Quantitative Analysis of Tat Peptide Binding to Import Carriers Reveals Unconventional Nuclear Transport Properties

A detailed study of nuclear import mediated by the HIV-1 Tat peptide (47YGRKKRRQRRR57, TatRRR) is reported. Fluorescence-based measurements, calibration of protein concentrations, and binding assays are exploited to address the physicochemical mechanisms of Tat peptide recognition by the classical importin α (Impα) and importin β (Impβ) receptors both in vitro and in intact cells. We show that TatRRR is an unconventional nuclear localization sequence that binds directly to both Impα and Impβ carriers in the absence of competitors (in vitro), whereas this property is silenced in the actual cellular environment. In the latter case, Impα/β-dependent nuclear import can be successfully restored by replacing the “RRR” stretch with “GGG”. We apply a recently developed method to determine quantitatively TatGGG affinity for each receptor. Based on these results, we can rationalize previous controversial reports on the Tat peptide and provide coherent guidelines for the design of novel intracellular targeting sequences.

A detailed study of nuclear import mediated by the HIV-1 Tat peptide ( 47 YGRKKRRQRRR 57 , Tat RRR ) is reported. Fluorescence-based measurements, calibration of protein concentrations, and binding assays are exploited to address the physicochemical mechanisms of Tat peptide recognition by the classical importin ␣ (Imp␣) and importin ␤ (Imp␤) receptors both in vitro and in intact cells. We show that Tat RRR is an unconventional nuclear localization sequence that binds directly to both Imp␣ and Imp␤ carriers in the absence of competitors (in vitro), whereas this property is silenced in the actual cellular environment. In the latter case, Imp␣/␤-dependent nuclear import can be successfully restored by replacing the "RRR" stretch with "GGG". We apply a recently developed method to determine quantitatively Tat GGG affinity for each receptor. Based on these results, we can rationalize previous controversial reports on the Tat peptide and provide coherent guidelines for the design of novel intracellular targeting sequences.
Cellular compartments are the defining feature of eukaryotic cells. The nucleus is surrounded by a double membrane called the nuclear envelope, which separates the genetic material and transcriptional activity from the translational and metabolic processes of the cytoplasm. Communication between nucleus and cytoplasm is mediated by nuclear pore complexes (1,2), large macromolecular assemblies that punctuate the nuclear envelope. Nuclear pore complexes form a selective barrier that inhibits translocation of large cargo molecules (Ͼ40 kDa) (3), unless they possess specific targeting signals called nuclear localization sequences (NLS). 4 The best characterized NLS consist of either one (monopartite) or two (bipartite) stretches of basic amino acids (4,5). Monopartite NLS are exemplified by the SV40 large T antigen NLS ( 126 PKKKRRV 132 ), and bipartite NLS are exemplified by the nucleoplasmin NLS ( 155 KRPAAT-KKAGQAKKKK 170 ). These sequences are specifically recog-nized by a heterodimer of proteins named importin ␣ (Imp␣) and importin ␤ (Imp␤) (6). Imp␣ binds the NLS specifically (7), whereas Imp␤ both enhances the affinity of Imp␣ for the NLS (8) and mediates the transfer of the cargo-Imp␣ complex across the nuclear pore complex (9). The cargo is then released in the nucleus upon RanGTP binding to Imp␤ (10). Because of the surfeit of known classical NLS-containing proteins, it is assumed that this pathway is the most prevalent in the cell; yet, to date no studies have empirically established the proportion of cargoes imported via this mechanism. Furthermore, there is evidence that NLS sequences with unconventional nuclear import properties may exist, most of which are derived from viral proteins. Among these, we have been interested for a long time in the HIV-1 Tat protein. Tat is an unusual transcriptional transactivator that attaches to cell surface heparan sulfate proteoglycans (11), enters cells by endolysosomal pathways (12), reaches the nucleus (13), and dramatically enhances the processivity of transcription directed by the viral long terminal repeat promoter element (14,15). The nuclear localization properties of Tat protein are commonly ascribed to the stretch 47 YGRK-KRRQRRR 57 (16,17) (also named Tat RRR hereafter). Notably, the same sequence was also shown to be responsible for the cell-penetrating properties of the full-length protein (for a review, see Ref. 18) and for its RNA-binding specificity (19 -21). As mentioned above, several reports ascribe to Tat NLS (and homologous viral sequences) novel nuclear import properties, albeit with contrasting results. In particular, Efthymiadis et al. (16) reported that the Tat NLS is able to mediate nuclear import in vitro in the absence of both Imp␣ and Imp␤, through binding to nuclear components. In turn, Truant and Cullen (17) observed that Tat NLS directly interacts with Imp␤ but not Imp␣ in vitro and showed that Imp␤ is both necessary and sufficient for the nuclear import of Tat into isolated nuclei. Contrary to these in vitro experiments, we recently demonstrated that passive diffusion is the dominant mechanism of Tat peptide-mediated nuclear transport in live cells (22). This apparent paradox is accounted for by the overwhelming binding affinity of the C-terminal RRR stretch toward negatively charged biomolecules (e.g. RNAs) that hinders Tat-peptide interactions with the transport machinery. Indeed, the NLS properties of Tat can be recovered in engineered mutants where the RRR stretch is substituted by other motifs (e.g. GGG; sequence: YGRKKRRQGGG, also named Tat GGG hereafter) (23). Overall, however, the molecular details of the nuclear import process mediated by wild-type and mutant Tat NLS remain elusive. We recently established a novel and straightforward method that combines fluorescence lifetime imaging microscopy (FLIM) and fluorescence recovery after photobleaching (FRAP) with in vivo calibration of protein concentrations, to gain access to both the thermodynamic (binding specificity and affinity) and kinetic (import rate) details of the nuclear transport process in intact cells (24). The broad applicability of the method was demonstrated for the interaction between NLS of SV40 and the transport receptor Imp␣ (24). Here, we extend this quantitative approach to the study of wildtype and mutant Tat-NLS interactions with the classical import carriers Imp␣, Imp␤, and their dual complex. We show that Tat RRR is not able to establish interactions with either Imp␣ or Imp␤ in the intact cellular environment, in keeping with our previous results (22,23). Conversely, we demonstrate that the Tat GGG mutant binds directly to both Imp␣ and Imp␤. Note that the conventional NLS from SV40 can establish direct interactions solely with Imp␣ (activated by Imp␤). Finally, by a complementary in vitro binding assay, we find that in the absence of competitors (i.e. intracellular cytosolic and nuclear factors) Tat RRR does bind to Imp␣ and Imp␤. Overall, these results indicate that Tat RRR is characterized by a nonclassical NLS that is silenced in the cellular environment but can be observed easily in vitro (in the absence of competitors) or restored in vivo in engineered mutants (Tat GGG ). We believe that these findings rationalize the picture of previous controversial results on Tat peptide nuclear import properties and can provide the basic knowledge for the rational design of localization sequences better tailored to the nucleus.

EXPERIMENTAL PROCEDURES
Plasmids and Cell Culture-Plasmids expressing the mCherrytagged NLS SV40 , Tat RRR and Tat GGG sequences were obtained by subcloning starting from their EGFP-tagged counterparts described previously (22). The cDNA encoding for mCherry obtained from the laboratory of Roger Y. Tsien (25) was amplified by PCR introducing the HindIII and EcoRI restriction sites at the 5Ј and 3Ј extremities, respectively. These sites were used to replace EGFP with mCherry. Tat MUT-AA RRR-mCherry and Tat MUT-AA GGG-mCherry mutants were obtained by site-directed mutagenesis using a QuikChange Lightning site-directed mutagenesis kit (Stratagene). In both constructs, the first moiety of Tat sequence MYGRKKRRQ was substituted with MYGRAARRQ. To introduce the two mutations, the following primer (Invitrogen) was used: 5Ј-ATGTATGGCAGGGCG-GCGCGGAGACAG-3Ј. The antisense primer has the respective reverse complementary sequence. The plasmid encoding for the EGFP-tagged importin ␣ (mouse full-length mNPI2) was kindly provided by Yoshihiro Yoneda (Department of Frontier Biosciences, Osaka University) (26). The plasmid encoding for the EGFP-tagged human importin ␤1 was kindly provided by Marilena Ciciarello (Institute of Molecular Biology and Pathology, National Research Council, Rome, Italy) (27). CHO-K1 were purchased from American Type Culture Collection (CCL-61 ATCC) and were grown in Ham's F12K medium supplemented with 10% of fetal bovine serum at 37°C and in 5% CO 2 . Transfections were carried out by using Lipofectamine Reagent (Invitrogen) according to the manufacturer's instruc-tions. For live imaging, ϳ10 5 cells were plated 24 h before experiments onto 35-mm glass bottom dishes (WillCo-dish GWSt-3522). For energy depletion measurements, cells were incubated in medium supplemented with sodium azide and 2-deoxy-D-glucose, as described elsewhere (22).
Cloning, Extraction, and Purification of Recombinant Proteins-Importin ␤ was amplified by using PCR and ligated to pGEX-6P1 vector into BamHI and NotI sites. Importin ␣ was subcloned into pGEX-6P1 vector into EcoRI and SalI sites. Expression of importin ␣ and importin ␤ recombinant proteins were induced in the Escherichia coli BL21 DE3 strain (Invitrogen) growing in the log-phase upon treatment with 1 mM isopropy-B-D-galactoside for 14 h at 20°C. Bacteria expressing recombinant proteins were recovered by centrifugation, resuspended in TE lysis buffer (50 mM Tris-HCl pH 8.3, 1 mM EDTA, 2 mM DTT, 500 mM NaCl, 1 mM PMSF, and protease inhibitors and lysed on ice by sonication. Lysates were clarified by centrifugation. The resulting supernatant was incubated with glutathione-agarose beads at 4°C with gentle rotation. cDNA encoding for Tat RRR -EGFP, Tat GGG -EGFP, NLS SV40 -EGFP, and EGFP (22) were cloned by PCR into pASK-IBA33plus His Tag vectors (IBA vectors). Protein expression is induced upon addition of 200 g anhydrotetracycline per 1 liter of E. coli shaking culture (A 550 nm of 0.5). Purification of His 6 tag proteins was performed according to standard protocols by using gravity flow nickel-nitrilotriacetic acid Superflow columns (IBA BiotagTechnology).
In Vitro Protein-Protein Binding Assays-First, His 6 -tagged fusion proteins were incubated with glutahione-agarose beads for 1 h at 4°C to avoid aspecific binding of fusion proteins to the matrix. GST-tagged importin ␣ and GST-tagged importin ␤ were incubated with a stoichiometric amount of His 6tagged fusion proteins (EGFP-HIS tag, NLS SV40 -EGFP-HIS tag, Tat RRR -EGFP-HIS tag, and Tat GGG -EGFP-HIS tag) in IP buffer (50 mM Hepes, pH 7.4, 150 mM NaCl, 1.5 mM MgCl 2 , 1 mM EGTA, 20 mM NaF, 10% glycerol, 1% Nonidet P-40, and protease inhibitors) at 4°C for 2 h. Then beads were washed four times with IP buffer and incubated with 100 l of 4ϫ SDS gel loading buffer at 95°C for 5 min. Proteins were analyzed by SDS-PAGE and Western blotting. Filter was incubated with anti-GFP monoclonal antibody (JL-8, Clontech, Mountain View, Ca). Purified proteins were also analyzed by Comassie Blue staining.
Fluorescence Microscopy and Concentration Analysis-Cell fluorescence was measured using a Leica TCS SP2 inverted confocal microscope (Leica Microsystems AG, Wetzlar, Germany) interfaced with an argon laser for excitation at 488 nm, and with a HeNe laser for excitation at 561 nm. Glass-bottomed Petri dishes containing transfected cells were mounted in a temperature-controlled chamber at 37°C (Leica Microsystems) and viewed with a 40 ϫ 1.25 numerical aperture oil-immersion objective (Leica Microsystems). Images were collected at low excitation power and monitoring emission by means of the acousto-optical beam splitter detection system of the confocal microscope. The following collection ranges were adopted: 500 -550 nm (EGFP) and 580 -650 nm (mCherry). The global concentrations of intracellular EGFP-and mCherrylinked proteins were determined by using the synthetic adduct fluorescein-glycine, as described thoroughly in a previous publication (24).
FRAP Experiments-Each FRAP experiment started with a four-time line-averaged image (prebleach) of the cell followed by a single-point bleach (nonscanning) near the center of the nucleus with laser pulse at full power to photobleach most of the nuclear fluorescence. Fluorescence recovery was measured by starting a time-lapse acquisition within few milliseconds after bleaching, with the imaging settings described above. Hence, under the assumption of fluorescence proportionality to concentration, the collected FRAP curves in both compartments were fitted to a monoexponential equation (Equation 1), where F 0 and F ∞ label the fluorescence intensity collected at time 0 and asymptotically after bleaching, respectively. Fluorescence values were normalized by the signal of the entire cell at the same time to correct for bleaching caused by imaging and by prebleach fluorescence to verify the presence of an immobile fraction of fluorescent molecules within the nucleus. As described in Ref. 24, the excess flux of cargo toward the nucleus solely due to active transport (⌽ C3N ), the concentration of cargo molecules bound to the importin carrier ([NLS:Imp]) in the cytoplasm, and the nuclear envelope permeability (P X ) are linked by Equation 2, where v C3N (m 3 /s) is the maximum rate for active transport toward the nucleus (i.e. the rate achievable when all of the cargo molecules are bound to import carriers). Thus, the calculated ⌽ C3N was plotted against the cytoplasmic cargo concentration (C NLS/Tat ) for each cell. Finally, if we assume a single binding equilibrium between the NLS/Tat cargo and the import carrier, [NLS:Imp] can be expressed as a function of C NLS/Tat , the global cytoplasmic concentration of import carriers (C Imp ), and the binding dissociation constant (K D *), according to Equation 3.
To recover the biochemically relevant parameter K D *, we fitted the ⌽ C3N versus C NLS/Tat curve with Equations 2 and 3, setting C Imp ϭ 1 M (see Ref. 24 for more details). FLIM Measurements-FLIM measurements were performed illuminating the sample with a 468-nm pulsed laser diode at 40 MHz repetition rate. Fluorescence emission was detected by means of fast photon-counting heads (H7422P-40, Hamamatsu) and time-correlated single photon counting electronics (SPC-830, Becker & Hickl, Berlin, Germany) at 500 -540 nm (band pass filter 510AF23, Omega Optical, Brattleboro, VT). Measurements were performed in living cells with the confocal system described previously with a 40ϫ oil immersion objective (Leica Microsystems). Additional measurements were carried out using a Leica TCS SP5 inverted confocal microscope (Leica Microsystems AG, Wetzlar, Germany) coupled to a PicoQuant single-molecule detection mod-ule. Laser power was adjusted to yield photon-counting rates of ϳ10 5 cps. Fluorescence decay was analyzed by the SPC Image (Becker & Hickl, Berlin Germany) software package. Time-correlated single photon counting detection was used to generate a lifetime map by fitting the fluorescence decay curve in each pixel of the image. We used EGFP as the "donor" fluorophore (high brightness and photostability and monoexponential lifetime) fused to transport carriers (Imp␣ and Imp␤) and mCherry as the "acceptor" (fast maturation, large absorption, and high photostability) fused to the candidate localization sequences (NLS, Tat GGG , and Tat RRR ). Fluorescence decay curves of biological samples containing only isolated donor molecules (i.e. donor alone or in the presence of a noninteracting acceptor) were fitted within a monoexponential decay model. The result of the fitting procedure is thus a single fluorescence lifetime ( F ), characteristic of that donor form. When a mix of unbound and bound donor molecules was present (i.e. donor in presence of an interacting acceptor), lifetime data were fitted to a bi-exponential decay law (Equation 4), where X B and X F (X B ϩ X F ϭ 1) are the amplitude coefficients corresponding to the individual lifetime components of bound ( B ) and free donor molecules ( F ). We set F to the calculated value for isolated donor (see above) and analyzed the distribution of the average lifetime m , according to the relation m ϭ (X B ϫ B ϩ X F ϫ F )/(X B ϩ X F ). A decrease in the m value highlights the appearance of a fraction of donor molecules bound to acceptor molecules. To quantitatively address the equilibrium constant (K D *) of Imp␣-Tat GGG and Imp␤-Tat GGG interactions, we used a combination of fluorophore concentration analysis and FLIM measurements, as described previously (24). Briefly, we calculated the characteristic lifetime of the complex ( B ) by using the ATP depletion assay (see supplemental Fig. 2); then, by setting B and F to the calculated values, we extracted X B and X F molar fractions from Equation 4. Finally, X B and X F were combined with EGFP/mCherry absolute concentrations to derive the actual K D * value of the studied interactions by Equation 5.

RESULTS
Subcellular Localization of Fusion Proteins-First, the fluorescent chimeras used in this study were individually transfected in CHO-K1 cells and analyzed by confocal microscopy (Fig. 1). EGFP-tagged Imp␣ and Imp␤ were detected in both the nucleus and the cytoplasm, with a local enrichment on the nuclear envelope (Fig. 1, A and B). This localization is consistent with their ability to shuttle across the nuclear envelope (28) and bind protein components of the nuclear pore complex (29). As expected, NLS SV40 -mCherry protein was predominantly localized in the nucleus (Fig. 1C), owing to the contribution of carrier-mediated active transport, whereas passively diffusing Tat RRR -mCherry is uniformly distributed across nuclear envelope, with a slight enrichment in nucleolar fluorescence (Fig.  1D). Mutation of the last three arginines into glycines conferred to the Tat peptide the ability to perform active transport; accordingly, Tat GGG -mCherry was localized predominantly in the nucleus (Fig. 1E). The behavior of mCherry-tagged localization signals is in keeping with previous results obtained on their EGFP-tagged counterparts (22,23).
Analysis of Imp␣/␤ Direct Binding to Tat Peptides in Intact Cells-As a first test, we performed FLIM analysis of cells expressing the donor molecule alone (EGFP-Imp␣ or Imp␤-EGFP, as shown in Fig. 2, A and B). As expected, the obtained lifetime decays are well fitted to a monoexponential function (see "Experimental Procedures") yielding the characteristic decay constant of the unquenched donor ( F ϭ 2.56 Ϯ 0.03 ns for EGFP-Imp␣, 2.57 Ϯ 0.03 ns for Imp␤-EGFP, mean Ϯ S.D. for n ϭ 12 cells). Subsequently, we measured EGFP lifetime in cells co-expressing EGFP-Imp␣ and NLS SV40 -mCherry (Fig.   2C). We quantitatively addressed this interaction in a previous publication (24); in this context, it is displayed as a control for the classical import pathway through the adaptor carrier Imp␣. As expected, two exponential components (Equation 4) are necessary for a satisfactory fit of the EGFP lifetime decays (supplemental Fig. S1), revealing the appearance of a fraction of NLS SV40 -bound Imp␣ undergoing FRET (Fig. 2C). The average lifetime values measured show FRET occurrence in all analyzed cells ( m ϭ 2.37 Ϯ 0.13 ns, n ϭ 45). An analogous set of measurements on cells co-expressing Imp␤-EGFP and NLS SV40 -mCherry (Fig. 2D), yielded no detectable FRET signal in n ϭ 24 analyzed cells (accordingly, EGFP lifetime was adequately fitted by a monoexponential function, F ϭ 2.55 Ϯ 0.13 ns, see also example in supplemental Fig. S1). This result confirms that this classical monopartite NLS does not bind directly to Imp␤; in this case, in fact, the transfected Imp␤-EGFP and NLS SV40 -

FIGURE 2. FLIM analysis of Imp␣/␤ direct binding to nuclear localization signals.
A and B, intensity image (gray), lifetime image (color), and lifetime distribution histogram (graphs) of EGFP-Imp␣ and Imp␤-EGFP, respectively. C and D, same FLIM analysis shown in A and B applied to NLS SV40 -mCherry co-transfected with either EGFP-Imp␣ or Imp␤-EGFP. E and F, FLIM analysis of Tat RRR -mCherry co-transfected with EGFP-Imp␣ and Imp␤-EGFP, respectively. G and H, Tat GGG -mCherry co-transfected with either EGFP-Imp␣ or Imp␤-EGFP, respectively. Scale bar in all images, 10 m. Monoexponential fit of decay curves is applied in A, B, D, E, and F ( F is displayed). A biexponential fit is applied elsewhere ( m is displayed). mCherry can form a ternary complex with the endogenous Imp␣. According to the molecular model, Imp␣ binds directly to NLS SV40 -mCherry, whereas importin ␤ binding domain (IBB)-EGFP turn binds to the IBB of Imp␣. Note that FRET efficiency varies nonlinearly with the distance between fluorophores; thus, we can argue that the adaptor protein Imp␣ leads to a distance between tagged NLS SV40 and Imp␤ in the ternary complex that strongly reduces FRET efficiency. The same FLIM-based approach was used to test the capability of Tatbased sequences to bind importins. Consistently with all the results we obtained so far, we found no detectable interaction of Tat RRR -sequence with either Imp␣ or Imp␤ in intact cells (examples are reported in Fig. 2, E and F), as EGFP decays were adequately fitted by a monoexponential function yielding lifetime value close to that of the donor alone ( F ϭ 2.55 Ϯ 0.03 ns, n ϭ 12 for Imp␣ and F ϭ 2.55 Ϯ 0.02 ns, n ϭ 12 for Imp␤). On the other hand, we observed direct interaction of Tat GGG mutant with both Imp␣ and Imp␤ (Fig. 2, G and H); FLIM measurements yielded shorter average lifetimes compared with the donor-only sample ( m ϭ 2.37 Ϯ 0.11 ns, n ϭ 21 for Imp␣ and m ϭ 2.32 Ϯ 0.13 ns, n ϭ 26 for Imp␤), revealing the presence of a fraction of Tat GGG -bound import carriers.
Calculation of Tat GGG -Imp␣ and Tat GGG -Imp␤ Effective Dissociation Constant (K D *) from FLIM Data-As shown recently for the case of NLS SV40 -Imp␣ interaction, affinity values can be extracted from FLIM data (24). Briefly, the K D * value can be calculated if the characteristic lifetime of the complex ( B ) is known ( F is easily derived by measuring the unquenched donor, as shown above). The B values for Tat GGG -Imp␣ and Tat GGG -Imp␤ complexes were derived by FLIM measurements in energy-depleted cells (supplemental Fig. S2) and combined to the corresponding F values to calculate the molar fraction of bound (X B ) and unbound (X F ) import carriers (Equation 4, "Experimental Procedures"). These X B and X F values can be used together with intracellular protein concentrations to derive the effective K D * through Equation 5. For what concerns Tat GGG binding to Imp␣, we found two characteristic ranges of affinity depending on Imp␣ cytoplasmic concentration (Table 1 and Fig. 3A). At an EGFP-Imp␣ concentration close to the endogenous value (Յ1 M (30)), we obtained K D * ϭ 26 Ϯ 5 M, whereas at high EGFP-Imp␣ expression levels (Ͼ10 M), we found K D * ϭ 175 Ϯ 35 M. As recently discussed for NLS SV40 -Imp␣ (24), the two ranges of Tat GGG affinity for Imp␣ can be interpreted as the result of endogenous Imp␤ intervention in altering the binding affinity. This hypothesis was further strengthened by an in vitro binding assay using purified importins and the recombinant fusion protein NLS SV40 -EGFP as a model substrate (supplemental Fig. S3). Remarkably, in the case of Tat GGG binding to Imp␤, we found only one characteristic equilibrium constant (K D * ϭ 320 Ϯ 75 M, Table 1). This K D * FIGURE 3. Affinity values for Tat GGG interaction with import carriers. A, the K D * value for Tat GGG interaction with Imp␣ is plotted against the calculated Imp␣ cytoplasmic concentration. Two average values of affinity are clearly distinguishable; K D * ϳ 26 M for low Imp␣ levels (ϳ1 M), where the endogenous Imp␤ may play a role in modulating Imp␣ autoinhibition (schematic); K D * ϳ 175 M for Imp␣ levels Ͼ10 M, where the contribution of endogenous Imp␤ can be considered negligible. B, in the case of the Tat GGG interaction with Imp␤, we find a broad distribution of affinity values (plot) but with no clear dependence on Imp␤ expression levels. This can be explained by the lack of any possible modulation for the direct interaction to Imp␤ carrier (schematic).

TABLE 1 Affinity constants derived by FLIM
For what concerns Tat GGG binding to Imp␣, were found for Tat GGG -Imp␣ binding depending on the Imp␣ cytoplasmic concentration (Յ1 M or Ͼ10 M). In the case of Tat GGG binding to Imp␤, only one affinity constant was extracted from FLIM data, independently of the carrier cytoplasmic concentration.
value shows no significant correlation with the cytoplasmic carrier concentration (Fig. 3B) and confirms the absence of additional partners modulating Imp␤ affinity toward NLS-endowed molecules. FRAP Analysis of Tat GGG Nuclear Import Kinetics and Binding Specificity-To this point, we showed that Tat GGG is a functional NLS capable of direct binding to both Imp␣ and Imp␤ carriers in live cells, albeit with different relative affinities. Here, we validated our conclusions by investigating the kinetics of Tat GGG nuclear import by FRAP and relating it to the thermodynamics of binding to Importins. Quantitative FRAP analysis of Tat GGG -GFP nucleocytoplasmic shuttling was performed as described in "Experimental Procedures" (example in Fig. 4A). By means of our mathematical model, we could derive the excess flux of cargo toward the nucleus solely due to active transport (⌽ C3N , mol/s) and plot it in Fig. 4B against the cytoplasmic cargo concentration (C cargo ) for each cell. We used the variability of protein expression levels to examine the relationship between cargo concentration and import fluxes. As demonstrated previously (24) and showed here by the green dots in Fig. 4B, NLS SV40 -GFP import fluxes follow a simple linear relationship with respect to the available cytoplasmic cargo concentrations up to 10 -15 M of NLS SV40 -GFP, where they begin to level off, reaching a plateau for high cargo concentrations. Fitting the ⌽ C3N plot to Equations 2 and 3 ("Experimental Procedures") yields an estimate of the maximum rate for active transport toward the nucleus (v CϾN ϳ 300 m 3 /s) and, in the case of NLS SV40 , the binding dissociation constant to Imp␣ (K D *ϳ16 M). Remarkably, replacement of NLS SV40 with Tat GGG leads to a similar "saturation-like" behavior but to much higher maximum rates of cargo delivery to the nucleus (v CϾN Ͼ 1000 m 3 /s; compare plateau levels of red and green dots in Fig. 4B). This evidence points at the presence of a different molecular mechanism for Tat GGG -driven import into the nucleus, relying on two importins rather than just one. Note that the difference between Tat GGG and NLS SV40 becomes particularly evident for cargo concentrations above ϳ50 M. On the basis of the affinities for import carriers calculated by FLIM, we know that this behavior can be linked to Tat GGG direct binding to Imp␤; above 50 M cargo concentration, in fact, we can assume that Tat GGG binding to Imp␣ (K D * ϳ 26 M, by FLIM) almost reached saturation, whereas its binding to Imp␤ (K D *ϳ320 M, by FLIM) starts to play a role in the nuclear import process. We confirmed this hypothesis by fitting Tat GGG import fluxes above 50 M cargo concentration; we obtained K D * ϭ 285 Ϯ 45 M (supplemental Fig. S4), in keeping with FLIM. In Fig. 4C, we show that the purported additional interaction of Tat GGG with Imp␤ is effectively leading to functional transport; Tat GGG -GFP nuclear accumulation above 50 M cargo concentrations is still sustained (K eq close to 2), whereas NLS SV40 -GFP distribution is almost uniform in the cells. (K eq drops to ϳ1 above 50 M cargo concentration.) Furthermore, in Fig. 4D, we show that sequential addition of arginine residues to Tat GGG decreases the relative affinity for the import carriers involved (see the change in slope for the ⌽ C3N curve) but not the maximum rate of transport allowed (plateau values of v C ϾN ), thus suggesting a similar import mechanism for all tested sequences. This behavior supports the hypothesis that the first eight residues of the Tat peptide (YGRKKRRQ) operate as an NLS, whereas the remaining three arginine residues (RRR) hinder active transport. We can further clarify this point by showing that mutations introduced within the first eight residues are able to completely abolish the importin-binding capability of Tat peptides (supplemental Fig. S5).
In Vitro Analysis of Imp␣/␤ Direct Binding to Tat Peptides-The relationship between the import properties of Tat GGG to those of its precursor Tat RRR is crucial and deserves further investigation. Because Tat RRR is unable to perform active transport in intact cells, we address its interaction with import carriers in vitro in the absence of competitors. To this end, we performed a binding assay using recombinant purified proteins. His 6 -tagged Tat RRR -EGFP, Tat GGG -EGFP, NLS SV40 -EGFP, and EGFP as well as GST-tagged Imp␣ and Imp␤ were expressed bacterially and purified as described under "Experimental Procedures" (Fig. 5A). The proteins were mixed, and GST fusion proteins were pulled down by means of glutathione-Sepharose beads. The bound proteins were detected by Western blotting by means of an EGFP antibody. Remarkably, both Imp␣ and Imp␤ interact with Tat RRR -EGFP in our assay (Fig. 5B, lanes 2  and 3). The Tat GGG mutant shows approximately the same ability of Tat RRR and NLS SV40 -EGFP to interact with Imp␣ (lane 5, compare with lanes 3 and 8), but partly loses its ability to directly target Imp␤ (lane 6); quantification of the obtained signals, in fact, reveals an almost 6-fold decrease of binding capability in the latter case. Despite the different experimental conditions, this behavior closely resembles what observed in the actual cellular environment, where Tat GGG direct binding to Imp␣ (K D * ϭ 175 M, in the range where Imp␤ contribution is not relevant) is less efficient than to Imp␤ (K D * ϭ 320 M). This set of results also definitely clarifies that the observed unconventional binding properties are maintained by the wildtype Tat peptide, but only in the absence of cytosolic and nuclear factors.

DISCUSSION
The thorough understanding of any signal-dependent nuclear import mechanism requires a quantitative analysis of both the thermodynamic and kinetic aspects of the phenomenon. To this end, we recently presented a method that combines FLIM and FRAP measurements with protein concentra-tion calibration and showed its application to the analysis of the well known NLS SV40 -Imp␣ interaction (24). Here, we apply the same approach to the study of Tat peptide-mediated nuclear transport. The motivation for this study is our demonstration that the mechanism driving Tat RRR nuclear permeation in live cells is passive diffusion (22), a result contrasting with previous in vitro studies that suggest that active processes are involved (16,17). We recently linked this discrepancy to the observation that the first eight residues of Tat peptide (YGRKKRRQ) can indeed operate as an NLS in engineered mutants (e.g. Tat GGG ), but the remaining three arginine residues (RRR) hinder active transport by promoting binding to intracellular moieties, including RNAs (23). Accordingly, we show here that mutation of the purported NLS stretch of Tat leads to inhibition of active import (i.e. inhibition of importin-binding capabilities) (supplemental Fig. S5). However, the identity of the nuclear import carriers potentially involved in Tat peptide transport is still a matter of debate. In this article, we combine FLIM microscopy and protein concentration calibration to directly monitor Tat peptide-importin interactions and measure the corresponding effective dissociation constant (K D *) in the actual cellular environment. In keeping with all our previous results in live cells, the wild-type Tat RRR sequence shows no detectable interaction with importins. On the contrary, we find that the mutated Tat GGG sequence is a direct target of both Imp␣ and Imp␤. It is worth noting that the K D * of Tat GGG -Imp␣ binding is dependent on Imp␣ expression level, analogously to what we observed for the NLS of SV40 (24). This effect is a consequence of the fact that endogenous Imp␤ can modulate this affinity through direct binding to the autoinhibitory IBB domain of Imp␣ (8,24). On the contrary, the K D * value of Tat GGG -Imp␤ binding is not dependent on the Imp␤ expression level, as expected for a nonmediated interaction. Thanks to FRAP experiments, we obtained independent proof of Tat GGG transport mechanism. By measuring Tat GGG import rate as a function of cargo concentration, we, in fact, recovered a saturation behavior markedly different from that of the classical NLS of SV40. In particular, the much higher Tat GGG import rates suggest the presence of a different molecular mechanism of transport that we argued relies on two importins rather than just one. Accordingly, fitting FRAP data to our model of nucleocy- FIGURE 5. In vitro binding assay. A, purified His-tagged proteins composed by Tat RRR , Tat GGG , and NLS SV40 sequences fused to EGFP and purified recombinant Imp␣ and Imp␤ fused to glutathione S-transferase. B, Western blot (WB) filter showing the direct interaction of Tat RRR and Tat GGG with Imp␣ and Imp␤. The NLS of SV40 was used as a control for the interaction with Imp␣ and not Imp␤, whereas the His-tagged EGFP protein was used as a control for the absence of interaction with import carriers. toplasmic shuttling revealed an additional (low affinity) interaction of Tat GGG with the import machinery (i.e. with Imp␤, as suggested by FLIM). This interaction proved to be functional, as showed by the sustained nuclear accumulation of Tat GGG at high cargo concentrations (Ͼ50 M), compared with the NLS SV40 case. Furthermore, the FRAP assay was used to test the effect of addition of arginines to Tat GGG ; in addition to showing the expected decrease of affinity for the import machinery, our results suggest that all the Tat mutants tested share the same import mechanism. We thus speculate that Tat GGG shares Tat RRR properties and that these are merely progressively unveiled by arginine substitution by restoring its capability to bind importins. Finally, we emphasize that the FRAP assay validates FLIM results in the absence of Imp␣/␤ overexpression. This in turn discounts the possibility that many other cellular importins bind Tat peptides (perhaps with high affinity) and thereby mediate transport under conditions when Imp␣/␤ are not overexpressed, as this would produce a detectable effect on the slope of ⌽ C3N versus C cargo . Our hypothesis of a dual functionality of Tat peptide sequence implies that the importinbinding capability observed for Tat GGG in living cells be fully recovered for Tat RRR in vitro. In the latter case, the absence of cellular components would make the RRR stretch irrelevant and let the "YGRKKRRQ" domain operate as an NLS. This prediction was tested and confirmed based on an in vitro binding assay. We observed that the wild-type Tat peptide can function as an NLS with unconventional properties because it is a direct target of both Imp␣ and Imp␤. Interestingly, we found that Tat RRR binds Imp␣ and Imp␤ with comparable affinity, whereas Tat GGG shows a clear preference for Imp␣. Although the latter result is consistent with the data reported in living cells (dissociation constants calculated by FLIM), the former reveals that the YGRKKRRQ and RRR domains act cooperatively in determining importin-binding specificity and affinity (as we already demonstrated for the complementary binding to intracellular moieties (23)). The observation that Imp␣ directly contributes to Tat peptide transport to the nucleus is new but somewhat expected, as it was recently showed that the "KKRR" domain is widely conserved as an optimal target of Imp␣ (31). We believe that these findings complement previous reports on the Tat peptide properties and lead to a coherent picture on the molecular details of its nuclear import process. More importantly, they provide useful knowledge for the rational design and the accurate in vivo testing of a new generation of localization sequences.