SEARCH

SEARCH BY CITATION

Keywords:

  • dendroaspin;
  • dynamics;
  • disintegrin;
  • disulfide bond;
  • folding;
  • rhodostomin;
  • Pichia pastoris

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Dendroaspin (Den) and rhodostomin (Rho) are snake venom proteins containing a PRGDMP motif. Although Den and Rho have different 3D structures, they are highly potent integrin inhibitors. To study their structure, function, and dynamics relationships, we expressed Den and Rho in Pichia pastoris. The recombinant Den and Rho inhibited platelet aggregation with the KI values of 149.8 and 83.2 nM. Cell adhesion analysis showed that Den was 3.7 times less active than Rho when inhibiting the integrin αIIbβ3 and 2.5 times less active when inhibiting the integrin αvβ3. In contrast, Den and Rho were similarly active when inhibiting the integrin α5β1 with the IC50 values of 239.8 and 256.8 nM. NMR analysis showed that recombinant Den and Rho have different 3D conformations for their arginyl-glycyl-aspartic acid (RGD) motif. However, the comparison with Rho showed that the docking of Den into integrin αvβ3 resulted in a similar number of contacts. Analysis of the dynamic properties of the RGD loop in Den and Rho showed that they also had different dynamic properties. These results demonstrate that protein scaffolds affect the function, structure, and dynamics of their RGD motif.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Venom toxins are small proteins with many disulfide bonds. They have multiple functions directed toward a variety of molecular targets, including a diversity of receptors, enzymes, and ion channels.1, 2 It is known that toxins with a similar fold can have different functions. In contrast, toxins with unrelated folds can also have similar functions.3–5 Many toxins are obtained from a variety of species, such as snakes, ticks, and leeches, and contain an arginyl-glycyl-aspartic acid (RGD) motif with different protein scaffolds.6–11 They specifically inhibit the integrin-binding function and are potent integrin antagonists. Many of these proteins have potential as therapeutic agents that can directly target integrins. Structural and functional studies of RGD-containing toxins suggest that the inhibitory potency of these proteins lies in subtle positional requirements of the RGD motif at the apex of a flexible loop, a structural feature for binding to integrins.12 However, little is known about the effect of protein scaffolds on the structure–function–dynamics relationships of RGD-containing proteins. Therefore, we selected two protein scaffolds with a PRGDMP motif to study the effect.

Dendroaspin (Den) is a snake venom protein isolated from Dendroaspis jamesoni kaimosae and is a short-chain neurotoxin homolog from the venom of Elapidae snakes with weak neurotoxicity.13 Den consists of 59 amino acids that include eight cysteine residues and a PRGDMP sequence at positions 42–47.14 Unlike neurotoxins, it contains an RGD motif and is a potent inhibitor of platelet aggregation. Den inhibited adenosine diphosphate (ADP)-stimulated platelets in a saturable manner with the inhibition constant of 172 nM.13 Rhodostomin (Rho) is obtained from Calloselasma rhodostoma venom and belongs to the family of disintegrins.15 It consists of 68 amino acids, which includes 12 residues of cysteine and a PRGDMP sequence at positions 48–53. We previously showed16 that Rho expressed in Pichia pastoris has the same function and structure as native protein. Rho binds to unstimulated and ADP-stimulated platelets in a saturable manner with the dissociation constants of 76 and 74 nM, respectively.17 It is 2.3 times more active than Den.

To compare the structure, function, and dynamics relationships of Den and Rho, it is essential to express them with the correct fold and high yield. Expression of three-fingered toxins with the correct fold has been obtained; however, the yield is less than 1 mg/L, or refolding of the protein is required.18–22 Many reports2, 16, 23 have shown that highly disulfide-bonded proteins can be expressed in P. pastoris with high yields. In the present study, we expressed a three-fingered toxin, Den, in P. pastoris with high yield and compared its backbone dynamics with those of Rho. This comparative study on the function, structure, and dynamics of Den and Rho serves as a basis for insight into the structure–function–dynamics relationships of integrin antagonists with different scaffolds.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Expression, purification, and characterization of Den and Rho

Den and Rho were expressed in P. pastoris X-33 strain using the pPICZαA vector. Recombinant Den and Rho expressed in P. pastoris were purified to homogeneity using Ni2+-chelating chromatography and C18 reversed-phase HPLC. Based on SDS–polyacrylamide gel electrophoresis, proteins produced in P. pastoris were homogenous (Supporting Information Fig. 1). The yields of Den and Rho produced in P. pastoris were 8–15 and 12–20 mg/L, respectively. In addition, the yield of 15N-labeled Den and Rho produced in P. pastoris was 5–10 and 14–17 mg/L. Mass spectrometry was used to determine their molecular weights. The experimental molecular weights of Den and Rho produced in P. pastoris were 7844.4 and 8417.4, which were in excellent agreement with the calculated values of 7844.8 and 8417.1, respectively. The values were calculated by assuming that all cysteines formed disulfide bonds, which indicated the formation of four disulfide bonds in Den and six disulfide bonds in Rho.

thumbnail image

Figure 1. 2D 1H-15N HSQC spectra of recombinant Den at pH 4. The protein concentration was 1 mM. Correlation peaks are labeled according to residues type and sequence number. The resonances of the side chains were labeled with “s” sign, and the peaks connected by lines correspond to Gln and Asn side-chain NH2 group.

Download figure to PowerPoint

Inhibition of platelet aggregation

The recombinant Den expressed in P. pastoris inhibited platelet aggregation with a KI of 149.8 ± 13.0 nM, which was consistent with the reported value of 172.0 ± 22.0 nM (Table I).13 The recombinant Rho expressed in P. pastoris inhibited platelet aggregation with a KI of 83.2 ± 10.4 nM, which was as potent as native Rho (Table I). Recombinant Den was 1.8 times less active than recombinant Rho in inhibiting platelet aggregation. These results showed that different scaffolds had little effect on their interactions with integrin αIIbβ3 of the platelets.

Table I. Summary of the Inhibition of Platelet Aggregation and Cell Adhesion by Den and Rho
ProteinIC50 (nM)
PlateletαIIbβ3/FgαVβ3/Fgα5β1/Fn
DEN149.8±13.077.4±20.332.5±9.4239.8±41.9
Rho83.2±10.421.0±11.213.0±5.7256.8±87.5

Inhibition of cell adhesion to fibrinogen and fibronectin

Inhibition of cell-expressing integrins αIIbβ3, αvβ3, and α5β1 to their ligands by Den and Rho was examined. The adhesions of CHO-expressing αIIbβ3 and αvβ3 to immobilized fibrinogen are integrins αIIbβ3- and αvβ3-dependent, respectively.24 They were consistent with the results of different integrin function-blocking mAbs.25, 26 In addition, the adhesion of K562 cells to fibronectin (FN) in the presence of 500 μM Mn2+ was shown to be predominantly α5β1-dependent.27 Their inhibitory constants are summarized in Table I. Den and Rho inhibited the adhesion of CHO cells that expressed integrin αIIbβ3 to immobilized fibrinogen with the IC50 values of 77.4 and 21.0 nM. Similar to the result of the platelet aggregation assay, Den was 3.7 times less active in inhibiting the expression of the integrin αIIbβ3 than was Rho. Den and Rho inhibited CHO cells that expressed αvβ3 from adhering to immobilized fibrinogen with the IC50 values of 32.5 and 13.0 nM. Den and Rho inhibited K562 cells from adhering to immobilized FN with the IC50 values of 239.8 and 256.8 nM, respectively. In contrast to their similar inhibitory activity of integrin α5β1, Den exhibited 3.7 and 2.5 times less active in inhibiting the expression of the integrins αIIbβ3 and αvβ3 than was Rho.

Structure determination

NMR spectroscopy was used to examine the folding of Den. Because there are 1H chemical shifts of Den at pH 4, we first measured 2D and 3D NMR spectra of Den under the same conditions (Figs. 1 and 2 and Supporting Information Figs. 2 and 3).28, 29 1H-15N HSQC spectra of recombinant Den at pH 4 showed that the NH chemical shifts were consistent with the reported values, which indicated that recombinant Den had the correct fold (Fig. 1). Only the HN chemical shift of I2 (Δδ = −0.67 ppm) downshifts more than 0.5 ppm because of six extra histidines at the N-terminus. Although the chemical shifts of the recombinant Den expressed in P. pastoris were consistent with those of native Den, it is still necessary to identify its disulfide pairings and secondary structures. Therefore, we performed NOESY experiments of Den at pH 2, 4, and 5.5 in 100% D2O to determine the four disulfide bonds of Den. Their pairings can be determined by searching Hβ to Hβ, Hβ to Hα, and Hα to Hα NOEs between different cysteines. Specifically, the NOEs between Hβ and Hβ of different cysteines can provide 98% uniqueness.30 NOESY spectra of Den at pH 4 were used to examine the NOE patterns of the disulfide bridges (Supporting Information Fig. 2). All four cysteine pairs of 3–22, 17–37, 39–51, and 52–57 were found from their Hβ/Hβ and Hβ Hα NOE patterns in the spectra. NMR analysis of the secondary structures of recombinant Den showed that it exhibited the double- and triple-stranded antiparallel β-sheets and three loops as do native proteins. The formation of double- and triple-stranded antiparallel β-sheets was characterized by the Hα-Hα, Hα-HN, and HN-HN NOE patterns of the connecting strands, the slowly exchanging amide protons, and the downfield-shifted α protons. Strip plots of 15N-edited NOESY of Den at pH 4 clearly showed the NOEs between Hα of C22 to HN of G38, Hα of K24 to HN of G36, Hα of N25 to HN of Y50, Hα of R35 to HN of N25, Hα of C37 to HN of Y23, HN of Y23 to HN of G36, HN of K24 to HN of Y50, and HN of N25 to HN of R34, which indicated the formation of triple-stranded antiparallel β sheets (Supporting Information Fig. 3). NOEs of W27 to I32 and M46, as well as a slow H/D exchange rate for the side-chain amino protons of N58, were found as native proteins, which suggested that they maintain their tertiary fold (Fig. 3). Based on our NMR studies, the recombinant Den produced in P. pastoris has the same three-fingered fold as native proteins.29

thumbnail image

Figure 2. Amide strip plots of Den and Rho. A: Amide strips from R43 to M46 of Den and (B) from R49 to M52 of Rho at pH 6.0. The dNN (i, i +1) and dαN (i, i +1) NOE connectivities are shown.

Download figure to PowerPoint

thumbnail image

Figure 3. Solution structures of Den. Stereoview of 20 lowest-energy NMR structures of Den were shown. The β-sheet secondary structure is shown in light blue.

Download figure to PowerPoint

The solution structure of Den was determined using NMR spectroscopy and the hybrid distance geometry-dynamical simulated annealing method. 1H-15N HSQC spectra of recombinant Den were recorded at pH 4, 4.8, and 5.2. Upon raising the pH from 4.0 to 5.2, many of the amide proton resonances decreased in intensity, and the R43, G44, and D45 residues showed similar NOE patterns. Therefore, NMR spectra at pH 4 were chosen for structure determination. NMR assignment of Den was obtained by analyzing standard 2D homonuclear and 3D heteronuclear NMR data. The 3D structure of Den was calculated using 773 experimentally derived restraints with an average of 13.1 restraints per residue (Table II). The 20 best structures of Den from 100 initial structures are shown in Figure 3. The backbone RMSD value of Den was 1.25 ± 0.44 Å, and the backbone RMSD values of Den for β-sheet regions (3–22, 17–37, 39–51, and 52–57) were 0.48 ± 0.10 Å. Based on Ramachandran analysis, all dihedral angles of Den were in the allowed region. A summary of the restraints and structural statistics is presented in Table II. Overall, the tertiary structure of Den has an elongated and asymmetric shape and consists of three two-stranded antiparallel β-sheets with many tight turns and loops.

Table II. Summary of Structural Restraints and Statistics for Den
Restraints used in the structure calculation
Distance and dihedral angle restraints
Intraresidue69
Sequential105
Medium range249
Long range271
Hydrogen bonds21
Dihedral angles58
Total773
Energy statistics
X-PLOR energy (kcal mol−1)
ENOE63.39 ± 6.11
Evdw3.40 ± 0.85
Geometric statistics
Deviations from idealized geometry
All backbone atoms (Å)1.25 ± 0.44
Backbone atoms (2–5, 13–17, 21–26, 34–38, and 49–54) (Å)0.48 ± 0.10
All heavy atoms (Å)2.03 ± 0.43
Heavy atoms (2–5, 13–17, 21–26, 34–38, and 49–54) (Å)1.21 ± 0.22
Ramachandran analysis
Most favored regions (%)74.5
Additionally allowed regions (%)25.5
Generously allowed regions (%)0.0

Structural difference between the RGD motif of Den and Rho

Because Den and Rho have different 3D scaffolds, we compared only their 3D conformations of the RGD loop that is the binding site for integrins. Ten of 20 of the Den and Rho structures were selected to align the nine-residue RGD loop, and the RMS deviations of the nine-residue backbone atoms of Den and Rho were 0.356 and 0.551 Å, respectively [Fig. 4(A,B)]. The RGD loop conformations of Den and Rho were found to have similar conformations forming a type I turn; however, there were differences in their backbone conformation (Fig. 4). This is consistent with the NOE amide strips from R43 to M46 of Den (Fig. 2).

thumbnail image

Figure 4. Structural alignment of the RGD loop. Structural alignments of the nine-residue RGD loop containing RI[P/A]RGDMPD sequences of (A) Den and (B) Rho. Ten of 20 Den and Rho structures were aligned, and the RMS deviations of the nine-residue backbone atoms of Den and Rho were 0.356 and 0.551 Å, respectively. The side chains of R, D, and others are shown in blue, red, and green, respectively. C: Superimposition of the nine-residue RGD loop of the average structure of Den and Rho. The side chains of R and D are shown in light blue and light red, respectively.

Download figure to PowerPoint

No structural differences between the integrin αvβ3 complexes of Den and Rho

The docking of Den into integrin αvβ3 was used to identify its integrin-interacting residues, because they were most active in inhibiting integrin αvβ3. The distance and the hydrogen bond restraints were derived from the X-ray structure of integrin αvβ3 complexed with a cyclic RGD pentapeptide (PDB code 1L5G), and eight key interactions were found between integrin αvβ3 and the R and D residues.31 Using these restraints, we docked Den into integrin αvβ3. A comparison with Rho showed that the docking of Den into integrin αvβ3 resulted in a similar number of contacts (Supporting Information Table SI). The key contacts—seven hydrogen bonds and two salt bridges between the R and D residues of the RGD motif and integrin α5β1—were the same.

Dynamics analysis of Den

1H-15N-correlated NMR spectroscopy was used to measure 15N R1, 15N R2, and 1H-15N NOE parameters for Den. The optimized value of τm for Den was determined to be 5.2 ns. This value was smaller than that estimated from the trimmed mean R2/R1 ratios, which was 3.6 ns. However, the values of the other parameters agreed with the values of long neurotoxin III, toxin α, and cardiotoxin (CTX), which were 4.6, 3.7, and 4.8 ns. Because Den is an axially symmetric ellipsoid, the axially symmetric model was used for the analysis. Starting from the initial estimates of τm, D///D-, θ and Ψ, we analyzed the 15N relaxation data by using axially symmetric models for rotational diffusion tensor. The obtained diffusion tensor of Den was asymmetric with D///D- = 1.259, which was higher than the reported values of 1.12 for long neurotoxin III and 1.06 for toxin α. The RGD-containing loop III, particularly the G44 residue, and the N-terminal region with low-NOE values showed distinctive backbone flexibility, which is consistent with the NMR structure of Den. The R1 and R2 values were very similar throughout the sequence, except that the residues G44, D45, and M46 in the loop III had lower NOE values, the G38 residue had a higher R2 value, and the residues R1, F40, and G44 had lower R1 values (Fig. 5 and Supporting Information Fig. S4).

thumbnail image

Figure 5. The relaxation parameters of Den. A: 15N R1 with error. B: 15N R2 with error. C: 1H-15N steady-state NOE with error. Gaps indicate the proline residues, and the β-sheet secondary structure is shown.

Download figure to PowerPoint

The square of the generalized order parameter (S2), the effective internal correlation time (τe), and a conformational exchange broadening parameter (Rex) for each backbone amide NH vector were determined using the model-free formalism (Fig. 6 and Supporting Information Fig. S5).32, 33 The average S2 values of Den overall and the β-sheet were 0.84 and 0.90, respectively. Compared to the rest of the Den backbone, only the S2 values of the residue R1 were lower than that of G48. Most residues of Den exhibited extensive flexibility with fast motion on the picasecond per nanosecond time scale. The N-terminal region was the most flexible region with an average S2 value of 0.26; however, the C-terminal was very rigid with an average S2 value of 0.90. The average S2 values of loops I and II were 0.87 and 0.81, and loop III, the flexible RGD loop, had an average S2 value of 0.69. The average S2 values of N-terminus and loops II and III were lower than the overall average value, which suggested that these regions were flexible. The tips of loop I (L7 and T9), loop II (N31, I32, I33, and R34), and loop III (D45, and M46) in Den exhibited flexible motions on the picasecond per nanosecond timescale.

thumbnail image

Figure 6. The model-free parameters of Den. Generalized order parameters (A) S2, (B) τm, and (C) Rex. Gaps indicate the proline residues, and the β-sheet secondary structure is shown. Only some fitting models resulted in the Rex and τe terms.

Download figure to PowerPoint

Comparison of the dynamic properties of Den and other three-fingered toxins

Backbone dynamics of two three-fingered toxins, toxin α, and CTX were determined using NMR spectroscopy.34, 35 A comparison of the backbone dynamics of these two toxins with that of Den showed that they were different from their average S2 values for different regions (Table III). It is consistent that the average S2 values of the three-fingered toxins' β-sheet were higher, and those of loops II and III and of the N-terminal region were lower than the overall average values. In contrast, the average S2 values of loop I and the C-terminal region were diverse and depended on individual three-fingered toxins. For example, the average S2 values of loop I and the C-terminal region of Den and toxin α were higher than their overall average value. In contrast, those of CTX were lower than its overall average value. Interestingly, their target-interacting loops had lower S2 values. For example, loop III of Den, the integrin-binding loop, had the lowest average S2 value. Loop II of toxin α and CTX also had the lowest average S2 values.

Table III. Average Order Parameter Values of Three-Fingered Toxins
ProteinsAverage S2
Overallβ-SheetsN-terminal regionLoop ILoop IILoop IIIC-terminal region
  • a

    N-terminal region (1), β-sheets (2–5, 13–17, 21–26, 34–38, and 49–54), loop I (6–12), loop II (27–33), loop III (39–48), and C-terminal region (55–59) of DEN are determined.

  • b

    N-terminal region (1), β-sheets (2–4, 14–16, 23–29, 34–40, and 49–54), loop I (5–13), loop II (30–33), loop III (41–48), and C-terminal region (55–61) of toxin α are determined.

  • c

    N-terminal region (1), β-sheets (2–4, 11–13, 20–25, 34–39, and 50–54), loop I (5–10), loop II (26–33), loop III (40–49), and C-terminal region (55-60) of CTX are determined.

  • d

    ND, not determined.

DENa0.840.900.260.870.810.690.90
Toxin αb0.800.83NDd0.840.700.750.84
CTXc0.800.840.720.780.770.790.76

Comparison of the dynamic properties of Den, Rho, and RGD-containing proteins

The backbone dynamics of many RGD-containing peptides and proteins, including Rho, echistatin, FN, and tenascin (TN), were determined using NMR spectroscopy.36, 35, 37, 38 A summary of these values of the residues R, G, D, and M/S is presented in Table IV. Although both Den and Rho have a PRGDMP amino acid sequence in the loop, the backbone dynamics of this sequence are very different from those of the RGDM motif, except that their M residues exhibited very similar backbone dynamics. The S2 values of the R, G, D, and M residues in the RGD loop of Den were 0.82, 0.47, 0.64, and 0.63, and in the RGD loop of Rho were 0.75, 0.63, 0.78, and 0.66. The S2 values of the Rho R and D residues were close to its overall average value. In contrast, the S2 values of Den D residue were lower than its overall average value, and the value of the G residue was 0.47, which indicated that the RGD loop of Den was more flexible than that of Rho.

Table IV. Relaxation Data and Dynamic Parameters of RGD-Containing Proteins
 R1 (s−1)R2 (s−1)NOES2τe (ns)Rex (s−1)IC50 (nM)
  • a

    Dendroaspin from this study.

  • b

    Dynamic properties and IC50 of rhodotomin are reported by Chen et al.38

  • c

    Dynamic properties of type III domain of fibronectin and tenascin are reported by Carr et al.37

  • d

    Dynamic properties of echistatin are reported by Chen et al.35

  • e

    IC50 of fibronectin is reported by Belvisi et al.39

  • f

    IC50 of tenascin is reported by Chung and Erickson.40

  • g

    IC50 of echistatin is reported by Chen et al.41

R43Dena1.95 ± 0.066.36 ± 0.170.69 ± 0.010.82 ± 0.02  32.5 ± 9.4a
R49Rhob1.69 ± 0.088.32 ± 0.130.48 ± 0.010.75 ± 0.010.11 ± 0.041.39 ± 0.1513.0 ± 5.7b
R78FN10c   0.463.25 835 ± 287e
R76TN3c   0.86  30f
G44Dena1.36 ± 0.053.63 ± 0.030.03 ± 0.010.47 ± 0.020.09 ± 0.01 32.5 ± 9.4a
G50Rhob1.55 ± 0.025.09 ± 0.030.32 ± 0.020.63 ± 0.010.81 ± 0.03 13.0 ± 5.7b
G79FN10c   0.392.75 835 ± 287e
G77TN3c   0.83  30f
G25Echd2.44 ± 1.279.06 ± 1.580.35 ± 0.020.50 ± 0.021.19 ± 0.054.44 ± 0.1520.7 ± 8.0g
D45Dena1.61 ± 0.055.17 ± 0.020.39 ± 0.010.64 ± 0.030.08 ± 0.03 32.5 ± 9.4a
D51Rhob1.72 ± 0.0310.03 ± 0.020.50 ± 0.020.78 ± 0.010.11 ± 0.012.94 ± 0.1613.0 ± 5.7b
D80FN10c   0.522.40 835 ± 287e
D78TN3c   0.77 1.3230f
D26Echd2.86 ± 1.015.88 ± 1.400.46 ± 0.020.46 ± 0.021.67 ± 0.081.17 ± 0.1320.7 ± 8.0g
M46Dena1.57 ± 0.064.99 ± 0.270.35 ± 0.010.63 ± 0.020.08 ± 0.01 32.5 ± 9.4a
M52Rhob1.51 ± 0.056.08 ± 0.010.44 ± 0.010.66 ± 0.010.07 ± 0.01 13.0 ± 5.7b
S81FN10c   0.442.41 835 ± 287e
M79TN3c   0.78  30f

In comparison with the backbone dynamics of other RGD-containing proteins, the Den and Rho relaxation parameters were very diverse. For example, the ranges of the S2 order parameters of R, G, D, and M/S residues were 0.46–0.86, 0.39–0.83, 0.46–0.78, and 0.44–0.78. The values of the τe and Rex parameters were also very diverse, and limited correlations were found from these parameters. However, it is consistent that their RGD loop was flexible and that the degree of flexibility depended on the proteins. For example, the RGD loops of Ech and FN were more flexible than that of TN (Table IV). Most of the G residue exhibited fast motion on the picasecond per nanosecond time scale, and the D residue exhibited slow motion on the microsecond per millisecond time scale and fast motion on the picasecond per nanosecond time scale. These dynamic properties may be important for their binding to integrins.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

The structure and dynamics relationships of the RGD motif of RGD-containing proteins affect their binding specificities and affinities on integrins.35–38 In the present study, we showed that Den and Rho have different tertiary folds with similar 3D conformation for the RGD motif. However, their ability to inhibit the integrins αvβ3, α5β1, and αIIbβ3 was different and to inhibit the integrin αvβ3 was significantly different. In contrast, docking analysis showed no interaction differences between their RGD motif and the integrin αvβ3. A difference was found, however, in the backbone dynamics of their RGD residues. These results showed that protein scaffolds affect their function and dynamics but not the conformation of the RGD motif, which suggests that the flexibility and the motions of the RGD residues may be important for their interaction with integrins.

The presence of the RGD amino acid sequence is not sufficient for its biological activity. Studies on identifying the criteria for biologically active RGD motif were well characterized.42 The RGD motif, the important binding site for integrin, lies in a mobile loop ranging from 5 to 11 residues joining two strands of β sheet.12 The RGD motif is located at the apex of a long loop, between two β strands of the protein, protruding 10–17 Å from the protein core.6–12 The flexibility and solvent exposure of the RGD loop are responsible for the recognition and fitting the interface between integrin α and β subunits. Studies35, 37, 38 also show that the dynamic properties of the RGD motif may be important in its interaction with integrins. It is known that the residues flanking the RGD motif of RGD-containing proteins affect their binding specificities and affinities on integrins.27, 36 Although Den and Rho have the same PRGDMP sequence in their RGD loop, they exhibited diverse specificity in inhibiting integrins. These results suggest that not only the flanking residues but also protein scaffold can affect the backbone dynamics of the RGD motif.

The synergy between structure and dynamics is essential to the function of integrin complexes.6–12 Many studies have compared the binding affinity of RGD-containing peptide with their 3D conformations and backbone dynamics.27, 36, 38, 43 As shown in Table IV, the residues of the RGD loop of these proteins exhibited extensive flexibility on fast motion on the picasecond per nanosecond time scale. In general, proteins and peptides with a flexible RGD loop exhibited lower binding affinity and interacted with many integrins. However, proteins and peptides with a specific RGD conformation exhibited higher binding affinity and interacted with specific integrins. This is consistent with our results that the RGD loop of the tenth-type III domain of FN (FN10) is more flexible than those of Den, Rho, and Ech, and FN10 exhibited lowest integrin-binding affinity (Table IV). The flexibility of the RGD loop in FN supports its recognition to diverse integrins. In contrast, the more rigid RGD loop in TN reflects their specificity for particular integrins.37 Analyses of the results indicate that the dynamic properties and 3D conformation of the RGD motif affect their specificity and binding affinity to integrins. However, backbone dynamics of individual RGD residues of these proteins were not correlated with their functions. For example, the S2 value for the RGD residues in Den and Rho was different: the S2 value of the R residue in Den was larger than Rho; and the S2 values of the G and D residues in Den were smaller then those of Rho. The correlation of structure and dynamics of the RGD-containing proteins with function is required for further investigation.

Toxins with unrelated folds can have similar functions. Three-fingered snake toxins show a wide spectrum of biological activities with a similar fold. In this study, we expressed Den in P. pastoris and used NMR spectroscopy to confirm its folding. Our analyses of the NOE patterns between different cysteines, the chemical shifts, and the secondary structures of Den indicate that the folding of Den expressed in P. pastoris was correct. The disulfide pairings of Den were in close spatial proximity, which is consistent with its reported disulfide pairings.13, 14 The present study provides the first direct evidence that highly disulfide-bonded three-fingered toxins can be expressed in P. pastoris with the correct fold and be labeled with the 15N isotope. Comparing the backbone dynamics of Den with those of toxin α and CTX showed that they were different from their average S2 values for different regions. In particular, the RGD loop of Den had the lowest average S2 value, which is consistent with the RGD loops of Rho and FN being the most flexible region in these proteins. Although the relaxation parameters of the RGD residues in proteins with different scaffolds are diverse, it is evident that most of the G residue exhibited fast motion on the picasecond per nanosecond time scale, and D residue exhibited slow motion on the microsecond per millisecond time scale and fast motion on the picasecond per nanosecond time scale. These results suggest that not only conformation but also the dynamic properties of the RGD motif are important for its biological activity.

In conclusion, this is the first comparison of the relaxation parameters of the RGD loop of different protein scaffolds. We expressed Den, a three-fingered toxin containing four disulfide bonds, in a P. pastoris expression system. We found that Den produced in P. pastoris has the same function and structure as those in native protein. This study provides the direct evidence that highly disulfide-bonded three-fingered toxins can be expressed in P. pastoris with the correct fold. This evidence may serve as the basis for uncovering the species and tissue in which evolution has used this adhesive domain. Our study also suggests that the inhibitory potency and selectivity of RGD-containing proteins cannot be determined only by the sequence content of the RGD loop. This study may serve as the basis for exploring the structure, dynamics, and functions relationships of integrins and their ligands.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Expression and purification of Den and Rho in P. pastoris

The expression and purification of Den and Rho in P. pastoris were done following previously described protocols.16, 44 Den and Rho expression in P. pastoris was achieved using yeast transfer vector (pPICZα A; Invitrogen) and an expression kit (EasySelect version A; Invitrogen) with minor modifications to the manufacturer's instructions. The structural gene of Den was amplified using a polymerase chain reaction (PCR) with the sense primer: 5′-AATTCGAATTCCATCATCATCATCATCATCATGG TAAGGAATGTGACTGTTCTT-3′, has Eco RI recognition and six histidine residues for facilitating purification. The antisense primer was 5′-CCGCGG CCGCGGTCAGTGGTATCTTGGACAGTCAGC-3′ with Sac II recognition and a TCA stop codon. The structural gene of Rho was amplified using PCR with the sense primer: 5′-AATTGAATTCCATCATCATCATCAT CATAGAATTTGTTTTAATCATCAATCTTCTCAACCA CAAACTACT-3′, which has Eco RI recognition and six histidine residues for facilitating purification. The antisense primer was 5′-AATTCCGCGGATTAT TACAAACTTCAGATTCACAACAAGACAACTTAATA CCTGGCTTAACAGTTGGACA-3′ with Sac II recognition and a TAA stop codon. The structural gene of Rho was amplified using a PCR with the sense primer: 5′-AATTGAATTCCATCATCATCATCATCAT AGAATTTGTTTTAATCATCAATCTTCTCAACCACA AACTACT-3′, which has Eco RI recognition and six histidine residues for facilitating purification. The antisense primer was 5′-AATTCCGCGGATTATTACAAACTTCAGATTCACAACAAGACAACTTAATAC CTGGCTTAACAGTTGGACA-3′ with Sac II recognition and a TAA stop codon. The PCR product was purified and then cloned into the Eco R1 and Sac II sites of pPICZαA. The recombinant plasmid was transformed into the DH5α strain, and the colony was selected using an agar plate with low-salt lateral binding (1% tryptone + 0.5% yeast extract + 0.5% NaCl + 1.5% agar [pH 7.5]) and 25 μg/mL of the antibiotic phleomycin D1 (Zeocin; Invitrogen). After the clone had been confirmed by sequencing the insert, 10 μg of plasmid was digested with Sac I to linearize the plasmid. The linearized construct was transformed into the Pichia strain X-33 using the heat-shock method, and the transformation was done with a kit (Pichia EasyComp; Invitrogen). The transformant was integrated at the AOX1 locus using single crossover, and the colony was selected using an agar plate with yeast peptone dextrose (1% yeast extract, 2% peptone, 2% glucose, and 2% agar) and 100 μg/mL of Zeocin. We picked the highest protein expression clone of Den and Rho from a number of clones with multiple copies of the Den gene insertion.

We produced 15N-labeled Den and Rho as follows: 100 μL of cell stock was grown at 30°C in 200 mL of 15N minimal medium [0.34% yeast nitrogen base (YNB) without ammonium sulfate and amino acids, 2% dextrose, 4 × 10−5% biotin, and 0.05% 15NH4Cl] in 100 mM potassium phosphate buffer [pH 6] with 100 μg/mL of Zeocin for 48 h. The cells were then transferred into 800 mL of 15N minimal medium. After another 48 h, the cell densities reached OD600 of 25–30. The cells were then centrifuged, collected, and grown in 1 L of 15N minimal methanol medium (0.34% YNB without ammonium sulfate and amino acids, 4 × 10−5 % biotin, 1% methanol, and 0.05% 15NH4Cl). Once every 24 h, 1% methanol was added to induce protein expression for 2 days. The supernatant was collected using centrifugation and dialyzed twice against 4 L of H2O. The final solution was loaded into a nickel-chelating column and eluted with a gradient of 200 mM imidazole. The recombinant Den and Rho produced in P. pastoris was further purified by reversed-phase C18 HPLC with a gradient of 15–20% acetonitrile. The purification of recombinant proteins was greater than 95% pure as judged using tricine–SDS–PAGE.

Mass spectrometry measurements

The molecular weights of Den and Rho were confirmed using an API 365 triple quadrupole mass spectrometer equipped with a TurboIonSpray source (PE Sciex, Thornill, Canada). Protein solutions (1–10 μM in 50–90% methanol or acetonitrile with 0.1% formic acid) were infused into the mass spectrometer using a syringe pump (Harvard Apparatus, South Natick, MA) at a flow rate of 12–20 μL/min to acquire full-scan mass spectra. The electrospray voltage at the spraying needle was optimized at 5000–5300 V. The molecular weights of proteins were calculated using computer software provided with the spectrometer.

Platelet aggregation assay

Venous blood (nine parts) from healthy donors who had not received any medication for at least 2 weeks was collected in 3.8% sodium citrate (one part). Blood was centrifuged at 150g for 10 min to obtain platelet-rich plasma (PRP) and allowed to stand for 5 min, after which PRP was collected. Platelet-poor plasma (PPP) was prepared from the remaining blood by centrifuging it at 2000g for 25 min. The PPP platelet count was measured on a hematology analyzer and diluted to 250,000 platelets/μL. A solution of 190 μL of PRP and 10 μL of either Rho or PBS buffer was incubated for 5 min using an aggregometer (Hema Tracer 601; Nikoh Bioscience, Tokyo) at 37°C. Ten microliters of 200 mM ADP was added to monitor the response of platelet aggregation using light transmission.

Cell adhesion assay

A cell adhesion assay was done using protocols previously described.27, 36, 43 Ninety-six-well microtiter plates (Costar; Corning) were coated with 100 μL of PBS buffer containing 200 μg/mL fibrinogen or 25 μg/mL FN and incubated overnight at 4°C. Nonspecific protein-binding sites were blocked by incubating each well with 200 μL of heat-denatured 1% BSA (Calbiochem) at room temperature for 1.5 h. The heat-denatured BSA was discarded, and each well was then washed twice with 200 μL of PBS.

CHO cells that expressed the integrins αvβ3 (CHO-αvβ3) and αIIbβ3 (CHO-αIIbβ3) were kindly provided by Dr. Y. Takada (Scripps Research Institute) and maintained in DMEM.36, 43 Human erythroleukemia K562 cells were purchased from ATCC and cultured in Roswell Park Memorial Institute (RPMI)-1640 medium containing 5% FCS. Harvested K562 cells were washed in PBS buffer containing 1 mM EDTA and resuspended in Tyrode's buffer (150 mM NaCl, 5 mM KCl, and 10 mM Hepes) [pH 7.35] containing 1 mM MgSO4, 2 mM CaCl2, and 500 μM MnCl2. CHO and human erythroleukemia K562 cells were diluted to 3 and 2.5 × 105 cells/mL, respectively, and 100 μL of the cells was used for the assay. Den and Rho (0.001–500 μM), which were used as inhibitors, were added to the cells and incubated at 37°C in a 5% CO2 atmosphere for 15 min. The treated cells were then added to the coated plate and reacted at 37°C (5% CO2) for 1 h. The reacting solution was then discarded, and nonadhered cells were removed by washing them twice with 200 μL of PBS. After the nonadhered cells had been removed by rinsing the wells with the same buffer, adhered cells were quantified using a crystal violet assay. The well was fixed with 100 μL of 10% formalin for 10 min and then dried. A solution of 50 μL of 0.05% crystal violet was added to the well at room temperature for 20 min. Each well was then washed four times with 200 μL of distilled water and dried. Colorizing solution (150 μL of 50% alcohol and 0.1% acetic acid) was then added. The resulting absorbance was read at 600 nm, and the readings were correlated with the number of adhering cells. Inhibition was defined as % inhibition = 100 – [OD600 (Rho protein-treated sample)/OD600 (untreated sample)] × 100. The reported IC50 values are the average of at least three separate experiments.

NMR spectroscopy

NMR experiments were done at 27°C on a spectrometer (Bruker Avance 600- and 700 MHz) equipped with pulse-field gradients and xyz-gradient triple-resonance probes. In these experiments, samples were dissolved in 10% D2O/90% H2O or 100% D2O at a concentration of 3 mM; pHs were adjusted to 4.0 or 6.0 with 100 mM KOD. NMR spectra of Ech at pH 4 were performed for resonance assignment. The data were processed with Topspin Version 1.3 software and analyzed with Aurelia software. 2D NOESY, TOCSY, and DQF-COSY NMR spectra were recorded in the phase-sensitive absorption mode with quadrature detection in both F1 and F2 dimensions. A concentration of 2 mM 15N-labeled Den and Rho was used for the 2D 1H-15N HSQC, 3D 1H-15N edited-TOCSY-, and NOESY-HSQC experiments. Mixing times of 30–90 and 60–150 ms were used for TOCSY and NOESY experiments, respectively. The center frequencies of double resonance experiments were 4.75 (1H) and 118 ppm (15N). The observed 1H chemical shifts were referenced with respect to the H2O or HOD signal, which was 4.754 ppm downfield from external sodium 3-trimethylsilylpropionate-2,2,3,3-d4 in D2O (0.0 ppm) at 300 K. The nitrogen chemical shift was referenced to external 15NH4Cl (3 mM in 1M HCl) at 300 K, which is 24.93 ppm downfield from liquid NH3.

Structure calculations

Structures were calculated using the program X-PLOR with the hybrid distance geometry-dynamical simulated annealing method.45 NOESY cross-peak intensities—categorized into strong, medium, weak, and very weak—were converted into distance constraints of 1.8–2.8, 1.8–3.6, 1.8–5.0, and 2.5–6.0 Å, respectively. Pseudoatom corrections were used for methylene, methyl, and aromatic protons, and an additional 0.5 Å was added to the upper-limit distances involving methyl protons. The dihedral angles ϕ were determined from the 3JNHα coupling constants. For 3JNHα values less than 5 Hz, ϕ values were restricted from –30° to –90°, and for 3JNHα values greater than 10 Hz, ϕ values were restricted from –100° to –170°. Two restraints were used for each NH-CO backbone hydrogen bond with dN[BOND]O restricted to 2.4–3.3 Å and dH[BOND]O to 1.7–2.3 Å. A family of 100 structures was generated using NOE distance, dihedral angle, and hydrogen bond restraints. The S-S covalent bonds were deleted and reintroduced as pseudo-NOE distances with the S-S distances constrained to the upper limit of 2.1 Å. During the first phase of dynamics at 2000 K, the value of the force constant of the NOE term was kept constant at 50 kcal/mol−1 Å−2. The repulsion term was gradually increased from 0.03 to 4.0 kcal/mol−1 Å−2 and the torsion angle term from 5 to 200 kcal/mol−1 rad−2. The simulated annealing refinement consisted of a 9-ps cooling dynamic followed by 200 cycles of Powell minimization. The 20 lowest-energy structures were accepted based on violations of distance restraints less than 0.5 Å, dihedral angle restraints less than 5°, a van der Waals energy cutoff value of 35 kcal/mol, and an NOE energy cutoff value of 55 kcal/mol. The structure figures were prepared using the MOLMOL or the PyMOL program.46, 47

Measurements of NMR dynamics

Backbone dynamics of Den and Rho acquired at two spectrometers (600 and 700 MHz). The 15N-spin-lattice (R1) and spin–spin (R2) relaxation rate constants and steady-state 1H-15N NOEs were measured from 1H-detected 1H-15N correlation spectra recorded with sensitivity-enhanced pulse sequences. A recycle delay of 6 s was used, and 128 complex T1 increments of 32 scans were acquired. A series of 10 experiments with relaxation delays of 30, 100, 150, 300, 450, 600, 800, 1000, 1500, and 3000 ms were done to measure T1. A series of 10 experiments with relaxation delays of 18, 36, 48, 72, 90, 100, 120, 150, 300, and 500 ms were done to measure T2. The longitudinal and transverse relaxation rate constants, R1 and R2, were obtained from exponential fits of the peak height data using least-squares fit software (SigmaPlot; Jandel Scientific). The reported Ri values are the mean values of two independent data sets. In the NOE experiment, two spectra—one with the NOE and one without—were collected. The NOE was calculated as the ratio of peak heights in spectra collected with and without NOE. The reported NOE value was the average value of three pairs of NOE experiments.

The heteronuclear 15N relaxation rate constants, R1 and R2, and the 1H-15N steady-state NOE values were analyzed using the FastModelFree program.48 In this approach, the overall and internal molecular motions were assumed to be independent, and the spectral density function for a molecule undergoing isotropic tumbling was calculated using the appropriate expression:

  • equation image

where 1/τ = 1/τm + 1/τe and S2 = S2sS2f, τm is the overall rotational correlation time of the molecule, τe is the effective correlation time for the motions on the slower of the two time scales, S2 is the square of the generalized order parameter, and S2s and S2f are the squares of order parameters for the motion on the slow and fast time scales, respectively.32

Molecular docking

The docking of Rho and Den to integrin αvβ3 was done with the docking program HADDOCK 1.3 using hydrogen bond and distance restraints.50 The starting structures for the docking were the average-minimized NMR structures of Rho and integrin αvβ3 (PDB code 1L5G). The interaction restraints were derived from an X-ray structure of integrin αvβ3 in complex with a cyclic pentapeptide {c[−RGDf(NMe)V−]} using the software iMoltalk.31, 51 The defined distance threshold was 4 Å, and the interaction restraints between the RGD motif and integrin were used for calculation. The input restraints between the R49/R43, G50/G44, and D51/D45 residues, and integrin αvβ3 were 34, 10, and 34, respectively. They were the contacts between the R49/R43 residue and the residues Asp150, Tyr178, Gln180, and Asp218 of integrin αv, between the G50/G44 residue and the residues R216 and A218 of β3, and between the D51/D45 residue and the residues Ser121, Tyr122, Ser123, Arg214, Asn215, Arg216, Asp217, and Ala218 and Mn2+ of the MIDAS of β3. An additional 0.5 Å distance was added to the upper and lower limits in interaction restraints. Using these restraints, the standard HADDOCK protocol for protein docking was done with minor modifications. This protocol combines three stages of molecular dynamics calculations, including heating and cooling with a progressive increase of the flexibility at the binding interface. In the first stage, 500 conformations were calculated using a rigid-body docking protocol. The best 100 structures in terms of their intermolecular energies were refined using semiflexible simulated annealing in the second stage. Both the side chains and the backbone atoms of the residues 46–54 of Rho, 40–48 of Den, 149–152, 176–181, 211–216, and 217–219 of αv, and 120–124 and 212–221 of β3 were defined as flexible, which allowed them to move in a semirigid-body docking protocol to search for conformational rearrangements. The resulting 100 structures with the lowest intermolecular energy values were refined with explicit water molecules in the last stage. The structures were classified using clustering based on the pairwise RMSD differences. Two clusters of structures were found by fitting them over the RGD residues with an average RMSD value <1.5 Å for the backbone atoms of all the amino acids in 100 integrin complexes.

Protein Data Bank accession numbers and NMR assignments

The co-ordinates of 20 calculated structures of Den and Rho have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/), under accession numbers 2LA1 and 2PJF, respectively. 1H and 15N resonances of Den and Rho have been deposited in the BioMagResBank databank under accession numbers BMRB-6915 and BMRB-5117.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

We are indebted to Drs. Wen-Mei Fu and Wenya Huang for their helpful comments.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information
  • 1
    Lewis RJ, Garcia ML ( 2003) Therapeutic potential of venom peptides. Nat Rev Drug Discov 2: 790802.
  • 2
    Anangi R, Chen C-Y, Cheng C-H, Chen Y-C, Chen C-C, Chu Y-P, Chang C-H, Chuang W-J ( 2007) Expression of snake venom toxins in Pichia pastoris. Toxin Rev 26: 169187.
  • 3
    Ménez A ( 1998) Functional architectures of animal toxins: a clue to drug design? Toxicon 36: 15571572.
  • 4
    Kini RM ( 2002) Molecular moulds with multiple missions: functional sites in three-finger toxins. Clin Exp Pharmacol Physiol 29: 815822.
  • 5
    Mouhat S, Jouirou B, Mosbah A, De Waard M, Sabatier JM ( 2004) Diversity of folds in animal toxins acting on ion channels. Biochem J 378: 717726.
  • 6
    Ruoslahti E ( 1996) RGD and other recognition sequences for integrins. Annu Rev Cell Dev Biol 12: 697715.
  • 7
    McLane MA, Sanchez EE, Wong A, Paquette-Straub C, Perez JC ( 2004) Disintegrins. Curr Drug Targets Cardiovasc Haematol Disord 4: 327355.
  • 8
    Mans BJ, Louw AI, Neitz AW ( 2002) Savignygrin, a platelet aggregation inhibitor from the soft tick Ornithodoros savignyi, presents the RGD integrin recognition motif on the Kunitz-BPTI fold. J Biol Chem 277: 2137121378.
  • 9
    Chen C-Y, Luo S-C, Kuo C-F, Lin Y-S, Wu J-J, Lin M-T, Liu C-C, Jeng W-Y, Chuang W-J ( 2003) Maturation processing and characterization of streptopain. J Biol Chem 278: 1733617343.
  • 10
    Shiu J-H, Chen C-Y, Chang L-S, Chen Y-C, Chen Y-C, Lo Y-H, Liu Y-C, Chuang W-J ( 2004) Solution structure of gamma-bungarotoxin: the functional significance of amino acid residues flanking the RGD motif in integrin binding. Proteins 57: 839849.
  • 11
    Wang FZ, Akula SM, Sharma-Walia N, Zeng L, Chandran B ( 2003) Human herpesvirus 8 envelope glycoprotein B mediates cell adhesion via its RGD sequence. J Virol 77: 31313147.
  • 12
    Blobel CP, White JM ( 1992) Structure, function and evolutionary relationship of proteins containing a disintegrin domain. Curr Opin Cell Biol 4: 760765.
  • 13
    McDowell RS, Dennis MS, Louie A, Shuster M, Mulkerrin MG, Lazarus RA ( 1992) Mambin, a potent glycoprotein IIb-IIIa antagonist and platelet aggregation inhibitor structurally related to the short neurotoxins. Biochemistry 31: 47664772.
  • 14
    Sutcliffe MJ, Jaseja M, Hyde EI, Lu X, Williams JA ( 1994) Three-dimensional structure of the RGD-containing neurotoxin homologue dendroaspin. Nat Struct Biol 1: 802807.
  • 15
    Dennis MS, Carter P, Lazarus RA ( 1993) Binding interactions of kistrin with platelet glycoprotein IIb-IIIa: analysis by site-directed mutagenesis. Proteins 15: 312321.
  • 16
    Guo R-T, Chou L-J, Chen Y-C, Chen C-Y, Pari K, Jen C-J, Lo S-J, Huang S-L, Lee C-Y, Chang T-W, Chuang W-J ( 2001) Expression in Pichia pastoris and characterization by circular dichroism and NMR of rhodostomin. Proteins 43: 499508.
  • 17
    Huang T-F, Sheu R-R, Teng C-M, Chen S-W, Liu C-S ( 1991) Triflavin, an antiplatelet Arg-Gly-Asp-containing peptide, is a specific antagonist of platelet membrane glycoprotein IIb-IIIa complex. J Biochem 109: 328334.
  • 18
    Chi L-M, Vyas AA, Rule GS, Wu W-G ( 1994) Expression of glutathione S-transferase-cardiotoxin fusion protein in Escherichia coli. Toxicon 32: 16791683.
  • 19
    Fiordalisi JJ, James PL, Zhang Y, Grant GA ( 1996) Facile production of native-like κ-bungarotoxin in yeast: an enhanced system for the production of a neuronal nicotinic acetylcholine receptor probe. Toxicon 34: 213224.
  • 20
    Kumar TK, Yang P-W, Lin S-H, Wu C-Y, Lei B, Lo S-J, Tu S-C, Yu C. ( 1996) Cloning, direct expression, and purification of a snake venom cardiotoxin in Escherichia coli. Biochem Biophys Res Commun 219: 450456.
  • 21
    Levandoski MM, Caffery PM, Rogowski RS, Lin Y, Shi QL, Hawrot E ( 2000) Recombinant expression of alpha-bungarotoxin in Pichia pastoris facilitates identification of mutant toxins engineered to recognize neuronal nicotinic acetylcholine receptors. J Neurochem 74: 12791289.
  • 22
    Drevet P, Lemaire C, Gasparini S, Zinn-Justin S, Lajeunesse E, Ducancel F, Pinkasfeld S, Courcon M, Tremeau O, Boulain JC, Ménez A ( 1997) High-level production and isotope labeling of snake neurotoxins, disulfide-rich proteins. Protein Expr Purif 10: 293300.
  • 23
    White CE, Kempi NM, Komives EA ( 1994) Expression of highly disulfide-bonded protein in Pichia pastoris. Structure 2: 10031005.
  • 24
    Takagi J, Kamata T, Meredith J, Puzon-McLaughlin W, Takada Y ( 1997) Changing ligand specificities of alphavbeta1 and alphavbeta3 integrins by swapping a short diverse sequence of the beta subunit. J Biol Chem 272: 1979419800.
  • 25
    Zhang XP, Kamata T, Yokoyama K, Puzon-McLaughlin W, Takada Y ( 1998) Specific interaction of the recombinant disintegrin-like domain of MDC-15 (metargidin, ADAM-15) with integrin alphavbeta3. J Biol Chem 273: 73457350.
  • 26
    Wattam B, Shang D, Rahman S, Egglezou S, Scully M, Kakkar V, Lu X ( 2001) Arg-Tyr-Asp (RYD) and Arg-Cys-Asp (RCD) motifs in dendroaspin promote selective inhibition of beta1 and beta3 integrins. Biochem J 356: 1117.
  • 27
    Rahman S, Aitken A, Flynn G, Formstone C, Savidge GF ( 1998) Modulation of RGD sequence motifs regulates disintegrin recognition of alphaIIb beta3 and alpha5 beta1 integrin complexes. Replacement of elegantin alanine-50 with proline, N-terminal to the RGD sequence, diminishes recognition of the alpha5 beta1 complex with restoration induced by Mn2+ cation. Biochem J 335: 247257.
  • 28
    Jaseja M, Lu X, Williams JA, Sutcliffe MJ, Kakkar VV, Parslow RA, Hyde EI ( 1994) 1H-NMR assignments and secondary structure of dendroaspin, an RGD-containing glycoprotein IIb-IIIa (alpha IIb-beta 3) antagonist with a neurotoxin fold. Eur J Biochem 226: 861868.
  • 29
    Klaus W, Broger C, Gerber P, Senn H ( 1993) Determination of the disulphide bonding pattern in proteins by local and global analysis of nuclear magnetic resonance data. Application to flavoridin. J Mol Biol 232: 897906.
  • 30
    Lipari G, Szabo A ( 1982) Model-free approach to the interpretation of nuclear magnetic resonance relaxation in macromolecules. I. Theory and range of validity. J Am Chem Soc 104: 45464559.
  • 31
    Xiong JP, Stehle T, Zhang R, Joachimiak A, Frech M, Goodman SL, Arnaout MA ( 2002) Crystal structure of the extracellular segment of integrin αVβ3 in complex with an Arg-Gly-Asp ligand. Science 296: 151155.
  • 32
    Clore GM, Szabo A, Bax A, Kay LE, Driscoll PC, Gronenborn AM ( 1990) Deviations from the simple two parameter model free approach to the interpretation of 15N nuclear magnetic relaxation of proteins. J Am Chem Soc 112: 49894991.
  • 33
    Lee C-S, Kumar TK, Lian L-Y, Cheng J-W, Yu C ( 1998) Main-chain dynamics of cardiotoxin II from Taiwan cobra (Naja naja atra) as studied by carbon-13 NMR at natural abundance: delineation of the role of functionally important residues. Biochemistry 37: 155164.
  • 34
    Connolly PJ, Stern AS, Turner CJ, Hoch JC ( 2003) Molecular dynamics of the long neurotoxin LSIII. Biochemistry 42: 1444314451.
  • 35
    Chen Y, Suri AK, Kominos D, Sanyal G, Naylor AM, Pitzenberger SM, Garsky VM, Levy RM, Baum J ( 1994) Three-dimensional structure of echistatin and dynamics of the active site. J Biomol NMR 4: 307324.
  • 36
    Shiu J-H, Chen C-Y, Chen Y-C, Chang Y-T, Chang Y-S, Huang C-H, Chuang W-J ( 2012) Effect of P to A mutation of the N-terminal residue adjacent to the RGD motif on rhodostomin: importance of dynamics in integrin recognition. PLoS One 7: e28833.
  • 37
    Carr PA, Erickson HP, Palmer AG3rd ( 1997) Backbone dynamics of homologous fibronectin type III cell adhesion domains from fibronectin and tenascin. Structure 5: 949959.
  • 38
    Chen C-Y, Shiu J-H, Hsieh Y-H, Liu Y-C, Chen Y-C, Chen Y-C, Jeng W-Y, Tang M-J, Lo S-J, Chuang W-J ( 2009) Effect of D to E mutation of the RGD motif in rhodostomin on its activity, structure, and dynamics: importance of the interactions between the D residue and integrin. Proteins 76: 808821.
  • 39
    Belvisi L, Bernardi A, Colombo M, Manzoni L, Potenza D, Scolastico C, Giannini G, Marcellini M, Riccioni T, Castorina M, LoGiudice P, Pisano C ( 2006) Targeting integrins: insights into structure and activity of cyclic RGD pentapeptide mimics containing azabicycloalkane amino acids. Bioorg Med Chem. 14: 169180.
  • 40
    Chung CY, Erickson HP ( 1994) Cell surface annexin II is a high affinity receptor for the alternatively spliced segment of tenascin-C. J Cell Biol 126: 539548.
  • 41
    Chen YC, Cheng CH, Shiu JH, Chang YT, Chang YS, Huang CH, Lee JC, Chuang WJ ( 2012) Expression in Pichia pastoris and characterization of echistatin, an RGD-containing short disintegrin. Toxicon 60: 13421348.
  • 42
    Torshin IY ( 2002) Structural criteria of biologically active RGD-sites for analysis of protein cellular function—a bioinformatics study. Med Sci Monit 8: BR301BR312.
  • 43
    Cherny RC, Honan MA, Thiagarajan P ( 1993) Site-directed mutagenesis of the arginine-glycine-aspartic acid in vitronectin abolishes cell adhesion. J Biol Chem 268: 97259729.
  • 44
    Chen C-Y, Cheng C-H, Chen Y-C, Lee J-C, Chou S-H, Huang W, Chuang W-J ( 2006) Preparation of amino-acid-type selective isotope labeling of protein expressed in Pichia pastoris. Proteins 62: 279287.
  • 45
    Brunger AT ( 1992) X-PLOR: version 3.1: a system for X-ray crystallography and NMR. New Haven: Yale University Press.
  • 46
    Koradi R, Billeter M, Wuthrich K ( 1996) MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph 14: 5155.
  • 47
    Schrodinger LLC ( 2010) The PyMOL molecular graphics system, Version 1.3r1.
  • 48
    Cole R, Loria JP ( 2003) FAST-Modelfree: a program for rapid automated analysis of solution NMR spin-relaxation data. J Biomol NMR 26: 203213.
  • 49
    Mandel AM, Akke M, Palmer AG, III ( 1995) Backbone dynamics of Escherichia coli ribonuclease HI: correlations with structure and function in an active enzyme. J Mol Biol 246: 144163.
  • 50
    Dominguez C, Boelens R, Bonvin AM ( 2003) HADDOCK: a protein-protein docking approach based on biochemical or biophysical information. J Am Chem Soc 125: 17311737.
  • 51
    Diemand AV, Scheib H ( 2004) iMolTalk: an interactive, internet-based protein structure analysis server. Nucleic Acids Res 32: W512W516.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and Methods
  7. Acknowledgements
  8. References
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
PRO_2169_sm_SuppInfo.pdf2653KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.