Proc. Nati. Acad. Sci. USAVol. 81, pp. 1135-1139, February 1984Cell Biology

Tumor cell proteinase visualization and quantification using afluorescent transition-state analog probe

(cancer proteinase/tumorigenesis/metastasis/synthetic proteinase inhibitor)

KAREN A. KOZLOWSKI*, FREDERICK H. WEZEMANt, AND RICHARD M. SCHULTZ*:Departments of *Biochemistry and tAnatomy, Loyola University of Chicago Stritch School of Medicine, 2160 South First Avenue, Maywood, IL 60153

Communicated by Elkan R. Blout, November 7, 1983

ABSTRACT The fluorescent proteinase transition-stateanalog inhibitor, dansyl-L-argininal (DnsArgH), may be a se-lective probe of cysteine and serine-type proteinases in a fibro-sarcoma tumor cell line (HSDM1C1). DnsArgH binds withhigh affinity to proteinases because of its transition-state ana-log properties, and on association it gives a dramatically in-creased fluorescent yield. The DnsArgH binding is inhibitedby the serine proteinase inhibitor diisopropyl fluorophosphateand by the cysteine proteinase inhibitor p-chloromercuriben-zoate. The fluorescence emission appears at its maximumsteady-state yield immediately on addition of DnsArgH to theHSDMIC, fibrosarcoma cells. The immediacy of the DnsArgHreaction supports the contention that DnsArgH binding maybe to cell surface-associated proteinases. Quantification of thecell proteinase concentration, by comparison of the fluores-cence yield obtained from DnsArH interactions with bovinetrypsin and papain, indicates 10- 5 to 10-16 mol of proteinaseper HSDMIC, cell. In fluorescence microscopy, DnsArgH flu-orescence appears distributed throughout the fibrosarcomacell without association to organelles. DnsArgH fluorescencefrom normal fibroblast controls (IMR-90) was found to be sub-stantially lower than in the transformed fibrosarcoma cells,supporting a hypothesis that proteinases have a role in malig-nancy.

Significant evidence exists for the role of proteolytic en-zymes in malignant transformation, tumor cell invasion, andtumor cell metastasis (1-3). For example, treatment of nor-mal cells with proteolytic enzymes can lead to the appear-ance of phenotypic characteristics in the untransformed cellsthat are similar to those of transformed cells (1, 4-7). In addi-tion, proteolytic enzymes secreted from transformed cellsmay degrade host tissues into which the tumor cells invade,and may, through the degradation of surrounding tissue, pro-mote tumor cell metastasis (1-3). Particular proteinases arereported to be secreted in larger amounts by transformedcells than their normal counterparts. These identified pro-teinases secreted by transformed cells include the serine-type proteinase plasminogen activator (1, 2, 8, 9), the lyso-somal cysteine-type proteinase cathepsin-B (1, 10), and col-lagenolytic proteinases (1, 3). The existence of a role forproteolytic enzymes in tumorigenesis and metastasis is sup-ported by experiments in which these processes are shownto be inhibited by inhibitors of proteinase enzymes (1, 2).Whereas a large amount of evidence exists for secretion

into the extracellular environment of relatively high concen-trations of proteinases by tumor cells, recent reports alsodocument the existence of proteinase activity either associ-ated with the outer cell membrane or membrane fractionswithin transformed cells (1, 2, 11-19). Similar to the types ofproteinases reported to be found secreted into extracellularfluid, both a neutral trypsin-like seine proteinase, whichmay be plasminogen activator (1, 2, 11-16), and a cathepsin-

B cysteine proteinase (1, 17) have been reported. A high con-centration of cell surface proteinases may provide an inti-mate mechanism by which the membrane of the transformedcell can both maintain malignant traits within the tumor andpromote the degradation of surrounding tissue during tumorcell invasion and metastasis.

In this communication we report the development of ageneral proteinase fluorescent probe, dansyl-L-argininal(DnsArgH), which is capable of detecting both serine andcysteine proteinases with a primary site specificity towardarginine. Thus, both the neutral trypsin-like serine protein-ases and the cathepsin-B-like cysteinyl proteinase, previous-ly reported to be associated with transformed cells should bedetected with DnsArgH. The fluorescent aldehyde binds toproteinases with a high affinity and selectivity because of thetransition-state-like properties of the aldehyde proteinase as-sociation complex (20-26), and only on association into theenzyme active site does DnsArgH give a high fluorescentyield. We have used DnsArgH to both visualize and quantifythe concentrations of cysteine and serine trypsin-like pro-teinases in fibrosarcoma cells and normal fibroblasts.

MATERIALS AND METHODSDnsArgH Synthesis. DnsArgH was prepared by first reac-

tion (Nw-NO2)-L-arginine methyl ester (Sigma) with dansylchloride by standard techniques to give Na-dansyl-(Nc@-N02)-L-arginine methyl ester, mp 114-116'C. The Ncr-dan-syl-(N'@-N02)arginine methyl ester was reduced with diiso-butyl aluminum hydride (Aldrich) in anhydrous tetrahydro-furan at -200C, according to the procedure of Ito et al. (27),to Na-dansyl-(N@-NO2)-L-argininal. This product wasworked-up as a bisulfite salt similar to the procedure ofThompson (28) in the synthesis of peptidylalaninal com-pounds, and then reacted with semicarbazide HCl in sodiumacetate at 80'C to protect the aldehyde function as semicar-bazone, yielding Na-dansyl-(N@-N02)-L-arginine semicarba-zone, mp 175-1780C. The Nw-NO2 group was removed byreduction in a Parr Hydrogenator (H2 at 20 psi; 1 psi = 6.89kPa) in methanol/water/acetic acid (16:2:1) over palladiumblack, yielding dansyl-L-arginine semicarbazone acetate, mp143-145°C (crystallized from ethanol/ethyl ether). Finally,the semicarbazone protection was removed in methanol/2 MHCI/formaldehyde, as described for the synthesis of leupep-tin (27). The solution was brought to pH 7.0, and the productwas extracted with methylene chloride, dried, and concen-trated, yielding dansyl-L-argininal-HCl, crystallized fromethanol/ethyl ether, mp 180-185°C. TLC on silica gel Gplates in CHCl3/CH30H/acetic acid (15:5:1) gave two spotsat Rf = 0.54 and 0.58 due to the probable presence of boththe free aldehyde and cyclized carbinolamine tautomericforms of argininal, as described for leupeptin (29).

Elemental analysis calculated for C18H26N503SIC1I (dan-

Abbreviations: DnsArgH, dansyl-L-argininal; iPr2P-F, diisopropylfluorophosphate; ClHgBzO, p-chloromercuribenzoate; Pi/NaCl,phosphate-buffered saline.tTo whom reprint requests should be addressed.

1135

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

1136 Cell Biology: Kozlowski et al.

syl-L-argininalbHCl): C, 50.52; H, 6.12; N, 16.36. Found: C,50.38; H, 6.08; N, 16.54 (Spang Microanalytical Laboratory,Eagle Harbor, MI).1H NMR (2H20) ppm: 1.9 (d, dansyl CH3-N-CH3), 3.1-3.6

(-CH2-), 5.3 (carbinolamine anomeric CH), 7.8 (aromatic).DnsArgH Binding to Trypsin and Papain. Mercuri-papain

(Sigma) was dissolved in 0.1 M sodium acetate, pH 5.5/0.005M cysteine/0.001 M EDTA, and incubated for 30 min to acti-vate the papain. The binding constant for DnsArgH to papa-in was determined against the substrate N-benzoyl-DL-argi-nine p-nitroanilide under first order conditions ([S] < Kn)using the equation below.

Il [I]vuVi

1

in which vu is the observed rate in the absence of DnsArgH,v; is the observed rate in the presence of DnsArgH at con-centration near or greater than Ki, and [I] is the concentra-tion of DnsArgH inhibitor. The binding constant forDnsArgH to bovine trypsin (Sigma) was determined againstthe substrate N-benzoyl-L-arginine ethyl ester using the inte-grated rate equation, under conditions in which the acidproduct of the reaction does not inhibit (30).

Cell Culture. Cultures of HSDM1Cl cells (American TypeCulture Collection) were maintained in Ham's F-10 mediumcontaining 15% horse serum with 2.5% heat-inactivated fetalcalf serum and antibiotics (penicillin and streptomycin).IMR-90 fibroblasts (American Type Culture Collection)were maintained in Eagle's minimal essential medium withnon-essential amino acids in Earle's balanced salt solutioncontaining 10% heat-inactivated fetal calf serum with antibi-otics. Cells were maintained at 37°C in 98% relative humid-ity, 5% C02/95% air.Fluorescence Spectroscopy of Cell Suspensions. Both

HSDM1C, and IMR-90 cells were maintained in serum-freemedium for 24 hr prior to recovery by mechanical means.Harvested cells were tested for viability by the trypan bluedye exclusion test, counted using a hemocytometer, washedwith phosphate-buffered saline (P1/NaCI), and aliquoted inP1/NaCl at cell concentrations of 5 X 104 cells per ml.DnsArgH (0.11 mM in cuvette) was added to suspensions ofeither HSDM1Cl or IMR-90 cells in a mechanically stirredcuvette (Instech, Fort Washington, PA). Fluorescence wasmeasured using a Perkin-Elmer MPF44B spectrofluorimeter.The suspensions were excited at 330 nm and emissions weremeasured between 400 and 600 nm. The emissions werestandardized in each run against the Raman band of waterand normalized against a spectrum of cells alone to correctfor light scattering. The relative quantum yield was obtainedby integrating the area under the emission peak. Other cellsuspensions (5 x 104 cells per ml) of either HSDM1Cl orIMR-90 cells were pre-incubated with 1 mM diisopropylfluorophosphate (iPr2P-F) or 1 mM p-chloromercuriben-zoate (ClHgBzO) for 20 min at room temperature prior toaddition of DnsArgH. Solutions of DnsArgH and cells alonegave emissions that coincided with the buffer base line at thesensitivity setting of the spectrofluorimeter. The spectroflu-orimeter settings were those that gave a full-scale recordingof the DnsArgH-proteinase fluorescence.

Analysis of Glutaraldehyde Effects on DnsArgH-Trypsin.The effect of glutaraldehyde on the fluorescence emission ofDnsArgH-trypsin complexes was evaluated to facilitate theuse of glutaraldehyde-fixed cells for fluorescence microsco-py. Trypsin (0.1 ,M) and DnsArgH (0.11 mM) were stirredin 1.5 ml of P,/NaCl to which 1.5 ml of 0.1 M cacodylate-buffered 2.5% glutaraldehyde was added, and fluorescencewas measured between 400 and 600 nm.

Fluorescence Microscopy. Cells maintained in 75-cm2flasks were recovered by mild (0.1%) trypsinization and re-

plated onto nonfluorescing 5-mm diameter Aklar plasticdiscs. Cells were grown in complete medium but were trans-ferred to serum-free medium for 24 hr prior to the followingprocedure. Discs were examined by inverted phase contrastmicroscopy to determine that the discs were preconfluent.Prior to further treatment, the discs containing the cells inlogarithmic-phase growth were fixed in cold 0.1 M cacody-late-buffered 2.5% glutaraldehyde (pH 7.4) for 1 hr. Thediscs were rinsed with Pi/NaCl (pH 7.4) and placed in Pi/NaCl containing either 1 mM iPr2P-F or 1 mM ClHgBzO for20 min. Other discs were not exposed to either iPr2P-F orClHgBzO prior to incubation in DnsArgH. The discs werethen removed from the incubation solution and placed onglass slides with a drop of P1/NaCl. Coverslips were placedon the slides and they were observed by dark-field fluores-cence microscopy using a UG-1 exciter filter and barrier fil-ter 50 (Zeiss). All fluorescent photomicrographs were ac-quired by standardizing the length of exposure so that semi-quantification of the cell fluorescence would be attained.

RESULTSPeptidyl aldehyde analogs of specific substrates of serineand cysteine proteinase form covalent hemiacetal or thio-hemiacetal complexes of high relative stability with the ser-ine or cysteine nucleophile, respectively, of the proteinaseactive site.

0//

R-C

H

{IA}0

+ X-E:;= R-C :

{IIA}0-

k2X-E -.-R-C-X-E

k-2 HH

{IIB} [1]

OH

R-C-X-E

H

The formation of such a reversible hemiacetal or thiohemi-acetal complex is shown in Eq. 1, in which RCH=O is thepeptidyl aldehyde inhibitor, E-X is the active site nucleo-phile (serine -OH or cysteine YS-) of the respective enzyme,IA is the tetrahedral oxanion intermediate, and IIB is thestable tetrahedral hemiacetal or thiohemiacetal complex be-tween aldehyde and enzyme (21, 26). Structure TIB containsa tetrahedrally liganded carbonyl carbon (Sp3 electronic con-figuration) similar to the proposed structure of the reactivecarbonyl carbon in the transition-state of peptide hydrolysis(20-26). Thus, the complex of peptidyl aldehydes with en-zyme mimics the transition-state configuration in substratehydrolysis, and DnsArgH is a transition-state analog inhibi-tor. Transition-state analog inhibitors are predicted to haveaffinity to enzyme active sites orders of magnitude higherthan even the natural substrates of the enzyme (31, 32). Inagreement with a mechanism of transition-state analog bind-ing for peptidyl aldehyde analogs of specific proteinase sub-strates, the affinity constants found for peptidyl aldehyde as-sociations are 102_105 times greater than those of analogouspeptide substrates to the enzyme (20-26). In accord with thisexpectation, we found an equilibrium dissociation constantof 4.37 x 10-5 M for the binding of DnsArgH to trypsin, and4.03 x 10-6 M for the binding of DnsArgH to papain. A simi-lar high affinity may be expected for the binding of DnsArgHto the trypsin-like serine proteinase and cathepsin B-likeproteinase previously identified in the cell membrane ofsome tumor cell lines.Model Studies with Trypsin and Papain. In our experi-

ments, bovine trypsin serves as a model for the serine-type

Proc. NatL Acad ScL USA 81 (1984)

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

Proc. Natl. Acad. Sci. USA 81 (1984) 1137

and papain as a model for the cysteine-type proteinases. Fig.1 shows a typical fluorescent emission spectra between 400nm and 600 nm on excitation at 330 nm from a solution of 0.1ttM trypsin or papain and 0.11 mM DnsArgH. Under theseconditions, the DnsArgH concentration is in excess of theequilibrium dissociation constants of DnsArgH to the re-spective enzymes, and the enzyme active sites are >90%saturated by DnsArgH. Neither DnsArgH nor the enzymealone has an observable fluorescence under the conditions ofobservation. Thus, DnsArgH only exhibits an observablefluorescence on binding to the enzyme.

iPr2P-F is known to specifically modify the active site ser-ine of serine proteinases, and ClHgBzO specifically modifiesthe active site cysteine of cysteine proteinases (33). Whentrypsin and papain were preincubated with 1 mM iPr2P-F andClHgBzO, respectively, prior to the addition of DnsArgH,no fluorescence was observed. These results support themechanism of Eq. 1, which requires an unmodified serine orcysteine active site nucleophile for formation of the high-af-finity adduct (structure IIB). Thus, DnsArgH binding is ac-tive-site directed, and its binding and fluorescence emissionis completely inhibited by iPr2P-F in the serine proteinaseand by ClHgBzO in the cysteine proteinase.To ascertain whether glutaraldehyde interferes with the

reaction of DnsArgH, the effect of glutaraldehyde on the re-action of DnsArgH with papain and trypsin was studied un-der the same conditions used in the treatment of cells formicroscopy. We observed a decrease offluorescence of only10% in the presence of 2.5% glutaraldehyde. Accordingly,the presence of the nonspecific aldehyde, glutaraldehyde,does not significantly inhibit the binding of the active-sitedirected DnsArgH to papain or trypsin, thus indicating thatglutaraldehyde fixation of cells for microscopic observationdoes not significantly alter binding of DnsArgH to cell pro-teinases.

Quantification of Proteinase Concentrations by Spectro-fluorimetry. Fig. 2 shows the fluorescence emission from asolution of 0.11 mM DnsArgH containing 1 x 105 HSDM1Clfibrosarcoma cells in 2.0 ml of Pi/NaCI and stirred within acuvette of spectrofluorimeter. The emission spectrum fromthe DnsArgH-treated cells is observed to be identical to that

60

50

40

30a) 20C

a) 10

30Cu

a) 40-

30'

20-

a

450 500 550 600

b

o

.W 56 50 60

Wavelength, nm

FIG. 1. Emission spectrum of trypsin-DnsArgH complex (a) andpapain-DnsArgH complex (b). (a) Trypsin (0.1 MiM) in a sodiumphosphate buffer (pH 7.4) containing 0.1 M NaCl (P/NaCl buffer)and 0.11 mM DnsArgH was excited at 330 nm and emissions were

scanned at 400-600 nm. Preincubation of trypsin with 1 mM iPr2P-Fgave no fluorescent increase on addition of DnsArgH. (b) Papain(0.1 MM) in 0.1 M sodium acetate buffer (pH 5.5) containing 0.16 mM

cysteine, 33 ,uM EDTA, and 0.11 mM DnsArgH was excited at 330nm and emissions scanned at 400-600 nm. Preincubation of papainwith 1 mM ClHgBzO gave no fluorescent increase on addition ofDnsArgH.

0 4co30/

45 500 .50 600

Wavelength, nm

FIG. 2. Emission spectrum of DnsArgH-treated fibrosarcomacells. To a suspension of HSDM1C1 cells (0.5 x 105 cells per ml) inPj/NaCl was added 0.11 mM DnsArgH. The cell suspension wasexcited at 330 nm and emissions were scanned at 400-600 nm (-).The cell suspension was preincubated with 1 mM iPr2P-F for 30 minprior to the addition of DnsArgH (.-.-.). The cell suspension waspreincubated with 1 mM CIHgBzO prior to the addition of DnsArgH(---). When the cells were preincubated with both iPr2P-F andClHgBzO for 30 min, there was no fluorescent increase on additionof DnsArgH.

found for papain and trypsin in solution (Figs. 1 and 2). Fur-thermore, the emission spectrum appears immediately onaddition of DnsArgH to the cells and remains at a constantfluorescence for at least 1 hr. No significant fluorescencewas observed under the same conditions and spectrofluorim-eter sensitivity on mixing 1 x 105 fibrosarcoma cells in P,/NaCl without added DnsArgH or from DnsArgH alone. Asolution of 1 x l~ IMR-90 fibroblasts (untransformed cells)in 2.0 ml of Pj/NaCl with 0.11 mM DnsArgH gave 25% of thefluorescence observed for the fibrosarcoma cells. The 4-foldhigher fluorescent yield observed in the fibrosarcoma cellsthan in normal fibroblasts is in agreement with the 4-foldhigher proteolytic activity observed by Hatcher et al. (13) intransformed cell membranes.When the serine proteinase inhibitor iPr2P-F is incubated

with the fibrosarcoma cells for 20 min prior to the addition ofDnsArgH, the fluorescence yield is decreased by '67% (Fig.2). When the cysteine proteinase inhibitor ClHgBzO is incu-bated with the fibrosarcoma cells for 20 min prior to the addi-tion of DnsArgH, the fluorescence is decreased ''56% (Fig.2). The results indicate, within experimental error, approxi-mately equal amounts of cysteine and serine proteinases.When the HSDM1C1 cells are incubated with both iPr2P-Fand ClHgBzO, no fluorescence increase occurs on additionof DnsArgH indicating that no nonspecific binding ofDnsArgH is occurring. The sum of both the iPr2P-F andClHgBzO inhibition of fluorescence is therefore ~100%.The fluorescence yield from 0.1 ,uM trypsin-DnsArgH or

papain-DnsArgH (Fig. 1) is %'~3 times that of the fluorescenceobtained from fibrosarcoma cells at 0.5 x 105 per ml. A sim-ple calculation indicates that -7 x 10-16 mol of trypsin orpapain units are present per fibrosarcoma cell. This calcula-tion is speculative, as it assumes an identical fluorescenceyield from the DnsArgH complex with fibrosarcoma protein-ases as for the DnsArgH complex with trypsin or papain.However, even if the calculation is in error by an order ofmagnitude, the proteolytic enzymes are apparently presentin relatively high concentration in the tumor cell.

Visualization of Proteinases by Fluorescence Microscopy.Dark-field fluorescence microscopy was used to visualizeand confirm the differences in DnsArgH binding to theHSDM1C1 fibrosarcoma cells, the iPr2P-F- and ClHgBzO-treated fibrosarcoma cells, and IMR-90 fibroblasts. A com-parison of fluorescence in the different experiments can bemade in a qualitative way from the photomicrographs ob-tained by fluorescence microscopy, as the fluorescent photo-micrographs were all obtained under identical conditions andexposure times. Fig. 3 shows the fluorescence of DnsArgH-treated fibrosarcoma cells in the absence of iPr2P-F orClHgBzO. The DnsArgH fluorescence appears to be widely

Cell Biology: Kozlowski et aL

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

1138 Cell Biology: Kozlowski et al.

FIG. 3. Fluorescent photomicrograph of HSDM1Cl fibrosarco-ma cells incubated with DnsArgH cells grown to logarithmic-phaseon nonfluorescing Aklar plastic discs. The cells were fixed with glu-taraldehyde and incubated with 0.11 mM DnsArgH at 370C for 1 hr.Discs mounted on glass slides were observed by dark-field fluores-cence microscopy using a UG-1 exciter filter and barrier filter 50(Zeiss). (x100.)

distributed throughout the image of the fibrosarcoma cells.The nuclei of the cells and other organelles do not fluoresce.Fig. 4 shows the DnsArgH-treated fibrosarcoma cells thatare preincubated for 20 min prior to the addition ofDnsArgHwith 1 mM iPr2P-F. Similar results were obtained usingClHgBzO as an inhibitor. In the iPr2P-F- or ClHgBzO-pre-treated cells, the fluorescence appears similarly distributedover the whole cell. However, the fluorescence is significantlydecreased in cells pretreated with iPr2P-F or ClHgBzO, con-firming the result obtained in the spectrofluorimeter (Fig. 2).Fig. 5 shows the fluorescence in DnsArgH-treated IMR-90fibroblasts. The fluorescence from these cells is significantlylower than that obtained from the fibrosarcoma cells. IMR-90 fibroblasts pretreated with inhibitors prior to incubationin DnsArgH gave barely observable cell fluorescence.

DISCUSSIONA fluorescent active-site directed proteinase inhibitor hasbeen used to locate and semi-quantify trypsin-like cysteineand serine proteinase concentrations in fibrosarcoma andnormal fibroblast cells, using both spectrofluorimetric andfluorescent microscopy techniques. The presence of proteo-lytic enzymes in tumor cells, especially associated with its

FIG. 4. Fluorescent photomicrograph of HSDM1Cl fibrosarco-ma cells preincubated with iPr2P-F followed by incubation inDnsArgH. Cells preincubated with 1 mM iPr2P-F for 30 min at roomtemperature prior to incubation in 0.11 mM DnsArgH at 37°C for 1hr. Dark-field fluorescence microscopy. (x1OO.)

FIG. 5. Fluorescent photomicrograph of normal fibroblasts incu-bated with DnsArgH. Cells grown to logarithmic-phase, fixed withglutaraldehyde, and incubated with 0.11 mM DnsArgH for 1 hr at37°C. Dark-field fluorescence microscopy. (x 100.)

cell surface, has been reported (11-17). However, the gener-alized pattern of distribution and the amounts of both serineand cysteine arginine-specific proteinases have not previous-ly been shown. Other techniques localize specific protein-ases both on and within cells (17, 34). However, unlike theuse of immunochemical techniques for the identification of aparticular type of protein, this technique allows one to probethe entire class of arginine-specific cysteine and serine pro-teinases. The known high affinity of peptidyl aldehydes toserine and cysteine proteinases with equilibrium dissociationconstants in the range of 10-5 to 10-10 M (20-26, 31, 32)argues that at the low DnsArgH concentration used in thisinvestigation the proteinase active sites will be near saturat-ed and nonspecific interactions are minimized. In addition,the complete inhibition of observable fluorescence by the ac-tive-site modifying agents iPr2P-F and ClHgBzO in serineand cysteine proteinase, respectively, shows that the bindingof DnsArgH occurs by an active-site directed mechanism.While the localization of the proteinase that interact with

DnsArgH cannot be pinpointed by the current data, the evi-dence may suggest that the association is primarily with out-er membrane-associated proteinases and not with intracellu-lar proteinases for the following reasons.

(i) The appearance of fluorescence in the spectrofluorime-ter is immediate-within the mixing time of the experiment.Furthermore, over the time course of observation (=1 hr),the fluorescence yield between 400 nm and 600 nm remainsconstant. These observations in combination with those dis-cussed below argue for a surface interaction, as it is expect-ed that membrane transport would be slower than the 20-30sec mixing time in the spectrofluorimeter experiment.

(ii) DnsArgH is cationic and would not be expected to rap-idly diffuse across lipid bilayers or membranes.

(iii) Whether iPr2P-F and ClHgBzO primarily react withcell surface proteinases or with both cell surface proteinasesand intracellular proteinases is unclear and may depend onthe physiological state and type of cell (35). However, it hasbeen recently shown by DiStefano et al. (14) that on treat-ment of Walker-256 cells with radiolabeled iPr2P-F, the labelis primarily found associated with proteinase on the outercell membrane. This result may support a similar cell mem-brane reaction for iPr2P-F with the fibrosarcoma cell.

(iv) The distribution of DnsArgH as shown by fluores-cence microscopy may support a cell membrane localiza-tion. If DnsArgH enters the cell, one could expect a local-ized fluorescence due to differences in proteinase concentra-tions associated with various organelles such as lysosomes.Our fluorescence photomicrographs reveal no localization offluorescence in lysosome-like structures even after iPr2P-F

Proc. NatL Acad Sci. USA 81 (1984)

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

Proc. NatL. Acad. Sci. USA 81 (1984) 1139

pretreatment and subsequent incubation in DnsArgH. iPr2P-F-uninhibitable lysosomal enzymes, such as cathepsin B,would be expected to be clearly visualized after serine pro-teinase inhibition with iPr2P-F by fluorescence microscopyhad DnsArgH entered the cell. By comparison, the uptake ofboth pepstatin and peptidyl substrates for lysosomal orrough endoplasmic reticulum localization of cathepsins Dand B, respectively, has been shown by Matthews et al. (34)and Graf et al. (36). Their photomicrographs show a granularintracellular fluorescence analogous to the localization oflysosomes by nonspecific staining (37). In contrast, theDnsArgH fluorescence in our photomicrographs does notappear discretely localized in regions of the cell and is simi-lar to photomicrographs of Sylven et al. (17) who visualizedcathepsin B at or near the cell surface using fluorescent anti-bodies. Similarly, a diffuse cellular fluorescence using anti-bodies to cathepsin D may also suggest cell membrane local-ization (38).Although the above arguments appear to support a cell

membrane localization for the DnsArgH-interacting protein-ases, based on the photomicrographs alone the argumentsmust be considered speculative. The major evidence for anouter surface interaction is the fast time course of the ap-pearance of the high steady-state fluorescence during inter-action with DnsArgH. However, the question of proteinaselocalization requires further investigation.The quantification of proteinase concentration per cell,

based on the fluorescence yield from DnsArgH-trypsin orDnsArgH-papain, indicated 7 x 10-16 to 7 x 10-17 mol ofproteinase per cell. Assuming the DnsArgH-reactive pro-teinases are primarily in the outer membrane, one can calcu-late the approximate percentage of total cell surface area oc-cupied by proteinase. Assuming a 15-gm radius for the cell,the total surface area is calculated to be 3000 /Im2. Then,assuming that each proteinase molecule occupies an area ofradius 15 A on the cell surface, the total surface area occu-pied by 7 x 10-17 mol of proteinase would be 300 pum2 or10% of the surface area. If the radius were slightly larger andthe outer membrane invaginated, the percentage surfacearea occupied by proteinase would be less. In any case, anassumption of a surface membrane localization for theDnsArgH-binding proteinases gives a high concentration ofsurface proteinase in the range of the concentrations foundfor protein lectins on cell surfaces (39) and for cholinergicreceptor protein in the subsynaptic membrane surface of theElectrophorus electroplax (40). A high concentration of pro-teinases may be essential for the migration potential and gen-eral proteolytic functions of the fibrosarcoma cell.

Considering the alternative possibility that the DnsArgHproteinase interaction is primarily intracellular, a similar cal-culation assuming a 15-,um cell radius and 7 x 10-17 mol ofproteinase per cell gives an intracellular proteinase concen-tration of 5 X 10-6 mmol/cm3. Accordingly, whether theDnsArgH-detected proteinase is localized within the cell, onthe outer surface, or associated with both regions of the cell,the fluorescence yield shows cell proteinase molecules to bein relatively high concentration.

In conclusion, we have shown that the fluorescent protein-ase transition-state analog inhibitor DnsArgH can apparentlyboth quantify and visualize serine and cysteine proteinases.The data show a high concentration of proteolytic enzymesin the tumor cell that is 4-fold greater than in control cells,supporting hypotheses arguing for a role of proteinases in ma-lignancy (1-3, 11-18). Proteinase probes such as DnsArgHshould be useful not only in the study of the role of protein-ases in transformed cells but, more generally, in the study ofproteinase activity during cell cycle events.

The authors wish to thank Katheryn Guzzino for her technicalassistance. Portions of this research will be submitted by K.A.K. inpartial fulfillment of the requirements for the Ph.D. degree at Loyola

University of Chicago. This investigation was supported by GrantsHL-24910, CA-34530, and BRSG 5368 from the National Institutesof Health.1. Quigley, J. P. (1979) in Surfaces of Normal and Malignant

Cells, ed. Hynes, R. 0. (Wiley, New York), pp. 247-285.2. Reich, E., Rifkin, D. B. & Shaw, E., eds. (1975) Proteases and

Biological Control, Cold Spring Harbor Conferences on CellProliferation (Cold Spring Harbor Laboratory, Cold SpringHarbor, NY), Vol. 2.

3. Strauli, P., Barrett, A. J. & Baici, A., eds. (1980) Proteinasesand Tumor Invasion (Raven, New York).

4. Nicholson, G. L. (1972) Nature (London) New Biol. 239, 193-197.

5. Burger, M. M. (1969) Proc. Natl. Acad. Sci. USA 62, 994-1001.

6. Burger, M. M. (1970) Nature (London) 227, 170-171.7. Martin, B. M. & Quigley, J. P. (1978) J. Cell. Physiol. 96, 155-

164.8. Rifkin, D. B., Loeb, J. N., Moore, G. & Reich, E. (1974) J.

Exp. Med. 139, 1317-1328.9. Unkeless, J. C., Tobia, A., Ossowski, T., Quigley, J. P., Rif-

kin, D. B. & Reich, E. (1973) J. Exp. Med. 137, 85-111.10. Recklies, A. D., Tiltman, K. J., Stoker, T. A. M. & Poole,

A. B. (1980) Cancer Res. 40, 550-556.11. Steven, F. S., Griffin, M. M., Itzhaki, S. & Al-Habib, A.

(1980) Br. J. Cancer 42, 712-721.12. Quigley, J. P. (1976) J. Cell Biol. 71, 472-478.13. Hatcher, V. B., Wertheim, M. S., Rhee, C. Y., Tsien, G. &

Burk, P. G. (1976) Biochim. Biophys. Acta 451, 499-510.14. DiStefano, J. F., Beck, G., Lane, B. & Zucker, S. (1982) Can-

cer Res. 42, 207-218.15. Spartaro, A. C., Morgan, H. R. & Bosmann, H. B. (1976) J.

Cell Sci. 21, 407-413.16. LaBombardi, V. J., Shaw, E., DiStefano, J. F., Beck, G.,

Brown, F. & Zucker, S. (1983) Biochem. J. 211, 695-700.17. Sylven, B., Snellman, 0. & Strauli, P. (1974) Virchows Arch. B

17, 97-112.18. Pietras, R. J. & Roberts, J. A. (1981) J. Biol. Chem. 256, 8536-

8544.19. Jakens, S. & Black, P. H. (1979) Proc. Natl. Acad. Sci. USA

76, 246-250.20. Thompson, R. C. (1973) Biochemistry 12, 47-51.21. Kennedy, W. P. & Schultz, R. M. (1979) Biochemistry 18,

349-356.22. Chen, R., Gorenstein, D. G., Kennedy, W. P., Lowe, G.,

Nurse, D. & Schultz, R. M. (1979) Biochemistry 18, 921-926.23. Brayer, G. D., Delbaere, L. T. J., James, M. N. G., Bauer,

C. A. & Thompson, R. C. (1979) Proc. Nati. Acad. Sci. USA76, 96-100.

24. Thompson, R. C. & Bauer, C. A. (1979) Biochemistry 18,1552-1558.

25. Westerik, J. 0. & Wolfenden, R. (1972) J. Biol. Chem. 247,8195-8197.

26. Frankfater, A. & Kuppy, T. (1981) Biochemistry 20, 5517-5524.

27. Ito, A., Takahashi, R., Muira, C. & Baba, Y. (1975) Chem.Pharm. Bull. 23, 3106-3113.

28. Thompson, R. C. (1977) Methods Enzymol. 46, 220-225.29. Maeda, K., Kawamura, K., Kondo, S., Aoyagi, T., Tekeuchi,

T. & Umezawa, H. (1971) J. Antibiot. 24, 402-404.30. Cornish-Bowden, A. (1976) Principals of Enzyme Kinetics

(Butterworth, Woburn, MA), pp. 142-152.31. Wolfenden, R. (1972) Acc. Chem. Res. 5, 10-18.32. Leinhard, G. E. (1973) Science 180, 149-154.33. Shaw, E. (1970) in The Enzymes, ed. Boyer, P. (Academic,

New York), Vol. 1, 3rd Ed., pp. 91-146.34. Matthews, I. T. W., Decker, R. S., Hornebeck, W. & Knight,

C. G. (1983) Biochem. J. 211, 139-147.35. Prouty, W. F. & Goldberg, A. L. (1972) J. Biol. Chem. 256,

8536-8544.36. Graf, M., Leemann, U., Ruch, F. & Strauli, P. (1979) Histo-

chemistry 64, 319-322.37. Allison, A. C. & Young, M. R. (1964) Life Sci. 3, 1407-1414.38. Poole, A. R., Dingle, J. T. & Barrett, A. J. (1972) J. Histo-

chem. Cytochem. 20, 261-265.39. Ozanne, B. & Sambrook, J. (1971) Nature (London) New Biol.

232, 156-160.40. Bourgeois, J.-P., Ryter, A., Menez, A., Fromageot, P., Bo-

quet, P. & Changeux, J.-P. (1972) FEBS Lett. 25, 127-133.

Cell Biology: Kozlowski et aL

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1