Membrane Binding Properties and Terminal Residues of the Mature Hepatitis C Virus Capsid Protein in Insect Cells

Abstract

The immature core protein (p23, residues 1 to 191) of hepatitis C virus undergoes posttranslational modifications including intramembranous proteolysis within its C-terminal signal sequence by signal peptide peptidase to generate the mature form (p21). In this study, we analyzed the cleavage site and other amino acid modifications that occur on the core protein. To produce the posttranslationally modified core protein, we used a baculovirus-insect cell expression model system. As previously reported, p23 is processed to form p21 in insect as well as in mammalian cells. p21 was found to be associated with the cytoplasmic membrane, and its significant portion behaved as an integral membrane protein. The protein was purified from the membrane by a simple and unique procedure on the basis of its membrane-binding properties and solubility in detergents. Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) analysis of purified p21 showed that the average molecular mass (m/z 19,307) of its single-charged ion differs by m/z 1,457 from that calculated for p23. To determine the posttranslational modifications, tryptic p21 peptides were analyzed by MALDI-TOF MS. We found three peptides that did not match the theoretically derived peptides of p23. Analysis of these peptides by MALDI-TOF tandem MS revealed that they correspond to N-terminal peptides (residues 2 to 9 and 2 to 10) starting with α-N-acetylserine and C-terminal peptide (residues 150 to 177) ending with phenylalanine. These results suggest that the mature core protein (molecular mass of 19,306 Da) includes residues 2 to 177 and that its N terminus is blocked with an acetyl group.

Hepatitis C virus (HCV), a member of the Hepacivirus genus belonging to the Flaviviridae family, is a causative agent of chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma (reviewed in reference 29). HCV possesses a 9.6-kb positive-strand RNA genome, which contains a large open reading frame encoding a viral polyprotein (∼3,010 amino acids). The polyprotein precursor is produced on the endoplasmic reticulum (ER) and co- and posttranslationally processed by cellular and viral proteases into putative structural (core, E1, E2, and p7) and nonstructural (NS2, NS3, NS4A, NS4B, NS5A, and NS5B) proteins (29).

The amino (N)-terminal core protein binds to RNA (11, 41) and encapsidates the viral genome to form a viral nucleocapsid (23). Furthermore, it has been implicated in the pathogenesis of HCV (reviewed in references 14, 24, and 29). In a viral particle, the nucleocapsid is enclosed by a viral envelope composed of a cellular lipid bilayer and two envelope glycoproteins, E1 and E2, which may mediate receptor binding and cell entry (29).

The most carboxyl (C)-terminal part (residues 174 to 191) of the core region (residues 1 to 191) in the polyprotein precursor functions as a signal sequence, which translocates the following E1 region into the ER lumen (16, 39). An immature core protein (p23, residues 1 to 191) is produced by cleavage of the translocating polyprotein precursor with host signal peptidase (SPase) after the signal sequence (16). Then, p23 again undergoes proteolysis to generate the mature core protein (p21), which is estimated to be between 173 to 181 amino acids in length (18, 25, 39, 44). Recently, McLauchlan et al. (32) showed that p23 is cleaved within the signal sequence by signal peptide peptidase (SPPase) in mammalian cells as previously suggested by Hüssy et al. (18), who used a baculovirus-insect cell expression system. Since p21 is exclusively found in HCV particles from the sera of HCV-infected patients (44), the signal sequence cleavage is thought to be important for the formation of mature virus particles.

The basic N-terminal part (amino acid residues 1 to 75) and middle part (residues 82 to 102) of the core protein interact with the viral genome (11, 23, 39, 41) and itself (31, 36), respectively. The hydrophobic C-terminal part of the core protein is thought to participate in its association with cellular membranes (17, 39) and the E1 envelope glycoprotein (26). Although the core protein associates with the ER (15, 30, 34, 39), mitochondria (37), and lipid droplets (1, 32, 34) in the cytoplasm, its small portion is likely to be localized in the nucleus (8, 19, 25, 35, 44). However, the precise functions of core protein localized in different subcellular sites remain to be elucidated.

SPPase, a polytopic intramembrane-cleaving protease belonging to the family of aspartic proteases, processes some signal peptides derived from newly synthesized secretary or membrane proteins into biologically active signal peptide fragments (reviewed in reference 43). SPPase genes are highly conserved in higher eukaryotes such as human, mouse, Drosophila melanogaster, Caenorhabditis elegans, and Arabidopsis thaliana (42). The signal sequence cleavage of HCV core protein occurs in insect cells (Spodoptera frugiperda Sf9) (18) as well as mammalian cells (25, 32, 39, 44). However, the precise cleavage site has not been determined. Furthermore, only limited information is available on other amino acid modifications to the core protein. A fraction of the core protein is phosphorylated at serine residues 53 and 116 in insect cells (40) as well as mammalian cells (27). On the other hand, the core protein is a substrate for tissue transglutaminase, which produces a covalently linked core protein dimer (28).

The baculovirus expression system is a most powerful eukaryotic system used to express recombinant proteins in insect cells such as Sf9 cells. Many proteins are posttranslationally modified in Sf9 cells in a manner similar to that of mammalian cells (20). In Sf9 cells, the HCV core, E1, and E2 structural proteins are produced from the core-E1-E2 polyprotein and assembled into virus-like particles (2). Moreover, the immature core protein is processed into its mature form in Sf9 cells (18). Thus, we chose this system as a model system to study the biochemical properties and posttranslational modifications of the core protein. When the core protein was expressed in its immature form (p23) in Sf9 cells, it was processed into the mature form (p21) found on the cytoplasmic membrane. We developed a novel method for purifying p21 from the membrane fraction of baculovirus-infected Sf9 cells. By direct mass spectral analyses of purified p21, we found that it begins with α-N-acetylserine (residue 2) and ends with phenylalanine (residue 177). These results indicate that the core protein undergoes at least three posttranslational modifications in insect cells: removal of the initiator methionine, N-terminal acetylation of the penultimate serine, and cleavage within the signal sequence.

MATERIALS AND METHODS

Generation of recombinant baculovirus expressing HCV core protein.

Recombinant baculovirus expressing p23 (residues 1 to 191) of HCV genotype 1b (HCR6 strain) was generated by using the Bac-to-Bac baculovirus expression system (Invitrogen). The core gene (nucleotides 342 to 914 followed by a stop codon) derived from HCV cDNA (pHCR6; GenBank accession no. AY045702) (44) was subcloned into the baculovirus expression vector pFastBac1 (Invitrogen) to generate pFastBac-core. The pFastBac-core was transformed into Escherichia coli strain DH10Bac (Invitrogen) to produce a recombinant bacmid (bMON-core) according to the manufacturer's protocol (Invitrogen). Sf9 insect cells were then transfected with the bMON-core by using DOTAP liposomal transfection reagent (Roche Molecular Biochemicals) and cultured in EX-CELL 400 medium (JRH Biosciences) at 27°C. The recombinant baculovirus expressing p23 (rAcNPV-core) was harvested at 3 days posttransfection and amplified by passaging once.

Subcellular fractionation of Sf9 cells expressing the HCV core protein.

Sf9 insect cells (1.4 × 10⁸ cells) were infected with the rAcNPV-core at a multiplicity of infection of 1 PFU per cell and cultured in a spinner flask containing 50 ml of EX-CELL 400 medium at 27°C. At 48 h postinfection, the cells were harvested and washed twice with ice-cold phosphate-buffered saline. They were then suspended in 5 packed-cell volumes of ice-cold hypotonic buffer (10 mM Tris-HCl [pH 7.5 at 20°C], 10 mM KCl, 1 mM MgCl₂, 0.5 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl fluoride [PMSF]) and recovered by centrifugation. Then, the cells were resuspended in 2 packed-cell volumes of hypotonic buffer and swollen on ice for 5 min. The swollen cells were homogenized in a Dounce homogenizer (Wheaton) with 20 strokes of a tight pestle. The homogenate (5.5 ml) was mixed with a 1/10 volume of 2.5 M sucrose, layered on a 2-ml sucrose cushion (10 mM Tris-HCl [pH 7.5], 10 mM KCl, 1 mM MgCl₂, 0.5 mM DTT, 0.34 M sucrose), and centrifuged in a swinging bucket rotor at 500 × g_av for 10 min at 4°C. Following centrifugation, the postnuclear supernatant (PNS) (4.8 mg of protein/ml; 6.8 ml) was recovered. The PNS was separated into a supernatant (S10 fraction) and pellet by centrifugation at 10,000 × g_av for 10 min at 4°C. The pellet was resuspended in 3 ml of isotonic buffer (10 mM Tris-HCl [pH 7.5], 10 mM KCl, 1 mM MgCl₂, 0.5 mM DTT, 0.25 M sucrose) and centrifuged at 10,000 × g_av for 5 min at 4°C. The resultant supernatant recovered was combined with the supernatant (S10 fraction) from the previous centrifugation. The pellet was then resuspended in isotonic buffer and referred to as the P10 fraction (5.0 mg of protein/ml; 2 ml). The S10 fraction was further separated into a supernatant (S100 fraction; 1.6 mg of protein/ml; 9.5 ml) and pellet by centrifugation in a Beckman Type 70.1 Ti rotor at 100,000 × g_av for 1 h at 4°C. The pellet was resuspended in isotonic buffer and referred to as the P100 fraction (3.1 mg of protein/ml; 2 ml).

As described above, uninfected Sf9 cell PNS was also prepared and subjected to subcellular fractionation.

Subcellular fractionation of HeLa cells expressing truncated HCV polyprotein.

Recombinant vaccinia virus LO-R6J20 expressing truncated HCV polyprotein (core-E1-E2-p7-NS2; residues 1 to 1032) under the control of the T7 promoter was previously described by Yasui et al. (44). HeLa cells were simultaneously infected with recombinant vaccinia viruses LO-R6J20 and LO-T7-1 expressing the T7 RNA polymerase or mock infected and cultured for 12 h at 37°C as previously described (44). The PNS was prepared from the homogenate of the infected or mock-infected cells and subjected to subcellular fractionation to obtain the P10, P100, and S100 fractions as essentially described above.

Analysis of the membrane-binding properties of the core protein.

The P10 membrane fraction (25 μg of protein) from Sf9 cells infected with the rAcNPV-core was incubated for 30 min on ice in 20 μl of TNED buffer (20 mM Tris-HCl [pH 8.0 at 20°C], 150 mM NaCl, 1 mM EDTA, 0.5 mM DTT) containing 1 mM PMSF in the presence or absence of additional EDTA (final concentration, 50 mM), additional NaCl (final concentration, 1 M), 1% (wt/vol) Triton X-100, 1% sodium deoxycholate, 60 mM octyl-β-glucoside, or 0.25% sodium N-lauroylsarcosine (sarcosyl). Similarly, the P10 membrane fraction was treated with sodium carbonate (100 mM Na₂CO₃, pH 11.5). Then, the mixtures were separated into supernatants (S15 fraction) and pellets (P15 fraction) by centrifugation at 15,000 × g_av for 10 min at 4°C. The S15 fractions were further separated into supernatants (S100 fraction) and pellets (P100 fraction) by centrifugation in a Beckman TLA-100.3 rotor with adapters for microcentrifuge tubes at 100,000 × g_av for 10 min at 4°C.

Equilibrium flotation centrifugation with membranes.

The P10 membrane fraction (25 μl; 125 μg of protein) was mixed with 5 volumes of 2.5 M sucrose in TNED buffer to give a final concentration of 2.1 M and placed on the bottom of a 0.8-ml ultraclear tube (Beckman). The sucrose mixture was overlaid with 300 μl of 1.9 M sucrose and 150 μl of 0.25 M sucrose in TNED buffer and centrifuged at 100,000 × g_av for 16 h at 4°C in a Beckman SW 55 Ti rotor with adapters. After centrifugation, 6 fractions (100 μl each) were collected from the top of the gradient, and the pellet was resuspended in 100 μl of isotonic buffer.

The P10 membrane fraction (75 μg of protein) was treated under various conditions in 60 μl of the mixture for 30 min on ice as described above. The mixture was combined with 1.5 volumes of 2.4 M sucrose in TNED or sodium carbonate to give a final concentration of 1.44 M and placed on the bottom of the 0.8-ml tube. The sucrose mixture was overlaid with 300 μl of 1.25 M sucrose and 150 μl of 0.25 M sucrose in TNED or sodium carbonate buffer. As described above, the step gradients were centrifuged and the fractions were collected.

Purification of p21 from the P10 membrane fraction.

All purification steps were carried out at 4°C or on ice. The P10 membrane fraction (3 mg of protein) was incubated in 1.2 ml of sodium carbonate for 30 min and mixed with 1.8 ml of 2.4 M sucrose in sodium carbonate. The resulting 1.44 M sucrose mixture was placed on the bottom of a tube and overlaid with 5 ml of 1.25 M sucrose and 2 ml of 0.25 M sucrose in sodium carbonate. The discontinuous gradient was centrifuged in a Beckman SW 41 Ti rotor at 100,000 × g_av for 16 h, and fractions were collected from the top. Fractions around the 1.25 M-0.25 M sucrose interface were pooled, diluted with 5 volumes of TNED, and centrifuged at 100,000 ×g_av for 1 h to precipitate the membranes. The pelleted membranes were resuspended in 0.55 ml of isotonic buffer and incubated for 5 min with several short sonications in a bath sonicator (UCD-200T Bioruptor; Cosmobio). The membrane suspension was referred to as the low-density membrane (LDM) fraction (1.9 mg of protein/ml; 0.57 ml). The LDM fraction (1 mg of protein) was incubated in 1 ml of TNED buffer containing 1% Triton X-100 and 1 mM PMSF for 30 min and centrifuged in a TLA-100.3 rotor with adapters at 100,000 × g_av for 30 min. The resulting pellet (TXP100-1 fraction) was suspended in 0.53 ml of TNEDG buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 0.5 mM DTT, 20% glycerol) containing 0.25% sarcosyl and sonicated as described above. Following centrifugation at 100,000 × g_av for 30 min, the resultant supernatant was recovered and referred to as the SRS100-1 fraction (0.77 mg of protein/ml; 0.53 ml). The SRS100-1 fraction (0.39 mg of protein; 0.5 ml) was diluted with 10 volumes of TNEDG-1% Triton X-100 to form aggregates of p21 and centrifuged in a TLA-100.3 rotor at 100,000 × g_av for 40 min. The resulting pellet (TXP100-2 fraction) was dissolved in 0.25 ml of TNEDG-0.25% sarcosyl and centrifuged again in a TLA-100.3 rotor with the adapters at 100,000 × g_av for 20 min. The resultant supernatant was referred to as the SRS100-2 fraction (0.04 mg of protein/ml; 0.25 ml).

Gel filtration column chromatography.

The SRS100-2 fraction (100 μl) was applied to a Sephacryl S-300HR column (inner diameter, 0.6 by 18.5 cm; Amersham Biosciences) that had been previously equilibrated with TNEMG buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 5 mM 2-mercaptoethanol, 20% glycerol) containing 0.25% sarcosyl, and the proteins were eluted with the same buffer at 2.7 ml/h. Fractions (∼0.1-ml) were collected, and aliquots were analyzed by Western blotting with anticore antibody. Protein standards used to calibrate the column were cytochrome c (12.4 kDa), chymotrypsinogen A (25 kDa), ovalbumin (45 kDa), bovine serum albumin (66 kDa), catalase (232 kDa), ferritin (440 kDa), and thyroglobulin (669 kDa). The exclusion limit was evaluated with dextran blue 2000.

Sucrose density gradient centrifugation.

The SRS100-2 fraction (20 μl) was incubated in 200 μl of TNED buffer containing 1% Triton X-100 on ice for 5 min. The mixture was overlaid onto a 4.8-ml 0.4 to 1.0 M sucrose gradient in TNED-0.1% Triton X-100 and then centrifuged at 150,000 × g_av for 2 h at 4°C in a SW 55 Ti rotor. After centrifugation, 20 fractions (250 μl each) were collected from the top of the gradient and a pellet was recovered.

To determine the positions of particles with sedimentation coefficients of 80, 60, and 40S in the 0.4 to 1.0 M sucrose gradient, 80S ribosome and 60/40S ribosomal subunits were prepared from the P100 microsomal fraction of HeLa cells. To dissociate the 80S ribosome from the microsomal membrane, the P100 fraction (0.13 mg of protein) was incubated in buffer A (20 mM Tris-HCl [pH 8.0], 70 mM KCl, 5 mM MgCl₂) containing 0.3% sodium deoxycholate. To prepare ribosomal subunits from the membrane-associated ribosome, the P100 fraction was incubated in buffer B (20 mM Tris-HCl [pH 8.0], 300 mM KCl, 3 mM MgCl₂) containing 0.3% sodium deoxycholate. The mixtures were layered on the 0.4 to 1.0 M sucrose gradient in buffer A (for ribosome) or B (for ribosomal subunits) and then centrifuged as described above. Twenty fractions were collected, and the 80S ribosome and 60/40S ribosomal subunits in the fractions were detected at a wavelength of 280 nm.

Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) analyses. (i) Determination of the monomeric molecular mass of p21.

Twenty microliters of the SRS100-2 fraction was diluted with 9 volumes of TNEDG-1% Triton X-100 buffer and centrifuged at 100,000 × g_av for 20 min at 4°C. The resultant pellet (TXP100-3 fraction) was washed with 200 μl of water and then dissolved in 2.5 μl of 0.1% trifluoroacetic acid (TFA)-60% acetonitrile. The sample (0.5 μl) was applied on a stainless steel MALDI target plate (Applied Biosystems), overlaid with 0.5 μl of a saturated matrix solution of sinapinic acid (Sigma) in 0.1% TFA-30% acetonitrile, and dried at room temperature. The average molecular mass of a single-charged p21 ion was determined by using the Voyager-DE PRO MALDI mass spectrometer (Applied Biosystems) equipped with a nitrogen laser (337 nm). Mass spectra were acquired in the linear positive-ion mode by using an accelerating voltage of 25,000 V, grid voltage of 93%, guide wire of 0.05%, and extraction delay time of 425 ns. Calibration mixture 3 (bovine insulin, E. coli thioredoxin, and horse apomyoglobin) of the Sequazyme Peptide Mass Standards Kit (Applied Biosystems) was used for external and internal standards. Spectra were obtained by accumulating at least 300 consecutive laser shots. Data were analyzed by using Data Explorer software version 4.3 (Applied Biosystems). The average molecular masses of single-charged ions for predicted p21 and p23 were calculated by using the Gpmaw program, version 4.0 (Lighthouse Data).

(ii) Peptide fingerprint analysis of tryptic peptides of p21.

In situ alkylation of p21 with acrylamide was performed as essentially described by Mineki et al. (33). The TXP100-1 fraction prepared as described above was dissolved in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer for in situ alkylation (50 mM Tris-HCl [pH 8.5], 2% SDS, 10 mM DTT, 0.005% bromophenol blue, 20% glycerol) and heated at 100°C for 3 min. After adding one-fifth volume of 35% acrylamide to the denatured sample, it was immediately resolved by electrophoresis in a SDS-12.5% polyacrylamide gel. The p21 band (∼0.5 μg of protein) was excised from the gel stained with Coomassie brilliant blue. The excised gel was chopped into pieces, rinsed with water, and treated several times with a destaining solution (50 mM ammonium bicarbonate, 50% methanol) at 40°C. The gel pieces were scraped with a pestle and dried in a centrifugal evaporator. Afterward, they were rehydrated with 8 μl of digestion solution (50 mM ammonium bicarbonate, 50% acetonitrile) containing modified trypsin (sequencing grade; 0.05 pmol/μl) (Roche Molecular Biochemicals) at 4°C for 10 min. Following the addition of 8 μl of digestion solution to the mixture, it was incubated overnight at 37°C. The resulting peptides were extracted from the gel by successive extractions with 0.1% TFA, 0.1% TFA-50% acetonitrile, and 0.1% TFA-80% acetonitrile. The peptide mixture was filtered through a 0.22-μm-pore-size Durapore membrane by using an Ultrafree-MC microcentrifuge tube (Millipore), dried, and dissolved in 20 μl of 0.1% TFA-30% acetonitrile. An aliquot (0.5 μl) was spotted onto the MALDI target plate and overlaid with a saturated solution (0.5 μl) of α-cyano-4-hydroxycinnamic acid (Sigma) in 0.1% TFA-50% acetonitrile. The Voyager-DE PRO MALDI mass spectrometer was employed for peptide mass fingerprinting in the reflector positive-ion mode by using an accelerating voltage of 20,000 V, grid voltage of 76%, mirror voltage ration of 1.12, guide wire of 0.002%, and extraction delay time of 225 ns. Calibration mixture 2, consisting of angiotensin I, adrenocorticotropin (ACTH; clip 1 to 17), ACTH (clip 18 to 39), ACTH (clip 7 to 38), and bovine insulin (Applied Biosystems), was used as an external standard. Peptide fingerprint searches were performed by using the FindPept tool program available at the ExPASy Molecular Biology Server (www.expasy.org).

(iii) MS/MS analysis of selected tryptic peptides.

The tryptic peptides of p21 were extracted from the polyacrylamide gel as described above and concentrated to approximately 20 μl in a centrifugal evaporator. The peptide mixture was separated on the PepMap C18 (3-μm particle size; 100-Å pore diameter) capillary column (inner diameter, 75 μm by 15 cm; LC Packings) by using the UltiMate Capillary LC System (LC-Packings). The sample (10 μl) was loaded onto the column preequilibrated with 90% solvent A (0.1% TFA-5% acetonitrile) and 10% solvent B (0.1% TFA-90% acetonitrile), and the column was washed with the same solvent mixture. Peptides were eluted with a linear gradient from 90% solvent A and 10% solvent B to 40% A and 60% B in 33 min at 0.3 μl/min and detected at a wavelength of 214 nm. The column fractions (0.1 μl each) were automatically collected, mixed with 8 mg of α-cyano-4-hydroxycinnamic acid per ml in 0.1% TFA-60% acetonitrile, and spotted on a MALDI target plate by the Probot Micro Fraction Collector (LC-Packings). MALDI-TOF tandem mass spectrometry (MS/MS) analysis of selected peptides was performed by using the Applied Biosystems 4700 Proteomics Analyzer equipped with a 200 Hz Nd:YAG laser (355 nm) and TOF/TOF optics. MS/MS spectra were obtained at a collision energy of 1 kV without the presence of gas. To probe their amino acid sequences and amino acid modifications, the resultant MS/MS spectra were analyzed by using the SeqMS program (12).

SDS-PAGE and Western blotting.

SDS-PAGE and Western blotting with mouse anticore monoclonal antibody (515S) were carried out as essentially described by Yasui et al. (44). The band intensities of the Western blots were measured by using the National Institutes of Health image program (version 1.62), and relative amounts of p21 in the fractions were estimated (see Fig. 4).

FIG. 4.

p21 forms high-molecular-weight complexes. (A) p21 (SRS100-2 fraction) (Fig. 3) was applied to a gel filtration column and eluted with TNEDG buffer containing 0.25% sarcosyl. The relative amounts of p21 in the fractions are shown. The positions of marker proteins are shown on top. (B) p21 (SRS100-2) was incubated in TNED buffer containing 1% Triton X-100 and then subjected to 0.4 to 1 M sucrose density gradient centrifugation. The relative amounts of p21 in the fractions are shown. The positions of the 80S ribosome and 40 and 60S ribosomal subunits are shown on top.

Protein Concentration.

Protein concentrations were determined by the methods of Bradford (5) by using bovine serum albumin as a standard.

RESULTS

Membrane-binding properties of HCV core protein.

To purify and characterize posttranslationally modified HCV core protein, recombinant p23 (residues 1 to 191) was expressed in Sf9 insect cells by using a baculovirus expression system. A PNS was prepared from a homogenate of Sf9 cells infected with recombinant baculovirus expressing p23 (rAcNPV-core) and analyzed by SDS-PAGE (Fig. 1A) and Western blotting with monoclonal anticore antibody (Fig. 1B). A large quantity of core protein was detected as a polypeptide with an apparent molecular mass of 21,000 Da (p21) in the PNS (lanes 2), suggesting that p23 is processed into p21 in insect cells. To investigate the cytoplasmic distribution of p21 in the cells, the PNS was separated into membrane fractions sedimenting at 10,000 × g (P10) and 100,000 × g (P100) and a cytosolic fraction (S100). These fractions were then analyzed by SDS-PAGE (Fig. 1A) and Western blotting (Fig. 1B). p21 was detected in the P10 (∼70%) (lanes 3) and P100 (∼30%) (lanes 4) membrane fractions, but not in the S100 fraction (lanes 5), suggesting that p21 associates with the cytoplasmic membrane. This subcellular distribution was similar to that of p21 in mammalian cells (34, 44). Furthermore, as shown in Fig. 1C, the molecular mass of p21 in the P10 fraction from Sf9 cells (lane 3) was identical to that of p21 (lane 2), which was expressed as a core-E1-E2-p7-NS2 polyprotein (residues 1 to 1032) in HeLa cells when a vaccinia virus expression system was used (44). To confirm the membrane association of p21, the P10 fraction was subjected to membrane flotation in a discontinuous sucrose density gradient to separate membranes from insoluble aggregates (Fig. 1D). The P10 fraction was adjusted to 2.1 M sucrose and placed on the bottom of a discontinuous density gradient consisting of 2.1, 1.9, and 0.25 M sucrose. After centrifugation to equilibrium, fractions were collected from the top of the gradient and analyzed by Western blotting with anticore antibody (Fig. 1D, left). The densities (grams per cubic centimeter) of the fractions were confirmed by measuring their refractive indexes (right). The majority of p21 in the P10 fraction was localized in the 1.9 M-0.25 M sucrose interface (fraction 2), indicating that p21 associates with intracellular membrane-bound organelle(s) in insect cells. Since the majority of p21 was found in the P10 membrane fraction, it was used for further analyses.

FIG. 1.

p21 is associated with cytoplasmic membranous organelle(s) in insect cells. (A) Sf9 cells were infected with recombinant baculovirus (rAcNPV-core) to express the HCV core protein (p23, residues 1 to 191). PNS was prepared from the homogenate of infected cells and fractionated into P10, P100, and S100 fractions as described in Materials and Methods. PNS (16 μg of protein; lane 1) from uninfected cells and PNS (16 μg; lane 2), P10 (5 μg; lane 3), P100 (3 μg; lane 4), and S100 (8 μg; lane 5) fractions from infected cells were analyzed by SDS-12.5% PAGE, followed by staining with Coomassie brilliant blue. The positions of marker proteins are shown on the left. (B) One-fifth the amount of the same samples shown in panel A was analyzed by Western blotting with anticore monoclonal antibody. The positions of marker proteins and p21 are shown on the left and right, respectively. (C) HeLa cells were mock infected or infected with recombinant vaccinia viruses LO-R6J20 and LO-T7-1 (J20/T7) to express a core-E1-E2-p7-NS2 polyprotein (residues 1 to 1032). P10 membrane fractions were prepared from the cells. The P10 fractions from mock-infected HeLa cells (10 μg; lane 1), J20/T7-infected HeLa cells (10 μg; lane 2), and rAcNPV-core-infected Sf9 cells (1 μg; lane 3) were immunoblotted with anticore monoclonal antibody. (D) The P10 fraction from rAcNPV-core-infected Sf9 cells was subjected to membrane flotation in a discontinuous sucrose density gradient consisting of 2.1, 1.9, and 0.25 M sucrose. After centrifugation, fractions were collected from the top of the gradient and analyzed by Western blotting with anticore monoclonal antibody (left panel). The densities (grams per cubic centimeter) of the fractions were determined by refractometry (right panel).

To characterize the membrane-binding properties of p21, the P10 membrane fraction was treated under various conditions and then separated into fractions sedimenting at 15,000 × g (P15) and 100,000 × g (P100) and a soluble fraction (S100) (Fig. 2A). When the P10 fraction (lane 1) was preincubated in TNED buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 0.5 mM DTT) (Fig. 2A, lanes 2 to 4), p21 was recovered in the P15 membrane fraction (lane 2). When EDTA (lanes 5 to 7) and NaCl (lanes 8 to 10) were further added to the reactions to final concentrations of 50 mM and 1 M, respectively, p21 was not released from the P15 fraction under either set of conditions (lanes 5 and 8). This suggests that divalent cation-mediating and ionic interactions are not involved in p21 association with the membrane and/or other membrane proteins. Even when the P10 fraction was treated with sodium carbonate (pH 11.5) (Fig. 2A, lanes 11 to 13) to release luminal contents and peripheral proteins from the membranes (13), the majority of p21 remained with the P15 fraction (lane 11). Thus, p21 is tightly associated with the membrane as an integral membrane protein and/or released as aggregates from the membrane. To assess its solubility in various detergents, the P10 membrane fraction was incubated in TNED buffer containing 1% Triton X-100 (Fig. 2A, lanes 14 to 16), 1% deoxycholate (lanes 17 to 19), or 60 mM octylglucoside (lanes 20 to 22). However, p21 in the P10 fraction was predominantly insoluble in these detergents and was recovered in the P15 fraction (lanes 14, 17, and 20). In contrast, when sarcosyl, a strong anionic detergent, was added to the mixture at a final concentration of 0.25% (lanes 23 to 25), almost all of p21 was found in the S100 soluble fraction (lane 25).

FIG. 2.

Membrane-binding properties of p21 associated with the P10 membranes. (A) The P10 fraction (lane 1) from the rAcNPV-core-infected Sf9 cells was incubated in TNED buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 0.5 mM DTT) in the presence (lanes 5 to 10 and 14 to 25) or absence (lanes 2 to 4) of the indicated additions or in sodium carbonate (lanes 11 to 13) and then separated into P15 (lanes 2, 5, 8, 11, 14, 17, 20, and 23), P100 (lanes 3, 6, 9, 12, 15, 18, 21, and 24), and S100 (lanes 4, 7, 10, 13, 16, 19, 22, and 25) fractions by centrifugation. The fractions were analyzed by Western blotting with anticore monoclonal antibody. (B to E) Following treatment of the P10 fraction as indicated, it was subjected to membrane flotation in a discontinuous sucrose density gradient consisting of 1.44, 1.25, and 0.25 M sucrose. The fractions were collected and analyzed by SDS-12.5% PAGE (upper panels) and Western blotting with anticore antibody (middle panels). The positions of marker proteins and p21 are shown on the left and right, respectively. A core protein that slowly migrated is indicated by an asterisk. The densities (grams per cubic centimeter) of the fractions are shown in the lower panels.

To investigate whether p21 remains associated with the membrane after treatment of the P10 membrane fraction under the selected conditions, the sucrose concentration was adjusted to 1.44 M, and it was subjected to membrane flotation in a discontinuous sucrose density gradient consisting of 1.44, 1.25, and 0.25 M sucrose (Fig. 2B to E). The fractions were collected from the top of the gradient and analyzed by SDS-PAGE (top panels) and Western blotting with anticore antibody (middle panels). The densities (grams per cubic centimeter) of the fractions are shown in the bottom panels. For the density gradient consisting of 1.44, 1.25, and 0.25 M sucrose, a large quantity of p21 in the P10 fraction preincubated in TNED buffer (Fig. 2B) precipitated with heavy membranes (fraction 7), whereas some p21 floated with light membranes (fractions 3 to 5). When the P10 fraction was treated with sodium carbonate (Fig. 2C), 60% of p21 floated with the LDMs around the 1.25 M-0.25 M sucrose interface (fractions 2 and 3). Fifteen percent of p21 was detected in bottom fractions 5 and 6, which corresponded to the positions of the solubilized proteins. In addition, residual p21 was detected in pellet fraction 7 together with a slowly migrating core protein (indicated by the asterisk in the middle panel), which was thought to be immature core protein (p23). These findings indicate that a significant portion of p21 in the P10 fraction behaved as an integral membrane protein. On the other hand, after treatment of the P10 membrane with 1% Triton X-100 (Fig. 2D), the majority of p21 was recovered in pellet fraction 7, suggesting that p21 is present as aggregates and/or has a tendency to aggregate even if released from the membrane. For sarcosyl treatment (Fig. 2E), p21 was recovered in bottom fractions 5 and 6, indicating that it was completely solubilized from the membrane.

Purification of p21 from the P10 membrane fraction.

On the basis of the membrane-binding properties of p21 and its solubility in detergents, we established a method to purify p21 from the P10 membrane fraction (Fig. 3A). During purification, fractions at respective steps were analyzed by SDS-PAGE (Fig. 3B) and Western blotting with anticore antibody (Fig. 3C). First, following treatment of the P10 membrane fraction (lanes 2) with sodium carbonate, it was subjected to membrane flotation by using a discontinuous gradient consisting of 1.44, 1.25, and 0.25 M sucrose, and p21 was recovered in the LDM fraction (lanes 3). The LDM fraction was incubated in TNED buffer containing 1% Triton X-100, and Triton-insoluble materials were then precipitated by ultracentrifugation. The resulting pellet (TXP100-1) was suspended in TNEDG buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 0.5 mM DTT, 20% glycerol) containing 0.25% sarcosyl. After ultracentrifugation, p21 was collected in the sarcosyl-soluble supernatant fraction (SRS100-1) (lanes 4). p21 in the SRS100-1 fraction easily aggregated with the decreasing sarcosyl concentration. When the SRS100-1 fraction was diluted with approximately 10 volumes of TNEDG buffer containing 1% Triton X-100 which had been included to keep other cellular membrane proteins soluble, p21 formed aggregates and could be recovered in the pellet fraction (TXP100-2) by ultracentrifugation. The pellet was resuspended in TNEDG buffer containing 0.25% sarcosyl and then ultracentrifuged. Finally, the supernatant (SRS100-2) containing p21 as the major polypeptide was collected (lanes 5). Two cycles of aggregation and solubilization of p21 allowed us to recover it at about 70% purity.

FIG. 3.

Purification of p21 from the P10 fraction. (A) Purification scheme for p21. The P10 fraction from rAcNPV-core-infected Sf9 cells was treated with sodium carbonate (pH 11.5) to release the luminal contents and peripheral proteins from the membranes. Following membrane flotation in the discontinuous sucrose density gradient, p21 was recovered in an LDM fraction (the 1.25 M-0.25 M sucrose interface). Based on the solubility of p21 in TNED (Fig. 2) or TNEDG (TNED plus 20% glycerol) buffer containing 1% Triton X-100 or 0.25% sarcosyl, the LDM fractions were further fractionated to obtain the Triton-insoluble (TXP100) and sarcosyl-soluble (SRS100) fractions as described in Materials and Methods. (B) The P10 fraction (10 μg; lane 1) from uninfected Sf9 cells and the P10 (10 μg; lane 2), LDM (5.7 μg; lane 3), SRS100-1 (3.5 μg; lane 4), and SRS100-2 (0.25 μg; lane 5) fractions from rAcNPV-core-infected cells were analyzed by SDS-12.5% PAGE. The positions of marker proteins and p21 are shown on the left and right, respectively. (C) One-eighth the amount of the same samples shown in panel B was analyzed by Western blotting with anticore monoclonal antibody.

To investigate the molecular mass of p21 in the SRS100-2 fraction, we subjected it to gel filtration chromatography on a Sephacryl S-300 column in the presence of 0.25% sarcosyl (Fig. 4A). p21 was eluted as a single peak at a molecular weight of approximately 150,000, suggesting that it occurs as an oligomer. As described above, p21 solubilized with sarcosyl formed aggregates when diluted 9- to 10-fold with TNED buffer containing 1% Triton X-100. To estimate the sedimentation coefficient of the p21 aggregates, they were subjected to 0.4 to 1.0 M sucrose density gradient centrifugation as described in Materials and Methods. As shown in Fig. 4B, they sedimented as a broad peak throughout fractions 6 to 15 corresponding to 40 to 130S, suggesting that several kinds of complexes with different sizes (e.g., 70 and 100S) were present. These data indicate that soluble p21 forms high-molecular-weight complexes and that these complexes are heterogeneous in size.

Posttranslational modifications of the core protein.

To estimate the monomeric molecular mass of p21, the protein in its purified form was analyzed by MALDI-TOF MS in linear mode (Fig. 5). To increase the purity of p21, the SRS100-2 fraction was subjected to p21 aggregate formation again, and the aggregates were recovered as a pellet (TXP100-3) as described in Materials and Methods. To determine the purity of the p21 pellet, it was solubilized in SDS-PAGE sample buffer and analyzed by SDS-PAGE. As shown in Fig. 5A, p21 was present at over 90% purity. For MALDI-TOF MS analysis, the pellet was washed with water and dissolved in 0.1% TFA-60% acetonitrile. Analysis of purified p21 by MALDI-TOF MS revealed that the average mass of a single-charged ion ([M+H]⁺) was m/z 19,307.1 (Fig. 5B). The observed average mass of the single-charged p21 ion differed from that calculated for p23 (m/z 20,764.0; residues 1 to 191). Therefore, posttranslational modifications including C-terminal signal sequence cleavage with SPPase are believed to be responsible for the difference in molecular masses. However, we cannot explain such differences only by cleavage within the signal sequence.

FIG. 5.

Determination of the monomeric molecular mass of p21. (A) p21 in the TXP100-3 fraction, which was prepared from the SRS100-2 (Fig. 3) fraction as described in Materials and Methods, was analyzed by SDS-12.5% PAGE. (B) p21 (TXP100-3) was cocrystallized in a matrix of sinapinic acid and analyzed by MALDI-TOF MS in the linear positive-ion mode. The average masses of single- and double-charged p21 ions ([M+H]⁺ and [M + 2H]²⁺, respectively) are shown above the peaks.

To identify modifications to the core protein, we analyzed tryptic p21 peptides (Fig. 6). Partially purified protein (TXP100-1) was subjected to in situ alkylation with acrylamide followed by SDS-PAGE. After in-gel digestion with trypsin, the tryptic fragments were analyzed by MALDI-TOF MS to produce peptide mass fingerprint data (Fig. 6A). To detect posttranslationally modified peptides, the identities of the peptide ions ([M+H]⁺) were assigned by matching measured monoisotopic mass values (m/z) to those calculated for the predicted tryptic p23 fragments. As shown in Fig. 6A and Table 1, 13 peptides derived from p21 matched with theoretically derived peptides of p23. Furthermore, three intense peptide ions with m/z values of 969.58 (peptide 5), 1097.62 (peptide 7), and 2964.51 (peptide 14) were observed, but their masses did not match those of p23. Other weak signals were probably the result of contaminants.

FIG. 6.

MALDI-TOF MS analyses of p21 tryptic peptides. (A) p21 (TXP100-1 fraction) (Fig. 3) was subjected to in situ alkylation with acrylamide followed by SDS-PAGE. After in-gel digestion of p21 with trypsin, the tryptic fragments were extracted and cocrystallized in a matrix of α-cyano-4-hydroxycinnamic acid and analyzed by MALDI-TOF MS in reflectron positive-ion mode. The signals of the analyzed peptides are numbered. Peptides unmatched with theoretically derived values for those of p23 are shown in bold. The monoisotopic masses of their single-charged ions ([M+H]⁺) are shown in Table 1. (B and C) Peptides 5 (m/z 969.5) and 14 (m/z 2964.5) shown in panel A were purified by reverse-phase capillary liquid chromatography and sequenced by MALDI-TOF MS/MS. The MS/MS spectra were interpreted by the SeqMS program. The parent ions are shown by [M+H]⁺. Deduced fragment ions are indicated above the signals. In panel C, the MS/MS spectra are given as differently expanded intensity axes (upper and lower panels). The arrows indicate the sequence elucidated from the b- and y-ion series, which are shown in boldface. Pam-C indicates propionamide cysteine.

TABLE 1.

Tryptic peptides of p21 identified by MALDI-TOF MS analyses^a

No.	Mass (m/z)		Position (residue)		Modification
	Observed	Calculated	Start	End
1	738.45	738.46	62	67
2	772.46	772.48	44	50
3	813.43	813.43	68	74
4	951.58	951.59	60	67
5b	969.51	969.51	2	9	α-N-acetylserine
6	1088.59	1088.58	51	59
7b	1097.62	1097.61	2	10	α-N-acetylserine
8	1312.63	1312.64	102	113
9	1468.74	1468.74	102	114
10	1624.85	1624.84	102	115
11	1629.93	1629.91	24	39
12	2353.34	2353.31	18	39
13	2838.58	2838.55	14	39
14b	2964.51	2964.47	150	177	Propionamide cysteined
15c	2994.70	2994.65	13 (14)	39 (40)
16	3076.50	3076.46	75	101	Methionine sulfoxided

^aThe identity of the peptides was assigned by matching the measured monoisotopic mass values of single-charged ions ([M+H]⁺) (Fig. 6A) to those calculated for the predicted p21 tryptic fragments (Fig. 7).

^bThe amino acid sequences of selected peptides were confirmed by MALDI-TOF MS/MS analysis (Fig. 6B and C).

^cPeptide 15 corresponds to residues 13 to 39 or 14 to 40.

^dArtificial modifications.

To identify peptides 5, 7, and 14, they were purified from the tryptic peptide mixture by C18 capillary column chromatography and analyzed by MALDI-TOF MS/MS. Peptide 5 was coeluted with the peptide 7 from the column with the retention time of 6 min (data not shown). The single-charged ion of peptide 5 at m/z 969.5 was subjected to MS/MS analysis, and the resultant fragment ions were assigned to predicted N-terminal and C-terminal ion series (21) (Fig. 6B). Then, the amino acid composition of peptide 5 was deduced mainly from the b- and y-ion series. The results indicated that peptide 5 contains α-N-acetylserine at its N terminus and that its sequence is S-T-N-P-(K/Q)-P-(K/Q)-R. Although the MS/MS spectrum of peptide 5 with Glu at the N terminus must give a very similar pattern to that shown in Fig. 6B, the presence of the y8 ion indicated that the N-terminal structure was N-acetyl Ser. The presence of ions b5-17, y3-17, y5-17, and y6-17, which lost ammonia from the R, K, or Q residue, ions b1-18 and b3-18, which lost water from the S or T residue, internal fragment ions [(K/Q)+P and N+P+(K/Q)], and immonium ions (P, N, and R) supports the validity of the above sequence. Since the sequence was found at the N terminus (residues 2 to 9) of the core protein, we concluded that its sequence is S-T-N-P-K-P-Q-R and that its N terminus is blocked with an acetyl group. The observed monoisotopic mass (m/z 969.58) completely matched the calculated mass (m/z 969.59) of protonated acetyl-S-T-N-P-K-P-Q-R (Table 1). In addition, peptide 7 was identified as acetyl-S-T-N-P-K-P-Q-R-K (residues 2 to 10), which was not cleaved by trypsin between R and K residues (data not shown).

Peptide 14 (m/z 2964.5) was eluted as a single peak from the column with the retention time of 27 min (data not shown) and was subsequently analyzed by MALDI-TOF MS/MS (Fig. 6C). Assignment of fragment ions to the predicted b- and y-ion series suggested that its sequence is 2A+(L/I)-H-G-V-R-V-V-E-D-G-V-N-Y-A-T-G-N-(L/I)-P-G-C-S-F-S+(L/I)-F (propionamide cysteine is underlined). The observation that the sequence was found in the C-terminal part of the core protein indicates that its sequence is A-L-A-H-G-V-R-V-V-E-D-G-V-N-Y-A-T-G-N-L-P-G-C-S-F-S-I-F (residues 150 to 177). Furthermore, its observed monoisotopic mass (m/z 2964.51) is almost identical to that calculated for the C-terminal peptide (residues 150 to 177; m/z 2964.47) which possesses cysteine alkylated with acrylamide (propionamide cysteine) (Table 1). In MALDI, the ionization of peptides is greatly enhanced by the presence of basic amino acids (R, K, and H), and the positive ion signals of R-containing peptides appear to be significantly higher than those for other peptides (22). Thus, the absence of y1 to y21 ions (except for y8) and presence of y22 to y25 ions support the idea that peptide 14 contains an R residue at position 7 (residue 156) and not a C-terminal R or K residue, which should be present at the C termini of tryptic peptides except for the C-terminal peptide. In addition, the absence of b1 to b3 ions is probably due to the lack of basic amino acids in these fragment ions. From this observation, peptide 14 was identified as the C-terminal peptide (residues 150 to 177) from p21.

Our MS/MS analysis of peptides 5 and 14 was reproducible for three independently purified preparations. The structure of p21 was deduced based on these findings (Fig. 7A). The mature core protein (p21) purified from insect cells includes residues 2 to 177, and its N terminus is blocked with an acetyl group. We identified the specific tryptic peptides with 77% coverage of the p21 primary structure (Table 1 and Fig. 7A). The observed average mass (m/z 19307.1) of protonated p21 ([M+H]⁺) shown in Fig. 5B is in agreement with the calculated average mass (m/z 19307.1).

FIG. 7.

Posttranslational modifications of the HCV core protein. (A) The predicted amino acid sequence and posttranslational modifications for the core protein are shown. p21 starts with α-N-acetylserine (residue 2) and ends with phenylalanine (residue 177). The positions of the tryptic peptides identified by MALDI-TOF MS analyses (Fig. 6 and Table 1) are underlined. The peptide shown with dashed lines corresponds residues 13 to 39 or 14 to 40. The matching peptides covered 77% of the proposed structure of p21. The calculated average molecular mass (m/z) of the single-charged p21 ion ([M+H]⁺) is shown. (B) The immature core protein (p23) undergoes co- and/or posttranslational modifications with the indicated enzymes to generate the mature core protein (p21). The N- and C-terminal amino acid sequences of p23 are shown. The signal sequence is depicted in the shaded box.

DISCUSSION

The major obstacle in analyzing the maturation and assembly of the HCV core, E1, and E2 structural proteins has been the lack of an efficient system for producing virus particles in mammalian cells. Although a few systems used for assembly of structural proteins into virus-like particles in mammalian cells have been developed by using Semliki forest virus and vesicular stomatitis virus vectors (4, 10), Blanchard et al. (3) noted that virus-like particle budding toward the ER lumen is abortive. Furthermore, expression levels of virus proteins in mammalian cells are low. On the other hand, the baculovirus-insect cell expression system has been applied to studies on the maturation and assembly of HCV structural proteins. Baumert et al. (2) reported that the core-E1-E2 polyprotein is processed into the core, E1, and E2 structural proteins in Sf9 insect cells and that they are capable of assembling into virus-like particles. Virus-like particles bud into the lumen of cytoplasmic vesicles or vacuoles derived from the ER and are enveloped with cellular membrane (2). Furthermore, the immature core protein is efficiently processed into its mature form in Sf9 insect cells (18) as well as mammalian cells (25, 32, 39, 44). These findings suggest that the core protein reaches maturation in the insect cells, although there is no evidence that the expressed 191-amino-acid HCV core is translocated properly and cleaved by SPPase correctly in insect cells. SPPase in insect cells is very similar to the version in mammalian cells (42). However, it is still possible that a subtle structural difference between mammalian and insect forms of the protease may alter the location of the cleavage site; any small differences in molecular weight for mature core generated in insect compared to mammalian cells would not be possible to detect by PAGE.

As previously reported (18), p23 is efficiently processed into p21 in Sf9 insect cells (Fig. 1). The subcellular fractionation study of cells expressing the core protein indicated that the majority of p21 associates with P10 membranes (Fig. 1). The distribution of p21 in the subcellular fractions was similar to that for mammalian cells (34). The P10 fraction may contain heavy membranous organelles such as mitochondria, lysosomes, peroxisomes, and also ER fragments. Since the core protein is associated with mitochondria (37), it is possible that a portion of p21 in the P10 fraction is associated with mitochondria. When P10 membranes were treated with sodium carbonate (pH 11.5) to release the luminal contents and peripheral proteins from the membranes (13), 60% of the p21 remained associated with the LDMs (Fig. 2), while 15% was solubilized. This indicates that a significant portion of p21 in the P10 fraction behaves as an integral membrane protein. In contrast, half the amount of p21 in the P100 fraction of Sf9 cells, which may contain microsomes derived from the rough and smooth ER, was solubilized from membranes with alkaline treatment (data not shown), indicating that the half of the p21 protein in the microsome fraction behaves as a peripheral membrane protein. To investigate the membrane-binding properties of p21, which was expressed as a core-E1-E2-NS2 polyprotein in HeLa cells in the vaccinia virus expression system (44), we analyzed whether p21 is released from mammalian P10 membranes with sodium carbonate. According to the results, 30% of p21 associated with P10 membranes of HeLa cells was resistant to the alkaline treatment (data not shown). Santolini et al. (39) reported that the majority of p21 that associates with the microsomal membrane of mammalian cells behaves as a peripheral membrane protein. McLauchlan et al. (32) suggested that C-terminal cleavage of the immature core protein with SPPase is required for release of the mature core protein from the ER membrane and targeting to the surface of lipid droplets. They also reported that the mature core protein that associates with the lipid droplets can be extracted with sodium carbonate. Therefore, it is possible that the membrane-binding properties of p21 vary with subcellular localization.

We developed a simple method for the purification of p21 from P10 membranes by using the following properties (Fig. 3) (i) A significant portion of p21 tightly associates with LDMs after alkaline treatment of P10 membranes. (ii) p21 dissociates from LDMs as aggregates by treatment with 1% Triton X-100. (iii) p21 is solubilized from Triton-insoluble aggregates with 0.25% sarcosyl and reassembles into high-molecular-weight complexes by decreasing the sarcosyl concentration. Sarcosyl-solubilized p21 is present as a small complex with an apparent molecular weight of 150,000 and is capable of forming heterogeneous high-molecular-weight complexes with a sedimentation coefficient of 40 to 130S in the presence of Triton X-100 (Fig. 4). In the absence of Triton X-100, almost all p21 forms aggregates, which can be precipitated through a 0.4 to 1.0 M sucrose linear gradient (data not shown). In HCV particles, the hydrophilic N-terminal part of p21, which interacts with genomic RNA (11, 23, 39, 41), might be embedded in the viral nucleocapsid, whereas the hydrophobic C-terminal part may underlie the lipid bilayer to ensure the structural integrity of the particles. Therefore, in the absence of membranes, Triton X-100 might stabilize the high-molecular-weight complexes and inhibit their aggregation via C-terminal-mediated hydrophobic interactions. Kunkel et al. (23) reported that a purified C-terminal deletion mutant (residues 1 to 124) of the core protein, which was expressed in E. coli, can be assembled into spherical nucleocapsid-like particles resembling the nucleocapsids of HCV particles but that full-length core protein (residues 1 to 179) forms only irregular particles. Therefore, it may be difficult to reconstitute nucleocapsid particles in vitro with full-length core protein and RNA in the absence of membrane. From the observations of Kunkel et al. (23), RNA is thought to be essential for self-assembly of the core protein into nucleocapsid-like particles. Thus, it should be investigated whether RNA is copurified with p21 in our system. When high-molecular-weight complexes were assembled with purified p21 in the presence of RNase A, they were pelleted during centrifugation on a 0.4 to 1.0 M sucrose gradient under the same conditions as the experiments depicted in Fig. 4B (data not shown). This suggests that these complexes might be associated with RNA copurified with p21 and that the association of the complexes with RNA also prevents their aggregation. These properties are similar to those of p21 (residues 1 to 173) synthesized by the in vitro translation in rabbit reticulocyte lysate reported by Santolini et al. (39). They indicated that p21 (residues 1 to 173) in rabbit reticulocyte lysate forms an ∼80S complex and that the complex aggregates by treatment with RNase A. Although they speculated that p21 (residues 1 to 173) is associated with the 60S ribosomal subunit through its hydrophilic N-terminal region (39), our results suggest that purified p21 is assembled into high-molecular-weight complexes in the absence of the ribosomal subunit. However, we have not well-characterized the high-molecular-weight p21 complexes. Specifically, we still need to analyze other components (e.g., RNA) of these complexes and their supramolecular structures by electron microscopy.

Hüssy et al. (18) expressed the core (residues 1 to 188)-dehydrofolate reductase-hexa-histidine tag fusion protein, which contains the SPPase cleavage site in the signal sequence but not the SPase cleavage site corresponding to the core-E1 junction, in Sf9 insect cells by using a baculovirus expression system and analyzed the N-terminal sequences of C-terminal dehydrofolate reductase fragments. Their findings indicated that the C-terminal fragments start with alanine at residue 180 or serine at residue 183. Therefore, they speculated that the mature core protein ends with leucine at residue 179 or 182. Here, we subjected p21 purified from baculovirus-infected Sf9 cells to MALDI-TOF MS analyses to obtain direct structural information and found that the mature core protein (molecular mass of 19,306 Da) starts with α-N-acetylserine (residue 2) and ends with phenylalanine (residue 177) (Table 1 and Fig. 5 to 7). No other posttranslational modifications were detected. SPPase is closely related to another intramembrane-cleaving aspartic protease, presenilin, which is a component of γ-secretase and is involved in processing of amyloid-β precursor protein, Notch, Erb-B4 receptors, and E-cadherin (43). γ-Secretase cleaves the transmembrane region of the amyloid-β precursor protein at several sites (9). Therefore, it is possible that SPPase also heterogeneously processes the signal sequence of the core protein at several sites (e.g., 177 or 178, 179 or 180, and 182 or 183). However, the major form of p21 from the insect cells ended with phenylalanine at residue 177, which is strictly conserved in all core proteins of HCV isolates (7). In addition, all core proteins start with the N-terminal methionine-serine sequence (7), which is potentially processed by methionine aminopeptidase (reviewed in reference 6). Methionine aminopeptidase is conserved in all organisms and specifically removes the initiator methionine from the N terminus of proteins having penultimate amino acids with small side chains (glycine, alanine, serine, cysteine, threonine, proline, and valine) (6). Kunkel et al. (23) mentioned that removal of the initiator methionine from recombinant core protein occurs in E. coli. In eukaryotes, proteins with N-terminal serine are most frequently acetylated with N^α-acetyltransferase NatA (reviewed in reference 38). Removal of the initiator methionine and subsequent N-terminal acetylation occur cotranslationally in eukaryotes (6, 38). Therefore, it is plausible that the N terminus of the core protein is cotranslationally processed by methionine aminopeptidase and N^α-acetyltransferase and that N-terminal modifications participate in stabilizing the protein.

In this study, we established a method for purification of the mature core protein (p21) from the membrane fraction of insect cells and also determined its posttranslational modifications by direct structural analyses with MALDI-TOF MS. We believe that the immature core protein undergoes co- and/or posttranslational modifications with at least three enzymes, methionone aminopeptidase, N^α-acetyltransferase, and signal peptide peptidase, to generate the mature form in mammalian as well as insect cells (Fig. 7B), because these enzymes are highly conserved in higher eukaryotes. Experiments are in progress to address the significance of these modifications on the biological activities of the core protein.