DNA AND CELL BIOLOGY Volume 11, Number 9, 1992 Mary Ann Liebert, Inc., Publishers Pp. 661-672

Molecular Cloning and Functional Characterization of Chicken Cathepsin D, a Key Enzyme for Yolk Formation HELMUT RETZEK,* ERNST STEYRER,*>t ESMOND J. SANDERS.Î JOHANNES and WOLFGANG J. SCHNEIDER*>§

NIMPF,*>§

ABSTRACT

Upon receptor-mediated endocytosis of very-low-density lipoprotein (VLDL) and vitellogenin into growing chicken oocytes, the protein moieties of these lipoproteins are proteolytically cleaved. Unlike the complete lysosomal degradation in somatic cells, enzymatic ligand breakdown in oocytes generates a characteristic set of polypeptides, which enter yolk storage compartments for subsequent utilization by the embryo. Here, we demonstrate directly that the catalyst for the intraoocytic processing of both apolipoprotein B and vitellogenin is cathepsin D. The enzyme was purified from oocytic yolk, preovulatory follicle homogenates, and liver by affinity chromatography. When plasma VLDL and vitellogenin were incubated with the purified enzyme, fragments indistinguishable from those found in yolk were generated from both precursors under identical, mildly acidic conditions. Amino-terminal sequencing of the pure enzyme demonstrated 88% identity with mammalian cathepsin Ds over 34 residues. On the basis of this information, a full-length clone specifying chicken preprocathepsin D was isolated from a chicken follicle cDNA library by screening with a human cathepsin D probe. Whereas previous studies have demonstrated that the receptors for lipoproteins in somatic cells and oocytes, respectively, of the chicken are the products of different genes, Southern and Northern blot hybridization experiments showed that the enzymes expressed in oocytes and liver are the product of a single gene, giving rise to a 3.3-kb transcript. The primary structure of the 335-residue mature protein suggests a high degree of conservation of known crucial features of aspartyl proteases; however, the absence of the so-called processing region and of an aromatic residue in a region thought to partake in catalysis raise questions with possible evolutionary implications.

INTRODUCTION

fully grown, has a volume of up to 15 cm3! Co- and postendocytic sorting and trafficking of the plasma precursor of oviparous species grow rapidly through proteins by the oocyte are likely to involve processes disreceptor-mediated uptake from the plasma of a set of tinct, or at least modified, from those operating in somatic complex proteins that are synthesized in and secreted from cells. Indeed, studies in Xenopus oocytes have indicated the liver under the influence of estrogen (Bergink et al, that in contrast to complete proteolytic degradation in 1974; Tata, 1986). In the chicken, a 95-kD plasma mem- typical lysosomes of somatic cells, VTG undergoes specific brane protein has been shown to function in the uptake of cleavage into large fragments (lipovitellins, phosvitin, and two major lipoproteins, very-low-density lipoprotein possibly others; for review, see Wallace, 1985) within a (VLDL) and vitellogenin (VTG) into the oocyte, thereby quasi-lysosomal compartment termed "light yolk platelets" forming the bulk of the yolk (Stifani et al, 1990a). The in- (Wall and Meleka, 1985; Opresko et al, 1980). These disflux of large amounts of these and other yolk components tinct organdíes harbor some of the enzymes found in clasvia receptor-mediated endocytosis poses an enormous chai- sical lysosomes, and appear to be the precursors to mature, lenge to the endocytic machinery of this cell which, when "heavy yolk platelets" which represent the final destination

Oocytes

•Department of Biochemistry and Lipid and Lipoprotein Research Group, and ÍDepartment of Physiology, The University of Alberta, Edmonton, Alberta, Canada T6G 2S2. tCurrent address: Department of Medical Biochemistry, University of Graz, A-8010 Graz, Austria. §Current address: Department of Molecular Genetics, University and Biocenter Vienna, A-1030 Vienna, Austria. 661

RETZEK ET AL.

662 for storage (Wall and Meleka, 1985). The proteolytic processing of VTG in light yolk platelets of Xenopus oocytes appears to be a prerequisite for ultimate deposition in heavy yolk platelets, since pepstatin A, a specific inhibitor of certain aspartic proteases such as cathepsin D, blocks both VTG cleavage and progression from light to heavy yolk platelets (Opresko and Karpf, 1987). Recently, we have become interested in postreceptor events governing the deposition of yolk proteins into growing oocytes of the chicken. In contrast to yolk of Xenopus oocytes, which mainly consists of VTG, the bulk of chicken yolk is a complex mixture of two lipoproteins,

VTG and VLDL. In the case of VLDL, which binds to the 95-kD oocyte receptor via apolipoprotein (apo) B (Nimpf et al, 1988), we could previously show that a pepstatin A-sensitive protease cleaves the apoB, thereby producing lipoprotein particles (termed "yolk VLDL") that contain a characteristic set of apo B-derived fragments (Nimpf et al, 1989 a). Yolk VLDL is stored, apparently without further processing, in the oocyte until use by the developing embryo. Conceivably, the embryo has direct access to these energy-rich particles via receptor-mediated endocytosis, since yolk VLDL retains its ability to bind to chicken lipoprotein receptors specific for apo B (Nimpf et al, 1989a). Thus, the initial proteolysis of VLDL, and presumably also the postendocytic processing of VTG (about which less is known), plays a significant role in organellar structure and function of the chicken oocyte, i.e., in the reproductive effort of the hen. In the present study, we have begun to delineate specific yolk protein processing pathways in the chicken oocyte by identification, purification, functional characterization, and cDNA cloning of a protease that is capable of catalyzing the bona fide intraoocytic breakdown of the protein components of both VLDL and VTG.

MATERIALS AND METHODS Materials We obtained pepstatin A (Cat. No. P4265), pepstatin A-agarose (Cat. No. P2032), and bovine hemoglobin (Cat. No. 2625) from Sigma; molecular mass standards were from BRL and Sigma. The molecular biology reagents and

enzymes were obtained from Bethesda Research Laboratories, New England Biolabs, Pharmacia, or Boehringer Mannheim. All other materials were from previously re-

ported

sources

(Stifani

et

al, 1990a,b).

Animals White Leghorn layers (8-18 months old) were obtained from the Department of Animal Sciences, University of Alberta, and maintained on layer mesh and fed ad libidum, with a light period of 12 hr. We used adult female New Zealand White rabbits for raising antibodies.

Preparation of substrates and enzymatic digestion VLDL from plasma (Nimpf et al, 1988) and yolk (Nimpf et al, 1989a), and VTG (Stifani et al, 1988) and

lipovitelin (Stifani

et al, 1990b) were prepared as described in the indicated references. The standard cathepsin D assay was performed at 37° C in a volume of 400 pi using 1 % hemoglobin as substrate in 200 mM glycine-HCl buffer pH 3.5. Undegraded substrate was precipitated by adding ice-cold trichloracetic acid (final concentration, 3%) and centrifugation at 13,000 x g for 10 min, and acid-soluble fragments were determined by absorbance at 280 nm. The amount of substrate degraded in the presence of 5 pg/ml of pepstatin A was subtracted from that degraded in its absence to determine cathepsin D activity. One unit (U) of cathepsin D is defined as an absorbance of 1.0 at 280 nm after incubation for 1 hr. Incubations with VLDL or VTG contained 30 pg of substrate protein, 0.7 units of purified enzyme from either yolk or liver, 120 mM citrate buffer with the pH values indicated in the figure legends, and 0 or 2 pg peptstatin A in a final volume of 50 pi. Incubations were performed at 37°C for 16 hr, and prepared for

NaDodSOi-polyacrylamide gel electrophoresis (PAGE) by precipitation with trichloroacetic acid (final concentration, 10%) in the presence of 0.03% sodium desoxycholate, followed by washing with 500 pi of acetone that had been precooled to -20°C, and drying under vacuum.

Purification of chicken cathepsin

D

From the Liver: The initial steps of the method of Barret (1970) were modified and combined with affinity chromatography on pepstatin A-agarose as follows. All operations were performed at 4°C. Fresh chicken liver (80 grams) was minced, mixed with 80 ml of ice-cold H20 containing 1 mM PMSF, 2 pM leupeptin, and 1 /¿g/ml aprotinin, and homogenized in a Waring Blendor. The resulting homogenate was adjusted to 0.15 M CaCl2, followed by slow addition of 1 M Na2HPO„ to a final concentration of 0.15 M with continuous blending action. The mixture was subjected to centrifugation at 15,000 x g for 10 min, the resulting supernatant filtered through cheesecloth, and the pH adjusted to 3.5 with 1 NHC1. Then, 0.9 volumes of acetone (precooled to -20°C) were slowly added under

stirring, and the precipitate removed by centrifugation as above. The resulting supernatant was precipitated with acetone at a final concentration of 64% (by volume), the precipitate recovered by centrifugation as above, and resuspended in 25 mMTris-HCl pH 7.8. Insoluble material was removed by centrifugation (15,000 x g, 10 min), and the solution mixed with 4 volumes of buffer containing 200 mM glycine-HCl, 0.5 MNaCl pH 3.5 (binding buffer) followed by a final centrifugation step to remove precipitating material. Pepstatin A-agarose (1 ml settled bed volume/100 ml sample), equilibrated in binding buffer was added to the solution and the suspension incubated at 4°C for 6 hr on a rotating platform. The affinity matrix was transferred to a column, and washed successively with 50 bed volumes each of binding buffer, binding buffer + 6 M urea, and binding buffer, respectively. Elution was performed with buffer containing 100 mM Tris-HCl, 0.05 M NaCl pH 8.4; the enzyme solution was adjusted to pH 5.5 with 0.5 M acetic acid and stored at -70°C.

CATHEPSIN D IN YOLK FORMATION From Developing Follicles: All operations were performed at 4°C. Follicles (3-5 mm in diameter) were dissected free of adherent connective tissue, rinsed quickly in 150 mM NaCl, and homogenized in ice-cold H20 containing the mixture of protease inhibitors described above (5 ml/g of follicles) with a Polytron homogenizer (2 times 30 sec at setting 7). The homogenate underwent one freezing/thawing cycle (-70°C/37°C) and was mixed with an equal volume of 50 mM Tris-HCl pH 7.8 followed by centrifugation at 100,000 x g for 30 min. The supernatant, 30 ml, was applied to a column (0.8 x 8 cm) of DEAE-cellulose (DE52, Whatman), equilibrated in 25 mM Tris-HCl pH 7.8. The column was washed with 10 bed volumes of the same buffer, and then a linear gradient of NaCl (0-200 mM) in the same buffer was applied. Active fractions were pooled and mixed with 4 volumes of binding buffer (see above), and affinity purification on pepstatin A-agarose performed as described for the liver en-

663

tion. Replicate filters were hybridized with 2 x 106 cpm/ ml of the probe in 50% (wt/vol) formamide, 6 x SSC (1 x SSC is 150 mM NaCl, 15 mM sodium citrate pH 7), 0.1% (wt/vol) SDS, 100 mg/ml dextranesulfate, 1 x Denhardt's solution (Sambrook et al, 1989), and 0.1 mg/ml sheared salmon sperm DNA at 37°C, and washed with 1 x SSC at 55 °C prior to autoradiography. DNA was prepared from lysates of purified plaques; inserts were excised with Eco RI and subcloned into the multiple cloning site of the plasmid pTZ19 (Pharmacia).

typically

zyme.

From Follicular Yolk: For preliminary digestion and other experiments (not shown), we prepared partially purified yolk cathepsin D by the following alternative procedure, which gave higher yields. Typically, 200 ml of yolk (obtained by extrusion from follicles 2-3 cm in diameter), were diluted with 400 ml of ice-cold H20 containing the above protease inhibitors, and the suspension was stirred vigorously for 30 min. Following centrifugation at 20,000 x g for 30 min, the supernatant was freeze-thawed as above, the centrifugation step repeated, and the solution subjected to DEAE-cellulose chromatography as described above. The active fractions emerging from the DEAE-cellulose column were pooled, adjusted to 15% (wt/vol) of ammonium sulfate pH 6.8, clarified by centrifugation (15,000 x g, 10 min), and applied to a column (1.5 x 10 cm) of phenyl-Sepharose equilibrated in 10 mM sodium phosphate, 15% ammonium sulfate pH 6.8. The column was washed with 5 bed volumes of the same buffer, and then eluted with a linear gradient (total of 30 bed volumes) of 15 to 0% ammonium sulfate in 10 mM sodium phosphate pH 6.8. Activity, eluting at approximately 9% ammonium sulfate, was pooled and dialyzed against 50 mM sodium acetate pH 4.8 and subjected to chromatography on CM-cellulose (CM 52, Whatman). The column (0.8 X 6 cm) had been equilibrated in the dialysis buffer, was washed with 2 bed volumes of the same buffer following sample application, and eluted under reverse flow with 50 mM sodium acetate pH 5.5. Active fractions were pooled and stored at -70°C.

DNA

sequencing

CsCl gradient-purified pTZ plasmids were linearized and digested with Exo III and SI nuclease to create a set of nested deletions. After filling the ends using Klenow fragment, the plasmids were religated with T4 ligase and transformed into competent E. coli JM 109. Cells were plated on LB plates containing 50 pg/ml ampicillin. Colonies were grown in 2 ml of LB (50 pg/ml ampicillin) overnight. Aliquots of the dense cultures were used for DNA preparation and the remaining cells were infected with 80 ^.1 of the helper phage M13K07 (10M09 pfu/ml). After addition of 2 ml of LB + ampicillin + kanamycin, the cells were grown overnight. Single-stranded plasmid DNA was prepared (Sambrook et al, 1989) and used as templates for sequencing. DNA sequencing was performed by the dideoxy method using the large fragment of DNA polymerase I and the 17-mer reverse sequence primer. Insert DNA was sequenced on both strands.

Southern and Northern analysis Genomic Southern blotting was performed using 10 pg of DNA prepared from chicken liver (Davis et al, 1986) for each restriction digestion. The DNA fragments were resolved by 0.8% agarose gel electrophoresis and transferred to Hybond N. Total RNA from the indicated tissues was obtained by the guanidinium isothiocyanate-CsCl method (Chirgwin et al, 1979), and poly(A)*RNA isolated by

oligo(dT)-cellulose chromatography (Sambrook et al, 1989). After electrophoretic separation on 1.1% agarose gels, the RNA was transferred to Hybond N. All blot hybridizations were performed in 50% (wt/vol) formamide, 6 x SSC, 1 x Denhardt's solution, 0.1 mg/ml salmon sperm DNA, 100 mg/ml of dextrane sulfate, and 0.1% NaDodSO,, at the temperature indicated in the figure legends.

Other methods cDNA cloning cDNA library was synthesized by random priming from mRNA derived from chicken follicles (4-6 mm in diameter) following surgical removal of thecal layers (Clontech). Recombinant phages were propagated in Escherichia coli Y 1088. Approximately 3 x 104 plaques were screened with a partial cDNA clone for human cathepsin D (pHG2CDl.l, kindly provided by Dr. J.M. Chirgwin). The probe was labelled with [32P]dCTP by nick translaA

Xgtll

The protein content of samples was determined by the method of Bradford (1976). Immunoglobulin G (IgG) fractions were purified from sera on columns of Protein A-Sepharose CL-4B (Stifani et al, 1990b). Affinity-purified rabbit anti-chicken cathepsin D IgG for electron microscopical immunocytochemistry was prepared as described (Robinson et al, 1988), with the Protein A-Sepharose-purified IgG as starting material. NaDodS04-PAGE was conducted on 4.5-18% gradient gels, according to the

664

RETZEK ET AL.

method of Laemmli (1970); samples were heated to 90°C for 5 min in the presence of 20 mMdithiothreitol. Aminoterminal protein sequence analysis was performed according to the manufacturer's protocols on an Applied Biosystems gas-phase Protein Sequencer model 470A coupled to an Applied Biosystems model 120A HPLC for on-line 3phenyl-2-thiohydantoin identification. All protein sequencing chemicals were obtained from Applied Biosystems.

RESULTS The approach to isolate the protease(s) from ovarian follicles was based on our previous finding that pepstatin A inhibited the breakdown of plasma VLDL-apoB to the fragments observed in yolk VLDL by a yolk "platelet" fraction (Nimpf et al, 1989a). In preliminary experiments, we isolated and partially purified an active fraction from the yolk of large follicles (see Materials and Methods), and subsequently determined that the highest specific activity of the pepstatin-inhibitable enzyme was present in small (3to 5-mm diameter) follicles. Thus, starting from a homogenate of small follicles, we used pepstatin A-Sepharose affinity chromatography as the principal step in the purification scheme. From 30 grams of tissue, we obtained —50 pg of enzyme (yield, 3%); upon analysis by NaDodS04PAGE (Fig. 1, lane 3), the highly active preparation showed three bands with apparent Mr values of 43,000, 30,000 and -14,000, respectively, in agreement with the known existence of single-chain (43 kD) and two-chain (30 kD and 14 kD) forms of cathepsin D in a variety of species (Hasilik and von Figura, 1984; Faust et al, 1985; Tang and Wong, 1987). The enzyme preparation also contained a contaminant which comigrated with the major protein of the fraction obtained in the penultimate purification step (Fig. 1, lane 4). This contaminant could be removed only under conditions that led to significant loss of enzyme activity. Since only limited amounts of starting material were available to us, and the yield was low, we attempted purification of the enzyme from chicken liver. A combination of the procedure by Barrett (1970) and pepstatin A-affinity chromatography as described in Materials and Methods resulted in highly active (2,500 U/mg) purified enzyme. From 80 grams of chicken liver, we isolated 1.4 mg (yield, -25%) of the material shown in Fig. 1 (lane 1). It contained the same three polypeptides as the enzyme obtained from developing follicles, and was devoid of the contaminant consistently present in the material from the ovarian source.

To confirm that the

purified enzymatic activity was due to cathepsin D, we performed amino-terminal protein sequencing of the 43-kD polypeptide. We determined the first 34 residues, which revealed 88% identity with aminoterminal sequences of several mammalian cathepsin Ds (30 matches out of 34, with one difference due to a conservative replacement, Val to He), and less similarity (16 mismatches) to chicken pepsinogen (Fig. 2). These data putatively identified the affinity-purified protein as chicken ca-

FIG. 1. Isolation of cathepsin D from chicken tissues. The enzyme was purified from liver (lane 1) and developing ovarian follicles (lanes 3 and 4) as described in Materials and Methods, and analyzed by NaDodS04-PAGE (4.5-18%) gel electrophoresis. Lane 1, 15 pg of purified liver enzyme; lane 2, Mr standards (from top to bottom: myosin, 200 kD, /3-galactosidase, 116 kD; phosphorylase b, 97 kD; bovine serum albumin, 68 kD; ovalbumin, 45 kD; a-chymotrypsinogen, 26 kD, and pMactoglobulin, 18 kD); lane 3, 15 pg of affinity-purified enzyme from follicles; lane 4, 40 pg of protein from the DEAE-cellulose

chromatography step preceding chromatography statin A-agarose (see Materials and Methods).

on

pep-

thepsin D and facilitated its further structural characterization via homology cloning. Because the main focus of our research program is on genes involved in oocyte growth and follicular development, we proceeded with molecular cloning of the enzyme from our follicle cDNA library in Xgtll. Screening of 30,000 recombinants with a human cathepsin D sub—

clone, pHG2CDl.l (Faust et al, 1985; kindly provided by

Dr. John M. Chirgwin) resulted in 12 hybridizing signals; sequencing of three of the inserts, phrCCD2, phrCCD3, and phrCCD5, confirmed extensive homology to human cathepsin D. The longest clone, phrCCD3, (insert length, 1.4 kbp) was completely sequenced, and confirmatory sequencing was performed on clones phrCCD2 ( 1.3 kbp) and phrCCD5 (—1.2 kbp), which were positioned within the boundaries of phrCCD3. The nucleotide sequence and predicted amino acid sequence are shown in Fig. 3. Because we had determined the amino-terminus of the mature protein (Fig. 2), we could unambiguously align the nucleotide sequence with preprocathepsin D; there was perfect agreement between the amino-terminal sequence as determined from the cDNA and from the purified enzyme. —

—

CATHEPSIN D IN YOLK FORMATION Chi. PG

665

TATESYEPMTNYMDASYYGTISIGTPQQDFSVIF ••••••

•••

•

•

•

•

•


500,000) was converted to smaller fragments with apparent Mr values ranging from 40,000 to 180,000. In other experiments (not shown), plasma VLDL was treated with different substrate/enzyme

ratios and for different times to establish that the results of Fig. 7 represent end points in the digestive action of the enzyme. It is obvious from these data that cathepsin Ds isolated from liver and oocytes produced identical fragments of apo B, and that the generated peptides correspond very well to those present in yolk VLDL (Fig. 7, A and B, lanes 4). Particularly good agreement between the two patterns was observed when proteolysis had proceeded in the pH range from 5.4 to 5.6. In control experiments, the presence of pepstatin A completely inhibited the proteolysis (Fig. 7, lanes 7). These results with VLDL encouraged us to investigate the cathepsin D-catalyzed cleavage of VTG. Inasmuch as this other major yolk precursor shares the oocyte receptor with VLDL in the chicken (Stifani et al, 1990a), and its cleavage has been shown to be pepstatin A-sensitive in Xenopus oocytes (Opresko and Karpf, 1987), the production from plasma VTG of VTG-fragments found in yolk would further support its function in yolk formation and deposition. The major intraoocytic processing products of VTG are two groups of polypeptides termed lipovitellins

RETZEK ET AL.

666

1

-63 90 -45

180 -15 270 16

360 46

89

GGGGCCCCGTCAGCTGACGCTCCGTGCaa^CTGAGGCC»XGCTACAA

MAPRGLLVLLLLALVGPC A

A~L

IRIPLTKFTSTRRMLTEVGSEIPDMNA

GCGGCACTCATCAGGATCO0CCTCACCAAATT«OCTCCAOSCGCO3^

ITQFLKFKLGFADL

-46

179

TTAGG£GCTCTAGTGCCTTCTAAGO?raC03œGCTAT

V*E

PTPEILKNYMDAQY

-16 269 15

359

ATCACCCAGTTCCTCAAGTTCAAGCTGGGriTl'GCrGACCTGG

+ YGEIGIGTPPQKFTVVFDTGSSNLWVPSVH TATGGTGAGATTG«»TTGGGACCCXœatfyVGAAGTTCACTGTGG^

CHLLDIACLLHHKYDASKSSTYVE[n]GTEFA

45 449 75

539

450

TGTCACCTGCTAGACATCGCCTGTTTGCTACACCACAAGTATGATGCAT^^

76 540

NLKIKNQIFG

ATCCACTATGGGACTGGGAœCTCTCTGGATTCCTGAGCCAGGACACAGTCAC^^

105 629

106 630

EAVKQPGITFIAAKFDGILGMAFPRISVDK GAGGCTGTGAAGCAœCAGGCATCACCTTTATTGCTGOCAAGTTCGATCGC^

719

136 720

GTCACACCTTTCTTTGATAATGTCATXXAGCAGAAœTGATTGAGAAAW

166 810

CCAGGCGCTGATCTGCTGCTItt^X^CTGACCCCAAATACTA^

IHYGTGSLSGFLSQDTVT

VTPFFDNVMQQKLIEKNIFSFYLNRDPTAQ

PGGELLLGGTDPKYYSGDFSWVffTJVTRKAYW

+ QVHMDSVDVANGLTLCKGGCEAIVDTGTSL

196 900

CJkGGTCCACATGGACTCGGTGGATGrrGCCAATGGGCTGACTCTTTGaVAA^

226 990

ATCACTGGCCCCACCAAGGAAGTGAAGGAGCrGCAAACAGCCATTGGT^^

256 1080 286

1170 316

L^G

ITGPTKEVKELQTAIGAKPLIKGQYVISCD

KISSLPVVTLMLGGKPYQLTGEQYVFKVSA

135

165 809

195 899

225 989

255 1079

285

AAGATCTCGTCTCTC