Identification of dipeptidyl peptidase IV as a protein shared by the plasma membrane of hepatocytes and liver biomatrix

Experimental

Cell Research 158 (1985) 509-518

Identification of Dipeptidyl Peptidase IV as a Protein Shared by the Plasma Membrane of Hepatocytes and Liver Biomatrix EARL F. WALBORG, Jr.,* SHIGEKI TSUCHIDA, DANIEL S. WEEDEN, MICHAEL W. THOMAS, ANITA BARRICK, KERRY D. McENTIRE, JAMES P. ALLISON and DOUGLAS C. HIXSON The University of Texas System Cancer Center, Science Park Research Division, Smithville, TX 78957, USA

The histotypic organization of liver parenchyma involves specific intercellular contacts and interaction of hepatocytes with supporting biomatrix. Evidence from this laboratory identified a peptide (HeplOS, apparent M, 105000) that is shared by the plasma membrane of rat hepatocytes and rat liver biomatrix. This report identifies HeplOS as a peptide component of dipeptidyl peptidase IV (DPPIV; EC 3.4.14.-). A monoclonal antibody (MAb 236.3) was shown to immunoprecipitate DPPIV from non-ionic detergent extracts of surface-labeled “‘1 hepatocytes. The immunoprecipitate contained two ‘251-labeled peptides: HeplO5 and a peptide of apparent M, 150000 (Hepl50). Proteolysis of ‘251-labeled HeplO5 and Hepl50 by Staphylococcus aureus V8 protease yielded essentially identical patterns of ‘251-labeled peptide degradation products, indicating that HeplO5 and Hepl50 are structurally related. Only Hepl50 exhibited DPPIV activity on transblot analysis, an observation that is consistent with the interpretation that it is the monomeric form of the enzyme. Heating (lWC, 5 min) of puritied HeplSO in the presence of sodium dodecylsulfate (SDS) resulted in its conversion to HeplO5 and the disappearance of any demonstrable enzymatic activity. ‘H-labeled diisopropyl fluorophosphate was incorporated into HeplO5, indicating that HeplO5 contains the active site for DPPIV. Purified rat liver biomatrix was shown to possess significant DPPIV activity. Taken together, these data indicate that HeplO5 s a peptide component of DPPIV. 0 1985 Academic PRSS, IIIC.

Recent reports from this laboratory [l, 21 described a peptide, designated HeplOS, that is present in rat liver biomatrix and the plasma membrane of rat hepatocytes: an observation that suggests that this peptide may play a role in maintenance of tissue architecture. This peptide is one component of a family of wheat germ agglutinin (WGA)-binding glycoproteins exhibiting apparent subunit molecular weights of approx. 105000 that are present at the surface of rat hepatocytes [3]. This family of glycoproteins, designated gp105 was absent from two solid rat hepatocellular carcinomas [3, 41. Furthermore, components immunologically related to gp105 exhibited structural alterations in three rat hepatocellular carcinomas propagated in the ascitic form [4]. Reports by Elovson [5] and Kreisel et al. [6], indicating that the serine protease dipeptidyl peptidase IV (EC 3.4.14.-) (DPPIV) is a component of the plasma membrane of rat hepatocytes, and that this enzyme exhibited lectin-binding * To whom offptint requests should be addressed. 33-858336

Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/85 $03.00

510 Walborg et al. affinities and a subunit molecular weight similar to gp105, prompted the investigation of this enzymatic activity in gp105 and its components. This report demonstrates that HeplO5 is a peptide component of DPPIV. MATERIALS

AND METHODS

Cells Hepatocytes were obtained by perfusion of the intact adult rat liver with collagenase, as previously described [7]. Rats of the ACW,, strain were obtained from the Science Park, Veterinary Resources Division of the University of Texas System Cancer Center, Bastrop, Texas.

Assay of DPP IV The enzymatic activity of non-ionic detergent extracts of hepatocytes was assayed using the calorimetric method of Hopsu-Havu & Sarimo [8], modified to use glycyl-L-proline+methoxy-bnaphthylamide (Vega Biochemical Co., Tucson, Ariz.) as substrate. The reaction was performed using 1.1 ml of sample and 0.2 ml of substrate. The released 4-methoxy-/I-napthylamine was diazotized with Fast Garnet GBC (Sigma Chemical Co., St. Louis, MO.) as described [8], except that the concentration of dye was reduced to 0.5 mg/ml. Following incubation for 30-60 mitt at 37”C, 1 ml of the reaction solution was added to 0.4 ml of the dye reagent. A standard curve was constructed using an alcoholic solution of 4-methoxy-/I-naphthylamine (Sigma). One unit of activity was defined as the amount of enzyme required to release 1 umol of product per mitt at 3PC. For calculation of specific activity, protein concentration was determined by the method of Peterson [9], using bovine serum albumin as a standard. For the assay of cell extracts, cells (lO’/ml) were solubihzed in 0.2 M Tris, 0.5 % Nonidet P-40 (NP40) (Particle Data Laboratories, Ltd., Elmhurst, Ill.), pH 7.8, insoluble components removed by centrifugation (30000 g, 10 min), and aliquots of the supematant solution used for assay.

Iodination of Cells Cells were surface-labeled with i2’I (carrier-free, Amersham/Searle Corp., Arlington Heights, Ill.) by the lactoperoxidase procedure of Keski-Oja et al. [lo], as previously described [ll]. Cells (2Xlo6/ml) were lysed in phosphate-buffered saline (PBS) [12], containing 0.5% NP-40 and 0.2% sodium azide and insoluble material removed by centrifugation (30000 g, 10 mitt).

Antisera and Monoclonal Antibodies Several of the polyclonal antisera and monoclonal antibodies used in these studies have been perviously described including: (1) a rabbit antiserum, designated anti-gplO5, raised against wheat germ agglutinin-binding components isolated from a detergent extract of rat hepatocytes [3]; (2) a rabbit antiserum, designated anti-biomatrix, raised against rat liver biomatrix [l, 21; and (3) two monoclonal antibodies (MAb), designated 236.3 and 236.4, reactive with components within the gplO5 set of plasma membrane glycoproteins [13]. Another rabbit antiserum, designated 8325, was raised against a crude preparation of rat (ACI) kidney DPPIV. It was observed that a partially purified preparation of y-glutamyl transpeptidase [14] also exhibited relatively high activity for DPPIV. The referenced isolation procedure for enzyme III [14] was followed through the (NHJ2S04 fractionation, but terminated prior to solubilization with bromelain. The procedure was modified to use 0.5 % Biton X-100 [IS] instead of Lubrol WX and the protein precipitated by 50% saturated (NHJ2S0., was dialysed against 50 mM K phosphate, containing 1 mM EDTA, pH 7.5. This preparation possessed 6x lo-’ units of DPPIV activity per mg protein. Protein concentration was determined as described above. A portion (5 mg in 1 ml) of this crude enzyme preparation, mixed with Bordetefla pertussis antigen [16] (Difco Laboratories, Detroit, Mich.), containing 1.1x 10s bacteria, was injected subcutaneously into an adult male New Zealand White rabbit (David Cunningham, Dripping Springs, Tex.). This initial immunization was followed by subsequent intramuscular injections (Zweek intervals) of 5 mg of the crude enzyme preparation emulsified with incomplete Freund’s adjuvant (Miles Laboratories, Inc., Elkhart, Ind.). The rabbit Exp Cell Res I58 (1985)

Rat liver dipeptidyl peptidase IV

511

was bled 10 days after each immunization and the serum tested for the production of antibodies capable of immunoprecipitating DPPIV activity (see below). The third bleed was utilized. All rabbit antisera were heat inactivated (SK, 15 min) before use.

Analysis of Zmmunoprecipitates Immunoprecipitations were performed using PBS/NP-40 extracts of hepatocytes [3] and formalinfixed Staphylococcus aureus (IgG Sorb, The Enzyme Center, Inc., Boston, Mass.) as the immunoabsorbent, as previously described [3]. Immunoprecipitates were assayed for enzymatic activity as described above. Immunoprecipitates of PBSMP-40 extracts of ‘251-labeled hepatocytes were also submitted to polyacrylamide gel electrophoresis (PAGE) according to the method of Laemmli [ 171. Samples were solubilized in sodium dodecyl sulfate (SDS)-containing sample buffer [12] with or without 2 % 2-mercaptoethanol and with or without heating of the samples at 100°Cfor 5 min. The gels were dried and submitted to autoradiographic analysis as previously described HI]. The following proteins were used as molecular weight standards rabbit muscle myosin (kf, 200000), Escherichia coli B-n-galactosidase (M, 116000), rabbit muscle phosphorylase A (i&f, SSOOO),bovine serum albumin (M, 66 000) and hen ovalbumin (M, 45 000). In one experiment ‘251-labeled immunoprecipitates were resolved on a Laemmli gel (samples not heated in sample buffer) and those portions of the gel containing components of apparent M, 105000 and 150000 sliced from the gel. These separate slices of gel were placed directly onto a second Laemmli get for electrophoretic resolution.

Peptide Mapping of HeplO5 and Hepl50 HeplO5 and Hepl50 were resolved by PAGE in the presence of SDS and the gel dried and submitted to autoradiography. The localized bands of HeplOS and Hepl50 were cut from the gel, inserted into sample wells of a slab gel (12.5 % acrylamide) and digested with Staphylococcus aureus VS protease (Sigma) as described by Cleveland et al. [18].

Analysis of Enzymatic Activity and Zmmunologic Reactivity of Peptides Resolved by Polyacrylamide Gel Electrophoresis Samples of PBYNP-40 solubilized, ‘251-labeled hepatocytes were placed in SDS sample buffer [I21 with or without 2 % 2-mercaptoethanol and with or without heating of the sample at 100°Cfor 5 min. Protein was electrophoretically transferred to nitrocellulose paper (Bio-Rad Laboratories, Richmond, Calif.) according to the method of Towbin et al. [19]. The transfer was accomplished using a Tkansblot, Model 1200(Bio-Rad) operated for 17 h at approx. 425 mA. The efficiency of the electrophoretic transfer was monitored by staining the acrylamide gels with Coomassie Brilliant Blue R-250 (BioRad). Enzymatic and immunologic detection of peptides resolved by SDS-PAGE was performed using the procedure of Muilerman et al. [20], except that 10 mM ‘Bis, 0.9% NaCI, 3% bovine serum albumin (fraction V, Sigma), pH 7.4, was used as the ‘standard’ buffer in the immunologic detection of proteins on nitrocellulose. When assaying the transfers for enzymatic activity, the ‘standard buffer’ contained 1% ovalbumin (Grade V, Sigma) instead of bovine serum albumin. Following the electrophoretic transfer, the nitrocellulose was incubated for 1 h at 37°C in the standard buffer. Transfers intended for detection of direct enzymatic activity were left in standard buffer at 4°C while transfers intended for detection of immunologic reactivity were incubated for 16 h at 4°C with a 1: 100 dilution of antiserum 8325. The transfer incubated with antiserum was washed ten times with standard buffer (each wash 15 mm); then incubated for 2 h at 23°C with a 1 : 500 dilution of the crude DPPIV preparation described above; and finally washed six times (each wash 10 min) with 0.2 M ‘Iris, pH 7.8. The transfers intended for detection of either direct enzymatic activity or immunologic reactivity were overlaid with filter paper (Whatman No. l), saturated with 1 mM glycyl-L-proline-4-methoxy-/Inaphthylamide in 0.2 M B-is, pH 7.8. Following an incubation period of 15 min at 37°C the falter paper was carefully removed and the nitrocelhrlose transfer dipped in 1 M acetate, pH 4.2, containing 0.5 mg Fast Garnet GBC/ml. After rinsing in 0.2 M MS, pH 7.8, the nitrocellulose transfer was airdried. Areas of the transfer having enzymatic activity were indicated by the red azo dye. A permanent record of the activity was obtained by photographing the stained nitrocellulose transfer. Exp Cell Res I58 (1985)

512 Walborg et al. PuriJication of DPPIV from Rat Liver DPPIV was puritied by a procedure that will be described separately (Tsuchida & Walborg, unpublished results). This procedure utilizes solubilization of a crude membrane fraction (the pellet obtained by centrifuging liver homogenate at 105000 g, 45 min) with non-ionic detergent and subsequent purification of the papain-solubilized form of the enzyme by salt fractionation. Further purification was achieved by gel filtration and ion-exchange chromatography. This procedure yielded a 1300-fold increase in the specific activity of the final product over that of the liver homogenate. The specific activity of the final product, determined as described above, was 28 units/mg protein.

Labeling of Proteins by [3H]diisopropyl Fluorophosphate ([3H]DFP) Purified rat liver DPPIV (see directly above) was labeled with [l ,3-3H]diisopropyl fluorophosphate (sp. act., 5.2 Ci/mmol; 0.25 mCi/0.25 ml; New England Nuclear, Boston, Miss.), using the method of Kenny et al. [21]. DPPIV (150 pl containing three units of DPPIV activity/ml of 10 mM B-is, pH 7.9) was incubated with the above r3H]DFP for 2 h at 37’C. 3H-labeled peptides resolved by PAGE were visualized by fluorography as previously described [12].

DPPIV Activity of Rat Liver Biomatrix Rat liver biomatrix was prepared by the method of Hixson et al. [2]. Subsequent to the described perfusion, the liver was perfused with 0.9% NaCl and frozen in situ by placing crushed dry ice over the exposed liver. A block of frozen liver was manually cut into small portions (approx. 3 mm3) using a razor blade and each of these portions of liver placed on preweighed nylon mesh (1 cm’ screens cut from ASTM-3-70-210 Nitex monofilament nylon screens obtained from Pow Equipment, Inc., Houston, Tex.). After thawing of the biomatrix preparation, the screens were weighed on an Electrobalance (Model 28, Cahn Instruments, Inc., Carritos, Calif.) to an accuracy of 0.01 mg. A number of screens were combined to yield approx. 3 mg biomatrix for enzyme assay. Enzymatic activity was assayed as described above, except that the incubation mixture was centrifuged (12000 g, 3 min) immediately following incubation to remove insoluble material.

RESULTS Enzymatic Activity of Rat Liver Biomatrix Biomatrix was found to exhibit m units/mg dry weight. Immunoprecipitation

significant

DPPIV activity,

namely 13

of Enzyme Activity from Detergent Extracts of Hepatocytes

The effect of NP-40 concentration on solubilization of enzymatic activity was investigated using concentrations of detergent ranging from 0.1 to 2.0%. A concentration of 0.5 % NP-40 yielded maximal solubilization with concentrations of 0.1, 1.O, 1.5, and 2.0 yielding 53, 83, 80, and 69 % of the activity solubilized in 0.5% NP-40. Cellular components solubilized in 0.5% NP-40 exhibited DPPIV activity (2l.Ok2.3 SD munits/mg protein). Fig. 1 demonstrates the effectiveness of the various antisera and monoclonal antibodies in immunoprecipitating enzymatic activity from detergent extracts of rat hepatocytes. In the case of MAb 236.3 two different preparations of the antibody yielded maximal precipitation at levels Y2-s of that obtained with the polyclonal antiserum 8325. The less than quantitative immunoprecipitation of DPPIV activity by MAb 236.3, as well as by anti-biomatrix and antigpl05 EXP Cell Res 1% (1985)

Rat liver dipeptidyl peptidase IV

513

Fig. I. Immunoprecipitation of DPPIV activity from nonionic detergent extracts of rat hepatocytes. Immunoprecipitation was performed using the following antisera: A-A, 8325; A-A, MAb 236.3; O-0, antigpl05 antiserum; O--O, anti-biomatrix antiserum; Cl-Cl, MAb 236.4. Immunoprecipitations were performed as described in Materials and Methods. The enzymatic activity immunoprecipitated by the antisera is expressed as the percentage of the total activity present in the non-ionic detergent extract of ld hepatocytes. Non-immune rabbit serum (10 pl) immunoprecipitated no enzymatic activity. Antiserum

added

(pl)

antisera, may be explained by partial denaturation of reactive epitopes of DPPIV caused by solubilization of the antigen in non-ionic detergent. It should be noted that two of the reactive antibodies were raised against whole cells (MAb 236.3) or insoluble purified biomatrix (anti-biomatrix antiserum). Components Immunoprecipitated from NP-40 Extract of Rat Hepatocytes by MAb 236.3 As shown in fig. 2, E, I, immunoprecipitates obtained using MAb 236.3 and NP40 extracts of surface-labeled “‘1 hepatocytes contained two predominant 1251labeled peptides exhibiting apparent M, 105000 and 150000 components designated HeplOS and HeplSO, respectively. Both HeplO5 and Hepl50 were present when samples submitted to PAGE were not heated (fig. 2, E, I), while samples heated (lOOT, 5 min) in the presence of 2-mercaptoethanol contained only HeplO5 (fig. 2, c). When immunoprecipitates were heated (lOOT, 5 min) in the absence of 2-mercaptoethanol, only HeplO5 and a component having an apparent M, 110000 were present (fig. 2, L). When Hepl50, purified by PAGE, was subsequently heated and submitted to electrophoretic resolution on a Laemmli gel, only HeplO5 was observed (fig. 3). Peptide Mapping of HeplO5 and Hepl50 As shown in fig. 4, digestion of 1251-labeledHeplO5 and Hepl50 with Staphylococcus aureus V8 protease yielded similar peptide patterns, i.e., two predominant 1251-labeledpeptides and a peptide of lower mobility that decreased in concentration in incubations having the higher concentration of protease. Exp Cell Res 158 (1985)

514 Walborg et al. &M,

- 150 -110

-105

ABCDEFGHIJKL ----++++++++

MAb 236.3

+*++++++----

2-mercaptoeihanol

-++--++--++ +

-+-+

-+-+-+-+-+-+

Heat. -+-+-+-

100’

5 min.

Autoradiography Enzyme

activity

Fig. 2. Effect of heating and 2-mercaptoethanol on the peptide composition and enzymatic activity of

immunoprecipitates (MAb 236.3) from non-ionic detergent extracts of surface-labeled “‘1 hepatocytes. Individual lanes contain A-D, NP-40 extracts of hepatocytes or EL, immunoprecipitates (MAb 236.3) from such extracts that have been submitted to PAGE (6 % gels) in the presence of SDS and subsequently transferred to nitrocellulose. The resolved peptides were visualized by autoradiography (A, 2 h; C, E, G, I, K, 24 h) or stained for enzymatic activity (B, D, F, H, J, L). Some samples were solubilized in SDS-sample buffer containing 2-mercaptoethanol without heating prior to PAGE resolution; while other samples were heated (KWC, 5 min) prior to PAGE. Other samples were solubilized in SDS sample buffer without 2-mercaptoethanol without heating or with heating at IOO’C for 5 min. The lanes contain A, E, D, 5x 104; C, 4x 10’; or E, L, lo5 cell equivalents of hepatocyte extract. (Left) Positions of the molecular weight standards; (right) estimated molecular weights of the observed peptide bands. Since gels used in this experiment exhibited a non-Smear relationship between mobility and log M, >116000, the molecular weight of HeplSO was assigned on the basis of its mobility on gels on which the molecular weight standards showed a linear relationship between mobility and log M,; e.g., see fig. 5.

Detection of Enzymatic and Immunologic Reactivity of HeplO5 and Hepl50 using Blotting Techniques Enzymatic activity of HeplOS and HeplSO resolved by SDS-PAGE in the presence or absence of 2-mercaptoethanol could be visualized on nitrocellulose transfers only if the samples were not heated prior to electrophoretic separation. Heating (lOO”C, 5 min) of NP-40 extracts of hepatocytes or immunoprecipitates of these extracts abolished the subsequent detection of enzymatic activity (compare B and D, F and G, or J and L of fig. 2). Enzymatic activity could be detected only in one discrete peptide component, HeplSO (fig. 2, B, F,J). Detection of immunologic reactivity in HeplO5 using the method of Muilerman et al. [20] was not successful using either polyclonal antiserum 8325 or MAb Exp Cell Res 158 (1985)

Rat liver dipeptidyl peptidase IV lo-+M,

‘. ,*.‘, ,

515

1 o-3xMr

(“;2,‘.:

200-

116-105 95-

66-

45-

3 ABCD

4

ABCD

'-

5

ABC

Fig. 3. Conversion of Hep 150 to Hep 105 upon heating in the presence of SDS. A, B, Hep 150; or C, D, Hep 105 were purified from immunoprecipitates (MAb 236.3) of NP-40 extracts of lZ51-labeled

hepatocytes as described in Materials and Methods. Samples of these peptides were applied to Laemmli gels (6% a&amide) A, C, without heating; or B, D, after heating at 100°Cfor 5 min. Each lane contained material representing approx. 1.5~ 10’ cells. The lanes shown were reproduced from autoradiograms exposed for A, B, 7; or C, D, 4 days. Fig. 4. V8 protease digestion of Hep 105 and Hep 150. A, B, Hep 150 and C, D, Hep 105 were purified from immunoprecipitates (MAb 236.3) of NP-40 extracts and submitted to digestion with A, C, 25 pg or B, D, 50 ug of V8 protease as described in Materials and Methods. The protease digests were submitted to resolution by PAGE (12.5 % acrylamide) in the presence of SDS and the resolved I?labeled peptides visualized by autoradiography (exposure, 6 days). The arrows indicate a peptide of lower mobility present in digests using the lower concentration of protease. Fig. 5. Labeling of peptide components of DPPIV by [3H]diisopropyl fluorophosphate (DFP). Purified papain-solubilized rat liver DPPIV (0.45 units) was labeled with r3H]DFP. Two-thirds of the labeled material was submitted directly to A,B, resolution by SDS-PAGE. One sample, B, was heated (IOO’C, 5 min) prior to electrophoresis, while the other sample, A, was not heated. One-third of the labeled material was immunoprecipitated with MAb 236.3 and the immunoprecipitate submitted C, to SDS-PAGE. The electrophoretic separation was performed using 7.5 % acrylamide gels. Sample buffer contained 2% 2-mercaptoethanol. The 3H-labeled components were visualized by fluorography (3-week exposure). (Left) Positions of the molecular weight standards; (right) estimated molecular weights of the observed peptides.

236.3. Since Hepl50 possessed direct enzymatic activity, no information could be gained regarding its immunologic reactivity using this method. Use of Diisopropyl Fluorophosphate to Identify Peptides Containing the Active Site of DPPIV As shown in fig. 5, A, [‘HI-DFP labeled predominantly one component of purified, papain-solubilized DPPIV, i.e., a component of apparent M, 105000, Exp Cell Res IS8 (1985)

516 Walborg et al. while 3H labeling of a minor component having an apparent M, 150000 was also observed. If the 3H-labeled components were heated in the presence of SDS prior to resolution by PAGE, only the M, 105000 component was observed (fig. 5, B). The immunoprecipitate of 3H-labeled, purified DPPIV by MAb 236.3 contained the component of apparent M, 105000. DISCUSSION Immunoprecipitation of DPPIV activity from non-ionic detergent extracts of rat hepatocytes by anti-biomatrix antiserum and MAb 236.3 indicated that Hep 105, a peptide shared by the hepatocyte plasma membrane and liver biomatrix, was associated with this enzymatic activity (fig. 1). Immunoprecipitates of extracts of surface-labeled hepatocytes by MAb 236.3 contained two labeled peptides: HeplO5 and Hepl50, having apparent M, 105000 and 150000, respectively (fig. 2). Hepl50 was present only when immunoprecipitates were not heated in the presence of SDS (fig. 2), suggesting conversion of Hepl50 to HeplO5. Such a conversion was, in fact, effected by heating of purified Hepl50 in the presence of SDS (fig. 3). Consistent with a structural relationship between HeplO5 and Hepl50 was the demonstration that proteolytic digestion of these peptides with V8 protease yielded substantially identical products (fig. 4). Omission of 2mercaptoethanol during the heating of immunoprecipitates was accompanied by the appearance of an additional component, apparent M, 110000 (fig. 2), indicating that conversion of Hepl50 to HeplO5 may be a stepwise process and that disulfide linkages may be involved, at least in part, in the conversion. Since conversion of Hepl50 to HeplO5 was not accompanied by the consistent appearance of identifiable low molecular weight peptides (figs 2, 3), conversion could occur by degradation or denaturation; however, the magnitude of the change in apparent molecular weight during the conversion suggests that degradation is a more plausible explanation. Detectable enzymatic activity was associated only with Hepl50 (fig. 2), indicating that conversion to HeplO5 results in loss of enzymatic activity. The inability to demonstrate an immunologic or enzymatic relationship between HeplO5 and Hepl50 (fig. 2) raised the possibility that DPPIV activity was associated with a structurally unrelated peptide, i.e., a peptide dissociated from Hepl50 during its degradation/denaturation to HeplO5. This possibility is excluded by the fact that DFP, a reagent that reacts with the active site of serine proteases [22], labeled a peptide presumably homologous to HeplO5 (fig. 5); thus indicating that HeplO5 is a peptide component of the enzymatically active Hepl50 and that HeplO5 bears the active site. The observation that the detergentand papain-solubilized forms of DPPIV exhibit similar apparent subunit molecular weights on SDS-PAGE is consistent with a previous report by Elovson [51. The discrepancy between the apparent molecular weight of HeplO5 and that reported by Kenny et al. [21] and Elovson [5] for a DPPIV-related peptide Exp Cell Res 158 (1985)

Rat liver dipeptidyl

peptidase IV

517

solubilized by heating in the presence of SDS probably results from the fact that the latter studies assumed a M, 130000 for Escherichia coli B-D-galactosidase used as a standard in the PAGE analysis. The studies reported herein assumed a M, 116000 [23] for this protein. Data reported by Elovson (fig. 4 in ref. [5]) clearly shows a DPPIV-related peptide having an apparent molecular weight less that of #&D-galactosidase; an observation consistent with that reported herein. Kreisel et al. [6] have shown that DPPIV activity is immunoprecipited by a polyclonal antiserum raised against a plasma membrane glycoprotein fraction (TC2W2; apparent M, 110000) isolated from rat liver. These authors also demonstrated that DPPIV activity was detected in a component resolved in SDScontaining buffer, but only if the sample was not heated prior to electrophoretic resolution. This enzymatically active component was assigned an apparent M, 220000 on the basis of its electrophoretic behavior in polyacrylamide gels in the presence of SDS. They interpreted their observations to indicate that the enzymatically active form resolved on polyacrylamide gels is the dimeric form of the enzyme (M, 230000 to 300000 [21, 24, 251); while the results reported herein are interpreted to indicate that it is the monomeric form of the enzyme, containing a single active site [21] or a renatured dimeric enzyme produced during transfer to nitrocellulose that possesses the observed activity on nitrocellulose transfers of SDSpolyacrylamide resolving gels. An explanation of the discrepancy in molecular weights reported here and by Kreisel et al. [6] must await further experimentation. Previous studies [ 1, 21, using indirect immunofluorescence to localize HeplOS, can now be interpreted to indicate that DPPIV is present on the hepatocyte plasma membrane and on the extracellular matrices associated with the bile canaliculi and the space of Disse; and similar studies on purified rat liver biomatrix [I, 21 can now be interpreted to indicate that DPPIV is present on some, but not all, the fibrous elements of rat liver,biomatrix. The microanatomical localization of DPPIV suggests that this enzyme may be involved in the interaction of hepatocytes with adjacent biomatrix or in the degradation of the collagenous elements of the extracellular matrix [26]; while the altered expression of HeplO5 in rat hepatocellular carcinomas [3] suggests that DPPIV may be involved in the loss of multicellular architecture characteristic of these carcinomas. This research was supported by grants CA 27377, CA 26891, and CA 31103 from the National Cancer Institute. Support was also obtained from the Paul and Mary Haas Foundation and the George and Mary Josephine Hamman Foundation. The housing and care of experimental animals was supported, in part, by National Cancer Institute Core Grant CA 16672.

REFERENCES 1. Hixson, D C, Ponce, M de L, McEntire, K D, Chesner, J E, Lund, J N, Allison, J P & Walborg, E F, Jr, Structural carbohydrates in the liver (eds H Popper, W Reutter, E Kiittgen & F Gudat) p. 577, MTP Press, Ltd, Lancaster (1983). Exp Cell Res IS.3 (1985)

518 Walborg et al. 2. Hixson, D C, Ponce, M de L, Allison, J P & Walborg, E F, Jr, Exp cell res 152 (1984) 402. 3. Hixson, D C, Allison, J P, Chesner, J E, Leger, M J, Ridge, L L & Walborg, E F, Jr, Cancer res 43 (1982) 3874. 4. Hixson, D C, Allison, J P, Chesner, J E, Leger, M J, Ridge, L L, Thomas, M W & Walborg, E F, Jr, Membranes in tumour growth (eds T Galeotti, A Cittadini, G Neri & S Papa) p. 19. Elsevier Biomedical Press, Amsterdam (1982). 5. Elovson, J, J biol them 255 (1980) 5807. 6. Kreisel, W, Heussner, R, Volk, B, Buchsel, R, Reutter, W & Gerok, W, FEBS lett 147(1982) 85. 7. Starling, J J, Capetillo, S C, Neri, G & Walborg, E F, Jr, Exp cell res 104 (1977) 177. 8. Hopsu-Havu, V K & Sarimo, S R, Hoppe-Seyler’s Z physiol them 348 (1%7) 1540. 9. Peterson, G L, Anal biochem 83 (1977) 346. 10. Keski-Oja, J, Mosher, D F & Vaheri, A, Biochem biophys res commun 74 (1977) 699. 11. Glenney, J R, Jr & Walborg, E F, Jr, J supramol struct 11 (1979) 493. 12. Glenney, J R, Jr, Allison, J P, Hixson, D C & Walborg, E F, Jr, J biol them 254 (1979) 9247. 13. Allison, J P & Hixson, D C, Proc Am Assoc. cancer res 24 (1983) 217. 14. Tate, S S & Meister, A, J biol them 250 (1975) 4619. 15. Tsuchida, S, Hoshino, K, Sato, T, Ito, N & Sato, K, Cancer res 39 (1979) 4200. 16. Gregoriades, G & Manesis, E K, Liposomes and immunobiology (ed B H Tom & H R Six) p. 271. Elsevier/North-Holland, New York (1980). 17. Laemmli, U K, Nature 227 (1970) 680. 18. Cleveland, D W, Fischer, S G, Kirschner, M W & Laemmli, U K, J biol them 252 (1977) 1102. 19. Towbin, H, Staehelin, T & Gordon, J, Proc natl acad sci US 76 (1979) 4350. 20. Muilerman, H G, ter Hart, H G J & Van Dijk, W, Anal biochem 120 (1982) 46. 21. Kenny, A J, Booth, A G, George, S G, Ingram, J, Kershaw, D, Wood, E J & Young, A R, Biochem j 155 (1976) 169. 22. Dixson, M & Webb, E C, Enzymes, 2nd edn, p. 473. Academic Press, New York (1964). 23. Fowler, A V & Zabin, I, Proc natl acad sci US 74 (1977) 1507. 24. Barth, A, Schulz, H & Neubert, K, Acta biol med germ 32 (1974) 157. 25. Ichinose, M, Maeda, R, Fukuda, T, Watanabe, B, Ishimaru, T, Izumi, M, Miyake, S & Takamori, M, Biochim biophys acta 719 (1982) 527. 26. Kojima, K, Hama, T, Kato, T & Nagatsu, T, J chromatog 189 (1980) 233. Received January 3, 1985

fip CellRes

158 (1985)

Printed

in Sweden