeficiency virus type 1 (HIV-1) infection is blocked by the alpha isoform of macaque TRIM5 (TRIM5alpha(rh)) or by the product of the owl monkey TRIM5-cyclophilin A gene fusion (TRIMCyp). Human TRIM5alpha potently restricts specific strains of murine leukemia virus (N-MLV) but has only a modest effect on HIV-1. The amino termini of TRIM5 orthologues are highly conserved and possess a coiled-coil domain that promotes homomultimerization. Here we show that heterologous expression of TRIM5alpha(rh) or TRIMCyp in human cells interferes with the anti-N-MLV activity of endogenous human TRIM5alpha (TRIM5alpha(hu)). Deletion of the cyclophilin domain from TRIMCyp has no effect on heteromultimerization or colocalization with TRIM5alpha(hu) but prevents interference with anti-N-MLV activity. These data demonstrate that TRIM5 orthologues form heteromultimers and indicate that C-terminal extensions alter virus recognition by multimers of these proteins.Journal of Virology 07/2005; 79(12):7883-8. · 5.08 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: To probe the genetic determinants controlling the interaction between the retroviral restriction gene Fv1 and its murine leukemia virus target, we set out to develop rapid, transient assays for Fv1 function. Cells were transfected or transduced with Fv1 expression plasmids which can produce green fluorescent protein via an internal ribosome entry site positioned between the Fv1 and green fluorescent protein coding sequences. Fv1 function was then assessed by comparing virus replication in green fluorescent protein-positive and -negative cells, using retroviral vectors encoding a second fluorescent marker, yellow fluorescent protein, or beta-galactosidase. Using this assay, we could show that Fv1 specificities were not as absolute as previously thought, since the Fv1(b) allele was capable of interacting with "nonrestricted" B- and NB-tropic viruses and by shuffling the n- and b-alleles of Fv1, it was possible to generate a Fv1 molecule capable of restricting N-, B-, and NB-tropic viruses equally efficiently. Further, we could show that the presence of nonrestricting Fv1 in the same cell as restrictive Fv1 abrogates restriction, implying competition for binding to the retroviral target.Journal of Virology 09/2000; 74(16):7422-30. · 5.08 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: The TRIM5 family of proteins contains a RING domain, one or two B boxes, and a coiled-coil domain. The TRIM5alpha isoform also encodes a C-terminal B30.2(SPRY) domain, differences within which define the breadth and potency of TRIM5alpha-mediated retroviral restriction. Because Macaca nemestrina animals are susceptible to some human immunodeficiency virus (HIV) isolates, we sought to determine if differences exist in the TRIM5 gene and transcripts of these animals. We identified a two-nucleotide deletion (Delta2) in the transcript at the 5' terminus of exon 7 in all M. nemestrina TRIM5 cDNA clones examined. This frameshift results in a truncated protein of 300 amino acids lacking the B30.2(SPRY) domain, which we have named TRIM5theta. This deletion is likely due to a single nucleotide polymorphism that alters the 3' splice site between intron 6 and exon 7. In some clones, a deletion of the entire 27-nucleotide exon 7 (Deltaexon7) resulted in the restoration of the TRIM5 open reading frame and the generation of another novel isoform, TRIM5eta. There are 18 amino acid differences between M. nemestrina TRIM5eta and Macaca mulatta TRIM5alpha, some of which are at or near locations previously shown to affect the breadth and potency of TRIM5alpha-mediated restriction. Infectivity assays performed on permissive CrFK cells stably transduced with TRIM5eta or TRIM5theta show that these isoforms are incapable of restricting either HIV type 1 (HIV-1) or simian immunodeficiency virus infection. The expression of TRIM5 alleles incapable of restricting HIV-1 infection may contribute to the previously reported increased susceptibility of M. nemestrina to HIV-1 infection in vivo.Journal of Virology 12/2007; 81(22):12210-7. · 5.08 Impact Factor
JOURNAL OF VIROLOGY, Aug. 2011, p. 7828–7835
Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Vol. 85, No. 15
Modulation of TRIM5? Activity in Human Cells by Alternatively
Spliced TRIM5 Isoforms?‡
Emilie Battivelli,1,2Julie Migraine,1,2Denise Lecossier,1,2Saori Matsuoka,1,2† Danielle Perez-Bercoff,1,2
Sentob Saragosti,1,2Franc ¸ois Clavel,1,2and Allan J. Hance1,2,3*
INSERM U941, Paris 75010, France1; Institut Universitaire d’He ´matologie, Universite ´ Paris Diderot, Ho ˆpital Saint-Louis,
Paris 75010, France2; and Service des Maladies Infectieuses et Tropicales, Assistance Publique—Ho ˆpitaux de
Paris, Ho ˆpital Bichat—Claude Bernard, Paris 75018, France3
Received 31 March 2011/Accepted 18 May 2011
TRIM5? is a restriction factor that can block an early step in the retroviral life cycle by recognizing and
causing the disassembly of incoming viral capsids, thereby preventing the completion of reverse transcription.
Numerous other isoforms of human TRIM5 exist, and isoforms lacking a C-terminal SPRY domain can inhibit
the activity of TRIM5?. Thus, TRIM5? activity in a given cell type could be dependent on the relative
proportions of TRIM5 isoforms expressed, but little information concerning the relative expression of TRIM5
isoforms in human cells is available. In this study, we demonstrate that mRNAs coding for TRIM5? represent
only 50% of total TRIM5 transcripts in human cell lines, CD4?T cells, and macrophages. Transcripts coding
for, in order of abundance, TRIM5? (TRIM5-iota), a previously uncharacterized isoform, TRIM5?, TRIM5?,
and TRIM5? are also present. Like TRIM5? and TRIM5?, TRIM5? and TRIM5? do not inhibit HIV-1
replication, but both have dominant-negative activity against TRIM5?. Specific knockdown of TRIM5? in-
creases TRIM5? activity in human U373-X4 cells, indicating that physiological levels of expression of trun-
cated TRIM5 isoforms in human cells can reduce the activity of TRIM5?.
Following the entry of retroviral capsids into the cytoplasm
of target cells, they can encounter cellular restriction factors
that block the early steps of the viral life cycle. One well-
characterized early restriction factor is TRIM5? (22, 27, 41,
46), whose recognition of the incoming capsid leads to rapid
capsid disassembly, preventing the completion of reverse tran-
scription (42). Human TRIM5? can potently inhibit N-tropic
murine leukemia virus (N-MLV) (12, 18, 31, 48) and moder-
ately inhibit equine infectious anemia virus and feline immu-
nodeficiency virus replication (8, 18, 33). The infectivity of
laboratory-adapted strains of HIV-1 is inhibited only 2- to
3-fold by the physiological levels of human TRIM5? expressed
in human cells (13, 17, 39, 44, 49), but HIV-1 expressing capsid
proteins derived from clinical isolates can be considerably
more sensitive to human TRIM5?, especially if TRIM5? ex-
pression in target cells is augmented by pretreatment with
alpha interferon (IFN-?) (1).
TRIM5? is the longest of numerous TRIM5 isoforms ex-
pressed in human cells (Fig. 1) and is composed of several
distinct domains. The RING domain (coded by exon 2) ex-
presses E3 ubiquitin ligase activity, is required for optimal
antiviral activity, and contributes to the rapid turnover of
TRIM5? (9, 15, 41). The function of the B-box 2 domain
(coded by exon 2) is not fully understood, but amino acid
changes in this region can influence TRIM5? turnover, higher-
order self-association of TRIM5? dimers, the formation of
TRIM5?-containing cytoplasmic bodies, and antiviral activity
(7, 10, 15, 20). The coiled-coil domain (coded by exons 2 to 4)
promotes the formation of homodimers and participates in
capsid recognition (16, 19, 24, 26, 30). These three domains
comprise the RING/B-box/coiled-coil (RBCC) tripartite motif
characteristic of all TRIM proteins. Finally, TRIM5? also has
a C-terminal SPRY domain (coded by exons 7 and 8) that
directly interacts with the viral capsid, and species-specific
polymorphisms in this region are a major determinant of the
spectrum of viruses recognized by a given ortholog (27, 28, 43).
As indicated in Fig. 1, mRNAs coding for numerous other
TRIM5 isoforms that code for proteins containing the RBCC
domains, but in which the SPRY domain is replaced by shorter
alternative sequences, have been described. Two of these trun-
cated TRIM5 isoforms, TRIM5? and TRIM5?, have been
shown to be devoid of antiviral activity against N-MLV and
HIV-1. Interestingly, however, overexpression of both of these
isoforms, as well as TRIM5? ?SPRY constructs, strongly in-
hibits the activity of TRIM5?, resulting, at least in part, from
the heterodimerization of these isoforms with TRIM5? (2, 23,
29, 30, 41, 49).
The ability of truncated TRIM5 isoforms to inhibit TRIM5?
suggests that TRIM5? activity in a given cell type could be
dependent on the relative proportions of TRIM5 isoforms
expressed, but little information concerning the relative pro-
portion of different TRIM5 isoforms expressed in human cells
is available. In this study, we demonstrate that mRNAs coding
for TRIM5? represent only 50% of total TRIM5 transcripts in
human cell lines, CD4?T cells, and macrophages. Transcripts
coding for, in order of abundance, TRIM5? (TRIM5-iota; a
previously uncharacterized isoform), TRIM5?, TRIM5?, and
* Corresponding author. Mailing address: INSERM U941, Ho ˆpital
Saint Louis, 1 Avenue Claude Vellefaux, 75475 Paris Cedex 10,
France. Phone: 33-1-57-27-67-56. Fax: 33-1-57-27-68-04. E-mail: allan
† Present address: AIDS Research Center, National Institute of
Infectious Disease, Tokyo, Japan.
‡ Supplemental material for this article may be found at http://jvi
?Published ahead of print on 1 June 2011.
TRIM5? are also present. Like TRIM5? and TRIM5?,
TRIM5? and TRIM5? do not inhibit HIV-1 replication, but
both have dominant-negative activity against TRIM5?. Spe-
cific knockdown of TRIM5? increases TRIM5? activity in hu-
man U373-X4 cells, indicating that relative abundance of dif-
ferent TRIM5 isoforms can modulate TRIM5? activity in
MATERIALS AND METHODS
Cell culture. The human cell lines used in this study were cultured as previ-
ously described (1, 25). Peripheral blood mononuclear cells (PBMCs) were
isolated by Ficoll-Hypaque density gradient centrifugation (Pharmacia Biotech)
from buffy coats obtained from healthy blood donors and washed twice in cold
phosphate-buffered saline (PBS) containing 0.3 mM EDTA and once in cold
PBS. Adherent cells were separated from nonadherent cells by adherence to
plastic for 1 h at 37°C in MDM medium (RPMI 1640 containing 100 U/ml
penicillin G, 100 ?g/ml streptomycin, 100 ?m nonessential amino acids, 50 ?g/ml
?-mercaptoethanol, 25 mM HEPES [all from Gibco BRL]) containing 2% heat-
inactivated human AB serum (ABS; PAA Laboratories, France). To obtain
monocyte-derived macrophages, adherent cells were washed 3 times with PBS
and cultured overnight in MDM medium containing 10% ABS; cells were de-
tached by incubation at 4°C for 1 h followed by vortexing and gentle scraping,
further purified by negative selection using the Dynal monocyte negative isola-
tion kit (Dynal Biotech) according to the manufacturer’s instructions, and cul-
tured in MDM medium containing 15% ABS at 106cells/ml for 7 days.
To obtain activated CD4?lymphocytes, nonadherent cells were positively
selected using an anti-CD4 monoclonal antibody (R&D Systems) coupled to
CELLection pan-mouse IgG Dynabeads (Dynal Biotech) according to the man-
ufacturer’s recommendations and resuspended at 106cells/ml in RPMI 1640
containing 10% heat-inactivated fetal calf serum (FCS), 100 U/ml penicillin G,
100 ?g/ml streptomycin, 5 U/ml recombinant human interleukin 2 (IL-2)
(Roche), and 100 ?l CD3/CD28-coated Dynabeads (Dynal Biotech) for 72 h.
Viruses. The production of the vesicular stomatitis virus (VSV)-pseudotyped
pNL4-3-based recombinant viruses containing a deletion in env, expressing
Renilla luciferase in place of nef protein, and in which the Gag-protease se-
quences were derived from clinical isolates (NRC2 or NRC10) or from NL4-3
has previously been described (1, 25). Recombinant VSV-pseudotyped N-tropic
and B-tropic murine leukemia virus (N-MLV and B-MLV, respectively) vectors
expressing yellow fluorescent protein (YFP) were prepared as previously de-
scribed (1, 3), using plasmids obtained from Jonathan Stoye.
Cell lines overexpressing TRIM5 isoforms. The techniques used to generate
CRFK cells and U373-X4 cells overexpressing LacZ and untagged or hemagglu-
tinin (HA)-tagged TRIM5? and TRIM5? by transduction with pLenti6/V5-D-
TOPO-based vectors have been previously described (1). Analogous techniques
were used to produce cell lines overexpressing untagged and HA-tagged TRIM5?
(CRFK cells and U373-X4 cells) and TRIM5? (CRFK cells), as described in the
U373-X4 cells expressing pre-miRNA targeting TRIM5?. Complementary oli-
gonucleotides encompassing a sequence in the coding region of TRIM5? down-
stream of the nominal exon 4 splice donor site (oligonucleotides T19 and T20;
see Table S1 in the supplemental material) or in the 3? untranslated region
(UTR) of TRIM5? (oligonucleotides T21 and T22) were synthesized, hybridized,
and cloned into the pcDNA 6.2-GW/EmGFP-miR vector (Invitrogen). The same
vector coding for a sequence targeting LacZ was used as a control. The expres-
sion cassettes carrying the pre-microRNAs (pre-miRNAs) targeting TRIM5? or
LacZ and coexpressing emerald GFP (EmGFP) were transferred to the pLenti6/
V5-DEST vector (Invitrogen), and lentiviral vectors were prepared according to
the manufacturer’s instructions. U373-X4 cells were transduced with the lenti-
viral vectors, and transduced cells were selected using 8 ?g/ml blasticidin. To
enrich for cells expressing high levels of the pre-miRNA, cell populations
strongly expressing the coexpressed GFP were isolated by cell sorting
(FACSVantage SE Diva; Becton-Dickinson) and maintained under selection
with 8 ?g/ml blasticidin. For each construct, two different cell lines were gener-
ated independently and are labeled A and B.
Infectivity assays. HIV-1 infectivity was measured by determining luciferase
activity in target cells 40 h after infection as previously described (1, 25). When
indicated, cells were incubated in the presence of 1,000 U/ml IFN-? for 24 h prior
to infection to increase TRIM5 expression. To measure MLV infectivity, the
proportion of YFP-expressing cells was determined by cytofluorometry 40 h after
infection, using previously described procedures (1).
FIG. 1. TRIM5 isoforms expressed by human cells. TRIM5 isoforms supported by cDNA sequences in GenBank as curated by AceView (45)
that include exons 2 to 4 are shown and are identified by the nomenclature used in this study, by AceView nomenclature, and by the Celera genome
annotation (47). Representative nucleotide sequence accession numbers are also given. In the splicing profile, filled rectangles indicate translated
regions, and open rectangles indicate nontranslated regions of each exon. Exon numbers as used in this study are indicated above the profile for
TRIM5?. The exons coding for the RING, linker 1 (L1), B-box 2 (B-box), coiled-coil (CC), linker 2 (L2), and SPRY domains of TRIM5? are
shown at the bottom of the figure. The human genome does not code for TRIM5? (AF220026); this sequence appears to result from illegitimate
recombination between TRIM5? and TRIM5? transcripts. TRIM5? has also been reported (32), but this putative isoform does not have an RBCC
VOL. 85, 2011 TRIM5 ISOFORMS IN HUMAN CELLS7829
Western blotting. Cell suspensions were washed two times with PBS, and 3 ?
106cells were pelleted and resuspended in 75 ?l of lysis buffer (1% NP-40, 150
mM NaCl, 1 mM EDTA, 1% protease inhibitor cocktail [Sigma], 20 mM
HEPES, pH 7.8). Lysates were cleared by centrifugation (11,000 ? g for 10 min),
and protein content was determined using the Bradford assay (Bio-Rad). Cel-
lular proteins (10 ?g) were separated by electrophoresis into 10% SDS-PAGE
gels and transferred to Protran BA85 membranes (Whatman) and blocked (Li-
Cor blocking buffer; 7 h, 20°C). To detect TRIM5 isoforms, the membranes were
probed sequentially with a goat polyclonal anti-TRIM5 antibody (1/1,000 dilu-
tion, 4°C, overnight) (ab4389; Abcam), a biotinylated horse anti-goat IgG anti-
body (1/5,000 dilution, room temperature, 1 h) (BA-9500; Vector Laboratories),
and streptavidin-conjugated horseradish peroxidase (1/100,000 dilution, room
temperature, 45 min) (SA-5004; Vector Laboratories) and revealed by using
ECL Advance (GE Healthcare). To detect HA-tagged TRIM5? isoforms, the
membranes were probed with a peroxidase-conjugated rat anti-HA epitope an-
tibody (1/10,000, 18 h, 4°C) (3F10; Roche) and revealed using ECL Advance.
Real-time PCR. Cell suspensions were washed with PBS, and mRNA was
extracted from 2 ? 106cells by using NucleoSpin RNA II kits (Macherey-Nagel,
Du ¨ren, Germany). cDNA synthesis was performed using random hexamers and
Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen,
Carlsbad, CA) according to the manufacturer’s instructions. TRIM5 isoforms
were quantified by real-time PCR by previously described techniques (25) using
the primers and TaqMan probes shown in Table S1 in the supplemental material.
To serve as standards, cDNAs for each TRIM5 isoform were cloned into pCR-
TOPO2.1 vectors (Invitrogen). Thus, the same serial dilutions of linearized
plasmids could be used as standards for the quantification of total TRIM5
(primers in exons 3 and 4 and probe spanning the splice junction) and a given
TRIM5 isoform. The construction of standards for TRIM5?, TRIM5?, and
TRIM5? is described in the supplemental material. Results are expressed as
percentages [(number of copies of each TRIM5 isoform)/(total number of copies
of TRIM5 mRNA) ? 100].
Statistical analysis. Results are expressed as means ? standard errors of the
means (SEMs) unless otherwise indicated. Groups were compared by analysis of
variance. Posttest comparisons (performed if the P value was ?0.05) were made
using Bonferroni’s multiple-comparison test. Pairwise comparisons were made
using Student’s unpaired t test.
Expression of mRNAs coding for TRIM5 isoforms in human
cells. In initial studies, we produced cDNA from 6 different
human cell lines and used real-time PCR to quantify the ex-
pression of mRNAs coding for TRIM5 isoforms. In these stud-
ies, the total expression of spliced TRIM5 mRNA was assessed
by quantifying all TRIM5 transcripts containing both exons 3
and 4 but in which intron 3 had been removed. TRIM5? was
the most abundant isoform in all cells evaluated, but its ex-
pression represented only 37 to 52% of all TRIM5 mRNAs
(Fig. 2). No significant differences in TRIM5? mRNA were
observed by comparing the different cell types.
Expression of mRNA coding for TRIM5?, an isoform re-
sulting from the failure to remove intron 7, ranged from 3 to
8% of total TRIM5 mRNAs. Expression of TRIM5? was sig-
nificantly lower in H9 cells and MT4-R5 cells than in CEM
cells and P4 cells (P ? 0.05 for all comparisons).
An Alu sequence in the reverse orientation is present in the
TRIM5 locus. This creates an alternative splice acceptor site
whose use in TRIM5? results in the replacement of the C-ter-
minal 195 amino acids of the SPRY domain by a 28-amino-acid
sequence derived from the Alu sequence (see Fig. S1 in the
supplemental material). The expression of TRIM5? mRNA
represented ?1% of the total TRIM5 mRNAs in 5 of the 6 cell
lines tested, although expression in HeLa-derived P4 cells (4%
of the total) was significantly greater (P ? 0.05 for all compar-
isons). Such exonized Alu sequences have been observed in
numerous human genes, but the expression of isoforms carry-
ing exonized Alu sequences is typically low, as was observed for
TRIM5? (21, 40).
Because the three isoforms currently identified as reference
sequence (RefSeq) by the NCBI could explain ?70% of all
spliced TRIM5 mRNAs, we quantified additional alternatively
spliced variants identified in expressed sequence tag (EST)
FIG. 2. Quantification of mRNAs coding for TRIM5 isoforms.
RNA was purified from the indicated human cell lines, and cDNA was
prepared using random hexamers and MMLV reverse transcriptase.
cDNA coding for each of the indicated TRIM5 isoforms was quanti-
fied using specific primers and TaqMan probes, and the results are
expressed as percentages of total TRIM5 cDNA, as measured using
primers and probes that detect sequences spanning exons 3 and 4.
Shown are the means ? SEMs for two (TRIM5?) or three (all other
isoforms) experiments, each performed using different cell prepara-
tions. Note that the y axis scales are different in the different graphs.
7830 BATTIVELLI ET AL.J. VIROL.
libraries. An isoform resulting from the failure to remove in-
tron 4 was present in all cell types. This isoform, which we have
named TRIM5-iota (TRIM5?) in keeping with the nomencla-
ture of Brennan et al. (4), represented 9 to 29% of total
TRIM5 mRNAs and therefore was the second most abundant
isoform after TRIM5?. TRIM5? expression was significantly
higher in MT4-R5 cells than in H9 cells (P ? 0.05).
The isoform resulting from skipping exon 5 and the failure
to remove intron 6 (TRIM5?) was detectable in all cell lines
but represented less than 0.2% of all TRIM5 mRNAs. The
isoform resulting from skipping exon 5 and subsequently splic-
ing as for TRIM5? (TRIM5ε) would be detected by the prim-
ers and probe used to quantify TRIM5?, but because this
message should be susceptible to nonsense-mediated degrada-
tion (6), it is unlikely to be present in substantial quantities. An
isoform resulting from skipping exon 6 and subsequently splic-
ing as for TRIM5? has been described (variant h), but it would
be detected by the primers and probe used to quantify
TRIM5?. Finally, an isoform resulting from skipping exon 5
and subsequently splicing as for TRIM5? has been identified
(variant e); this variant is also subject to nonsense-mediated
degradation and would be detected by the primers and probe
used to quantify TRIM5?.
The amino acid sequences of all these variants, including
TRIM5? and TRIM5?, are identical to that of TRIM5? up
to the splice donor site in exon 4 used by TRIM5?. Thus, all
of these proteins contain identical RING, B-box 2 and
coiled-coil regions and also express the same 17 amino acids
corresponding to the N-terminal portion of the L2 linker in
TRIM5? (Fig. 1).
The proportions of TRIM5? and TRIM5? transcripts in nat-
ural HIV-1 target cells were similar to those observed in the
human cell lines. These transcripts represented, respectively,
55.6% ? 9.5% and 10.2% ? 3.6% of total TRIM5 transcripts
in mitogen-activated human CD4?T cells (n ? 4) and
46.2% ? 1.7% and 13.7% ? 2.4% of total TRIM5 transcripts
in monocyte-derived human macrophages (n ? 2).
We also attempted to detect expression of the TRIM5? pro-
tein. When extracts from U373-X4 cells overexpressing
TRIM5? were evaluated by Western blotting using antibodies
recognizing epitopes in the N-terminal region of human
TRIM5, TRIM5? expression was easily detectable, but the af-
finity of these antibodies was not sufficient to detect either
TRIM5? or TRIM5? in untransduced cells (data not shown).
Thus, this approach could not be used to confirm that the
TRIM5? protein was expressed in human cell lines. Studies
evaluating U373-X4 cells in which TRIM5? mRNA had been
knocked down by using a small interfering RNA (siRNA)
approach did indicate that functionally significant levels of
TRIM5? protein were expressed by these cells (see below).
TRIM5? and TRIM5? do not inhibit early steps in HIV-1
replication. CRFK cells overexpressing human TRIM5? in-
hibit the infectivity of the NL4-3 HIV-1 strain by 2-fold (Fig.
3). As previously described, the recombinant HIV-1 strains
NRC2 and NRC10, which express Gag-protease sequences
derived from clinical isolates, showed greater sensitivity to
TRIM5? (1). TRIM5? and TRIM5?, which lack a SPRY do-
main, or TRIM5 proteins in which the SPRY domain has been
truncated have been previously shown to not inhibit HIV-1
infectivity (2, 23, 29, 30, 41, 49). In both TRIM5? and TRIM5?,
the SPRY domain is replaced by a short C-terminal sequence.
As expected, CRFK cells overexpressing TRIM5? (Fig. 3) or
TRIM5? (data not shown) did not inhibit HIV-1 replication,
including the viruses that were relatively sensitive to TRIM5?.
Following the inhibition of CypA-CA interactions by 5 ?M
Debio-025, TRIM5? or TRIM5? still had no effect on HIV-1
replication (data not shown).
Overexpression of TRIM5? inhibits the antiviral activity of
TRIM5?. TRIM5 isoforms lacking a SPRY domain, including
TRIM5?, TRIM5?, and TRIM5? ?SPRY, have been shown to
multimerize with cognate TRIM5? and inhibit its activity (2, 23,
express an intact L2 linker region, and Javanbakht et al. have
shown that the L2 region contributes to the multimerization of
TRIM5? (16). Thus, it was unclear whether TRIM5? would also
have dominant-negative effects on TRIM5?. To evaluate this
question, we stably expressed TRIM5? and TRIM5? in U373-X4
cells and evaluated their effect on the ability of TRIM5? consti-
tutively expressed by this cell line to inhibit N-MLV replication.
U373-X4 cells or those transduced with a control vector inhibited
N-MLV infection, but not B-MLV infection, by ?2 logs, an effect
that is entirely due to the expression of TRIM5? by this cell line
(1). As previously reported, overexpression of TRIM5? or HA-
tagged TRIM5? completely inhibited TRIM5? activity (Fig. 4).
TRIM5? also had a strong inhibitory effect on TRIM5?. The
replication of B-MLV was somewhat greater than that of N-MLV
for cells expressing HA-tagged TRIM5? (P ? 0.05). Western
blotting using an antibody recognizing HA-tagged proteins
showed that both HA-TRIM5? and HA-TRIM5? were strongly
expressed and that the level of expression of HA-TRIM5? was
somewhat greater than that of HA-TRIM5? (data not shown).
Inhibiting TRIM5? expression improves the activity of
TRIM5?. To evaluate whether physiological levels of TRIM5?
expression could also inhibit TRIM5? activity against HIV-1,
we produced U373-X4 cell lines stably expressing pre-miRNA
FIG. 3. TRIM5? does not inhibit HIV-1 replication. Untransduced
CRFK cells and those transduced with lentiviral vectors resulting in
expression of LacZ, HA-tagged TRIM5?, or HA-tagged TRIM5? were
infected with VSV-pseudotyped recombinant HIV-1 strains carrying
sequences encoding Renilla luciferase in the place of nef and Gag-
protease sequences derived from NL4-3 or clinical isolates (NRC2 and
NRC10). Forty hours after infection, medium was removed, the cells
were lysed, and luciferase activity was determined by luminometry.
The results, expressed relative to those obtained for untransduced
CRFK cells, are the means ? SEMs for three experiments. ?-gal,
VOL. 85, 2011 TRIM5 ISOFORMS IN HUMAN CELLS7831
sequences targeting sequences in the coding region (TRIM5?-
KD-coding) or 3? untranslated region (TRIM5?-KD-UTR) of
TRIM5? and compared the sensitivities of these cells and
U373-X4 cells transduced with a control pre-miRNA construct
to infection with HIV-1. Evaluation of TRIM5 mRNA levels in
these stably transduced cell lines by real-time PCR indicated
that both TRIM5?-knockdown constructs reduced TRIM5?
mRNA levels by 69 to 76% compared to that of cells express-
ing the control pre-miRNA (mean ? SEM, 73.5% ? 1.2%;
n ? 4) Expression of the miRNAs targeting TRIM5? had no
significant effect on the expression of TRIM5? (0.1% ? 6.7%
reduction in mRNA levels; n ? 4).
As shown in Fig. 5A, knockdown of TRIM5? expression
increased the ability of U373-X4 cells to inhibit HIV-1 infec-
tion. Consistent with the idea that physiological levels of
TRIM5? were inhibiting TRIM5? activity, TRIM5? knockdown
led to a 2-fold inhibition of the infectivity of NL4-3 but a 3-fold
inhibition of the infectivity of NRC2 and NRC10 (P ? 0.001
for both comparisons), viruses that are more sensitive to hu-
man TRIM5?. Transducing cells stably expressing the pre-
miRNA targeting the TRIM5? 3?-UTR with a vector resulting
in expression of TRIM5? lacking this target sequence improved
the infectivity of HIV-1 in these cell lines (Fig. 5B). Taken
together, these results indicate that even physiological levels of
TRIM5? expression can inhibit TRIM5? activity in U373-X4
The ability of TRIM5? to inhibit the early steps of HIV-1
replication is dependent on the recognition of the incoming
capsid by its C-terminal SPRY domain, whereas TRIM5 iso-
forms lacking a SPRY domain can inhibit TRIM5? activity
through the formation of heterodimers. In this study, we dem-
onstrated that mRNAs coding for TRIM5? represent only half
of all TRIM5 transcripts present in human cells, including
CD4?lymphocytes and macrophages, and that TRIM5?, an
isoform lacking a SPRY domain, is the second most abundant
isoform expressed. Like TRIM5? and TRIM5?, TRIM5? does
not inhibit the early steps of HIV-1 replication but does inhibit
the antiviral activity of TRIM5?. Specific knockdown of
TRIM5? improved the antiviral activity of endogenous levels of
FIG. 4. Overexpression of TRIM5? and TRIM5? inhibits TRIM5? activity. U373-X4 cells were transduced with retroviral vectors, resulting in
the expression of LacZ, HA-tagged or untagged TRIM5?, or HA-tagged or untagged TRIM5?, as indicated. Stably transduced cell lines were
selected, and 5 ? 104cells were infected with the indicated amounts of VSV-pseudotyped vectors expressing Gag sequences of N-MLV or B-MLV.
YFP-expressing cells were quantified by cytofluorometry 40 h after infection. Titers of MLV vectors were determined by using CRFK cells, and
results are expressed as the amounts required to infect 50% of cells in tissue culture (50% tissue culture infective dose [TCID50]). Results are the
means ? SEMs of 3 (cells expressing HA-tagged isoforms) or 4 (other cell lines) independent experiments.***indicates a P value of ?0.001 and
*indicates a P value of ?0.05 in a comparison of N-MLV and B-MLV.
7832 BATTIVELLI ET AL.J. VIROL.