Obtustatin and viperistatin represent the shortest known snake venom monomeric disintegrins. In the present study, we have produced recombinant full-length wild-type and site-directed mutants of obtustatin to assess the role of the K21TS23 tripeptide and C-terminal residues for specific inhibition of the α1β1 integrin. Thr22 appeared to be the most critical residue for disintegrin activity, whereas substitution of the flanking lysine or serine residues for alanine resulted in a less pronounced decrease in the anti-α1β1 integrin activity of the disintegrin. The triple mutant A21AA23 was devoid of blocking activity towards α1β1 integrin-mediated cell adhesion. The potency of recombinant KTS-disintegrins also depended on the residue C-terminally adjacent to the active motif. Substitution of Leu24 of wild-type obtustatin for an alanine residue slightly decreased the inhibitory activity of the mutant, whereas an arginine residue in this position enhanced the potency of the mutant over wild-type obtustatin by 6-fold. In addition, the replacements L38V and P40Q may account for a further 25-fold increase in α1β1 inhibitory potency of viperistatin over KTSR-obtustatin.

INTRODUCTION

Integrins are αβ heterodimeric type-I transmembrane glycoproteins that mediate a range of regulated and dynamic interactions between cells and extracellular adhesion molecules. In humans, 18 α and eight β integrin subunits combine in a restricted manner to form at least 24 dimers, each of which exhibits a distinct ligand-binding profile [13]. Integrins transduce bidirectional signals across the plasma membrane, between the ECM (extracellular matrix) outside a cell and the cytoskeleton inside the cell [4,5], and integrin-mediated cell attachment to the ECM is a basic requirement to build a multicellular organism. Among the ligands of integrins are fibronectin, vitronectin, laminin and collagen. The collagen-binding integrins (α1β1, α2β1, α10β1 and α11β1) contain an approx. 200-amino-acid-residue (I- or A-) domain inserted between the second and the third Ca2+-binding motifs within the N-terminal part of the α subunit that plays a central role in ligand binding [6], and play an important role in the physiology of many organs and in innate immunity, inflammation and autoimmunity [7]. The αI-domain-containing integrins that mediate direct adhesion to collagens seem to have a relatively recent history, since they have been found only in the vertebrate clade [8]. Among them, integrin α1β1 has been recognized as a high-affinity receptor for collagen type IV, and binds also, albeit with less avidity, to other collagens such as type I, XIII and XVI [9,10]. Integrin α1β1 is highly expressed on microvascular endothelial cells, and blocking of its adhesive properties significantly reduces the vascularization ratio and tumour growth in the mouse [1113], suggesting potential stategies to therapeutically target these receptors to alleviate or treat disease.

Viperidae snakes have developed in their venoms an efficient and almost complete repertoire of β1- and β3-integrin receptor antagonists [14]. Thus, with the exception of the α2β1 integrin, which is targeted by a number of C-type lectin-like proteins [15], inhibitory motifs towards β1 and β3 integrins have evolved in different members of the disintegrin family [1619]. Among them, RGD (Arg-Gly-Asp) blocks integrins α8β1, α5β1, αvβ1, αvβ3 and αIIbβ3; MLD (Met-Leu-Asp) targets the α3β1, α4β1, α4β7, α7β1 and α9β1 integrins; VGD (Val-Gly-Asp) and MGD (Met-Gly-Asp) impair the function of the α5β1 integrin; KGD (Lys-Gly-Asp) inhibits the αIIbβ3 integrin with a high degree of selectivity;WGD (Trp-Gly-Asp) has been reported to be a potent inhibitor of the RGD-dependent integrins α5β1, αvβ3 and αIIbβ3; and KTS (Lys-Thr-Ser) and RTS (Arg-Thr-Ser) represent selective α1β1 inhibitors [14,2022].

KTS/RTS-disintegrins, 41-residue monomeric polypeptides cross-linked by four conserved disulfide bonds, represent the shortest snake venom disintegrins described to date [13]. The disintegrin obtustatin, isolated from the venom of Vipera lebetina obtusa, was the first described α1β1 antagonist [11,23]. It blocked the adhesion of α1β1-expressing K562 cells to immobilized collagens IV and I with an IC50 of 2 nM and 0.5 nM respectively, but was inactive against other integrin receptors, including the α1β1 structurally closely related collagen receptor α2β1 integrin. Obtustatin (5 μg/disk) inhibited 84% of FGF (fibroblast growth factor) 2-induced angiogenesis in vivo in the chicken [11] and the quail [24] chorioallantoic membrane assay. Moreover, in the Lewis lung [11] and B16F10 melanoma [24] syngenic mouse models, this disintegrin reduced tumour growth by approx. 50% and 65% respectively, when administered intraperitoneally every other day at a dose of 5 mg/kg of body weight. Subsequently, the KTS-disintegrins viperistatin (from V. palestinae venom) [12] and lebestatin (from Macrovipera lebetina transmediterranea venom) [25,26], and the RTS-disintegrin jerdostatin (recombinantly expressed from a Protobothrops jerdonii cDNA library) [27] have been proven to be very potent and highly selective inhibitors of the α1β1 integrin. NMR studies of the solution structure [28] and internal motions [29] of obtustatin have revealed that, in contrast with all known disintegrin structures in which the RGD adhesion motif is located at the apex of an eleven residue hairpin loop, the active K21TS23 tripeptide is oriented towards a side of a nine residue integrin-binding loop. The integrin-binding loop of obtustatin exhibits lateral local motions within a range of approx. 35° and with maximal displacement of approx. 5 Å (1 Å=0.1 nm), and the integrin-binding loop and the C-terminal tail of the disintegrin display concerted motions in the 100–300 ps timescale [29]. The shape and size of its short integrin-binding loop along with its composition, flexibility and the distinct orientation of the KTS tripeptide may underlie the structural basis of the unique selectivity and specificity of obtustatin for integrin α1β1.

Structural diversity among the α1β1-specific disintegrins appears to be segregated within their C-terminal halves (see Figure 3 in [13]). Viperistatin differs from obtustatin by only three amino acids at positions 24, 38 and 40, and is approx. 25-fold more potent than obtustatin in inhibiting the interaction of human α1β1 integrin with collagen IV in vitro [12]. In the present study we report the production and structure–function correlations of recombinant KTS-disintegrins.

EXPERIMENTAL

Antibodies, ECM and cell-adhesion proteins

Polyclonal antiserum against the snake venom disintegrin eristostatin was developed commercially (Chemicon), and the IgG fraction was purified using a Protein A-based purification kit (Pierce Biotechnology). Mouse anti-human β1 integrin subunit monoclonal antibodies B44 and Lia1/2 were purchased from Chemicon and Beckman Coulter respectively. Goat anti-rabbit IgG conjugated with HRP (horseradish peroxidase) and AP (alkaline phosphatase) were purchased from Cell Signaling Biotechnology and Sigma respectively. Bovine collagens type IV and type I were purchased from Chemicon and Chrono-Log respectively. Human plasma fibronectin was from Sigma. Human VCAM-1 (vascular cell adhesion molecule-1) Ig domain was kindly provided by Dr P. Wainreb (Biogen, Cambridge, MA, U.S.A.).

Cell lines

K562 and Jurkat cell lines were purchased from the American Type Culture Collection and cultured in RPMI 1640 and DMEM (Dulbecco's modified Eagle's medium) respectively, supplemented with 10% (v/v) fetal bovine serum (Gibco). K562 cells lines transfected with α1 and α2 integrins were kindly provided by Dr M. Hemler (Dana-Farber Cancer Institute, Boston, MA, U.S.A.) and were cultured in RPMI 1640 containing 500 μg/ml G418 sulfate (Mediatech).

Purification of anti-adhesive proteins from snake venoms

The following snake venom proteins were purified on a C18 column eluted with a linear acetonitrile gradient using the previously described two-step reverse-phase HPLC protocol [12,30]. The disintegrins obtustatin, VLO4 and VLO5 were purified from the venom of V. lebetina obtusa purchased from Latoxan. The disintegrin eristostatin was isolated from Eristocophis macmahoni venom (Latoxan). The disintegrin viperistatin and the C-type lectin-like protein VP12 were purified from the venom of V. palestinae kindly provided by Dr A. Bdolah (Zoology Department, Tel-Aviv University, Tel Aviv, Israel).

Cloning and expression of recombinant obtustatin and its mutants

A synthetic gene for obtustatin was designed based on the 41-amino-acid sequence of obtustatin [23] and optimal codons for Escherichia coli expression, and was commercially synthesized by Genscript. The obtustatin gene was PCR-amplified and ligated into the pGex-4-T1 vector for expression as a thrombin-cleavable GST (gluthatione transferase)-fusion protein using the pUC57 plasmid containing the synthetic gene as a template. Forward (5′-GTTCCGCGTGGATCCTGTACGACGGGTCCTTGTTGTCGT-3′) and reverse (5′-TCGACCCGGGAATTCTTAGCCGGGGTAAAGTGGACAATC-3′) primers (Oligos Etc.) contained restriction sites (underlined) for BamHI and EcoRI respectively. Recombinant obtustatin molecules contained the additional N-terminal dipeptide (Gly-Ser).

Site-directed mutagenesis was performed by PCR using the pGex-4-T1 vector containing obtustatin as a template and the following primers (Oligos Etc.) (underlining in the primers indicates mutated codons: mutant (KAS), forward primer 5′-GCTGGCACGACGTGTTGGAAGGCTTCTCTTACATCC-3′ and reverse primer 5′-CTTCCAACACGCGTGCCAGCTGGCTTAAGTTT-3′; mutant (ATS), forward primer 5′-CCAGCTGGCACGACGTGTTGGGCTACATCTCTTACA-3′ and reverse primer 5′-CCAACACGTCGTGCCAGCTGGCTTAAGTTTGCA-3′; mutant (KTA), forward primer 5′-GGCACGACGTGTTGGAAGACAGCTCTTACATCCCAT-3′ and reverse primer 5′-TGTCTTCCAACACGTCGTGCCAGCTGGCTTAAG-3′; mutant (AAA), forward primer 5′-CCAGCTGCCACGACGTGTTGGGCTGCTGCTCTT ACATCCCAT-3′ and reverse primer 5′-CCAACACGTCGTGCCAGCTGGCTTAAGTTTGCA-3′; mutant (KTSR), forward primer 5′-ACGACGTGTTGGAAGACATCTAGAACATCCCATTAC-3′ and reverse primer 5′-AGATGTCTTCCAACACGTCGTGCCAGCTGGCTT-3′; mutant (KTSA), forward primer 5′-ACGACGTGTTGGAAGACATCTGCTACATCCCATTAC-3′ and reverse primer 5′-AGATGTCTTCCAACACGTCGTGCCAGCTGGCTT-3′. The sequences of cDNAs of obtustatin and its mutants were verified at the Kimmel Cancer Center (Thomas Jefferson University, Philadelphia, PA, U.S.A.) and were used to transform E. coli BL21 (DE3) strain. The GST-fusion protein was induced and isolated as described previously [31]. After digestion with thrombin, recombinant obtustatin and its mutants were purified by reverse-phase HPLC as described previously [12]. The purity of the recombinant proteins was verified by SDS/PAGE, and Western blot analysis using an anti-eristostatin polyclonal antibody, which cross-reacts with native obtustatin, and a HRP-based chemiluminescence detection kit (Cell Signaling).

Cell-adhesion studies

Cell-adhesion studies were performed using an assay developed previously in our laboratory [32]. Briefly, integrin ligands and monoclonal antibody B44 were immobilized on to a 96-well plate in PBS (or in 0.02 M acetic acid for collagens I and IV) overnight at 4 °C. Wells were blocked for 1 h at room temperature (20 °C) with 1% (w/v) BSA in HBSS (Hanks balanced salt solution) buffer containing 3 mM MgCl2. Cells were labelled with 12.5 μM CMFDA (5-chloromethylfluoresceine diacetate) (Invitrogen) in HBSS buffer containing 3 mM MgCl2 for 30 min at 37 °C. Cells were freed from unbound reagent by washing in the same buffer. The labelled cell suspension (1×106 cells/ml) were pre-mixed with increasing concentrations of disintegrins, incubated for 15 min at room temperature and added to the wells of the ligand-coated plate. The plate was incubated for 30 min at 37 °C and unbound cells were removed by washing. Bound cells were lysed by adding 0.5% Triton X-100. In parallel, the standard curve was prepared in the same plate using known concentrations of labelled cells. Plates were read using an FLx800 fluorescence plate reader (BioTek) selecting the bottom reading option, and an excitation wavelength of 485 nm using a 530 nm emission filter.

ELISA

Soluble α1β1 integrin was recombinantly produced in an insect cell-expression system and purified as described [33]. Integrin interaction experiments were carried out using an ELISA-type protocol according to Eble et al. [33]. Briefly, the CB3 [IV] collagen fragment was immobilized on a 96-well plate overnight at 4 °C in PBS at a concentration of 20 μg/ml. The plate was washed three times with TBS (Tris-buffered saline; 20 mM Tris and 150 mM NaCl), pH 7.5 containing 2 mM MnCl2 (TBS/Mg) and blocked with 1% (w/v) BSA in TBS/Mg at room temperature for 1 h. Then, 30 nM of soluble α1β1 integrin, dissolved in the same buffer supplemented with 1 mM MnCl2, was mixed with different concentrations of recombinant disintegrins. After incubation for 15 min at room temperature the mixtures were added to the plate and incubated for 2 h at room temperature. The plate was then washed twice with 50 mM Hepes (pH 7.5) containing 150 mM NaCl2, 2 mM MgCl2 and 1 mM MnCl2, and bound integrin was fixed with 2.5% (v/v) glutardialdehyde (Sigma). The α1β1 integrin was detected using a rabbit anti-β1 antiserum. After washing, goat anti-rabbit IgG conjugated with AP was added. 4-Nitrophenyl phosphate disodium salt hexahydrate (Sigma) was added as a substrate for AP and colour was developed at room temperature. The plate was read using an ELISA plate reader at 405 nm.

Cell-free disintegrin-binding assay for LIBS (ligand-induced binding site) expression

Disintegrins were labelled with FITC as described previously [32]. Anti-LIBS monoclonal antibody B44 (0.5 μg/ml) or the blocking conformational-insensitive anti-β1 integrin subunit monoclonal antibody Lia1/2 (0.5 μg/ml) were immobilized on to a 96-well plate. Wells were blocked with 1% (w/v) BSA as described above (Cell-adhesion studies). α1K562 cells (α1-transfected K562) were harvested and washed with HBSS containing 3 mM Mg2+. The cell pellet was lysed with 0.5% Triton X-100 in HBSS containing 3 mM Mg2+ and the protein concentration in the lysate was estimated using the BCA (bicinchoninic acid) assay kit (Pierce). FITC–disintegrin was mixed with cell lysate (2 mg of protein/ml) to a total sample volume of 100 μl. Samples were incubated for 30 min at 37 °C and transferred to a 96-well plate with immobilized B44 antibody. To measure non-specific binding of FITC–disintegrin, the same experiment was carried out using non-α1-transfected K562 cells. After incubation for 30 min at 37 °C, the plates were washed three times with HBSS containing 3 mM Mg2+ and 1% (w/v) BSA. Finally, plates were read (using the bottom reading option) using a fluorescent plate reader FLx800 (BioTek) at an excitation wavelength of 485 nm and a 530 nm emission filter. In parallel, a standard curve of bound disintegrin was prepared on the same plate using known concentrations of FITC–disintegrin.

Molecular modelling

The structure of obtustatin mutants were calculated using NOE (nuclear Overhauser effect) information and disulfide bond restraints of the NMR solution structure of obtustatin (PDB code 1MPZ) and an energy minimization protocol supplied with the program CNS [34]. For the final ensembles, a set of 50 structures was calculated and the converged structures selected corresponded to a plateau in the values of the overall calculated energy. For the analysis of the mutants the lowest energy structure was used as a representative model, and the mutations were introduced using the protocol supplied with the program PyMOL (DeLano Scientific; http://pymol.sourceforge.net/).

RESULTS

Antagonistic profile of native and recombinant obtustatin

Obtustatin recombinantly produced in E. coli exhibited the same chromatographic behaviour on reverse-phase HPLC as obtustatin purified from viper venom and showed the same inhibitory saturation pattern and IC50 on α1K562 cell adhesion to immobilized collagen IV as the native (venom-isolated) disintegrin (Figure 1A and Table 1). The selectivity of recombinant and native obtustatins was also indistinguishable from one another (Figure 1B). In particular, neither native nor recombinant obtustatin blocked the adhesive functions of the α2β1 integrin that is structurally and functionally highly related to α1β1 integrin. The recombinant production of fully active obtustatin allowed us to assess the previously suggested relevance of synthetic KTS-bearing peptides for specific α1β1 integrin inhibition [11,12,23] employing the site-directed mutated molecules displayed in Figure 2. Specifically, substitutions within the KTS tripeptide motif showed significant impairment of the α1β1 integrin inhibitory ability of the recombinant proteins (Figure 3). Using a cell-adhesion assay, the IC50 values of the single alanine mutants A21TS23, KAS and KTA had a 25-fold 82-fold and 13-fold respectively, decrease inhibiting the α1K562 cell adhesion to immobilized collagen IV (Figure 3A and Table 1). The triple mutant A21AA23 was devoid of inhibitory activity up to a concentration of 10 μM (Figure 3A and Table 1). Similar features were observed when the single and triple obtustatin mutants were tested for their ability to inhibit the binding of a recombinant soluble form of the α1β1 integrin to the CB3 fragment of collagen IV in a cell-free ELISA-type assay (Figure 3C).

Comparison of the effect of recombinant and native obtustatin on integrin-mediated cell adhesion to different ECM proteins

Figure 1
Comparison of the effect of recombinant and native obtustatin on integrin-mediated cell adhesion to different ECM proteins

ECM proteins (2 μg/ml) were immobilized overnight at 4 °C on a 96-well plate in 0.02 M acetic acid for collagen and in PBS for VCAM-1 and fibronectin. After blocking the plate with BSA, CMFDA-labelled cells pre-incubated in the presence or absence of snake venom proteins were added to the wells, and incubated for 30 min at 37 °C. Non-binding cells were washed out, adhered cells were lysed, and the plate was read using a fluorescence plate reader. (A) Effect of different concentrations of recombinant and native obtustatin on the inhibition of adhesion of α1K562 cells to collagen IV. (B) Effect of recombinant (a) and native (b) obtustatin, VP12 (c), VLO5 (d) and VLO4 (e) on adhesion of the indicated cell lines to their respective immobilized ligands. Venom proteins were applied at a concentration of 20 μg/ml. VP12 belongs to subgroup of snake venom C-type lectin-like proteins that specifically target the α2β1 integrin [50]. VLO5 and VLO4 are dimeric disintegrins from V. lebetira obtusa which antagonize the function of integrins α4β1 and α4β1 respectively [30,51]. The observed small inhibition activity of native obtustatin on α4β1 integrin binding to VCAM-1 was due to a slight contamination with VLO5, which has an almost identical reverse-phase retention time as obtustatin. Error bars represent the S.D. from three independent experiments.

Figure 1
Comparison of the effect of recombinant and native obtustatin on integrin-mediated cell adhesion to different ECM proteins

ECM proteins (2 μg/ml) were immobilized overnight at 4 °C on a 96-well plate in 0.02 M acetic acid for collagen and in PBS for VCAM-1 and fibronectin. After blocking the plate with BSA, CMFDA-labelled cells pre-incubated in the presence or absence of snake venom proteins were added to the wells, and incubated for 30 min at 37 °C. Non-binding cells were washed out, adhered cells were lysed, and the plate was read using a fluorescence plate reader. (A) Effect of different concentrations of recombinant and native obtustatin on the inhibition of adhesion of α1K562 cells to collagen IV. (B) Effect of recombinant (a) and native (b) obtustatin, VP12 (c), VLO5 (d) and VLO4 (e) on adhesion of the indicated cell lines to their respective immobilized ligands. Venom proteins were applied at a concentration of 20 μg/ml. VP12 belongs to subgroup of snake venom C-type lectin-like proteins that specifically target the α2β1 integrin [50]. VLO5 and VLO4 are dimeric disintegrins from V. lebetira obtusa which antagonize the function of integrins α4β1 and α4β1 respectively [30,51]. The observed small inhibition activity of native obtustatin on α4β1 integrin binding to VCAM-1 was due to a slight contamination with VLO5, which has an almost identical reverse-phase retention time as obtustatin. Error bars represent the S.D. from three independent experiments.

Table 1
Inhibition of α1K562 cell adhesion to immobilized collagen IV by wild-type obtustatin and viperistatin and the different obtustatin mutants
Type of mutation IC50 (nM)* 
Wild-type recombinant/native obtustatin 2.00 
Wild-type native viperistatin 0.08 
ATS-obtustatin 51.30 
KAS-obtustatin 164.18 
KTA-obtustatin 26.08 
AAA-obtustatin >10000 
KTSR-obtustatin 0.33 
KTSA-obtustatin 2.56 
Type of mutation IC50 (nM)* 
Wild-type recombinant/native obtustatin 2.00 
Wild-type native viperistatin 0.08 
ATS-obtustatin 51.30 
KAS-obtustatin 164.18 
KTA-obtustatin 26.08 
AAA-obtustatin >10000 
KTSR-obtustatin 0.33 
KTSA-obtustatin 2.56 
*

Data represent the mean from three independent experiments.

Amino acid sequences of obtustatin and its mutants recombinantly expressed in E. coli

Figure 2
Amino acid sequences of obtustatin and its mutants recombinantly expressed in E. coli

Mutated residues are shown in bold. (A) Wild-type obtustatin; (BE) mutants affecting the KTS motif; and (FH) mutations for converting obtustatin into viperistatin.

Figure 2
Amino acid sequences of obtustatin and its mutants recombinantly expressed in E. coli

Mutated residues are shown in bold. (A) Wild-type obtustatin; (BE) mutants affecting the KTS motif; and (FH) mutations for converting obtustatin into viperistatin.

Effect of recombinant obtustatin and its mutants on the interaction of α1β1 integrin with collagen IV

Figure 3
Effect of recombinant obtustatin and its mutants on the interaction of α1β1 integrin with collagen IV

(A) Effect of wild-type obtustatin and its KTS motif mutants on the adhesion of α1K562 cells to immobilized collagen IV. (B) Effect of wild-type viperistatin and obtustatin and the KTSA and KTSR mutants on the adhesion of α1K562 cells to immobilized collagen IV. (C) Effect of wild-type obtustatin and its mutants on the binding of soluble α1β1 integrin to the immobilized CB3 fragment of collagen IV using the ELISA-type assay described in the Experimental section. Error bars represent the S.D. from three experiments. wt, wild-type.

Figure 3
Effect of recombinant obtustatin and its mutants on the interaction of α1β1 integrin with collagen IV

(A) Effect of wild-type obtustatin and its KTS motif mutants on the adhesion of α1K562 cells to immobilized collagen IV. (B) Effect of wild-type viperistatin and obtustatin and the KTSA and KTSR mutants on the adhesion of α1K562 cells to immobilized collagen IV. (C) Effect of wild-type obtustatin and its mutants on the binding of soluble α1β1 integrin to the immobilized CB3 fragment of collagen IV using the ELISA-type assay described in the Experimental section. Error bars represent the S.D. from three experiments. wt, wild-type.

Next, we checked the hypothesis that the residue immediately C-terminal to the KTS motif may be partly responsible for the increased inhibitory potential of viperistatin over obtustatin [12]. To this end, Leu24 (obtustatin) was replaced with alanine and with arginine (as in viperistatin). The K21TSA24 mutant (Figure 2G) showed only a slight decrease in IC50 in both ELISA-type and cell-adhesion assays. On the other hand, the L24R mutant (Figure 2F) exhibited a 6-fold enhancement in its α1β1 inhibitory potential (Table 1, and Figures 3B and 3C); however, in the adhesion, but not in the ELISA-type, assay this latter effect was concentration-dependent since the difference in activity of the KTSR mutant disappeared at disintegrin concentrations higher than 10 nM (compare Figures 3B and 3C).

KTS-disintegrins induced an LIBS epitope upon binding to α1β1 integrin

Ligand engagement to integrins may trigger conformational changes leading to the exposure of cryptic epitopes (LIBS) recognized by monoclonal antibodies [36]. The ability of KTS-disintegrins to induce LIBS was assessed using two different assay systems. Using a solid-phase adhesion assay, we showed that both obtustatin and viperistatin significantly induced conformational alterations in α1β1 integrin reported by expression of a neoepitope on the β1 subunit recognized by the B44 monoclonal antibody (Figure 4A); however, the ability to induce LIBS was significantly lower for viperistatin than for obtustatin. On the other hand, wild-type obtustatin and its KTSR mutant had similar effects on monoclonal antibody B44 epitope exposure, suggesting that the distinct LIBS-inducing activity of obtustatin and viperistatin may not reside in their integrin-binding loops. On the other hand, using a cell-free binding assay, in which FITC-labelled disintegrins complexed with α1β1 integrin in the cell lysate were detected following binding to the immobilized B44 antibody, obtustatin, viperistatin and the mutated recombinant KTSR-obtustatin showed similar abilities inducing LIBS epitope on α1β1 integrin (Figure 4B) as in the adhesion assay (Figure 4A). Binding of FITC–disintegrins to the anti-β1 blocking antibody Lia1/2, used as a control, was negligible and was substracted as background from each reading. The outcome of this experimental setup ruled out the potential controversy concerning the possibility that internalization of the ligand–integrin complex influenced the results gathered through the adhesion assay.

Effect of wild-type viperistatin and obtustatin and the KTSR-obtustatin mutant on LIBS expression

Figure 4
Effect of wild-type viperistatin and obtustatin and the KTSR-obtustatin mutant on LIBS expression

(A) Experiments were performed using cell-adhesion assays. To this end, monoclonal antibody B44 recognizing a LIBS epitope on the β1 subunit was immobilized on a 96-well plate (0.5 μg/ml) by overnight incubation at 4 °C in PBS. After blocking the plate with BSA, CMFDA-labelled α1K562 cells pre-incubated in the presence (1 mM in HBSS containing 3 mM Mg2+) or the absence of disintegrins were added and incubated at 37 °C for 30 min. Unbound cells were removed by washing, adhered cells were lysed with 0.5% Triton X-100, and the fluorescence was read using a fluorescence plate reader. The error bars represent the S.D. from three experiments. *P<0.05 between control and samples incubated with disintegrins, as determined by an unpaired Student's t test. **P<0.05 between samples incubated with wild-type or mutated obtustatin molecules and with viperistatin. (B) Experiments were performed using a cell-free binding assay. α1K562 cells were lysed with HBSS containing 3 mM Mg2+ and 0.5% Triton X-100. FITC–disintegrins were mixed with 100 μl of cell lysate (2 mg/ml) for 30 min at 37 °C followed by incubation for another 30 min at 37 °C with monoclonal antibody B44 immobilized on to the wells of a 96-well plate. The same experiment was carried out using the lysate of non-transfected K562 cells as a control of the specific binding of the FITC–disintegrin to integrin α1β1 (♦). After washing with HBSS containing 3 mM Mg2+, fluorescence was read using a fluorescence plate reader. Error bars represent the S.D. from three independent experiments. wt, wild-type.

Figure 4
Effect of wild-type viperistatin and obtustatin and the KTSR-obtustatin mutant on LIBS expression

(A) Experiments were performed using cell-adhesion assays. To this end, monoclonal antibody B44 recognizing a LIBS epitope on the β1 subunit was immobilized on a 96-well plate (0.5 μg/ml) by overnight incubation at 4 °C in PBS. After blocking the plate with BSA, CMFDA-labelled α1K562 cells pre-incubated in the presence (1 mM in HBSS containing 3 mM Mg2+) or the absence of disintegrins were added and incubated at 37 °C for 30 min. Unbound cells were removed by washing, adhered cells were lysed with 0.5% Triton X-100, and the fluorescence was read using a fluorescence plate reader. The error bars represent the S.D. from three experiments. *P<0.05 between control and samples incubated with disintegrins, as determined by an unpaired Student's t test. **P<0.05 between samples incubated with wild-type or mutated obtustatin molecules and with viperistatin. (B) Experiments were performed using a cell-free binding assay. α1K562 cells were lysed with HBSS containing 3 mM Mg2+ and 0.5% Triton X-100. FITC–disintegrins were mixed with 100 μl of cell lysate (2 mg/ml) for 30 min at 37 °C followed by incubation for another 30 min at 37 °C with monoclonal antibody B44 immobilized on to the wells of a 96-well plate. The same experiment was carried out using the lysate of non-transfected K562 cells as a control of the specific binding of the FITC–disintegrin to integrin α1β1 (♦). After washing with HBSS containing 3 mM Mg2+, fluorescence was read using a fluorescence plate reader. Error bars represent the S.D. from three independent experiments. wt, wild-type.

DISCUSSION

The α1β1 integrin is widely considered to represent a target for anti-angiogenic, and thus anticancer, therapeutic intervention. Delineating structure–function correlations of specific α1β1 inhibitors is therefore of major pharmaceutical relevance. Our previous studies using linear synthetic peptides indicated a role for the KTS and RTS tripeptide motifs of obtustatin, viperistatin and jerdostatin in the selective inhibition of integrin α1β1 binding to collagens I and IV in vitro [11,12,26] and angiogenesis in vivo [11]. Although not so widely spread among integrin ligands as the RGD tripeptide, the KTS motif is present in human and mouse ADAM-9 (a disintegrin and metalloproteinase-9) disintegrin-like domain at a topologically equivalent position in snake venom disintegrins [37] (UniProtKB/Swiss-Prot entry Q13443). This motif is also localized in the C-terminal part of VP3, a structural protein of human polyomavirus JC, which after activation induces a multifocal leukoencephalopathy in the brain [38]. These findings strongly suggest that the KTS motif may play an important role in certain human pathophysiological conditions. To establish structure–function correlations of KTS-bearing proteins, we have set up a bacterial system for the recombinant expression of fully active obtustatin. A major conclusion from comparison of the α1β1 inhibitory assays using point mutations which gradually converted obtustatin into viperistatin is that substitution of leucine for arginine at position 24, immediately C-terminal to the KTS motif, did not compromise the α1β1 selectivity but accounted for only 40% of the 25-fold enhanced activity of viperistatin over obtustatin. A corollary of this conclusion is that the amino acid adjacent to the KTS motif confers inhibitory potency rather than receptor specificity. Furthermore, the partial enhancement of the inhibitory activity of the KTSR-obtustatin mutant indicated that the other two different residues between viperistatin and obtustatin (L38V and P40Q; Figure 2) may contribute to the distinct inhibitory features of KTS-disintegrins. To provide a structural ground for the functional role of the KTSX motif residues, and to investigate structural determinants of the enhanced inhibitory activity of viperistatin over obtustatin, we generated molecular models of obtustatin and viperistatin (Figure 5). Among the KTS motif residues, mutation of the central Thr22 for alanine had the major functional impact (Table 1). This effect can be rationalized due to the critical role of this residue in maintaining the conformation of the integrin-binding loop [23,28]. A network of interactions between the solvent-shielded Hγ2 of Thr22 and residues Thr25 and Ser26 explains the restrained mobility of Thr22 [29]. On the other hand, the 6-fold gain-of-activity conferred by the L24R mutation indicates that the positively charged extended side chain of arginine is better suited than the solvent-exposed alkyl side chain of leucine for establishing productive interactions with residues within the ligand-binding site of the α1β1 integrin. The size and shape of the integrin-binding loop has been proven to be a determinant for conferring integrin specificity to cyclic RGD peptides with a known conformation [39] and to the short-sized RGD-disintegrins echistatin and eristostatin [40]. Thus the critical importance of Thr22 may be linked to a structure-stabilizing mechanism of the integrin-binding loop, which supports the correct conformational presentation of the KTS sequence to be recognized by α1β1 integrin, rather than to a direct role in receptor binding.

Molecular models of obtustatin and viperistatin

Figure 5
Molecular models of obtustatin and viperistatin

(A) NMR solution structure of obtustatin (PDB code 1MPZ) showing the locations of residues (rendered in stick representation) which were sequentially mutated in the different recombinant proteins shown in Figure 2 to convert recombinant obtustatin into viperistatin. (B) Computer modelling of the transformation of obtustatin into viperistatin. Figures were created using PyMol.

Figure 5
Molecular models of obtustatin and viperistatin

(A) NMR solution structure of obtustatin (PDB code 1MPZ) showing the locations of residues (rendered in stick representation) which were sequentially mutated in the different recombinant proteins shown in Figure 2 to convert recombinant obtustatin into viperistatin. (B) Computer modelling of the transformation of obtustatin into viperistatin. Figures were created using PyMol.

Although a tripeptide at the tip of a mobile loop represents the primary integrin inhibitory determinant of RGD-disintegrins, both the potency and receptor selectivity of disintegrins appears to be finely tuned by residues flanking the active sequence [16,40,41]. Similarly, the presence of an arginine residue at the C-terminal side of the primary KTS-integrin-recognition motif of viperistatin enhances its inhibitory activity towards α1β1 integrin by approx. 4-fold in comparison with obtustatin, which contains a leucine residue in this position; however, this non-conservative mutation does not affect the integrin specificity of KTS-disintegrins, whereas structural evidence gathered on synthetic peptides and RGD-disintegrins showed that the amino acid immediately C-terminal to the RGD sequence influence the integrin-recognition selectivity of the disintegrin through modulation of the RGD loop conformation [39].

It is also worth noting that substitution of leucine for arginine at position 24 (sequence F in Figure 2) was not enough for switching the inhibitory activity of obtustatin to that of viperistatin (Table 1, and Figures 3B and 3C). The partial (40%) gain-of-function of recombinant KTSR-obtustatin is in the range of inhibitory activity of lebestatin (IC50 of 0.4 nM), another α1β1-specific KTSR-disintegrin isolated from the venom of M. lebetina transmediterranea [25,26]. Besides position 24, obtustatin and lebestatin sequences differ only at position 38, which contains leucine and serine residues respectively. In line with structural studies on obtustatin showing that its integrin-binding loop and C-terminal tail displays concerted motions [29], our results suggest that these structural elements act as a secondary determinant of integrin-binding specificity/potency. Relevant to this point, obtustatin and viperistatin have different amino acids at positions 38 and 40, which are L38V and P40Q respectively (Figure 2). We argue that these two substitutions, along with the active site L24R mutation, may create a distinct chemical environment responsible for the increased inhibitory activity towards α1β1 integrin of viperistatin over obtustatin. Furthermore, in the present study we show that obtustatin, and to a lesser extent viperistatin, elicit the exposure of LIBS in α1β1 recognized by monoclonal antibody B44 (Figure 4). Our working hypothesis is that a proline residue at position 40 may contribute to the better exposure of conformational changes in the α1β1–obtustatin complex.

Concluding remarks and significance

The α1β1 integrin is an attractive target for the development of novel therapeutic strategies against pathological angiogenesis in cancer and certain inflammatory diseases. The KTSX sequence found in viper venom disintegrins represents a specific motif targeting the α1β1 integrin. The NMR solution structure of the KTS-disintegrin obtustatin along with the structure–function correlations reported in the present study may provide the ground for designing a new generation of synthetic lead compounds with potential application for anti-angiogenic intervention.

A further novel finding of the present study is the demonstration that ligand engagement elicits in α1β1 the exposure of cryptic epitopes recognized by monoclonal antibody B44. Ligand-induced binding sites had been reported in β3 and α4β1 integrins. Our results suggest that conformational regulation may represent a general feature of integrins for transducing ligand-specific signals. LIBS offer further possibilities for developing immunochemical weapons to fight undesired up-regulation of the α1β1 integrin in pathological conditions.

Finally, although the present study focused on the structural determinants of snake venom disintegrins for specific inhibition of α1β1 integrin, the occurrence of the KTS tripeptide in certain proteins, such as human ADAM-9 and VP3 capsid protein of the human polyomavirus JC, suggests a hitherto unrecognized role for this motif in physiology and pathophysiology. Functions of ADAM family proteins are not fully characterized. Interactions with collagen receptors may regulate their adhesive and/or enzymatic functions. On the other hand, the presence of the KTS sequence on a capsid protein of a virus attacking the human nervous system suggests that the α1β1 integrin may serve as part of the host-cell entry strategy of the pathogen. Further investigations on novel KTS-bearing proteins are needed for expanding our understanding of the basic principles underlying its functionality in diverse biological contexts.

Abbreviations

     
  • ADAM

    a disintegrin and metalloproteinase

  •  
  • AP

    alkaline phosphatase

  •  
  • CMFDA

    5-chloromethylfluoresceine diacetate

  •  
  • ECM

    extracellular matrix

  •  
  • GST

    gluthatione transferase

  •  
  • HBSS

    Hanks balanced salt solution

  •  
  • HRP

    horseradish peroxidase

  •  
  • LIBS

    ligand-induced binding site

  •  
  • TBS

    Tris-buffered saline

  •  
  • VCAM-1

    vascular cell adhesion molecule-1

FUNDING

This work was supported in part by the National Institutes of Health [grant number CA100145] (to C. M.); the American Heart Association [grant number 0230163N] (to C. M.); the German Research Council [grant number EB177/4-2] (to J. A. E.); and the Ministerio de Ciencia e Innovación, Madrid, Spain [grant number BFU2007-61563] (to J. J. C.).

References

References
1
Calvete
J. J.
Platelet integrin GPIIb/IIIa: structure–function correlations. An update and lessons from other integrins
Proc. Soc. Exp. Biol. Med.
1999
, vol. 
222
 (pg. 
29
-
38
)
2
Hynes
R. O.
Integrins: bidirectional, allosteric signaling machines
Cell
2002
, vol. 
110
 (pg. 
673
-
687
)
3
Takagi
J.
Springer
T. A.
Integrin activation and structural rearrangement
Immunol. Rev.
2002
, vol. 
186
 (pg. 
141
-
163
)
4
Berrier
A. L.
Yamada
K. M.
Cell-matrix adhesion
J. Cell. Physiol.
2007
, vol. 
213
 (pg. 
565
-
573
)
5
Luo
B. H.
Carman
C. V.
Springer
T. A.
Structural basis of integrin regulation and signaling
Annu. Rev. Immunol.
2007
, vol. 
25
 (pg. 
619
-
647
)
6
Eble
J. A.
Collagen-binding integrins as pharmaceutical targets
Curr. Pharm. Des.
2005
, vol. 
11
 (pg. 
867
-
880
)
7
McCall
K. D.
Zutter
M. M.
Collagen receptor integrins: rising to the challenge
Curr. Drug Targets
2008
, vol. 
9
 (pg. 
139
-
149
)
8
Hughes
A. L.
Evolution of the integrin α and β families
J. Mol. Evol.
2001
, vol. 
52
 (pg. 
63
-
72
)
9
Leitinger
B.
Hogg
N.
Integrin I domains and their function
Biochem. Soc. Trans.
1999
, vol. 
27
 (pg. 
826
-
832
)
10
Dickeson
S. K.
Mathis
N. L.
Rahman
M.
Bergelson
J. M.
Santoro
S. A.
Determinants of ligand binding specificity of the α1β1 and α2β1 integrins
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
32182
-
32191
)
11
Marcinkiewicz
C.
Wainreb
P. H.
Calvete
J. J.
Kisiel
D. G.
Mousa
S. A.
Tuszynski
G. P.
Lobb
R. R.
Obtustatin: a potent selective inhibitor of α1β1 integrin in vitro and angiogenesis in vivo
Cancer Res.
2003
, vol. 
63
 (pg. 
2020
-
2023
)
12
Kisiel
D. G.
Calvete
J. J.
Katzhendler
J.
Fertala
A.
Lazarovici
P.
Marcinkiewicz
C.
Structural determinants of the selectivity of KTS-disintegrins for the α1β1 integrin
FEBS Lett.
2004
, vol. 
577
 (pg. 
478
-
482
)
13
Calvete
J. J.
Marcinkiewicz
C.
Sanz
L.
KTS and RTS-disintegrins: anti-angiogenic viper venom peptides specifically targeting the α1β1 integrin
Curr. Pharm. Des.
2007
, vol. 
13
 (pg. 
2853
-
2859
)
14
Sanz
L.
Bazaa
A.
Marrakchi
N.
Pérez
A.
Chenik
M.
Bel Lasfer
Z.
El Ayeb
M.
Calvete
J. J.
Molecular cloning of disintegrins from Cerastes vipera and Macrovipera lebetina transmediterranea venom gland cDNA libraries: insight into the evolution of the snake venom integrin-inhibition system
Biochem. J.
2006
, vol. 
395
 (pg. 
385
-
392
)
15
Ogawa
T.
Chijiwa
T.
Oda-Ueda
N.
Ohno
M.
Molecular diversity and accelerated evolution of C-type lectin-like proteins from snake venom
Toxicon
2005
, vol. 
45
 (pg. 
1
-
14
)
16
Niewiarowski
S.
McLane, M. A.
Kloczewiak M.
Stewart
G. J.
Disintegrins and other naturally occurring antagonists of platelet fibrinogen receptors
Semin. Hematol.
1994
, vol. 
31
 (pg. 
289
-
300
)
17
Marcinkiewicz
C.
Functional characteristic of snake venom disintegrins: potential therapeutic implication
Curr. Pharm. Des.
2005
, vol. 
11
 (pg. 
815
-
827
)
18
Calvete
J. J.
Marcinkiewicz
C.
Monleon
D.
Esteve
V.
Celda
B.
Juarez
P.
Sanz
L.
Snake venom disintegrins: evolution of structure and function
Toxicon
2005
, vol. 
45
 (pg. 
1063
-
1074
)
19
Koh
D. C.
Armugam
A.
Jeyaseelan
K.
Snake venom components and their applications in biomedicine
Cell. Mol. Life Sci.
2006
, vol. 
63
 (pg. 
3030
-
3041
)
20
Scarborough
R. M.
Rose
J. W.
Hsu
M. A.
Phillips
D. R.
Fried
V. A.
Campbell
A. M.
Nannizzi
L.
Charo
I. F.
Barbourin. A GPIIb-IIIa-specific integrin antagonist from the venom of Sistrurus m. barbouri
J. Biol. Chem.
1991
, vol. 
266
 (pg. 
9359
-
9362
)
21
Marcinkiewicz
C.
Calvete
J. J.
Senadhi
V. K.
Marcinkiewicz
M. M.
Raida
M.
Schick
P.
Lobb
R. R.
Niewiarowski
S.
Structural and functional characterization of EMF10, a heterodimeric disintegrin from Eristocophis macmahoni venom that selectively inhibits α5β1 integrin
Biochemistry
1999
, vol. 
38
 (pg. 
13302
-
13309
)
22
Calvete
J. J.
Fox
J. W.
Agelan
A.
Niewiarowski
S.
Marcinkiewicz
C.
The presence of the WGD motif in CC8 heterodimeric disintegrin increases its inhibitory effect on αIIbβ3, αvβ3, and α5β1 integrins
Biochemistry
2002
, vol. 
41
 (pg. 
2014
-
2021
)
23
Moreno-Murciano
M. P.
Monleón
D. D.
Calvete
J. J.
Celda
B.
Marcinkiewicz
C.
Amino acid sequence and homology modeling of obtustatin, a novel non-RGD-containing short disintegrin isolated from the venom of Vipera lebetina obtusa
Protein Sci.
2003
, vol. 
12
 (pg. 
366
-
371
)
24
Brown, M. C.
Staniszewska I.
Del Valle, L.
Tuszynski , G. P.
Marcinkiewicz
C.
Angiostatic activity of obtustatin as α1β1 integrin inhibitor in experimental melanoma growth
Int. J. Cancer
2008
, vol. 
123
 (pg. 
2195
-
2203
)
25
Bazaa
A.
Marrakchi
N.
El Ayeb
M.
Sanz
L.
Calvete
J. J.
Snake venomics: comparative analysis of the venom proteomes of the Tunisian snakes Cerastes cerastes, Cerastes vipera and Macrovipera lebetina
Proteomics
2005
, vol. 
5
 (pg. 
4223
-
4235
)
26
Olfa
K. Z.
José
L.
Salma
D.
Amine
B.
Najet
S. A.
Nicolas
A.
Maxime
L.
Raoudha
Z.
Kamel
M.
Jacques
M.
, et al. 
Lebestatin, a disintegrin from Macrovipera venom, inhibits integrin-mediated cell adhesion, migration and angiogenesis
Lab. Invest.
2005
, vol. 
85
 (pg. 
1507
-
1516
)
27
Sanz
L.
Chen
R. Q.
Perez
A.
Hilario
R.
Juarez
P.
Marcinkiewicz
C.
Monleon
D.
Celda
B.
Xiong
Y. L.
Pérez-Payá
E.
Calvete
J. J.
cDNA cloning and functional expression of jerdostatin, a novel RTS-disintegrin from Trimeresurus jerdonii and a specific antagonist of the α1β1 integrin
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
40714
-
40722
)
28
Moreno-Murciano
M. P.
Monleón
D.
Marcinkiewicz
C.
Calvete
J. J.
Celda
B.
NMR solution structure of the non-RGD disintegrin obtustatin
J. Mol. Biol.
2003
, vol. 
329
 (pg. 
135
-
145
)
29
Monleón
D.
Moreno-Murciano
M. P.
Kovacs
H.
Marcinkiewicz
C.
Calvete
J. J.
Celda
B.
Concerted motions of the integrin-binding loop and the C-terminal tail of the non-RGD disintegrin obtustatin
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
45570
-
45576
)
30
Bazan-Socha
S.
Kisiel
D. G.
Young
B.
Theakston
R. D.
Calvete
J. J.
Sheppard
D.
Marcinkiewicz
C.
Structural requirements of MLD-containing disintegrins for functional interaction with α4β1 and α9β1 integrins
Biochemistry
2004
, vol. 
43
 (pg. 
1639
-
1647
)
31
Smith
D. B.
Johnson
K. S.
Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase
Gene
1988
, vol. 
67
 (pg. 
31
-
40
)
32
Marcinkiewicz
C.
Vijay-Kumar
S.
McLane
M. A.
Niewiarowski
S.
Significance of RGD loop and C-terminal domain of echistatin for recognition of αIIbβ3 and αvβ3 integrins and expression of ligand-induced binding site
Blood
1997
, vol. 
90
 (pg. 
1565
-
1575
)
33
Eble
J. A.
Kassner
A.
Niland
S.
Morgelin
M.
Grifka
J.
Grassel
S.
Collagen XVI harbors an integrin α1β1 recognition site in its C-terminal domains
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
25745
-
25756
)
34
Brunger
A. T.
Adams
P. D.
Clore
G. M.
DeLano
W. L.
Gros
P.
Grosse-Kunstleve
R. W.
Jiang
J. S.
Kuszewski
J.
Nilges
M.
Pannu
N. S.
, et al. 
Crystallography and NMR system: a new software suite for macromolecular structure determination
Acta Crystallogr. D Biol. Crystallogr.
1998
, vol. 
54
 (pg. 
905
-
921
)
35
Reference deleted
36
Humphries
M. J.
Monoclonal antibodies as probes of integrin priming and activation
Biochem. Soc. Trans.
2004
, vol. 
32
 (pg. 
407
-
411
)
37
Weskamp
G.
Krätzschmar
J.
Reid
M. S.
Blobel
C. P.
MDC9, a widely expressed cellular disintegrin containing cytoplasmic SH3 ligand domains
J. Cell Biol.
1996
, vol. 
132
 (pg. 
717
-
726
)
38
Trowbridge
P. W.
Frisque
R. J.
Identification of three new JC virus proteins generated by alternative splicing of the early viral mRNA
J. Neurovirol.
1995
, vol. 
1
 (pg. 
195
-
206
)
39
Pfaff
M.
Tangemann
K.
Müller
B.
Gurrath
M.
Müller
G.
Kessler
H.
Timpl
R.
Engel
J.
Selective recognition of cyclic RGD peptides of NMR defined conformation by αIIbβ3, αvβ3 and α5β1 integrins
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
20233
-
20238
)
40
McLane
M. A.
Vijay-Kumar
S.
Marcinkiewicz
C.
Calvete
J. J.
Niewiarowski
S.
Importance of the structure of the RGD-containing loop in the disintegrins echistatin and eristostatin for recognition of αIIbβ3 and αvβ3 integrins
FEBS Lett.
1996
, vol. 
391
 (pg. 
139
-
143
)
41
Lu
X.
Rahman
S.
Kakkar
V. V.
Authi
K. S.
Substitutions of proline 42 to alanine and methionine 46 to asparagine around the RGD domain of the neurotoxin dendroaspin alter its preferential antagonism to that resembling the disintegrin elegantin
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
289
-
294
)