Thermodynamic Analysis of H1 Nuclear Import

The nuclear import of H1 linker histones is mediated by a heterodimer of transport receptors, known as importinβ and importin7. Interestingly, both importins separately interact with H1, but only as a dimer they facilitate the translocation through the nuclear pore. We identified the H1 binding site of importin7, comprising two extended acidic loops near the C terminus of importin7. The analysis of the H1 import complex assembly by means of isothermal titration calorimetry revealed that the formation of a receptor heterodimer in vitro is an enthalpy-driven process, whereas subsequent binding of H1 to the heterodimer is entropy-driven. Furthermore, we show that the importinβ binding domain of importin7 plays a key role in the activation of importin7 by importinβ. This process is allosterically regulated by importinβ and accounts for a specific tuning of the activity of the importinβ·importin7 heterodimer. The results presented here provide new insights into cellular strategies to even energy balances in nuclear import and point toward a general regulation of importinβ-related nuclear import processes.

The nuclear transport machinery represents one of the major transport systems in eukaryotic cells connecting the nucleus and the cytosol by bridging the nuclear envelope. Nuclear pore complexes (NPCs) 2 are embedded in the nuclear envelope and provide the channels for nuclear transport. NPCs have a dual function; whereas solutes, ions and small macromolecules  are allowed to diffuse passively through the nuclear pore, larger macromolecules or such, whose transfer has to be tightly controlled, are transported by specific, soluble transport receptors in a signal-dependent manner (for review, see Refs. [1][2][3][4]. The majority of known transport signals are specifically recognized by members of the importin␤/karyo-pherin␤ protein superfamily shuttling between nucleus and cytosol, also known as importins and exportins. They are composed of a number of helix-turn-helix motifs termed HEAT tandem repeats, which pack side by side in an almost parallel fashion, forming elongated molecules with a superhelical twist (5)(6)(7)(8). Besides one or more substrate binding sites, these proteins additionally comprise a binding site for the small GTPase Ran (9). Ran represents the central mediator of directionality of nucleocytoplasmic trafficking. Its low intrinsic GTPase activity allows for a rigid control of GTP hydrolysis. The cytosolic GTPase-activating protein RanGAP1 acts as activator of Ran; the additional factors RanBP1 and RanBP2 act as GTPase enhancers and further stimulate Ran's GTP hydrolysis rate. Due to the nucleocytoplasmic compartmentalization of the regulatory factors, with RanGAP in the cytosol and RanGEF in the nucleus, a steep gradient of RanGTP is established across the nuclear envelope with high concentrations in the nucleus and low ones in the cytosol. The RanGTP gradient across the nuclear envelope is thought to provoke a unidirectional translocation of cargoes through the NPC; in terms of nuclear import, binding of nuclear RanGTP to import receptors with associated cargo usually irreversibly terminates import processes by release of the cargo and initiates re-export of the import receptors.
The nuclear import pathways known so far are subdivided in three classes. First, the receptor adaptor pathways with the well characterized example of the import of classical nuclear localization signal bearing substrates by importin␤ (Imp␤) and importin␣ (Imp␣). Second, the single receptor pathways, with a single importin directly binding its cargo as found in the import of small ribosomal proteins by the nuclear import receptors Imp␤, transportin1 (Trn1), importin5 (Imp5), and importin7 (Imp7) (10) or the nuclear import of the parathyroid hormonerelated protein PTHrP by Imp␤ (11). The third class applies a receptor pair, with each importin participating equally in substrate recognition and translocation through the pore, the socalled co-import pathways. A prominent example for such coimport pathways is the nuclear import of H1 linker histones. The functional import receptor for H1 is the Imp␤⅐Imp7 heterodimer (12,13). Only upon dimerization of Imp␤ and Imp7, a cooperative binding to H1 occurs (13). The ternary Imp␤⅐Imp7⅐H1 complex then translocates through the nuclear pore and after translocation sticks to the nuclear basket. Upon binding of nuclear RanGTP to Imp␤, the ternary complex is thought to dissociate into Imp␤, which recycles into the cytosol in the RanGTP-bound state, and an Imp7⅐H1 dimer, which travels on inside the nucleus to the DNA gyres. There, binding of RanGTP to Imp7 possibly represents the signal for transfer of H1 onto the chromatin. RanGTP⅐Imp7 is finally exported back to the cytosol completing the import cycle (10,12).
Strikingly, substrates of co-import pathways tend to exhibit an extended, basic surface, that may be involved in unspecific interactions or aggregation in the cytosol as well as in the nucleus. As the need for coverage and protection of such basic regions during transport arises, only a pair of import receptors is thought to provide a joint substrate binding site large enough to also display a chaperoning activity (12). Interestingly, importins involved in co-import often are functional in other types of import, like Imp␤ and Imp7 as described above.
Because the mechanisms of complex formation of Imp␤ and Imp7 and its specific role in H1 import remain elusive, the question arises of how such versatile molecules like Imp␤ and Imp7 accommodate to their different roles in the co-import pathway of H1 in comparison with common single receptor import. Hence, it is of high interest to thermodynamically characterize H1 import to see whether the import receptors Imp␤ and Imp7 modulate each other upon dimerization and how their cooperativity in binding H1 might be explained and, finally, to investigate if there are general thermodynamic indications on importin versatility.
H1.11L from Gallus gallus inserted into pET-13a (a generous gift from V. Ramakrishnan, Cambridge, UK) was expressed in E. coli BL21 (DE3) in 2YT medium with 0.4 mM isopropyl 1-thio-␤-D-galactopyranoside. The cells were resuspended in a lysis buffer containing 50 mM Tris/HCl, pH 7.5, 100 mM NaCl, and protease inhibitors. After cell lysis in a microfluidizer 110S (Microfluidics, Newton, MA) and clarification by centrifugation at 30,000 ϫ g, 4°C for 45 min, the lysate was titrated to pH 8.8 and subsequently loaded onto a DEAE-Sepharose anion exchange column (GE Healthcare) in a buffer containing 50 mM Tris/HCl, pH 8.8, and 100 mM NaCl. The flow-through containing H1.11L was collected. After adjusting the NaCl concentration to 500 mM, the flow-through was loaded onto a SP-Sepharose cation exchange column (XK 26/20, GE Healthcare) in buffer A containing 50 mM Tris/HCl pH 8.8 and 500 mM NaCl. H1.11L was eluted from the column in a gradient from buffer A to buffer B (50 mM Tris/HCl pH 8.8, 1 M NaCl) and finally purified with a Superdex S75 gel filtration column (XK 26/60, GE Healthcare) in a buffer containing 20 mM Tris/HCl, pH 7.3, and 100 mM NaCl.
Binding Assays-Pulldown assays on Ni-NTA-Sepharose. 2 nmol of His-tagged proteins were immobilized in binding buffer (20 mM Tris/HCl, pH 7.3, 300 mM NaCl, 2 mM ␤-mercaptoethanol) on 50 l of HisTrap-Ni-NTA-Sepharose (GE Healthcare), and after copious washing in binding buffer with 40 mM imidazole, 4 nmol of the corresponding purified binding partners were applied to the Ni-NTA-Sepharose. Complexes were eluted with binding buffer containing 300 mM imidazole and analyzed via SDS-PAGE followed by Coomassie staining.
Pulldown assays on GSH-Sepharose (GE Healthcare) were performed equally with the following exceptions. Washing was done with binding buffer (20 mM Tris/HCl pH 7.3, 300 mM NaCl), and elution was performed in binding buffer containing 20 mM reduced glutathione.
The binding assays for gel filtration analysis of Imp␤/Imp7 interactions were prepared in gel filtration buffer containing 20 mM Tris/HCl pH 7.6 and 100 mM NaCl. For interaction analysis of import receptors with H1 the salt concentration in the gel filtration buffer was increased to 300 mM NaCl. To achieve a good separation of formed complexes from the monomers, the protein quantities were 2-4 nmol of the larger binding partners; the smaller ones were each added in a 1.5-fold molar excess. Complex formation was tested on an analytical Superdex 200 column (S200 10/300 GL, GE Healthcare) and analyzed by SDS-PAGE with subsequent Coomassie staining.
In Vitro Nuclear Import Assays and Immunofluorescence Detection-Preparation of the fluorescent import substrates 4z-rpL23a and 6z-H1 0 was done as described (10,13).
The permeabilization of HeLa cells, the preparation of the energy regenerating system, and the Ran mix were done according to the protocol of Adam et al. (18) and modifications described by Jäkel and Görlich (10). Recombinant nuclear import receptors were applied in concentrations as indicated in the figures. As a positive control, 10 l of reticulocyte lysate were added to the import assay; the negative control lacked any of the recombinant import factors and the reticulocyte lysate. Import was allowed to progress for 15 min at 37°C (import of rpL23a) and 30 min at 30°C (import of H1) before the cells were washed and fixed in 3.7% formaldehyde. For immunofluorescence detection of His-tagged proteins the nuclei were subsequently permeabilized with 0.5% Triton X-100 and washed. Upon blocking with 0.2% fish gelatin the cells were incubated with a ␣-His antibody from Mus musculus and after copious washing incubated with the secondary antibody, a ␣-mouse antibody (Capra hircus) labeled with Alexa 594. Finally nuclear import was monitored by fluorescence microscopy with a Zeiss Axioskop20 microscope. Confocal laser scanning microscopy was done with a Leica DM IRE2 microscope, a Leica TCS SP2 spectral detector, and Leica confocal software. The immunofluorescence signals of 40 randomly chosen nuclei per experiment were quantified with ImageJ. For the comparison of different import assays, the immunofluorescence signal of the negative control was subtracted from all other values before the final evaluation.
Isothermal Titration Calorimetry-For determination of binding isotherms of Imp␤, Imp7, and RanGDPNP, all proteins were in a buffer containing 20 mM Tris/HCl, pH 7.3, 100 mM NaCl, and 1 mM MgCl 2 . The isotherms were recorded at 25°C. For isothermal titration calorimetry (ITC) experiments with H1.11L and for determination of the specific heat capacity of the Imp␤⅐Imp7-interaction (additional binding isotherms were recorded at 15°C and 10°C) all proteins were in a buffer containing 20 mM Tris/HCl, pH 7.3, 200 mM potassium acetate, 20 mM potassium phosphate, and 1 mM MgCl 2 . For H1.11L interactions all isotherms were recorded at 10°C. ITC experiments were carried out with a VP-ITC microcalorimeter (Microcal). The protein concentrations were 5 M in the sample cell and 50 -100 M in the injection syringe. Data were analyzed with Origin 7.0, and the fitting curves were calculated according to one set of binding site models.

RESULTS
Previous studies revealed the Imp␤⅐Imp7 heterodimer to be the only functional import receptor for H1 linker histones (12,13). The binding sites on Imp␤ for Imp7 and H1 0 have already been identified (12) and include HEAT repeats 4 -9 (Imp7) and HEAT repeats 6 -19 (H1 0 ). They overlap and comprise the central part of Imp␤. The interaction site of Imp7 for Imp␤ has also been identified previously (13). The H1 binding site of Imp7 was unknown as was the binding site of Imp␤ for H1 subtypes other than H1 0 . Because evidence was given that during H1 import Imp␤/Imp7 might not only act as import receptors but as chaperones as well (13,19), the question arose as to whether the properties of the binding sites of Imp␤⅐Imp7 for the linker histone H1 support the hypothesized chaperoning activity of the Imp␤⅐Imp7 heterodimer during nuclear import. An energy characterization of the formation of a trimeric Imp␤⅐Imp7⅐H1 complex was of high interest as well to investigate if there are energy requirements for a receptor dimer in H1 nuclear import.
In this study numerous deletion mutants of both Imp␤ and Imp7 were constructed to locate their binding sites for H1.2 and H1.11L, respectively, and to investigate the effects of dimerization of Imp␤ and Imp7 on subsequent H1 binding. Additionally, the energy requirements of trimeric Imp␤⅐Imp7⅐H1 complex formation were characterized by ITC.
The newly designed Imp␤ deletion constructs used in this study comprise fragments with incomplete binding sites either for Ran or Imp7 and H1 0 (see Fig. 1A). In contrast to most previous studies, they were designed in accordance with a crystal structure of Imp␤ (Ref. 8 and supplemental Fig. 1), with special attention to the HEAT repeat organization of the molecule. Only complete HEAT repeats were deleted, to maintain correct folding of the resulting fragments. The Imp␤ deletions should allow for the identification of the H1.2 and H1.11L binding site of Imp␤. The construction of fragments of Imp7 focused on N-terminal deletions (see Fig. 3A), as sequence analysis revealed two extended and rather disordered acidic regions (aa 882-912 and 927-957) at the C terminus of Imp7, which may be involved in substrate recognition. For a complete list of all tested constructs, see supplemental Fig. 2.
Binding Sites of Imp␤ for RanGTP and Imp7 Overlap but Allow for Simultaneous Binding-Although the binding sites of Imp␤ for RanGTP and Imp7 have been elucidated previously, showing an overlap of both sites within Imp␤ (5,7,12), the question of whether binding of Imp␤ to Imp7 and RanGTP are mutually exclusive in vitro has not been addressed so far.
To elucidate the organization of binding sites on Imp␤, the binding capabilities of Imp␤ fragments for RanGTP and Imp7 or Imp7_1002-C, containing solely the importin␤ binding domain of Imp7 (IBB 7 , aa 1008 -1038), were tested in pulldown assays via the recovery of Imp7 and Ran by the immobilized Imp␤ constructs. Gel filtration chromatography was applied to confirm the stability of complexes observed in pulldown experiments (see Figs. 1 and 3). The analysis of truncated Imp␤ fragments confirms the previously published minimal binding site for Imp7, represented by an N-terminal-located part of Imp␤ comprising HEAT repeats 6 -9 (aa 210 -396, Fig. 1). Further N-terminal deletions abolish binding to Imp7_1002-C completely, as is demonstrated for Imp␤_304-C (Fig. 1, A and B).
Therefore, a prominent overlap of the binding sites of Imp␤ for Imp7 and RanGTP, comprising aa 210 -396, could be confirmed. Because RanGTP was previously shown to dissociate the Imp␤⅐Imp7 heterodimer, the question was whether RanGTP achieves this by a competitive displacement mechanism, pointing to interactions of Imp7 and RanGTP with identical amino acid side chains of Imp␤. Surprisingly, in our hands RanGDPNP did not necessarily dissociate the Imp␤⅐Imp7 heterodimer (Fig. 2). In fact, simultaneous binding of Imp7 and RanGTP to Imp␤ was observed in gel filtration experiments, where a significant binding of a purified Imp␤⅐RanGDPNP heterodimer to Imp7 could be verified. In this experiment an Imp7 construct with N-terminal His tag was used which abolishes binding of Imp7 to RanGDPNP (not shown). This suggests that simultaneous binding of RanGTP and Imp7 to Imp␤ is possible in vitro, yet probably with reduced affinity.

Any H1 Subtype Interacts with the Same Binding Site on Imp␤-
With respect to the three-dimensional structures of Imp␤ (5,8), several Imp␤ fragments with intact HEAT repeats were tested in binding to H1. Because the question remained unanswered of whether the contribution of RanGTP to any H1 import complex disassembly is identical, various H1 subtypes were applied in the binding experiments. Despite previous find-ings, where a stable complex formation with H1 0 was only demonstrated for an Imp␤⅐Imp7 heterodimer and the single import receptors were shown to bind H1 0 with strongly reduced affinity (12), a significant and stable complex formation of full-length Imp␤ with H1.2 could be observed in the absence of Imp7 (Fig. 1C). Binding of Imp␤ to H1.2 is mediated by a large portion of Imp␤ involving HEAT repeats 4 -14 (aa 127-641, Fig. 1, A and C), as is shown for Imp␤_127-641. Further deletions of either the N terminus or the C terminus, represented by Imp␤_N-396 and Imp␤_210-C, respectively, abolish H1.2 binding completely. Thus, the binding sites for both Imp7 and H1.2 on Imp␤ do overlap (see Fig. 4, upper), implying a cooperative binding mode of the Imp␤⅐Imp7 heterodimer and H1.2. In addition, the H1.2 binding site is very similar to the one for H1 0 (12), pointing to a common mechanism of binding of Imp␤ to several H1 subtypes. Supporting this hypothesis, a difference in binding of several subtypes of H1 to Imp␤ could not be observed when repeating the experiments with H1 0 and H1.11L from G. gallus (data not shown). All tested histones display significant similarities to each other with sequence identities between 43% (H1.2) and 41% (H1.11L) in comparison to H1 0 (see supplemental Fig. 3). Consecutively, a similarity in the primary structure corresponds to similar if not the same requirements for binding sites on Imp␤⅐Imp7.
Moreover, the RanGTP binding site of Imp␤ (HEAT repeats 1-8 and 14 -15, aa 20 -378 and 639 -682) (5, 7) overlaps with both Imp7 and H1 binding sites (see Fig. 4), as published previously (12). In contrast to the interaction between Imp␤ and RanGTP, which allowed for simultaneous binding of Imp7, a simultaneous binding of H1.2 and RanGTP to Imp␤ could not be observed (data not shown), implying a mutual exclusivity of binding to Imp␤ and, therefore, at least in part identical requirements for interaction sites on Imp␤. This probably includes the acidic loop within HEAT 8 of Imp␤ (aa 333-343), for this loop has been reported to be involved in a competitive displacement of substrates bound to Imp␤ by RanGTP FIGURE 1. The minimal H1 binding site of Imp␤ is located between aa 127 and 641. A, schematic representation of Imp␤ deletion constructs and their qualitative binding capacities for Imp7 and H1, summarizing both H1 0 , H1.11L (data not shown) and H1.2. Imp␤_N-641 (aa 1-641, HEAT-repeats 1-14), Imp␤_N-726 (aa 1-726, HEATs 1-16), Imp␤_32-C (aa 32-876, HEATs 2-19), Imp␤_127-C (aa 127-876, , Imp␤_210-C (aa 210 -876, HEATs 6 -19), and Imp␤_127-641 (aa 127-641, HEATs 4 -14) all contain both the Imp7 binding site (aa 143-409) and the H1 0 binding site (aa 203-876). The fragments Imp␤_N-396 (aa 1-396), with an incomplete H1 0 binding site, and Imp␤_304-C (aa 304 -876), lacking a major part of the interaction site for Imp7, were previously described (Bednenko et al. (15)). The levels of binding are indicated by ϩ and Ϫ, with ϩϩϩ representing wild-type level of binding by full-length Imp␤. As a control, binding of the Imp␤ fragments to Imp7 is indicated as well. The binding sites for RanGTP, Imp7 and H1 all overlap. B, wild-type level of Imp␤ binding to Imp7 or IBB 7 , respectively, is given when aa 127-641 of Imp␤ are present (SDS-PAGE of gel filtration analysis of Imp␤ fragments and Imp7 or IBB 7 ). Further truncations reduce the binding capability (cf. Imp␤_210-C, Imp␤_)N-396). For Imp␤_304-C no binding can be detected. C, wild-type level of Imp␤ binding to H1.2 requires aa127-876 (SDS-PAGE of gel filtration experiments with Imp␤ fragments and H1.2, results for H1 0 identical, data not shown) The lanes show the peak fractions of the Imp␤ fragments. Imp␤_127-641 is still able to significantly bind to H1 yet with slightly reduced affinity. Where co-eluted, H1.2 is marked by an asterisk; the masses of the molecular weight standard (MW) are as indicated. (8). Because Imp7 lacks an acidic loop in a comparable position but instead displays two large, acidic loops near its C terminus, we next determined their role in H1 nuclear import.
H1 Binding by Imp7 Is Conferred by Its Acidic C Terminus-Because the H1 interaction site of Imp7 has not been identified previously, deletion mutants of Imp7 were designed, allowing for the identification of the binding site on Imp7 for the linker histone H1.2. Their ability of binding to H1.2 was analyzed via pulldown assays and gel filtration experiments. As a control for correct folding of the C-terminal fragments, their binding capability to Imp␤ was also tested (data not shown), as IBB 7 is located at the very C terminus (Fig. 3A and Ref. 13).
The C terminus of Imp7 including two prominent acidic loops (aa 882-912 and 927-957, respectively) could be identified to be crucial in mediating H1.2 binding (Fig. 3). The minimal H1.2 binding site of Imp7 in pulldown assays comprises aa 824 -1001. Both Imp7_824-C, lacking the Ran interaction site, and Imp7_N-1001, lacking IBB 7 , are able to bind H1.2 (Fig. 2B). However, stable complex formation of Imp7_824-C and H1.2 could not be observed in gel filtration experiments (data not shown). A level of H1.2 binding comparable with full-length Imp7 requires aa 665-1001 (Fig. 4). Again, as already reported above in the analysis of Imp␤, a difference in binding several subtypes of H1 by Imp7 could not be observed (data not shown). Interestingly, the IBB 7 domain (1008 -1038) is located in close vicinity to the H1 binding site, yet its presence is dispensable for H1 binding to Imp7. This implies a linear organization of binding sites on Imp7, in contrast to Imp␤, indicating cooperative effects of IBB 7 on H1 binding by Imp7 upon heterodimerization with Imp␤. To further characterize the Imp␤ and Imp7 fragments, not only binding to H1 was investigated but nuclear import of H1 as well.
The IBB Domain of Imp7 Induces Cooperativity of Imp␤ and Imp7 in Binding H1-The import capability of truncated fragments of both Imp␤ and Imp7 was tested in in vitro nuclear import assays to elucidate the specific role of each receptor during H1 import. Concerning nuclear import of substrates of single receptor pathways, the activity of an IBB 7 -deficient fragment of Imp7, Imp7_N-1001, remains unaffected, as this fragment is fully operational in nuclear import of the small ribosomal protein rpL23a (Fig. 5A, panel 5), which is known to be a common substrate for such nuclear import pathways (10). When focusing on H1 import, previous findings that only a dimer of Imp␤ and Imp7 is able to function as import receptor for H1 (13) could be confirmed, as none of the import receptors is capable of mediating nuclear import alone (Fig. 5B, panels 3  and 4). Both importins need to be present to regain H1 import capability (Fig. 5B, panel 5). In addition, the heterodimerization before H1 binding is apparently necessary to gain full import competence; adding Imp7 to a preincubated Imp␤⅐H1 complex FIGURE 2. Binding of Imp7 and RanGDPNP to Imp␤ are not mutually exclusive. A, complex formation analysis of Imp␤, Imp7, and RanGDPNP by analytical gel filtration. Preparation of the Imp␤⅐RanGDPNP complex. Imp␤ (12 nmol) was preincubated with RanGDPNP (16 nmol) at 20°C for 15 min, and the Imp␤⅐RanGDPNP complex was subsequently purified by gel filtration (upper chromatogram). The purified Imp␤⅐RanGDPNP complex was then incubated with Imp7 (Imp␤⅐RanGDPNP: 3 nmol, Imp7:2 nmol) at 20°C for 15 min and analyzed by gel filtration. The formation of a trimeric RanGDPNP⅐Imp␤⅐Imp7 complex is demonstrated in the lower chromatogram (light gray, absorption at ϭ 254 nm; dark gray, absorption at ϭ 280 nm). B, the SDS-PAGE of the peak fraction of the RanGDPNP⅐Imp␤⅐Imp7 preparation reveals a complex formation of Imp␤⅐RanGDPNP and Imp7. Imp7, Imp␤, RanGDPNP, and the masses of the molecular weight standard (MW) are indicated. (Fig. 5B, panel 6) does not rescue H1 import capability of Imp␤⅐Imp7. Conversely, the addition of Imp␤ to a preformed Imp7⅐H1 complex allows the reconstitution of H1 import (Fig.  5B, panel 7), yet with a slightly diminished import capability in comparison to the full-length proteins as the receptor concentrations had to be increased in that experiment (0.5 M, fulllength proteins 0.2 M). Notably, a complete loss of H1 import activity can be observed when applying Imp␤ and Imp7_N-1001 (Fig. 5B, panel 8) despite the findings that heterodimerization of Imp␤ and Imp7 is dispensable in in vitro reconstitution assays. Thus, the formation of an Imp␤⅐Imp7 heterodimer is a prerequisite for H1 import in vivo, supporting previous data (13). For further investigation of the specific role of each importin in H1 nuclear import, the reconstitution of H1 import was tested with Imp␤_127-641 and Imp7_598-C. Imp␤_127-641 has an incomplete RanGTP interaction site (20) but comprises any other domain needed for H1 import, namely the interaction sites for Imp7, H1, and the nucleoporins of the NPC. Imp7_598-C comprises the binding sites for H1 and Imp␤ but lacks the Ran binding site completely. The results show that even at high concentrations of the import receptors a co-import of H1 by Imp7 and Imp␤_127-641 cannot be reconstituted (Fig. 5B, panels 9 and 10). Because Nup153, a nucleoporin at the nuclear basket of the NPC, is known to be the final binding partner of Imp␤ before its release into the nucleus upon RanGTP binding to Imp␤, this might be due to an initial blocking of the NPC at the nuclear basket in the very first round of import. Confocal laser scanning microscopy (Fig. 6A) reveals a prominent staining of the nuclear envelope. Although a precise localization of the fluorescence signal is not possible, the rim staining points to a clogging of the NPC. Thus, Ran binding to Imp␤ and dissociating the Nup153⅐Imp␤⅐Imp7⅐H1 complex is indispensable for termination of H1 import. When performing the identical experiment with Imp␤ and Imp7_598-C (Fig.  5B, panels 11 and 12) applying low receptor concentrations, the import capability is lost as well. However, an increase of the receptor concentration allows a reconstitution of H1 import. An immunofluorescence detection with an anti-His antibody (Fig. 6B) shows that His-Imp7C598 accumulates in the nucleus, demonstrating that this deletion is trapped in the nucleus. A quantification of the immunofluorescent signals (Fig. 6C) backs this observation. When subtracting the mean background signal of FIGURE 3. The H1 binding site of Imp7 is located between aa 665 and 1001 including the C-terminal acidic loops of Imp7 and does not overlap with IBB 7 . A, schematic representation of the soluble Imp7 fragments and their qualitative binding capacities to Imp␤ and H1. Four constructs, namely Imp7_598-C (aa 598 -1038), Imp7_665-C (aa 665-1038), Imp7_824-C (aa 824 -1038), and Imp7_1002-C (aa 1002-1038) contain the Imp␤ binding site (aa 1008 -1038, in the following referred to as IBB 7 ), thus, putatively allowing dimerization with Imp␤, whereas five other deletion constructs do not: Imp7_N-1001 (aa 1-1001), Imp7_876 -1001 (aa 876 -1001), Imp7_N-916 (aa 1-916), Imp7_917-1001 (aa 917-1001), and Imp7_876 -916 (aa 876 -917). The latter three additionally lack one or both of the acidic loops of full-length Imp7. The ability of the fragments to interact with Imp␤ and H1 is indicated by ϩ and Ϫ in comparison with wild-type level of binding of full-length Imp7 (ϩϩϩ). Wild-type levels of H1 binding requires amino acids 665-1001. The acidic loops of Imp7 are both indispensable for binding H1 and are solely involved in the interaction with H1 but not in binding either RanGTP or Imp␤. B, SDS-PAGE of binding assays of Imp7 fragments and H1.2. The interaction of glutathione S-transferase-fused fragments of Imp7 with H1.2 was tested in pulldown experiments as described under "Experimental Procedures." The lanes show the elution fractions. The interaction of the other Imp7 fragments with H1.2 was analyzed by gel filtration. Here the peak fractions of the Imp7 fragments are shown. Where co-eluted, H1.2 is marked by an asterisk; the masses of the molecular weight standard (MW) are as indicated.
the negative control from the mean values of the signals of the reconstitution experiments (see supplemental Table 1), the resulting noise-reduced values show that only Imp7_598-C is subject to nuclear trapping. This is demonstrated by the extraordinarily high nuclear immunofluorescence signal when applying 2 M Imp␤-His⅐His-Imp7_598-C (1.71-fold in comparison to 2 M Imp␤-His⅐His-Imp7, Fig. 6C). In contrast to that, all other experiments including the control experiment with Imp␤-His⅐Imp7_598-C show no increase of nuclear immunofluorescence. Consequently, Imp7_598-C is retained efficiently inside the nucleus. Thus, raising the concentration of Imp7_598-C causes an increase of the immunofluorescence signal but in terms of import compensates for the nucleoplasmic trapping. Therefore, the trimeric import complex of Imp␤⅐Imp7⅐H1 is dissociated into RanGTP⅐Imp␤ and Imp7⅐H1. In summary, the interaction between Imp␤ and RanGTP terminates the H1 nuclear import process, whereas the RanGTP binding site of Imp7 is dispensable. As in the absence of IBB 7 , the H1 nuclear import capability is lost; the heterodimerization of Imp␤ and Imp7 before H1 binding is a prerequisite for H1 import.
Thermodynamic Analysis of the Interactions of Imp␤, Imp7, H1, and RanGTP-To answer the questions concerning possible thermodynamic reasons for an Imp␤⅐Imp7 heterodimer in H1 nuclear import, ITC was used to characterize the binding properties of the different complexes by determining their binding constants.
At first, the heterodimerization of Imp␤ and Imp7 was investigated by characterizing the interaction of Imp␤ and Imp7_1002-C, for reasons of convenience referred to as IBB 7 (Figs. 7A and 8A). This fragment was chosen for the interaction studies because the thermodynamic characteristics of fulllength Imp7 and IBB 7 in binding to Imp␤ did not significantly change besides a slightly reduced affinity of full-length Imp7 to Imp␤ (data not shown). The heterodimerization of Imp␤ and IBB 7 is an exothermic process with a binding enthalpy of about Ϫ67 kJ/mol at 25°C, a rather high affinity (K D ϭ 71 nM), and an entropy change ⌬S of Ϫ89.6 J/(K⅐mol). The relation between Gibbs free energy and the changes in enthalpy and entropy reveals that dimerization of Imp␤ and Imp7 is enthalpy-driven despite a significant entropy-enthalpy compensation, which is typical for protein-protein interactions: ⌬G ϭ ⌬H Ϫ T⌬S ϭ Ϫ66.9 kJ/mol ϩ 298 K ϫ 89.6 J/(K⅐mol) ϭ Ϫ40.2 kJ/mol.
The entropy decrease is completely compensated by the enthalpy change. Because the binding isotherm of Imp␤ and IBB 7 at 15°C displays a binding enthalpy of Ϫ54 kJ/mol (not shown), the interaction has a specific heat capacity ⌬C p of Ϫ1.3 kJ/(K⅐mol). The decrease of ⌬C p with decreasing temperature points to the involvement of hydrophobic interactions between Imp␤ and Imp7. When comparing the binding isotherms of Imp␤ fragments with IBB 7 (Figs. 7A and 8A), a decrease of affinity of Imp␤ deletions to IBB 7 is coherent with extending N-and C-terminal deletions of Imp␤ (Fig. 7A), surprisingly including Imp␤_127-C (K D ϭ 258 nM) and Imp␤_N-396 (K D ϭ 216 nM), which apparently bound to Imp7 as efficiently as full-length Imp␤ in pulldown and gel filtration experiments (Fig. 1B). In the case of the interaction between Imp␤_210-C and IBB 7 , the affinity dramatically decreases from 71 nM (full-length Imp␤) to 1.86 M (Imp␤_210-C, Fig. 7A). Thus, for dimerization of Imp␤ and Imp7 at the level of the full-length proteins, the complete N terminus of Imp␤ needs to be present. When deleting the five C-terminal HEAT-repeats 15-19 (aa 642-876), the affinity of Imp␤ to IBB 7 remains unaffected compared with the binding of the full-length protein, as is displayed by Imp␤_N-641. In fact, the affinity even increases (Fig. 7A, Fig. 8A).
Within Imp␤ the interaction sites for Imp7 (HEATs 1-14 for wild-type binding) and RanGTP (5, HEATs 1-15) show a prominent overlap. Because the gel filtration analysis of the interaction between RanGTP⅐Imp␤ and Imp7 revealed simultaneous binding of both Imp7 and RanGTP to Imp␤, ITC was used to check for differences of the binding constants of Imp␤ and IBB 7 and a preincubated RanGDPNP⅐Imp␤ complex and IBB 7 , respectively, to elucidate the influence of RanGTP on the Imp␤⅐Imp7 heterodimer (Fig. 7B). The binding isotherms demonstrate a 7-fold reduced affinity of IBB 7 to Imp␤⅐RanGDPNP with regard to Imp␤ in the absence of Ran. Although monomeric Imp␤ binds IBB 7 with a K D value of 71 nM, the affinity of the Imp␤⅐Ran complex to IBB 7 is significantly lowered, with a K D value of only about 500 nM (Fig. 7A). The corresponding binding enthalpies do not differ, and although a competition of IBB 7 and RanGTP for parts of the binding sites on Imp␤ cannot be excluded, the similar ⌬H values might indicate that RanGTP binding to Imp␤ does not reduce the interaction sites on Imp␤ for IBB 7 . Remarkably, the entropy significantly changes from Ϫ89.6 J/K⅐mol (Imp␤ ϩ IBB 7 ) to Ϫ177.1 J/(K⅐mol) (Imp␤⅐RanGDPNP ϩ IBB 7 ).
In conclusion, the results suggest that the binding sites of Imp␤ for Imp7 and RanGTP are parallel-organized, meaning that they do not only overlap but allow for simultaneous binding of Imp7 and RanGTP to Imp␤ as well. Yet due to a significant reduction of the affinity of Imp␤ to Imp7 upon binding RanGTP, a dissociation of the heterodimer at the nuclear basket caused by RanGTP binding appears likely. An allosteric effect of RanGTP on the dissociation of Imp␤ and Imp7 is plau- FIGURE 4. Spatial organization of the binding sites of Imp␤ and Imp7. The binding sites of Imp␤ for Imp7 and H1 are parallel-organized, meaning that they overlap and allow for simultaneous binding. In contrast, the binding sites of Imp7 for RanGTP (RanBS 7 ), Imp␤ (IBB 7 ), and H1 are sequentially organized and non-overlapping.
sible. The binding sites for Imp7 and H1 on Imp␤ overlap but allow simultaneous binding as well. Therefore, we next addressed the question of if the specific organization of binding sites on Imp␤ for Imp7 and H1 accounts for cooperativity of Imp␤⅐Imp7 in H1 binding, or due to the linear organization of Imp7, if cooperativity of the heterodimer is induced by a conformational switch in Imp7.
Heterodimerization of Imp␤ and Imp7 Partly Balances the Energy Requirements for H1 Binding in Vitro-To understand the interactions within the trimeric Imp␤⅐Imp7⅐H1 complex, the binding constants of each of the participating subcomplexes were determined by isothermal titration calorimetry in a buffer containing relatively high amounts of salt ions (see "Experimental Procedures."). This buffer was required to circumvent precipitation of the ternary Imp␤⅐ Imp7⅐H1 complex and truncated versions thereof at the moment of complex formation. The experiments reveal that H1 binding is an endothermic process (Fig. 7C), in contrast to heterodimerization of Imp␤ and Imp7. An increase of the entropy of about 309 J/K⅐mol (Imp7 ϩ H1) and 360 J/K⅐mol (Imp␤ ϩ H1) points toward a compensation of the binding enthalpy by an entropyincreasing displacement of salt counterions, which are bound on the surface of H1 by the importins. When titrating each of the import receptors with H1, it becomes evident that the affinities of Imp␤ and Imp7 to H1 are very similar to each other, with k D values of 336 nM (Imp␤) and 273 nM (Imp7), respectively (Fig. 7B). The affinity of a preincubated Imp␤⅐Imp7 heterodimer to H1 is about 2-3 times higher in comparison to the single receptors (k D value of 126 nM, Fig. 7B).
Considering the stoichiometry of binding of the single receptors to H1, which is 2:1 (n ϭ 0.5) in both cases, the binding enthalpies and entropies have to be halved to determine the enthalpies of one importin binding to H1. In that case the sums of the single enthalpies and entropies, respectively, equal the enthalpy and entropy of the heterodimer binding to H1. This implies that both receptors interact with different, non-overlapping regions of H1.
Notably, the process of H1 binding by Imp␤⅐Imp7 in vitro is endothermic but exergonic: ⌬G ϭ ⌬H Ϫ T⌬S ϭ 58.7 kJ/mol Ϫ 283 K ϫ 339.9 J/(K⅐mol) ϭ Ϫ37.5 kJ/mol. Thus, the formation of the trimeric Imp␤⅐Imp7⅐H1 complex in vitro is entropydriven. The observed entropy increase here is most likely due to the solvation of salt ions, which have been bound to H1 FIGURE 5. The IBB 7 domain of Imp7 is only required for receptor-heterodimer import. A, in vitro nuclear import assays with rpL23a show that the purified importins Imp␤ and Imp7 are functional (panels 3 and 4). Imp7_N-1001 (panel 5) displays a nearly wild-type level of import capability when compared with full-length Imp7. DAPI, 4Ј,6-diamidino-2-phenylindole; FITC, fluorescein isothiocyanate. B, H1 import can only be reconstituted when both Imp␤ and Imp7 are present (panels 3-5). Heterodimerization of Imp␤ and Imp7 before H1 binding is necessary for reconstitution of the full H1 import capability (panels 5-7). Imp7_N-1001 has no import capability in the co-import of H1 with Imp␤ (panel 8). Whereas import of H1 cannot be reconstituted with Imp7 and Imp␤_127-641 (panels 9 and 10), the import of H1 by Imp␤ and Imp7_598-C (panels 11 and 12) can be achieved with increasing receptor concentrations (retic, reticulocyte lysate). Receptor concentrations are as indicated.
before binding to Imp␤/Imp7 and, therefore, turns the trimerization into an exergonic process. In contrast to that, it appears likely that the entropic contribution to the interaction in a living cell is much less, since H1 is not soluble in the cytosol in the absence of cytosolic chaperones and, thus, does not bind so many salt ions. Hence, focusing on enthalpy changes possibly reflects the in vivo situation better, with less aberrations.
When taking the binding enthalpy of the heterodimerization of Imp␤ and Imp7 under identical conditions into account (⌬H Imp␤⅐Imp7 ϭ Ϫ47 kJ/mol at 10°C), it is obvious that the heterodimerization process compensates in part for the endothermic process of the subsequent H1 binding by Imp␤⅐Imp7 in vitro: ⌬H Imp␤⅐Imp7 ϩ ⌬H Imp␤⅐Imp7⅐H1 ϭ Ϫ47.5 kJ/mol ϩ 58.7 kJ/mol ϭ 11.2 kJ/mol. It is tempting to confer this result to the in vivo situation, but a direct transfer of the complete heat energy from the Imp␤/Imp7 dimerization onto H1 binding is unlikely, whereas an energy transfer by entropy storage is probable.
Consecutively, we investigated by isothermal calorimetry experiments (Fig. 7D) if additional effects account for cooperativity of H1 binding by Imp␤⅐Imp7 besides possibly balancing the energy requirements of substrate binding by dimerization of Imp␤ and Imp7. Compared with the binding isotherm of Imp7 and H1, the isotherm of an Imp7 fragment lacking IBB 7 but displaying a complete H1 binding site, Imp7_N-1001, shows that in the absence of IBB 7 the affinity of Imp7 to H1 does not significantly change. Thus, in contrast to Imp␣, where an autoinhibitory function of IBB ␣ before binding to Imp␤ has been reported (22,23), an autoinhibitory effect of IBB 7 on Imp7 can be excluded. In that case one would expect an increase in affinity to H1 when deleting IBB 7 . When preincubating Imp7_N-1001 with Imp␤ before titration with H1, the resulting binding isotherm shows no stimulated affinity but, rather, a slightly decreased one, with a k D value of 488 nM (Fig. 7B). Hence the IBB 7 domain is a prerequisite for cooperativity of Imp␤ and Imp7 in H1 binding, and this cooperative effect is not simply due to the neutralization of an autoinhibition of Imp7 by Imp␤. Therefore, it appears likely that IBB 7 is a kind of messenger of an induced-fit mechanism of one or even both of the receptors in binding H1. This question was addressed by recording binding isotherms of H1 together with one receptor in full-length and the other in a truncated version that is only able to bind the first receptor but not H1. Thus, it is possible to check which receptor is activated by dimerization. In a first step a potential activation of Imp␤ was examined by titrating a preincubated Imp␤⅐IBB 7 complex with H1. IBB 7 has already been shown to be defect in binding H1 (Fig. 3B). A stimulating effect of IBB 7 on the binding affinity of Imp␤ to H1 cannot be observed; the resulting k D of 408 nM (Figs. 7D and 8B) resembles the k D of full-length Imp␤ with H1 (336 nM, Fig. 7C and 8B). Note that the stoichiometry in this experiment turned out to be 3:1 (Imp␤⅐IBB 7 :H1), indicating a non-physiological mode of interaction. To analyze a possible activation of Imp7 by Imp␤, a binding isotherm of Imp␤_N-396 and H1 was recorded initially, as a minor part of the Imp␤ binding site for H1 is still present in this deletion. Although it did not bind H1 in pulldown assays and gel filtration experiments FIGURE 6. The Ran binding site of Imp␤ mediates the dissociation of the H1 import complex from the NPC. A, in vitro nuclear import assays with subsequent confocal laser scanning microscopy reveal a prominent staining of the nuclear envelope when applying Imp␤_127-641⅐Imp7. DAPI, 4Ј,6-diamidino-2-phenylindole; FITC, fluorescein isothiocyanate; Retic, reticulocyte lysate. B, an immunofluorescence detection of His-tagged proteins in H1 nuclear import shows a nuclear trapping of His-Imp7_598-C. C, a quantification of nuclear-localized immunofluorescence demonstrates that Imp7_598-C accumulates in the nucleus with increasing receptor concentrations. For the noise reduction the value of the immunofluorescence signal in the negative control was subtracted from all other values before the evaluation. The immunofluorescence signal of the full-length heterodimer Imp␤-His⅐His-Imp7 was set to 1. Receptor concentrations are as indicated. (Fig. 1C), a weak binding of Imp␤_N-396 to H1 could be observed in ITC (Fig. 7E) with a k D value of 658 nM (fulllength Imp␤: 336 nM, Fig. 7B). Because the stoichiometry in this experiment turned out to be about 4 to 1, meaning that 4 molecules of Imp␤_N-396 bind to one molecule of H1, the binding enthalpy for a single molecule of Imp␤_N-396 has to be divided by 4, resulting in a ⌬H value of roughly Ϫ20 kJ/mol. This indicates a loss of interaction surface of about 40% in comparison to full-length Imp␤ with H1 (33.4 kJ/mol for a single molecule of Imp␤). The binding isotherm of a preincubated Imp␤_N-396⅐Imp7 complex and H1 (Fig. 7E) reveals two distinct binding events which have separately been analyzed. The first represents the interaction of Imp␤_N396 and H1 with a stoichiometry of n ϭ 0.25. Thus, it can be clearly distinguished from the second binding event, the interaction of Imp7 with H1. This interaction displays a 1:1 stoichiometry, which points toward a binding of Imp7 to H1, which is similar to that of Imp7 to H1 when complexed with full-length Imp␤. Remarkably, the affinity of Imp7 is now significantly increased by Imp␤_N-396 with a k D of 166 nM, in comparison to Imp7 and H1 without Imp␤ (k D ϭ 273 nM, Figs. 7C and 8B). The binding enthalpy of roughly Ϫ35 kJ/mol is ϳ40% lower than that of Imp␤/ Imp7 and H1 (ϳ59 kJ/mol) due to only little contributions of Imp␤_N-396 to the binding enthalpy. The entropy changes are reduced as well, with ⌬S ϭ 253.3 J/K⅐mol (Imp␤/Imp7 ϩ H1: ⌬S ϭ 339.9 J/K⅐mol), indicating less salt ions being released from the surface of H1. Note that the receptor fragment Imp␤_N-396 has already been saturated before (n ϭ 0.25), and its energetic contribution can be considered low. This is probably also true in the case of a re-ordered binding of Imp␤_N-396 to H1, since then the dissociation energy of the initial binding of Imp␤_N-396 to H1 has to be taken into account as well. In conclusion, the experiments clearly indicate that the partial storage of energy within Imp␤ upon heterodimerizing with Imp7 (Fig. 7A) does not directly affect the interaction of Imp␤⅐Imp7 with H1. In fact, the results show that the cooperativity of Imp␤⅐Imp7 is mediated mainly by IBB 7 conferring an activation of Imp7 by Imp␤.

DISCUSSION
In the nuclear import process of the linker histone H1 the heterodimerization of the nuclear import receptors Imp␤ and Imp7 before H1 binding is an important step upon which both importins act cooperatively in binding their substrate (13). By dissecting the single steps of H1 nuclear import in this study, the IBB 7 domain could be identified to be the mediator of cooperativity of Imp␤ and Imp7 in H1 binding. The entropy decrease on the heterodimerization of Imp␤ and Imp7 (Fig. 7A) indicates a partial storage of energy within Imp␤⅐IBB 7 , most likely due to a compression of the superhelical structure of Imp␤ upon binding to IBB 7 . Such structural changes have recently been reported for interac- FIGURE 7. Energetic characterization of the trimeric Imp␤⅐Imp7⅐H1 import complex by ITC. A, the heterodimerization of Imp␤ and IBB 7 is an exothermic reaction. The N terminus of Imp␤ is necessary for wild-type level of binding Imp7. Imp␤ fragments with extending N-terminal deletions, Imp␤_127-C and Imp␤_210-C, show an increasing loss of affinity to IBB 7 . A decrease of affinity is also coherent with extended C-terminal deletions, as is demonstrated by Imp␤_N-396. Wild-type levels of IBB 7 binding are only given with HEAT repeats 1-14 being present (Imp␤_N-641). B, upon association with RanGDPNP, Imp␤ is still able to bind Imp7, yet with significantly reduced affinity. C, binding of H1 to importins is an endothermic reaction. Imp␤ and Imp7 act cooperatively in binding H1. Upon dimerization the affinity of Imp␤⅐Imp7 to H1 is more than doubled in comparison to the single receptors. Note, whereas the single receptors bind two at a time to H1 (n ϭ 0.5), a stoichiometry with n ϭ 1 is only achieved with Imp␤⅐Imp7. D, when Imp7_N-1001, lacking IBB 7 , is titrated with H1 (Imp7_N-1001 ϩ H1), the resulting binding isotherm is similar to full-length Imp7 (Imp7 ϩ H1). Therefore, IBB 7 has no significant autoinhibitory effect on Imp7 in the absence of Imp␤. In the absence of IBB 7 , as demonstrated for preincubated Imp␤ and Imp7_N-1001, which subsequently have been titrated with H1 (Imp␤ /Imp7_N-1001 ϩ H1), a cooperative effect of both receptors on binding H1 cannot be observed. Notably, the presence of IBB 7 has no stimulating effect on Imp␤ affinity to H1 (Imp␤⅐IBB 7 ϩ H1); instead, the stoichiometry of 3:1 (n ϭ 0.33) indicates a non-physiological binding (marked by an asterisk). E, the affinity of Imp7 to H1 is stimulated by Imp␤. Imp␤_N-396 only weakly binds to H1 with a stoichiometry of n ϭ 0.25 but significantly enhances the affinity of Imp7 when preincubated with Imp7 before titration with H1. The first binding event represents binding of Imp␤_N-396 to H1 with n ϭ 0.25 (blue fitting curve), and the second binding event represents the interaction of Imp7 and H1 with n ϭ 1 (green curve). The stoichiometry indicates that Imp7 in complex with Imp␤_N-396 adopts the same role as to when bound to full-length Imp␤. FIGURE 8. Summary of binding affinities derived from ITC experiments. A, binding affinities of full-length Imp␤ and truncated fragments thereof to IBB 7 . The affinity of Imp␤ to IBB 7 is set to 100%; the other affinities represent the quotients of the affinity of full-length Imp␤ and the affinity of the respective mutant. B, binding affinities of Imp␤, Imp7, Imp␤⅐Imp7, and of different truncated complexes thereof to H1. The affinity of Imp␤/ Imp7 to H1 is set to 100%; the other affinities are calculated as in described in A. not only transporters but potent cytosolic chaperones as well. In the context of chaperoning activity of Imp␤ and Imp7, the RanGTP-mediated dissociation of the trimeric Imp␤⅐Imp7⅐H1 complex after the passage through the NPC becomes even more momentous.
Because not only Imp␤ but also Imp7 can bind RanGTP with high affinity, the question arises whether in the presence of RanGTP an Imp7⅐H1 complex would immediately dissociate at the nuclear basket as well. The reduced affinity of Imp7 alone to H1 in comparison with the Imp␤⅐Imp7 dimer supports such an assumption. In that case none of the importins would serve as a chaperone for H1 in the karyoplasm; other chaperones would then be required to allow a safe passage of linker histones to the nucleosomes. In fact, nuclear chaperones for H1 linker histones do exist (for review, see Ref. 27), and their involvement in completing the delivery of H1 histones to the nucleosomes is likely. Hence, we propose a more detailed model of nuclear import of H1 linker histones including these assumptions (see supplemental Fig. 5).
In summary, the nuclear import of linker histones is allosterically regulated in two major steps of the import process, namely cargo recognition by the Imp␤⅐Imp7 heterodimer and dissociation of the trimeric Imp␤⅐Imp7⅐H1-complex by RanGTP inside the nucleus at the nuclear pore complex. Imp␤ allosterically activates Imp7 via its IBB 7 , and this activation is the cause for cooperativity. This suggests that a conformational change of Imp7 upon binding to Imp␤ is likely, pointing to an induced fit mechanism of binding H1. Additionally, the dimerization of Imp␤ and Imp7 before H1 binding is an energetically favorable process. Because the heterodimerization of both importins is able to compensate for the energy requirements of H1 binding in vitro, the necessity of an import receptor dimer for the nuclear import of linker histones is underlined.
Imp␤ turns out to be a general allosteric regulator of coreceptors and adaptors in nuclear import, not only including H1 import, as demonstrated in this study for Imp7, but also classical nuclear localization signal import with Imp␣. Strikingly, Imp␤ is not always the modulator of co-receptors and adaptors but can also be a target of modulation processes, as was demonstrated for the nuclear import of U1 small nuclear ribonucleoproteins (28). In that case the IBB domain of SPN1 (IBB SPN1 ) modulates Imp␤ in a specific way, which allows for the dissociation of Imp␤ from Nup153 independent of RanGTP, in contrast to the Nup153⅐Imp␤⅐IBB ␣ complex. Hence, concerning modulations by or of Imp␤ and the changing energy requirements of different import pathways depending on these modulations, Imp␤ appears to be the center of a cross-talk of nuclear import processes.
Because not only H1 is imported into the nucleus by the Imp␤⅐Imp7 heterodimer but also the integrase of human immunodeficiency virus-1 (HIV-1-IN) (29), in this case with Imp7 being the protagonist and Imp␤ only a minor player, it will be of major importance to obtain crystal structures of H1 and HIV-1-IN in complex with Imp␤⅐Imp7 to understand how in detail Imp7 is regulated by Imp␤ and how HIV-1-IN adopts the features of cellular cargoes such as H1 to make use of a very similar import pathway as well.