ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 342, No. 2, June 15, pp. 373–382, 1997Article No. BB970139

Viral Infection

II. Hemin Induces Overexpression of p67 as It Partially Prevents Appearanceof an Active p67-Deglycosylase in Baculovirus-Infected Insect Cells

Debabrata Saha, Shiyong Wu,1 Avirup Bose, Nabendu Chatterjee, Arup Chakraborty,2

Madhumita Chatterjee, and Naba K. Gupta3

Department of Chemistry, University of Nebraska, Lincoln, Nebraska 68588

Received March 7, 1997, and in revised form April 3, 1997

is more efficiently recognized by the p67-DG antibod-ies since these antibodies were prepared against theThe roles of p67-deglycosylase (p67-DG) in the regu-active form of p67-DG. q 1997 Academic Presslation of protein synthesis in baculovirus-infected in-

Key Words: p67; glycoproteins; baculovirus expres-sect cells were studied. Like vaccinia viral infection,sion system; deglycosylation; p67-DG; hemin.baculovirus infection of insect cells also induced the

appearance of a p67-DG. However, p67-DG activitycould not be detected because these cells do not con-tain a detectable level of p67. The baculovirus expres-sion vector system (BEVS), however, promotes sig- Bose et al. (1) reported that vaccinia viral infectionnificant expression of cloned p67-cDNA. The expres- induced the appearance of a p67-DG4 activity in thesion of p67 was significantly enhanced by the addition infected animal cells (KRC-7 and L929). The activatedof hemin to the growth medium. Maximum enhance- p67-DG deglycosylated p67 and inactivated it. This al-ment was observed at 5 mM hemin. Data suggest that lowed eIF-2 kinase(s) to phosphorylate eIF-2 a-subunithemin prevents the activation of latent p67-DG inside and inhibit host protein synthesis. Previously, we re-the cell and does not have any effect on p67 gene tran- ported that the reticulocyte lysate also contains a p67-scription. To gain a better understanding of the mech- DG in latent form (2). During heme deficiency, thisanism of p67-DG activation and hemin stimulation of deglycosylase was activated. This inhibited proteinp67 synthesis, we have now purified p67-DG from ba- synthesis in the reticulocyte lysate. The mechanisms ofculovirus-infected insect cells. We prepared antibod- induced appearance of p67-DG in viral-infected animalies against this protein. These antibodies reacted with cells (1) and activation of a latent p67-DG in reticulo-a 105-kDa protein in cell extracts from the uninfected cyte lysate (2) are not clear.insect cells (Sf9), KRC-7, and L929 (animal cells). In

In the present work, we have expressed a cloned p67addition, these antibodies reacted with an additionalcDNA (3) in a baculovirus expression system. The p6760-kDa protein in the cell extracts of baculovirus-in-expression was enhanced up to fivefold by inclusion offected Sf9 cells and vaccinia virus-infected KRC-7 andhemin (5 mM) in the growth medium. The expressedL929 cells. Data are also presented to show that thep67 was purified to homogeneity and was found to beantibodies against p67-DG reacted more efficientlyidentical to native p67 in several reactions tested.(40%) with the 60-kDa protein in both hemin-deficient

To better understand the mechanism of p67 deglyco-reticulocyte lysate and hemin-deficient baculovirus-sylation, p67-DG has now been purified to near homo-infected cells. We suggest that hemin prevents the con-geneity (MrÇ60 kDa) and prepared polyclonal antibod-version of an inactive p67-DG into an active form pos-

sibly by covalent modification such as protein phos- ies against this protein. These antibodies reacted withphorylation or protein glycosylation. The active form a 105-kDa protein in the cell extracts from both unin-

fected insect and animal cells but reacted with an ad-

1 Present address: HHMI, University of Michigan, Medical Center,4 Abbreviations used: p67, eIF-2-associated 67-kDa protein; p67-Ann Arbor, MI-48109.

2 Present address: Division of Infectious Diseases, University of DG, p67-deglycosylase; eIF-2, eukaryotic initiation factor 2; HRI,heme-regulated protein synthesis inhibitor (eIF-2 kinase); GlcNAc,Pittsburgh School of Medicine, Pittsburgh, PA 15213.

3 To whom correspondence should be addressed. Fax: (402) 472- N-acetylglucosamine; AcMNPV, Autographa californica nuclearpolyhedrosis virus; pfu, plaque-forming unit.9402.

3730003-9861/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID ABB 0139 / 6b38$$$201 05-17-97 01:22:28 arcal

374 SAHA ET AL.

(m.o.i.) of 10 in a spinner culture at 277C. The production of p67 wasditional 60-kDa protein in both baculovirus-infectedmonitored by immunoblotting. The cells were harvested at 12 to 120insect cells and vaccinia virus-infected animal cells.h postinfection. The cells were lysed in a lysing buffer containing 10Similar proteins are also present in both heme-supple- mM Tris–HCl (pH 7.5), 130 mM NaCl, 1% Triton X-100, 10 mM

mented and heme-deficient reticulocyte lysate. These NaPPi, and a protease inhibitor cocktail (benzaminidine, phenan-throline, aprotinin, leupeptin, pepstatin, PMSF). The lysates werestudies have led to proposed a tentative mechanismkept in ice for 45 min. p67 production was analyzed by immunoblot-of p67-DG activation and hemin inhibition of p67-DGting using p67 polyclonal antibodies.during viral infection.

Purification of p67 from infected Sf9 cells. As the p67 levelreached an optimal level in the transfected Sf9 cells (approximately72 h after infection), the cells were harvested and lysed in the lysingMATERIALS AND METHODSbuffer containing protease inhibitors. The lysates were centrifuged

Most of the materials and the experimental procedures used in at 10,000g for 30 min and the cell debris were removed. The cellthis study were the same as described previously (2–7). Three-sub- supernatant was then dialyzed in a buffer containing 20 mM potas-unit eIF-2 and p67 were purified according to the procedure described sium phosphate (pH 7.6), 10 mM b-mercaptoethanol, 50 mM EDTA,by Datta et al. (4). HRI was purified as described by Trachsel et al. 10% glycerol, and the protease inhibitor cocktail. The dialyzed solu-(8). The p67 poly- and monoclonal antibodies and eIF-2a antibodies tion containing 5 mg proteins was then applied onto a hydroxylapa-were prepared as described. As reported, p67 monoclonal antibodies tite column (Bio-Rad-HTP cartridge for FPLC). The column wasspecifically recognize the GlcNAc moieties on p67 (5). washed with five column volumes of the same buffer. The proteins

were eluted with a linear salt gradient of 200–400 mM potassiumSf9 cell culture. Sf9 cells were cultured in Grace’s insect mediumphosphate (pH 7.6). The fractions were analyzed by immunoblotting(Gibco BRL) [contains 10% fetal calf serum (Sigma), 10 mg/ml genta-using p67 polyclonal antibodies. The fractions containing expressedmycin (Gibco BRL), and 0.1% v/v surfactant Pluronic F-68 (Gibcop67 were pooled and dialyzed against buffer A [50 mM Tris–HCl (pHBRL)] at 277C either in monolayers or in suspension cultures. In7.8), 100 mM KCl, 5 mM b-mercaptoethanol, 100 mM EDTA, and 10%suspension cultures, the cells were incubated in spinner culture flask(v/v) glycerol]. Approximately, 1 mg of the dialyzed protein was thenwith constant stirring at 50–60 rpm (9). The cell viability was moni-injected onto a Mono-S HR 5/5 anion exchange column and wastored using trypan blue and was maintained §97% all the time. Theeluted with a linear salt gradient of 100–400 mM KCl in buffer A.surfactant was used to reduce the shear sensitivity of the cells.The p67 was eluted approximately at 200 mM KCl. The eluted proteinProper aeration was required when large volume of the cells (ú100was analyzed by 15% SDS–PAGE and immunoblotting using p67ml) was cultured. In several experiments, 5 mM hemin (Sigma) wasantibodies.added to the growth medium.

Analysis of p67 deglycosylation in Sf9 cell lysates. The cell lysatesConstruction of recombinant baculovirus. The pGEM-p67 (3) waswere prepared. The cell lysates were incubated with or without exog-mutated at the region of protein synthesis initiation site fromenously added p67 at 377C. At different intervals, the aliquots of theACATGG to CCATGG using a PCR method (10) to generate ancell extracts were analyzed for p67 using p67 mono- and polyclonalunique NcoI site. The mutated gene sequence was confirmed by DNAantibodies (2).sequencing using Sequencing II kit (USB) and Sequagel Sequencing

Determination of RNA levels. The Northern blot technique wasSystem (National Diagnostics).the same as described (12). The Sf9 cells were cultured and infectedThe entire p67 open reading frame was excised from mutatedwith the recombinant virus in the presence or absence of hemin atpGEM-p67 using NcoI and HindIII digestion and purified by agarosean m.o.i. of 10 for 0, 12, 24, 48, and 72 h postinfection and were thengel. This fragment contained p67 gene sequence from 1 to /1474 (30harvested. The total RNA was isolated using guanidium isothiocya-bp after the stop codon) relative to the ATG initiation codon. Thenate method (13). Approximately 10 mg of the total RNA was used.pBlueBac III vector (Invitrogen) was linearized using NcoI and Hin-

dIII restriction enzymes and dephosphorylated using calf intestine Assay of p67 activity to inhibit HRI-catalyzed phosphorylation ofalkaline phosphatase. Purified p67 gene sequence was then ligated eIF-2 a-subunit. The assay was carried out using the procedureto the vector to form pBB-p67 using T4 DNA ligase. The recombinant described previously (6). The purified native p67 and the expressedtransfer vector (pBB-p67) was used to transform competent Esche- protein were dialyzed against PSS IV buffer [20 mM Tris–HCl (pHrichia coli cells (DH5aF cells; Gibco BRL) for a large-scale preparation 7.8), 50 mM EDTA, 5 mM b-mercaptoethanol, and 10% (v/v) glycerol]of the plasmid. The plasmid was purified using the DNA cleanup containing 100 mM KCl. The assay mixtures contained, in additionsystem (Promega) and was used in cotransfection with linear to the standard buffer: 1 mg eIF-2, 0.04 mg HRI, 0.5 mg native p67,AcMNPV wild-type viral DNA (Invitrogen) into Sf9 cells using cat- or the expressed protein as indicated. As earlier, eIF-2 a-subunitionic liposome (Invitrogen) (11). Linear AcMNPV DNA (1 mg) and was phosphorylated using [g-32P]ATP (259 TBq/mmol, ICN Biochem-pBB-p67 (3 mg) were used to transfect 2 1 107 Sf9 cells. The trans- ical) and the radioactively labeled protein was analyzed by SDS–fected cells were cultured in complete TNM-FH (Invitrogen) medium. PAGE followed by autoradiography.After 4–5 days of incubation, the virus secreted into the medium was Purification of p67-deglycosylase. Sf9 cells were infected withcollected and analyzed for plaque formation. The virus sample at wild-type baculovirus (AcMNPV) at an m.o.i. of 10. After 72 h ofdifferent dilutions (1003–1007) were used to infect a fresh culture of infection, the cell lysates were prepared according to the procedureSf9 cells. After 1 h of infection, the cells were covered with 1.5% described above. The lysate was then subjected to 0–40% ammoniumSeaPlaque agarose (FMC) in complete TNM-FH medium with 25 mg/ sulfate cut. The precipitate was then dialyzed in PSS IV containingml gentamycin and 150 mg/ml X-gal. After 4–5 days of incubation at 100 mM KCl for 12 h. The dialyzate was clarified by centrifugation277C, the blue plaques were observed and counted to determine the at 10,000g for 15 min and injected onto a Mono-Q H/R anion exchangeviral titer and also screened to identify the recombinant viruses. The column previously equilibrated with buffer A containing 100 mMsingle plaque was then used to infect a fresh culture of Sf9 cells. After KCl. The column was washed with 10 ml of the same buffer and was4–5 days, the cells were harvested and the cell extracts were analyzed then eluted using a linear KCl gradient (100–500 mM, total volumeby immunobloting using p67 poly- and monoclonal antibodies. The 30 ml) in buffer A. The fractions containing p67-DG activity elutedsupernatant from the cells expressing p67 was used as viral source. at approximately 175–200 mM KCl in buffer A.These viruses were plaque purified and subsequently amplified. The p67-DG was assayed using a synthetic substrate p-nitrophe-

nyl-N-acetyl-b-D-glucosaminide following the procedure described byExpression of recombinant protein. The Sf9 cells were culturedat a density of 2 1 106/ml and infected with the recombinant virus (14). One unit is defined as the amount of enzyme which hydrolyzes

1 mmol of p-nitrophenyl glycoside per minute at 307C. The p67-DGin the presence or absence of hemin at a multiplicity of infection

AID ABB 0139 / 6b38$$$202 05-17-97 01:22:28 arcal

375EXPRESSION OF p67 IN BACULOVIRUS-INFECTED INSECT CELLS

activity during purification was also analyzed by standard immu-noblot procedure using both p67 mono- and polyclonal antibodies.The purity of p67-DG was analyzed using a 15% SDS–PAGE. Thegel was stained with Coomassie blue.

Preparation of p67-DG polyclonal antibodies. The polyclonal an-tibodies against the purified p67-DG were prepared following theprocedure as described (5). The antibodies were purified from mouseserum according to the manufacturer’s (Bio-Rad) procedure usingBio-Rad DEAE–Affigel blue column chromatography.

Characterization of p67-DG polyclonal antibodies. The p67-DG FIG. 2. Effect of hemin on the expression of p67 in the baculoviruspolyclonal antibodies were characterized using standard immunoblot system. The experimental procedures were same as described in theprocedure. Purified p67-DG and cell extracts from uninfected and legend to Fig. 1 except that hemin at varying concentrations wasinfected insect cells were used as antigens. added to the growth medium. The p67 in the cell lysates (25 mg

protein) was measured by immunoblotting using p67 monoclonal an-Detection of p67-DG in vaccinia viral-infected animal cells. KRC-tibodies. The concentrations of hemin added were: lane 1, 0; lane 2,7 (rat hepatoma) and L929 (mouse fibroblast) cells were infected with1 mM; lane 3, 3 mM; lane 4, 5 mM; lane 5, 10 mM; lane 6, 20 mM.25 or 5 pfu per cell of vaccinia virus as described (1). After 4 h of

infection the cells were harvested and lysed. Four hundred micro-grams of proteins in the cell extracts was incubated with p67-DGpolyclonal antibodies overnight at 47C. The antigen–antibody com-

the experiments described in lanes 1–3, we used nativeplexes were then immunoprecipitated with protein A–agarose. Thep67 at three different concentrations (lane 1, 0.5 mg,immunoprecipitates were then analyzed by SDS–PAGE followed by

immunoblotting with p67-DG polyclonal antibodies. lane 2, 1.0 mg, lane 3, 1.5 mg). The p67 was not detectedEffects of hemin on p67-DG antibody-reactive materials in unin- in uninfected Sf9 cells (lane 4) and in Sf9 cells infected

fected and baculovirus-infected insect cell lysates. Sf9 cells were with wild-type baculovirus (lane 5). The p67 was alsogrown in the presence or absence of hemin and were infected with not detected in the culture medium or in the cell debrisbaculovirus. The p67-DG antibody-reactive materials in the cell ex-

(data not shown). The expressed protein migrated simi-tracts were analyzed using standard immunoblot procedures.larly to native p67 (lane 6).Effects of hemin on p67-DG antibody-reactive materials in hemin-

Hemin addition to the growth medium significantlysupplemented and hemin-deficient rabbit reticulocyte lysates. Thereaction mixtures contained (total volume 40 ml): 15 ml (500 mg) retic- increased the p67 yield (Fig. 2). The cells were grownulocyte lysate, 10 mM Tris–HCl, pH 7.8, 1 mM Mg(OAc)2, 100 mM for 72 h in the presence or in the absence of hemin.KCl, 0.2 mM GTP, 5 mM creatine phosphate and 0.4 mM creatine The p67 levels in the cell lysates were analyzed usingphosphokinase. Hemin (20 mM) was added to the reaction mixture

standard immunoblot procedures as described in theas indicated. The reactions were incubated at 377C for 20 min. Thereaction mixtures were immunoprecipitated using p67-DG polyclonal legend to Fig. 1. The p67 was present in the cell lysateantibodies and protein A–agarose. The immunoprecipitates were without hemin (lane 1) and addition of hemin in in-subsequently analyzed by immunoblotting using p67-DG polyclonal creasing concentrations up to 5 mM (lanes 2–4) led toantibodies. an increased appearance of p67. Further increases in

hemin concentration to 10 mM (lane 5) and 20 mM (laneRESULTS 6) led to significant decrease in p67. Scanning by Molec-

Expression of p67 in Sf9 cells in the absence and ular Dynamics Image Quant Version 3.3 showed anpresence of hemin. The recombinant baculovirus approximate fivefold increase in p67 production in cellclones, expressing p67, were characterized by immu- lysates supplemented with 5 mM hemin (lane 4) overnoblot analysis of total cellular proteins using p67 poly- that produced in the absence of hemin (lane 1).clonal antibodies. The results are shown in Fig. 1. In Appearance of a p67-DG activity in baculovirus-in-

fected insect cells and hemin inhibition of this appear-ance. The p67-DG activity in different cell extractswas analyzed. The cell extracts were incubated at 377Cand the remaining p67 was measured by immunoblot-ting using p67 mono- and polyclonal antibodies. Theimmunoblots were quantitated by scanning using Mo-lecular Dynamics Image Quant version 3.3. As pre-viously mentioned (2), the p67 monoclonal antibodiesare specific for the glycosyl residues on p67 and can beused for assays of p67-DG activity. The results areshown in Figs. 3A–3D. The insect cells do not containFIG. 1. Expression of p67 in baculovirus-infected Sf9 cells. The Sf9detectable level of p67. Therefore, in the experimentscells were infected with baculovirus containing p67 cDNA. The p67

in the cell lysates was measured by immunoblotting using p67 poly- described in Figs. 3A and 3B, 0.5 mg exogenous p67clonal antibodies. Purified p67 was used for comparison. Lane 1, 0.5 was added. The p67-deglycosylation was measured inmg p67; lane 2, 1.0 mg p67; lane 3, 1.5 mg p67; lane 4, 25 mg proteins the extracts from uninfected insect cells (Fig. 3A, lanesfrom uninfected Sf9 cell lysate; lane 5, 25 mg proteins from wild-type

1–8) and also from the insect cells infected with wild-baculovirus-infected Sf9 cell lysate; lane 6, 25 mg protein from p67recombinant baculovirus-infected cell lysate. type baculovirus (Figs. 3B, lanes 1–8). In the extracts

AID ABB 0139 / 6b38$$$202 05-17-97 01:22:28 arcal

376 SAHA ET AL.

from uninfected cells, both p67 mono- and polyclonal of incubation at 377C. However, when the extracts fromantibody-reactive materials remained relatively stable. the baculovirus (wild-type)-infected insect cells wereApproximately, 75% of both the p67 monoclonal (Fig. used, the rate of disappearance of the p67 monoclonal3A, lane 4) and the p67 polyclonal (Fig. 3A, lane 8) antibody-reactive materials was significantly fasterantibody-reactive materials were present after 20 min (Fig. 3B, lanes 2–4) than the rate of disappearance of

the p67 polyclonal antibody-reactive materials (Fig.3B, lanes 6–8). Approximately 50% of the p67 mono-clonal antibody reactive materials disappeared after 5min of incubation (Fig. 3B, lane 2) and this disappear-ance was almost totally complete after 20 min of incu-bation (Fig. 3B, lane 4). On the other hand, the p67polyclonal antibody-reactive materials were signifi-cantly more stable. Approximately, 80% of the p67 poly-clonal reactive materials were present after 5 min ofincubation (Fig. 3B, lane 6) and more than 50% of thep67 polyclonal antibody-reactive materials were pres-ent after 20 min of incubation (Fig. 3B, lane 8). In theexperiments described in Figs. 3C and 3D, the insectcells were grown in the absence (Fig. 3C) and in thepresence (Fig. 3D) of 5 mM hemin and were then in-fected with baculovirus containing p67 cDNA. Thesecell extracts contained significant amounts of ex-pressed p67. The disappearance of this expressed p67was monitored. As shown in Fig. 3C, the extracts fromthe insect cells grown in the absence of hemin activelydeglycosylated endogenous p67 (lanes 2–4). The p67polyclonal antibody-reactive materials in this cell ex-tracts were, however, significantly stable (lanes 6–8).After 20 min of incubation, 90% of the p67 monoclonalantibody-reactive materials disappeared (lane 4). Un-der similar experimental conditions, less than 20% ofthe polyclonal antibody-reactive materials disappeared(lane 8). However, when the insect cells were grown inthe presence of 5 mM hemin, both the p67 monoclonal-reactive (Fig. 3D, lanes 1–4) and p67 polyclonal-reac-tive (Fig. 3D, lanes 5–8) materials were relatively sta-ble. Approximately, 50% of the monoclonal antibody-reactive materials (lane 4) and 90% of the polyclonalantibody-reactive materials (lane 8) were detected after20 min of incubation.

In the experiments described in Fig. 4, we studied themechanism of hemin prevention of p67 deglycosylation.We grew the insect cells in the absence of hemin andFIG. 3. Appearance of a p67-DG activity in baculovirus-infected

insect cells and hemin inhibition of this activity. The p67-DG activity then infected the cells with baculovirus containing p67.was measured in different cell extracts using both p67-mono- and The cell lysates were prepared. The cell lysates werepolyclonal antibodies. The experimental conditions were the same

then incubated with exogenous p67 in the absenceas described under Materials and Methods. In the experiments de-(lanes 1–4) and in the presence (lanes 5–8) of 5 mMscribed in lanes 1–4 in all the panels (A–D), p67-monoclonal antibod-

ies and in lanes 5–8 in all the panels (A–D), p67-polyclonal antibod- hemin. At different intervals, the remaining p67 wasies were used. The reactions were incubated at 377C for different assayed using p67 monoclonal antibodies. As shown,intervals: Lanes 1 and 5, 0 min; lanes 2 and 6, 5 min; lanes 3 and p67 was almost totally deglycosylated after 20 min of7, 10 min; lanes 4 and 8, 20 min. (A) The extracts from uninfected

incubation both in the absence (lane 4) and in the pres-Sf9 cells (25 mg protein) contained 0.5 mg exogenous p67. (B) Theextracts (25 mg) from wild-type baculovirus-infected Sf9 cells con- ence (lane 8) of hemin. These results along with thetained 0.5 mg p67. (C) The extracts (25 mg) from p67-recombinant results presented in Fig. 3, suggest that baculovirusbaculovirus-infected Sf9 cells grown in the absence of hemin. No infection induces appearance of a p67-DG activity inexogenous p67 was added. (D) The extracts (25 mg) from the p67-

the infected insect cells and hemin partially preventsrecombinant baculovirus-infected Sf9 cells grown in the presence of5 mM hemin. No exogenous p67 was added. appearance of this activity. The hemin effect may be

AID ABB 0139 / 6b38$$$203 05-17-97 01:22:28 arcal

377EXPRESSION OF p67 IN BACULOVIRUS-INFECTED INSECT CELLS

were used. The results are shown in Fig. 6D. The HRIefficiently phosphorylated the eIF-2 a-subunit (lane 1).This phosphorylation was similarly inhibited by eithernative p67 (lane 2) or by the expressed proteins (lane3, synthesized in the absence of hemin; lane 4, synthe-sized in the presence of hemin).

Purification of p67-deglycosylase from wild-type ba-culovirus-infected insect cells. p67-deglycosylase waspurified to near homogeneity from insect cells infected

FIG. 4. Effect of addition of exogenous hemin on p67-deglycosyla-with wild-type baculovirus. During purification the fac-tion in the extracts from the p67-recombinant baculovirus-infectedtor preparations were assayed using a synthetic sub-Sf9 cells. The cells were grown in the absence of hemin. The reaction

conditions were same as described in the legend to Fig. 3C except strate p-nitrophenyl N-acetyl-b-D-glucosaminide fol-that 5 mM hemin was added to the reactions described in lanes 5–8. lowing a standard procedure (14). Table I summarizesThe reactions were incubated for different intervals: lanes 1 and 5, the results of purification. About 30-fold purification0 min; lanes 2 and 6, 5 min; lanes 3 and 7, 10 min; lanes 4 and 8,

over the crude cell extract was obtained. In SDS–20 min.PAGE (Fig. 7), the purified protein showed a singlepolypeptide band at approximately 60 kDa. During pu-rification, the factor preparations were also analyzedat the transcription level or posttranslationally duringfor p67-deglycosylation activity following standard im-activation of a latent p67-DG as reported previously inmunoblot procedure using both p67 mono- and poly-reticulocyte lysate (2). However, once activated in theclonal antibodies (Fig. 8). The results obtained usingabsence of hemin, the activated p67-DG deglycosylatesboth assay methods agreed indicating that these twop67 even in the presence of hemin. Hemin does notactivities copurified. The results shown in Fig. 8 de-directly inhibit p67-DG activity.scribes the p67-deglycosylation activity of the purifiedEffect of hemin on p67 transcription. We exploredpreparation. As earlier, both p67 mono- and polyclonalthe possibility that hemin may alter the p67 transcrip-antibodies were used in standard immunoblot proce-tion and thus alter the p67 yield. For these experi-dure. Exogenous p67 (0.5 mg) was added and the reac-ments, we analyzed p67-mRNA level in infected celltions were further incubated with purified p67-DG.extracts at different time intervals in the absence (Fig.

5A) or in the presence of (Fig. 5B) of hemin. As shown(Fig. 5), p67-mRNA level at similar intervals remainedessentially the same in the absence (Fig. 5A) or pres-ence of hemin. These results suggest that the hemineffect is posttranscriptional.

Effect of temperature on p67 deglycosylation in ex-tract from infected insect cells. In the experiments de-scribed in Figs. 3 and 4, we assayed p67 deglycosylationat 377C. The deglycosylation was almost totally com-plete after 20 min of incubation. To explain the accumu-lation of p67 inside the insect cells grown at 277C, wecompared the rates of p67 deglycosylation at 27 and377C. Our results showed that the rate of p67 deglyco-sylation was slow at 277C compared to 377C (data notshown).

Purification of p67 from Sf9 cells. The expressedp67 was purified from baculovirus-infected insect cellsgrown in the presence or in the absence of hemin asdescribed under Materials and Methods. The purifiedprotein was analyzed by SDS–PAGE and immunoblot-ting (Fig. 6). The expressed proteins (Fig. 6A, lanes 2and 3) migrated similarly as native p67 (Fig. 6A, lane FIG. 5. Effect of hemin on p67-transcription. Northern blot analysis1) and also reacted with p67 monoclonal (Fig. 6B, lanes of p67-mRNA was performed as described under Materials and Meth-

ods. Sf9 cells were infected with recombinant baculovirus in the ab-2 and 3) and polyclonal antibodies (Fig. 6C, lanes 2 andsence or presence of hemin. At different intervals, the cells were3). Lane 1 in Figs. 6B and 6C show the reactivities ofharvested and total RNA was isolated. Approximately 10 mg of thenative p67. The purified expressed proteins were also total RNA was used. A, lanes 1–5: cells were infected in the absence

analyzed for eIF-2 kinase inhibitory activity. Equal of hemin. B, lanes 1–5: cells were infected in the presence of hemin.In both panels: lane 1, 0 h; 2, 12 h; 3, 24 h; 4, 48 h; 5, 72 h.concentrations of native p67 or the expressed proteins

AID ABB 0139 / 6b38$$$203 05-17-97 01:22:28 arcal

378 SAHA ET AL.

FIG. 7. SDS-PAGE of purified p67-DG. The gel (15% SDS) was runin a minigel apparatus (Bio-Rad) and stained with Coomassie blueR-250. Lane 1, broad range molecular weight marker (Bio-Rad); lane2, purified p67-DG.

approximately 80% of the p67 polyclonal antibody-reac-tive material remained intact (Fig. 8B, lane 2). Theseresults show that the purified p67-DG did, indeed, deg-lycosylate p67.

Further characterization of p67-DG. p67-DG degly-FIG. 6. Characterization of expressed p67. (A) SDS–PAGE: Lane 1, cosylates p67 (Fig. 8). It is expected that such deglyco-native p67; lane 2, expressed p67 synthesized in the presence of hemin; sylation would eliminate p67 activity to protect eIF-2 a-lane 3, expressed p67 synthesized in the absence of hemin. In each

subunit from phosphorylation by eIF-2 kinases. Theseexperiment, 1 mg protein was used. (B) Immunoblot analysis usingresults were obtained (Fig. 9). HRI efficiently phos-p67 monoclonal antibodies: Lane 1, native p67; lane 2, expressed p67

synthesized in the presence of hemin; lane 3, expressed p67 synthe- phorylated eIF-2 a-subunit (lane 1) and purified p67-sized in the absence of hemin. In each experiment, 0.5 mg protein was DG alone did not have any significant effect. Additionused. (C) Immunoblot analysis using p67 polyclonal antibodies. The of purified p67 almost fully inhibited this phosphoryla-experimental conditions were the same as in B except that p67 poly-

tion reaction (lane 3). However, when p67 was preincu-clonal antibodies were used. (D) Inhibition of eIF-2 a-subunit phos-bated with increasing levels of p67-DG, p67 activity tophorylation by native p67 and the expressed proteins. The reaction

conditions were as described under Materials and Methods. The reac- protect eIF-2 a-subunit was progressively lost (lanestion mixtures were incubated at 377C for 10 min and the reactions 4–6).were terminated by adding a solution containing 2% SDS, 50 mM b-

Preparation and characterization of p67-DG poly-mercaptoethanol, 50 mM Tris–HCl, pH 7.0, 50% (v/v) glycerol andclonal antibodies. Polyclonal antibodies against puri-0.2% bromophenol blue. The reaction mixtures were heated at 1007C

for 3 min and analyzed by SDS–PAGE followed by autoradiography. fied p67-DG were prepared in mice as described underLanes: 1, eIF-2 alone; 2, eIF-2 / native p67; 3, eIF-2 / the expressed Materials and Methods. These antibodies reacted withprotein (0 hemin); 4, eIF-2 / the expressed protein (/ hemin). purified p67-DG in standard immunoblot experiment

(Fig. 10).Reactivity of p67-deglycosylase polyclonal antibodiesMore than 90% of the p67 monoclonal antibody-reac-

tive material disappeared during 20 min incubation toward protein molecules in insect cells before and afterbaculovirus infection. Standard immunoblot experi-(Fig. 8A, lane 2). Under similar incubation conditions,

TABLE I

Purification of p67-DG from Baculovirus-Infected Insect Cells

Volume Protein Total activity Specific Activity % FoldFractions (ml) (mg) (mmol/min) (mmol/min/mg protein) Yield purification

Lysate 9.0 82.0 0.695 0.0085 100 10–40% AmSO4 2.0 12.2 0.580 0.048 83 5.6Mono-Q 1.0 1.3 0.331 0.254 47 30

AID ABB 0139 / 6b38$$$203 05-17-97 01:22:28 arcal

379EXPRESSION OF p67 IN BACULOVIRUS-INFECTED INSECT CELLS

FIG. 8. Assay of the purified p67-DG. 0.5 mg of purified p67 and0.5 mg of purified p67-DG were incubated at 377C for 0 and 20 min.The p67 deglycosylation was measured by standard immunoblottingusing p67-mono (A) and polyclonal (B) antibodies. Lane 1, (A and B)for 0 min; lane 2, (A and B) for 20 min.

FIG. 10. Characterization of p67-DG polyclonal antibodies: Thepreparation of p67-DG polyclonal antibodies has been described un-ments using extracts from insect cells were performed.der Materials and Methods. 0.25 mg purified p67-DG was used anThese polyclonal antibodies reacted with only one high-antigen in standard immunoblot experiment.molecular-weight (105 kDa) protein in the uninfected

cell extracts (Fig. 11A, lane 1). Upon baculovirus infec-tion, these antibodies reacted with the same high-mo- pitated with p67-DG antibodies and protein A–aga-lecular-weight (105 kDa) protein and an additional low- rose. The immunoprecipitates were subsequently ana-molecular-weight protein (60 kDa) in the infected cell lyzed by p67-DG antibodies. As shown (Fig. 11B), theextracts. This low-molecular-weight protein has the extracts from both uninfected KRC-7 (lane 1) and L929same size as purified p67-DG (60 kDa). cells (lane 4) reacted with a high-molecular-weight pro-

Detection of p67-DG polyclonal antibody reactive ma- tein (105 kDa) and there was no significant change interials in animal cells before and after vaccinia viral the level of this protein after vaccinia viral infection.infection. Bose et al. (1) reported that vaccinia viral KRC-7 cells were resistant to 5 pfu per cell of the virusinfection induces appearance of a p67-DG activity in but were sensitive when 25 pfu per cell of the virusinfected cells. Two different cell extracts (KRC-7 and were used. L929 cells were, however, sensitive to theL929) were used. These extracts were coimmunopreci- virus at 5 pfu per cell of the virus. The results show

that productive viral infection in KRC-7 (infected with25 pfu per cell of virus, lane 3) and L929 (infected with5 pfu per cell of virus, lane 5) cells induced appearanceof a nearly 60-kDa protein which was immunoreactivewith p67-DG polyclonal antibodies. This protein wasnot present in mock-infected KRC-7 cells (lane 1), inviral-resistant KRC-7 cells infected with 5 pfu per cellof the virus (lane 2), and in mock-infected L929 cells(lane 4). The additional polypeptides bands observed inall lanes represent the heavy chain of immunoglobulin.

Effect of hemin on p67-deglycosylase polyclonal anti-body-reactive materials in baculovirus-infected insectcell lysates and in rabbit reticulocyte lysates. As re-ported (1), hemin prevented appearance of p67-DG inbaculovirus-infected insect cells. Hemin also preventedFIG. 9. p67-DG catalyzed deglycosylation of p67 and consequentactivation of p67-DG in reticulocyte lysate (2).inhibition of p67 activity to protect eIF-2 a-subunit from HRI-cata-

lyzed phosphorylation. The reactions were carried out in two steps. We have now examined the effect of hemin on p67-In step 1, the reactions contained 20 mM Tris-HCl, pH 7.8, 50 mM DG level in both baculovirus infected insect cells (Fig.EDTA, 5 mM b-mercaptoethanol, 10% (v/v) glycerol and 100 mM KCl. 12A) and in reticulocyte lysate (Fig. 12B). StandardIn addition, the reactions contained: lane 1, none; lane 2, 0.5 mg p67-

immunoblot procedure was used for baculovirus-in-DG; lane 3, 0.5 mg p67; lane 4, 0.5 mg p67 / 0.25 mg p67-DG; lanefected insect cell extracts (Fig. 12A). The reticulocyte5, 0.5 mg p67 / 0.5 mg p67-DG; lane 6, 0.5 mg p67 / 1.0 mg p67-DG.

All the reactions were incubated at 377C for 20 min. To the reaction cell lysates (500 mg) were first coimmunoprecipitatedmixtures, 1.0 mg three-subunit eIF-2, 0.04 mg HRI, and [g-32P] ATP with p67-DG polyclonal antibodies and protein A–(259 TBQ/mmol, ICN Biochemical) were then added (step 2). The agarose. The immunoprecipitates were then analyzedreactions were further incubated at 377C for 10 min. Radioactively

by standard immunoblotting using p67-DG antibod-labeled eIF-2 was then analyzed by SDS–PAGE followed by autora-diography. ies. The two additional bands observed in lanes 2 and

AID ABB 0139 / 6b38$$$203 05-17-97 01:22:28 arcal

380 SAHA ET AL.

3 (Fig. 12B) represent the heavy chain of immuno-globulin.

The p67-DG antibodies reacted with an Ç60-kDaprotein in cell lysates from both baculovirus-infectedinsect cells (Fig. 12A, lanes 4–5) and in reticulocytelysates (Fig. 12B, lanes 2–3). The molecular weight ofimmunoreactive materials in reticulocyte lysate wasslightly higher than the molecular weight of the immu-noreactive materials in infected insect cells cells (Fig.

FIG. 12. Effects of hemin on p67-DG polyclonal antibody-reactivematerials in baculovirus-infected insect cell lysates and in rabbitreticulocyte lysates. (A) The reaction conditions using Sf9 cells havebeen described. Lanes: 1, purified p67-DG (2 mg); 2, hemin-deficientuninfected Sf9 cell lysates (25 mg); 3, hemin-supplemented uninfectedSf9 cell lysates (25 mg); 4, hemin-deficient wild-type baculovirus-infected Sf9 cell lysates (25 mg); 5, hemin-supplemented wild-typebaculovirus-infected Sf9 cell lysates (25 mg). (B) The reaction condi-tions using reticulocyte lysate have been described. Lanes: 1, purifiedp67-DG (2 mg); 2, hemin-deficient reticulocyte lysate; 3, hemin-sup-plemented reticulocyte lysate.

FIG. 11. (A) Detection of p67-DG polyclonal antibody-reactive ma-terials in the uninfected and baculovirus-infected insect cell extracts.Both uninfected and infected cell extracts were used as antigen instandard immunoblot experiment. Lanes: 1, 25 mg of the uninfected 12B). In both cases (Fig. 12A, lane 5; Fig. 12B, lane 3),cell extracts; 2, 25 mg of the infected cell extracts. (B) Detection however, the level of this immunoreactive material wasof p67-DG polyclonal antibody-reactive materials in vaccinia viral- significantly lower (Ç40%) when hemin was present.infected animal cells. KRC-7 (rat hepatoma) and L929 (mouse fibro-

Scanning by Molecular Dynamics Image Quant versionblast) cells were infected with 25 or 5 pfu per cell of vaccinia virus.3.3 showed following ratios: (Fig. 12A) lane 4, 1.0; laneAfter 4 h of infection, the cells were harvested and lysed. 400 mg

proteins in the cell extracts was incubated with p67-DG polyclonal 5, 0.6; (Fig. 12B) lane 2, 1.0; lane 3, 0.58.antibodies overnight at 47C. The antigen–antibody complexes werethen immunoprecipitated with protein A–agarose. The immunopre-

DISCUSSIONcipitates were then analyzed by SDS–PAGE followed by immu-noblotting with p67-DG polyclonal antibodies. Lanes: 1, mock-in- In this paper, we report efficient expression of p67fected KRC-7 cells; 2, KRC-7 cells infected with 5 pfu per cell of in insect cells using baculovirus expression system andvaccinia virus; 3, KRC-7 cells infected with 25 pfu per cell of vaccinia

this expression was enhanced severalfold by the addi-virus; 4, mock-infected L929 cells; 5, L929 cells infected with 5 pfuper cell of vaccinia virus. tion of hemin in the growth medium. This baculovirus

AID ABB 0139 / 6b38$$$204 05-17-97 01:22:28 arcal

381EXPRESSION OF p67 IN BACULOVIRUS-INFECTED INSECT CELLS

system has previously been used to express many N-and O-linked glycosylated proteins (15–17). However,several reports indicated that these glycosylations maybe similar but not identical to the native proteins. Itshould be noted that most of the expressed proteinsbear complex glycosyl residues. In contrast, the p67used in our studies contain only monosaccharides(GlcNAc). Our results indicate that the reactivity ofthe expressed protein to the p67 monoclonal antibodiesindicates that the expressed protein contains a GlcNAc-containing epitope similar to native p67. However, ourattempt to detect p67 in the uninfected insect cells wasunsuccessful. There are several possibilities: (i) p67may be present in a catalytic amount which is undetect-able by our standard experimental procedures. (ii) Ourantibodies may not detect the p67 present in the insectcells since these antibodies were raised in mouse.

The results presented in this paper suggest that ba-culovirus infection, like vaccinia viral infection (1), in-duces appearance of a p67-DG activity in the infected

FIG. 13. A proposed mechanism of regulation of p67-DG activitycells. The extracts from the uninfected insect cells didduring viral infection.not show any p67-DG activity (Fig. 3A). After baculovi-

rus infection, the extracts from the infected cells ac-tively deglycosylated exogenous p67 (Fig. 3B). Thisp67-DG activity was observed even when wild-type ba- lecular-weight protein (105 kDa) in the extracts fromculovirus (i.e., without p67 cDNA insert) was used (Fig. uninfected insect and several animal cells (KRC-7,3B). Addition of hemin to the growth medium partially L929, and reticulocytes). These antibodies also reacted

with an additional 60-kDa protein when the same in-prevented the appearance of this p67-DG activity. Thesect cells were infected with baculovirus and animalcells grown in the presence of hemin possessed signifi-cells (KRC-7 and L929) were infected with a vacciniacantly lower p67-DG activity (Fig. 3D). However, oncevirus. These antibodies also reacted with a nearly 60-activated in the absence of hemin inside the cells, thekDa protein in the reticulocyte lysate.activated p67-DG deglycosylated p67 even in the pres-

These results suggest that viral infection inducesence of hemin (Fig. 4). These results are in agreementappearance of a nearly 60-kDa p67-DG. The origin ofwith the results reported previously using reticulocytethis p67-DG is not clearly understood. Several possi-lysates (2). Chefalo et al. have also reported (19) thatbilities are:hemin induced overexpression of several coexpressed

proteins such as eIF-2 a and interleukin b. Presum- (i) A p67-DG activity is present in the animal cellsably, hemin prevented HRI inhibition of protein syn- in a precursor form and viral infection activates thisthesis in infected cells. In our experiments, we have precursor protein possibly by protease cleavage. Sev-overexpressed p67 in the presence of hemin. Increased eral cell extracts analyzed in this work showed thep67 synthesis may also increase coexpression of other presence of a high-molecular-weight protein immuno-proteins. reactive with p67-DG antibodies. Viral infection in

The hemin effect was further examined. We provide each case produced an additional low-molecular-weightevidence that hemin does not have any effect on p67 protein which was also immunoreactive with p67-DGtranscription. p67 mRNA level was the same inside the antibodies. This protein migrated similarly as p67-DG.cells grown in the presence or absence of hemin. The However, no precursor–product relationship was ap-results presented in this paper suggest that hemin ef- parent. These results may not support a hypothesisfect is posttranscriptional. that the high-molecular-weight protein is the precursor

We have now purified this p67-DG activity to near of p67-DG.homogeneity. Our results suggest that p67-DG activity (ii) p67-DG is host gene coded. This gene is activatedmay be nonspecific since this enzyme was purified us- during viral infection. We have previously reported theing a synthetic substrate p-nitrophenyl N-acetyl-b-D- presence of p67-DG activity in reticulocyte lysate. Itglucosaminide. However, our results also showed that may be that the stress conditions such as viral infectionp67-DG deglycosylates p67 both in vivo and in vitro. and heme deficiency induce expression of this hostPolyclonal antibodies has been raised against this puri- gene.

(iii) p67-DG is viral gene coded. This gene is acti-fied p67-DG. These antibodies reacted with a high-mo-

AID ABB 0139 / 6b38$$$204 05-17-97 01:22:28 arcal

382 SAHA ET AL.

2. Chakraborty, A., Saha, D., Bose, A., Chatterjee, M., and Gupta,vated by some cellular component immediately afterN. K. (1994). Biochemistry 33, 6700–6706.viral infection.

3. Wu, S., Gupta, S., Chatterjee, N., Hileman, R. E., Kinzy, T. G.,Denslow, N., Merrick, W. C., Chakrabarti, D., Osterman, J. C.,As mentioned above, we have now purified p67-DGand Gupta, N. K. (1993). J. Biol. Chem. 288, 10796–10801.

to near homogeneity. Using this protein, we are now 4. Datta, B., Chakrabarty, D., Roy, A. L., and Gupta, N. K. (1988).attempting to clone the p67-DG cDNA. The availability Proc. Natl. Acad. Sci. USA 85, 3324–3328.of p67-DG cDNA will enable us to answer the above 5. Datta, B., Ray, M. K., Chakrabarti, D., Wylie, D., and Gupta,

N. K. (1989). J. Biol. Chem. 264, 20620–20624.questions and establish the origin of p67-DG.6. Ray, M. K., Datta, B., Chakraborty, A., Chattopadhyay, A.,Another significant observation is that hemin addi-

Meza-Keuthen, S., and Gupta, N. K. (1992). Proc. Natl. Acad.tion reduces the p67-DG level (Ç40%) in baculovirus-Sci. USA 89, 539–543.infected insect cells and also in reticulocyte lysate

7. Ray, M. K., Chakraborty, A., Datta, B., Chattopadhyay, A., Saha,(Figs. 6A and 6B). In both baculovirus-infected insect D., Bose, A., Kinzy, T. G., Wu, S., Hileman, R. E., Merrick, W. C.,cells and also in reticulocyte lysate (2), hemin prevents and Gupta, N. K. (1993). Biochemistry 32, 5151–5159.

8. Trachsel, H., Ranu, R. N., and London, I. M. (1978). Proc. Natl.the activation of p67-DG. The mechanism of activationAcad. Sci. USA 75, 3654–3658.is not known and may involve covalent modifications

9. Luckow, V. A., and Summers, M. D. (1988). Bio/Technology 6,such as protein phosphorylation or glycosylation. The47–55.reduced level of p67-DG-inactive form is less immuno-

10. Saiki, R. K., Celfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R.,reactive with the p67-DG antibodies. These antibodies Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988). Science 233,were prepared against the active form of p67-DG. Our 487–491.suggested mechanism for regulation of p67-DG is 11. Kitts, P. A., Ayres, M. D., and Possee, R. D. (1990). Nucleic Acids

Res. 18, 5667–5672.shown in Fig. 13.12. Gupta, S., Wu, S., Chatterjee, N., Ilan, J., Ilan, J., Osterman,

J. C., and Gupta, N. K. (1995). Gene Expression 5, 113–122.13. Chomczynnski, P., and Sacchi, N. (1987). Anal. Biochem. 162,ACKNOWLEDGMENTS

156–159.This work was supported by NIGMS Grant GM22079, an American 14. Li, Y.-T., and Li, S.-C. (1972). Methods Enzymol. 28, 702–713.

Heart Association (Nebraska Chapter) grant, and a Nebraska State 15. O’Rielly, D. R., Miller, L. K., and Luckow, V. A. (1992). Baculovi-grant for Cancer and Smoking Diseases. We also thank Professor rus Expression Vectors: A Laboratory Manual, Freeman, NewJohn W. B. Hershey (University of California, Davis) for critical read- York.ing and many valuable suggestions during the preparation of the

16. Stiles, B., and Wood, H. A. (1983). Virology 131, 230–241.manuscript.17. Thomsen, D. R., Post, L. E., and Elhammer, A. P. (1990). J. Cell.

Biochem. 43, 67–79.18. Hart, G. W., Haltwinger, R. S., Holt, G. D., and Kelly, W. G.REFERENCES

(1989). Annu. Rev. Biochem. 58, 841–874.19. Chefalo, P. J., Yang, J. M., Ramaiah, K. V. A., Gehrke, L., and1. Bose, A., Gupta, S., Saha, D., Chatterjee, N., and Gupta, N. K.

(1997). Arch. Biochem. Biophys. 342, 362–372. Chen, J. J. (1994). J. Biol. Chem. 269, 25788–25794.

AID ABB 0139 / 6b38$$$204 05-17-97 01:22:28 arcal