and in Vivo - Semantic Scholar

THEJOURNAL OF BIOLOGICAL

CHEMISTRY

Vol. 267, No. 20,Issue of July 15, p p . 14118-14121, 1992 Printed in U.S.A.

The All-D-configuration Segment Containing the IKVAV Sequence of Laminin A Chain Has Similar Activities to the All-L-peptidein Vitro and in Vivo* (Received for publication, March 9, 1992)

Motoyoshi Nomizu$, Atsushi Utani§, Norio Shiraishig,Maura C. Kibbeyg, Yoshihiko Yamadas, and Peter P. Roller$T From the $Laboratory of Medicinal Chemistry, DevelopmentalTherapeutics Program, Division of Cancer Treatment, National Cancer Institute and the §Laboratory of Developmental Biology, National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20892

Laminin is a basement membrane glycoprotein that era1 active sites of laminin were identified by using proteolytic has diverse biological activities. A sequence on the A peptides from the native protein andby using synthetic pepchaincontaining IKVAV (Ile-Lys-Val-Ala-Val)has tide segments (4-9). In particular, an IKVAV- (Ile-Lys-Valbeen shown to promote neurite outgrowth,cell adhe- Ala-Val) containing sequence of the A chain of laminin was sion, and tumor growth and metastasis. Here we have found to be one of the most potent sites. IKVAV segmentdetermined the structural requirements of this syncontaining peptides have been found to play a role in cell theticpeptide for biological activity.Twelve-amino adhesion and neurite growth promotion (6, IO), in the proacid-long all-L- (LAM-L) and all-D-peptide (LAM-D) motion of metastasis and collagenase IV production (11, 12), segments as well as an alternatingD- and L-amino acid- in the stimulation of angiogenesis,’ and in cell growth (14) containingpeptide (LAM-DL), whichincluded the IKVAV sequence (residues 2097-2108), were synthe- and tumor promotion(15). Current evidence suggests that the sized. Circular dichroism spectral analysis revealed a region of laminin encompassing the A chain IKVAV site, in mirror image conformationof LAM-Dand LAM-Lwith combination with segments of the B1 and B2 chains, formsa mainly &sheet and to a minor extent a-helical struc- triple-stranded a-helical coiled-coil structure (8, 16, 17). Alhave ture. LAM-DLdid not exhibit any significant ordered though a number of integrin and non-integrin receptors conformational features. LAM-D and LAM-L showed been reported as laminin-binding proteins (18, 19), it is unsimilar cell attachment activities forrat pheochromo- clear whether multiple receptors interact with the IKVAV cytoma cells(PC12), whereasLAM-DLwas inactive. A site because a 110-kDa protein has been described that has peptideanalogwithrandomized IKVAV sequence limited distribution (20). Presumably, the structural conforactive IKVAV site plays an important (LAM-RM) was also inactive. A similar trend wasob- mation surrounding the served in competition experimentsof the four peptides role for the specificity of the receptor-ligand interactions. with laminin in analogous cell attachment assays. in In Substitution withD-configuration amino acids into peptide vivo experiments, both LAM-D and LAM-L were ca- ligands has been widely used for probing the conformational pable of increasing tumor growth when subcutaneously requirements of ligand-receptor interactions.For example, injected into mice with murine melanoma cells the low and high affinity receptorsof atrial natriuretic peptide B16F10. Results indicate that theconformational sta- can be distinguished by systematic D-amino acid replacements tus of the IKVAV domain is a contributing factor in in the ligand (21). Others reported oncomparable antimicrodetermining the biological activity but that there is no bial and channel-forming activitiesof all-L- and all-D-amino strict requirement for a specific chirality. There is a acid-containing surface-activepeptides, such as magainin, likely sequence specificityto theIKVAV region. cecropin and melittin, where these agents presumably act as amphiphilic a-helical peptides inlipid membranes with minimally stringent chiral environments (22,23). In studies aimed at probing the natureof sense/antisense peptide interactions, Laminin is a large (Mr = 900,000) extracellular multido- it was found that all-D-configuration amino acid-containing main glycoprotein. It is amajor component of basement peptides showed equally good affinities toward their complemembranes, which underlie all epithelia and surround muscle, mentary peptides as their all-L-analogs (24). peripheral nerve, andfat cells (1, 2). The majorform of In this structure activity study we wish to report the cell laminin consists of A, B1, and B2 chains, which are held attachment activity and tumor-promoting activity of all-L-, together by disulfide bonds, and hasa cruciform shape, which alI-D-, and alternating L- and D-amino acid-containing 12can be seen when examined by electron microscopy (3). Lam- mer peptides encompassing theIKVAV region of the laminin inin has multiple biological activities including promotionof A chain. For comparison, a peptide containing a randomized cell adhesion, growth, differentiation, migration, neurite out- IKVAV region was also evaluated. growth, tumor metastasis, andcollagenase IV induction. Sev* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ll To whom correspondence should be addressed: Laboratory of Medicinal Chemistry, NCI, NIH, Bldg. 37, Rm. 5C02, Bethesda, MD 20892. Tel.: 301-496-3597; Fax: 301-402-2275.

MATERIALS AND METHODS

Synthesis of Peptides-The peptides were synthesized by the solidphase method (25) using a Biosearch automated peptide synthesizer model9600based on t-butoxycarbonyl strategy. Deprotection and D. S. Grant, J. L. Kinsella, R. Fridman, B. A. Piasecki, Y. Yamada, M. Zain, and H. K. Kleinman, unpublished data.

14118

14119

Laminin All-D-peptideSegment

El

cleavage from the resin was achieved by treatment with anhydrous M-L : H-Ala-Ala-Ser Ile-Lys-Val-Ala-Val Ser-Ala-Asp-Arg-OH HF (26), and the crude peptideswere purified by gel filtration (SephM-0 : H - h - h - a U - I y I - y n l - U - y n l &pU-&-&g-OH adex G-10, eluted with 1 M aqueous acetic acid) and reverse phase M - D L : H-Ala-Ah-Ser U-Lys-ynl-Ala-ynl Ser-Ah-Asp-&g-OH HPLC' (using Vydac 5Cls column and a gradient of water/acetonitrile containing 0.05% trifluoroacetic acid). Identity of the synthetic pepM W r : H-Ala-Ala-Ser Val-Val-Ilc-Ala-Lyt Ser-Ala-Asp-Arg-OH tides was confirmed by fast atom bombardment massspectral analysis, and amino acid analysis. Massspectra were measured in glycerol FIG. 1. List of synthetic peptides. LAM-L, all-L-configuration matrix ona VG 7070E-HF double focusing mass spectrometer. Amino laminin A chain segment (residues 2097-2108) 12-mer; LAM-D, allacid analyses were performed at the Protein Structure Laboratory, D-configurationlamininsegment; LAM-DL,alternating L- and DUniversity of California, Davis,CA. configurationlamininsegment; and LAM-RM,all-L-configuration Circular Dichroism (CD)Measurements-The peptides were dis- laminin segment withthe IKVAV segmentscrambled.Observed mass solved in either 50 mM ammonium phosphate buffer (pH 7.0), 50% spectral protonated molecular ions forthe four peptides were respec(v/v) trifluoroethanol (TFE) in the same buffer, or MeOH (0.20 mg tively, 1187.6,1187.9,1187.9, and 1187.3 (calculated isotopic MH' of peptide/ml). Maximumsolubility of the peptides was approxi- 1187.7). Note: D-configuration amino acidsare underlined. mately 1mg/ml buffer. Spectra were recorded from 260 to 185nmon a Jasco model J500A/DP-501N CD spectropolarimeter in Hellma cells, witha 1-or atom bombardment mass spectrometry, and amino acid analysis. The stereochemical antipodal relationship of the all+ 0.1-mm path length, at room temperature. by CD Cell Adhesion Assay-Mouse laminin was extracted and purified and all-D-configuration peptides was demonstrated from the Engelbreth-Holm-Swarmtumor, using methods previously spectropolarimetry (Fig. 2 A ) . described (1). The conformational tendencies of the four peptides were Rat pheochromocytomaPC12cells(obtainedfrom G. Guroff, further examined by CD spectropolarimetry. The all-L-conNICHD, NIH) were used for cell adhesion assays.Cell adhesion was the all-D-peptide performed in 96-well dishes (Immulon2, Dynatech, Alexandria,VA). figuration native segment peptide (LAM-L), (LAM-D), and therandomizedpeptide(LAM-RM) were Various concentrations of either laminin or peptides, which were dissolved in H20, were added to the dishes and dried at room tem- found to be largely in the random conformation in aqueous perature overnight. The laminin- or peptide-coated wells were gently buffer (Fig. 2, A and B ) . In 50% TFE, the LAM-L peptide at washed once withPBS, and any uncoated surfaces were subsequently 168 PM (0.2 mg/ml) concentration was determined to be 45% blocked with 0.1 ml of Dulbecco's minimum essential medium/well containing 1% heat-inactivated bovineserumalbuminfor 1 h at (3-sheet, 7% a-helical, and 47% random coil, using the Prov(28). A similar but 37 "C. Then the solution was removed gently, and the cells (5 X lo4) encher spectral deconvolution program were added to each well in a total volumeof0.1mlof serum-free mirror image CD spectrum was obtained with LAM-D in 50% Dulbecco'sminimum essential medium containing 0.02%bovine TFE. At five times higher concentration (1 mg/ml, 840 PM semm albumin and incubated for 30 min at 37 "C in 5% Con, 95% peptide) in 50% TFE, however, the same peptide (LAM-L) air. Wells were then washed with PBS to remove unattached cells. exhibited a negative ellipticity with a maximum at 219 nm Attached cells werestained by 0.1% crystal violet followed by measand a small broad positive ellipticity peak below 205 nm, urement of A at 560 nm. indicating a largely @-sheet structure and probably extensive Inhibition Assay-96-well dishes (Immulon 2) were coated with laminin (1 pglwell) and blocked with Dulbecco's minimum essential aggregation (Fig. 2C). In MeOH, the spectrum of the LAM-L medium and inactivatedbovineserumalbuminasabove.Various peptide was also indicative of (3-sheet structure with negative amounts of peptides were mixed with PC12 cells and added to each ellipticity maximum a t 216 nm and a distinct positive maxiwell in the total volume of 0.1 ml. After incubation for 30 min, wells mum at 195 nm. The spectrum of the randomized peptide were washed with PBS to remove unattached cells. Attached cells (LAM-RM) at 168 p~ concentration (0.2 mg/ml) in the buffer were counted as described above. Coating Efficiency-Laminin (0.2 and 1 pg/lOO pllwell) and the as well as in 50% aqueous TFE is similar to the spectrumof synthetic peptides (1 pg/lOO pllwell) were added to 6 wells each of LAM-L, indicating that scrambling of the five central residues the 96-well dishes and dried at room temperature overnight. Half of does not affect the conformational tendency. On the other laminin- or peptide-coatedwells were gently washedthree times with of the LAM-DL, which containedan PBS (200 p1 each). Coated laminin and the peptides were extracted hand,thespectrum twicewith PBS (200 and 100 pl) containing 3% Triton X-100. alternating L- and D-amino acid configuration,does not show Fluorescamine in acetone (20 p1,l mg/ml) was addedto the extracted any distinct ordered characteristics. solutions (total 300 rl), and fluorescence emission at 475nmwas Cell attachment activity of the four peptideswas examined determined using a SLM Instruments, Inc. model 8000c spectrofluo- a t various concentrations using rat pheochromocytoma PC12 rometer, 45-plcuvette, with excitation at 396 nm. Coating efficiencies cells(Fig. 3). All-D-peptide (LAM-D) was found to have (%) were determinedusingfluorescence intensity of PBS-washed approximately 80%cell attachment activity of LAM-L; a t t h e samples relativeto non-washed samples fromthe wells. were found to be inactive Tumor Cell Growth Assay-Matrigel, a mixture of basement mem- same time LAM-DL and LAM-RM brane components,was prepared from the transformed mouse Engel- under the same experimental conditions. In competition asbreth-Holm-Swarm tumor as described(27). says, LAM-L, LAM-D, LAM-DL, and LAM-RM were tested MousemelanomaB16F10cells (1 X lo5) weremixedwith the for their inhibitory activity on PC12 cell attachment to a peptides (0.5 mg) and Matrigel (7 mg) and injected subcutaneously into C57BL6N mice (3-month-oldfemales). Tumorvolume (length X laminin substrate (Fig. 4). In this assay, LAM-D alsoshowed was about 80% as potent as LAMwidth X height) was measured with calipers 7, 14, 17,19, and 20 days an inhibitory activity that L at 5 pg/well, whereas LAM-DL was foundto be ineffective. after injection. Mice were sacrificed on the 20th day. The randomized peptide, LAM-RM, showed only very weak inhibitory activity. RESULTS It has been shown recently that subcutaneous co-injection Four peptides, which are listed in Fig. 1, were prepared by of reconstituted basement membrane matrix Matrigel with the solid-phase synthetic strategy (25). The purity and iden- various tumor cells can stimulate tumor growth (15, 29, 30). tity of the peptides were confirmed by analytical HPLC, fast In our experiments, the LAM-L and LAM-D peptides were co-injected subcutaneously with B16F10 melanoma cells into The abbreviations used are: HPLC,high performance liquidchro- mice, with and without the basement membrane matrix Mamatography; PBS, phosphate-buffered saline;TFE, trifluoroethanol; LAM-L, all-L-configurationlaminin A chain segment; LAM-D, all-D- trigel. As shown inFig. 5A, LAM-L alonewas able to stimulate configuration laminin segment; LAM-DL,alternating L- and D-COn- more tumor growth than that observed in the absenceof the peptide when subcutaneously injected with B16F10 melanoma figuration laminin segment; LAM-RM, all-L-configuration laminin segment with the IKVAV segment scrambled. cells into mice. LAM-D also stimulated the tumor growth with

14120

Laminin All-D-peptideSegment

IB 1

-I5 uLA"L

Y - O L I n 50% TFE

I n buffer l

I

190230

C

210

250

190

,

,

210

,

,

,

230

,

250

WAVELENGTH (nm) FIG. 2. CD spectra of the synthetic peptides. The concentration of all the peptides was 0.2 mg/ml (168 p

of LAM-L, which was 1 mg/ml (840 p

~ )in,

~ )except ,

for the spectrum

50%aqueous TFE illustrated in panel C. deg, degree.

100

80 n e \ .

U

0.2

e.5

I

5

PEPTIDE COlTED ( I r g h l l )

FIG.3. Attachment of PC12 cells to peptide-coated plates. Cell attachment activity with 0.2 pg of laminin/well is taken as 100%. The data are given as means rt S. D. (n = 10). Coating efficiencies of laminin (0.2 pglwell) and the synthetic peptides (1 pg/well) were 71.0% (laminin), 7.2% (LAM-L),6.1% (LAM-D),3.0% (LAM-RM), and 2.3% (LAM-DL).

a

60

I :

V

U c

5

VI -I -I

40

w

V

20

similaractivitycompared withLAM-L.Moreover, LAM-L and LAM-D enhanced tumor growtha to similar degree when subcutaneously co-injected with Matrigel (Fig. 5 B ) . DISCUSSION

In this paper, we have shown that the synthetic all-Dconfiguration amino acid-containing 12-mer peptide, encompassing the IKVAV sequence of the laminin A chain, exists in a mirror image conformation to the all-L-peptide analog. Interestingly, the D-peptide was as biologically active as the a major L-peptide. In 50% TFE, LAM-L and LAM-D exist to extent in a ,&sheet (45%)conformation and toa minor extent (7%) in an a-helical conformation. LAM-DL,which consists of L- and D-amino acids at alternating positions, possesses only random coil structure and does not have any significant cell attachment activity. These results suggest that conformational statusof the IKVAV domain isa contributing factor in determining the biological activity; however, there i s no strict requirement for a chiral environment. One L-configuration peptide was evaluated in which the IKVAV segment was scrambled, but the molecule retained the same hydropathic characteristicsoverall. This peptidewas shownto have similar /3-sheetla-helical conformation tendencies to the native sequence. In contrast, it showed only negligible cell attachment activity (Fig. 3), indicating that there is some sequence specificity in the laminin-receptor interaction. Cell attachment activity of peptides was measured in comparison with native Engelbreth-Holm-Swarm laminin. When

0

5

IO

PEPTIDE ADDED @g/weI 1)

FIG.4. Effect of the all-L-peptide (LAM-L),the all-D-peptide (LAM-D),and the alternating D- and L-amino acids containing peptide (LAM-DL) on blocking cell (PC12) attachment to a laminin substrate. The data are given as means f S.D. ( n = 10). Coating efficiency of laminin (1 pglwell) was 49.4%. a, at ahigher concentration of LAM-DL(15 pg/well) cell attachment efficiency was found to be 101 ? 6%. taking into account the relative coating efficiencies of the IKVAV segment peptides to laminin, the molar ratio was approximately 500 (Fig. 3). The considerably higher cell attachment activity of native laminin, however, is not surprising. First,most likely, laminininteracts with cell surface receptors through ancillary sites also. Second, the structural conformation of the surrounding region in the vicinity of the IKVAV site appears to be important. The IKVAV region of laminin has beenproposed earlier to be part of a triplestranded a-helicalcoiled-coil structure consistingof the three chains, A, B1, and B2. Based on the criteria of Parry (31), Beck et al. (8) predicted that the probability of forming a coiled-coil structure is the highest for the B1, intermediate

Laminin All-D-peptide Segment

14121

moting activity of these peptides could be the result of increased collagenase IV levels. The IKVAV-containing peptide also has mitogenic activity of certain cell types4andcan promote angiogenesis.' Therefore, the enhanced tumor formation associated with the peptides may also be a manifestation of the elevated cell growth rate of the transformed B16F10 cells and increasedblood supply. It is generally knownthat D-amino acid-containing peptides are considerably more resistant to proteolytic cleavage than their L-configuration analogs. Since the IKVAV-containing peptide of the laminin A chain has a variety of activities, the LAM-D peptide is potentially useful for therapeutic applications such asnerve regeneration and angiogenesis.

I-

Acknowledgments-we thank Dr. J. S. Driscoll, National Cancer Institute, andDr. H. K. Kleinman,National Institute of Dental Research, for encouragement and critical readingof the manuscript. We thank Dr. Yoshiaki Omata, National CancerInstitute, for assisting in the coating efficiency determinations. REFERENCES 1. Timpl, R., Rohde, H., Robey, P. G., Rennard, S. I., Foidart, J.-M., and

16

15

16

17

M

19

20

DAYS POST INJECTION

FIG. 5. Effect of the LAM-L and the LAM-D on B16F10 melanoma tumorgrowth. A, in absenceof Matrigel;B, in presence of Matrigel ( M G ) .Frequent tumor measurements were taken (length X width X height) with calipers,and the data are expressedas tumor

volume (rnrn3). The data are given as means k S.D. (n = 3-6). a and b indicate tumor volumes that are significantly different from the corresponding tumor volume obtained the in absence of peptide using the Student's t test. a, = p < 0.05;b, = p < 0.1. for the B2, and the lowest for the A chain; and very recent results confirmed these predictions (32). The low a-helical content (7%) of the LAM-L and LAM-D peptides in ourCD analysis agrees with their prediction. Our preliminary data using 30-amino acid-long synthetic peptides (13)also suggest that the addition of the A chain t o a heterodimer of B1 and B2 chains does not increase a-helicity but substantially increases the thermal stabilityof the resulting disulfide-linked h e t e r ~ t r i m e r .Recently, ~ severalall-D-peptides(cecropin A, magainin, and melittin) were shown to exhibit antibacterial potency nearly identical to that of the all-L-enantiomer (22, 23). These peptides possessed an a-helical conformation, and this activitywas caused bychannel forming on thecell surface. The IKVAV-containing peptides of various lengths have been shown t o have a variety of biological functions such as the promotion of cell adhesion, migration, neurite outgrowth, and the increase of metastatic potential of tumor cells (6, 7, 11, 12,14, 15). Theseactivitiesare likely tobemediated through the interaction of the peptides witha cellular receptor(s). Although such a receptor has not been clearly identified, our resultssuggest that thereceptor should recognize the conformations associated with both D- and L-peptides. This is rather unique because no receptor has been reported to interact with a ligand consisting entirely of D-amino acids. LAM-D and LAM-L increase tumor growth of melanoma B16F10 cells in C57BL4N mice. Because an IKVAV-containingpeptide has beenshown to increasecollagenase IV activity and plasminogen activation (11, 12), the tumor-growth-proM. Nomizu, unpublished results.

Martin, G. R. (1979) J. Biol. Chem. 254,9933-9937 2. Chun , A E Jaffe, , R., Freeman, I. L., Vergnes, J. P., Braginski, J. E., and Car!in,'B.'(1979) Cell 16,277-287 3. Engel, J., Odermatt, E., Engel, A,, Madri, J., Furthmayr, H., Rohde, H., and Timpl, R. (1981) J. Mol. Biol. 150,97-120 4. Graf, J., Iwamoto, Y., Sasaki, M., Martin, G. R., Kleinman, H. K., Robey, F. A., and Yamada, Y. (1987) Cell 48,989-996 5. Charonis, A. S., Skubitz, A. P. N., Koliakos, G. G., Reger, L. A,, Dege, J., Vogel, A. M., Wohlhueter, R., and Furcht, L. T. (1988) J. Cell Bid. 107, 1253-1260 6. Tashiro K. Se he1 G. C Weeks B. Sasaki M. Martin G . R. Kleinman, H. K.: anh &m,ha. Y:'(1989) J. diol. C h t k 2 6 4 , 16i74-16i82 7. Tashiro, K., Sephel, G. C., Greatorex, D., Sasaki, M., Shirashi, N.! Martin, G. R., Kleinman, H. K., and Yamada, Y. (1991) J. Cell Phys~ol.146, 451-459 8. Beck, K., Hunter, I., and Engel, J. (1990) FASEB J. 4 , 148-160 9. Yamada, K. M. (1991) J. Biol. Chem. 266, 12809-12812 10. Sephel, G. C., Tashiro, K., Sasaki, M., Greatorex, D., Martin, G. R., Yamada, Y., and Kleinman, H. K. (1989) Biochem. Biophys. Res. Commun. 162,821-829 11. Kanemoto, T., Reich, R., Royce, L., Greatorex, D., Adler, S. H., Shiraishi, N., Martin, G. R., Yamada, Y., and Kleinman, H. K. (1990) Proc. Natl. Acad. Sei. V. S. A. 87,2279-2283 12. Stack, S., Gray, R. D., and Pizzo, S. V. (1991) Biochemistry 30,2073-2077 13. Nomizu, M., Utani, A., Shiraishi, N., Yamada, Y., and Roller, P. P. (1991) J. Chem. SOC. Chem. Commun. 1434-1435 14. Kubota, S., Tashiro, K., and Yamada, Y. (1992) J. Bid. Chem. 267,4285A9RR

15. Swz&y, T. M., Kibbey, M. C., Zain, M., Fridman, R., and Kleinman, H. K. (1991) Cancer Metastasis Reu. 10, 245-254 16. Paulsson, M., Deutzmann, R., Timpl, R., Dalzoppo, D., Odermatt, E., and Engel, J. (1985) EMBO J. 4 , 309-316 17. Deutzmann, R., Aumailley M., Wiedemann, H., Pysny, W., Timpl, R., and Edgar, D. (1990) Eur. J.'Biochem. 191,513-522 18. Mecham, R. P. (1991) FASEB J . 5 , 2538-2546 19. Albelda, S.,and Buck, C. A. (1990) FASEB J. 4,2868-2880 20. Kleinman, H. K., Weeks, B. S., Cannon, F. B., Sweeney, T. M., Sephel, G. C., Clement B., Zain, M., Olson M. 0. J., Jucker, M., and Burrous, B. A. (1991) A h . Biochem. Biophyi. 290, 320-325 21. Scarhorou h, R M , McEnroe, G. A., Arfsten, A., Kang, L.L., Schwartz, K., and fewidki, J. A. (1988) J . Biol. Chem. 2 6 3 , 16818-16822 22. Wade, D., Boman A., Wahlin, B., Drain, C. M. Andreu D. Boman H. G. and Merrifield,'R. B. (1990) Proc. Natl. A c h . Sei. b. d. A. 87,' 47611 4765 23. Bessalle, R., Kapitkovsky, A,, Gorea, A,, Shalit, I., and Fridkin, M. (1990) FEBS Lett. 274, 151-155 24. Fassina, G., Roller, P. P., Olson, A. D., Thorgeirsson, S. S., and Omichinski, J. G. (1989) J. Biol. Chem. 264, 11252-11257 25. Merrifield, R. B. (1963) J. Am. Chem. SOC.85, 2149-2154 26. Sakakibara, S., Shimonishi, Y., Kishida, Y., Okada, M., and Sugihara, H. (1967) Bull. Chem. SOC.Jpn. 40, 2164-2167 27. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., Star, V. L., Cannon, F. B., Laurie, G. W., and Martin, G. R. (1986) Biochemistry 28, 312-318 28. Provencher, S. W., and Glockner, J. (1981) Biochemistry 20,33-37 29. Fridrnan, R., Giaccone, G., Kanemoto T., Martin G. R., Gazdar A. F. and Mulshine, J . L. (1990) Proc. Natl. .&ad. Sci. V.' S. A. 87 669i-6706 30. Fridman, R., Kibbey, M. C., Royce, L. S., Zain, M., Sweengy, T. M., Jicha, D. L., Yannelli, J. R., Martin, G. R., and Kleinman, H.K. (1991) J. Natl. Cancer Inst. 8 3 , 769-774 31. Parry, D. A. D. (1982) Biosci. Rep. 2 , 1017-1024 32. Engel, J., Hunter, I., Schulthess, T., Beck, K., Dixon, T. W., and Parry, D. A. D. (1991) Biochem. SOC.Trans. 19, 839-843

' Y. Yamada, unpublished data.