Characterization of Three Endogenous Peptide Inhibitors for Multiple Metalloproteinases with Fibrinogenolytic Activity from the Venom of Taiwan Habu (Trimeresurus mucrosquamatus)

ARTICLE NO.RC989017

Characterization of Three Endogenous Peptide Inhibitors

for Multiple Metalloproteinases with Fibrinogenolytic

Activity from the Venom of Taiwan Habu

(Trimeresurus mucrosquamatus)

Kai-Fa Huang,* Chin-Chun Hung,* Shih-Hsiung Wu,*

,

_{† and Shyh-Horng Chiou*}

,

_†

,1

†Institute of Biochemical Sciences, National Taiwan University, Taipei; and *Institute of Biological Chemistry,

P.O. Box 23-106, Academia Sinica, Taipei, Taiwan

Received May 23, 1998

talloproteinases present in crotalid and viperid venoms.

Three small peptide components were isolated and

_{q 1998 Academic Press}

purified from the venom of Taiwan habu

(Trimere-surus mucrosquamatus), which show specific

activ-ity to inhibit the strong proteolytic activactiv-ity of

multiple metalloproteinases present in the crude

_{Venoms of the snake families of Viperidae and}

Cro-venom. Using multiple chromatographies coupled

talidae cause shock, intravascular clotting, systemic

with successive ultrafiltrations, three inhibitors,

and local hemorrhage, edema and necrosis upon

victim-i.e. pyroglutamate-lysine-tryptophan (pyroGlu-Lys-Trp),

ized prey. It has also been well known that snake

ven-pyroglutamate-asparagine-tryptophan

(pyroGlu-Asn-oms contain complex mixtures of pharmacologically

ac-Trp) and pyroglutamate-glutamine-tryptophan

(pyro-tive peptides and proteins including some potent

pro-Glu-Gln-Trp) were obtained in good yields and high

teolytic enzymes (1-3). The common symptom of

homogeneity. The yields of these peptide fractions

hemorrhage caused by snakebites is usually the result

were estimated to be about 0.65 mg, 0.55 mg and 0.42

of structural destruction of capillary basement

mem-mg from 250 mem-mg total lyophilized crude venom, which

branes via proteolytic degradation by these proteases

corresponded to the approximate concentrations of 8.4

(4). Death is usually the result of the combined effects

mM, 7.3 mM and 5.4 mM respectively in venom

secre-of several components in the venom. In order to avoid

tion. Detailed and unambiguous structural

determina-the auto-digestion of venom gland by its own

degrada-tion was established by amino acid analyses, mass

spectrometry and microsequencing of purified pep-

tive proteases present in gland secretions, snake

ven-tides. Further functional characterization of these

oms of various species have been reported to contain

three tripeptides showed that they could weakly in-

some self-defensive enzyme inhibitors such as serine

hibit three metalloproteinases previously isolated

protease inhibitors (5-7), citrate and small peptides (8),

from the same venom. The inhibitory activities were

_{that can selectively bind to these proteases thereby}

similar among these tripeptides and their IC

50

(concen-

partially inhibit and minimize their proteolytic

activi-tration for 50% inhibition) were estimated in a range

_{ties in crude venoms. Venom proteinases so far}

charac-of 0.20-0.95 mM, which is much more effective than

_{terized have been shown to be a heterologous group of}

citrate, another venom protease inhibitor of low mo-

_{proteins, generally belonging to two major classes of}

lecular-weight component. Since these tripeptides are

_{proteases, i.e. thrombin-like enzymes and}

metallopro-the endogenous peptide inhibitors present in metallopro-the lu-

_{teinases with a wide range of molecular masses (9-11).}

men of venom glands, it is conceivable that they may

Early studies have indicated that, a few snake

spe-act as a self-defensive mechanism against the

auto-di-cies of Viperidae and Crotalidae families contained

en-gestive deleterious effect of the strong

metalloprotein-dogenous small peptides present as a form of

pyrogluta-ases in vivo, particularly several zinc-dependent

me-mate(pyroGlu)-containing tripeptides, e.g.

pyroGlu-Asn-Trp and pyroGlu-Gln-Trp (12,13). These small

N-terminally blocked peptides were later found in

venoms of various species of snakes, and shown to be

1_{Corresponding address: Institute of Biological Chemistry,}

Acade-relatively good inhibitors of zinc-dependent

metallopro-mia, P.O. Box 23-106, Taipei, Taiwan. Fax: (886)-2-3635038. E-mail:

Vol. 248, No. 3, 1998 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Cleavage of the N-terminal pyroglutamate from peptides and

se-concentration present in various snake venoms of

quence analysis. Cleavage of the N-terminal pyroglutamate from

Crotalidae and Viperidae families (16), has also been

peptides was accomplished using sequencing grade pyroglutamate

identified as a low molecular-weight component in

_{aminopeptidase (Boehringer Mannheim). About 5-10 mg peptides}

crude venom to play a role as an endogenous inhibitor

were dissolved in 45ml digestion buffer (100 mM sodium phosphate, 10 mM EDTA, 5 mM dithiothreitol and 5% glycerol, pH 8.0). After

of venom enzymes and proteinases (17). Therefore it

addition of 5ml pyroglutamate aminopeptidase (0.25mg/ml) and

incu-would be of interest to compare the inhibitory activities

bation for 18 hours at 47C first and then 4 hours at 25 7C, peptides in

of these small endogenous components in snake

ven-reaction mixture were separated by reverse-phase high performance

oms on the proteinases isolated from the same venom

_{liquid chromatography. Sequencing of deblocked peptides was}

car-in order to provide some car-insights car-into the protracted

ried out by automated Edman degradation with a microsequencing

sequenator (Model 477A, Applied Biosystems) as described (21).

issue on how snakes develop the self-defensive

mecha-nism in protecting the venom gland itself from the dam-

_{Determination of protein and peptide concentration. Protein}

con-aging effect of its secreted metalloproteinases. Pre-

centration was determined by dye-binding assays (22) or estimated from the absorbance of the purified protein at 280 nm, using an

viously we have identified three metalloproteinases

extinction coefficient (E1%_{) of 10. Concentrations of}

tryptophan-con-with strong fibrinogenolytic activity from the Taiwan

taining tripeptides were measured by absorbance at 280 nm using

habu (Trimeresurus mucrosquamatus), designated as

_{a molar extinction coefficient of 5600 M}01_cm01_.

TM-1, TM-2, and TM-3 (18,19). In this report we have

FTC-caseinolytic activity assay of metalloproteinases and its

inhi-focused on the isolation and characterization of three

bition by peptides and citrate. Caseinolytic activity of

metallopro-pyroglutamate tripeptides and compare their inhibi-

_{teinases was measured using a fluorescence substrate, FTC-casein}

tory activities with citrate ion on these three metallo-

as reported previously (23), with or without the addition of small

proteinases. One novel tripeptide, pyroGlu-Lys-Trp,

peptide inhibitors or sodium citrate. Reaction mixtures, containing 5ml enzyme (0.2-0.4mg/ml) and 5ml of the indicated concentrations

has for the first time been identified and shown to be

of inhibitor fractions (or replacing this with the assay buffer, 100 mM

a genuine inhibitor for these metalloproteinases.

Tris.HCl/10 mM CaCl2, pH 8.0, as positive control), was incubated at

room temperature for 15 min, followed by adding 5ml FTC-casein (10 mg/ml in 50 mM Tris.HCl, pH 7.2) and 35ml of the assay buffer

MATERIALS AND METHODS

for a prolonged incubation at 377C for 70 min. Proteolysis was termi-nated by adding 120 ml of 5% TCA and mixing extensively. The

Materials. Lyophilized venom powder was obtained from the

lo-reaction mixture was allowed to stand at room temperature for 1 h cal snake farm. The substrate FTC-casein was purchased from Sigma

and the TCA-insoluble protein was sedimented by centrifugation at Chemical Company (St. Louis, MO). Gel suspensions of TSK

DEAE-13,000 rpm for 2.5 min. A 90-ml aliquot of the supernatant fraction 650(M) anion-exchange resin were purchased from Merck

(Darms-was diluted with 0.5 M Tris buffer, pH 8.5, with vigorous mixing to tadt, Germany).

ensure the entire sample was at the proper pH. Fluorescence was

Isolation and purification of small peptides from crude venom. An _{measured using an excitation wavelength at 490 nm and an emission}

anion-exchange chromatography on an open column (2.5 1 46 cm) _{wavelength at 522 nm on a Hitachi’s F-4010 fluorescence} spectropho-packed with TSK DEAE-650(M) gel suspension was employed to _{tometer. Percent inhibition for the caseinolytic activity of} metallopro-separate small peptide components. Venom powder dissolved in _{teinases was estimated from the fluorescence yields of assayed} sam-0.025 M ammonium bicarbonate, pH 7.8 starting buffer (total 5 ml) _{ples as compared with that of control.}

was applied to the column and then eluted in a linear gradient of 0.025-0.5 M ammonium bicarbonate, followed by 0.5-1.0 M

ammo-nium bicarbonate, pH 8.0 buffer, similar to those described pre-

_{RESULTS AND DISCUSSION}

viously (18). Two fractions rich in small peptides were obtained from

the above chromatography and ultrafiltered through membranes

(YM-10, Amicon) and the flow-throughs were collected and lyophi-

_{The venoms of various snakes have been shown to}

lized. Reverse-phase chromatography was performed on a Hitachi’s

_{possess very strong and stable proteolytic enzymes}

liquid chromatograph using C18column (0.46 1 25 cm, Vydac).

Elu-with fibrinolytic or fibrinogenolytic activity, notably in

tion was carried out in a linear gradient of 5-95% acetonitrile in 0.1%

the snake families of Crotalidae and Viperidae (24-26).

TFA and the eluates were monitored at either 214 nm or 280 nm.

Major fractions eluted from the column were collected and lyophi-

Currently several proteolytic enzymes of snake origin

lized.

_{have been used clinically as potential antithrombotic}

agents (27-30) with the special concern on how to

modu-SDS – polyacrylamide gel electrophoresis and amino acid analyses.

The purity of purified polypeptides was checked by SDS-polyacryl-

_{late and mitigate unwanted toxic effects of local and}

amide slab gel electrophoresis (5% stacking/15% resolving gel) as

_{systemic hemorrhage or bleeding caused by these}

pro-described (20) with some modifications (i.e., 5% cross-linking N, N

*-teinases. Most venom proteases with such hemorrhagic

methylenebisacrylamide in the gel solution). The amino acid

compo-side effects consist mainly of metalloproteinases (31).

sitions were determined with a Beckman 6300 amino acid analyzer

using a single-column system based on a conventional ion-exchange

In this study we have made an effort to look for natural

chromatography system.

_{proteinase inhibitors present in the crude venom of}

Taiwan habu and their effects on three strong

hemor-Mass spectrometry analyses. Fast atom bombardment (FAB)

mass spectra were obtained on an Autospec mass spectrometer (Mi-

_{rhagic metalloproteinases isolated previously from the}

cromass, UK) fitted with a cesium ion gun operated at 25 kV. Samples

_{same venom (18,19). Successful identification and}

char-were dissolved in 5% acetic acid for loading onto the probe tip coated

acterization of three pyroglutamate tripeptides could

with monothioglycerol as matrix. Multiple scans at 6 seconds per

account for the resistance of snake venom components

decay were acquired and summed for data processing using the Opus

FIG. 1. Isolation of small peptides from the crude venom of Taiwan habu (Trimeresurus mucrosquamatus) on TSK DEAE-650(M) column. About 250 mg of lyophilized venom powder was dissolved in the starting buffer of 0.025 M ammonium bicarbonate, pH 7.8, and applied to the column equilibrated with the same buffer. Elution was carried out in three steps similar to that described in the previous report (18,19) with some modifications. The column eluates (3.5 ml/tube) were monitored for absorbance at 280 nm. Fractions I-XII, as indicated, were collected and ultrafiltered through YM-10 Amicon membrane, the flowthroughs were collected, and lyophilized for further analysis and purification. The bars marked with Vf and VIf represent the flowthrough fractions obtained after ultrafiltrations through membrane.

Isolation and purification of peptides from crude

twelve fractions and ultrafiltered through membranes

with a molecular weight cut-off of 10,000. The

flow-venom of Taiwan habu.

Fractionation of T.

mucros-quamatus venom on a TSK DEAE-650 anion-exchange

throughs from ultrafiltration were collected,

concen-trated and examined for their UV and fluorescence

column (Fig. 1) using ammonium bicarbonate as

elu-tion buffer yielded an eluelu-tion pattern superior to that

emission spectra. Flow-through fractions Vf

and VIf

,

obtained from fractions V and VI in Fig. 1, which were

reported previously (19). It is to be noted that fractions

I-V, eluted at the initial buffer, contained multiple pro-

eluted by the initial buffer and the first-step gradient

respectively, were shown to have a high absorbance at

teolytic enzymes with molecular weights of about

24,000 as judged by SDS-gel electrophoresis (data not

280 nm and an emission maximum of about 358 nm

detected on a Hitachi F-4010 fluorescence

spectropho-shown), corresponding to those of metalloproteinases

reported by our group (19). Fractions VI-X, eluted at

tometer when excited at 295 nm.

High performance liquid chromatographies (HPLC) of

the first-step gradient, possessed also multiple

protein-ases with molecular weights in a range of 27,000 to

these two flow-throughs were further carried out on a

Hitachi liquid chromatograph (Fig. 2) in order to purify

32,000, corresponding to a family of serine proteases

described in the previous report (32). We collected

these fractions in a homogeneous form. HPLC showed

TABLE 1

Amino Acid Compositions, Sequences, and Fluorescence Emission Wavelengths of Small Peptide Fractions

Compositiona _{Sequence residue}d

lemic Asx Glx Lys Trp (nm) 2nd 3rd Vf — 44.6% 45.0% n.d.b 358.2 Lys Trp VIf-a 46.8% 50.2% — n.d. 358.0 Asn Trp VIf-b — 95.5% — n.d. 358.0 Gln Trp L-Trp 354.4

a_{Data of amino acid analyses are expressed as mol%, and V}

f, VIf-a and VIf-b are fractions collected from reverse-phase HPLC.

b_{n.d., not determined.}

c_{lemi, the emission wavelength with the highest fluorescence yield using excitation at 295 nm. L-Trp is used as the standard for}

wavelength comparison.

d_{The N-terminal first residue was blocked and the second and third residues were determined by the treatment of peptides with}

Vol. 248, No. 3, 1998 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 2. High-performance liquid chromatographies (HPLC) on a reverse-phase C18column of fractions Vf and VIf obtained from TSK

DEAE-650(M) column after ultrafiltrations. About 50ml (2 mg/ml) of fraction Vf (A) or fraction VIf (B) was injected into a C18column

equilibrated with 5% acetonitrile in 0.1% TFA, and the bound materials were eluted with a linear gradient of 5-95% acetonitrile in 0.1% TFA for 30 min. After HPLC the fraction VIf eluted as two peaks of equal amounts, denoted as VIf-a and VIf-b respectively as indicated in (B), whereas fraction Vf eluted as one major homogeneous peak (A).

that fraction VIf

can be split into two peaks at elution

in peak VIf-a and VIf-b produced molecular ions with

times of 19.4 and 19.9 min with equal yields, denoted as

a mass/charge ratio of about 430 and 444 (Fig. 3),

corre-VIf-a and VIf-b respectively (Fig. 2B). Fraction Vf

was

sponding to those molecular masses of

pyroGlu-Asn-eluted as one major peak (Fig. 2A).

Trp and pyroGlu-Gln-Trp respectively. These two

pyro-glutamate tripeptides are identical to those isolated

previously from snakes of A. halys blomhoffii, B.

jara-Amino acid compositions and mass spectrometry.

raca, C. adamanteus and T. flavoviridis (12). On the

Amino acid analyses showed that peak Vf

contained

other hand, the molecular ion emitted from the sample

equal amounts of Glx (Glu or Gln) and Lys. Likewise,

of peak Vf

showed a mass/charge ratio of about 444

peak VIf-a contained Glx and Asx (Asp or Asn), however

(Fig. 3), corresponding to a peptide of pyroGlu-Lys-Trp,

only Glx was identified in peak VIf-b (Table 1).

FIG. 3. Mass spectrometry of small peptides in peaks Vf, VIf-a and VIf-b from HPLC. Approximately 1mg of each peptide sample dissolved in 5% acetic acid was analyzed by mass spectrometry (FAB-MS). The X-axis designates the mass/charge ratio, which is representative of the molecular weights of these peptides. The number above the major peak in each spectrum shows the measured molecular weight of major molecular ion peak by FAB-MS.

reported before. It is concluded that these small pep-

judged by the fluorescence spectra and

microsequenc-ing after deblockmicrosequenc-ing of the N-terminal residue (Table

tides are all composed of three amino acids and

Vol. 248, No. 3, 1998 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

same venom of Taiwan habu, were employed to

in-vestigate the inhibitory activities of these tripeptide

samples using a fluorescence substrate, fluorescein

thiocarbamoyl (FTC)-casein. In the absence of

tri-peptides, FTC-casein was almost completely

hy-drolyzed after incubation with metalloproteinases at

37

7C for 2 h, whereas the FTC-caseinolytic activity

of metalloproteinases was dose-dependently

inhib-ited by increasing the tripeptide inhibitors. As

shown in Fig. 4, there are no much activity difference

among three pyroglutamate tripeptides to inhibit

each metalloproteinase within the concentration

range used in this study. The IC50

(concentration for

50% inhibition) of these tripeptides for TM-1, TM-2

and TM-3 were estimated to be 0.20 mM, 0.25 mM

and 0.58 mM for pyroGlu-Lys-Trp, respectively; 0.37

mM, 0.46 mM and 0.95 mM for pyroGlu-Asn-Trp,

respectively; and 0.24 mM, 0.31 mM and 0.71 mM

for pyroGlu-Gln-Trp, respectively. These IC50

values

are generally weaker than most conventional

prote-ase inhibitors of larger sizes such as trypsin or

chy-motrypsin inhibitors. However they still exhibit the

high specificity for these venom metalloproteinases

and not the thrombin-like serine proteases present

in the same venom (32, data not shown). Moreover

when comparing IC50

of these tripeptides with that

of citrate ion (17), another endogenous proteinase

inhibitor present in snake venoms, these tripeptides

are actually 100-fold more effective than citrate. It

is also noteworthy that TM-1 and TM-2 appear to

be slightly more susceptible to tripeptide inhibitors

than TM-3, which corroborates our previous

conclu-sion that TM-1 and TM-2 are more closely related

to each other than to TM-3, and both of them are

distinguishable from TM-3 in their partial protein

sequences (19).

In conclusion we have isolated and characterized

three endogenous tripeptides which show specific

in-hibition against the strong metalloproteinases

pres-ent in the same venom. Characterization of these

FIG. 4. Dose-dependent inhibition of T. mucrosquamatus

metal-pyroglutamate tripeptides may provide a molecular

loproteinase activities by three pyroglutamate tripeptides. The

per-basis to account for the resistance of snake venom

centage inhibition of FTC-caseinolytic activity (1 mg/ml) by

tripep-components to these metalloproteinases in vivo.

Fur-tides on TM-1, TM-2 and TM-3 metalloproteinases isolated from T.

mucrosquamatus venom (18,19), was measured and estimated ac-

_{ther elucidation of the mechanistic aspects of these}

cording to that as described in Materials and Methods. Incubation

_{tripeptides in relation to the inhibition of venom}

me-of FTC-casein substrates with each metalloproteinase isme-oform in the

talloproteinases could open avenues for exploring

absence of tripeptides was used as control for 100% FTC-caseinolytic

the potential of using these peptide inhibitors as

activity and each data point was the average of two experiments.

modulators in alleviating the toxic side effect of

bleeding or hemorrhage associated with several

venom protease-based antithrombotic agents.

VIf-b and Vf

from 250 mg crude venom were estimated

to be about 0.65 mg, 0.55 mg and 0.42 mg, respectively,

REFERENCES

using a molar extinction coefficient for tryptophan of

5600 M

01

_cm

01

_{at 280 nm.}

1. Jimenez-Porras, J. M. (1968) Ann. Rev. Pharmacol. 8, 299 – 318.

Inhibition of metalloproteinase activities by tripep-

_{2. Markland, F. S. Jr. (1988) in Hemostasis and Animal Venoms}

tide inhibitors.

Three metalloproteinase isoforms,

(Pirkle, H., and Markland, F. S., Eds.), pp. 149 – 172, Dekker,

New York.

TM-1, TM-2 and TM-3 isolated previously from the

3. Tu, A. T. (1982) in Rattlesnake Venoms: Their Actions and Treat- 18. Huang, K.-F., Hung, C.-C., and Chiou, S.-H. (1993) Biochem.

Mol. Biol. International 31, 1041 – 1050.

ment (Tu, A. T., Ed.), pp. 247 – 312, Dekker, New York.

4. Ownby, C. L., Bjarnason, J., and Tu, A. T. (1978) Am. J. Pathol. 19. Huang, K.-F., Hung, C.-C., Pan, F.-M., Chow, L.-P., Tsugita, A., 93, 201 – 218. and Chiou, S.-H. (1995) Biochem. Biophys. Res. Commun. 216,

223 – 233. 5. Hokama, Y., Iwanaga, S., Tatsuki, T., and Suzuki, T. (1976) J.

Biochem. 79, 559 – 578. 20. Laemmli, U. K. (1970) Nature 227, 680 – 685.

6. Strydom, D. J. (1977) Biochim. Biophys. Acta 491, 361 – 369. 21. Chiou, S.-H., Hung, C.-C., and Huang, K.-F. (1992) Biochem.

Biophys. Res. Commun. 187, 389 – 396.

7. Ritonja, A., Meloun, B., and Gubensek, F. (1983) Biochim.

Bio-phys. Acta 746, 138 – 145. _{22. Bradford, M. (1976) Anal. Biochem. 72, 248 – 254.}

8. Bieber, A. L. (1979) in Handbook of Experimental Pharmacology _{23. Twining, S. S. (1984) Anal. Biochem. 143, 30 – 34.} (Lee, C. Y., Ed.), Vol. 52, pp. 295 – 306, Springer, Berlin.

24. Ouyang, C. (1957) J. Formosan Med. Assoc. 56, 435 – 448. 9. Pirkle, H., and Stocker, K. (1991) Thromb. Haemostas. 65, 444 –

25. Meaume, J. (1966) Toxicon 4, 25 – 58. 450.

26. Jimenez-Porras, J. M. (1970) Clin. Toxicol. 3, 389 – 431. 10. Markland, F. S. (1991) Thromb. Haemostas. 65, 438 – 443.

27. Esnouf, M. P., and Tunnah, G. W. (1967) Brit. J. Haematol. 13, 11. Kini, R. M., and Evans, H. J. (1992) Toxicon 30, 265 – 293.

581 – 590. 12. Kato, H., Iwanaga, S., and Suzuki, T. (1966) Experientia 22, 49 –

28. Bell, W. R., Pitney, W. R., and Goodwin, J. F. (1968) Lancet i, 50.

490 – 493. 13. Lo, T. B. (1972) J. Chin. Biochem. Soc. 1, 39 – 46.

29. Cercek, B., Lew, A. S., Hod, H., Jano, J., Lewis, B., Reddy, 14. Robeva, A., Politi, V., Shannon, J. D., Bjarnason, J. B., and Fox,

K. N. N., and Ganz, W. (1987) Thromb. Res. 47, 417 – 426. J. W. (1991) Biomed. Biochim. Acta 50, 769 – 773.

30. Pollak, V. E., Glas-Greenwalt, P., Olinger, C. P., Wadhwa, N. K., 15. Francis, B., and Kaiser, I. I. (1993) Toxicon 31, 889 – 899.

and Myre, S. A. (1990) Am. J. Med. Sci. 299, 319 – 325. 16. Freitas, M. A., Geno, P. W., Sumner, L. W., Cooke, M. E.,

Hudi-31. Bjarnason, J. B., and Fox, J. W. (1994) Pharmacol. Ther. 62, burg, S., Ownby, C. L., Kaiser, I. I., and Odell, G. V. (1992)

Tox-325 – 372.

icon 30, 461 – 464.

17. Francis, B., Seebart, C., and Kaiser, I. I. (1992) Toxicon 30, 32. Hung, C.-C., Huang, K.-F., and Chiou, S.-H. (1994) Biochem.

Biophys. Res. Commun. 205, 1707 – 1715.