structural basis for dual inhibitory role of tamarind ... · tki-fxa complex revealed that the...

Structural basis for dual inhibitory role of tamarind Kunitzinhibitor (TKI) against factor Xa and trypsinDipak N. Patil, Anshul Chaudhary, Ashwani K. Sharma, Shailly Tomar and Pravindra Kumar

Department of Biotechnology, Indian Institute of Technology Roorkee, Roorkee, Uttarakhand, India

Keywords

factor Xa inhibitor; protein–protein

interaction; tamarind Kunitz type inhibitor;

versatile reactive site; X-ray structure

Correspondence

S. Tomar, Department of Biotechnology,

Indian Institute of Technology Roorkee,

Roorkee, Uttarakhand 247667, India

Fax: +91 1332 286151

Tel: +91 1332 285849

E-mail: [email protected]

P. Kumar, Department of Biotechnology,

Indian Institute of Technology Roorkee,

Roorkee, Uttarakhand 247667, India

Fax: +91 1332 286151

Tel: +91 1332 285072

E-mail: [email protected]

(Received 23 July 2012, revised 9 October

2012, accepted 19 October 2012)

doi:10.1111/febs.12042

A Kunitz type dual inhibitor (TKI) of factor Xa (FXa) and trypsin was

found in tamarind. It also shows prolongation of blood coagulation time.

The deduced 185 amino acid sequence of TKI by cDNA cloning and

sequence analysis revealed that it belongs to the Kunitz type soybean tryp-

sin inhibitor (STI) family; however, it has a distorted Kunitz signature

sequence due to insertion of Asn15 in the motif. TKI exhibited a competi-

tive inhibitory activity against both FXa (Ki = 220 nM) and porcine pancre-

atic trypsin (Ki = 3.2 nM). The crystal structure of TKI shows a b-trefoilfold similar to Kunitz STI inhibitors; however, a distinct mobile reactive

site, an inserted residue and loop b7b8 make it distinct from classical

Kunitz inhibitors. The crystal structure of TKI-trypsin and a 3D model of

TKI-FXa complex revealed that the distinct reactive site loop probably

plays a role in dual inhibition. The reactive site of TKI interacts with an

active site and two exosites (36 loop and autolysis loop) of FXa. Apart from

Arg66 (P1), Arg64 (P3) is one of the most important residues responsible

for the specificity of TKI towards FXa. Along with the reactive site loop

(b4b5), loops b1 and b7b8 also interact with FXa and could further confer

selectivity for FXa. We also present the role of inserted Asn15 in the stabil-

ization of complexes. To the best of our knowledge, this is the first structure

of FXa inhibitor belonging to the Kunitz type inhibitor family and its

unique structural and sequence features make TKI a novel potent inhibitor.

Database

The complete nucleotide of TKI was deposited in the NCBI gene databank with accession no.

HQ385502. The atomic coordinates and structure factor files for the structure of TKI and

TKI:PPT complex have been deposited in the Protein Data Bank with accession numbers

4AN6 and 4AN7, respectively

Structured digital abstract

� TKI and TKI bind by x-ray crystallography (View interaction)

� TKI and PPT bind by x-ray crystallography (View interaction)

Abbreviations

API-A, arrowhead protease inhibitor A; APTT, activated partial thromboplastin time; AT, antithrombin; BAPNA, N-benzoyl-L-arginine-p-

nitroanilide; BbCI, Bauhinia bauhinioides cruzipain inhibitor; BbKI, Bauhinia bauhinioides kallikrein inhibitor; BTPNA, N-benzoyl-L-tyrosyl-p-

nitroanilide; BuXI, Bauhinia ungulata factor Xa inhibitor; BvTI and BvvTI, Bauhinia variegata trypsin inhibitor; CTI, Copaifera langsdorffıi

trypsin inhibitor; DrTI, Delonix regia trypsin inhibitor; EcTI, Enterolobium contortisiliquum trypsin inhibitor; ETI, Erythriana caffra trypsin

inhibitor; FXa, factor Xa; MKMLP, Murraya koenigii miraculin-like protein; PPE, porcine pancreatic elastase; PPT, porcine pancreatic trypsin;

PT, prothrombin time; STI, soybean trypsin inhibitor; TAP, tick anticoagulant peptide; TFPI, tissue factor pathway inhibitor; TKI, tamarind

Kunitz type inhibitor; WCI, winged bean chymotrypsin inhibitor.

FEBS Journal 279 (2012) 4547–4564 ª 2012 The Authors Journal compilation ª 2012 FEBS 4547

Introduction

The formation and interactions of specific protein–pro-tein complexes have enormous importance in biological

processes [1,2]. Protein–protein complexes of serine

proteases and their inhibitors are extensively studied

models [1]. The Kunitz soybean trypsin inhibitor (STI)

superfamily of serine proteases has been widely studied.

The plant Kunitz type inhibitors are present in compar-

atively large quantities in the seeds of the Leguminosae

subfamilies, Mimosoideae, Caesalpinioideae and Papi-

lionoideae [3,4]. These inhibitors have molecular mass

of about 20 kDa with one or two polypeptide chains

and few cysteines [5–7]. Most of them are a single poly-

peptide chain with two disulfide bridges and a single

reactive site [4,8,9]. The Kunitz STIs are included as

Kunitz-P inhibitors in the I3A family of the MEROPS

database (http://merops.sanger.ac.uk) [10].

Several structures of Kunitz STI type inhibitors have

been reported showing that they have a b-trefoil foldconsisting of 12 anti-parallel b-strands connected by

long loops. Their exposed reactive site loop has a char-

acteristic canonical conformation with Arg/Lys at the

P1 position [6,11–15]. The conserved Asn (Asn13 in

STI) residue on loop b1 interacts with P2 and P1′ ofthe reactive site loop (loop b4b5) and maintains the

canonical conformation of the reactive site loop

[11,12,16]. The structure of the STI complex with por-

cine pancreatic trypsin (PPT) reveals that P1 residue

Arg63 enters into the S1 pocket of trypsin and makes a

salt-bridge with Asp189 of the enzyme and inhibits it in

a substrate-like manner [11]. Moreover, some structural

reports showed a variation in Kunitz type inhibitor

structures. For example, Copaifera langsdorffıi trypsin

inhibitor (CTI) has a b-trefoil fold composed of two

non-covalently bound polypeptide chains with only a

single disulfide bridge [17]. Delonix regia trypsin inhibi-

tor (DrTI) has one amino acid insertion between P1

and P2 of the reactive site distorting its conformation

[6]. Bauhinia bauhinioides cruzipain inhibitor (BbCI), a

Kunitz type inhibitor, has a conservative b-trefoil foldbut lacks disulfide bonds [18]. The crystal structure of

Murraya koenigii miraculin-like protein (MKMLP)

from seeds of Murraya koenigii with trypsin inhibitory

activity also shows a conservative b-trefoil fold but has

seven cysteines forming three disulfide bridges, Asn65

at P1 instead of Arg/Lys, Asn13 replaced by Ala13 and

carbohydrate moieties linked to Asn64 [19]. Arrowhead

protease inhibitor A (API-A), a double-headed arrow-

head protease inhibitor similar to Kunitz STI inhibitor,

has a b-trefoil fold but possesses two reactive sites [20].

The plant Kunitz type inhibitors are not restricted to

trypsin inhibition only, but also show inhibition of

several serine and other proteases. For example, beyond

the trypsin inhibitory activity of STI, it acts as an anti-

inflammatory by inhibiting human neutrophil elastase

[21] and has an anti-invasive property on ovarian cancer

cells [22]. Enterolobium contortisiliquum trypsin inhibi-

tor (EcTI) inhibits trypsin, chymotrypsin, factor XIIa,

plasma kallikrein, human neutrophil elastase, plasmin

and the invasion of gastric cancer cells [23–25]. Bauhi-nia ungulata factor Xa inhibitor (BuXI) blocks the

activity of trypsin, chymotrypsin, plasma kallikrein,

plasmin, factor XIIa and factor Xa (FXa) [26]. Bauhi-

nia bauhinioides kallikrein inhibitor (BbKI) inhibits

trypsin, chymotrypsin, plasma kallikrein and plasmin

[27,28]. BbCI shows inhibitory properties against

enzymes of two different classes, cysteine and serine

proteases [28,29]. Bauhinia variegata trypsin inhibitor

(BvvTI) has trypsin inhibitory, anti-HIV-1-RT activity

and inhibits the proliferation of nasopharyngeal cancer

CNE-1 cells [30]. These several plant Kunitz type inhibi-

tors are important molecules to gain knowledge of the

basic principle of protein–protein interactions and

might have applications in pharmaceuticals as potential

drugs. So, it is necessary to search for novel multivalent

Kunitz inhibitors having pharmaceutical significance.

We have previously reported purification of tama-

rind Kunitz STI type inhibitor (TKI) and its prelimin-

ary crystallographic study as well its complex

formation with trypsin and the crystallization of the

complex [31,32]. TKI was also reported as an in vitro

and in vivo potential bio-insecticide against different

insect pests and a weak elastase inhibitor and was

shown to be a non-competitive inhibitor [33]. How-

ever, in the present study, we show that TKI inhibits

FXa and trypsin competitively. The inhibition con-

stants (Ki) of TKI for FXa and trypsin were calcu-

lated. TKI was also assayed for its anticoagulant

property. To gain structural insights into a dual inhibi-

tory role of this novel TKI, cDNA cloning and the

detailed 3D crystal structures of free TKI and its com-

plex with trypsin were determined and described.

Moreover, the complex with FXa has been modelled

to study the mode of inhibition of FXa by TKI. Fur-

thermore, we describe the role of inserted residue

Asn15 in the stabilization of the complexes.

Results

cDNA cloning of TKI and sequence analysis

The first strand of cDNA was obtained from highly

pure RNA. It was amplified using forward degenerate

4548 FEBS Journal 279 (2012) 4547–4564 ª 2012 The Authors Journal compilation ª 2012 FEBS

TKI diverse reactive site role in dual inhibition D. N. Patil et al.

and oligo(dT)18 primers. An approximately 700 bp

fragment was obtained, which included the 3′ UTR

region with poly(A+) tail. A 555 bp ORF was

obtained after sequencing of the TKI gene, which

coded for a polypeptide of 185 amino acids with a cal-

culated molecular mass of 20 575 Da. The sequence of

TKI showed significant similarity to reported Kunitz

type inhibitors in an NCBI BLAST search, which dem-

onstrates that TKI belongs to the soybean Kunitz fam-

ily inhibitors. The primary amino acid sequence of

TKI showed that it has a distorted Kunitz signature

motif ([LIVM]-x-D-x-[EDNTY]-[DG]-[RKHDENQ]-x-

[LIVM]-(x)5-Y-x-[LIVM]). The Kunitz signature motif

is found in the soybean Kunitz type trypsin inhibitor

family and in TKI the observed pattern is

VHDTDGKPVLNNAGQYYI which is located at the

N-terminal of the TKI sequence (Fig. 1). The sequence

analysis revealed that the distorted Kunitz signature

sequence was due to insertion of Asn at position 15.

The insertion of Asn is also observed in Bauhinia var-

iegata trypsin inhibitor (BvTI) (Fig. 1). A protein

sequence BLAST with non-redundant database showed

that TKI shares 45%, 44%, 42%, 39% and 33%

sequence identity with BuXI [26], EcTI [25], BvTI [34],

Erythriana caffra trypsin inhibitor (ETI) [12] and STI

[11] respectively. The chain length of these proteins

ranges from 172 to 185 residues. The four cysteines

are linked by two disulfide bonds in TKI as in ETI

and STI (Cys 42–86, Cys 134–144). Multiple sequence

alignment of the representative Kunitz type inhibitors

from different plants revealed that the TKI sequence

forms a close relationship with classical Kunitz family

members and the closest relationship with BuXI as

shown in Fig. 1.

Inhibitory properties and Ki determination

The inhibitory activity of TKI was determined against

trypsin, chymotrypsin, elastase and FXa by measuring

the hydrolytic activity toward N-benzoyl-L-arginine-p-

nitroanilide (BAPNA),N-benzoyl-L-tyrosyl-p-nitroanilide

(BTPNA), N-Suc-(Ala)3-nitroanilide and CH3OCO-D-

CHA-Gly-Arg-pNA-AcOH, respectively. TKI showed

inhibition of trypsin and FXa but did not show any sig-

nificant inhibitory activity against chymotrypsin and

showed very weak inhibition of elastase (data not

shown). The dissociation constants (Ki) for TKI

against trypsin (Fig. 2B) and FXa (Fig. 2C) were

determined from a Dixon plot. Analysis of the Dixon

plot revealed that TKI is a competitive inhibitor with

Ki value 3.2 9 10�9M and 2.2 9 10�7

M for trypsin

and FXa activity, respectively, which clearly explains

that TKI is a potent inhibitor of porcine trypsin and

human FXa. The inhibition of human FXa by TKI

indicated that TKI could inhibit blood coagulation.

Therefore the anticoagulation property of purified

TKI was tested. The incubation of increasing concen-

trations of TKI with fresh human plasma showed

Fig. 1. Sequence analysis of TKI. Multiple sequence alignment of TKI with other Kunitz STIs. Arrows indicate b sheets and TT turns.

Absolute conserved residues among all proteins are shown with a black background. Cysteine residues involved in formation of disulfide

bonds are represented with grey numbering. The insertion of one residue between the Kunitz signature sequence (represented as KSS) is

shown by a star and the reactive site is shown as RS in a rectangular box.


D. N. Patil et al. TKI diverse reactive site role in dual inhibition

significant increase in activated partial thromboplastin

time (APTT) and prothrombin time (PT). TKI (10 lM)extended the APTT of normal plasma 2.6-fold and of

PT 5-fold (Fig. 2A).

Quality and overall structure of TKI

The crystallographic data and refinement statistics for

the free TKI structure are summarized in Table 1. The

TKI structure was determined at 1.94 �A resolution

and contains two molecules per asymmetric unit (TKI

A and TKI B). The free TKI structure was refined to

an Rcryst of 19.04% and Rfree of 24.02%. The overall

model is well defined by electron density; however, the

electron densities of residues 1, 153–155 in chain A,

residues 1, 138–141 in chain B and C-terminal residues

(177–185) from both the chains were not observed. A

section (residues 122–130) of the final 2F0 � Fc

electron density map is shown in Fig. 3C. The final

model consists of 343 residues and 157 solvent mole-

cules in both the molecules. The Ramachandran plot

analysis was done using MOLPROBITY server [35]. The

A

B

C

Fig. 2. (A) Anticoagulation activity of TKI. For blood coagulation

time measurements APTT (black) and PT (grey) were determined

by a standard procedure. (B), (C) Dixon plots for the determination

of the dissociation constant (Ki) value of TKI against (B) trypsin and

(C) FXa. The enzyme assays were carried out at two different

concentrations of substrate for trypsin and FXa, respectively. The

reciprocal of velocity was plotted against different concentrations

of TKI and the Ki value was determined from the intersection of

the two regression lines. All the experiments were done in

triplicate and average values were used.

Table 1. Crystallographic data and refinement statistics for free TKI

and its complex with PPT. Rsym = RhklRni¼1jIhkl ;i ��Ihkl j=RhklR

ni¼1Ihkl ;i .

The values on the Ramachandran plot were obtained with the

MOLPROBITY server [35].

Free TKI

TKI–PPT

complex

Crystallographic data

Space group P212121 P21221

Wavelength 1.54179 1.54179

Resolution 50.0–1.93 91.6–2.22

Cell dimensions

a (�A) 40.43 61.08

b (�A) 60.42 67.12

c (�A) 105.53 91.61

Unique reflections 19 581 17 821

Completeness (%)

(last shell)

97.5 (75.7) 92.9 (40.3)

Rsym (%) (last shell) 4.3 (25.7) 7.6 (40.0)

I/r (last shell) 19.4 (3.6) 11.86 (2.0)

Multiplicity (last shell) 3.0 (2.4) 3.3 (2.0)

Refinement 29.05–1.94 45.22–2.23

No. of reflections

(working/test)

17 690 (16726/964) 16 883 (15972/911)

No. of residues 343 379

Water molecules 157 156

Resolution range (�A) 1.94 2.23

Rcryst (%) 19.04 19.95

Rfree (%) 24.02 24.58

Average B factors (�A2) A 21.95 A 26.71

B 23.09 B 32.78

Water atoms 36.15 41.93

All atoms 23.27 29.95

Rmsd on bond lengths (�A) 0.008 0.009

Rmsd on bond angles (�A) 1.43 1.36

Ramachandran plot (%)

Favoured 96.1 95.6

Allowed 3.6 4.1

Outliers 0.3 0.3



refined model shows that 96.1% of all residues are in

favoured regions and 99.7% of all residues are in the

allowed regions (Table 1).

The overall structure of TKI is similar to that of

other Kunitz type inhibitors. It is devoid of helical

structures and is composed of b-strands and turns.

The protein shows a b-trefoil fold consisting of 12

anti-parallel b-strands connected with long loops simi-

lar to STI, an archeal member of the Kunitz type

trypsin inhibitor family [36,37]. These b-strands form

six double stranded b-hairpins. Three b-hairpins form

the barrel structure and the remaining form a cap-like

structure on the barrel (Fig. 3A). TKI consists of two

disulfide bonds, Cys42-Cys86 and Cys134-Cys144, at

the surface that are also conserved in other Kunitz

type inhibitors stabilizing the 3D structure [11–13].The TKI structure does not have a 310-helix similar

to STI due to deletion of three amino acids at the

respective positions. This type of deletion is also

observed in other Kunitz type inhibitors like ETI,

winged bean chymotrypsin inhibitor (WCI) and DrTI

[6,12,13].

A B

C

Fig. 3. (A) Overall fold of TKI shown in a cartoon diagram. Magenta and cyan represent b-strands of the lid and b-barrel respectively. Two

disulfide bonds (C42–C86, C134–C144) are shown in green. The reactive site loop P4–P3′ (Ser63–His69) is represented in blue showing that

residue Arg66 is P1. The structure was submitted to the PDB database (PDB ID code 4AN6). The architecture of the reactive site loop of

TKI with the networks of hydrogen bond interactions around reactive site residues is highlighted. Reactive site loop residues are shown as

cyan carbon sticks. The vital residue Asn14 which holds the canonical conformation is shown as a light pink carbon stick. The black dashed

lines represent hydrogen bonds and water molecules are shown as red spheres. (B) Conformationally variable loops of TKI A and TKI B are

shown by stars. The variable reactive site loop is highlighted showing the difference in P1, P3 and P3′ residues. (C) Stereo view of the final

2F0 � Fc electron density map for a section (residues 122–130) of free TKI structure.



The geometry and conformation of the reactive

site loop

The TKI reactive site loop P4–P3′ (Ser63-Arg64-

Ala65-Arg66-Ile67-Ser68-His69) is devoid of secondary

structures and disulfide bonds. The Arg66 present at

P1, which is identical to its closely related ETI and

STI, is defined well in the electron density map. The

reactive site loop of TKI has an overall high B factor

for P1–P4 residues. The side chains of residues at P1,

P3 and P3′ are flexible and may fit optimally into the

active site of cognate proteases by adapting the confor-

mation. The Asn14 present on loop b1, which is a con-

served residue in the STI family (Asn13 in STI),

stabilizes the reactive site loop (loop b4b5) in a similar

manner as was shown in the other member of the STI

family [11,16]. The ND2 atom of Asn14 forms hydro-

gen bonds with the P2 and P1′ carbonyl O which are

present at both sides of the scissile bond. Asn14 also

interacts with P4 and P2′ and maintains the canonical

conformation of the reactive site loop, which is a char-

acteristic feature of this family. This conserved Asn is

called a spacer residue and is also conserved in the

other serine protease inhibitor families such as Kazal,

ecotin, Streptomyces subtilisin inhibitor, grasshopper

and potato II [16].

Comparison of free TKI with homologous

structures

The structural studies demonstrated that TKI adopted

an overall b-trefoil fold with two disulfide bonds simi-

lar to the Kunitz inhibitors. Although the overall fold

of TKI showed structural similarity with Kunitz STIs,

the two molecules in the asymmetric unit of free TKI

show variations in loop regions (Fig. 3B). Two of the

connecting loops (L6 residues 95–108 and L7 residues

112–119) of TKI show a high degree of deviation

among themselves. Apart from connecting loops, the

two b sheets B1 (129–137) and B2 (140–150) exhibit

different orientations between TKI A and TKI B.

ETI [12], STI [11], CTI [17], WCI [13] and DrTI [6]

share closest structural similarity with TKI. TKI exhib-

its 39%, 33% and 31% sequence identity with ETI, STI

and WCI respectively. The superposition of TKI with

ETI, STI and WCI shows rmsd values of 1.16 �A (for

117 Ca atoms), 0.95 �A (for 113 Ca atoms) and 1.5 �A

(for 123 Ca atoms), respectively. The superposition

of Ca atoms of the reactive site loop residues of TKI

(P4–P3′) with ETI, STI and WCI exhibit rmsd values of

0.8 �A, 0.7 �A and 0.7 �A respectively. The superposition

of the reactive site loop residues of TKI A with TKI B

shows different conformations with rmsd of 0.8 �A.

CTI, an inhibitor of the Kunitz family different

from classical inhibitors [17], shares 41% sequence

identity with TKI and superposition of TKI with CTI

gives an rmsd of 0.55 �A for 117 Ca atoms. DrTI is a

Kunitz family member with a distorted reactive site

loop [6] and shows 36% sequence identity with TKI

and an rmsd of 1.02 �A for 107 Ca atoms.

Overall structure of TKI–PPT complex

The structure of TKI–PPT complex was determined at

2.23 �A resolution and refined to an Rcryst of 19.95%

and Rfree of 24.58%. The refined model of the complex

consists of one molecule per asymmetric unit. The

TKI–PPT model lacks three segments (residues 26–28,96–104 and 136–141), one N-terminal residue and 10

residues of the C-terminal of TKI in the electron den-

sity. The complex structure consists of 379 residues,

156 solvent molecules and a Ca2+ atom. The refined

model shows that 95.6% of all residues are in favoured

regions and 99.7% of all residues are in the allowed

regions of the Ramachandran plot (Table 1).

The overall structure of TKI–PPT complex is similar

to that of STI–PPT. TKI blocks the active site of PPT

and P1 residue forms a hydrogen bond network with

S1 pocket residues. After binding to PPT, a conforma-

tional change at the P1 (Arg66) position and also

reduction in the temperature factor of the reactive site

loop were observed. One Ca2+ atom is observed in the

complex model which interacts with OE1 and OE2 of

Glu70, carbonyl O of Asn72 and Val75 and OE2 of

Glu80 residues of PPT. This Ca2+ atom is also

observed at a similar position in the STI–TKI model.

The overall complex structure and detailed interaction

of TKI with PPT is shown in Fig. 4. The 2F0 � Fc

electron density map clearly shows that the reactive

site loop of TKI enters into the S1 pocket and inter-

acts with Asp189 of PPT (Fig. 5).

Comparison of free TKI and TKI–PPT complex

structure

The comparison between the free TKI and TKI–PPTcomplex shows slight conformational changes in the

reactive site loop. The superposition of the free TKI A

and TKI B with the TKI–PPT gave an rmsd of 0.4 �A

(for 123 Ca atoms) and 0.5 �A (for 131 Ca atoms)

respectively and its reactive site loop showed decreased

B factors compared with both TKI A and TKI B. The

superimposition of TKI–PPT complex with TKI A

and TKI B is shown in Fig. 6. The reactive site loop

of TKI A and TKI B is superimposed onto the TKI–PPT complex showing rmsd of 0.59 �A and 0.6 �A



respectively. The flexible P1 residue (Arg66) of free

TKI becomes more rigid with lower B factor in the

TKI–PPT complex structure. This illustrates that P1

stabilizes the complex by making interactions with res-

idues of the S1 pocket and other residues of PPT. The

superposition of the reactive site loop of TKI–PPTcomplex with TKI A and TKI B discloses that there is

immense divergence at position P1. The distance

between Ca of P1 in TKI–PPT and TKI A and TKI B

is 1.4 �A and 0.8 �A respectively. The difference is also

examined at the side chain of P1. The side chain of

Arg66 of TKI in the complex structure is � 5 �A and

4 �A apart from the side chain of Arg66 of TKI A and

TKI B respectively (Fig. 6A). Another residue

(Arg64), P3 of the reactive site, also shows conforma-

tional variation. P3 becomes rigid after forming a

complex with trypsin and participates in stabilization

of the complex by interacting with residues of PPT

(Fig. 6B). Compared with free TKI, His69 of the reac-

tive site loop (P3′) of TKI–PPT shows conformational

rotation. It moves away from Ser68 to prevent short

contact and acquires a position to interact with Ser39

and His40 of PPT through a water molecule (W132),

which indicates that it may help in stabilization of the

complex (Fig. 6C).

Comparison of TKI–PPT with other Kunitz type

inhibitor complexes with trypsin

The superposition of TKI–PPT on STI–PPT gives an

rmsd of 0.82 �A for 311 Ca atoms and their reactive

site loop gives an rmsd of 0.46 �A. This is smaller than

the rmsd between the reactive site loop of free STI and

TKI models. The interface between TKI and trypsin

covers an area of about 1025 �A2 of TKI which is lar-

ger than the corresponding surface of STI (937 �A2) in

the STI–PPT complex as calculated by the program

PDBSUM [38]. There are notable differences in the inter-

actions of the interacting residues of TKI with PPT

compared with the STI–PPT structure. The interac-

tions of Asn15, Arg64 (P3), Ser68 (P2′), His69 (P3′)and Ser116 of TKI with PPT make it different from

the STI–PPT complex. The details of the interaction of

TKI–PPT are discussed in the next section.

TKI also shares 31% sequence identity with Kunitz

type WCI [13]. The complex structures of WCI mutant

F64Y/L65R (PDB ID 3I29) and L68R (PDB ID

2QYI) with bovine trypsin superimposed on the TKI–PPT structure shows an rmsd of 1.5 �A (for 323 Ca

atoms) and 1.8 �A (for 335 Ca atoms). The interactions

of residue Arg at P1 are similar. Double-headed API-

A has less sequence identity (28%) with TKI and its

complex with two trypsins (PDB ID 3E8L) superim-

posed with TKI–PPT shows an rmsd of 1.1 �A for 230

Ca atoms [20].

Mode of interaction between TKI and PPT

In the TKI–PPT structure, six residues of TKI (Thr3,

Arg64 (P3), Arg66 (P1), Ser68 (P2′), Ser75 and

Ser116) interact with PPT residues directly and four

residues of TKI [Asn15, Gly17, Arg60 and His69

(P3′)] interact through water with residues of PPT

(Fig. 4B). A total of 10 residues of TKI interact with

20 residues of PPT forming a dense network of

hydrogen bonds to produce a stable complex. The

A

B

Fig. 4. (A) Overall structure of the TKI–PPT complex. TKI and PPT

are shown in red and blue respectively. Residue P1 enters the S1

pocket and interacts with residues of PPT (purple dashes); one Ca2

+ atom is also observed (green) which interacts with PPT residues

(purple dashes). (B) Detailed hydrogen bond interactions between

TKI reactive site residues and PPT residues. Interacting residues of

TKI are labelled in red. The residues of PPT are labelled in blue.

Water molecules are in orange and hydrogen bonds are

represented as black dashes.



detailed interactions between TKI and PPT are shown

in Fig. 4 and Table 2. There are 21 hydrogen bonds

between TKI and PPT involving four residues of the

reactive site. The P1 residue Arg66 interacts with

PPT and forms a large number of comprehensive

hydrogen bonds (Table 2). The side chain of Arg66

enters into the S1 pocket of PPT and interacts with

residues of the S1 pocket. The guanidinium group of

Arg66 forms a salt-bridge with the carboxylate group

of Asp189 of PPT. Arg66 of TKI interacts with

Asp189, Ser190, Gly193, Ser195, Ser214 and Gly219

of PPT. Arg66 also interacts with Gln192 and Gly216

of PPT through water (W83). The scissile bond was

found to be uninterrupted in electron density. The P3

A B C

Fig. 6. Comparison of free TKI and the TKI–PPT complex structure. Superposition of free TKI A and TKI B with TKI–PPT, represented in

blue, cyan and red respectively, shows variation at the reactive site loop. Conformation of P1 and P3 residue is highlighted and shows

variation of Ca as well as NH1 and NH2 atoms (A, B). Additionally, the conformational rotation of P3′ residue, His69, is also observed. It

moves to a position from where it interacts with Ser39 and His40 of PPT through a water molecule (C).

Fig. 5. Stereo view of the 2F0 � Fc

electron density map around the

interaction of TKI with PPT of the TKI–PPT

complex structure contoured at 1.0r. The

reactive site loop residues of TKI and the

S1 pocket residues of PPT are shown as

red and blue sticks respectively and the

disulfide bond is yellow. Hydrogen bonds

are represented as red dashed lines.



residue Arg64 interacts with Gly216 and Asn97 of

PPT. Arg64 also interacts with Thr98, Gln175,

Gln192 and Gly218 through water. Ser68 at the P2′position also participates in making hydrogen bonds

with His40 CO and Phe41 of PPT. The P3′ residue

His69 shows a slight conformational change com-

pared with free TKI where it interacts with Ser39 and

His40 of PPT.

Molecular docking of TKI with FXa

Rigid body docking was performed using program

CLUSPRO. Figure 7 shows the detailed interactions of

the best model of the TKI–FXa complex taken from

the 30 highest ranking structures calculated by CLUS-

PRO. The docking studies of TKI A and TKI B with

FXa show interactions with the S1 and S4 pocket of

FXa in a classical L-shaped substrate-like conforma-

tion (Fig. 7B). Our studies of TKI–FXa show that

TKI B has a greater number of favourable stabilizing

interactions than TKI A and the total buried surface

of TKI B between the TKI–FXa complex is 1253 �A2,

which is more than for TKI A as calculated by the

program PDBSUM [38]. Various types of interaction of

residues of TKI with FXa residues were calculated by

the PIC server. Mainly, residues from three loops (loop

b1, loop b4b5 and loop b7b8) interact with FXa. Resi-

dues Gln18, Arg60, Arg66, His69, Glu79 and Lys136

of TKI form salt-bridges with Lys148, Glu147,

Asp189, Glu39, Arg222 and Glu97 of FXa, respec-

tively (Table 3). Of the seven residues of the reactive

site loop, six residues interact with FXa residues and

participate in stabilization of the complex. The

detailed interactions are shown in Table 3 and

Fig. 7C. The Arg66 (P1) of TKI intrudes inside the

substrate binding pocket S1 of FXa making direct

interactions with residues of the S1 pocket.

The positively charged side chain of Arg64 (P3) of

TKI accesses the S4 pocket of FXa parallel to the

p-electron-rich aromatic ring of Phe174 of FXa and

may participate in a cation–p interaction (Fig. 7B).

The cation–p interaction can take place in parallel, in

perpendicular or in a T-shaped orientation [39,40]. The

PIC server predicts a cation–p interaction with Tyr99,

Phe174 and Trp215 of the S4 pocket. Arg64 further

stabilizes the complex by forming hydrogen bonds

with Glu97, Thr98 and Gly216. One of the residues

Gln192, reported as a vital residue in FXa for specific-

ity and binding [41,42], interacts with four important

residues of TKI forming six hydrogen bonds (Table 3).

Two of them are reactive site loop residues Ser63 (P4)

and Ala65 (P2); one is a spacer residue Asn14 and one

is an insertion residue Asn15 which is observed in TKI

only. The residues from some of the loops also take

part in stabilization of the complex, such as Ala73 and

Ser75 of the b4b5 loop, Ser116 and Ala118 of loop

b7b8 and Asn14, Asn15, Gly17 and Gln18 of loop b1.

Discussion

A Kunitz type protease inhibitor from tamarind seeds

has been cloned and characterized both biochemically

Table 2. Total interactions of residues of TKI with PPT.

TKI PPT Water Distance (�A)

Thr3 OG1 Lys60 NZ 2.8

Asn15 NH2 and CO W25 3.0, 3.3

Gln192 NE2 2.6

NH2 and CO W25 3.0, 3.3

Asn143 ND2 3.2

NH2 and CO W25 3.0, 3.3

W154 3.0

Lys145 CO 3.0

W25 3.0, 3.0

W154 3.0

Gly148 CO 3.5

Gly17 N W140 3.2

Ser149 OG 3.0

Arg60 NH2 W140 3.4

Ser149 OG 3.0

Arg64 NH2 Asn97 CO 2.8

CO Gly216 N 3.1

NH1 and NH2 W151 3.0, 3.4

Thr98 CO 2.7

NH1 and NH2 W151 3.0, 3.4

Gln175 CO 2.7

CO W83 3.2

Gln192 OE1 2.9

CO W83 3.2

Gly219 CO 2.8

CO Gly193 N 2.7

Arg66 CO Ser195 N 2.9

N OG 3.0

N Ser214 CO 3.0

NH1 Ser190 OG 2.9

NH1 CO 3.0

NH1 Asp189 OD1 2.9

NH2 OD2 2.8

NH2 Gly219 CO 3.2

NE W83 3.2

Gly216 CO 2.9

NE W83 3.2

Gln192 OE1 2.9

Ser68 OG His40 CO 3.5

N Phe41 CO 3.2

His69 ND1 W132 2.6

Ser39 OG 2.9

ND1 W132 2.6

His40 CO 3.0

Ser75 OG Asn97 ND2 3.3

Ser116 OG Tyr217 OH 3.4



and structurally. In this report, the dual protease

inhibitor activities of the purified TKI against human

FXa and trypsin have been investigated. The 3D

structures of TKI and TKI–trypsin complex were

determined. Additionally, a homology model of the

TKI–FXa complex structure was constructed. The 3D

structure of TKI shows a b-trefoil fold similar to other

Kunitz STI type inhibitors. However, there are many

distinct features of the TKI structure and interactions

of the TKI reactive site with FXa and trypsin which

make it a novel Kunitz type inhibitor.

Based on sequence homology, TKI has been classi-

fied as a Kunitz type protease inhibitor. The atomic

structure of TKI illustrates that it possesses a b-trefoilfold similar to Kunitz (STI) type inhibitors which

makes TKI a member of the Kunitz type inhibitor

family. The TKI–trypsin complex was crystallized and

its structure was determined to reveal the key

structural features of TKI important for trypsin inhibi-

tion. Additionally, biochemical studies showed that

TKI is a potent competitive inhibitor of FXa and

trypsin with equilibrium dissociation constants (Ki) in

the nano-molar range. The competitive inhibition of

trypsin by TKI has been further confirmed by a struc-

tural study of the TKI–trypsin complex, which clearly

shows the binding of the TKI reactive site loop in the

active site of trypsin. This inhibitory study of TKI

towards trypsin is contrary to the previously published

report which showed that TKI was a non-competitive

inhibitor of trypsin [33].

Analysis of the TKI reactive site showed that it has

scant sequence identity with other Kunitz type inhibi-

tors (Fig. 8A) as well as FXa inhibitors. Comparison

of the reactive site of TKI with that of other physio-

logical or non-physiological inhibitors of FXa such as

antithrombin (AT) [43], protein Z-dependent protease

inhibitor [44], tissue factor pathway inhibitor (TFPI)

[45], BuXI [26], antistasin [46] and tick anticoagulant

peptide (TAP) [47] shows that the reactive site loop

does not have much similarity with these inhibitors

except Arg at P1 as in AT, BuXI and antistasin and

Arg at P3 as in TAP and antistasin that has P1 Arg

and P3 Arg residues. This comparison reflects the

uniqueness of the TKI reactive site loop. Furthermore,

the interactions of distinct reactive site residues of TKI

with trypsin in the complex of TKI–PPT are responsi-

ble for the difference in stabilization pattern compared

with the STI–PPT complex (Table 2).

Comparison of the model of the TKI–FXa complex

with the TKI–PPT complex structure reveals the

structural features of TKI that could be accountable

for the dual inhibitory property. The inhibition of

FXa as well as trypsin is probably due to the unique

reactive site which is flexible and can have diverse

conformations. This is evident from the two molecules

TKI A and TKI B present in the asymmetric unit

A

C

B

Fig. 7. Molecular docking of TKI B with

FXa. (A) Overall fold of TKI B–FXa showing

P1 blocking the S1 pocket of FXa. TKI B

and FXa are shown in red and cyan

respectively. (B) The surface view of

TKI–FXa complex showing Arg66 and

Arg64 of TKI B intruding inside the S1 and

S4 pocket of FXa respectively in an

L-shaped substrate like manner. (C) The

detailed hydrogen bond interactions in the

TKI B–FXa docked model are represented

as black dashes and red and cyan carbon

sticks for TKI B and FXa residues

respectively.



that show structural variation in the reactive site

loop. Molecular modelling studies show that six resi-

dues of the reactive site of TKI interact with S1, S4

and two secondary binding site/exosites (acidic 36

loop and autolysis loop) residues and provide specific-

ity for FXa (Table 3). The serine residue at P4 may

be crucial in FXa complex formation as it interacts

with two vital residues, Tyr99 and Gln192, which

define specificity for FXa [41,42,48]. There is an

uncommon residue Arg at the P3 position in Kunitz

STI inhibitors which occupies the S4 hydrophobic

pocket, the prime determinant site for specificity of

FXa [49,50], and interacts with residues of the cation-

binding hole of S4. The side chain of Arg64 is paral-

lel to the hydrophobic ring of Phe174 giving rise to

the possibility of forming the cation–p interaction

(Fig. 7B). A similar role of Arg at P3 was observed

in the TAP–FXa complex structure [47] and in the

modelled antistasin–FXa complex [46]. These observa-

tions support the hypothesis that the Arg64 (P3) of

TKI could be one of the most significant residues,

which provides the specificity towards FXa. The resi-

due Arg at P1 which is observed in most Kunitz

inhibitors is able to make contact with S1 pocket resi-

dues of FXa in a similar way as in trypsin.

Another two residues Ser and His at P2′ and P3′that interact with exosites of FXa are non-polar in

other Kunitz type inhibitors such as STI, ETI, DrTI,

CTI, BuXI, BvTI, EcTI and WCI. Ser68 interacts with

acidic 36 loop and autolysis loop residues of FXa.

His69 makes a salt-bridge with acidic 36 loop residue

(Table 3). It was reported earlier that AT forms salt-

bridges and hydrogen bonds with the 36 loop and the

autolysis loop residues of FXa [43]. In the TAP–FXa

complex structure, residues of TAP form hydrogen

bonds and salt-bridges with the positively charged resi-

dues of the autolysis loop [47]. Similar interactions

were also shown in the modelled complex of antistasin

–FXa [46] and in the second Kunitz domain of TFPI–FXa [45]. Therefore Ser and His could also be the key

determinants towards the specificity for FXa. These

observations indicate that reactive site residues of TKI

favour formation of a stable complex with FXa. Muta-

genesis studies will further enlighten the effects on

inhibitory activity and will provide an approach to

designing an improved novel inhibitor that can act as

anticoagulant in vivo. Along with Arg64 (P3), other

residues such as Ser63, Ala73, Ser75, Ala118 and

Lys136 of TKI also interact with residues of S4 and

could further stabilize the complex. Such types of

interaction are not observed in the TKI–PPT complex

as ‘96-KETY-99′ in FXa for specificity is unavailable

in PPT (see Fig. 8B). TKI inhibits both proteases

despite less sequence identity (37%) of human FXa

with porcine trypsin, as it has specificity for the S4

pocket of FXa which is not observed in most Kunitz

STIs.

Table 3. TKI interactions with FXa.

TKI B FXa d (�A)

(A) Salt-bridges

Gln18 OE1 Lys148 NZ 2.6

Arg60 NH2 Glu147 OE1 3.9

NH2 OE2 2.7

NH1 OE2 2.7

Arg66 NH2 Asp189 OD1 3.3

NH2 OD2 2.6

NH1 OD1 3.6

His69 ND1 Glu39 OE1 3.0

Glu79 OE2 Arg222 NH1 2.8

Lys136 NZ Glu97 OE2 2.7

(B) Reactive site loop (P4–P3′) interactions

Ser63 OG Tyr99 OH 3.4

CO Gln192 NE2 3.2

Arg64 NH1 Glu 97 CO 2.7

NH2 Thr 98 CO 2.6

CO Gly216 N 3.0

Phe174 Cation–p interaction

Ala65 CO Gln192 NE2 3.0

Arg66 NH1 Asp 189 OD1 3.5

NH2 OD2 2.7

NH2 OD1 3.3

NH1 Ala 190 CO 3.1

NH2 CO 2.7

CO Gly193 N 2.7

CO Asp194 N 3.2

CO Ser 195 N 2.9

N OG 3.3

N Ser 214 CO 3.3

NH2 Gly218 CO 2.7

NE CO 2.9

Ser68 OG Glu39 OE2 3.0

OG Phe41 CO 3.3

CO Arg143 NH2 2.7

His69 ND1 Glu39 OE2 2.9

(C) Gln192 of FXa interaction with TKI residues

Gln192 NE2 Ser63 CO 3.2

Gln192 NE2 Ala65 CO 3.0

Gln192 OE1 Asn14 CO 3.5

Gln192 OE1 Asn15 ND2 2.9

NE2 OD1 2.9

OE1 N 2.7

(D) Additional interactions

Ala73 CO Lys96 NZ 2.8

Ser75 OG Glu97 OE1 2.9

Ser116 OG Ser173 CO 2.8

Ala118 N Glu97 OE2 3.1

Asn15 ND2 Glu147 OE2 3.2

Gly17 CO Lys148 NZ 2.6



Besides the versatile reactive site, two other distinct

features of TKI include insertion of the residue Asn15

in the Kunitz signature motif and a flexible loop b7b8.The inserted residue disrupts the Kunitz signature

motif in TKI. The crystal structure of free TKI shows

that Asn15 (loop b1) has different conformations in

the TKI A and TKI B molecules. The Ca distance

between Asn15 of TKI A and TKI B is 0.9 �A and

their side chains are 2.8 �A apart from each other

(Fig. 9A). The superposition of TKI A and TKI B on

STI and ETI showed that Asn15 is not superimposed

with any residue of ETI and STI; on the contrary it

points outwards from an exposed loop at the protein

surface. The possible role of this inserted residue lies

in the TKI–PPT complex structure and the modelled

TKI–FXa complex. In TKI–PPT complex structure,

Asn15 forms a hydrogen bond with Gln192 and

Asn143 through W25 and further with Lys145 and

Gly148 via a second water molecule (W154). Conse-

quently, the formation of a network of hydrogen

bonds by Asn15 through two water molecules contrib-

utes to a stronger interaction of TKI with PPT and

hence the stabilization of the TKI–PPT complex

(Fig. 9C and Table 2). Moreover, in modelled TKI–FXa complex, Asn15 again plays a vital role as it

forms three hydrogen bonds with Gln192 of FXa, a

key determinant for specificity of substrate/inhibitor

[41,42] (Fig. 9D). The interaction of Asn15 with

Glu147 of the autolysis loop is also helpful for provid-

ing stability with FXa (Table 3). Together, these

results indicate that TKI has a key residue insertion

that could be important in recognition by its proteases

and stabilization of the complexes. CTI also has an

insertion of Asp at the same place as Asn (Fig. 9B);

however, the implication of this insertion in CTI has

not been reported.

Superposition of TKI A on TKI B shows conforma-

tional variation at the loop b7b8 (Asp112-Pro119).

This loop adopts a unique conformation compared

with the loop of STI. The loop b7b8 of TKI is closer

Fig. 8. (A) Web logo representation showing the variation in residue reactive sites in TKI compared with other Kunitz STI type inhibitors

[76]. The y axis represents sequence conservation in bits and the x axis represents the reactive site (P4–P3′). The respective reactive site

residues of TKI are shown below the x axis. (B) Interaction of residues of TKI with residues of porcine trypsin and FXa are compared

(chymotrypsin numbering) and shown in a grey background. The specificity of TKI towards FXa may be obtained due to the interaction of

TKI with S4 and two exosites, namely autolysis and acidic 36 loop (shown in an oval shape), and the difference in FXa and trypsin residues

at this site gives the selectivity towards them. The aromatic S4 pocket formed by residues Tyr99 (Leu99), Phe174 (Gly174) and Trp215

(Trp215) of FXa, which may form a cation–p interaction with the substrate/inhibitor, is absent in trypsin (respective residues of trypsin are in

parentheses). The S1 pocket is shown as a square which has the same residues except Ala190 in FXa instead of Ser190 of trypsin.



to the reactive site loop and its flexible nature indicates

that it may interact with residues of proteases to which

TKI binds and stabilizes the complex. In the TKI–PPT complex structure, Ser116 interacts with residue

Tyr217 of PPT. These observations suggest that the

flexible loop b7b8 probably plays an important role in

the formation of a complex. However, the interaction

of the b7b8 loop with PPT was not observed in STI–PPT, which indicates that it may not be very impor-

tant in complex formation with PPT but it might be

crucial for stabilization of complexes with other serine

proteases. For example, in the modelled TKI–FXa

complex, two residues of this loop interact with FXa

residues (Table 3). One of them is Glu97 of the S4

pocket of FXa which demonstrates that this loop may

be crucial in complex formation with FXa and could

enhance the affinity of TKI towards FXa. So, conclu-

sively, residues of loop b1, b4b5 (reactive site) and

b7b8 of TKI are able to interact with S1, S4 pockets

and two exosites of FXa and confer specificity for

FXa along with trypsin.

In summary, comparative analyses with other

Kunitz type inhibitors at sequence and 3D structural

levels allow us to clarify and propose the structural

and mechanistic basis for the dual targeting capability

of TKI, which binds and inhibits FXa as well as tryp-

sin. However, the distinct reactive site of TKI may

possibly inhibit other trypsin like serine proteases and

needs further investigation. Moreover, inhibition of

FXa and prolongation of blood clotting time point

towards the potential clinical application and use of

TKI in antithrombotic therapy.

Materials and methods

RNA extraction and cDNA cloning

In order to obtain the complete amino acid sequence of TKI,

3-month-old Tamarindus indica seeds were collected locally.

Total RNA was isolated from the seeds as described previ-

ously [51] with slight modifications such as the use of 4%

polyvinylpyrrolidone in the extraction buffer and slightly

higher pH extraction buffer [52]. The first strand of cDNA

was reverse transcribed from total RNA using oligo(dT)18adaptor primer (5′-CCAGTGAGCAGAGTGACGACTC

GAGCTCAAGCTTTTTTTTTTTTTTTTTT-3′). The TKI

A

C

B

D

Fig. 9. Insertion of Asn15 within the Kunitz signature sequence of TKI and its role in mediating interaction between inhibitor and its cognate

proteases. (A) Superposition of TKI A and TKI B with STI and ETI shows that Asn15 is not superimposed with any residues of ETI and STI.

The Asn15 residue in TKI A and TKI B do not superimpose onto each other due to conformational variation and their side chains are 3 �A

apart. (B) Superposition of TKI A with CTI shows that CTI also has insertion of Asp instead of Asn and it may participate in complex

formation with its cognate proteases. (C) Interactions of Asn15 with residues of PPT through water molecules are shown. Hydrogen bonds

and water molecules are presented as black and red spheres respectively. (D) In modelled TKI–FXa complex, Asn15 forms three hydrogen

bonds with Gln192, a key determinant for specificity towards its substrates/inhibitor, and one hydrogen bond with Glu147 of FXa.



coding sequence was amplified using a forward degenerate

primer (F1: 5′-GAYACHGTNCAYGAYACHGAYGG-3′)

which was designed on the basis of the previously reported

N-terminal sequence of TKI [53], and the oligo(dT)18adaptor primer was used as the reverse primer (R1). Then

amplified TKI gene was cloned into pGEM-T Easy Vector

according to the instructions provided (Promega, Madison,

WI, USA) and transformed using DH5a competent cells.

The obtained recombinant plasmid containing TKI gene

insert (pGEMT-TKI) was sequenced in both directions using

M13 forward and reverse primers.

Analysis of TKI gene

The ORF finder tool from NCBI was used to deduce the

primary amino acid sequence of TKI from the TKI gene

sequence. Similar sequences available in the database were

identified by the program BLAST [54]. Multiple sequence

alignment was performed using CLUSTAL W [55]. The predic-

tion of the Kunitz signature and probable motifs involved

in post-translational modifications were done using the

MOTIF scan server [56].

Purification and crystallization of TKI

The purification and crystallization of TKI were reported

earlier [31]. However, better diffracting crystals for free

TKI were obtained using 25% (v/v) polyethylene glycol

3350 (PEG 3350) and 0.2 M magnesium formate at pH 7.0.

TKI–PPT and TKI–FXa complex formation and

purification

The TKI–PPT complex formation, purification and crys-

tallization were done as described previously [32]. Further-

more, the crystals used for this complex were obtained

using 0.2 M ammonium phosphate dibasic and 20% PEG

3350. For TKI–FXa complex formation, the purified

100 lL TKI (15 mg�mL�1 in 20 mM Tris buffer pH 8.0)

was mixed with 100 lL human FXa (2.5 mg�mL�1 con-

centrated from 1 mg�mL�1 FXa). The TKI–FXa mixture

was incubated at 37 °C for 2 h. The complex was purified

using a calibrated Superdex-75 GL 10/300 (GE Health-

care, Uppsala, Sweden) size-exclusion column on an

AKTA Purifier FPLC system (GE Healthcare). The frac-

tions were analysed on 15% reducing SDS/PAGE and

concentrated to 4.5 mg�mL�1 using an Amicon Ultra-4

10 kDa cutoff concentrator (Millipore, Carrigtwohill, Co.,

Cork, Ireland).

Anticoagulant assay

APTT [57] and PT [58] assays were performed using nor-

mal human plasma and commercially available kits (Liqui-

plastin for PT and Liquicelin-E for APTT, Tulip

Diagnostics Pvt. Ltd, Roorkee, India) in the presence and

absence of TKI. For in vitro APTT assays, normal human

plasma (100 lL), kit reagent (100 lL) and TKI (0–10 lM,previously diluted in 100 lL of 0.1 M Tris/HCl buffer pH

8.0) were mixed and incubated for 3 min at 37 °C. To initi-

ate the coagulation process, 100 lL of 0.025 M CaCl2 was

added, and the time for clot formation was recorded. To

measure the in vitro PT assay, normal human plasma

(100 lL) was incubated with different concentrations of

TKI (0–10 lM previously diluted in 100 lL of 0.1 M Tris/

HCl buffer at pH 8.0) for 2 min at 37 °C. Then, 100 lL of

the kit reagent was added, and the clotting time was mea-

sured. Both the experiments were carried out in triplicate

and the average value was taken.

Assay of inhibitory activity

Trypsin inhibitory activity of TKI was determined by mea-

suring the residual hydrolytic activity of PPT towards a

BAPNA substrate (Sigma-Aldrich, St Louis, MO, USA)

and chymotrypsin inhibitory activity was assayed using

BTPNA substrate (Sigma-Aldrich) at pH 8.0. FXa inhibi-

tory assay was performed as described previously with

modification [59–61]. 0.25 lM human FXa (Pierce, Thermo

Scientific, USA) was incubated with different concentra-

tions of TKI (1–60 lM) in 50 mM Tris buffer pH 8.0 and

300 mM NaCl at 37 °C for 30 min on a waterbath. After

incubation, the chromogenic substrate CH3OCO-D-CHA-

Gly-Arg-pNA-AcOH (Sigma-Aldrich Co, F3301) was

added and incubated for 3 min at 37 °C on a waterbath.

The reaction was stopped by adding 2% citric acid. Elas-

tase inhibitory activity of TKI was performed as described

previously [62] by measuring the residual activity of porcine

pancreatic elastase (PPE) using the substrate N-Suc-(Ala)3-

nitroanilide (Sigma-Aldrich). The colour formation by the

liberated chromophoric group pNA by enzymatic hydroly-

sis of BAPNA/BTPNA/FXa substrate/PPE substrate was

measured spectrophotometrically at 410 nm. All assays

were performed in triplicate.

Ki determination

The inhibition constant (Ki) of TKI against both trypsin

and FXa was determined from a Dixon plot [63,64]. The

enzyme inhibition was characterized at two different sub-

strate concentrations ([S1] and [S2]). BAPNA (4 and 6 mM)

and FXa chromogenic substrate (5 and 7 mM) were used as

a substrate for trypsin and FXa activity, respectively. Stud-

ies were performed by pre-incubating the fixed amount of

respective enzyme [trypsin (0.014 nM), FXa (0.5 nM)] with

increasing concentrations of TKI (0–15 nM for trypsin

inhibition and 0–500 nM for FXa inhibition) for 15 min at

37 °C followed by addition of two different substrate

concentrations. The reciprocal velocity (1/v) versus inhibi-

tor concentrations [I], for each substrate concentration [S1]



and [S2], were plotted (Dixon plots). Ki was calculated

from the intersection of the two regression lines.

Data collection and processing

The X-ray diffraction data were collected at 1.94 �A for free

TKI and 2.23 �A resolution for the TKI–PPT complex at

home source. Data sets for free TKI and TKI–PPT com-

plex were collected at 100 K. Indexing, integration of all

the diffraction images and scaling of the diffraction data

were carried out using HKL2000 [65] and the data collection

statistics are summarized in Table 1.

Structure determination and refinement

The structures of free TKI and its complex with trypsin

were solved by the molecular replacement method using the

program MOLREP in the CCP4I suite [66]. The best solution

for free TKI and the complex were obtained with the crys-

tal structure of a trypsin inhibitor from Erythriana caffra

seeds (PDB ID code, 1TIE) [12] and the complex structure

of soybean (PDB ID code, 1AVW) as the search model

respectively [11]. The rigid body refinement was followed

by iterative cycles of restrained atomic parameter refine-

ment using REFMAC5 [66,67]. Further model visualization,

refinement, model building and fitting of the electron den-

sity map were carried out using the molecular graphics pro-

gram COOT [68,69]. The details of the refinement statistics

are included in Table 1. The final structures were validated

using MOLPROBITY server [35]. Figures were generated using

the PYMOL [70] and ESPRIPT program [71].

Rigid body docking

As crystals of TKI with FXa were not obtained, rigid body

docking was performed to examine the mode of interaction

of TKI with FXa. The fully automated online protein–pro-

tein docking program CLUSPRO [72,73] was used to model

the TKI–FXa complex with good surface complementarity

and low desolvation energies. The PDB file for FXa (PDB

code 1FAX) [74] was submitted as a receptor structure,

and the coordinates of TKI (PDB code 4AN6) as the

ligand structure. The server was executed with default

parameters. The top-ranked complexes from CLUSPRO were

used for further analysis based on the prior knowledge of

active site interactions. Interactions between TKI and FXa

residues were calculated using the PIC server [75].

Acknowledgements

The authors thank Macromolecular Crystallographic

Facility (MCU) at IIC, IIT Roorkee for the structure

determination. This work has been supported finan-

cially by the Council of Scientific and Industrial

Research (CSIR), New Delhi, India (No. 38(1228)/09/

EMR II dated Dec 01, 2009). DNP thank MHRD

Government of India, for financial support. We are

grateful to Dr S. Karthikeyan, IMTECH-Chandigarh,

for providing help in the structure solution of TKI–PPT complex. Dr Madhulika, MD (pathology), helped

in providing the facility for blood coagulation assays.

DNP also thanks the CSIR for financial support. The

authors are grateful to Shivendra Pratap for help with

the anticoagulation assay and docking studies and to

P. Supriya, D. Sonali, P. Nandita and P. Selva Kumar

for critical reading of the manuscript.

References

1 Otlewski J, Jaskolski M, Buczek O, Cierpicki T,

Czapi�nska H, Krowarsch D, Smalas AO, Stachowiak

D, Szpineta A & Dadlez M (2001) Structure–function

relationship of serine protease–protein inhibitor

interaction. Acta Biochim Pol 48, 419–428.

2 Jones S & Thornton JM (1996) Principles of protein–

protein interactions. Proc Natl Acad Sci USA 93,

13–20.

3 Oliva MLV, Souza-Pinto JC, Batista IFC, Araujo MS,

Silveira VF, Auerswald EA, Mentele R, Eckerskorn C,

Sampaio MU & Sampaio CAM (2000) Leucaena

leucocephala serine proteinase inhibitor: primary

structure and action on blood coagulation, kinin

release and rat paw edema. Biochim Biophys Acta 1477,

64–74.

4 Oliva MLV, Silva MCC, Sallai RC, Brito MV &

Sampaio MU (2010) A novel subclassification for

Kunitz proteinase inhibitors from leguminous seeds.

Biochimie 92, 1667–1673.

5 Norioka N, Hara S, Ikenaka T & Abe J (1988)

Distribution of the Kunitz and the Bowman–Birk

family proteinase inhibitors in leguminous seeds

(biological chemistry). Agric Biol Chem 52, 1245–1252.

6 Krauchenco S, Pando SC, Marangoni S & Polikarpov I

(2003) Crystal structure of the Kunitz (STI)-type

inhibitor from Delonix regia seeds. Biochem Biophys Res

Commun 312, 1303–1308.

7 Lingaraju MH & Gowda LR (2008) A Kunitz trypsin

inhibitor of Entada scandens seeds: another member

with single disulfide bridge. Biochim Biophys Acta 1784,

850–855.

8 Lehle K, Wrba A & Jaenicke R (1994) Erythrina caffra

trypsin inhibitor retains its native structure and function

after reducing its disulfide bonds. J Mol Biol 239,

276–284.

9 Koide T & Ikenaka T (1973) Studies on soybean

trypsin inhibitors. Eur J Biochem 32, 417–431.

10 Rawlings ND (2010) Peptidase inhibitors in the

MEROPS database. Biochimie 92, 1463–1483.



11 Song HK & Suh SW (1998) Kunitz-type soybean

trypsin inhibitor revisited: refined structure of its

complex with porcine trypsin reveals an insight into the

interaction between a homologous inhibitor from

Erythrina caffra and tissue-type plasminogen activator.

J Mol Biol 275, 347–363.

12 Onesti S, Brick P & Blow DM (1991) Crystal structure

of a Kunitz-type trypsin inhibitor from Erythrina caffra

seeds. J Mol Biol 217, 153–176.

13 Dattagupta JK, Podder A, Chakrabarti C, Sen U,

Mukhopadhyay D, Dutta SK & Singh M (1999)

Refined crystal structure (2.3 A) of a double-headed

winged bean alpha-chymotrypsin inhibitor and location

of its second reactive site. Proteins 35, 321–331.

14 Ravichandran S, Sen U, Chakrabarti C & Dattagupta

JK (1999) Cryocrystallography of a Kunitz-type serine

protease inhibitor: the 90 K structure of winged bean

chymotrypsin inhibitor (WCI) at 2.13 A resolution.

Acta Crystallogr D Biol Crystallogr 55, 1814–1821.

15 Meester PD, Brick P, Lloyd LF, Blow DM & Onesti S

(1998) Structure of the Kunitz-type soybean trypsin

inhibitor (STI): implication for the interactions between

members of the STI family and tissue-plasminogen

activator. Acta Crystallogr D Biol Crystallogr 54,

589–597.

16 Dasgupta J, Khamrui S, Dattagupta JK & Sen U

(2006) Spacer Asn determines the fate of Kunitz (STI)

inhibitors, as revealed by structural and biochemical

studies on WCI mutants. Biochemistry 45, 6783–6792.

17 Krauchenco S, Nagem RAP, Da Silva JA, Marangoni

S & Polikarpov I (2004) Three-dimensional structure of

an unusual Kunitz (STI) type trypsin inhibitor from

Copaifera langsdorffii. Biochimie 86, 167–172.

18 Hansen D, Macedo-Ribeiro S, Verassimo P, Yoo Im S,

Sampaio MU & Oliva MLV (2007) Crystal structure of

a novel cysteinless plant Kunitz-type protease inhibitor.

Biochem Biophys Res Commun 360, 735–740.

19 Gahloth D, Selvakumar P, Shee C, Kumar P & Sharma

AK (2010) Cloning, sequence analysis and crystal

structure determination of a miraculin-like protein from

Murraya koenigii. Arch Biochem Biophys 494, 15–22.

20 Bao R, Zhou CZ, Jiang C, Lin SX, Chi CW & Chen Y

(2009) The ternary structure of the double-headed

arrowhead protease inhibitor API-A complexed with

two trypsins reveals a novel reactive site conformation.

J Biol Chem 284, 26676–26684.

21 Ribeiro JKC, Cunha DDS, Fook JM & Sales MP

(2010) New properties of the soybean trypsin inhibitor:

inhibition of human neutrophil elastase and its effect

on acute pulmonary injury. Eur J Pharmacol 644,

238–244.

22 Inagaki K, Kobayashi H, Yoshida R, Kanada Y,

Fukuda Y, Yagyu T, Kondo T, Kurita N, Kitanaka T

& Yamada Y (2005) Suppression of urokinase

expression and invasion by a soybean Kunitz trypsin

inhibitor are mediated through inhibition of

Src-dependent signaling pathways. J Biol Chem 280,

31428–31437.

23 De Paula CAA, Coulson-Thomas VJ, Ferreira JG,

Maza PK, Suzuki E, Nakahata AM, Nader HB,

Sampaio MU & Oliva MLV (2012) EcTI, a plant

proteinase inhibitor, decreases in vitro cell adhesion and

invasion by inhibition of Src-FAK signaling pathways.

J Biol Chem 287, 170–182.

24 Nakahata AM, Mayer B, Ries C, de Paula CAA,

Karow M, Neth P, Sampaio MU, Jochum M & Oliva

MLV (2011) The effects of a plant proteinase inhibitor

from Enterolobium contortisiliquum on human tumor

cell lines. Biol Chem 392, 327–336.

25 Batista IFC, Oliva MLV, Araujo MS, Sampaio MU,

Richardson M, Fritz H & Sampaio CAM (1996)

Primary structure of a Kunitz-type trypsin inhibitor

from Enterolobium contortisiliquum seeds.

Phytochemistry 41, 1017–1022.

26 Oliva MLV, Andrade SA, Juliano MA, Sallai RC,

Torquato RJ, Sampaio MU, Pott VJ & Sampaio CAM

(2003) Kinetic characterization of factor Xa binding

using a quenched fluorescent substrate based on the

reactive site of factor Xa inhibitor from Bauhinia

ungulata seeds. Curr Med Chem 10, 1085–1093.

27 Oliva MLV, Mendes CR, Santomauro-Vaz EM,

Juliano MA, Mentele R, Auerswald EA, Sampaio MU

& Sampaio CAM (2001) Bauhinia bauhinioides plasma

kallikrein inhibitor interaction with synthetic peptides

and fluorogenic peptide substrates related to the

reactive site sequence. Curr Med Chem 8, 977–984.

28 Luiza Vilela Oliva M, da Silva Ferreira R, Gasperazzo

Ferreira J, Alessandra Andrade de Paula C, Salas CE

& Uemura Sampaio M (2011) Structural and functional

properties of Kunitz proteinase inhibitors from

leguminosae: a mini review. Curr Protein Pept Sci 12,

348–357.

29 Oliveira C, Santana LA, Carmona AK, Cezari MH,

Sampaio MU, Sampaio CAM & Oliva MLV (2001)

Structure of cruzipain/cruzain inhibitors isolated from

Bauhinia bauhinioides seeds. Biol Chem 382, 847–852.

30 Fang EF, Wong JH, Bah CSF, Lin P, Tsao SW & Ng

TB (2010) Bauhinia variegata var. variegata trypsin

inhibitor: from isolation to potential medicinal

applications. Biochem Biophys Res Commun 396,

806–811.

31 Patil DN, Chaudhry A, Sharma AK, Tomar S &

Kumar P (2009) Purification, crystallization and

preliminary crystallographic studies of a Kunitz-type

proteinase inhibitor from tamarind (Tamarindus indica)

seeds. Acta Crystallogr Sect F Struct Biol Cryst

Commun 65, 736–738.

32 Tomar S, Patil DN, Datta M, Tapas S, Chaudhary A,

Sharma AK & Kumar P (2009) Crystallization and

preliminary X-ray diffraction analysis of the complex of



Kunitz-type tamarind trypsin inhibitor and porcine

pancreatic trypsin. Acta Crystallogr Sect F Struct Biol

Cryst Commun 65, 1179–1181.

33 Ara�ujo CL, Bezerra IWL, Oliveira AS, Moura FT,

Macedo LLP, Gomes CEM, Barbosa AEAD, Macedo

FP, Souza TMS & Franco OL (2005) In vivo

bioinsecticidal activity toward Ceratitis capitata (fruit

fly) and Callosobruchus maculatus (cowpea weevil) and

in vitro bioinsecticidal activity toward different orders

of insect pests of a trypsin inhibitor purified from

tamarind tree (Tamarindus indica) seeds. J Agric Food

Chem 53, 4381–4387.

34 De Souza AF, Torquato RJS, Tanaka AS & Sampaio

CAM (2005) Cloning, expression and characterization

of Bauhinia variegata trypsin inhibitor BvTI. Biol Chem

386, 1185–1189.

35 Chen VB, Arendall WB, Headd JJ, Keedy DA,

Immormino RM, Kapral GJ, Murray LW, Richardson

JS & Richardson DC (2010) MolProbity: all-atom

structure validation for macromolecular

crystallography. Acta Crystallogr D Biol Crystallogr 66,

12–21.

36 Murzin AG, Lesk AM & Chothia C (1992) beta-Trefoil

fold. Patterns of structure and sequence in the Kunitz

inhibitors interleukins-1 beta and 1 alpha and fibroblast

growth factors. J Mol Biol 223, 531–543.

37 McLachlan AD (1979) Three-fold structural pattern in

the soybean trypsin inhibitor (Kunitz). J Mol Biol 133,

557–563.

38 Laskowski RA (2009) PDBsum new things. Nucleic

Acids Res 37, D355–D359.

39 Minoux H & Chipot C (1999) Cation–p interactions in

proteins: can simple models provide an accurate

description? J Am Chem Soc 121, 10366–10372.

40 Lin Z & Johnson ME (1995) Proposed cation-pi

mediated binding by factor Xa: a novel enzymatic

mechanism for molecular recognition. FEBS Lett 370,

1–5.

41 Wang SX, Hur E, Sousa CA, Brinen L, Slivka EJ &

Fletterick RJ (2003) The extended interactions and Gla

domain of blood coagulation factor Xa. Biochemistry

42, 7959–7966.

42 Rezaie AR & Esmon CT (1995) Contribution of residue

192 in factor Xa to enzyme specificity and function.

J Biol Chem 270, 16176–16181.

43 Johnson DJD, Li W, Adams TE & Huntington JA

(2006) Antithrombin-S195A factor Xa-heparin structure

reveals the allosteric mechanism of antithrombin

activation. EMBO J 25, 2029–2037.

44 Wei Z, Yan Y, Carrell RW & Zhou A (2009) Crystal

structure of protein Z-dependent inhibitor complex

shows how protein Z functions as a cofactor in the

membrane inhibition of factor X. Blood 114, 3662–3667.

45 Burgering MJM, Orbons LPM, van der Doelen A,

Mulders J, Theunissen HJM, Grootenhuis PDJ, Bode

W, Huber R & Stubbs MT (1997) The second Kunitz

domain of human tissue factor pathway inhibitor:

cloning, structure determination and interaction with

factor Xa. J Mol Biol 269, 395–407.

46 Lapatto R, Krengel U, Schreuder HA, Arkema A, de

Boer B, Kalk KH, Hol WGJ, Grootenhuis PDJ,

Mulders JWM & Dijkema R (1997) X-ray structure of

antistasin at 1.9 A resolution and its modelled complex

with blood coagulation factor Xa. EMBO J 16,

5151–5161.

47 Wei A, Alexander RS, Duke J, Ross H, Rosenfeld SA

& Chang CH (1998) Unexpected binding mode of tick

anticoagulant peptide complexed to bovine factor Xa.

J Mol Biol 283, 147–154.

48 Rezaie AR (1996) Role of residue 99 at the S2 subsite

of factor Xa and activated protein C in enzyme

specificity. J Biol Chem 271, 23807–23814.

49 Rai R, Sprengeler PA, Elrod KC & Young WB (2001)

Perspectives on factor Xa inhibition. Curr Med Chem 8,

101–119.

50 Maignan S & Mikol V (2001) The use of 3D structural

data in the design of specific factor Xa inhibitors. Curr

Top Med Chem 1, 161–174.

51 Tai HH, Pelletier C & Beardmore T (2004) Total RNA

isolation from Picea mariana dry seed. Plant Mol Biol

Rep 22, 93a–93e.

52 Meng L & Feldman L (2010) A rapid TRIzol-based

two-step method for DNA-free RNA extraction from

Arabidopsis siliques and dry seeds. Biotechnol J 5,

183–186.

53 Rao DH & Gowda LR (2008) Abundant class III

acidic chitinase homologue in tamarind (Tamarindus

indica) seed serves as the major storage protein. J Agric

Food Chem 56, 2175–2182.

54 Altschul SF, Madden TL, Sch€affer AA, Zhang J,

Zhang Z, Miller W & Lipman DJ (1997) Gapped

BLAST and PSI-BLAST: a new generation of protein

database search programs. Nucleic Acids Res 25,

3389–3402.

55 Thompson JD, Higgins DG & Gibson TJ (1994)

CLUSTAL W: improving the sensitivity of progressive

multiple sequence alignment through sequence

weighting, position-specific gap penalties and weight

matrix choice. Nucleic Acids Res 22, 4673–4680.

56 Hulo N, Bairoch A, Bulliard V, Cerutti L, Cuche BA,

De Castro E, Lachaize C, Langendijk-Genevaux PS &

Sigrist CJA (2008) The 20 years of PROSITE. Nucleic

Acids Res 36, D245–D249.

57 Hoffmann JJ & Meulendijk PN (1978) Comparison of

reagents for determining the activated partial

thromboplastin time. Thromb Haemost 39, 640–645.

58 Quick AJ (1966) The prothrombin consumption time

test. Am J Clin Pathol 45, 105–109.

59 Batista IFC, Ramos OHP, Ventura JS, Junqueira-de-

Azevedo ILM, Ho PL & Chudzinski-Tavassi AM



(2010) A new factor Xa inhibitor from Amblyomma

cajennense with a unique domain composition. Arch

Biochem Biophys 493, 151–156.

60 Van Wijk EM & Smit JLJ (1984) A chromogenic

micro-assay for factor X-activity. Thromb Res 34,

263–268.

61 Oliva ML, Sallai RC, Sampaio CA, Fritz H, Auerswald

EA, Tanaka AS, Torquato RJ & Sampaio MU (1996)

Bauhinia serine proteinase inhibitors: effect on factor X,

factor XII and plasma kallikrein. Immunopharmacology

32, 85–87.

62 Moon JY, Yim EY, Song G, Lee NH & Hyun CG

(2010) Screening of elastase and tyrosinase inhibitory

activity from Jeju Island plants. EurAsia J BioSci 4,

41–53.

63 Dixon M (1953) The determination of enzyme inhibitor

constants. Biochem J 55, 170–171.

64 Segel IH (1975) Enzyme Kinetics. Wiley, New York,

p. 465.

65 Otwinowski Z & Minor W (1997) Processing of X-ray

diffraction data collected in oscillation mode. Methods

Enzymol 276, 307–326.

66 Bailey S (1994) The CCP4 suite: programs for protein

crystallography. Acta Crystallogr D Biol Crystallogr 50,

760–763.

67 Murshudov GN, Vagin AA & Dodson EJ (1997)

Refinement of macromolecular structures by the

maximum-likelihood method. Acta Crystallogr D Biol

Crystallogr 53, 240–255.

68 Emsley P & Cowtan K (2004) Coot: model-building

tools for molecular graphics. Acta Crystallogr D Biol

Crystallogr 60, 2126–2132.

69 Emsley P, Lohkamp B, Scott WG & Cowtan K (2010)

Features and development of Coot. Acta Crystallogr D

Biol Crystallogr 66, 486–501.

70 DeLano WL (2002) The PyMOL Molecular Graphics

System. Schrodinger, LLC, New York.

71 Gouet P, Courcelle E & Stuart DI (1999) ESPript:

analysis of multiple sequence alignments in PostScript.

Bioinformatics 15, 305–308.

72 Comeau SR, Gatchell DW, Vajda S & Camacho CJ

(2004) ClusPro: an automated docking and

discrimination method for the prediction of protein

complexes. Bioinformatics 20, 45–50.

73 Comeau SR, Gatchell DW, Vajda S & Camacho CJ

(2004) ClusPro: a fully automated algorithm for protein

–protein docking. Nucleic Acids Res 32, W96–W99.

74 Brandstetter H, K€uhne A, Bode W, Huber R, Von der

Saal W, Wirthensohn K & Engh RA (1996) X-ray

structure of active site-inhibited clotting factor Xa.

J Biol Chem 271, 29988–29992.

75 Tina KG, Bhadra R & Srinivasan N (2007) PIC:

protein interactions calculator. Nucleic Acids Res 35,

W473–W476.

76 Crooks GE, Hon G, Chandonia JM & Brenner SE

(2004) WebLogo: a sequence logo generator. Genome

Res 14, 1188–1190.



structural basis for dual inhibitory role of tamarind ... · tki-fxa complex revealed that the...

Documents