Eur. J. Biochern. 211, 357-365 (1993) 0 FEBS 1993

Evidence of serine-protease activity closely associated with Drosophila alcohol dehydrogenase Joan FIBLA, Silvia ATRIAN and Roser GONZALEZ-DUARTE Departament de Genktica, Universitat de Barcelona, Spain

(Received July 17, 1992) - EJB 92 1019

With the use of monoclonal antibodies against alcohol dehydrogenase (ADH) we detected ADH proteolysis in different Drosophila mehogaster tissues during development [Visa, N., Fibla, J., Santa- Cruz, M. C . & Gonzilez-Duarte, R. (1992) J . Histochem. Cytochern. 40,39-491. We now report the analysis of t h s proteolytic activity in crude homogenates and in purified ADH preparations of several Drosophilu species. Our results indicate that in non-denaturing IEF gels the proteolytic activity co- migrates with native ADH electromorphs of all the species analyzed. In addition, we show that it co- purifies with ADH and is responsible for the instability of apparently homogeneous ADH prep- arations in the presence of SDS. When purified ADH preparations were analyzed, the endogenous proteolytic activity yielded the same banding pattern as that obtained with crude homogenates. Even after rechromatography on Sephacryl S-200, the usual last step in our standard purification protocol, the proteolytic activity remained associated with the ADH fractions. Among the various agents which could explain the ADH-linked proteolytic effect, a pre-existing nicked state of the enzyme or chemical proteolysis have been ruled out. The kinetics observed on pure ADH preparations, the effect of specific protease inhibitors and substrate specificity have led us to ascribe this activity to the subtilase serine-protease family. Given that proteolysis is evident even in rechromatographed Sephacryl S-200 fractions, if incubated in SDS for enough time, we propose two alternative hypotheses to explain this phenomenon. First, the proteolytic activity may come from a protease which is inseparable from the ADH active forms and second, the ADH itself may behave as a subtilase when it adopts a particular conformation. Moreover, the previously reported differential banding pattern during development suggests a role for this activity in vivo, in which fatty acids could produce the inducer effect attributed to SDS in vitro.

Drosophila alcohol dehydrogenase (ADH, alcoho1:NAD' oxidoreductase) is the best characterized member of the short- chain dehydrogenase family of structurally related enzymes of prokaryotic, insect and human origin, all involved in different catalytic reactions (Persson et al., 1991). Drosophilu ADH is a non-metalloenzyme dehydrogenase which acts as a dimer composed of two identical subunits, each of 255 amino acids, able to oxidize primary and secondary alcohols using NAD' as cofactor. It is a soluble enzyme present at moderately high concentrations (i.e. up to 1 % of total soluble protein) in most

Correspondence to R. Gonzilez-Duarte, Departarnent de Genktica, Universitat de Barcelona, Av. Diagonal 645, E-08028 Barcelona, Spain

Fax: + 34-3-41 10969. Abbreviations. ADH, alcohol dehydrogenase; ADH-G1 , ADH-

G2 and ADH-G3, ADH contained in the GI, G2 and G3 groups of fractions, respectively; NaCl/Pi, phosphate-buffered saline; PI, isoelectric point; PhMeS02F, phenylmethylsulphonyl fluoride; TosLysCH2C1, Nu-tosyl-l-lysyl chloromethane; TosPheCHzCl, Na- tosyl-l-phenylalanyl chlorornethane.

Enzymes. Alcohol dehydrogenase (EC 1.1.1.1); chymotrypsin (EC 3.4.21.1); hexokinase (EC 2.7.1.1); lysozyme (EC 3.2.1.17); peroxi- dase (EC 3.11.1.7); phosphofructokinase (EC 2.7.1.11); subtilisin (EC 3.4.21.14); trypsin (EC 3.4.21.4); 20P-hydroxysteroid dehydrogenase (EC 1.1 .I .53).

Drosophila species. Its biological role has been extensively reported and there is ample agreement about its essential contribution to alcohol metabolism and detoxification (Chambers, 1988,1991).

Purification protocols have been successfully developed for the enzyme from D. melunogaster and other Drosophilu species (Thatcher, 1977; Juan and Gonzalez-Duarte, 1980; Ribas de Pouplana et al., 1991). Pure protein solutions used for biochemical and structural characterization of this enzyme are reasonably stable. However, when apparently homo- geneous preparations are incubated in the presence of SDS- containing buffers, proteolytic peptides progressively appear until the monomeric form of ADH eventually becomes undetectable on SDSjPAGE gels. Moreover, an ADH degra- dation pattern has also been detected following immunochemical analysis of ADH in several Drosophila or- gans and at different developmental stages (Visa et al., 1992).

It is well known that many proteins which are resistant to attack by proteolytic enzymes can be rendered susceptible under denaturing conditions. In addition these conditions can promote the processing of an inactive precursor. Highly purified yeast proteins, such as hexokinase and phospho- fructokinase, can undergo rapid proteolytic degradation in the presence of SDS; it has been claimed that this is due to a

358

closely linked protease which forms stable complexes with a subpopulation of enzyme molecules (Rustum et al., 1971; Diezel et al., 1972). However, this proteolytic phenomenon has not been reported in Drosophila enzymes, although autoproteolysis of some larval and pupal proteins has been described (Mitchell et al., 1985; Mitchell and Petersen, 1987).

The aim of the present study is to characterize the proteo- lytic activity associated with Drosophila native ADH in crude homogenates and in purified ADH preparations.

MATERIALS AND METHODS

Fly strains and antisera

The following Drosophila melanogaster strains were used : AdhUF, AdhF and Adh’ homozygous for the ‘ultrafast’, ‘fast’ and ‘slow’ Adh alelles, and Adhn4, a null strain cross-reacting- material negative (CRM -). Wild stocks from D . mulleri and D. bocquetti were supplied by Dr. A. Fontdevila (Universitat Autbnoma de Barcelona, Spain) and Dr. F. Lemeunier (Centre National de la Recherche Scientifique, France) respec- tively and D. lebanonensis was obtained from Umei Drosophila Stock Center (Sweden), and kept under standard laboratory conditions.

Monoclonal antibodies MMBB8 and LLBE8 against Drosophila ADH were obtained and characterized as pre- viously described (Fibla et al., 1989). The epitope recognized by MMBB8 has been mapped using synthetic peptides and been found to correspond to the peptide comprised between Leu189 and Val199 of ADH (unpublished results). Thus, any protein fragment comprising this peptide will be recognized by mAb MMBB8.

ADH and peptide purification

ADH was purified by conventional methods already estab- lished in our laboratory (Juan and Gonzalez-Duarte, 1980; Ribas de Pouplana et al., 1991). An additional purification step was introduced : a second exclusion chromatography with Sephacryl S-200 or with Superose 6 or Superose 12 and 6 connected in series using an FPLC system. Fractions were concentrated using low-binding Ultrafree-CL filters (Milli- pore) with a 10-kDa exclusion pore. Proteolytic fragments of ADH were resolved by reverse-phase HPLC on a column (15Ox4mm) of Delta-Pack (C4, 30nm, 5 pm), using acetonitrile step gradients (30-60% and 60-100%) in 0.1% trifluoroacetic acid.

Protein determination and ADH activity assays

Drosophila crude homogenates were prepared as in Fibla et al. (1989), total soluble protein was determined following Bradford (1976), ADH protein was measured by a direct ELISA assay, using a dilution 1 : 50 of mAb MMBB8 (Fibla et al., 1989) and alcohol dehydrogenase activity was recorded following Juan and Gonzilez-Duarte (1980).

Protein electrophoresis, blotting and immunostaining

Native isoelectrofocusing gel electrophoresis was carried out on Mini-IEF cell equipment from BioRad, following manufacturer’s instructions. A pH gradient of 3.5 - 10 was established in 7% polyacrylamide gels polymerized upon a gel support film (BioRad). After electrophoresis, the gels were

incubated for ADH activity (Winberg et al., 1986) or pro- cessed for two-dimensional analysis.

Polyacrylamide gel electrophoresis in the presence of SDS (SDS/PAGE) was carried out according to Laemmli (1970), in 15% acrylamide gels. Proteins were visualized by 0.1% Coomassie brillant blue or conventional silver staining (Wray et al., 1981). Two-dimensional analysis was performed accord- ing to O’Farrell(l975) with the following modifications: strips of IEF gels were incubated in sample buffer (60 mM Tris/HCl pH 6.8, containing 10% glycerol, 5% 2-mercaptoethanol and 2% SDS) for different times, then heated at 100°C for 2 min and loaded onto an SDS/PAGE gel.

Protein blotting was performed according to Dunn (1986). Briefly, after electrophoresis, SDS/PAGE gels were soaked in 50 mM Tris/HCl pH 7.4, 20% glycerol solution for 30 min and proteins were electrotransferred in 10 mM NaHC03, 3 mM Na2C0, pH 9.9,20% methanol solution to nitrocellu- lose filters for 1 h at 100 V constant current, using BioRad Mini-Transblot cell equipment. Filters were then saturated with 10% Molico dried milk (Nestle) in phosphate-buffered saline (NaCl/Pi, 130 mM NaCl, 3 mM KCl, 1.5 mM KH2P04, 8 mM Na2HP04, pH 7.0), for 2 h at 37°C and incubated in anti-ADH antibody solution at room tempera- ture for 1 h. The washing steps were performed using a solu- tion of 0.5% Molico dried milk and 0.05% Tween 20 in NaC1/ Pi (solution A). After a l-h incubation with rabbit anti-(mouse IgG) coupled to peroxidase (Boehringer; 1 : 3000 dilution in solution A), immunoblots were developed in a solution con- sisting of 50 mg diaminobenzidine/HCl, 2 ml 1 YO CoClz and 100 ~ 1 3 % Hz02 in 100 ml NaCl/Pi. The reaction was stopped by several washes in distilled water.

Quantification of protein bands in SDSlPAGE gels and immunoblots was performed by scanning densitometry in an Ultroscan XL (Pharmacia-LKB) and data were processed by the GelScan program (Pharmacia-LKB).

Proteolytic assays

The standard proteolytic assay was carried out by in- cubating samples containing approximately 5 pg ADH in an SDS solution (sample buffer or 2% SDS in distilled water) for different times. The reaction was stopped at 100°C in a water bath. Samples were then loaded onto an SDS/PAGE gel, or injected into an analytical HPLC column. Minor modifi- cations to this procedure are indicated in the text when re- quired.

ADH susceptibility to the serine-proteases trypsin (Worthington), chymotrypsin (Serva) and subtilisin Calsberg (Boehringer), was also tested. Several commercial proteins such as bovine serum albumin, yeast alcohol dehydrogenase, chymotrypsinogen A, lysozyme, myoglobin and P-lactoglo- bulin (all from Sigma) were assayed as substrates for proteo- lytic activity present in ADH fractions.

The following protease inhibitors were used : benzamidine and pepstatin A were supplied by Sigma, EDTA was from Merck and Nix-tosyl-l-phenylalanyl chloromethane (TosPhe- CH2C1), Ncl-tosyl-l-lysyl-chloromethane (TosLysCH2Cl) and phenylmethylsulphonyl fluoride (PhMeS02F) were from Boehringer.

Sequence comparisons

The SWISSPROT data base was used to access the protein sequences compared. Painvise alignements were obtained using the GAP program. Both data bases and programs were

3 59

A

66 - 45 - !3!= 18 - 17 - 14 - 3 -

B

IEF 4 b

3.5 5.1 6.6 8.5 9.0

1 2 3 4 5 6 7 8 9

66 - 45 - % = 18 - 17 - 14 -

3 -

Fig. 1. Instability of D. rnelunoguster ADH-F enzyme monitored by two- dimensional analysis. (A) ADH cross-reacting bands after a two- dimensional blot of crude homogenate of D. melanogaster AdhF, immunodetected with mAb MMBBS. The sample was fractionated in an IEF strip, which was subsequently incubated in sample buffer and laid upon the polyacrylamide gel to run the second dimension of the electrophoresis. At the top of the figure there is a scheme of ADH-F electromorphs (PI 6.4 and 6.9) as would be detected after incubation for alcohol dehydrogenase activity. (B) SDSjPAGE gel blotted and immunodetected with mAb MMBBS. Slices of the IEF gel containing the most active electromorph of ADH-F (PI 6.9) were incubated with sample buffer as follows: lane 1, heated at 100°C in a boiling water bath for 2 min and then kept at room temperature for 30 min; lane 2, kept at room temperature for 30 min and then heated at 100°C for 2 min; lane 3, heated at 100°C for 5 min and loaded into the gel. In the following lanes, slices were kept at room temperature for 30 min in the presence o f lane 4,2 mM EDTA; lane $10 pM benzamidine; lane 6, 1 1M pepstatin A; lane 7, 1 mM PhMeS02F; lane 8, 10 pM TosPheCH,Cl; lane 9,lO pM TosLysCH2C1. The electrophoretic mo- bility of the following protein standards is represented on the left: bovine serum albumin (66 kDa), yeast alcohal dehydrogenase (45 kDa), carbonic anhydrase (29 kDa), chymotrypsinogen A (25 kDa), lysozyme (18 kDa), myoglobin (17 kDa), fl-lactoglobulin (14 kDa) and insulin (3 kDa).

supported by the GCG (Wisconsin) Software Package (v7.1), (Devereux et al., 1984).

RESULTS Association of a proteolytic activity with native ADH

Our previous results, using specific monoclonal antibodies against Drosophila ADH (Visa et al., 1992) demonstrated discrete proteolytic ADH degradation in a wide range of larval and adult tissues. We have taken advantage of the availability

66,

45 - 29- 25-

18 - 17 - 14 -

3 -

IEF t b

3.5 5.1 6.6 8.5 9.0

D. bocquetti

D. lebanonensis

D. mulleri

Fig. 2. Interspecific analysis of the proteolytic activity associated with active ADH electromorphs. ADH cross-reacting bands on blots immunodetected with mAb MMBBS. The active electromorphs of D. bocquetti (PI 7.1), D . Iebanonensis (PI 7.5) and D . mulleri (PI 6.8, 7 and 7.4) were sliced and incubated with sample buffer for 30min before loading onto the SDS/PAGE gel. The position of each electromorph for each species detected after incubation for alcohol dehydrogenase activity is represented at the top of the figure. The electrophoretic mobility of marker proteins is indicated as in Fig. 1.

of Drosophila mAbs, and of the enzyme purification methods already established in our laboratory to analyze this proteo- lytic process in depth. Crude homogenates of D. melanogaster AdhF were fractionated in non-denaturing IEF gels, which were subsequently processed for two-dimensional analysis (SDS/PAGE) and immunoblotted with the mAb MMBBS (Fig. IA). Isoelectrofocusing of ADH-F gives three electromorphs, but only two @I 6.4 and 6.9) are easily ob- served. As could be expected, two 27-kDa spots, correspond- ing to the monomeric form of each electromorph, were immunodetected in the two-dimensional blot. Besides, peptide fragments showing a similar pattern to that already reported (Visa et al., 1992) appeared below each 27-kDa spot. Frag- mentation could have originated in vivo or during the homogenization step. In both cases, the two-dimensional blot would resolve into a pattern of scattered spots with different PI. Instead, the discrete number of spots obtained, all in a column below each electromorph, strongly suggested that the degradation process had taken place after the enzyme had been focused on the gel. When the IEF strip was just soaked in boiling SDS-containing buffer, prior to the second dimension PAGE, no degradation bands were observed. This indicated that the cleavage had taken place while the IEF strip had been kept under SDS. Further insight into the agents responsible for ADH proteolysis was obtained with a slight modification of standard two-dimensional analysis as follows. Crude homogenate of D. melanogaster AdhF was fractionated in a preparative non-denaturing IEF gel. The band corresponding to the most active electromorph @I 6.9) was cut out from the gel and sliced in 0.2 x 0.5-cm pieces. Each piece was incubated in sample buffer under different conditions and loaded into

360

G1 G2 G3

Fig. 3. Stability of purified ADH fractions monitored by SDS/PAGE. (A) Elution profile of the Sephacryl S-200, last step of the Drosophilu ADH purification protocol. Total protein content was measured by absorbance at 280 nm (-), ADH protein was evaluated by a direct ELISA assay (......) and ADH activity was measured spectro- photometrically at 340 nm (-----). ADH-GI, ADH-G2 and ADH-G3 fractions were pooled as described. (B) Proteolysis was assayed in aliquots of ADH-G1, ADH-G2 and ADH-G3 containing approxi- matelly 5 pg ADH. ADH-G1 had previously been concentrated 10- fold. Samples were mixed in SDS buffer and immediately heated at 100°C in a boiling water bath (lane 1) or incubated for 15 min (lane 2) or overnight (lane 3) in SDS buffer at room temperature. (C) Aliquots of the ADH-G3 were gel-filtered through a second Sephacryl S-200, or a Superose 6 or Superose 12 and Superose 6 in series columns. Fractions containing ADH were then pooled, concentrated 10-fold, and aliquots containing approximately 5 pg ADH protein were tested for proteolytic degradation as in (8). Detection of peptide fragments was by Coomassie blue staining of the SDSjPAGE gels.

the well of an SDS/PAGE gel. After electrophoresis, proteins were blotted and immunodetected with mAb MMBB8 (Fig. 1 B). As shown in lane 2, the proteolytic pattern observed under this new experimental approach reproduced that of Fig. 1A and showed that the potential proteolytic activity migrated with the ADH electromorphs on the IEF gel. In addition, the absence of degradation fragments when the slice was soaked in sample buffer and immediately loaded (lane l), allowed us to rule out a nicked state for the native ADH.

Both chemical and enzymatic mechanisms could be in- volved in the ADH cleavage. It has been reported that in the

presence of SDS, chemical protein cleavage can take place at 100°C (Kowit and Maloney, 1982; Rittenhouse and Marcus, 1984). However, in our case boiling the sample for 2 or 10 min in the presence of SDS did not generate any detectable frag- ments (Fig. lB , lanes 1 and 3) . Further, the well reported proteolytic effect of trace metals in the electrophoretic re- agents, was ruled out using EDTA (lane 4). Then, enzymatic proteolysis remained as the most feasible agent for ADH fragmentation. Indeed, the absence of degradation at high temperatures (lane 3) and the preservation of the monomeric form when specific protease inhibitors were added to the sample (lanes 4 - 9), strongly supported this hypothesis. Of all the protease inhibitors used, only PhMeS02F, a well known serine-protease inhibitor, blocked ADH proteolysis. The fact that other serine-protease inhibitors specific for trypsin-like enzymes did not affect this proteolytic activity, led us to in- clude it within the subtilase group of proteases (Siezen et al., 1991). Two-dimensional immunoblots were also performed with different D . rnelanogaster Adh alleloenzymes (data not shown) and other Drosophila species (Fig. 2) . All our results were consistent with co-migration of the serine-protease ac- tivity and active ADH forms on native IEF gels. They also showed that it was not restricted to D. melanogaster, nor to a single allelic form of ADH, nor to a particular electromorph.

Evidence of an SDS-induced instability in Drosophila ADH preparations

To characterize this proteolytic activity further, analyses were carried out using purified ADH preparations. DrosophiZa ADH has been regularly purified in our laboratory from dif- ferent species since 1980. Highly concentrated samples of purified enzyme preparations have provided material for bio- chemical and structural studies (Juan and Gonzalez-Duarte, 1980; Vilageliu and Gonzdlez-Duarte, 1984; Atrian and

361

Chymo- &Lacto- Albumin, Lysozyme ttyprinogen A globulin Myoglobin Bovine Yeast-ADH

M r 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

Fig. 5. Degradation of different substrates by Drosophila proteolytic activity monitored by SDS/PAGE. 2 pg purified proteins of commercial source were mixed with aliquots of ADH-G3 containing 0.3 pg ADH; 2 x sample buffer was added 1 : 1 to each solution and immediately heated at 100°C in a boiling water bath (lane 1) or incubated at room temperature for 3 h (lane 3). Control samples were incubated for 3 h in sample buffer without SDS (lane 2). Detection of peptide fragments was by Coomassie Blue staining. Marker proteins (Mr) are shown as in Fig. 1

Gonzalez-Duarte 1985; Winberg et al., 1986; Ribas de Pouplana et al., 1991) after fulfilling several criteria of hom- ogeneity, such as SDS/PAGE, IEF gels, amino acid compo- sition, absorption spectra and also catalytic stability at 4°C. Nevertheless, ADH instability was detected in the presence of an SDS-containing buffer (Fig. 3). A single 27-kDa band corresponding to the ADH monomer was observed only when the sample was boiled immediately after the sample buffer had been added, whereas a discrete pattern of bands appeared when, previous to boiling, the sample was incubated in sample buffer for 15 min at room temperature. A set of nine bands were distinguished in a high-resolution SDS/PAGE gel, corre- sponding to 23.9, 23.0, 21.7, 19, 1, 15.8, 12.5, 10.9, 9.3 and 4.7 kDa. To rule out an artefact of the PAGE assay or any effect induced by sample buffer components other than SDS, aliquots of purified enzyme were incubated in 2% SDS and injected into an HPLC column, as described in Material and Methods. A single peak was obtained if the sample had been injected immediately after detergent addition (Fig. 4A), while this initial peak practically disappeared after a 3-h SDS incu- bation (Fig. 4B). SDSjPAGE analysis of aliquots of each peak revealed peptide fragments which corresponded to the pre- viously described pattern. Therefore, SDS appeared to be the only agent responsible for proteolysis, which allowed us to rule out the effects of other sample buffer compounds.

Proteolytic activity in Drosophila ADH preparations

In an attempt to obtain D. rnelanogaster ADHF free from proteolytic activity, fractions eluted from Sephacryl S-200 (Pharmacia-LKB), the last step of a standard purification procedure, were pooled in three groups according to the level of ADH activity. ADH contained in the G2 group of fractions (ADH-G2), had activity of 2 3.0 unit/ml. ADH contained in the G1 group of fractions (ADH-Gl), eluting before the main peak, and ADH contained in the G3 group of fractions (ADH- G3), eluting after the main peak, contained those fractions with lower enzymatic activity (Fig. 3A). Stability of ADH- G1, ADH-G2 and ADH-G3 was tested in the presence of SDS following the procedure already described (Fig. 3B). ADH- G2 showed higher stability, after a 15-min incubation only faint bands were detected. ADH-GI had been concentrated

10-fold to operate with comparable amounts of ADH, this is approximatelly 5 pg for each proteolytic assay. Purification of ADH from D . lebanonensis following the same rationale yielded a similar Sephacryl-S200 elution pattern. So the same three pools were obtained, although their proteolytic banding pattern differed from those of D. melanogaster (data not shown).

In an attempt to separate both activities further chromatography was performed with the ADH-G3 pool. Ali- quots of ADH-G3 were rechromatographed on Sephacryl S- 200, or processed through Superose 6, or Superose 12 and Superose 6 in series. Fractions containing ADH activity were pooled and tested for proteolytic activity (Fig. 3C). In all cases, a degradation pattern was observed in the presence of SDS, which was more evident after overnight incubation. The fact that both ADH activity and proteolysis remained together, even after a second step of gel filtration with different supports, seems to argue in favour of a very tight association between these two activities. Moreover, ADH samples obtained from D. melanogaster homogenates following a dif- ferent purification protocol, initially designed for medium- chain dehydrogenases (Danielsson et al., 1992) which includes a DEAE-Sepharose ion-exchange step, yielded an identical pattern of proteolysis.

The separation of both activities was also approached by conventional two-dimensional electrophoresis (IEF + SDS/ PAGE) with aliquots of ADH-G3 from D. melanogaster AdhF and D. lebanonensis. The first dimension of control strips showed one major electromorph when stained for total protein or incubated to reveal ADH activity. After the second dimen- sion, degradation peptides again appeared in a vertical array below the active ADH spot in both species. Again, in this assay, proteolysis was precluded in the absence of SDS and in the presence of PhMeSOzF (data not shown).

Characterization of the proteolytic activity

The proteolytic activity of ADH-G3 was assayed on differ- ent proteins: yeast ADH, /?-lactoglobulin, bovine serum albu- min, chymotrypsinogen A, lysozyme and myoglobin (Fig. 5). All underwent complete degradation after a 3-h incubation in the presence of SDS, whereas controls without SDS remained

362

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 A

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 B

C 1 2 3 1 ' 2 3 4

Fig. 6. Effect of Drosophila proteolytic activity and subtilisin on purified ADH. (A) Silver-stained SDS/PAGE of ADH-G3 aliquots (0.7 pg ADH) mixed 1 : 1 with 2 x sample buffer and incubated at room temperature for the following times: lanes 3 - 12,l min, 3 min, 5 min, 7 min, 10 min, 15 min, 30 min, 60 min, 120 min, and 180 min, respec- tively. Controls are shown (lanes 1, 2 and 13). ADH-G2 incubated with 2 x sample buffer (1 : 1) for 10 min (lane l), ADH-G3 heated at 100°C in a boiling water bath after sample buffer addition (lane 2) or incubated at room temperature for 10 rnin in the presence of 1 mM PhMeS02F (lane 13). (B) Subtilisin activity on ADH-G2 was tested in the same conditions using a concentration of 0.5 pg/ml of subtilisin. Reaction mixtures in lanes 3 - 12 were incubated for the same times as in (A). Controls are: (lane 1) ADH-G2 heated at 100°C after sample buffer addition; lane 2, incubated for 10 rnin at room temperature in the presence of 1 mM PhMeS02F. (C) Silver-stained SDS/PAGE of ADH proteolytic pattern produced by subtilisin and Drosophila endogenous protease activity. Lane 2, an aliquot of ADH-G2 (0.7 pg ADH) was incubated with subtilisin (0.5 pg/ml) in sample buffer for 15 min. Lane 3, ADH-G3 (0.7 pg ADH) was incubated for 15 rnin at room temperature in a 1 : 1 mixture with sample buffer; lane 1, ADH- G2 incubated at room temperature for 15 rnin used as control. (D) Immunoblot assay of ADH proteolytic pattern produced by subtilisin and Drosophila endogenous protease activity. ADH cross-reacting bands after a treatment of ADH-G2 with subtilisin (lanes 1 and 3) or endogenous protease activity (lanes 2 and 4) immunodetected with mAb LLBE8 (lanes 1 and 2) and mAb MMBB8 (lanes 3 and 4). Samples were processed as in C. Arrows in C and D indicate apparent equivalent bands which have been shown different by immunodetec- tion.

unaltered. These results stressed the essential contribution of SDS to the proteolytic activity. To establish the kinetics of this proteolysis, aliquots of ADH-G3 were incubated in SDS solution for different times, at room temperature. Four high- molecular-mass fragments (23.9, 23.0, 21.7 and 19.1 kDa)

were observed after a 1-min incubation (Fig. 6A). Longer incubation times (2 h onwards) resulted in the progressive appearence of smaller peptides at the expense of larger ones, including the ADH monomer. The same pattern was produced when samples were incubated at 4°C. Moreover, this proteo- lytic activity was restricted to a pH range of 5-7, far from the optimum pH values (8.5 - 9.0) for ADH activity. Protease inhibitors were again used to characterize the proteolytic ac- tivity in purified ADH solutions and PhMeSOzF was, as in crude homogenates, the only effective compound (Fig. 6A, lane 13). The ample range of substrates for this proteolytic activity and the inhibitory effect of PhMeS02F are both fea- tures associated with proteases of the subtilase family. More- over, trypsin and chymotrypsin failed to degrade ADH incu- bated under the same standard conditions, whereas the com- mercial subtilisin-Calsberg enzyme produced a discrete ADH degradation pattern. These results encouraged us to carry out additional assays to assess the proteolytic effect of subtilisin on ADH and to compare it with the Drosophila endogenous proteolytic activity.

Additional bands, besides that of ADH, were occasionally detected on SDS gels (Fig. 3 and 6A). These three bands, with molecular masses between 30 - 45 kDa, were potential candidates for proteolytic contaminants. Two of those, ap- pearing on Fig. 3, were discarded on the basis of the lack of correlation between their intensities and the level of proteo- lytic activity on those samples. The third one, of 45 kDa (Fig. 6A), was separated from ADH through an FPLC step with Superose 6, and then assayed as a proteolytic agent for ADH in the presence of SDS. Results obtained allowed us to rule out any contribution of the 45-kDa polypeptide to the proteolytic process.

Effect of subtilisin on stable fractions of Drosophila ADH

Subtilisins belong to the broad specificity group of alkaline serine proteases. They are irreversibly inhibited by PhMeSOzF and considerably resistant to denaturing agents, such as SDS and high temperature (Ottesen and Svendsen, 1970). The known similarity with the features of the Drosophila proteo- lytic activity led us to analyze the degradation pattern pro- duced in pure ADH (Fig. 6B). In the case of subtilisin, there was a time lag before ADH fragmentation began, but as soon as it became detectable all size fragments appeared at once. The differential behaviour could easily be attributable to the different proteolytic activities involved and to the kinetics operating in each case. ADH proteolytic patterns produced by subtilisin and Drosophila are compared in Fig.6C. The number and size of the proteolytic fragments were similar, although immunodetection of these peptides with specific mAbs revealed differences in their antigenic nature (Fig. 6D), which were even more obvious when comparing those of low molecular mass (see arrow in Fig. 6D).

DISCUSSION

Apparently homogeneous Drosophila alcohol dehydro- genase preparations show a closely linked proteolytic activity which has not been previously described and which supports recent findings in histological and ontogenic analysis (Visa et al., 1992). Different types of experiments have allowed us to characterize this proteolytic process as being associated to an SDS-dependent serine protease activity related to the subtilase family, which produces a discrete ADH banding pattern in

363

Table 1. Alignment of D. melunogaster ADHF (Mew with (a) B. subtilis umylosucchuriticus subtilisin (subt) and (b) 2Ob-hydroxysteroid dehydrogenase (20b). The alignments were obtained using the GAP program. Vertical bars denote identity matches and colons denote conservative substitutions. Asp, His and Ser catalytic residues are boxed in the subtilisin sequence and equivalent positions in Drosophila ADH are signaled by an arrow. A Lys-Arg subtilase precursor-processing signal (Barr, 1991) is also found at the N-terminus of Drosophila ADH, as underlined.

(a)

Melf x Subt: Percent Similarity: 39.916 Percent Identity: 21.008

1

1

51

46

99

94

14 5

137

185

185

235

234

(b)

1 . 1 . SFTLTNKNVIFVAGLGGIGLDTSKELLEDLKNLVILDRIENPAAIAELK 50

. . . I :. : . I.: . . : : I 1 1 . I : . . : ::: ..... AQ~VPYGISQI~PALHSQGYTGSNV~AVIDSG@SHPDL~ 45 AINPKVTVTFYPYDVTVPIAETTKLLKTIFAQLKTVDVL. -1NGAGILDD 98 : .. I . . I I : * . . . I . : 1 1 I . . : : I1 : I : : . ,

. : : . I . . . I . : : I .... : I . :.. I I I : I I : : .

.: : : I 1 1 . [ I . . ( .I1 1 1 . : . .. : I . [ 1 1 . :

GGASFVPSETNPYQDG..S@THVAGTIAALNNSIGVLGVAPSASLYAV 93

HQIERTIAVNYTGLVN. ... TTTAILDFWDKRKGGPGGPGGIICNIGSVTGFN 144 KVLDSTGSGQYSWIINGIEWAISNNMDVINMSLGGPSG.. ..... STALK 136 AIYQVPVYSG .. TKAAWNFTSS ........ LAKLAPITGVTAYTVNPGI 184 TWDKAVSSGIWAAAAGNEGSSGSSSTVGYPAKYPSTIAVGA..VNSSN 184

TRTTLVHTFNSWLDVEPQVAEKLLAHPTQPSLACAENFVKIELNQNGAI 234 - 1

I . . : . . I I I I :.... . I . . . I : . : . - . . . - . . . . - 1 : QRASFSSA. GSELDVMAPGVSIQSTLPGGTYGAYNGT~J~ATPHVAGAAAL 23 3

WKLDLGTLEAIQWTKHWDSGI ........... .......... 255 ILSKHPTWTNAQVRDRLESTATYLGDSFYYGKGLINVQAAAQ 275

. . I : . . I .:::I.

Melf x 20b: Percent Similarity: 44.813 Percent Identity: 20.332

1 SFTLTNKNVIFVAGLGGIGLDTSKELLKRDLKNLVILDRIENPAAIAELK 50

1 MNDLSGKTVIITGGARGLGAEAARQAVAAGAR..WLADVLDEEGAATAR 48

51 AINPKVTVTFYPYDVTVPIAETTKLLKTIFAQLKTVDVLINGAGI ..... 95 49 ELGD..AARYQHLDVTIE.EDWQRWAYAREEFGSVDGLVNNAGISTGMF 95

96 LDDHQIER ... TIAVNYTGLNTTTAILDFWDKRKGGPGGPGGIICNI~SVTG 142 96 LETESVERFRKWDINLTGVFIGMKTVI...PAMKDAGGGSIVNISSAAG 142

143 FNAIYQVPVYSGTKAAWNFTSSLAKLAPITGVTAYTVNPGITRTTLVHT 192

143 LMGLALTSSYGASKWGVRGLSKLAAVELGTDRIRVNSVHPGMTYTPMTAE 192

193 FNSWLDVEPQVAEKLLAHPTQPSLACAENFVXAIELNQNGAIWKLDLGTL 242

193 TGIR.QGEGNYPNTPMGRVGNEPGEIAGAWKLL. .... SDTSSYVTGAE 236 243 EAIQWTKHWDSGI ...... 255 237 LAVDGGWTTGPTVKYVMGQ 255

. I . . [ . [ ] : . : I I : ( :..:: : : : I : / . : :..: I . :

.:.. .. : . . I I I : . .: ::: .:: . I 1 . I : I . I 1 1

I : . . : I1 . : . : I . [ / : . . .... .. . . I : : . I I I . I I : I . . I

. . .. .. .. l::.l : I .:.. I ... : . . [ : I I : I I . : . .

I . . : : . I . . ::. :::..... . . I : . . [ / : ......

I : : .. .... . .

crude homogenates and in purified preparations. Co-mi- gration of ADH and the endogenous proteolytic activity was a constant feature every time crude homogenates were re- solved by two-dimensional analysis. When purified ADH so- lutions were proccessed in order to obtain ADH protease-free preparations, we observed that proteolytic activity was present in all ADH fractions obtained after successive gel filtration steps. Some ADH fractions from these purified preparations

appeared to be more stable than others, but even those, given enough time at the standard conditions, generated the proteo- lytic pattern. Although the hypothesis of a minor contaminant with protease activity co-purifying with ADH cannot be dis- carded, since it appears to be favoured when analyzing the patterns shown in Fig. 3, the unsuccessful attempts to obtain fractions with protease activity devoid of ADH, following many different purification procedures, point to their tight

364

association. The additional proteins we occasionally detected in purified preparations, and that were once considered candi- dates of proteolytic agents, have also been ruled out.

Two different phenomena could account for these results, both of which have already been reported in the literature. Several yeast enzyme purification rationales led to apparently homogeneous preparations which eventually showed an SDS- dependent protease contamination which was tightly adsorbed to a subpopulation of enzyme molecules. Rustum et al. (1971) concluded that traces of an active protease formed some kind of complex with a small percentage of yeast hexokinase molecules. Attempts to isolate it in free form were unsuccessful. Moreover, this protease was completely stable at 4°C or at room temperature, was sensitive to boiling and PhMeS02F treatment, and was activated in 1 YO SDS produc- ing a specific hexokinase cleavage. Other examples of SDS- activated linked proteases are those that affect yeast phosphofructokinase (Diezel et al., 1972) and bovine pituitary gland fibroblast growth factor (Ho et al., 1990). In fact, all the features ascribed to these proteases apply to the Drosophila proteolytic activity. It is unfortunate that proteases are still mostly unknown in Drosophila and the very few reported, presumably involved in food digestion, belong to the trypsin- like family (Chen, 1978). Therefore, we could consider that there is a subtilase in Drosophila closely associated with an active ADH subpopulation, which remains latent until it is activated by SDS.

On the other hand, the assumption of Mitchell et al. con- cerning the autoproteolytic potential of some Drosophila pro- teins, Hsp70 and several wing polypeptides, could provide us with an alternative explanation (Mitchell et al., 1985; Mitchell and Petersen, 1987). They argued that if a contaminant pro- tease was involved in proteolysis, it would have to share the isoelectric point and molecular size with each protein assayed. As this is improbable, they assumed that their protein had become a protease. The acquisition of proteolytic activity by non-protease proteins could be sustained by the structural potential of their primary structure, providing Ser, His and Asp residues were in suitable positions to produce a tertiary folding that would mimic the serine protease active domain. Moreover, SDS would provide the proteins with the structural changes needed to fix them in a stable ‘protease confor- mation’, besides facilitating the cleavage of the substrate by extensive denaturation. To establish any primary structure similarity of ADH with subtilases, we compared the sequence of D . melanogaster ADH with that of Bacillus subtilis subtilisin (Table 1 ) . The degree of similarity found is by no means trivial because it amounts to 40% with 21% identity, both polypeptides being approximately of the same size, whereas the values obtained for a member of the short-chain dehydro- genase family (Persson et al., 1991), 20P-hydroxysteroid de- hydrogenase, are of the same order, 44% similarity and 20% identity. Of the three residues: Asp, His and Ser which consti- tute the subtilisin catalytic triad, two are found in ADH at equivalent positions. It has been reported that, even in the absence of one of those residues, peptide hydrolysis proceeds at a catalytic rate in subtilisin, then the absence of His in Drosophila ADH would not preclude its potential proteolytic activity.

The previously reported differential banding pattern dur- ing development could suggest a role in vivo for this protease activity. On the one hand, the SDS inducer effect in vitro could be produced in vivo by fatty acids. These physiological compounds have been considered as possible mediators in protein conformational changes (Mitchell et al., 1985) supply-

ing the intracellular requirements for the expression of thls ‘latent’ proteolytic activity. The participation of Drosophila ADH in fat metabolism is well documented (Heinstra et al., 1990; Freriksen et al., 1991), so it would be reasonable to assume that these compounds are the true inductors of a degradative process in vivo bound to the turnover of the en- zyme.

In summary, we present two alternative hypotheses to explain an ADH-associated proteolysis in Drosophila. At pre- sent we cannot discard one or the other. The autoproteolytic hypothesis, supported by structural analysis, would be in agreement with previous information reported in the literature about dual activities coexisting in a single enzyme (Brady et al., 1990) whose functional implications have yet to be elucidated. Or we could also be dealing with an evolutionary relic, cryptically maintained, of the ADH structure. On the other hand, if there was a subpopulation of native ADH molecules intimately associated with a latent serine protease, the molecular nature of this interaction is totally unknown at present. Results obtained with purified enzyme correlate well with previous findings of proteolytic susceptibility of ADH in crude homogenates detected with anti-ADH monoclonal antibodies. Our results contribute to the analysis of ADH turnover in vivo and provide new data concerning the physio- logical behaviour of ADH. Moreover, this information could be of relevance when trying to interpret new data concerning the structural features and chemical interactions of this en- zyme.

We are grateful to Dr. L. Enjuanes for helpful discussions. HPLC analyses were carried out at the Serveis Cientifico-Tkcnics de la Universitat de Barcelona. We are obliged to Dr. I. Casals for designing HPLC analyses. R. Rycroft’s encouraging correction of the manu- script is highly appreciated. This work was supported by grants from Fondo de Investigaciones Sanitarias (FIS, 88/0833 and 89/0179) and Dir. Gen. de Investigacidn Cientifica y T2cnica (DGICYT, PB88-0188). JF was a recipient of a Formacibn Personal Investigador (FPI) fellow- ship from the Ministerio de Educacidn y Ciencia.

REFERENCES Atrian, S. & Gonzalez-Duarte, R. (1985) Biochem. Genet. 23, 891 -

Barr, P. J. (1991) Cell 66, 1-3. Bradford, M. M. (1976) Anal. Biochem. 72,248-254. Brady, L., Brzozowski, A. M., Derewenda, Z. S., Dodson, E.,

Dodson, G., Tolley, S., Turkenburg, J. P., Christiansen, L., Huge- Jensen, B., Norskov, L., Thim, L. & Menge, U. (1990) Nature 343,

911.

767 - 770. Chambers, G. K. (1988) Adv. Genet. 25, 39 - 107. Chambers, G. K. (1991) Comp. Biochem. Physiol. 99B, 723 - 730. Chen, P. S . (1978) in Biochemistry ofinsects, pp. 145-203, Academic

Danielsson, O., Eklund, H. & Jornvall, H. (1992) Biochemistry 31,

Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acids Res.

Diezel, W., Nissler, K., Heilmann, W., Kopperschlager, G. &

Dunn, S . (1986) Anal. Biochem. 157, 144-153. Fibla, J., Enjuanes, L. & Gonzalez-Duarte, R. (1989) Biochem. Bio-

Freriksen, A., Seykens, D., Scharloo, W. & Heinstra, P. W. H. (1991)

Heinstra, P. W. H., Seykens, D., Freriksen, A. & Geer, B. W. (1990)

Ho, P. L., Carpenter, M. R., Smillie, L. B. & Gambarini, A. G. (1990)

Press, New York.

3751 - 3759.

12, 387-395.

Hofmann, E. (1972) FEBS Lett. 27, 195-197.

phys. Res. Commun. 160,638-646.

J. Biol. Chem. 266, 21 399 - 21 403.

Insect Biochem. 20, 343 - 348.

Biochem. Biophys. lies. Commun. 170,769-714.

365

Juan, E. & Gonzllez-Duarte, R. (1980) Biochem. J . 189, 105-110. Kowit, J. D. & Maloney, J. (1982) Anal. Biochem. 123, 86-93. Laemmli, U. (1970) Nature 227,680-685. Mitchell, H. K., Petersen, N. S. & Buzin, C . H. (1985) Proc. Natl

Mitchell, H. K. & Petersen, N. S. (1987) J. Biol. Chem. 262, 14298-

O’Farrell, P. H. (1975) J . Biol. Chem. 250,4007-4021. Ottesen, M. & Svendsen, I. (1970) Methods Enzymol. 19, 199-215. Persson, B., Krook, M. & Jornvall, H. (1991) Eur. J. Biochem. 200,

537 - 543. Ribas de Pouplana, Ll., Atrian, S., Gonzllez-Duarte, R., Fothergill-

Gilmore, L. A,, Kelly, S. M. & Price, N. C . (1991) Biochem. J.

Acad. Sci. USA 82,4969-4973.

14304.

276,433-438.

Rittenhouse, J. & Marcus, F. (1984) Anal. Biochem. 138,442-448. Rustum, Y. M., Massaro, E. J. & Barnard, E. A. (1971) Biochemistry

Siezen, R. J., deVos, W. M., Leunissen, J. A. M. & Dijkstra, B. W.

Thatcher, D. R. (1977) Biochem. J. 163,317-323. Vilageliu, L1. & Gonzalez-Duarte, R. (1984) Biochem. Genet. 22,797 -

Visa, N., Fibla, J., Santa-Cruz, M. C. & Gonzalez-Duarte, R. (1992)

Winberg, J.-O., Hovik, R., McKinley-McKee, J. S., Juan, E. &

Wray, W., Boulikas, T., Wray, V. P. & Hancock, R. (1981) Anal.

10, 3509-3516.

(1991) Protein Eng. 4, 719-737.

815.

J. Histochem. Cytochem. 40, 39-49.

Gonzalez-Duarte, R. (1986) Biochem. J . 235,481 -490.

Biochem. 118,197-203.