The RTP Site Shared by the HIV-1 Tat Protein and the 11 S

The RTP Site Shared by the HIV-1 Tat Protein and the11 S Regulator Subunit a is Crucial for their Effects onProteasome Function Including Antigen Processing

Xiaohua Huang1†, Ulrike Seifert2†, Ulrike Salzmann2, Peter Henklein2

Robert Preissner2, Wolfgang Henke3, Alice J. Sijts2

Peter-Michael Kloetzel2 and Wolfgang Dubiel1*

1Division of Molecular BiologyDepartment of SurgeryMedical Faculty ChariteHumboldt UniversityMonbijoustr. 2A, 10117 BerlinGermany

2Institute of BiochemistryMedical Faculty ChariteHumboldt UniversityMonbijoustr. 2A, 10117 BerlinGermany

3Institute of LaboratoryMedicine andPathobiochemistry, MedicalFaculty Charite, HumboldtUniversity, Monbijoustr. 2A10117 Berlin, Germany

The human immunodeficiency virus-1 Tat protein inhibits the peptidaseactivity of the 20 S proteasome and competes with the 11 S regulator/PA28 for binding to the 20 S proteasome. Structural comparison revealeda common site in the Tat protein and the 11 S regulator a-subunit (REGa)called the REG/Tat-proteasome-binding (RTP) site. Kinetic assays foundamino acid residues Lys51, Arg52 and Asp67 forming the RTP site of Tatto be responsible for the effects on proteasomes in vitro. The RTP siteidentified in REGa consists of the residues Glu235, Lys236 and Lys239.Mutation of the REGa amino acid residues Glu235 and Lys236 to Alaresulted in an REGa mutant that lost the ability to activate the 20 S protea-some even though it still forms complexes with REGb and binds to the20 S proteasome. The REGa RTP site is needed to enhance the presen-tation of a cytomegalovirus pp89 protein-derived epitope by MHC class Imolecules in mouse fibroblasts. Cell experiments demonstrate that theTat amino acid residues 37–72 are necessary for the interaction of theviral protein with proteasomes in vivo. Full-length Tat and the Tat peptide37–72 suppressed 11 S regulator-mediated presentation of the pp89epitope. In contrast, the Tat peptide 37–72 with mutations of amino acidresidues Lys51, Arg52 and Asp67 to Ala was not able to reduce antigenpresentation.

q 2002 Elsevier Science Ltd. All rights reserved

Keywords: Tat; HIV-1; 20 S proteasome; 11 S regulator; antigen processing*Corresponding author

Introduction

The human immunodeficiency virus-1 (HIV-1)Tat protein is a trans-activator in viral replication.1

In addition, many immunosuppressive functionshave been attributed to Tat.2 The viral proteininduces apoptosis in T-cells.3 It inhibits the phago-cytosis of apoptotic tumor cells by accessory cells4

and possibly prevents proper processing andmajor histocompatibility (MHC) presentation oftumor-associated antigens.5 Moreover, Tat sup-

presses antigen-driven T-cell proliferation.6 Certainmechanisms of immunosuppression may involveinhibitory effects of Tat on the proteasome system.

The 20 S proteasome is the core particle (CP) ofthe 26 S proteasome involved in the production ofantigenic peptides presented by MHC class Imolecules.7 It is a target for a number of viralproteins.8 The CP consists of two outer rings com-posed of seven different a-subunits and two innerrings with seven different b-subunits, three ofwhich harbor the active sites. Isolated CP fromeukaryotic cells is in an inactive (latent) state dueto the closed central channel. The N termini of theCP a-subunits close the gate into the CP.9 Deletionof nine amino acid residues from the N terminusof the a3-subunit yields an activated CP complex.The unique role of the a3-tail can be explainedby its contact with all other subunits of thea-ring.10 The transformation into an active stateoccurs when the latent CP assembles with the 19 S

0022-2836/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved

† These authors contributed equally to this work.

E-mail address of the corresponding author:[email protected]

Abbreviations used: HIV-1, human immunodeficiencyvirus-1; CP, core particle; 11 S REG, 11 S regulator/PA28;RTP site, REG/Tat-proteasome-binding site; CTL,cytotoxic T-cells; MHC, major histocompatibility;MCMV, murine cytomegalovirus.

doi:10.1016/S0022-2836(02)00998-1 available online at http://www.idealibrary.com onBw

J. Mol. Biol. (2002) 323, 771–782

Figure 1. The basic domain of Tat is necessary for the inhibition of the CP peptidase activity and for competitionwith the 11 S REG. (a) Sequences of Tat peptides 1 to 4 (Tatpep1 to Tatpep4) used in these studies. (b) Increasingamounts of full-length Tat or of Tatpep1 to Tatpep4 were added to constant amounts of the CP. Peptidase activitywas measured as DF (F: fluorescence) per minute and is expressed as relative peptidase activity (1.0 ¼ 20DF/minuteper mg of CP). Curves were calculated using the model described in Materials and Methods. Calculated kinetic con-stants Vmax, K1 and K2 are summarized in the Table. K1 and K2 are expressed as mg/ml. The data are representative offour independent experiments. (c) The basic domain of Tat is necessary for effective competition with the 11 S REG.Increasing amounts of Tatpep1 to Tatpep4 were added to constant amounts of the CP and the isolated 11 S REG.Peptidase activity was estimated as in (b) and is expressed as relative activity (1 ¼ 20 times baseline activity with CPalone). Calculations were performed as in (b). The data are representative of four independent experiments.

772 HIV-1 Tat Protein Suppresses Antigen Presentation

Figure 2 (legend opposite)

HIV-1 Tat Protein Suppresses Antigen Presentation 773

regulator (PA700) to form the 26 S proteasome8 orwith the 11 S regulator (PA28)11 or when it is acti-vated by chemical treatment.12 Recently, it hasbeen shown that the mechanism by which the“gate” is opened by the 19 S regulator requires the19 S complex Rpt2/S4 ATPase.13

Crystallographic studies revealed the mechan-ism by which the 11 S regulator (11 S REG) opensthe gate. The 11 S REG originally identified as anactivator of the CP14,15 consists of two subunits,REGa and REGb.16 The REGa forms a heptamericring in which the C termini protrude from thebody of the ring structure.17 The C termini ofthe REGa are important for the interaction withthe CP.18,19 Crystallographic studies on a complexcomposed of 11 S REG from Trypanosoma brucei(PA26) and CP from Saccharomyces cerevisiae visual-ized the gating of the CP.11 C-terminal amino acidresidues of PA26 insert into pockets locatedbetween proteasome a-subunits. The activationloop of PA26 (amino acid residues 98–106, corre-sponding to amino acid residues 141–149 ofhuman REGa20) enforces the transition to the opengate conformation by interacting with the Ntermini of the CP a-subunits.

Interferon-g treatment and REGa overexpressionenhance antigen presentation of a murine cytome-galovirus (MCMV) pp89 protein-derived epitope,indicating a role of the 11 S REG in antigenprocessing.21 However, recent data show that thisis epitope-specific. While the production andpresentation of selective epitopes is increased bythe 11 S REG, the generation of other epitopes isnot affected by the 11 S REG.7

Previously, we have shown that the Tat proteininhibits the CP. It competes with the 11 S REG forbinding to the CP, thereby blocking the 11 S REG-mediated activation of the CP. We have also foundthat the Tat protein binds to the 19 S regulatorycomplex of the 26 S proteasome stimulating thedegradation of ubiquitin conjugates.22

Here, we show that the HIV-1 Tat REG/Tat-proteasome-binding (RTP) site composed of Lys51,Arg52 and Asp67 is essential for inhibition of theCP and for competition with the 11 S REG. TheRTP site is also found in REGa and is involved inthe activation of the CP by REGa. The RTP site ofthe Tat protein is needed to suppress the 11 S

REG-mediated presentation of an MCMV pp89protein-derived epitope.

Results

The basic region of Tat is important for theeffects of the viral protein on the CP

In order to identify the Tat region that exertsinhibition of CP and competition with the 11 SREG, the effect of synthesized Tat peptide 37–72(Tatpep1) was compared to that of full-length Tatwith isolated human CP. Tatpep1 has been chosen,because it freely penetrates into cells23 and containsthe basic region responsible for many Tatactivities.2 To narrow down the “active” site inTat, truncations of Tatpep1 were synthesized (Tat-pep2 to Tatpep4) and tested. To estimate kineticparameters, the data were fit with a model outlinedin Materials and Methods. Figure 1(a) shows thesequences of Tatpep1 to Tatpep4 used in theseexperiments. Tatpep2 is missing the basic domain(amino acid residues 49–57); Tatpep3 is withoutthe C-terminal and Tatpep4 without the coreregion. The baseline CP activity in all experimentswas determined to be approximately 20 DF/minute per mg. The inhibition constants expressedin mg/ml obtained with Tatpep1 and full-lengthTat were identical (Figure 1(b)). Compared to full-length Tat, approximately two times higher molarconcentration of Tatpep1 is required to exert thesame effect on the CP. Tatpep3 or Tatpep4 had littleimpact on the inhibition of CP activity. In contrast,deletion of the basic region in Tatpep2 led to acomplete loss of CP inhibition (Figure 1(b)).

The same Tat peptides were tested for theirability to compete with the 11 S REG. Inhibition of11 S REG-activated CP by Tat peptides wasmeasured at a molar ratio of CP:11 S REG ofapproximately 1:6 (20 times stimulation of CPbasic activity). Tatpep1 caused inhibition of theCP–11 S REG complex activity at higher concen-trations. Tatpep2, Tatpep3 and Tatpep4 were notable to compete efficiently with the 11 S REG. Thisis reflected by high K1 and K2 constants and byVmax2 values for inhibition approximating zero(Figure 1(c)). In all cases Tat peptides exerted

Figure 2. Identification of a common site in Tat and REGa. (a) A solid surface representation of a portion of humanREGa (PDB code: AVO) and Tat (PDB code: TBC) is shown demonstrating the identified structure on the surface ofthe two proteins. The solvent-accessible surfaces of the residues Lys236 and Lys239 of REGa and Lys51 and Arg52 ofTat are colored in light blue and acidic residues Glu235 (REGa) and Asp67 (Tat) are yellow. (b) The REGa heptamer(red cylinders, a-helices; green ribbons, loop regions) and the Tat structures. The N and C termini and the activationloop for one of the seven subunits are designated. The three residues (Glu235, Lys236 and Lys239) given in stickrepresentation form the RTP site. For Tat, the backbone is illustrated as red ribbon, which is white in the region fromamino acid residues 37–72. The stick representation of residues Lys51, Arg52 and Asp67 visualizes the formation ofthe RTP site of Tat. (c) The homology modeled human CP structure is shown as secondary structure sketch (redcylinders, a-helices; yellow arrows, b-sheets; green ribbons, loops). The main chain of the REGa heptamer is tracedby violet ribbon. The atoms of residues 235, 236 and 239 are indicated. A heptameric Tat ring can be constructed thatfits to the CP a-ring by orienting residues 51, 52 (light blue) and 67 (yellow) of seven Tat molecules (red ribbon)towards the proteasomal a-ring the way the corresponding residues of REGa are positioned.


stimulation of proteasome peptidase activity at lowconcentrations.

Identification of a common site in Tat andin REGa

The available structural data of Tat24 and ofREGa17 were compared (see Materials andMethods). The data revealed a similar site com-posed of three charged residues in both the pro-teins. This site is formed in the Tat protein byamino acid residues Lys51, Arg52 and Asp67 andin REGa by amino acid residues Glu235, Lys236and Lys239. The site is localized on the surface ofeach protein (Figure 2(a)). It is part of a chargeda-helix in REGa (Figure 2(b)). Figure 2(c) showsthe human REGa structure17 and a putativehuman CP, which has been constructed on thebasis of the yeast CP structure data.25 The bindingbetween heptameric REGa and CP was simulatedaccording to the structure of the PA26–CPcomplex.11 The model clearly reveals that theidentified site in the REGa contacts the CP a-ring.The REGa amino acid residues Glu235, Lys236and Lys239 are in close proximity to the REGabinding (C-terminal amino acid residues) as wellas activation domains (amino acid residues 241–249). A ring composed of seven Tat monomersthat mimics REGa interaction with the CP a-ringwas modeled. In the heptameric Tat model,residues Lys51, Arg52 and Asp67 were positionedtowards the 20 S a-ring as the corresponding

residues Glu235, Lys236 and Lys239 of REGa(Figure 2(c)). It is known that Tat forms homo-oligomers.26 However, at the moment there is nodata on the existence of a homo-heptameric Tatand its binding to CP. Because the common siteidentified in REGa and in Tat is involved in bind-ing to the 20 S proteasome (see below) we call itREG/Tat-20 S proteasome-binding (RTP) site.

Tat amino acid residues Lys51, Arg52 andAsp67 are essential for competition with the11 S REG

To verify our hypothesis that the RTP site of Tatis responsible for the effects on the CP, Tatpep5was synthesized. In Tatpep5 the amino acidresidues Lys51, Arg52 and Asp67 were substitutedwith Ala (Figure 3(a)). In vitro experimentsrevealed that Tatpep5 did not inhibit the CP(Figure 3(b)). Tatpep5 behaved like Tatpep2, whichlacks the complete basic domain. To obtain a betterillustration of 11 S REG/Tat competition, Far-Western blots were conducted. The immobilizedCP was incubated with recombinant REGa in thepresence or in the absence of Tat peptides. REGabinds to the CP in the absence of Tat peptides(Figure 3(c)). A 30-fold molar excess of Tatpep1completely abolished binding of REGa to the CP.A 1:1 molar ratio of Tat and REGa had no impacton REGa-CP binding (data not shown). A 30-foldmolar excess of Tatpep2 was far less effective

Figure 3. Competition betweenREGa and Tat peptides for bindingto the CP. (a) In Tatpep5 aminoacid residues Lys51, Arg52 andAsp67 (RTP site) were substitutedby Ala. (b) Impact of Tatpep1, Tat-pep2 and Tatpep5 on CP peptidaseactivity. The activity was deter-mined as described in the legendto Figure 1 except that measure-ments were carried out at roomtemperature. (c) Visualization ofcompetition between REGa and Tatpeptides by Dot-blots. ImmobilizedCP was incubated with recombi-nant REGa in the absence (control)and in the presence of Tatpep1, Tat-pep5 or Tatpep2. After washing,nitrocellulose was probed withanti-REGa antibody.


compared to Tatpep1. Tatpep5 had no effect onREGa-CP interaction.

Site-directed mutagenesis revealed a functionof REGa amino acid residues Glu235 andLys236 in CP activation

To study the role of the identified RTP site for11 S REG/CP interaction, the REGa amino acidresidues Glu235 and Lys236 were changed to Alausing site-directed mutagenesis. As expected,recombinant REGawt alone activates the CP withlow affinity (18.3 mg/ml) (Figure 4(a)). In contrast,REGam did not activate peptide cleavage by theCP even though it binds with similar affinity tothe CP (Figure 5(c)). In Figure 4(b) stimulation ofthe CP peptidase activity by equimolar amountsof REGawt and REGb after 30 minutes of pre-incubation is shown. An activation constant ofapproximately 7 nM (for REGawt þ REGb) can beestimated. In contrast, the combination of REGam

and REGb did not activate the peptidase activityof the CP. The influence of REGam on CPactivation by the heteromeric 11 S REG complexwas studied. In Figure 4(c) increasing amounts ofREGam were added to constant amounts of CP(1 mg/ml), REGawt (2 mg/ml) and REGb (2 mg/ml). The mixture was pre-incubated for anadditional 30 minutes at 37 8C before adding thefluorigenic substrate and measuring fluorescence.Addition of REGam to the CP-REGawt/REGbcomplex led to an inhibition of peptidase activity.These data suggest that REGam might substitutefor REGawt in the heteromeric 11 S REG complexafter 30 minutes of pre-incubation.

Data shown in Figure 4 are consistent with thefollowing interpretation of functional changes ofREGam. There is an exchange of REGawt withREGam leading to the formation of REGam/REGb complexes, which can bind to, but do notopen the pore of the CP. A prerequisite of thishypothesis is that REGam forms heteromeric

Figure 4. Mutation of REGa residues Glu235 and Lys236 to Ala leads to a REGa mutant that does not activate the CP.(a) Increasing amounts of REGawt or REGam were added to constant amounts of the CP. After 30 minutes pre-incubation the fluorigenic peptide was added and fluorescence was measured in ten-minute intervals at roomtemperature. The curves and the kinetic constants were calculated with the same model as in Figure 1. (b) Increasingequimolar amounts of REGawt and REGb or REGam and REGb were pre-incubated with constant amounts of theCP. Fluorescence was measured and calculations were performed as in (a). (c) Increasing amounts of REGam wereadded to constant amounts of the CP and REGawt and REGb and pre-incubated for an additional 30 minutes at37 8C. Fluorescence was measured as in (a). The experimental data represent three to four independent experiments.


complexes with REGb similar to those formed byREGawt. This was tested by glycerol gradient cen-trifugation (Figure 5(a)). Obviously, REGam isable to form homomeric complexes, which sedi-ment into the same fractions as those composed ofREGawt. Mutations in REGam did not affect itsability to associate with REGb. The sedimentationof REGawt/REGb and REGam/REGb complexeswas very similar (Figure 5(a)). A second prerequi-site of the above hypothesis is a complex formationbetween the REGam/REGb heteromers and theCP. Isolated human CP (18 mg/ml) was incubatedwith REGawt/REGb (12 mg/ml each) or REGam/REGb (12 mg/ml each) and the mixtures wereseparated by glycerol gradient centrifugation. Thedistribution of REGam/REGb in glycerol gradientfractions is very similar to that of REGawt/REGb(Figure 5(b)). To show binding of REGam alone tothe CP, dot blots were performed. REGawt andREGam bind equally to the immobilized CP(Figure 5(c)).

Amino acid residues 37–72 of the Tat proteinare essential for binding to proteasomes invivo

To study the biological relevance of the Tateffects, binding of the viral protein to proteasomeswas investigated in HeLa cells. A Tat cDNA con-struct encoding a N-terminal Flag-tag was trans-fected into HeLa cells. An anti-Rpt2/S4 antibodywas used to precipitate the 19 S regulator and the26 S proteasome and an anti-a6/C2 antibody forprecipitation of the CP and the 26 S proteasome.The expressed Tat protein was detected with theanti-Flag antibody in the anti-Rpt2/S4-immuno-precipitates as well as in the anti-a6/C2-immuno-

precipitates (Figure 6(a)). These data indicatethat wild-type Tat interacts with endogenous20 S-related particles.

From our in vitro experiments we assumed thatTatpep1 is essential for binding to the CP. There-fore a Tat deletion mutant was constructed missingamino acid residues 37–72. The Tat deletionmutant was not able to bind HeLa cell proteasomesas revealed by non-denaturing electrophoresis(Figure 6(b)) or by glycerol gradient centrifugation(Figure 6(c)). Figure 6(c) shows that wild-type Tatbinds to 20 S and 26 S particles. These data are inline with our earlier in vitro findings demonstratingthat wild-type full-length Tat binds to the CP aswell as to the 19 S regulator.22

Tat interferes with antigen presentation of anMCMV pp89 protein-derived epitope

MHC class I antigen presentation was analyzedwith mouse B8 cells expressing the MCMV pp89protein, which is processed to the epitope 168–176in a proteasome-dependent and 11 S REG-enhanced manner.21 Mouse B8 cells transformedwith wild-type Tat were not viable under our con-ditions. Therefore, and because full-length Tat andTat peptides are taken up by cells,23 B8 cells wereincubated with Tatpep1, Tatpep5 and full-lengthTat and then tested for their susceptibility to lysisby pp89168 – 176-epitope-specific cytotoxic T-cells(CTLs). While untreated B8 cells were efficientlylyzed, B8 cells incubated with Tatpep1 (Figure7(a)) or full-length Tat (Figure 7(b)) revealed areduced CTL recognition, indicating a diminishedpresentation of the antigenic peptide. In contrast,incubation of B8 cells with Tatpep5, which lacksthe RTP site, did not impair CTL recognition.

Figure 5. Mutation of REGa resi-dues Glu235 and Lys236 to Aladoes not prevent complex for-mation. (a) Glycerol gradient centri-fugations with recombinantREGawt, REGam, REGawt andREGb or REGam and REGb. Four-teen 0.86 ml fractions were collectedfrom the bottom and proteinsimmunoblotted with anti-REGaand anti-REGb antibodies. (b) Gly-cerol gradient centrifugation in thepresence of purified CP and recom-binant REGawt and REGb orREGam and REGb. Eighteen0.68 ml fractions were collectedfrom the bottom and used forimmunoblots with anti-a6/C2 andanti-REGb antibodies. (c) Bindingof REGawt or REGam to the CP.Purified CP was dot-blotted tonitrocellulose and incubated with24 mg/ml of REGawt or REGam.After washing, nitrocellulose sliceswere probed with an anti-REGaantibody.


These data demonstrate that Tat can down-regulateMHC class I antigen presentation. As inferred bybiochemical data, this is most likely by competingwith 11 S REG and the RTP site is needed for theeffect. In contrast, B8 cells expressing REGam lack-ing the RTP site are less efficiently lyzed by CTLsthan B8 cells expressing REGawt (Figure 7(c)).

Discussion

The Tat amino acid residues Lys51, Arg52 andAsp67 are essential for inhibition of the CPand for competition with the 11 S REG

In an earlier study we have shown that full-length Tat inhibits the CP and competes with the11 S REG.22 Here, we narrowed down the Tat siteresponsible for the effects on proteasomes fromamino acid residues 37–72 to the basic domain

(amino acid residues 49–57) and further to aminoacid residues Lys51, Arg52 and Asp67, whichform the RTP site.

Tatpep1 inhibited the CP peptidase activity andcompeted with the 11 S REG in a manner similarto that of full-length Tat. At low Tatpep1 concen-trations a stimulation of CP peptidase activity inthe presence of the 11 S REG complex wasobserved. Similar kinetics have also been seenwith full-length Tat22 as well as with the protea-some inhibitor PI31, which also competes with11 S REG for binding to the CP.27 It is assumedthat under non-saturating conditions of 11 S REG,Tat protein binds to the free CP a-ring prior tocompeting with the 11 S REG. We hypothesizethat a hybrid complex, 11 S-CP-Tat, might form,which is more active than the single occupied CP(11 S-CP). This, however, might not be relevant forcellular conditions, because 11 S-CP complexesrarely occur in cells.28

The kinetics were sigmoid and described using asimple model that assumes two binding sites andthe occupation of both binding sites is necessaryfor the inhibition by Tat (Tat-CP-Tat) and for thestimulation by 11 S REG (11 S-CP-11 S) or by Tat inthe presence of 11 S REG (11 S-CP-Tat). Althoughthe calculated data fit well with the experimentalresults, more complex models assuming morethan two binding sites can be applied to fit sigmoidkinetics. Principally 14 binding sites correspondingto 14 CP a-subunits are possible. In our model onehas to assume that seven binding sites each aresummarized in two classes of binding sites. Inter-estingly, in most cases K1 and K2, the affinities ofour two binding sites, were equal. One mightspeculate, therefore, that binding of Tat to one site(one CP a-ring) did not change the affinity of theother site.

The deletion of the basic domain led to a com-plete loss of the peptide’s ability to interfere withCP peptidase activity. Modification of the RTP siteled to the synthesis of Tatpep5, which has no effecton CP activity in the concentration range testedand does not compete with the 11 S REG. Thesedata demonstrate that Tat amino acid residuesLys51, Arg52 and Asp67 are essential for theobserved effects, perhaps by keeping the viral pro-tein in a conformation that fits into the bindingsite on the CP a-ring. Inhibition of the CP activityby Tat is most likely due to inhibition of the opengate form and interference with substrate access toor product release from the proteolytic chamber ofthe CP. This, however, does not lead to a suppres-sion of protein breakdown in cells, since the 11 SREG is not rate-limiting for proteolysis.21 In con-trast, our finding that Tat binding to the 19 S regu-latory complex accelerates ubiquitin conjugatedegradation22 suggests that overall proteolysismight even be increased. Given the fact that mostof cellular proteasomes occur as 19 S-CP-19 S,19 S-CP-11 S and free CP complexes28 and that Tatdoes not compete with the 19 S complex,22 weassume that Tat would form mostly 19 S-CP-Tat

Figure 6. Amino acid residues 37–72 of Tat are essen-tial for binding to HeLa cell proteasomes. (a) A TatcDNA pcDNA3 construct encoding full-length Tat withan N-terminal Flag-tag was transfected into HeLa cells.The 19 S regulator, the 26 S proteasome and the CP wereimmunoprecipitated by anti-Rpt2/S4 and anti-a6/C2antibodies, respectively. Immunoblots of the precipitateswere tested with an anti-Flag antibody. Control:immunoprecipitation using the anti-a6/C2 antibodyand untreated HeLa cell lysate probed with anti-Flagantibody. (b) After transfections HeLa cell lysate proteinswere separated by non-denaturing gel electrophoresis,blotted to nitrocellulose and probed with an anti-Flagantibody. The position of the 26 S proteasome (26 S), the19 S regulator (19 S) and the CP (20 S) were identifiedby anti-Rpt2/S4 and anti-a6/C2 antibodies. There is nobinding of TatD37–72 to large protein complexes. Underthe conditions used the small TatD37–72 migrated outof the gel. (c) After transfection, aliquots of HeLa celllysates were centrifuged in glycerol gradients. Eighteen0.68 ml fractions were collected from the bottom andused for immunoblots with the anti-Flag antibody. Thepositions of the 26 S proteasome (26 S) and CP (20 S)were determined by the anti-Rpt2/S4 and anti-a6/C2antibodies.


hybrid complexes or Tat-CP-Tat complexes undercellular conditions. Like the 11 S-CP-Tat hybridcomplex, the 19 S-CP-Tat complex might be activetoo.

Interestingly, a modeled heptameric Tat ring inwhich the RTP site is oriented towards the CP asthat of REGa, fits to the heptameric a-ring of theCP. Although homo-heptameric Tat complexeshave not been described so far, it is known thatTat forms oligomers on its own.26

The REGa residues Glu235 and Lys236 areinvolved in CP activation

Comparison of the available structural data ofTat24 and of REGa17 revealed the RTP site in thetwo proteins. Lys51 and Arg52 are conserved inknown HIV-1 Tat variants and are essential forboth inhibition of CP and competition with the11 S REG. Similarly, Glu235 and Lys236 are con-served in all REGs (a, b and g). At the momentwe do not know the function of the conserved Gluand Lys residues in REGb and REGg. REGam orthe combination of REGam and REGb were notable to stimulate the CP, although the amino acidsubstitution in REGam did not change its capa-bility to form the 11 S REG complex or to bind tothe CP. Therefore, we conclude that the substi-tution of Glu235 and Lys236 for Ala results in afailure to open the pore of the CP. The REGa RTP

site might be necessary to fix the REGa in position,so that the activation loop can enforce the opengate conformation. Interestingly, mutation of a19 S subunit, Rpt2/S4 (rpt2K229R, S241F), resultsin a closed gate conformation of the CP withoutan impact on the 26 S complex formation.13

Tat inhibits 11 S REG-mediated antigenpresentation by MHC class I molecules

Recently, it became clear that the productionof certain epitopes such as moloney MuLVGagl85 – 93-epitope,29 the tyrosine kinase JAK1355 – 363-epitope30 and the MCMV pp89168 – 176-epitope21 isaffected by the 11 S REG.

The presentation of the MCMV pp89168 – 176-epitope can be stimulated by over-expressingREGa in B8 cells.21 This is supported by our datawith transfected B8 cells expressing REGawt. Incontrast, B8 cells expressing REGam were lyzedsignificantly less by effector CD8þ cells as com-pared to B8 cells expressing REGawt. These dataindicate that an intact RTP site of REGa is neededfor the enhancement of the pp89 epitopeproduction.

Full-length Tat as well as Tatpep1 inhibits thepresentation of the pp89 epitope, most likely bycompeting with the endogenous 11 S REG. Bothfull-length Tat and Tatpep1 are taken up by cells23

and Tatpep1 is essential for binding to intracellular

Figure 7. Tat suppresses the pres-entation of an MCMV pp89-derivedantigen by MHC class I molecules.(a) MCMV pp89-expressing mouseB8 cells were incubated without(control) and with Tatpep1 or Tat-pep5. After incubation, B8 cellswere tested for their susceptibilityto lysis by pp89-specific CTLs. Theassay was performed with 5000 tar-get cells and increasing numbers ofeffector cells (E/T ratio). Data areexpressed as means ^SEM, n ¼ 3.Statistical analysis revealed signifi-cant differences between controland Tatpep1 and between Tatpep1and Tatpep5 at E/T ratios 3 and 9(t-test, p , 0.01). (b) Inhibition ofantigen presentation by full-lengthTat. B8 cells were incubated withindicated Tat concentrations andCTL assays were performed as in(a). Data are expressed as means^SEM, n ¼ 3. Statistical analysisrevealed significant differencesbetween control and full-length Tatat E/T ratios of 3 and 9 (t-test,p , 0.05). (c) CTL assays with trans-fected B8 cells expressing REGawtor REGam. Three positive clonesexpressing REGawt (n ¼ 3) andfour expressing REGam (n ¼ 4)

were tested at an E/T ratio of 27:1. The data are means ^SEM. There is a significant difference between cell lysisobtained with REGawt as compared to REGam (t-test, p , 0.05).


proteasomes, as shown by our in vivo experiments.Tatpep5 does not interfere with B8 cell lysis,indicating that the Tat RTP site is necessary for thereduction of 11 S REG-mediated antigen presen-tation by MHC class I molecules.

Tat concentrations used here are higher than theTat concentrations estimated in AIDS patients.3 Atthe moment we do not know whether the Tat-mediated immunosuppression has a quantitativerole during disease progression. The exactrelevance of the described Tat effects to immuno-deficiency during AIDS or whether they can beused for directed immunosuppression has to bestudied in the future.

Materials and Methods

Materials

The CP and the 11 S REG were isolated from humanred blood cells as described.31 The CP complex wasstored at 270 8C and thawed only once. Tat peptides 1to 5 (Tatpep1–Tatpep5) were synthesized using FMOCstrategy on a 433A peptide synthesizer (ABI). Full-lengthTat protein was obtained from the NIH AIDS ReagentProgram (contributor Dr J. Brady).

Kinetic studies

For CP inhibition, 1 mg/ml of isolated CP was incu-bated at 37 8C with 100 mM Suc-Leu-Leu-Val-Tyr-MCA(Bachem) as substrate in a final volume of 100 ml. Thefluorescence was measured at 37 8C with a microtiterplate reader (Fluoroscan II, Labsystems) at 355 nm exci-tation and 460 nm emission over a 60-minute period infive minute intervals. During this period, the reactionwas linear. Full-length Tat and Tat peptides 1–5 wereadded at the final concentrations indicated. For 11 SREG competition experiments, 1 mg/ml of CP and1.6 mg/ml of isolated 11 S REG were incubated. Kineticstudies with recombinant wild-type REGa (wt), REGband mutated REGa (m) proteins were performed undersimilar conditions except that fluorescence was moni-tored at room temperature. Indicated amounts of REGaor REGam or REGa and REGb or REGam and REGbwere pre-incubated with 1 mg/ml of CP for 30 minutesat 37 8C. In the case of competition with REGam, indi-cated amounts of REGam were added to constantamounts of CP, REGa and REGb after 30 minutes ofpre-incubation and then the mixture was incubated foran additional 30 minutes at 37 8C. After pre-incubationthe fluorigenic substrate was added and fluorescencewas measured in ten-minute intervals. The data inFigures 1 and 4 were analyzed using the following rateequation: v ¼ Vmax1 þ Vmax2[Tat]/(K1 þ [Tat])(K2 þ [Tat]).The kinetic constants Vmax1, Vmax2, K1 and K2 are definedas:

Vmax1 ¼ ðKTat1 KTat

2 =K11 S REG1 =K11 S REG

2 Þ½11 S REG�2;

Vmax2 ¼ VmaxðKTat1 =K11 S REG

1 þ KTat2 =K11 S REG

2 Þ½11 S REG�;

K1 ¼ KTat1 ð1 þ ½11 S REG�=K11 S REG

1 Þ;

K2 ¼ KTat2 ð1 þ ½11 S REG�=K11 S REG

2 Þ

Here K1Tat and K2

Tat or K111 S REG and K2

11 S REG are the

apparent dissociation constants for Tat or 11 S REG onthe symmetrical 20 S particle.

Comparison of Tat and REGa structures

The automatic superposition procedure contained inthe dictionary of interfaces in proteins (DIPs) was usedto compare the structural data of Tat (PDB code: TBC)and of REGa (PDB code: AVO).32 For similarityscreening33 of REGa and Tat, the Tat structure wasdissected into pieces and exclusively solvent-accessiblesurface patches of Tat were taken into account.

The homology modeling of human CP structure wasbased on multiple sequence alignment and detailedanalysis of conserved local structures in CP subunits.25

Aligning the modeled structure of human CP with therecently deposited structure of mammalian proteasome(1IRU) revealed a very high degree of similarity for all1472 atoms (the rms (root-mean-square distance) valueequals to 0.29 A).

Cloning strategies

The cDNAs of full-length Tat (86 aa, HXB3, HIV-1GST-Tat Expression Vector, NIH AIDS Reagent Program,contributor Dr A. Rice) and of the deletion mutant,TatD37–72, were generated by PCR and subcloned intoBam HI and Xho I sites of pcDNA3 vector (Invitrogen)encoding an N-terminal Flag-tag. Constructs encodingglutathione-S-transferase (GST)-REGa and GST-REGbwere produced as described.34 Amino acid residuesGlu235 and Lys236 of REGa were substituted for Ala bysite-directed mutagenesis using the Quickchangee Site-directed Mutagenesis Kit (Stratagene). For stable trans-formation of B8 cells, cDNAs of REGawt and REGamwere sub-cloned into Bam HI and Xho I sites of thepcDNA3.1/HisC vector encoding for a N-terminalXpress tag (Invitrogen).

Cell culture

HeLa cells were cultured in RPMI 1640 medium understandard conditions. Transient transfections with full-length Tat and TatD37–72 were performed with the Per-fecte Transfection Kit (Invitrogen) as recommended bythe manufacturer.

B8 mouse fibroblast cells, which stably express theMCMV pp89 protein were cultured using Iscove’s MEM(Biochrom) with 125 mg/ml G418 (Gibco BRL).

Effector CD8þ cells specific for the immunodominantpp89 nonamer YPHFMPTNL presented by H-2Ld

MHC class I molecules were generated, cultured andre-stimulated as described.35

Transfected B8 cells were established using calciumphosphate precipitation. Expression of REGawt orREGam protein was tested by Western blots with ananti-Xpress antibody (Invitrogen).

Cell lysis, immunoprecipitation and non-denaturing electrophoresis

HeLa and B8 cell extracts were prepared using stan-dard procedures. Briefly, 24 hours after transfection themedium was removed and cells were rinsed with ice-cold 1 £ PBS. Ice-cold triple-detergent lysis buffer(50 mM Tris–HCl (pH 8.5), 150 mM NaCl, 0.02% (w/v)sodium azide, 0.1% (w/v) SDS, 1% (v/v) NP-40, 0.5%


(w/v) sodium deoxycholate) with freshly added PMSF(1 mg/ml) was added to the cell culture plates on ice.Then cells were collected and disrupted by repeatedaspiration through a 21-gauge needle. After centri-fugation at 15,000g for 20 minutes at 4 8C, aliquots ofthe cell lysate (supernatant) were used for Westernblots, immunoprecipitation and non-denaturingelectrophoresis.

Immunoprecipitation of the CP, the 19 S regulator andthe 26 S proteasome were performed with an anti-a6/C2 antibody (a gift from K. Hendil) and anti-Rpt2/S4antibody (a gift from C. Gordon). To reduce backgroundcaused by non-specific adsorption, cell lysates (300 ml)were pre-adsorbed to 30 ml protein A–agarose (Pharma-cia) for one hour at 4 8C. After centrifugation (oneminute, 14,000 rpm) 1–3 mg of anti-a6/C2 or anti-Rpt2/S4 antibody were added to the supernatant and incu-bated for one hour at 4 8C. After adding 60 ml of proteinA–agarose the mixture was incubated overnight at 4 8C.Protein A–agarose was washed five times with 1 ml ofice-cold triple-detergent lysis buffer (see above). Thenthe precipitate was boiled for five minutes in 20 ml of 1 £sample buffer. Immunoprecipitates were separated bySDS-PAGE and blotted to nitrocellulose. The blots wereprobed with anti-Flag antibody (Stratagene).

For non-denaturing electrophoresis, 2 ml of cellextracts were separated on a 4–15% (w/v) Phast gel(Pharmacia Biotech., Inc.) at 300 V hours. Proteins wereblotted onto nitrocellulose and probed with an anti-Flagantibody. Western blots were developed by ECLtechnique (Amersham).

Recombinant REGawt, REGb and REGam

pGEX-expression constructs encoding GST-REGawt,GST-REGb and GST-REGam were transformed intoE. coli DH5a and induced with 1 mM IPTG for 90minutes at 37 8C. Protein isolation as well as thrombincleavage was performed as recommended by the manu-facturer (Novagen).

Density gradient centrifugation and Far-Western blots

For binding studies with expressed Tat, 300 ml offreshly prepared HeLa cell lysate were immediatelyloaded on glycerol gradients. Recombinant REGawt,REGb and REGam proteins were pre-incubated with orwithout isolated CP for 30 minutes at 37 8C. Sampleswere separated by glycerol gradient centrifugation(10–40% (w/v) glycerol) as described.31 Fractions werecollected as indicated in the Figures. Proteins were pre-cipitated by TCA and immunoblotted. The anti-REGaand anti-REGb antibodies were produced as described.34

Dot-blots were performed as recommended by themanufacturer (Bio-Rad). Isolated CP (1 mg) was blottedto nitrocellulose. The membrane was washed, thenblocked for two hours at room temperature. Nitrocellu-lose slices with CP dots were incubated with REGawt(24 mg/ml) or REGam (24 mg/ml) for 30 minutes at37 8C. For competition with Tat peptides, REGawt wasincubated in the presence of 96 mg/ml of Tatpep1(approximate molecular mass 4000 kDa), 48 mg/ml ofTatpep2 (approximate molecular mass 2600 kDa) or96 mg/ml of Tatpep5 (approximate molecular mass4000 kDa). After washing, nitrocellulose slices wereprobed with the anti-REGa antibody.

Cytotoxic T-cell assays

To study effects of Tat on antigen presentation, B8 cellsexpressing the pp89 protein were incubated overnightwith Tatpep1 (2 mg/ml), Tatpep5 (2 mg/ml) or full-lengthTat (1.6 mg/ml). In addition, transfectant B8 cells expres-sing the pp89 protein and REGawt or REGam weretested in the CTL assay.

CTL assays were performed with effector CD8þ cellsspecific for the immunodominant pp89 nonamer YPHF-MPTNL presented by H-2Ld MHC class I molecules atday five after re-stimulation.35 Target cells were labeledfor two hours with 51Cr and then a four-hour cytolyticassay was performed with 5000 target cells and gradednumbers of effector T-cells in three replicate cultures.

Acknowledgements

We thank K. Hendil for the anti-a6/C2 antibody,C. Gordon for the anti-Rpt2/S4 antibody andP. Aichele for help with the CTL system. We thankK. Ferrell and M. Seeger for critical reading of themanuscript. The work was supported by a grantfrom the Deutsche Forschungsgemeinschaft (DU229/4) to W.D.

References

1. Watson, K. & Edwards, R. J. (1999). HIV-1 trans-activating (Tat) protein. Biochem. Pharmacol. 58,1521–1528.

2. Rubartelli, A., Poggi, A., Sitia, R. & Zocchi, M. R.(1998). HIV-I Tat: a polypeptide for all seasons.Immunol. Today, 19, 543–545.

3. Westendorp, M. O., Frank, R., Ochsenbauer, C.,Stricker, K., Dhein, J., Walczak, H. et al. (1995). Sensi-tization of T cells to CD95-mediated apoptosis byHIV-1 Tat and gp120. Nature, 375, 497–500.

4. Zocchi, M. R., Poggi, A. & Rubartelli, A. (1997). TheRGD-containing domain of exogenous HIV-1 Tatinhibits the engulfment of apoptotic bodies bydendritic cells. AIDS, 11, 1227–1235.

5. Poggi, A., Rubartelli, A. & Zocchi, M. R. (1998).Involvement of dihydropyridine-sensitive calciumchannels in human dendritic cell function. Com-petition by HIV-1 Tat. J. Biol. Chem. 273, 7205–7209.

6. Viscidi, R. P., Mayur, K., Lederman, H. M. & Frankel,A. D. (1989). Inhibition of antigen-induced lympho-cyte proliferation by Tat protein from HIV-1. Science,246, 1606–1608.

7. Kloetzel, P. M. (2001). Antigen processing by theproteasome. Nature Rev. Mol. Cell Biol. 2, 179–187.

8. Ferrell, K., Wilkinson, C. R. M., Dubiel, W. & Gordon,C. (2000). Regulatory subunit interactions of the 26 Sproteasome, a complex problem. Trends Biochem.Sci., 25, 83–88.

9. Groll, M., Bajorek, M., Kohler, A., Moroder, L.,Rubin, D. M., Huber, R. et al. (2000). A gated channelinto the proteasome core particle. Nature Struct. Biol.7, 1062–1067.

10. Groll, M., Ditzel, L., Lowe, J., Stock, D., Bochtler, M.,Bartunik, H. D. & Huber, R. (1997). Structure of 20 Sproteasome from yeast at 2.4 A resolution. Nature,386, 463–471.


11. Whitby, F. G., Masters, E. I., Knowlton, J. R., Yao, Y.,Wang, C. C. & Hill, C. P. (2000). Structural basis forthe activation of 20 S proteasomes by 11 S regulators.Nature, 408, 115–120.

12. Coux, O., Tanaka, K. & Goldberg, A. L. (1996). Struc-ture and functions of the 20 S and 26 S proteasomes.Annu. Rev. Biochem. 65, 810–847.

13. Kohler, A., Cascio, P., Leggett, D. S., Woo, K. M.,Goldberg, A. L. & Finley, D. (2001). The axial channelof the proteasome core particle is gated by the Rpt2ATPase and controls both substrate entry andproduct release. Mol. Cell, 7, 1143–1152.

14. Chu-Ping, M., Slaughter, C. A. & DeMartino, G.(1992). Identification, purification, and characteriz-ation of a protein activator (PA28) of the 20 S protea-some (macropain). J. Biol. Chem. 267, 10515–10523.

15. Dubiel, W., Pratt, G., Ferrell, K. & Rechsteiner, M.(1992). Purification of an 11 S regulator of the multi-catalytic protease. J. Biol. Chem. 267, 22369–22377.

16. Dubiel, W. & Kloetzel, P. M. (2000). The 20 S protea-some activator PA28 or 11 S regulator. In Proteasomes:The World of Regulatory Proteolysis (Hilt, W. & Wolf,D. H., eds), pp. 129, Landes Bioscience, Georgetown,TX.

17. Knowlton, J. R., Johnston, S. C., Whitby, F. G.,Realini, C., Zhang, Z., Rechsteiner, M. & Hill, C. P.(1997). Structure of the proteasome activatorREGalpha (PA28alpha). Nature, 390, 639–643.

18. Song, X., von Kampen, J., Slaughter, C. A. &DeMartino, G. (1997). Relative functions of the alphaand beta subunits of the proteasome activator, PA28.J. Biol. Chem. 272, 27994–28000.

19. Li, J., Gao, X., Joss, L. & Rechsteiner, M. (2000). Theproteasome activator 11 S REG or PA28: chimerasimplicate carboxyl-terminal sequences in oligo-merization and proteasome binding but not in theactivation of specific proteasome catalytic subunits.J. Mol. Biol. 299, 641–654.

20. Zhang, Z., Clawson, A., Realini, C., Jensen, C.,Knowlton, J. R., Hill, C. P. & Rechsteiner, M. (1998).Identification of an activation region in the protea-some activator REGalpha. Proc. Natl Acad. Sci. USA,95, 2807–2811.

21. Groettrup, M., Soza, A., Eggers, M., Kuehn, L., Dick,T. P., Schild, H. et al. (1996). A role for the proteasomeregulator PA28alpha in antigen presentation. Nature,381, 166–168.

22. Seeger, M., Ferrell, K., Frank, R. & Dubiel, W. (1997).HIV-1 tat inhibits the 20 S proteasome and its 11 Sregulator-mediated activation. J. Biol. Chem. 272,8145–8148.

23. Vives, E., Brodin, P. & Lebleu, B. (1997). A truncatedHIV-1 Tat protein basic domain rapidly translocatesthrough the plasma membrane and accumulates inthe cell nucleus. J. Biol. Chem. 272, 16010–16017.

24. Bayer, P., Kraft, M., Ejchart, A., Westendorp, M.,Frank, R. & Rosch, P. (1995). Structural studies ofHIV-1 Tat protein. J. Mol. Biol. 247, 529–535.

25. Gille, C., Goede, A., Preissner, R., Rother, K. &Frommel, C. (2000). Conservation of substructuresin proteins: interfaces of secondary structuralelements in proteasomal subunits. J. Mol. Biol. 299,1147–1154.

26. Tosi, G., Meazza, R., De Lerma, B. A., D’Agostino, A.,Mazza, S., Corradin, G. et al. (2000). Highly stableoligomerization forms of HIV-1 Tat detected bymonoclonal antibodies and requirement of mono-meric forms for the transactivating function on theHIV-1 LTR. Eur. J. Immunol. 30, 1120–1126.

27. Zaiss, D. M. W., Standera, S., Holzhutter, H.,Kloetzel, P. M. & Sijts, J. A. M. (1999). The protea-some inhibitor PI31 competes with PA28 for bindingto 20 S proteasomes. FEBS Letters, 457, 333–338.

28. Tanahashi, N., Murakami, Y., Minami, Y., Shimbara,N., Hendil, K. B. & Tanaka, K. (2000). Hybrid protea-somes. Induction by interferon-gamma and contri-bution to ATP-dependent proteolysis. J. Biol. Chem.275, 14336–14345.

29. van Hall, T., Sijts, A., Camps, M., Offringa, R., Melief,C., Kloetzel, P. M. & Ossendorp, F. (2000). Differen-tial influence on cytotoxic T lymphocyte epitopepresentation by controlled expression of eitherproteasome immunosubunits or PA28. J. Expt. Med.192, 483–494.

30. Dick, T. P., Ruppert, T., Groettrup, M., Kloetzel, P. M.,Kuehn, L., Koszinowski, U. H. et al. (1996). Coordi-nated dual cleavages induced by the proteasomeregulator PA28 lead to dominant MHC ligands. Cell,86, 253–262.

31. Seeger, M., Gordon, C., Ferrell, K. & Dubiel, W.(1996). Characteristics of 26 S proteases from fissionyeast mutants, which arrest in mitosis. J. Mol. Biol.263, 423–431.

32. Preissner, R., Goede, A. & Frommel, C. (1998).Dictionary of interfaces in proteins (DIP). Data bankof complementary molecular surface patches. J. Mol.Biol. 280, 535–550.

33. Preissner, R., Goede, A. & Frommel, C. (1999).Homonyms and synonyms in the dictionary of inter-faces in proteins (DIP). Bioinformatics, 15, 832–836.

34. Soza, A., Knuehl, C., Groettrup, M., Henklein, P.,Tanaka, K. & Kloetzel, P. M. (1997). Expression andsubcellular localization of mouse 20 S proteasomeactivator complex PA28. FEBS Letters, 413, 27–34.

35. Del Val, M., Munch, K., Reddehase, M. J. &Koszinowski, U. H. (1989). Presentation of cytomega-lovirus immediate-early antigen to cytolytic Tlymphocytes is selectively prevented by viral genesexpressed in the early phase. Cell, 58, 305–315.

Edited by R. Huber

(Received 10 June 2002; received in revised form 2 September 2002; accepted 6 September 2002)


Documents

The RTP Site Shared by the HIV-1 Tat Protein and the 11 S