Spectrin: an alternate target for cytoskeletal drugs

Abstract Cytoskeletal drugs having enormous therapeutic potential act on the cytoskeletal components like actin, tubulin either by promoting polymerization or destabilizing the same. Here we present the interaction of the popular cytoskeletal drugs such as taxol, latrunculin and cytochalasin with spectrin, a huge protein with multi domains that forms the cytoskeletal network. Particularly, the actin binding domain of spectrin regulates the dynamics of the actin cytoskeleton. We followed the binding of these drugs to its actin binding domain and intact spectrin as well. These drugs bind with moderate affinity (Kb ∼ 104 M−1) and the interaction with actin binding domain is entropy driven and hydrophobic in nature as determined by Van’t Hoff plot. The docking studies and molecular dynamics simulations further corroborate the experimental findings. Particularly the higher binding constants in the case of latrunculin and cytochalasin to the actin binding domain of spectrin suggest the binding sites are presumably located in its actin binding domain. Communicated by Ramaswamy H. Sarma


Introduction
Cytoskeletal drugs are small molecules that target actin or tubulin thereby affecting their dynamics (Jordan & Wilson, 2004). These drugs are potent chemotherapeutic agents and inhibit cell proliferation by acting on the polymerization dynamics of spindle microtubules, the rapid dynamics of which are essential to proper spindle function (Jordan & Wilson, 2004;Dumontet & Jordan, 2010;Pathak et al., 2021). Among the most successful microtubule-targeted chemotherapeutic drugs are paclitaxel, colchicine and the Vinca alkaloids (Vandecandelaere et al., 1997;Weisenberg et al., 1968;Wilson, 1970;Wani et al., 1971;Noble et al., 1958). Usually these drugs bind to one of three main sites on tubulin, the paclitaxel site, the Vinca domain and the colchicine domain (Parness & Horwitz, 1981;Arnal & Wade, 1995;Snyder et al., 2001;Kumar, 1981;Mitra & Sept, 2008;Morris & Fornier, 2008;Rai & Wolff, 1996;Owellen et al., 1972;Massarotti et al., 2012). The other commonly used drugs are latrunculins and cytochalasins that alter the state of actin polymerization or the organization of actin filaments (Morton et al., 2000;Allingham et al., 2006;Spector et al., 1989;Braet et al., 2008;Foissner & Wasteneys, 2007;Spector et al., 1983;Yarmola et al., 2000;MacLean-Fletcher & Pollard, 1980). However the mode of action of latrunculin seems to be less complex than that of cytochalasin. Latrunculin associates to actin monomers thereby preventing its polymerization to filamentous actin (Morton et al., 2000;Cou e et al., 1987). On the contrary cytochalasins interfere with the normal dynamics of the actin cytoskeleton by binding to the barbed end of actin filaments (Casella et al., 1981;Cooper, 1987;Nair et al., 2008). Understanding the mechanism of action of these commonly used drugs is crucial for the interpretation of data in complex systems, e.g. eukaryotic cells. The interaction of these drugs with cytoskeletal proteins has thus been an area of very active research in light of its medical significance (Bhattacharyya et al., 2008;Dorleans et al., 2007;Banerjee et al., 1997;Garland, 1978;Lambeir & Engelborghs, 1981;Dorleans et al., 2009;Ganesh et al., 2004;Schiff et al., 1979;Manfredi & Horwitz, 1984;Horwitz, 1992). Even though the binding sites of these above-mentioned cytoskeletal drugs in actin and tubulin are well characterized, they also interact with other proteins associated with the erythrocyte cytoskeleton. Earlier reports suggest existence of high affinity cytochalasin B binding sites in human, bovine and rabbit red blood cells (Hebert & Carruthers, 1992;Rampal et al., 1980;Lin & Snyder, 1977). These sites are located on membrane proteins exposed to the cytoplasmic side of the cell membrane and bind cytochalasins A, B, C, D, E and dihydrocytochalasin.
Spectrin is a component of the membrane skeletal network in all metazoan cells known to date and is the major protein of the RBC cytoskeleton, called erythroid spectrin. Newer studies suggest a direct interaction of erythroid spectrin and tubulin (Nigra et al., 2017). In fact it has already been reported that some tubulin associated proteins themselves interact with spectrin directly (Nunez et al., 1979;Jacob et al., 1972;Jacob, 1975). Spectrin is known to interact with several small molecule hydrophobic ligands including the drugs like antitumor antibiotic mithramycin, the local anesthetic dibucaine and the anticancer drug imatinibmesylate (Majee et al., 1999;Das et al., 2016;Mondal & Chakrabarti, 2002). It may be possible that spectrin can also directly interact with the cytoskeletal drugs. In this context the actin binding domain is of great significance (Karinch et al., 1990). The spectrin links the actin cytoskeleton to the cytoplasmic surface of plasma membranes via its actin binding domain and regulates the dynamic state of the actin cytoskeleton (Korsgren & Lux, 2010). To probe the interaction of erythroid spectrin and particularly the actin binding domain with these drugs, we have used fluorescence spectroscopy, molecular docking techniques and molecular dynamics simulations. We have found the taxol binds spectrin as well as the ABD with similar affinity and the binding stoichiometry of one, like tubulin. The binding with intact spectrin is associated with a negative change in enthalpy indicating involvement of van der Waals forces and hydrogen bonding in the complexation process. However, for latrunculin B and cytochalasin B the binding is tenfold greater for the ABD compared to intact spectrin and is entropy drivenindicating the hydrophobic forces predominant in the complexation process.

Isolation and purification of erythroid spectrin
Erythroid spectrin in its dimeric state from ovine blood was isolated and purified to homogeneity following previously published protocol (Patra et al., 2015;Patra et al., 2014). Purity of preparations was checked by SDS-PAGE with 7.5% acrylamide concentration under reducing conditions and protein concentration was measured using Bradford method with BSA as standard ( Figure S1B) (Bradford, 1976).

Expression and purification of the actin binding domain of b-spectrin
After careful literature survey the polypeptide sequence of the ABD of bI-spectrin was determined from the cDNA and polypeptide sequence of human erythroid spectrin (EMBL Data Bank accession numbers J05244 and J05500) (Sahr et al., 1990;Winkelmann et al., 1990). pET151/D-TOPO plasmids with the sequences of interest inserted under control of Lac operon and T7 viral promoter with N-terminal hexahistidine tag and Ampicillin selection marker was purchased from Invitrogen. Sequences were optimized for expression in E. coli. The corresponding plasmid map is provided in the supporting information ( Figure S1A). Plasmids were sequentially electroporated into XL1-Blue and BL21 (DE3) cells for cloning and protein expression respectively (Bose & Chakrabarti, 2019).
The ABD of bI-spectrin had appreciable solubility issues and was mostly incorporated into inclusion bodies. As such expression, extraction, purification and refolding methods were standardized for urea denatured extraction of the domains.
Cells were grown in L.B. media with 100 mg/ml Ampicillin at 37 C and protein expression was induced in a log phase culture, O.D. is 0.6 at 600 nm, with 0.5 mM Isopropyl b-D-1thiogalactopyranoside (IPTG) at 25 C for four hours (Maniatis et al., 1982). Cell pellets were obtained by centrifugation (2000 g). The pellets were lysed by pulsed sonication for ten minutes in a buffer containing 8 M urea, 10 mM Tris-HCl, 100 mM NaCl, pH 8.0, and lysate was clarified by centrifugation at 12,000 Â g to remove cell debris. Lysate was incubated with half its volume of Ni-NTA resin equilibrated with the same buffer for 2 hr at 25 C. The resin was then washed successively with buffer and 20 mM immidazole in the same buffer to elute nonspecific binders. Finally, the proteins were eluted in 200 mM immidazole in the same buffer. The supernatant was collected and loaded in the affinity column. The ABD was obtained in purified form by elution with immidazole buffer (Bose & Chakrabarti, 2019).
The resulting unfolded protein was then sequentially dialyzed against 7 M, 6 M, 5 M, 4 M, 3 M, 2 M, 1 M, and 0.5 M urea in the same buffer before being dialyzed in pure buffer without urea to yield folded protein. The resulting solution was centrifuged at 7,000 Â g for 5 min to remove any misfolded proteins and the supernatant was passed through 0.22 lm syringe filter to yield final pure folded protein. The purity of the sample was checked by 15% SDS-PAGE analysis ( Figure 1C). The fluorescence and CD spectra of the purified actin binding domain have been shown in the Figure S1D and E respectively. Their concentration was measured using the Bradford method with BSA as standard. The protein was stored at 4 C for not more than 48 hr (Bose & Chakrabarti, 2019).

Fluorescence measurements
Steady state fluorescence measurements were performed using a Cary Eclipse (Varian) fluorescence spectrophotometer equipped with a thermostated cell holder (set to 298 K) and 1 cm path length quartz cuvette. The excitation and emission silts were set at 5 nm each and intrinsic fluorescence was monitored by selectively exciting Tryptophan residues of spectrin and the actin binding domain at 295 nm (Sandu et al., 2021;Wang et al., 2018a;Ren et al., 2017;Marsch et al., 2018;Roy et al., 2016). Quenching experiments were performed in a buffer containing 10 mM Tris, 20 mM KCl at pH 7.8 by titrating spectrin and the actin binding domain with increasing concentration of the ligands. The concentrations of spectrin were maintained between 0.2 mM and 0.45 mM and that of the actin binding domain is 4.79 mM. The concentration of the drugs varied from 2-60 mM for cytochalasin B, 2.5-40.5 mM for latrunculin B and 2.3 À 53.6 mM for taxol. The fluorescence emission spectra were recorded between 310 to 450 nm and the spectra of the buffer was subtracted to generate the respective final spectra from each titration.

Evaluation of fluorescence data
To obtain the corrected fluorescence signal in the quenching experiments the observed intensities were subjected to dilution as well as inner filter correction. The spectra reported in the present study are subtracted from the buffer baseline. Inner filter correction was done using the following equation (Wu et al., 2019;Lyu & Wang, 2021) Where 'F c ' and 'F o ' are the corrected and observed fluorescence intensities respectively, and 'A ex ' is the absorbance at excitation wavelength and 'A em ' is the absorbance at emission wavelength.
Where 'F 0 'and 'F' stand for the relative fluorescence intensities in the absence and presence of quencher, DF represents the change in the fluorescence intensity upon addition of quencher, '[Q]' is the quencher concentration, f a denotes the fractional accessibility of the fluorophore to the quencher and 'K SV ' is the Stern-Volmer quenching constant.

Evaluation of binding parameters
From the fluorescence quenching data the binding constants and the number of binding sites on erythroid spectrin and the ABD for the drug binding were evaluated following the Scatchard equation, stated as follows - Where 'r' is the ratio between the concentration of bound ligand and the total number of available binding sites, 'c' is the concentration of free ligand, 'K b ' is the association constant and 'n' is the number of binding sites per protein molecule (Lehrer, 1971;Scatchard, 1946).

Evaluation of thermodynamic parameters
Thermodynamic parameters, 'DH o ' (change in enthalpy), 'DS o ' (entropy change), and 'DG o ' (Gibbs free energy variation), were determined using the Van't Hoff equations, stated as follows- Where 'R' and 'T' are the universal gas constant and absolute temperature respectively and 'K b ' is the (apparent binding constant) determined from the Scatchard equation. DH o and DS o were derived from the slope and intercept of the plot of lnK b versus 1/T using equation (6) (Castellan, 1971;Ross & Subramanian, 1981;Van Holde et al., 2006;Wang et al., 2018b;Li et al., 2017).

Protein structure modeling
Molecular modeling studies were performed to investigate binding of the drugs to different structural domains of erythroid spectrin. The crystal or NMR structures of intact dimeric spectrin are not available. We built the homology model of actin binding domain of spectrin by using Modbase (Pieper et al., 2004). The amino acid sequence of the actin binding domain of erythroid spectrin was taken from Uniprot (Bairoch & Apweiler, 2000). The homologous sequences for ABD in the protein structure database has been identified and used as templates (sequence identity is 68%) followed by the alignment of the target and template sequences. In the next step the framework structure for ABD was built followed by the addition and optimization of side chain atoms and loops. The modeled ABD was further refined and optimized and the overall quality was evaluated using using SAVES server version 6

Molecular docking studies
For docking study of these three drugs with the actin binding domain of spectrin, three dimensional structures of the drugs were extracted from PDB structural data base [5EQI for cytochalasin B, 2Q0U for latrunculin B and 5SYF for taxol]. Docking studies were performed with Auto Dock 4.2 software, which calculated the interaction between the drugs with the actin binding domain domain of b-spectrin (Morris et al., 2009). The rigid docking mode was adopted. The probable conformers of the drugs that bind different the actin binding domain domain were obtained using the Lamarckian genetic algorithm and empirical binding free energy function. For the docking process, a maximum number of 50 conformers were considered for ligand binding to the protein, and the conformer having the lowest free -energy of binding was chosen for further analysis. The lowest energy binding domain was visualized using PyMOL Molecular graphics system (DeLano, 2002).

Molecular dynamics simulation
The molecular dynamics (MD) simulations in GROMACS (GROningen MAchine for Chemical Simulations) package 2019 were performed with the selected docked complexes of actin binding domain of erythroid spectrin along with drug molecules having the lowest free energy of binding and the desired binding residues (Berendsen et al., 1995;Hess et al., 2008). The protein and all the ligand topologies were prepared using the Gromos54a7 force field and PRODRG server respectively. The systems were made in a cubic box using periodic boundary conditions (10 Å distance from the edge of the box) and solvated by applying the extended simple point charge (SPC/E) water model. The systems were neutralized by replacing solvent waters by required numbers of Na þ and Cl À ions. All the bond lengths were controlled by LINCS algorithm (Hess et al., 1997). All long-range electrostatic interactions were determined using the smooth Particle Mesh Ewald (PME) method (Essmann et al., 1995). Energy minimizations were performed using the steepest descent algorithm till convergence. To mimic the cellular environments the systems were prepared at 310 K temperature and 1 atm bar pressure during equilibration using isochoric-isothermal (NVT) Berendsen thermostat and isothermal-isobaric (NPT) Parrinello-Rahman ensemble method respectively (Berendsen et al., 1984;Parrinello & Rahman, 1981). Each MD simulation run was carried out for 200 ns in three replicates and the snapshots saved at 2 ps intervals. Before analysis of the MD results, the periodicity was removed by using the 'gmxtrjconv' and 'gmx convert-tpr' tools in GROMACS. The root mean square deviation (RMSD) of backbone atoms, root mean square fluctuation (RMSF) of Ca residues, radius of gyration of the protein molecule and accessible surface area were calculated using GROMACS programs 'gmxrmsd', 'gmxrmsf', 'gmx gyrate' and 'gmxsasa' respectively throughout the simulation run. The most stable conformation was obtained from 'gmx cluster' for further electrostatic and van der Waals interaction energy calculation using DS software tools. The superimposition with native protein along with docked stable conformations was done using Superimposition server (Maiti et al., 2004). DS Visualizer and VMD were used for structure visualization.

Accessible surface area calculations
The accessible surface area (ASA) of the amino acid residues in the uncomplexed protein and their docked complexes with ligands were calculated using the program NACCESS (Hubbard & Thornton, 1993).The structures corresponding to the minimum score as generated from Autodock analysis of the protein ligand docked structures were chosen in each case. Additionally, structures obtained from the MD studies are also considered for the analysis. The change in relative ASA (all atoms) for residue, i was calculated using: If a residue lost more than 10 Å 2 ASA when going from the uncomplexed to the complexed state it was considered as being involved in the interaction (Ghosh et al., 2009;Dutta et al., 2010).

Statistical analysis
The experiments were carried out at least times and the values are reported as the mean ± S.D. The data were analyzed by oneway analysis of variance (ANOVA) to check the significant differences between the means of different parameters. Tukey test was performed to compare the means. P < 0.001 is considered for statistically significant difference between the means.

Binding of the drugs to erythroid spectrin
We titrated erythroid spectrin and the actin binding domain with all the three drugs separately. Upon addition of increasing concentrations of the drugs there is a significant quenching of tryptophan fluorescence for both erythroid spectrin and the actin binding domain. (Figures S2 and S3 respectively) The quenching data is fitted to the Stern Volmer equation and the quenching constants (K sv ) have been calculated from the Stern Volmer plot of F o /F vs [drug]. The fluorescence spectra in presence and absence of the drugs are shown in supplementary Figures (S2 and S3). In order to establish the mechanism of quenching we studied the effect of temperature on the quenching constant and found no significant increase in K sv with rise in temperature (Table S1A, supplementary data) indicating the mechanism to be static in nature. Furthermore we had evaluated the quenching constant using modified Stern Volmer equation. The linearity was also maintained for modified Stern Volmer plot with an intercept value of 1. Moreover no significant change in the K sv values was found with rise in temperature (Table S1B, supplementary data).

Evaluation of the binding affinities and stoichiometry
The binding of the drugs to both erythroid spectrin and its actin binding domain was determined by the quenching of tryptophan fluorescence at four different temperatures. The binding constants (K b ) at 25 C were evaluated using Scatchard plot, shown in Figure 1. Spectral intensity data are taken from supplementary Figures S2 and S3. K b is of the order of 10 4 M À1 with one binding site for all the drugs upon complexation with the ABD. The binding affinity of taxol with the intact spectrin is same as that of the ABD as well. On the contrary the binding constant of latrunculin B and cytochalasin B is tenfold less for erythroid spectrin. Table 1 summarizes the estimated K b and n value.

Evaluation of thermodynamic parameters
The thermodynamic parameters associated with the binding of the drugs to erythroid spectrin and its actin binding domain as well were evaluated using Van't Hoff's plot, shown in Figure 2, and summarized in Table 2.

Molecular docking studies
Spectrin is a linear combination of the actin binding domain, rod domain containing repeat motifs of the 3-helix bundle and the self-associating domain formed by the N-terminal domain of a-spectrin with the C-terminal domain of b-spectrin. The modeled ABD was validated running the VERIFY 3 D program in SAVES server version 6 ( Figure S6). The modeled ABD contains five tryptophan residues and the relative ASA values are provided in the Table S3 and Figure S7 in the supporting information. The negative binding energies and low inhibition constant values derived from Autodock indicates favorable binding interactions of the actin binding domain with the drugs (Table S2A in the supporting information). The pictorial 2 D representations of the interacting residues in the drug ABD complexes have been shown in the supporting information ( Figure S8). The couple of hydrophobic residues is found close to latrunculin, cytochalasin and taxol in their respective docked conformation with the actin binding domain. The hydrophobic interaction primarily governs the complexation process as confirmed by the van't Hoff plot though the existence of other forces like hydrogen bonding and Van der Waals forces are not ruled out. The docked pose of the drugs to the actin binding domain of spectrin is shown in Figure 3.  To further probe the nature of binding site in the ABD of b-spectrin we calculated the change in relative ASA of each of the complete interacting residues in the complexes and the uncomplexed ABD itself. Interestingly the change in relative ASA of the interacting residues is quite high for all the ABD drug complexes. The results of these calculations are given in Table S2B in the supporting information.

Molecular dynamics studies
The stability of actin binding domain of spectrin and its interaction patterns with drug molecules were analyzed by MD studies (Figure 4). The RMSD and gyration of backbone atoms plotted against 200 ns time scale throughout the trajectory. The actin binding domain of spectrin protein had higher RMSD (range $4Å) than three protein-drug molecule complexes. The results indicate that drug molecules interact strongly with actin binding domain and stabilize the domain and protein-ligand complexes formed. Moreover, actin binding domain with taxol complex had least RMSD (range $3Å) and got stable after 30 ns in comparison to two other protein-drug complexes. The same trend also reflected in protein folding or gyration. The radius of gyration of actin binding domain was least when forming a compact stable complex with taxol molecules. Whereas, the free actin binding domain showed much higher radius of gyration or less compactness in folding than the drug molecule protein Table 2. Thermodynamic parameters associated with the binding drugs to erythroid spectrin and its actin binding domain as derived from Van't Hoff plot. Values are expressed as mean ± standard deviation. The significant difference between mean values is determined by ANOVA.

Discussions
In the present paper we have shown that the cytoskeletal drugs indeed interact with spectrin; to the best of our knowledge this is the first report showing the interaction of these drugs with spectrin. It is well established that taxol targets microtubules whereas latrunculin and cytochalasin act on actin and prevent its polymerization (Jordan & Wilson, 2004;Dumontet & Jordan, 2010;Pathak et al., 2021;Vandecandelaere et al., 1997;Weisenberg et al., 1968;Wani et al., 1971;Morton et al., 2000;Allingham et al., 2006;Spector et al., 1989;Braet et al., 2008;Foissner & Wasteneys, 2007;Spector et al., 1983;Yarmola et al., 2000;MacLean-Fletcher & Pollard, 1980). The mechanism of action of taxol is different from other antimitotic drugs. Taxol promotes tubulin polymerization and halts entry to anaphase (Jordan & Wilson, 2004;Nair et al., 2008).Taxol binds stoichiometrically to the functional taxoidbinding site located in the b-subunit of the tubulin with high binding constant in the order of 10 7 M À1 (Buey et al., 2005;Schiff & Horwitz, 1980;Rao et al., 1994;Combeau et al., 1994;Prota et al., 2013). The binding pocket resides in a deep hydrophobic cleft near the surface of b-tubulin. The taxol mainly interacts with the protein via three hydrogen bonds and multiple hydrophobic contacts. Taxol adopts a T-shaped conformation and induces structuring of the M-loop into a short helix, crucial prerequisite for lateral tubulin interactions (Snyder et al., 2001;Ganesh et al., 2004). The binding leads to a significant increase in the dynamics and flexibility of the portion of b-tubulin that surrounds the bound nucleotide and makes contact with the a-monomer of the next dimer in the protofilament (Mitra & Sept, 2008). Taxol binds spectrin with moderate affinity, the K b value in the order of 10 4 M À1 . Usually, an increase of the binding affinity can be accomplished by making either DH o more negative or DS o more positive. Our Van't Hoff's plot reveals that binding is entropy driven with hydrophobic forces predominant in the interaction process with ABD (Ross & Subramanian, 1981;Mavani et al., 2022). We could also see the hydrophobic residues Ile 75, Leu 78 and Tyr 79 of ABD in close proximity of taxol in its docked conformation and molecular dynamics analysis as well. The huge changes in accessible surface area are observed for these residues upon complexation from docking and molecular dynamics simulation studies ( Figure 6C). Besides hydrogen bonding and van der Waal forces further stabilize the complexation process ( Figure S8C). The RMS fluctuations are minimal for taxol actin binding domain complexes. We could conclude that there may be a similar hydrophobic cleft in ABD which can accommodate taxol molecule in 1:1 ratio. But for intact spectrin the high negative enthalpy value indicates the involvement of hydrogen bonding and van der Waals forces. Both the experimental and computational studies indicate that taxol binds better to the protein compared to other two drugs.
Actin, a highly conserved protein is organized into two structurally related domains, which can be further subdivided into subdomains 1 to 4. The smaller cleft, between subdomains 1 and 3, mediates the interactions of actin with most ABPs and series of drugs including cytochalasins and latrunculins. The efficacy of latrunculins surpasses that of the classic actin-depolymerizing drug cytochalasin D. As per literature survey the factors contributing to the diversity of latrunculins are as follows (i) macrolide ring size, (ii) chirality of the thiazolidine ring or C16 epimerization, and (iii) number of rings or, in other words, macrolide ring opening (Schiff & Horwitz, 1980;Rao et al., 1994). Latrunculin B destabilize F actin by binding to  G actin in 1:1 stoichiometric ratio at the site adjacent to the nucleotide-binding cleft (Massarotti et al., 2012;Spector et al., 1983). Even though the pyran ring hydroxy group is clearly essential for cytotoxicity and actin binding, a couple of hydrophobic residues are also involved in the interaction process with the macrocyclic loop of latrunculin B (El Sayed et al., 2006;Amagata et al., 2008;Furstner et al., 2007;Helal et al., 2013). The binding site in actin constitutes a hydrophobic region mostly conserved in these cases. Earlier in the review by Chakrabarti et al. the fluorescence quenching and electronmicroscopic measurements demonstrated that in spectrin the phospolipid PE binding site which is primarily hydrophobic is localized at one end of dimeric spectrin, either at the selfassociation domain/or at the actin binding domain (Chakrabarti et al., 2006). Our experimental findings confirmed that hydrophobic interactions play predominant role in the complexation process for ABD (Ross & Subramanian, 1981). Moreover we found a Leu 95 and Phe 157 of the ABD surrounding latrunculin B in its docked conformation. The MD simulation studies point out that the alkyl group of Val 139 is in close proximity of the macrolide ring of latrunculin B ( Figure  S8B). However, for intact spectrin the binding constant is tenfold less in the order of 10 3 M À1 . The Van't Hoff's plot indicates that the hydrogen bonding and VanderWaals interaction are primarily involved in the complexation process. It is, therefore, conceivable that the binding sites are located in the ABD of spectrin. Cytochalasins bind to the barbed end of actin filaments, which inhibits both the association and dissociation of subunits at that end. According to literature reports, the dissociation constant of cytochalasin to monomeric actin is quite low in nM range and the stoichiometry of binding is about one cytochalasin per actin filament (Cooper, 1987;Brown & Spudich, 1981;Flanagan & Lin, 1980). Our studies suggest that cytochalasin B do bind to spectrin but with lesser affinity compared to that of actin. However, the binding of cytochalasin B to ABD is more pronounced than that of intact spectrin. The binding with ABD is entropy driven, the hydrophobic forces predominate in the complexation process. Presumably the binding site for cytochalasin B is located in the actin binding domain of b-spectrin. The cytochalasin binding site in actin is located in the hydrophobic cleft between subdomains 1 and 3 of actin. According to the crystal structure by Nair et al, cytochalasin D binds along the length of the largely hydrophobic cleft between subdomains 1 and 3 (Nair et al., 2008). The cleft has dimensions 20 Å by 10 Å. The concave face of the bicyclic isoindolone ring faces the actin cleft while its macrocycle is placed close to the back face of actin (Nair et al., 2008). In the docked conformation of cytochalasin B with ABD, we have noticed a hydrophobic patch surrounding the cytochalasin B. Also, there is a huge change in the accessible surface area of interacting hydrophobic residues ( Figure 6A and Table S2B) upon binding of cytochalasin B to ABD. The binding affinity decreases in the case of intact spectrin. However, the existence of hydrogen bonding and Vander Waals interaction is apparent in case of intact spectrin. Extensive studies are necessary to identify the exact location of the binding site of these drugs in spectrin.

Conclusion
An understanding of the mechanism for commonly used drugs is vital to facilitate interpretation of data in complex eukaryotic systems. Our studies reveal that the common cytoskeletal drugs bind to spectrin and its actin binding domain with moderate affinity. In particular the actin targeting drugs cytochalasin and latrunculin interact better with the actin binding domain as reflected in the binding constant values. Presumably the spectrin can be considered as a secondary target for these cytoskeletal drugs. Detailed studies such as the need to identify the exact location of the binding site should be the subject of future studies.