Pitstop‐2 and its novel derivative RVD‐127 disrupt global cell dynamics and nuclear pores integrity by direct interaction with small GTPases

Abstract Clathrin‐mediated endocytosis (CME) is an essential cell physiological process of broad biomedical relevance. Since the recent introduction of Pitstop‐2 as a potent CME inhibitor, we and others have reported on substantial clathrin‐independent inhibitory effects. Herein, we developed and experimentally validated a novel fluorescent derivative of Pitstop‐2, termed RVD‐127, to clarify Pitstop‐2 diverse effects. Using RVD‐127, we were able to trace additional protein targets of Pitstop‐2. Besides inhibiting CME, Pitstop‐2 and RVD‐127 proved to directly and reversibly bind to at least two members of the small GTPase superfamily Ran and Rac1 with particularly high efficacy. Binding locks the GTPases in a guanosine diphosphate (GDP)‐like conformation disabling their interaction with their downstream effectors. Consequently, overall cell motility, mechanics and nucleocytoplasmic transport integrity are rapidly disrupted at inhibitor concentrations well below those required to significantly reduce CME. We conclude that Pitstop‐2 is a highly potent, reversible inhibitor of small GTPases. The inhibition of these molecular switches of diverse crucial signaling pathways, including nucleocytoplasmic transport and overall cell dynamics and motility, clarifies the diversity of Pitstop‐2 activities. Moreover, considering the fundamental importance and broad implications of small GTPases in physiology, pathophysiology and drug development, Pitstop‐2 and RVD‐127 open up novel avenues.


| INTRODUCTION
Clathrin-mediated endocytosis (CME) is an essential gateway of material exchange and communication of eukaryotic cells with their environment. 1 Its intimate involvement with a multitude of physiologic 2,3 and pathophysiologic 4-6 processes has instigated an intensive search for small molecule compounds capable of inhibiting the clathrinmediated uptake. Among the most recent and potent inhibitors is Pitstop-2, a compound which has been demonstrated to bind directly to the terminal β-propeller domain of the clathrin heavy chain causing an arrest of coated pit dynamics. 7 However, substantial effects other than inhibition of clathrin functions and CME, were subsequently reported by several groups including ours. In effect, the recommendation was expressly stated that Pitstop-2 be used with caution, and that its effects should not be interpreted to conclude anything pertaining to the function of the N-terminal domain of clathrin. 8 First, Pitstop-2 was shown to have an inhibitory effect on the clathrinindependent endocytosis. 9 Then, its ability to disrupt the clathrin function at kinetochores of dividing cells was reported. 10 Subsequently, nonspecificity of the compound was demonstrated in a system where all four potential binding sites at the clathrin terminal domain were inactivated by site-directed mutagenesis. 8 This was followed by further experimental evidence we provided, which demonstrated drastic disruptive effects of Pitstop-2 on the ultrastructure and functional integrity of nuclear pore complexes (NPCs), the mediators of all nucleocytoplasmic transport. 11 NPCs are made up of multiple copies of 30 different types of proteins termed nucleoporins (Nups), a third of which is rich in phenylalanine-glycine (FG) repeats, termed FG-Nups, which control NPC transport and selectivity. [12][13][14] Pitstop-2 disrupted the structural configuration of NPCs and impaired the binding of the import receptor importin-ß to FG-Nups, 11 an essential process for selective receptor-mediated transport across NPCs. 13 The effects of the inhibitor on the NPC functions were subsequently confirmed in an independent study. 15 Some of the activities mentioned above were ascribed to the inhibition of nonendocytic functions of the clathrin heavy chain. 10 Others, however, could not be attributed to the loss of known clathrin functions, 8,11 and alternative protein targets of Pitstop-2 remained unexplored. The aim of this work is to identify the possible alternative targets for Pitstop-2 which could clarify the reported activities. To facilitate the identification of alternative, clathrin-independent, protein targets of Pitstop-2, we designed and synthesized a novel fluorescent derivative of the original inhibitor (lambdaEM $520 nm), 7 termed RVD-127. It was synthesized on the basis of the procedure published for the original Pitstop-2. 16 Among a small library, a fluorescent derivative named RVD-127 containing an N,N-dimethylamino substituent in the 5-position of the naphthyl ring and a MeO-group instead of a para-Br substituent in the styrene moiety was synthesized ( Figure S1) and validated in diverse experiments. In addition, our docking studies with the crystal structure of the N-terminal of clathrin reveal that RVD-127 occupies the same binding site as Pitstop-2 ( Figures S2a,b). From these observations, we hypothesized that RVD-127 will show a similar bioactivity as the lead structure.
We first verified that the fluorescent RVD-127 is able to reproduce the cell biological effects of Pitstop-2. We then reasoned that it should enable us to visualize the spatial distribution of the targets within the cells. Figure 1 shows that RVD-127 was able to significantly decrease the uptake of the classical substrate for CME transferrin in endothelial cells (Figure 1a). It also inhibited both the intrinsic association of importin-β (Alexa-488 labeled) with the FG-Nups in NPCs, and thus its natural concentration in the nuclear envelope, as well as its nucleoplasmic translocation and accumulation (Figure 1b In our previous study, 11 to investigate whether Pitstop-2 had an impact on the NPC interaction with the major active transport receptor importin-β, we incubated digitonin-permeabilized mammalian cells with bacterially expressed and Alexa-488-labeled importin-β. To illustrate the specificity of the importin-β binding, we performed a concomitant immunostaining with mAb414, a monoclonal antibody which specifically binds several FG-Nups. As a result, we were able to quantify the extent of importin-β binding to the NPCs in presence of Pitstop-2 using the antibody signal as a normalization control. Confocal laser scanning microscopy revealed that importin-β was not only able to bind to the NPCs of the digitonin-permeabilized EA.hy926 but also accumulated in the nucleoplasm in presence of the Pitstop-2 negative control or solvent. 11 In presence of Pitstop-2, the situation was markedly different. The binding of importin-β to the NPCs was dramatically reduced while the mAb414 signal remained largely unaffected. Moreover, the intranuclear accumulation of importin-β was largely abolished. Hence, the observations made with RVD-127 in this study are consistent with our previous observations with the original inhibitor Pitstop-2. 11 These observations confirm that although the aforementioned chemical modifications of the original inhibitor apparently lower the affinity of RVD-127 to its targets, they do not abrogate neither the intended inhibitory effect on CME nor the effect on NPC.
2.2 | Pitstop-2 and RVD-127 cause dissociation of importin-β from NPC by direct binding to the small GTPase ran Having confirmed that RVD-127 is able to reproduce the effects of the unmodified inhibitor, we performed confocal microscopy fluorescence imaging of cells stained with RVD-127 to trace its binding sites.
The aforementioned disruption of importin-ß binding (Figure 1b) by the inhibitors implies that they may act on NPCs by directly or indirectly targeting FG-Nups, which are known to mediate importin-ß binding and selective NPC transport. FG-Nups are particularly abundant inside the NPC channel, but they also occur elsewhere across the NPC. Fluorescence imaging analysis in Figure 2 shows RVD-127 staining occurs throughout the cell, indicating that it targets an abundant and highly mobile component in the cell, which will be discussed later on. The analysis also reveals that despite the lack of immediate colocalization with the FG-Nups specific antibody mAb414, RVD-127 localizes along the entire NPC axis (Figure 2d). It was shown that multivalent binding of importins and exportins for different FG-Nups might be comparable for most of FG-Nups along the NPC axis (or at least it was not proven vice versa), and this affinity can change only upon cargo-complex formation or its dissociation upon RanGTP binding. The RVD-127 distribution pattern across the cytoplasmic and nucleoplasmic sides of the NPC correlates fairly well with the NPC dimensions and the distribution of FG-Nups. 17,18 This implies that RVD-127 targets a dynamic key player in bidirectional nucleocytoplasmic transport, closely associated with NPCs and importin-ß, and the small GTPase Ran is an immediate candidate. The presence of the small GTPase Ran is essential throughout the multistep bidirectional receptor-mediated transport cycle across the NPC as thoroughly demonstrated and explained in previous works. 19,20 Ran is also highly mobile and occurs throughout the cell. 21 It shuttles between the cytoplasm and the nucleus in its guanosine diphosphate (GDP) and GTP-bound states. A RanGTP:RanGDP gradient from the nucleus: cytoplasm drives importin-β-cargo directionality and powers the receptor-mediated transport cycle. [19][20][21] The gradient is generated by the spatial separation of the guanine nucleotide exchange factor and Ran GTPase activating protein, which reside in the nucleus and cytoplasm, respectively. 21 In brief, a protein bearing nuclear localization signal is recognized in the cytosol by an import receptor such as importin-α. Subsequent binding to importin-β specifically targets the formed cargo-receptor-complex to the NPC and mediates the translocation through the NPC channel by interaction with FG-Nups. 20,22 Binding of RanGTP to importin-β inside the nucleus dissociates importin-β and leads it back to the cytosol through the NPC channel. 19,20 RanGTP is hydrolyzed in the cytoplasm to RanGDP by the Ran GTPase activating protein, in particular at the cytoplasmic NPC filaments, releasing importinβ for the next transport cycle. 13,23 Eventually, RanGDP returns to the nucleus where it is converted to RanGTP by the chromatin-associated guanine nucleotide exchange factor RCC1. 21 Ran is also required for importin-β interaction with certain FG-Nups, 19 and the translocation of the importin-cargo-complex through the NPC. 24 The cell physiological which explains the abundance of this soluble and mobile small GTPase throughout the cell, with an estimated 10 7 copies per cell. 21 We therefore asked if Ran was implicated in RVD-127-induced effects and tested whether RVD-127 is able to directly associate with purified recombinant Ran. When Ran was immobilized on Ni-NTA affinity beads and immersed in RVD-127 solution in a bead halo assay, 26 a distinct staining of the bead edge was readily detected confirming our hypothesis of direct interaction of RVD-127 with small GTPase Ran (Figure 2e). Although the bead halo assay does not allow distinguishing between the GTP-bound or GDP-bound conformation of Ran, extended time-lapse imaging of live cells treated with RVD-127 shows gradual accumulation of the inhibitor, particularly within the nucleus (Figure 2f), implying that it induces a GDP-bound conformation. In this state, the small GTPase is imported into the nucleus for subsequent nucleotide exchange as mentioned above. 21 Another argument in favor of the assumed RVD-127-induced locking of Ran in the GDP-bound state comes from affinity studies. The affinity of Ran is 10-fold higher for GDP than GTP, whereas both forms of Ran have equal affinity to RCC1. 27 Ran GTP is faster binding and much slower dissociating from Ran. 27 Also, GTP affinity to Ran is higher than that to GDP, when Ran is in complex with importin-β. 28 Further evidence for Ran being a direct target of RVD-127 and Pitstop-2 comes from computational docking analysis of the inhibitors to the crystal structure of RanGDP ( Figure S4). Taken together, we conclude that Ran is directly implicated in the Pitstop-2/RVD-127-mediated effects on importin-β dissociation from NPCs and on NPCs themselves, as observed in this study and our previous work. However, we could not rule out other targets and, therefore, moved on to pinpointing potential protein targets of Pitstop-2, without chemical modifications. This was to ensure that any identified target should not result from the chemical modification of the original inhibitor, in order to obtain the fluorescent derivative. We adopted a complimentary approach to Pitstop-2 target search to verify the results obtained with RVD-127, using the drug affinity responsive target stability (DARTS) assay ( Figure S3). 29 The underlying principle of the assay is based on the notion that specific binding of a small molecule inhibitor to its target protein stabilizes the target and, thus, reduces its susceptibility to proteolysis. We reasoned that if the effects of Pitstop-2 on the NPCs can be observed in digitonin-permeabilized cells or isolated nuclear envelopes, 11 its binding to both the target and the off-target proteins is likely to be retained when the cells are lysed completely. Sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) separation of the cell lysates treated with progressively increasing concentration of a mixture of proteases in presence of the inhibitor vs solvent has revealed a distinct candidate band in the lysates supplemented with Pitstop-2 ( Figure S3).
Mass-spectrometric analysis of the protein band protected from proteolysis has revealed that it was predominantly composed of several isoforms of actin (PDB accession number P60709). The result was reproduced in three independent experiments. This finding together with the originally observed "freezing" of the clathrin-coated pit dynamics, 7 upon of exposure of cells to Pitstop-2, has prompted us to explore a possible effect of Pitstop-2 on actin dynamics. Cytoskeletal actin dynamics is essential for endocytosis, 4,30 and if disrupted, whatever the cause may be, it should consequently give rise to "freezing" of highly dynamic cellular processes including endocytosis. This probably accounts for the "unspecific," yet highly potent inhibitory effect of Pitstop-2 on clathrinindependent endocytosis, reported previously in clathrin knockdown cells. 9 Another possible lead to a consequential actin dynamics inhibition by Pitstop-2, as an implication of Ran inhibition, comes from previous experimental evidence demonstrating a close cross-talk between the actin cytoskeleton and Ran-mediated nuclear transport. 31 Indeed, we have recently shown that reversible interference with the functional NPC integrity was promptly paralleled by drastic inhibition of cell motility and migration of both normal cells and highly aggressive lung cells. 32 Hence, based on the outcome of the DARTS assay we hypothesized that the arrest of clathrin-coated pit dynamics might stem from  Rac1. The latter is a key regulator of cellular dynamics, 34 and was recently shown to increase cell stiffness, 35 which probably explains the substantial decrease in cell stiffness we observed with QNM following exposure to Pitstop-2. To test our hypothesis, a similar bead halo assay as performed with Ran, was also performed with Rac1, confirming a direct binding of RVD-127 to Rac1, as shown in Figure 5.  (Figure 5a, bottom). If the inhibitor acts upstream of actin polymerization by blocking the activation of nucleators by Rac1 effectors, 36 then its activity should be detected in a classical effector pull-down assay. To corroborate this, we tested whether the chemically unmodified Pitstop-2 is able to interfere with the primary function of Rac1 GTPase, its ability to interact with a downstream effector in its GTP-bound state. 37  By extending the spatial and temporal observation framework of Pitstop-2 activity from single clathrin-coated pits to the whole cells observed over hours rather than seconds we found that the previously reported Pitstop-2-induced "freezing" of the clathrin-coated pits 7 is probably a drastic, albeit reversible arrest of the overall cell motility caused by Pitstop-2 interaction with Rac1. The effect of the inhibitor on the cell motility is manifested at significantly lower concentration of Pitstop-2 (7.5 μM) than the inhibition of CME (30 μM).
This speaks strongly against the possibility that the defect of cell motility is caused by a disrupted clathrin-mediated integrin recycling. 44

| CONCLUSIONS
Pitstop-2 is a potent inhibitor of CME, but its inhibitory activities are not restricted to CME. This study sheds light onto the seemingly unspecific inhibitory effects of Pitstop-2, reported on previously. 8,9,11 The diverse inhibitory effects share a common theme, which is the requirement of small GTPases, be it directly or indirectly. We demon-

| Synthesis of RVD-127
To a suspension of pseudothiohydantoin (1.5 g, 13.23 mmol) in DMF was added NaH (318 mg, 13.23 mmol) portionwise and the resultant milky white suspension was allowed to stir at room temperature for 30 min ( Figure S1

| Drug affinity responsive target stability
The assay was performed essentially as described previously. 29 Briefly, the cells were cultured to a confluent state, washed with PBS and lysed using M-PER lysis buffer (Thermo Fischer Scientific GmbH, Schwerte, Germany). Protein concentration was determined using BCA-assay kit (Thermo Fischer Scientific GmbH) and adjusted to 1 mg/ml. The lysate was split into two equal portions which were supplemented either with Pitstop-2 (Abcam, plc., Cambridge, UK) (355 μg/mg protein) or with a corresponding amount of solvent (DMSO). The mixtures were equilibrated with the compounds for 1 h at room temperature and each was split into seven equal portions.
Each portion was supplemented with a corresponding amount of Pronase (Roche, Mannheim, Germany) and proteolysis was carried out for 30 min at room temperature. Reaction was stopped by precipitating the proteins with acetone. Precipitated proteins were dissolved in 1Â Laemmli SDS loading buffer and separated by linear 10% SDS-PAGE with subsequent Coomassie staining. Candidate band was excised out of the gel and subjected to mass spectrometry identification.

| Data analysis
Data analysis of colocalization of RVD-127 and mAb414 staining was

| Code availability
R script written to analyze the data using linear mixed effects model and its validation can be provided upon request.

CONFLICT OF INTEREST
The authors declare no competing financial interest.

DATA AVAILABILITY STATEMENT
Data available on request from the authors.