Ti

MLC

a

ARRAA

KTRSUNA

1

k2T

lBPm

(

h0

Virus Research 184 (2014) 30–38

Contents lists available at ScienceDirect

Virus Research

j ourna l ho me pa g e: www.elsev ier .com/ locate /v i rusres

he conserved sumoylation consensus site in TRIM5� modulates itsmmune activation functions

arie-Édith Nepveu-Traversy, Lionel Berthoux ∗

aboratory of Retrovirology, Department of Medical Biology and BioMed Research Group, Université du Québec à Trois-Rivières, 3351 Boulevard des Forges,P500, Trois-Rivières, QC G9A 5H7, Canada

r t i c l e i n f o

rticle history:eceived 22 October 2013eceived in revised form 19 January 2014ccepted 17 February 2014vailable online 25 February 2014

eywords:RIM5�estrictionumoylationbiquitylationF-�BP-1

a b s t r a c t

TRIM5� is a type I interferon-stimulated anti-retroviral restriction factor expressed in most primates andhomologous proteins are expressed in other mammals. Through its C-terminal PRYSPRY (B30.2) domain,TRIM5� binds to incoming and intact post-fusion retroviral cores in the cytoplasm. Following this directinteraction, the retroviral capsid core is destabilized and progression of the virus life cycle is interrupted.Specific recognition of its viral target by TRIM5� also triggers the induction of an antiviral state involvingthe activation of transcription factors NF-�B- and AP-1. In addition to PRYSPRY, several other TRIM5�domains are important for anti-retroviral function, including a RING zinc-binding motif. This domainhas “E3” ubiquitin ligase activity and is involved in both the direct inhibition of incoming retrovirusesand innate immune activation. A highly conserved sumoylation consensus site is present between theRING motif and the N-terminal extremity of TRIM5�. No clear role in restriction has been mapped tothis sumoylation site, and no sumoylated forms of TRIM5� have been observed. Here we confirm thatmutating the putatively sumoylated lysine (K10) of the Rhesus macaque TRIM5� (TRIM5�Rh) to an argi-nine has only a small effect on restriction. However, we show that the mutation significantly decreasesthe TRIM5�-induced generation of free K63-linked ubiquitin chains, an intermediate in the activationof innate immunity pathways. Accordingly, K10R decreases TRIM5�-mediated activation of both NF-�Band AP-1. Concomitantly, we find that K10R causes a large increase in the levels of ubiquitylated TRIM5�.Finally, treatment with the nuclear export inhibitor leptomycin B shows that K10R enhances the nuclearlocalization of TRIM5�Rh, while at the same time reducing its level of association with nuclear SUMO bod-

ies. In conclusion, the TRIM5� sumoylation site appears to modulate the E3 ubiquitin ligase activities ofthe adjacent RING domain, promoting K63-linked ubiquitin chains at the expense of auto-ubiquitylationwhich is probably K48-linked. Consistently, we find this sumoylation site to be important for innateimmune activation by TRIM5�. In addition, lysine 10 regulates TRIM5� nuclear shuttling and nuclearlocalization, which may also be related to its role in innate immunity activation.

. Introduction

Proteins belonging to the TRIM5 (tripartite motif 5) family are

nown for their species-specific anti-retroviral activity (Sayah et al.,004; Stremlau et al., 2004; Towers, 2007). Restriction mediated byRIM5 proteins relies mainly on their ability to recognize incoming

Abbreviations: CA, capsid; HEK, human embryonic kidney; HTLV, human T-celleukemia virus; IF, immunofluorescence; IP, immunoprecipitation; LMB, leptomycin; MOI, multiplicity of infection; NB, nuclear body; PBS, phosphate buffer saline;ML, promyelocytic leukemia; SUMO, small ubiquitin-like modifier; TRIM, tripartiteotif; VSV G, glycoprotein of the vesicular stomatitis virus; WT, wild type.∗ Corresponding author. Tel.: +1 819 376 5011x4466; fax: +1 819 376 5084.

E-mail addresses: nepveutr@uqtr.ca (M.-É. Nepveu-Traversy), berthoux@uqtr.caL. Berthoux).

ttp://dx.doi.org/10.1016/j.virusres.2014.02.013168-1702/© 2014 Elsevier B.V. All rights reserved.

© 2014 Elsevier B.V. All rights reserved.

intact capsids early after entry of the virus (Sebastian and Luban,2005). TRIM5� orthologs differ from each other mostly in the C-terminal hypervariable PRYSPRY domain that confers specificity toretroviral capsid (CA) (Keckesova et al., 2004; Perron et al., 2004;Song et al., 2005; Yap et al., 2004), e.g. HIV-1 is inhibited by rhesusmacaque TRIM5� (TRIM5�Rh), but not by the human version of theprotein (Yap et al., 2004). Interactions between TRIM5� proteinsand CA lead to the inhibition of early steps of the virus life cycleby different mechanisms: disruption of capsid integrity (Black andAiken, 2010; Perron et al., 2007; Stremlau et al., 2006), proteasomaldegradation (Kutluay et al., 2013; Lukic et al., 2011; Rold and Aiken,

2008) and impairment of nuclear import (Campbell et al., 2007).In addition, recent studies have shown that TRIM5�-CA interac-tions or the over-expression of TRIM5� lead to the induction ofan antiviral state that includes the activation of NF-�B and AP-1

ux / V

(TK

Pdip2tst2rcK2

tisimcntetswt(w

silRtsoi

2

2

o(wpGTcme2HTAiKlw

M.-É. Nepveu-Traversy, L. Bertho

Pertel et al., 2011; Tareen and Emerman, 2011b; Uchil et al., 2013).his pathway also involves the generation of “free” (unconjugated)63-linked ubiquitin chains (Pertel et al., 2011).

TRIM5� contains several domains located upstream of theRYSPRY domain, including RING, B-Box-2 and Coiled-coil (RBCC)omains (Reymond et al., 2001). The RING domain has an

ntrinsic E3 ubiquitin ligase activity essential for the auto-olyubiquitylation of the protein (Diaz-Griffero et al., 2006; Li et al.,013; Lienlaf et al., 2011; Maegawa et al., 2010), which is assumedo proceed through K48-linked ubiquitin chains. This process istimulated by contact with restriction-sensitive viruses and it con-ributes to the destabilization of retroviral cores (Kutluay et al.,013; Rold and Aiken, 2008). Deleting the RING domain or dis-upting its zinc-binding properties by mutagenesis abrogates theapacity of TRIM5� to activate NF-�B and AP-1 and to stimulate63-linked poly-ubiquitylation (Pertel et al., 2011; Uchil et al.,013).

Similar to the ubiquitylation process, sumoylation is a post-ranslational modification involved in many cellular mechanismsncluding cell signaling, transcription and regulation of proteintability (Zhao, 2007). Three main isoforms of SUMO are foundn mammalian cells, including SUMO-1, which can only achieve

ono-sumoylation, and SUMO-2/3 that are able to form SUMOhains (Vertegaal, 2007). SUMO proteins are localized mostly in theucleus and modify target proteins through covalent attachmento a lysine present in a specific consensus site, �-K-X-D/E (Gocket al., 2005; Sampson et al., 2001). TRIM5� carries such a sumoyla-ion consensus motif, that is conserved in all known orthologs (nothown), upstream of the RING domain. Despite the fact that TRIM5�as never identified as a sumoylated protein, recent publica-

ions proposed a role for SUMO-1 in TRIM5�-mediated restrictionArriagada et al., 2011; Lukic et al., 2013). However, this conclusionas disputed by another report (Brandariz-Nunez et al., 2013).

The main objective of this study was to determine whether theumoylation consensus site could have a role in TRIM5�-mediatedmmune activation. Our results suggest that the putatively sumoy-ated lysine 10 (K10) of TRIM5�Rh modulates the activity of theING domain by decreasing auto-ubiquitylation while increasinghe generation of K63-linked ubiquitylation. In conclusion, the con-ensus sumoylation site of TRIM5�Rh is involved in the regulationf ubiquitylation pathways that control the activation of innatemmunity.

. Materials and methods

.1. Plasmid DNAs and mutagenesis

pMIP-TRIM5�Rh expresses a C-terminal FLAG-tagged versionf the Rhesus macaque TRIM5� and has been described beforeBérubé et al., 2007; Sebastian et al., 2006). Directed mutagenesisas performed to introduce K10R, using the following mutagenicrimers: 5′-TTCCTCGAGATGGCTTCTGGAATCCTGCTTAATGTAAGG-AGGAGGTGACCTGT (forward) and 5′-TCCTGAATTCTTACTTA-CGTCGTCATCCTTGTAATC (reverse). The K10R substitution wasonfirmed by Sanger sequencing. The vector production plas-ids pMD-G, p�R8.9, pCL-Eco and pTRIP-CMV-GFP have all been

xtensively described elsewhere (Aiken, 1997; Berthoux et al.,003, 2004; Naviaux et al., 1996; Zufferey et al., 1997). pRK5-A-Ubiquitin WT and KO (Lim et al., 2005) were obtained fromed Dawson (Johns Hopkins University, Baltimore, MD) throughddgene. KO ubiquitin bears the following mutations eliminat-

ng all possibilities of ubiquitin chain formation: K6R, K11R, K27R,29R, K33R, K48R and K63R. pCEP4-NF-�B-Luc expresses firefly

uciferase under the control of an NF-�B-dependent promoter,hile pCEP4-�NF-�B-Luc is transcriptionally deficient due to the

irus Research 184 (2014) 30–38 31

deletion of the NF-�B binding site (Tareen and Emerman, 2011a).Both constructs were kind gifts from M. Emerman (University ofWashington, Seattle, WA). pHTS-AP1, a kind gift from J. Luban(University of Massachusetts Medical School, Worcester, MA),expresses firefly luciferase under control of an AP-1-dependentpromoter (Pertel et al., 2011).

2.2. Cell lines

Human embryonic kidney (HEK) HEK293 T cells, HeLa humanadenocarcinoma cervical cells and CRFK feline kidney cells weremaintained in Dulbecco’s modified Eagle’s medium supplementedwith 10% fetal bovine serum and antibiotics at 37 ◦C, 5% CO2. All cellculture reagents were from Hyclone (Thermo Scientific, Logan, UT,USA).

2.3. Virus production

MLV and HIV-1-based vectors were produced through tran-sient transfection of HEK293 T cells using polyethylenimine (MW25,000; Polysciences, Warrington, PA) and collected as previouslydescribed (Bérubé et al., 2007; Pham et al., 2010). To produce theMLV-based MIP vectors, cells were transfected with the relevantpMIP plasmid and co-transfected with pCL-Eco and pMD-G. All sta-bly transduced cell lines were produced as previously described(Bérubé et al., 2007; Pham et al., 2010). Successfully trans-duced cells were selected with puromycin, using concentrationsof 2 �g/ml (HeLa cells) and 4 �g/ml (CRFK cells). To produce GFP-expressing HIV-1 vectors, cells were co-transfected with p�R8.9,pMD-G and pTRIP-CMV-GFP as described previously (Bérubé et al.,2007).

2.4. Viral challenges

HeLa and CRFK cells were plated in 24-well plates at 50,000cells per well and infected the next day with different amountsof HIV-1TRIP-CMV-GFP (nicknamed “HIV-1-GFP”) or using a definedmultiplicity of infection (MOI). Two days post-infection, cells weretrypsinized and fixed in 2% formaldehyde in a PBS solution. The% of GFP-positive cells was then determined by analyzing 10,000cells using a FC500 MPL cytometer with the CXP software (BeckmanCoulter).

2.5. Western blotting

Cells were lysed in RIPA buffer (NaCl 150 mM, Triton 1%, SDS0.1%, TRIS 50 mM (pH 8.0), sodium deoxycholate 0.5% and Com-plete protease inhibitor cocktail (Roche, Bale, Switzerland)). Wholecell lysates were then boiled in protein sample buffer (60 mMTris–Cl pH6.8, 10% glycerol, 0.002% bromophenol blue, 2% SDS, 2%beta-mercaptoethanol) and resolved by SDS-PAGE. After transfer tonitrocellulose membranes, blots were probed with various antibod-ies as specified throughout the text and visualized using secondaryantibodies coupled to horseradish peroxydase (Santa Cruz, Dal-las, TX) and a chemiluminescence detection system (SuperSignalWest Femto, Thermo scientific, Waltham, MA). The horse radishperoxidase-conjugated anti-actin antibody (Santa Cruz) was usedto control for equal loading across lanes. Images were recordedon a UVP (Upland, California) EC3 imaging system, and densito-metry analyses were performed using the area density tool of theVisionWorks LS software (UVP).

2.6. Immunoprecipitation and auto-ubiquitylation

HEK293 T cells were co-transfected with either WT or mutantFLAG-tagged TRIM5�, and a plasmid expressing HA-tagged WT

32 M.-É. Nepveu-Traversy, L. Berthoux / Virus Research 184 (2014) 30–38

Fig. 1. Contribution of the TRIM5�Rh sumoylation consensus site to HIV-1 restriction. (A) HeLa cells or (B) CRFK cells stably expressing WT TRIM5�Rh or the K10R mutantw “HIV-t in bote same

uw1t11bDp1tewrC(

2

wTppBs((

ere challenged with increasing amounts of VSV G-pseudotyped HIV-1TRIP-CMV-GFP (wo days later. WT and K10R TRIM5�Rh protein expression levels were analyzed

xpression was analyzed as a loading control. The images show two sections of the

biquitin (pRK5-HA-Ubiquitin-WT). The “empty” vector pMIPas used as a negative control. Cells were infected with HIV-

TRIP-CMV-GFP (HIV-1-GFP; MOI of 0.2) 6 h before lysis. Two days postransfection, cells were harvested, lysed with the IP buffer (NaCl50 mM, Tris-Cl 30 mM (pH 7.5), sodium deoxycholate 0.5%, NP-40% and Complete protease inhibitor cocktail (Roche)) and incu-ated overnight with a polyclonal rabbit anti-FLAG (Cell Signaling,anvers, MA) used at a 1:50 dilution. Sepharose beads coupled torotein A (SIGMA, St Louis, MI) were then added and incubated for0–16 h. Beads were pelleted by centrifugation, washed 4 times inhe IP buffer and then resuspended in protein sample buffer. West-rn blots were performed as described above. Blots were probedith a monoclonal mouse anti-FLAG antibody (M2, SIGMA) and

evealed using a goat anti-mouse HRP secondary antibody (Santaruz). To detect ubiquitin, a rabbit polyclonal antibody against HASanta Cruz) was used.

.7. NF-�B and AP-1 reporter assay

HEK293 T cells were plated in 12-well plates and co-transfectedith increasing amounts of either WT or mutant FLAG-tagged

RIM5� and a fixed concentration of either pCEP4-NF-�B-Luc,CEP4-�NF-�B-Luc or pHTS-AP1. Cells were lysed with RIPA 48 host transfection and assessed for luciferase activity using the

rightGlow Luciferase kit (Promega). Luminescence was mea-ured with a Synergy HT multi-detection microplate readerBioTek, Winooski, VT) and analyzed with the Gen5 softwareBioTek).

1-GFP”) in triplicates. The percentage of cells expressing GFP was analyzed by FACSh cell lines by western blotting using an anti-FLAG antibody (right panels). Actin

gel that were cut and pasted together to exclude irrelevant lanes.

2.8. K63 ubiquitin chains formation assay

HEK293T cells were seeded in 12-well plates and co-transfectedwith either WT or mutant FLAG-tagged TRIM5 and either WT or KOpRK5-HA-ubiquitin. The “empty” plasmid pMIP was used as a neg-ative control. Cells were lysed in RIPA buffer 48 h post-transfectionand processed for western blotting. A monoclonal human/rabbitchimeric K63-linked ubiquitin chains-specific antibody (clone Apu,Millipore) was used (1:1000 dilution) to detect K63 ubiquitinchains.

2.9. Immunofluorescence microscopy

60,000 CRFK cells were plated on microscope cover glasses(Fisherbrand) in 6-well plates. The next day, cells were exposed toleptomycin B (LMB; 20 ng/ml, Enzo LifeScience) for 12 h. Cells werewashed with PBS, fixed for 10 min in 4% formaldehyde-PBS at 37 ◦C,washed three times in PBS and permeabilized with 0.1% Triton X-100 for 2 min on ice. Cells were then washed twice with PBS andtreated with 10% normal goat serum (Vector laboratories) in PBSfor 30 min at RT. This saturation step was followed by a 4 h incu-bation with a rabbit polyclonal FLAG antibody (Cell Signaling) andan antibody against endogenous SUMO-1 (mouse monoclonal anti-body, Invitrogen), both at a 1:200 dilution in PBS with 10% normalgoat serum. Fluorescent staining was done by incubating with an

Alexa594-conjugated goat anti-rabbit antibody and an Alexa488-conjugated goat anti-mouse antibody (each diluted 1:200). Cellswere washed 4 times in PBS before mounting in Vectashield (VectorLaboratories). Hoechst-33342 (0.8 �g/ml; Molecular Probes) was

M.-É. Nepveu-Traversy, L. Berthoux / Virus Research 184 (2014) 30–38 33

Fig. 2. The K10R mutation increases TRIM5�Rh auto-ubiquitylation levels. HEK293Tcells were co-transfected with WT or K10R FLAG-tagged TRIM5�Rh and with an HA-tagged version of human ubiquitin. Two days later, transfected cells were infectedwith HIV-1TRIP-CMV-GFP (MOI of 0.2) for 6 h before lysis, or left uninfected. Con-trols (CTL) were transfected with WT FLAG-TRIM5�Rh alone or with HA-ubiquitinalone and were not infected. FLAG-tagged proteins were immunoprecipitated asdescribed in Section 2. The IP products were analyzed by western blotting first usingala

awwp

3

3r

TpBrc(W5(iTd

3

aeTHSipttfaitc

Fig. 3. The K10R mutation decreases the ability of TRIM5�Rh to form K63-linkedubiquitin chains. (A) HEK293T cells were transfected with either the “empty”pMIP or with pMIP expressing FLAG-tagged WT or K10R pMIP-TRIM5�Rh, and co-transfected with the plasmid expressing WT or KO HA-Ubiquitin. KO Ubiquitin hasall its lysines mutated to arginines. Two days later, whole protein lysates were ana-lyzed by western blotting using rabbit antibodies directed at K63-linked ubiquitinchain or at FLAG. (B) Density of bands detected using the K63 chains-specific anti-body (ranging from 75 to >250 kDa) normalized to the control consisting of cellsco-transfected with WT or KO HA-Ub and the “empty” pMIP. Graph shows the aver-age data from 4 independent experiments with standard deviations, and P-values

mouse anti-FLAG antibody and subsequently with a rabbit anti-HA antibody. Allanes shown are from a single western blot experiment and were exposed the samemount of time but some irrelevant lanes were cut out.

dded along with the penultimate PBS wash to reveal DNA. Picturesere generated using a Zeiss AxioObserver microscope equippedith an Apotome module and the Axiovision software. Imagingarameters were set to identical values across samples.

. Results

.1. Mutating lysine 10 decreases the capacity of TRIM5˛Rh toestrict HIV-1 in HeLa and CRFK cells

Human HeLa and feline CRFK cells stably expressing WT or K10RRIM5�Rh were challenged with increasing amounts of a VSV G-seudotyped HIV-1 vector expressing GFP (HIV-1-GFP) (Fig. 1A and). Cells transduced with the “empty” vector were used as a non-estrictive control. WT and K10R protein expression levels wereomparable in either cell line, as analyzed by western blottingFig. 1A and B; see Suppl. data for the uncut version). As expected,

T TRIM5�Rh strongly inhibits HIV-1 infection in both cell lines, by0- to 100-fold on average. The K10R mutation slightly increasedup to 5-fold and 3-fold in HeLa and CRFK cells, respectively) thenfectious power of HIV-1-GFP compared to cells expressing WTRIM5�Rh (Fig. 1A and B). Thus, K10R causes a significant but smallecrease in the restriction of HIV-1.

.2. The K10R mutation increases TRIM5˛Rh auto-ubiquitylation

The level of TRIM5�Rh auto-ubiquitylation was assessed usingn IP-based protocol as previously published by others (Lienlaft al., 2011; Maegawa et al., 2010). FLAG-tagged WT and K10RRIM5�Rh were transiently expressed by transfection of humanEK293 T cells with the corresponding pMIP-based plasmids.imultaneously, cells were co-transfected with a construct express-ng HA-ubiquitin. 2 days later, cells were lysed and FLAG-taggedroteins were pulled down. Cells transfected with the empty vec-or were used as a negative IP control (Fig. 2; see Suppl. data forhe uncut version). IP pellets were analyzed by western blottingor the presence of FLAG-tagged and HA-tagged proteins. Using

FLAG antibody, we could detect a ∼55 kDa band correspond-ng to TRIM5�Rh in all lanes except for the cells transfected withhe empty vector. When FLAG-TRIM5�Rh and HA-ubiquitin wereo-transfected, ubiquitylated proteins were detected within the

were calculated using a two-way ANOVA test.

∼70 to 300 kDa range using an HA antibody (Fig. 2), including acharacteristic “ladder” between ∼70 and 100 kDa. No detectableubiquitylation products were seen in the absence of HA-ubiquitin,and a faint amount were detected in lysates from cells transfectedwith HA-ubiquitin alone, suggesting low levels of unspecific IPof non-FLAG-tagged ubiquitylated proteins in our experimentalconditions (Fig. 2). The amounts of unconjugated TRIM5�Rh werelower in cells co-transfected with TRIM5�Rh and HA-ubiquitin,compared with cells transfected with TRIM5�Rh alone (Fig. 2).Obviously, co-transfecting TRIM5�Rh and ubiquitin stimulatedubiquitylation of the former by the latter, thus resulting in theobserved decrease in unconjugated products. In the presence ofTRIM5�Rh-sensitive HIV-1, poly-ubiquitylation of WT TRIM5�Rhwas markedly increased, confirming that interactions betweenTRIM5� and incoming capsids result in an increase in TRIM5�turnover (Rold and Aiken, 2008). We found that K10R also causedan increase in the poly-ubiquitylation of TRIM5�Rh proteins, alongwith a decrease in the levels of unconjugated protein (Fig. 2),suggesting that the mutation increased RING-dependent auto-ubiquitylation. Interestingly, exposure of cells expressing K10RTRIM5�Rh to HIV-1 further amplified poly-ubiquitylation (Fig. 2). Insummary, the data in Fig. 2 show that mutating the putative sumoy-

lation site in TRIM5�Rh increases the auto-ubiquitylation activityof the adjacent RING domain, albeit without reaching saturationlevels.

34 M.-É. Nepveu-Traversy, L. Berthoux / Virus Research 184 (2014) 30–38

Fig. 4. The K10R mutation affects TRIM5�Rh-dependent activation of NF-�B. (A) HEK293T cells were transfected with increasing amounts of either the WT or the indicatedmutants of pMIP-TRIM5�Rh and co-transfected with a constant amount (0.6 �g) of a reporter plasmid for NF-�B activity expressing luciferase. As controls, cells were co-transfected with the empty pMIP plasmid and the NF-�B reporter construct, or were co-transfected with WT pMIP-TRIM5�Rh and the activation-deficient mutant of thereporter construct (�NF-�B-Luc). 2 days later, cells were lysed and the luciferase activity was measured. Bars show the average from triplicate transfections with standarddeviations. Statistical analysis was done using a 2-way ANOVA test from an experiment done in triplicate. (B) The cellular lysates prepared to quantify luciferase activity in( 3 lysaa nd B.

3s

euweocpbate3

A) were also used to analyze TRIM5� expression. For this, we pooled together then anti-FLAG antibody. (C and D) Repetition of the experiment shown in panels A a

.3. The K10R mutation affects the ability of TRIM5˛Rh totimulate the formation of K63-linked poly-ubiquitin chains

It was previously demonstrated that the RING domain wasssential for TRIM5� to stimulate the formation of K63-linkedbiquitin chains, which contribute to the NF-�B signaling path-ay activation through phosphorylation of the TAK1 kinase (Pertel

t al., 2011). To analyze the consequence of the K10R mutationn this TRIM5� function, FLAG-TRIM5�Rh and HA-ubiquitin wereo-transfected in HEK293 T cells and the presence of K63-linkedoly-ubiquitin was assessed by western blotting using an anti-ody specific for this type of ubiquitylated chains (Fig. 3A). As

control we used KO-Ub, a version of ubiquitin bearing lysine-o-arginine mutations at all potentially ubiquitylated lysines (Limt al., 2005). K63-linked chains were detected between ∼70 and00 kDa, and such poly-ubiquitylation products were present in

tes from each triplicate transfection and analyzed them by western blotting using

much smaller amounts in cells transfected with KO HA-Ub (Fig. 3A).A ∼130 kDa unspecific band was detected in all lanes at a sim-ilar density and thus probably represents a non-ubiquitylatedprotein recognized by the antibody we used. We performeda densitometry-based quantitative analysis of K63-linked poly-ubiquitylation that removed the artifact caused by the unspecificband at 130 kDa (Fig. 3B), using blots from 3 different experiments.Co-transfection of WT TRIM5�Rh and WT ubiquitin resulted in astrong stimulation of K63-linked poly-ubiquitylation, comparedwith cells co-transfected with WT TRIM5�Rh and KO ubiquitin orcompared with cells co-transfected with the empty pMIP and WTubiquitin (Fig. 3A and B). In contrast, K10R-TRIM5�Rh activated the

generation of K63-linked chains only poorly when co-transfectedwith WT ubiquitin (Fig. 3A and B). As exemplified in the experi-ment shown in Fig. 3A, there was no association between TRIM5�Rhexpression levels and induction of K63-linked chains. Indeed, in this

ux / Virus Research 184 (2014) 30–38 35

pewtK

3a

�p2lKtmcmpadbBup2wyceiawowTtwwfatamTwwaeN

3a

eadl21losw

Fig. 5. The K10R mutation affects TRIM5�Rh-dependent activation of AP-1. (A)HEK293T cells were transfected with 1 or 3 �g of either the WT or the indicatedmutants of pMIP-TRIM5�Rh and co-transfected with a constant amount (0.6 �g) ofa reporter plasmid for AP-1 activity expressing luciferase. As a control, cells wereco-transfected with the empty pMIP plasmid and the AP-1 reporter construct. 2 dayslater, cells were lysed and the luciferase activity was measured. Bars show the aver-age from triplicate transfections with standard deviations. Statistical analysis wasdone using a 2-way ANOVA test from an experiment done in triplicate. (B) The cel-

M.-É. Nepveu-Traversy, L. Bertho

articular experiment K10R was clearly expressed at higher lev-ls than the WT control, yet less K63-linked poly-ubiquitin chainsere detected. Therefore, the consensus sumoylation site is impor-

ant for the capacity of TRIM5�Rh to stimulate the production of63-linked ubiquitin chains.

.4. The K10R mutation decreases the ability of TRIM5˛Rh toctivate the NF-�B pathway

TRIM5�Rh activates an innate immunity pathway that is NF-B activation-dependent and that can contribute to increasing theroduction of interferons and other antiviral factors (Pertel et al.,011; Tareen and Emerman, 2011a; Uchil et al., 2013). The lower

evels of K63-linked polyubiquitin shown in Fig. 3 indicated that10R was less efficient at inducing K63-linked ubiquitylation and

hus one would expect reduced levels of NF-�B activation with thisutant. To investigate this possibility, we co-transfected HEK293T

ells with a mixture of WT or mutant TRIM5�Rh and a reporter plas-id expressing luciferase under the control of an NF-�B-dependent

romoter. As a control, WT TRIM5�Rh was also co-transfected withn NF-�B binding domain-deleted control (Fig. 4). The C35A RINGomain mutant of TRIM5�Rh was also included, as it is known toe deficient in NF-�B activation (Tareen and Emerman, 2011a).ecause optimal transactivation of luciferase expression dependspon the relative ratios between the luciferase-expressing reporterlasmid and the plasmid expressing TRIM5� (Tareen and Emerman,011a), we co-transfected increasing amounts of the latter plasmidith a fixed amount of the former. We performed a statistical anal-

sis of luciferase activity from triplicate transfections (Fig. 4A) andoncomitantly analyzed TRIM5� expression levels (Fig. 4B). In gen-ral, increasing the amount of transfected TRIM5�Rh resulted in anncrease in luciferase expression, but luciferase activity was lowert 3 �g of WT or K10R TRIM5�Rh-expressing plasmid comparedith 1 �g, probably reflecting a saturation effect as reported by

thers (Tareen and Emerman, 2011a). As expected, C35A TRIM5�Rhas a weak activator of NF-�B-dependent transactivation (Fig. 4A).

he K10R mutation caused a ∼20–30% reduction in NF-�B activa-ion compared with the WT control when 1 or 3 �g of TRIM5�Rhere transfected (Fig. 4A). However, this effect was not observedhen smaller amounts (0.1 or 0.3 �g) of TRIM5�Rh were trans-

ected (Fig. 4A). The reduced luciferase transactivation observedt the higher concentrations of K10R pMIP-TRIM5�Rh was not dueo reduced protein expression, as evidenced by a western blottingnalysis of the protein lysates (Fig. 4B). We repeated this experi-ent (Fig. 4C and D) and obtain very similar results. Specifically, WT

RIM5�Rh activated NF-�B more efficiently than K10R TRIM5�Rhhen relatively high amounts of the TRIM5-expressing plasmidere used (in this case, 0.3 �g or above) but not when a smaller

mount (0.1 �g) was transfected. In conclusion, K10R causes a mod-st but significant decrease in the capacity of TRIM5�Rh to activateF-�B in certain experimental conditions.

.5. The K10R mutation decreases the ability of TRIM5˛Rh toctivate AP-1-dependent transcription

AP-1 is a transcription factor composed of heterodimers of sev-ral possible proteins, the most well-known of them being c-Fosnd c-Jun (Hess et al., 2004). AP-1 is activated in response to virusetection by toll-like receptors (Zhong et al., 2006), and also fol-

owing TRIM5� over-expression (Pertel et al., 2011; Uchil et al.,013). Thus, we examined the impact of the K10R mutation on AP-

activation by TRIM5� over-expression, using an AP-1-dependent

uciferase expression system (Pertel et al., 2011). Co-transfectionf 3 �g of WT TRIM5�Rh and 0.6 �g of the AP-1 reporter con-truct led to luciferase levels that were significantly higher thanhen the K10R mutant was used (Fig. 5A). Indeed, the capacity of

lular lysates prepared to quantify luciferase activity in (A) were also used to analyzeTRIM5� expression. For this, we pooled together the 3 lysates from each triplicatetransfection and analyzed them by western blotting using an anti-FLAG antibody.

K10R TRIM5�Rh to activate AP-1 was intermediate between thatof its WT counterpart and that of the RING domain mutant C35A(Fig. 5A). When lower amounts of the TRIM5�Rh expressing plasmidwere used (1 �g instead of 3 �g), we saw no significant differencebetween K10R mutant and the WT control, mirroring our NF-�Bactivation data (Fig. 4). K10R TRIM5�Rh was expressed at levelsslightly higher than its WT counterpart in this experiment (Fig. 5B),ruling out an expression defect as the explanation for the decreasedAP-1 activation phenotype.

3.6. K10R TRIM5˛Rh shows altered nuclear/cytoplasmicdistribution and reduced co-localization with SUMO-1 in thenucleus

It was recently shown that TRIM5� is able to shuttle betweencytoplasm and nucleus and that TRIM5� nuclear bodies (NBs) canbe observed upon treatment with leptomycin B (LMB), an inhibitorof CRM-1 dependent nuclear export (Diaz-Griffero et al., 2011). IFmicroscopy was performed in CRFK cells stably expressing FLAG-tagged TRIM5�Rh WT and K10R, in the presence or absence of LMB.As expected, LMB treatment resulted in a fraction of WT and K10RTRIM5�Rh being present in the nucleus while no nuclear TRIM5�staining was observed in the absence of the drug (Fig. 5A). NuclearTRIM5� bodies were recently shown to co-localize with PML bod-ies (Diaz-Griffero et al., 2011). PML, also called TRIM19, is the main

determinant for the formation of PML bodies, nuclear structuresthat can harbor multiple different proteins and have known rolesin antiviral defense (Geoffroy and Chelbi-Alix, 2011). PML is heav-ily sumoylated, as are many proteins localizing at PML bodies, and

36 M.-É. Nepveu-Traversy, L. Berthoux / Virus Research 184 (2014) 30–38

Fig. 6. Decreased nuclear co-localization of K10R TRIM5�Rh and SUMO-1. (A) CRFK cells stably expressing either WT or K10R FLAG-tagged TRIM5�Rh and treated or not withLMB were processed for IF microscopy. A rabbit polyclonal anti-FLAG antibody was used for the detection of TRIM5�Rh (red) and a mouse monoclonal antibody was used forthe detection of endogenous SUMO-1 (green). White arrows point to examples of co-localizations of nuclear TRIM5�Rh and SUMO-1. (B) Average number of TRIM5�Rh NBsper cell with standard deviation. The total number of nuclei analyzed is shown on top of each bar, and statistical analysis was done using a Student’s t-test. (C) Quantitativeanalysis of the frequency of TRIM5�Rh-SUMO-1 co-localization in nuclei. The percentage of WT and K10R TRIM5�Rh NBs co-localizing with SUMO-1 is shown. Each barr g nucu r. (Fot

S(sSfctttRTotb1dp

4

ore

epresents the % of co-localization observed in 25 randomly chosen nuclei (excludinsing a Student’s t-test. The total number of NBs analyzed is shown on top of each bao the web version of the article.)

UMO is a major determinant for the formation of PML bodiesMuller et al., 1998). Thus, we postulated that the TRIM5� con-ensus sumoylation site was important for its co-localization withUMO when present in the nucleus. As shown Fig. 6A, a significantraction of TRIM5�Rh NBs co-localized with SUMO-1 in LMB-treatedells, and this was true for both the WT and K10R variants of the pro-ein. We then analyzed multiple randomly selected cells to quantifyhe amounts of TRIM5�Rh NBs found in nuclei (Fig. 6B), as well ashe percentage of these NBs co-localizing with SUMO-1 (Fig. 6C).esults show that in the presence of LMB, ∼5 WT and ∼11 K10RRIM5�Rh NBs were observed per cell (Fig. 6B), on average. Webserved no obvious changes in the apparent number or aspect ofhe cytoplasmic bodies. In addition, 38.6% of WT TRIM5�Rh NBsut only 18.3% of K10R TRIM5�Rh NBs co-localized with SUMO-

(Fig. 6C). Therefore, the K10R mutation affects the subcellularynamics of TRIM5�Rh, specifically its shuttling between cyto-lasm and nucleus and its association with SUMO-1 NBs.

. Discussion

Previous studies focusing on the consensus sumoylation sitef TRIM5� concluded that lysine 10 had little or no influence onetroviral restriction, depending on the cellular context (Arriagadat al., 2011; Brandariz-Nunez et al., 2013; Lukic et al., 2013). Our

lei that had no TRIM5� NBs) with standard deviation. Statistical analysis was doner interpretation of the references to color in this figure legend, the reader is referred

own data are consistent with these conclusions (Fig. 1). The slightdecrease in restriction activity for the K10R mutant could be relatedto its increased poly-ubiquitylation levels (Fig. 2), which may resultin premature TRIM5� ubiquitylation instead of ubiquitylation fol-lowing the interception of a restriction-sensitive capsid. Despite ithaving little involvement in the direct inhibition of incoming retro-viruses, the sumoylation site is conserved in all known primateorthologs of TRIM5� (not shown). Thus, we hypothesized that itwas important for a different function of TRIM5�, the activationof innate immune responses (Tareen and Emerman, 2011b). Thisfunction of TRIM5� is largely dependent upon the ubiquitin lig-ase activity associated with the RING domain (Pertel et al., 2011;Tareen and Emerman, 2011a; Uchil et al., 2013), which is adja-cent to the consensus sumoylation site. Therefore, we analyzed theeffects of mutating lysine 10 on the E3 ubiquitin ligase activity ofthe TRIM5�Rh RING domain.

It is established that a single RING domain can recruit differ-ent E2s, resulting in different ubiquitylation patterns (Christensenet al., 2007; Kim et al., 2007). The RING domain of TRIM5� functionsas a ubiquitin ligase in two different types of ubiquitylation: auto-

ubiquitylation that is presumed to be K48-linked (though this hasnot been demonstrated so far) (Diaz-Griffero et al., 2006; Li et al.,2013; Lienlaf et al., 2011; Maegawa et al., 2010) and the genera-tion of K63-linked ubiquitin chains which may be unconjugated

ux / V

(tlamtoiae(�UirEYcesdK

usctw(Tto

imttnptuosRi

obvtiTPliSCllhKiSPtt

M.-É. Nepveu-Traversy, L. Bertho

Pertel et al., 2011). Our results strongly suggest that the puta-ive sumoylation site in TRIM5� modulates these two ubiquitinigase activities. Indeed, the K10R mutation leads to an increase inuto-ubiquitylation (Fig. 2) accompanied by a decrease in the for-ation of K63-linked chains (Fig. 3). The observed phenotype of

he TRIM5� K10R mutant is markedly different from the behaviorf mutants affecting the RING finger zinc binding motif, includ-ng C15A, H32A and C35A. Indeed, such RING domain mutationsbrogate both auto-ubiquitylation (Lienlaf et al., 2011; Maegawat al., 2010) and the generation of K63-linked ubiquitin chainsPertel et al., 2011), and they also abrogate the activation of NF-B and AP-1 (Pertel et al., 2011; Tareen and Emerman, 2011a;chil et al., 2013). In contrast, the consensus sumoylation site

s not required for RING-mediated ubiquitylation but it acts as aegulator. Auto-ubiquitylation of TRIM5� is accomplished by the2 ubiquitin-conjugating enzyme UbcH5b (Langelier et al., 2008;amauchi et al., 2008) while the formation of free K63-conjugatedhains is dependent on the E2 heterodimer Ubc13/Uev1a (Pertelt al., 2011). Our data open the possibility that the consen-us sumoylation site promotes interactions of the TRIM5� RINGomain with Ubc13/Uev1a rather than UbcH5b, thereby inducing63-linked ubiquitylation.

The K10R mutation had a more significant effect on K63-linkedbiquitylation than on NF-�B or AP-1 activation. However, a recenttudy in which endogenous ubiquitin was depleted and cells wereoncomitantly made to express K63-mutated ubiquitin revealedhat NF-�B could be activated by two distinct pathways, one ofhich is completely independent of K63-linked ubiquitin chains

Xu et al., 2009). Hence, it is possible that over-expression ofRIM5� triggers both pathways and that mutating lysine 10 affectshe K63-linked ubiquitylation-dependent pathway but not the sec-nd one.

Because K10R decreases both the direct restriction of incom-ng viruses and the activation of innate immunity pathways, one

ight suspect that this mutation simply affects the overall struc-ure of TRIM5�, hence affecting its functions as well. However,wo observations point against this conclusion. First, K10R doesot affect TRIM5� expression levels nor its capacity to form cyto-lasmic or nuclear bodies (Figs. 1 and 6). Second, K10R increaseshe RING-associated E3 activity that leads to self-ubiquitylation, annlikely observation if K10R was affecting the general functionsf the RING domain. In that regard, it is worth noting again thepecificity of the K10R phenotype, compared to mutations of theING zinc finger that decrease all RING-mediated activities includ-

ng self-ubiquitylation.What is the role, if any, of the sumoylation pathway in the

bserved K10R mutation phenotype? Although TRIM5� has nevereen found to be sumoylated, it is possible that sumoylation occursery transiently. In that regard, it is interesting that TRIM5� is ableo shuttle in and out of the nucleus, although its steady-state local-zation is largely cytoplasmic (Diaz-Griffero et al., 2011). NuclearRIM5� localizes partly at PML NBs (Diaz-Griffero et al., 2011) andML itself, as well as many proteins associating with it, is sumoy-ated (Muller et al., 1998). This raises the possibility that TRIM5�s sumoylated only while transiently present in the nucleus or inUMO/PML NBs, which would be undetectable by classical analyses.onsistent with a link between SUMO and the nuclear shuttling and

ocalization of TRIM5�, we find that the K10R mutation doubles theevel of nuclear localization in LMB-treated cells, but decreases byalf its association with SUMO-1 (Fig. 6). Ubc13, the E2 enzyme for63-linked ubiquitylation, is present in the nucleus and its activ-

ty is sometimes dependent on the modification of its target by

UMO-2/-3 (Poulsen et al., 2013). For instance, sumoylation of theML NB-associated HTLV-1 protein Tax regulates its ubiquityla-ion, thereby modulating its capacity to activate NF-�B throughhe involvement of K63-linked ubiquitin chains (Fryrear et al.,

irus Research 184 (2014) 30–38 37

2012; Journo et al., 2013; Zane and Jeang, 2012). We speculate thatTRIM5� transiently shuttles to the nucleus where it is sumoylated,localizes at PML/SUMO bodies and recruits Ubc13 to stimulate K63-linked ubiquitylation.

5. Conclusions

Mutating the putatively sumoylated lysine 10 in TRIM5� affectsthe balance between self-ubiquitylation and K63-linked ubiquity-lation. It also decreases the capacity of TRIM5� to activate NF-�Band AP-1. In addition, the consensus sumoylation site seems tobe a determinant for the transient co-localization with SUMO inthe nucleus, which may be linked to the observed modulation ofubiquitylation.

Acknowledgements

We are grateful to Natacha Mérindol for a critical reading of thisarticle. We thank Michael Emerman, Jeremy Luban and Ted Dawsonfor sharing reagents. This work was funded by Canadian Institutesfor Health Research operating grant MOP-102712.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.virusres.2014.02.013.

References

Aiken, C., 1997. Pseudotyping human immunodeficiency virus type 1 (HIV-1) by theglycoprotein of vesicular stomatitis virus targets HIV-1 entry to an endocyticpathway and suppresses both the requirement for Nef and the sensitivity tocyclosporin A. J. Virol. 71 (8), 5871–5877.

Arriagada, G., Muntean, L.N., Goff, S.P., 2011. SUMO-interacting motifs of humanTRIM5alpha are important for antiviral activity. PLoS Pathog. 7 (4), e1002019.

Berthoux, L., Sebastian, S., Sokolskaja, E., Luban, J., 2004. Lv1 inhibition of humanimmunodeficiency virus type 1 is counteracted by factors that stimulate syn-thesis or nuclear translocation of viral cDNA. J. Virol. 78 (21), 11739–11750.

Berthoux, L., Towers, G.J., Gurer, C., Salomoni, P., Pandolfi, P.P., Luban, J., 2003. As2O3

enhances retroviral reverse transcription and counteracts Ref1 antiviral activity.J. Virol. 77 (5), 3167–3180.

Bérubé, J., Bouchard, A., Berthoux, L., 2007. Both TRIM5alpha and TRIMCyp have onlyweak antiviral activity in canine D17 cells. Retrovirology 4, 68.

Black, L.R., Aiken, C., 2010. TRIM5alpha disrupts the structure of assembled HIV-1capsid complexes in vitro. J. Virol. 84 (13), 6564–6569.

Brandariz-Nunez, A., Roa, A., Valle-Casuso, J.C., Biris, N., Ivanov, D., Diaz-Griffero,F., 2013. Contribution of SUMO-interacting motifs and SUMOylation to theantiretroviral properties of TRIM5alpha. Virology 435 (2), 463–471.

Campbell, E.M., Dodding, M.P., Yap, M.W., Wu, X., Gallois-Montbrun, S., Malim, M.H.,Stoye, J.P., Hope, T.J., 2007. TRIM5 alpha cytoplasmic bodies are highly dynamicstructures. Mol. Biol. Cell 18 (6), 2102–2111.

Christensen, D.E., Brzovic, P.S., Klevit, R.E., 2007. E2-BRCA1 RING interactions dictatesynthesis of mono- or specific polyubiquitin chain linkages. Nat. Struct. Mol. Biol.14 (10), 941–948.

Diaz-Griffero, F., Gallo, D.E., Hope, T.J., Sodroski, J., 2011. Trafficking of some oldworld primate TRIM5alpha proteins through the nucleus. Retrovirology 8, 38.

Diaz-Griffero, F., Li, X., Javanbakht, H., Song, B., Welikala, S., Stremlau, M., Sodroski, J.,2006. Rapid turnover and polyubiquitylation of the retroviral restriction factorTRIM5. Virology 349 (2), 300–315.

Fryrear, K.A., Guo, X., Kerscher, O., Semmes, O.J., 2012. The Sumo-targeted ubiquitinligase RNF4 regulates the localization and function of the HTLV-1 oncoproteinTax. Blood 119 (5), 1173–1181.

Geoffroy, M.C., Chelbi-Alix, M.K., 2011. Role of promyelocytic leukemia protein inhost antiviral defense. J. Interferon Cytokine Res. 31 (1), 145–158.

Gocke, C.B., Yu, H., Kang, J., 2005. Systematic identification and analysis ofmammalian small ubiquitin-like modifier substrates. J. Biol. Chem. 280 (6),5004–5012.

Hess, J., Angel, P., Schorpp-Kistner, M., 2004. AP-1 subunits: quarrel and harmonyamong siblings. J. Cell Sci. 117 (Pt 25), 73–5965.

Journo, C., Bonnet, A., Favre-Bonvin, A., Turpin, J., Vinera, J., Cote, E., Chevalier, S.A.,Kfoury, Y., Bazarbachi, A., Pique, C., Mahieux, R., 2013. Human T cell leukemia

virus type 2 tax-mediated NF-kappaB activation involves a mechanism inde-pendent of Tax conjugation to ubiquitin and SUMO. J. Virol. 87 (2), 1123–1136.

Keckesova, Z., Ylinen, L.M., Towers, G.J., 2004. The human and African green monkeyTRIM5alpha genes encode Ref1 and Lv1 retroviral restriction factor activities.Proc. Natl. Acad. Sci. U.S.A. 101 (29), 10780–10785.

3 ux / V

K

K

L

L

L

L

L

L

M

M

N

P

P

P

P

P

8 M.-É. Nepveu-Traversy, L. Bertho

im, H.T., Kim, K.P., Lledias, F., Kisselev, A.F., Scaglione, K.M., Skowyra, D., Gygi,S.P., Goldberg, A.L., 2007. Certain pairs of ubiquitin-conjugating enzymes (E2s)and ubiquitin–protein ligases (E3s) synthesize nondegradable forked ubiqui-tin chains containing all possible isopeptide linkages. J. Biol. Chem. 282 (24),17375–17386.

utluay, S.B., Perez-Caballero, D., Bieniasz, P.D., 2013. Fates of retroviral core com-ponents during unrestricted and TRIM5-restricted infection. PLoS Pathog. 9 (3),e1003214.

angelier, C.R., Sandrin, V., Eckert, D.M., Christensen, D.E., Chandrasekaran, V.,Alam, S.L., Aiken, C., Olsen, J.C., Kar, A.K., Sodroski, J.G., Sundquist, W.I.,2008. Biochemical characterization of a recombinant TRIM5alpha protein thatrestricts human immunodeficiency virus type 1 replication. J. Virol. 82 (23),11682–11694.

i, X., Kim, J., Song, B., Finzi, A., Pacheco, B., Sodroski, J., 2013. Virus-specific effectsof TRIM5alpha(rh) RING domain functions on restriction of retroviruses. J. Virol.87 (13), 7234–7245.

ienlaf, M., Hayashi, F., Di Nunzio, F., Tochio, N., Kigawa, T., Yokoyama, S., Diaz-Griffero, F., 2011. Contribution of E3-ubiquitin ligase activity to HIV-1 restrictionby TRIM5alpha(rh): structure of the RING domain of TRIM5alpha. J. Virol. 85 (17),8725–8737.

im, K.L., Chew, K.C., Tan, J.M., Wang, C., Chung, K.K., Zhang, Y., Tanaka, Y., Smith,W., Engelender, S., Ross, C.A., Dawson, V.L., Dawson, T.M., 2005. Parkin mediatesnonclassical, proteasomal-independent ubiquitination of synphilin-1: implica-tions for Lewy body formation. J. Neurosci. 25 (8), 9–2002.

ukic, Z., Goff, S.P., Campbell, E.M., Arriagada, G., 2013. Role of SUMO-1 and SUMOinteracting motifs in rhesus TRIM5alpha-mediated restriction. Retrovirology 10,10.

ukic, Z., Hausmann, S., Sebastian, S., Rucci, J., Sastri, J., Robia, S.L., Luban, J., Campbell,E.M., 2011. TRIM5alpha associates with proteasomal subunits in cells while incomplex with HIV-1 virions. Retrovirology 8, 93.

aegawa, H., Miyamoto, T., Sakuragi, J., Shioda, T., Nakayama, E.E., 2010. Contri-bution of RING domain to retrovirus restriction by TRIM5alpha depends oncombination of host and virus. Virology 399 (2), 212–220.

uller, S., Matunis, M.J., Dejean, A., 1998. Conjugation with the ubiquitin-relatedmodifier SUMO-1 regulates the partitioning of PML within the nucleus. EMBO J.17 (1), 61–70.

aviaux, R.K., Costanzi, E., Haas, M., Verma, I.M., 1996. The pCL vector system: rapidproduction of helper-free, high-titer, recombinant retroviruses. J. Virol. 70 (8),5701–5705.

erron, M.J., Stremlau, M., Lee, M., Javanbakht, H., Song, B., Sodroski, J., 2007. Thehuman TRIM5alpha restriction factor mediates accelerated uncoating of the N-tropic murine leukemia virus capsid. J. Virol. 81 (5), 2138–2148.

erron, M.J., Stremlau, M., Song, B., Ulm, W., Mulligan, R.C., Sodroski, J., 2004.TRIM5alpha mediates the postentry block to N-tropic murine leukemia virusesin human cells. Proc. Natl. Acad. Sci. U.S.A. 101 (32), 11827–11832.

ertel, T., Hausmann, S., Morger, D., Zuger, S., Guerra, J., Lascano, J., Reinhard,C., Santoni, F.A., Uchil, P.D., Chatel, L., Bisiaux, A., Albert, M.L., Strambio-De-Castillia, C., Mothes, W., Pizzato, M., Grutter, M.G., Luban, J., 2011. TRIM5 isan innate immune sensor for the retrovirus capsid lattice. Nature 472 (7343),361–365.

ham, Q.T., Bouchard, A., Grutter, M.G., Berthoux, L., 2010. Generation of humanTRIM5alpha mutants with high HIV-1 restriction activity. Gene Ther. 17 (7),

859–871.

oulsen, S.L., Hansen, R.K., Wagner, S.A., van Cuijk, L., van Belle, G.J., Streicher, W.,Wikstrom, M., Choudhary, C., Houtsmuller, A.B., Marteijn, J.A., Bekker-Jensen,S., Mailand, N., 2013. RNF111/Arkadia is a SUMO-targeted ubiquitin ligase thatfacilitates the DNA damage response. J. Cell Biol. 201 (6), 797–807.

irus Research 184 (2014) 30–38

Reymond, A., Meroni, G., Fantozzi, A., Merla, G., Cairo, S., Luzi, L., Riganelli, D., Zanaria,E., Messali, S., Cainarca, S., Guffanti, A., Minucci, S., Pelicci, P.G., Ballabio, A.,2001. The tripartite motif family identifies cell compartments. EMBO J. 20 (9),2140–2151.

Rold, C.J., Aiken, C., 2008. Proteasomal degradation of TRIM5alpha during retrovirusrestriction. PLoS Pathog 4 (5), e1000074.

Sampson, D.A., Wang, M., Matunis, M.J., 2001. The small ubiquitin-like modifier-1 (SUMO-1) consensus sequence mediates Ubc9 binding and is essential forSUMO-1 modification. J. Biol. Chem. 276 (24), 21664–21669.

Sayah, D.M., Sokolskaja, E., Berthoux, L., Luban, J., 2004. Cyclophilin A retrotranspo-sition into TRIM5 explains owl monkey resistance to HIV-1. Nature 430 (6999),569–573.

Sebastian, S., Luban, J., 2005. TRIM5alpha selectively binds a restriction-sensitiveretroviral capsid. Retrovirology 2, 40.

Sebastian, S., Sokolskaja, E., Luban, J., 2006. Arsenic counteracts human immun-odeficiency virus type 1 restriction by various TRIM5 orthologues in a celltype-dependent manner. J. Virol. 80 (4), 2051–2054.

Song, B., Gold, B., O‘Huigin, C., Javanbakht, H., Li, X., Stremlau, M., Winkler, C., Dean,M., Sodroski, J., 2005. The B30.2(SPRY) domain of the retroviral restriction factorTRIM5alpha exhibits lineage-specific length and sequence variation in primates.J. Virol. 79 (10), 6111–6121.

Stremlau, M., Owens, C.M., Perron, M.J., Kiessling, M., Autissier, P., Sodroski, J., 2004.The cytoplasmic body component TRIM5alpha restricts HIV-1 infection in OldWorld monkeys. Nature 427 (6977), 848–853.

Stremlau, M., Perron, M., Lee, M., Li, Y., Song, B., Javanbakht, H., Diaz-Griffero, F.,Anderson, D.J., Sundquist, W.I., Sodroski, J., 2006. Specific recognition and accel-erated uncoating of retroviral capsids by the TRIM5alpha restriction factor. Proc.Natl. Acad. Sci. U.S.A. 103 (14), 5514–5519.

Tareen, S.U., Emerman, M., 2011a. Human Trim5alpha has additional activi-ties that are uncoupled from retroviral capsid recognition. Virology 409 (1),113–120.

Tareen, S.U., Emerman, M., 2011b. Trim5 TAKes on pattern recognition. Cell HostMicrobe 9 (5), 349–350.

Towers, G.J., 2007. The control of viral infection by tripartite motif proteins andcyclophilin A. Retrovirology 4, 40.

Uchil, P.D., Hinz, A., Siegel, S., Coenen-Stass, A., Pertel, T., Luban, J., Mothes, W.,2013. TRIM protein-mediated regulation of inflammatory and innate immunesignaling and its association with antiretroviral activity. J. Virol. 87 (1), 257–272.

Vertegaal, A.C., 2007. Small ubiquitin-related modifiers in chains. Biochem. Soc.Trans. 35 (Pt 6), 1422–1423.

Xu, M., Skaug, B., Zeng, W., Chen, Z.J., 2009. A ubiquitin replacement strategy inhuman cells reveals distinct mechanisms of IKK activation by TNFalpha andIL-1beta. Mol. Cell 36 (2), 302–314.

Yamauchi, K., Wada, K., Tanji, K., Tanaka, M., Kamitani, T., 2008. Ubiquitination of E3ubiquitin ligase TRIM5 alpha and its potential role. FEBS J. 275 (7), 1540–1555.

Yap, M.W., Nisole, S., Lynch, C., Stoye, J.P., 2004. Trim5alpha protein restrictsboth HIV-1 and murine leukemia virus. Proc. Natl. Acad. Sci. U.S.A. 101 (29),10786–10791.

Zane, L., Jeang, K.T., 2012. The importance of ubiquitination and sumoylation on thetransforming activity of HTLV Tax-1 and Tax-2. Retrovirology 9, 103.

Zhao, J., 2007. Sumoylation regulates diverse biological processes. Cell. Mol. Life Sci.64 (23), 3017–3033.

Zhong, B., Tien, P., Shu, H.B., 2006. Innate immune responses: crosstalk of signalingand regulation of gene transcription. Virology 352 (1), 14–21.

Zufferey, R., Nagy, D., Mandel, R.J., Naldini, L., Trono, D., 1997. Multiply attenuatedlentiviral vector achieves efficient gene delivery in vivo. Nat. Biotechnol. 15 (9),871–875.