Partial redistribution of the Autographa californica ... - CiteSeerX

Journal of General Virology (2002), 83, 685–694. Printed in Great Britain ...................................................................................................................................................................................................................................................................................

Partial redistribution of the Autographa californica nucleopolyhedrovirus chitinase in virus-infected cells accompanies mutation of the carboxy-terminal KDEL ER-retention motif Giles P. Saville,1 Carole J. Thomas,1† Robert D. Possee2 and Linda A. King1 1 2

School of Biological and Molecular Sciences, Gipsy Lane Campus, Oxford Brookes University, Oxford OX3 0BP, UK NERC Institute of Virology and Environmental Microbiology, Mansfield Road, Oxford OX1 3SR, UK

During virus infection of insect cells, the Autographa californica nucleopolyhedrovirus chitinase is localized primarily within the endoplasmic reticulum (ER), which is consistent with the presence of a carboxy-terminal ER retention motif (KDEL). Release of chitinase into the extracellular medium appears to be concomitant with terminal cell lysis, rather than by active secretion. In this study, we have shown that mutation of the KDEL motif induces a partial redistribution of the chitinase at both early and late times post-infection. Deletion of the KDEL motif or substitution with glycine residues allowed chitinase to move through the secretory pathway, accumulating to detectable levels in the extracellular medium by 24 h post-infection ; more than 48 h prior to cell lysis. Deletion of the KDEL motif did not compromise enzyme activity, with the modified enzyme exhibiting characteristic endo- and exo-chitinolytic activity. Trichoplusia ni larvae infected with the modified virus were found to liquefy approximately 24 h earlier than larvae infected with a control virus in which the chitinase KDEL motif had not been deleted.

Introduction Acute infection of larvae with the Autographa californica nucleopolyhedrovirus (AcMNPV) induces a characteristic terminal liquefaction of all insect tissues (Slack et al., 1995 ; Hawtin et al., 1995). This process has been attributed to the synergistic action of two virus-encoded gene products, viral cathepsin (v-cath ; Rawlings et al., 1992 ; Slack et al., 1995) and chitinase (chiA ; Hawtin et al., 1995). Deletion of either gene abrogates the liquefying process (Slack et al., 1995 ; Hawtin et al., 1997), whereas dual infection of larvae with two single knock-out mutant viruses causes complementation resulting in restored liquefaction (Hawtin et al., 1997 ; Thomas et al., 1998). Chitinase and cathepsin have been found in several, but not all, baculovirus genomes including Bombyx mori nucleopolyhedrovirus (Bm)NPV (Ohkawa et al., 1994 ; Author for correspondence : Linda King. Fax j44 1865 483242. e-mail laking!brookes.ac.uk † Present address : Royal Veterinary College, Hawkshead Lane, North Mymms, Hatfield AL9 7TA, UK.

0001-8040 # 2002 SGM

Gomi et al., 1999), Spodoptera exigua (Se)NPV (Dai et al., 2000) and Lymantria dispar (Ld)NPV (Kuzio et al., 1999). Evidence to date indicates that baculoviruses which express chitinase and cathepsin cause terminal host liquefaction, resulting in the complete loss of integrity of host tissues permitting mature polyhedra to escape into the environment and promoting horizontal virus transmission. In this manner, the dissemination of progeny virus as a result of host liquefaction is achieved. The genome of AcMNPV has been shown to contain a single chiA (Ayres et al., 1994), encoding an enzyme that has been shown to possess an unusual endo- and exo-chitinolytic activity (Hawtin et al., 1995 ; Thomas et al., 2000). The chiA of AcMNPV is located upstream of lef-7, and in a back-to-back orientation with v-cath, with their respective promoters transcribing in opposing directions (Ayres et al., 1994 ; Hawtin et al., 1995 ; Slack et al., 1995). Both chiA and v-cath are transcribed from late baculovirus promoters (TAAG). It has been postulated that the single chiA of AcMNPV may have originated via gene transfer from the bacterium Serratia marcescens, which inhabits the insect gut (Hawtin et al., 1995). Although there is no phylogenetic evidence of such horizontal GIF

G. P. Saville and others

gene transfer (Kang et al., 1998), the bacterial and viral chitinases share a high level of amino acid identity (60n5 %) and the virus chiA has a GjC ratio that is more similar to that of the bacterium than to the rest of the AcMNPV genome (Hawtin et al., 1995). In a previous study we have shown that chitinase enters the secretory pathway of insect cells following cleavage of the amino-terminal signal peptide (Thomas et al., 1998). Chitinase is then retained within the endoplasmic reticulum (ER) until terminal lysis, when it can be detected in the culture medium (Thomas et al., 1998). Localization of chitinase within the ER was found to be consistent with the presence of a carboxyterminal KDEL ER-retention motif (Thomas et al., 1998). This motif is commonly found in proteins resident in the ER and consists of the tetrapeptide lysine–aspartic acid–glutamic acid–leucine (Lewis & Pelham, 1992). KDEL has been shown to function as an ER-retention motif in both animal and plant cells and acts as a signal to return proteins that have moved into the Golgi back to the ER, via the retrograde pathway (Lewis & Pelham, 1992 ; Hawes et al., 1999 ; Toyooka et al., 2000). The motif has been identified as a conserved region in other baculovirus chitinases including those of Orgyia pseudotsugata (Op)NPV and Choristoneura fumiferana Cf(NPV), and as RDEL in BmNPV (Thomas et al., 1998). The functional nature of these motifs in viral chitinases has yet to be elucidated although we have proposed that retention of chitinase in the ER may permit the virus to complete its replication cycle before cellular and tissue disintegration are induced. In this study we have investigated the biological significance of the chitinase KDEL motif by producing a number of modified viruses in which the motif has been deleted or the amino acids substituted. The effects on the intracellular distribution of chitinase and its biological effect on insect larvae have been investigated.

Methods

Production of plasmid transfer vectors. The transfer vector pAcpolhchiA contained a copy of the entire chiA under control of the polh promoter, and was produced by amplification of the chiA coding sequence by PCR using the forward primer 5h ATAAGGATCCATGTTGTACAAATTGTTAAACG 3h and the reverse primer 5h TACAGGGATCCTTACAGTTCATCTTTAGGTTTAAACTG 3h. The amplified product of 1543 bp contained BamHI restriction sites (bold) at either end to facilitate cloning into the standard baculovirus transfer vector pBacPAK9 (Clontech) to produce pAcpolhchiA, which was subsequently sequenced to confirm the integrity of the amplified chiA DNA. Further transfer vectors containing chiA with deletions or substitutions of the KDEL coding sequence were also produced. ChiA with a deletion of the coding sequence for KDEL was prepared using the following primers 5h ATAAGGATCCATGTTGTACAAATTGTTAAACG 3h (forward) and 5h TACAGGGATCCTTAAGGTTTAAACTGTCGTTTATCGC 3h (reverse) to produce pAcpolchiA∆KDEL. ChiA containing a deletion of the sequence encoding KDEL plus a further five amino acid residues was produced using the primers 5h ATAAGGATCCATGTTGTACAAATTGTTAAACG 3h (forward) and 5h GGTTGIG

TGGATCCTTAGTTTATCGCGTTGAGCAAGTCG 3h (reverse) to produce pAcpolchiA∆KDEL+&. Finally, a chiA containing a substitution of the KDEL coding sequence with four non-polar glycine residues was prepared using primers 5h ATAAGGATCCATGTTGTACAAATTGTTAAACG 3h (forward) and 5h TACAGGGATCCTTAGGAGGCGGGGGAGTTTATCGC 3h (reverse) to produce the transfer vector pAcpolhchiA∆KDEL/%G. All transfer vectors were sequenced to confirm the integrity of the respective deletions or substitution of the KDEL motif.

Virus propagation and insect cell culture. Recombinant viruses were routinely propagated in Spodoptera frugiperda (Sf9) cells maintained in Sf900II serum-free medium (King & Possee, 1992). Plaque-assays containing viruses with the lacZ coding region were identified by addition of 2 % X-Gal to the culture medium.

Generation of recombinant baculoviruses AcpolhchiA, AcpolhchiA∆KDEL, AcpolhchiA∆KDELM5 and AcpolhchiA∆KDEL/4G. Recombinant polyhedrin-negative AcMNPV were produced that contained a normal or modified chiA under control of the strong, very late polyhedrin gene promoter. In all of the modified viruses, the natural chiA had been disrupted by insertional inactivation with lacZ (Thomas et al., 1998). Sf9 cells (1i10') were co-transfected with pAcpolhchiA, pAcpolhchiA∆KDEL, pAcpolhchiA∆KDEL+& or pAcpolhchiA∆KDEL/%G (500 ng) and AcchiA−.lacZ virus DNA (Thomas et al., 1998). Progeny virus was titrated by plaque-assay and plaques were picked after staining with X-Gal and neutral red to identify the polyhedrin-negative, lacZpositive plaques and were subjected to plaque-purification to produce seed stocks of the recombinant viruses : AcpolhchiA, AcpolhchiA∆KDEL, AcpolhchiA∆KDEL+& and AcpolhchiA∆KDEL/%G. Recombinant viruses were characterized initially by examining virus-infected cells for the presence of chitinase by Western blot analysis and enzyme assay (described below). High-titre stocks of recombinant viruses were prepared in Sf9 cells as described in King & Possee (1992).

SDS–PAGE and Western blotting. Sf9 cells (1i10') were infected with either recombinant virus or AcMNPV at an m.o.i. of 5 p.f.u. per cell, or were mock-infected with culture medium. Cells were harvested at 48 h post-infection (p.i.), pelleted (3000 g) and the culture medium removed to a fresh tube. Cell pellets were washed three times in PBS and subjected to freeze–thaw lysis before analysis. SDS–PAGE and Western blot analysis of cell pellets or cell culture medium were carried out as described previously (King & Possee, 1992 ; Thomas et al., 1998). Membranes were probed with primary anti-chitinase antiserum (1 : 10 000), and a secondary anti-guinea pig IgG antibody conjugated to alkaline phosphatase (1 : 10 000), before blots were developed with 5bromo-4-chloro-3-indoyl phosphatase (BCIP) and nitro blue tetrazolium (NBT) as described by Thomas et al. (1998).

Enzyme assays. Sf9 cells (1i10' cells\ml) were infected with virus at an m.o.i. of 5 p.f.u. per cell or were mock-infected and harvested at appropriate times p.i. Cells were pelleted and subjected to three rounds of freeze–thaw lysis before assaying for chitinase activity using the microtitre plate method of McCreath & Gooday (1992), as described by Thomas et al. (1998), permitting the levels of exo-chitinase, endochitinase and N-acetylglucosamidase activity to be determined. The release of a fluorescent aglycone as a result of enzymatic activity in the presence of 4-methylumbelliferylglycosides of N-acetylglucosamine oligosaccharides (4MU-[GlcNAc] – ) was read on a Labsystems Fluoro"% skan II fluorimeter. The fluorescent aglycone is released in the presence of the following enzymatic activities : N-acetylglucosamidase (substrate 1, 4MU-GlcNAc), exo-chitinase (substrate 2, 4MU-[GlcNAc] ) and endo# chitinase (substrate 3, 4MU-[GlcNAc] ). $

Localization of AvNPV chitinase

Fig. 1. (a) Coding sequence of the AcMNPV chitinase gene illustrating translational start (ATG) and stop sites (TAA). (b) Carboxy-terminal modifications of AcMNPV chiA (1), in recombinant viruses (2) AcpolhchiA∆KDEL, (3) AcpolhchiA∆KDEL+5 and (4) AcpolhchiA∆KDEL/4G are illustrated.

Virus-infected cells were assayed for cathepsin activity using a cysteine protease assay (Ohkawa et al., 1994). Cells (3i10() were infected with virus at an m.o.i. of 10 p.f.u. per cell or were mock-infected and harvested at 40 h p.i. Prior to analysis cell lysates were subjected to three rounds of freeze–thaw and assayed as described by Hawtin et al. (1997).

Confocal microscopy. Confocal laser scanning microscopy (CLSM) was used to examine the distribution of chitinase in virusinfected Sf9 cells. Sterile coverslips (13 mm) were seeded with cells and infected with recombinant virus (m.o.i. l 5 p.f.u. per cell) or were mockinfected with medium. Cells were incubated at 28 mC and after harvesting were fixed with acetone–methanol (50 : 50) before immunostaining for chitinase with primary anti-chitinase antiserum (1 : 1000 dilution), and a secondary FITC-conjugated whole-molecule anti-guinea pig IgG antibody (Sigma) (1 : 64 dilution) as described by Thomas et al. (1998). Samples were mounted with Citifluor antifadant and examined on a Zeiss CLSM (Axiovert 100) using appropriate filter sets.

Virus infection of insect larvae. Third instar Trichoplusia ni larvae (n l 50) were infected with virus at a dose of 1i10% p.f.u., via microinjection into the haemolymph at the final posterior proleg. Control larvae were inoculated with sterile PBS in place of virus inoculum. Larvae were subsequently housed in individual containers at 28 mC and were maintained on sterile semi-synthetic diet (Hunter et al., 1984). Those larvae that died as a result of the micro-injection were discarded. Larvae were monitored at frequent intervals (6–8 h) for signs of virus infection and liquefaction.

Results Production of recombinant baculoviruses

The AcMNPV chiA was genetically modified by deletion or substitution of the sequences encoding the carboxy-terminal KDEL motif. In order to generate recombinant viruses GIH


efficiently, the modified chiA was inserted into a standard baculovirus transfer vector, pBacPAK9. Following co-transfection of Sf9 cells with parental virus DNA, modified chiA genes were transferred to the polh locus under control of the very late polh promoter. The parental virus used in the cotransfection was AcMNPV in which the chiA, at its normal locus, had been interrupted with the lacZ coding region (AcchiA−.lacZ ; Thomas et al., 1998), so that the recombinant viruses produced in this study expressed either wild-type or mutant chitinase from the polh promoter only. Previous studies have shown that disruption of chiA with lacZ abrogates all chitinolytic activity (Thomas et al., 1998). Recombinant viruses were generated in which the chiA contained a deletion in the sequences encoding the carboxy-terminal KDEL motif (AcpolhchiA∆KDEL), a KDEL deletion with an additional deletion of five further amino acid residues (AcpolhchiA∆KDEL+&), and a substitution of the KDEL motif with four non-polar glycine residues (AcpolhchiA∆KDEL/%G) (Fig. 1). A further virus with the entire chiA coding sequence, including KDEL, under the polyhedrin promoter was also generated (AcpolhchiA) to act as a control virus in all experiments (Fig. 1). Assessment of chitinase protein production by AcpolhchiA, AcpolhchiA∆KDEL, AcpolhchiA∆KDELM5 and AcpolhchiA∆KDEL/4G

Cells were either mock-infected or infected with recombinant virus (m.o.i. l 5 p.f.u. per cell). At 48 h p.i. the cell pellets and cell culture medium were analysed for chitinase by Western blot analysis. Fig. 2 illustrates that a protein of approximately 58 kDa was identified in the cell pellet of all virus infections at 48 h p.i. but was not present in mockinfected cells. On examination of the cell culture medium, a specific band at 58 kDa was observed in AcpolhchiA∆KDEL-, AcpolhchiA∆KDEL+&- and AcpolhchiA∆KDEL/%G-infected cells that was not present in AcpolhchiA-, AcMNPV- or mockinfected cells (Fig. 2). The appearance of chitinase in the cell culture medium at 48 h p.i., prior to cell lysis, was consistent with the deletion or substitution of the KDEL motif in these viruses. The greatest accumulation of chitinase in the culture medium was observed in AcpolhchiA∆KDEL-infected cells (Fig. 2, lane KD) and this virus was examined in more detail. A more detailed analysis of the temporal production of chitinase in AcMNPV-, AcpolhchiA- and AcpolhchiA∆KDELinfected Sf9 cells was carried out. Fig. 3 illustrates that chitinase was identified in the cell pellets of all virus infections, from 24 h p.i. in AcMNPV-infected samples (panel a), and from as early as 12 h p.i. in AcpolhchiA- (panel c) and AcpolhchiA∆KDELinfected cells (panel e). Estimated levels of chitinase synthesis in virus-infected cell pellets were comparable for all viruses. Upon examination of the culture medium, no specific band was identified in the AcMNPV-infected samples (Fig. 3 ; panel b), and chitinase was only detected in the culture medium of GII

Fig. 2. Detection of chitinase in virus-infected Sf9 cells. Virus-infected cell pellets (p) and cell culture medium (c) were assayed at 48 h p.i. for chitinase by SDS–PAGE and Western blot analysis. Cells were infected with virus at an m.o.i. of 5 p.f.u. per cell with AcMNPV (WT), AcpolhchiA (polh), AcpolhchiA∆KDEL+5 (KDj5), AcpolhchiA∆KDEL/4G (KD/4G), AcpolhchiA∆KDEL (KD) or were mock-infected (M). Samples (1 ml) were harvested at 48 h p.i. and pelleted (3000 g, 5 min). Cell pellets were resuspended in 50 µl of sterile PBS and subjected to three rounds of freeze–thaw lysis, prior to 15 µl of cell pellet or cell culture medium being loaded onto an SDS–PAGE gel. Membranes were probed with primary anti-chitinase antiserum (1 : 10 000) and secondary alkaline phosphataseconjugated antibody (1 : 10 000). Blots were developed with BCIP and NBT. Protein size markers are in kDa.

AcpolhchiA-infected (panel d) samples at 96 h p.i., which was concomitant with cellular lysis as detected by visual examination of cells under the light microscope. These results were consistent with the presence of the chiA KDEL motif in both viruses. Accumulation of chitinase was identified in the culture medium of AcpolhchiA∆KDEL-infected cells from as early as 12 h p.i. (panel f). These results indicated that deletion of the chiA KDEL motif was sufficient to cause a partial redistribution of the intracellular localization of chitinase, resulting in some chitinase trafficking through the secretory pathway to the culture medium of infected cells. Visual examination of infected cells under the light microscope indicated that cell lysis could not account for the appearance of chitinase in the culture medium at the early time points of 12–72 h p.i. (data not shown).

Cysteine protease activity in AcMNPV-, AcpolhchiA-, AcpolhchiA∆KDEL- and mock-infected cells

Virus-infected cells were assayed for cysteine protease activity to confirm the integrity of v-cath in AcpolhchiA- and AcpolhchiA∆KDEL-infected cells using the assay of Ohkawa et al. (1994). Cells (Sf9) were infected with recombinant virus (m.o.i. l 10 p.f.u. per cell) or were mock-infected and harvested at 40 h p.i. Samples were assayed in the presence or absence of the specific cysteine protease inhibitor E64 (Slack et al., 1995). The levels of cysteine protease activity associated with AcpolhchiA- and AcpolhchiA∆KDEL-infected cells were


Fig. 3. Temporal synthesis of chitinase in AcMNPV-, AcpolhchiA- and AcpolhchiA∆KDEL-infected Sf9 cells. Virus-infected (m.o.i. l 5 p.f.u. per cell) cell pellets (p) and cell culture medium (c) were analysed for chitinase by SDS–PAGE and Western blot analysis. Samples (1 ml) were harvested at the time-points indicated and cells pelleted by centrifugation (3000 g, 5 min). Cell pellets were resuspended in 1 ml of sterile PBS and subjected to three rounds of freeze–thaw lysis. Samples (15 µl) of cell pellet and cell culture medium were then analysed on a 10 % SDS–PAGE gel. AcMNPV (a), AcpolhchiA (c) and AcpolhchiA∆KDEL (e) cell pellets, and AcMNPV (b), AcpolhchiA (d) and AcpolhchiA∆KDEL (f) cell culture medium were probed with primary antichitinase antibody (1 : 10 000), and secondary alkaline phosphatase-conjugated antibody (1 : 10 000). Blots were developed with BCIP and NBT. Protein size markers are in kDa. A positive control (jve) of AcMNPV-infected cell pellet at 48 h p.i. is shown on the right of each gel.

Fig. 4. Cysteine protease activity of recombinant viruses. Cells (Sf9) were infected with AcMNPV, AcpolhchiA or AcpolhchiA∆KDEL at an m.o.i. of 10 p.f.u. per cell or were mock-infected. Cells (1i107) were harvested at 40 h p.i., pelleted, washed in PBS and assayed in the presence and absence of the cysteine protease inhibitor E64. Data are from three individual experiments.

shown to be comparable to those found in AcMNPV-infected cells (Fig. 4). All samples showed a significant reduction of cysteine protease activity when assayed in the presence of E64, confirming the integrity of v-cath in the recombinant viruses.

Mock-infected cells exhibited low-levels of cysteine protease activity, which were reduced only slightly in the presence of E64, which was in agreement with previous results (Slack et al., 1995 ; Hawtin et al., 1997 ; Thomas et al., 1998). GIJ


Fig. 5. N-Acetylglucosamidase, exo- and endo-chitinolytic activity in (a) mock-, (b) AcMNPV-, (c) AcpolhchiA- and (d) AcpolhchiA∆KDEL-infected Sf9 cells. Cells (1i106) were infected with virus at an m.o.i. of 5 p.f.u. per cell and harvested at the time-points illustrated. Cells were pelleted, washed in PBS and subjected to three rounds of freeze–thaw lysis prior to assaying for chitinase by microtitre plate assay. Assays were read on a Labsystems Fluoroskan II fluorimeter. Data shown are from three individual experiments.

Detection of chitinolytic activity in AcMNPV-, AcpolhchiA- and AcpolhchiA∆KDEL-infected Sf9 cells

The chitinolytic activity exhibited during the infection of Sf9 cells with recombinant virus was assayed using a microtitre plate assay. This assay permitted the levels of N-acetylglucosamidase (substrate 1), exo-chitinase (substrate 2) and endo-chitinase (substrate 3) activity to be determined. Low levels of N-acetylglucosamidase activity were detected in both mock- and virus-infected cells, which remained constant throughout virus infection (Fig. 5). The levels of N-acetylglucosamidase activity associated with AcpolhchiA∆KDEL infection were similar to those found during GJA

infection with the control viruses AcMNPV and AcpolhchiA, but were lower than those of mock-infected cells (Fig. 5). NAcetylglucosamidase activity was detected from 0 h p.i., indicating that it was unlikely to be associated with virus infection and may be attributed to host cell enzymatic activity. In particular, virus-infected samples showed reduced levels of this activity from the onset of virus infection (0 h p.i. l 1 h p.i. with viral inoculum), and this may have been a result of immediate early viral gene expression suppressing the levels of host cell enzyme activity. In AcMNPV-infected cells, high levels of exo- and endochitinase activity were observed (Fig. 5b), with endo-chitinolytic activity found to be higher throughout the time-course of


AcpolhchiA∆KDEL

AcpolhchiA (a)

(d)

(b)

(e)

(c)

(f)

24 h

48 h

72 h

Fig. 6. Confocal laser scanning microscopy of AcpolhchiA- (panels a–c) and AcpolhchiA∆KDEL- (panels d–f) infected Sf9 cells. Cells were infected with virus and harvested at 24 (a, d), 48 (b, e) and 72 (c, f) h p.i. prior to fixation with acetone–methanol (50 : 50). Samples were immunostained with anti-chitinase antibody (1 : 1000) and FITC-conjugated secondary antibody (1 : 64) before mounting on Citifluor antifadant. Samples were examined on a Carl Zeiss Axiovert 100 CLSM using appropriate filter sets. Bar, 5 µm.

infection studied. This was consistent with previous studies (Hawtin et al., 1995 ; Thomas et al., 1998). In the control virus, AcpolhchiA (Fig. 5c), levels of exo- and endo-chitinase activity rose sharply at 21 h p.i., reaching a plateau at 30 h p.i. Levels of exo- and endo-chitinase were comparable until 30 h p.i., with higher detectable levels of endo-chitinase observed from

36 h p.i. Levels of exo- and endo-chitinase activity detected in cells infected with AcpolhchiA∆KDEL (Fig. 5d) were slightly lower than those associated with AcpolhchiA infection (Fig. 5c). Increasing levels of activity were observed from 12 h p.i. (Fig. 5d), with higher levels of endo- rather than exo-chitinase activity ; maximum levels of activity were detected at 24 h p.i. GJB


Fig. 7. Effects of recombinant virus infection in T. ni larvae. Cohorts of third instar larvae (n l 50) were infected with AcMNPV (WT), AcpolhchiA (P) or AcpolhchiA∆KDEL (∆KD), with 1i104 p.f.u. via micro-injection, or were mock-infected (M). Larvae were maintained on sterile semi-synthetic diet at 28 mC. Larval condition was monitored on a daily basis and larval death and liquefaction recorded. Healthy larvae typically pupated at 8 days p.i. Mean larval numbers from duplicated experiments are given as percentages ; raw data were used to calculate significance levels in a single factor ANOVA analysis (given in text).

These data indicated that modification of chiA, and its placement under the control of the polh promoter, had not affected the characteristic exo- and endo-chitinolytic activity of AcMNPV chiA. Furthermore, we were unable to prove that the chitinase accumulated in the culture medium of AcpolhchiA∆KDEL-infected cells (as demonstrated by Western blot) contained active enzyme, as the culture medium was found to produce a very high autofluorescence which masked any enzyme-induced fluorescence. Localization of chitinase and modified chitinase in Sf9 cells infected with AcpolhchiA and AcpolhchiA∆KDEL using CLSM

The localization of chitinase and modified chitinase in virusinfected cells was examined by CLSM. Virus-infected cells were harvested at 24, 48 and 72 h p.i. and immunostained with anti-chitinase polyclonal antibody and a secondary FITCconjugated antibody as described in Methods and in Thomas et al. (1998). The distribution of chitinase staining in AcpolhchiA- and AcpolhchiA∆KDEL-infected cells (Fig. 6a–f) was perinuclear, which was not present in mock-infected cells (data not shown) as has been previously reported for AcMNPV-infected cells (Thomas et al., 1998). Chitinase staining was observed from 24 h p.i. (panel a) ; increasing in intensity until 72 h p.i. (panel c) A more intense level of chitinase staining was observed in AcpolhchiA-infected cells (panels a–c) than in cells infected with AcpolhchiA∆KDEL (panels d–f). The heavily stained perinuclear area associated with AcpolhchiA-infected cells exhibited highly distended regions of ER, with vacuolar-like areas containing high levels of chitinase ; these distended regions were markedly less apparent during AcpolhchiA∆KDEL infection (Fig. 6, compare GJC

panels b and e). The lower level of chitinase staining associated with AcpolhchiA∆KDEL-infected cells was consistent with the movement of chitinase from the ER into the secretory pathway of the cell, which was in agreement with the deletion of KDEL from the chiA of AcpolhchiA∆KDEL. Effects of AcpolhchiA and AcpolhchiA∆KDEL infection upon survival and liquefaction of T. ni larvae

The effect of recombinant virus infection upon T. ni was examined to confirm that the characteristic liquefaction associated with virus infection had not been compromised as a result of the manipulation of chiA and its overexpression from the polyhedrin gene promoter. Third instar larvae were infected with 1 µl of recombinant virus at 1i10% p.f.u.\µl or were mock-infected. Those larvae infected with AcpolhchiA or AcpolhchiA∆KDEL exhibited a similar phenotype to larvae infected with AcMNPV. A whitening of the larval cuticle was observed from 4 days p.i., followed by characteristic liquefaction from 5–8 days p.i. No signs of virus infection or subsequent liquefaction were observed in mock-infected larvae. Healthy larvae typically pupated from 8 days p.i. Furthermore, the efficacy of the recombinant viruses was compared to AcMNPV- and mock-infected larvae. Fifty individual larvae (third instar) were infected with AcpolhchiA, AcpolhchiA∆KDEL or AcMNPV, or were mock-infected as previously described. The number of individual larvae which succumbed to virus infection and liquefied was recorded (Fig. 7). A significant difference in the number of larvae which had succumbed to virus infection was observed from 5 days p.i. When virus-infected and mock-infected individuals were compared, a significant difference (F l 265n3 ; P 0n001) in the number of larval deaths was observed using a single factor


analysis of variance (ANOVA). At this time-point the virus AcpolhchiA∆KDEL had killed a significantly greater number of larvae (60 %) than either AcMNPV infection (30 %) or AcpolhchiA infection (24 %). At 7 days p.i., the number of dead larvae was shown to be significantly higher for AcpolhchiA∆KDEL-infected individuals (60 %) than AcpolhchiA(40 %) or AcMNPV- (40 %) infected larvae (F l 137n6 ; P 0n001). Larval death and liquefaction was observed to occur at a greater frequency in AcpolhchiA∆KDEL-infected individuals from 5 days p.i., with all infected larvae liquefying by 7 days p.i. Therefore, larval death and liquefaction in AcpolhchiA∆KDEL-infected larvae occurred up to 24 h earlier than for AcpolhchiA-infected larvae, and killed a significantly greater number of individuals than either AcpolhchiA or AcMNPV between 5 and 7 days p.i. (F l 137n6 ; P 0n001) These data offered evidence that the virus AcpolhchiA∆KDEL had a greater killing efficacy for T. ni larvae than either AcpolhchiA or wild-type AcMNPV.

Discussion This paper describes the genetic modification of the sequence encoding the KDEL motif identified at the carboxy terminus of the AcMNPV chiA (Thomas et al., 1998). The KDEL motif is considered to be an ER-retention signal which localizes chitinase to the ER of infected cells until the terminal stages of baculovirus infection. This conserved region was modified by substitution of KDEL to GGGG (AcpolhchiA∆KDEL/%G) or by deletion (AcpolhchiA∆KDEL, AcpolhchiA∆KDEL+&), and the modified gene was placed under the control of the polh promoter at the polyhedrin gene locus in recombinant baculoviruses. This produced non-occluded recombinant virus expressing mutant chitinase from the polyhedrin locus. Accordingly, the virus AcpolhchiA was constructed which expressed an intact chiA from the polyhedrin promoter to act as a control virus. The effects of these carboxy-terminal mutations of chitinase upon the localization and biological activity of the enzyme were assessed as well as the in vivo efficacy of the viruses to T. ni. Chitinase protein synthesis was detected in the intracellular fraction in all recombinant virus infections, but was not present in mock-infected samples. The accumulation of chitinase in the culture medium of cells infected with AcpolhchiA∆KDEL/%G, AcpolhchiA∆KDEL and AcpolhchiA∆KDEL+&, at a time-point (48 h p.i.) prior to cellular lysis, indicated that modification of KDEL was sufficient to permit trafficking of the enzyme through the secretory pathway of virus-infected cells. Chitinase was not observed in the culture medium of AcpolhchiAinfected cells at 48 h p.i. A more detailed time-course of infection indicated that chitinase was only detectable in the culture medium of AcpolhchiA-infected cells at a time-point relating directly to cell lysis (72–96 h p.i). The accumulation of chitinase in the culture medium of the AcpolhchiA∆KDELinfected cells was detected from 12 h p.i., and confirmed that

deletion of KDEL was sufficient to permit chitinase to enter the secretory pathway from the ER, and to pass to the extracellular fraction of the infected cell. The biological activity of chitinase was assessed by a microtitre plate assay (McCreath & Gooday, 1992), which could differentiate between the exo- and endo-chitinolytic activities of the enzyme. The intracellular activity of exo- and endo-chitinase in cells infected with the control virus AcpolhchiA was higher than that associated with the modified virus AcpolhchiA∆KDEL. We suggest that this is a result of the retention of the enzyme in the intracellular fraction of the infected cell, while the lower levels detected in AcpolhchiA∆KDEL infection are a result of the secretion of chitinase to the extracellular fraction during infection. The higher levels of chitinolytic activity in AcpolhchiA-infected cells than AcMNPV are likely to be a result of the stronger polh promoter driving expression of chitinase in AcpolhchiA. The distribution of chitinase observed in infected cells by CLSM indicated that the deletion of KDEL, and accumulation of chitinase in the culture medium of the cell, were sufficient to reduce the heavy ER-associated staining observed in AcpolhchiA infection. This was reflected in the more uniform perinuclear distribution of chitinase in AcpolhchiA∆KDEL infection, with reduced areas of heavily stained and distended ER, and was consistent with the lower levels of intracellular chitinolytic activity associated with AcpolhchiA∆KDEL infection. The effects of the modified viruses with chiA under the control of the polh promoter were also examined in vivo in T. ni larvae. The phenotype of recombinant virus-infected larvae was shown to be similar to that associated with AcMNPV infection, with characteristic liquefaction observed at the terminal stage of virus infection. This indicated that the expression of chiA from the polh promoter and the modification of KDEL did not abrogate the liquefaction of infected larvae, and was in agreement with the retained dual enzymatic activity of exo- and endo-chitinase in the recombinants. The recombinant virus AcpolhchiA∆KDEL was shown to cause a significantly greater number of the larval cohort examined to succumb to viral infection, and resulted in total larval liquefaction of infected individuals by 7 days p.i. This virus caused a lethal infection in a greater number of individuals than AcMNPV or AcpolhchiA, and induced larval death and liquefaction up to 24 h earlier than the control virus AcpolhchiA. These data indicate that the modification of the AcMNPV chiA by manipulation of the carboxy-terminal KDEL motif is sufficient to alter the localization of the enzyme during baculovirus infection. Whilst the dual chitinolytic activity of the enzyme was not compromised, the distribution of chitinase in vitro was altered and the biological activity to T. ni larvae was enhanced. This offers evidence of the functional nature of the AcMNPV chiA KDEL motif as an ER-retention signal, and whilst the partial redistribution of chitinase as a result of a GJD


KDEL manipulation occurs, the potential effects both in vitro and in vivo warrant further investigation. We thank Sue Mann for assistance with cell culture. These studies were funded by the Perry Foundation and the BBSRC Bio-imaging Initiative.

References Ayres, M. D., Howard, S. C., Kuzio, J., Lopez-Ferber, M. & Possee, R. D. (1994). The complete DNA sequence of Autographa californica

nuclear polyhedrosis virus. Virology 202, 586–605. Dai, X., Hajos, J. P., Joosten, N. N., van Oers, M. M., IJkel, W. F. J., Zuidema, D., Pang, Y. & Vlak, J. M. (2000). Isolation of a Spodoptera

exigua baculovirus recombinant with a 10n6 kbp deletion that retains biological activity. Journal of General Virology 81, 2545–2554. Gomi, S., Majima, K. & Maeda, S. (1999). Sequence analysis of the genome of Bombyx mori nucleopolyhedrovirus. Journal of General Virology 80, 1323–1337. Hawes, C. R., Brandizzi, F. & Andreeva, A. V. (1999). Endomembranes and vesicle trafficking. Current Opinion in Plant Biology 2, 454–461. Hawtin, R. E., Arnold, K., Ayres, M. D., Zanotto, P. M. D. A., Howard, S. C., Gooday, G. W., Chappell, L. H., Kitts, P. A., King, L. A. & Possee, R. D. (1995). Identification and preliminary characterization of a chitinase

gene in the Autographa californica nuclear polyhedrosis virus genome. Virology 212, 673–685.

King, L. A. & Possee, R. D. (1992). The Baculovirus Expression System. A

Laboratory Guide. London : Chapman and Hall. Kuzio, J., Pearson, M. N., Harwood, S. H., Funk, C. J., Evans, J. T., Slavicek, J. M. & Rohrmann, G. F. (1999). Sequence and analysis of the

genome of a baculovirus pathogenic for Lymantria dispar. Virology 253, 17–34. Lewis, M. J. & Pelham, H. R. B. (1992). Ligand-induced redistribution of a human KDEL receptor from the Golgi complex to the endoplasmic reticulum. Cell 68, 353–364. McCreath, K. J & Gooday, G. W. (1992). A rapid and sensitive microassay for determination of chitinolytic activity. Journal of Microbiological Methods 14, pp 229–237. Ohkawa, T., Majima, K. & Maeda, S. (1994). A cysteine protease encoded by the baculovirus Bombyx mori nuclear polyhedrosis virus. Journal of Virology 68, 6619–6625. Rawlings, N. D., Pearl, L. H. & Buttle, D. J. (1992). The baculovirus Autographa californica nuclear polyhedrosis virus genome includes a papain-like sequence. Biology and Chemistry 373, 1211–1215. Slack, J. M., Kuzio, J. & Faulkner, P. (1995). Characterization of v-cath, a cathepsin L-like proteinase expressed by the baculovirus Autographa californica multiple nuclear polyhedrosis virus. Journal of General Virology 76, 1091–1098. Thomas, C. J., Brown, H. L., Hawes, C. R., Lee, B. Y., Min, M. K., King, L. A. & Possee, R. D. (1998). Localization of a baculovirus-induced

chitinase in the insect cell endoplasmic reticulum. Journal of Virology 72, 10207–10212.

Hawtin, R. E., Zarkowska, T., Arnold, K., Thomas, C. K., Gooday, G. W., King, L. A., Kuzio, J. A. & Possee, R. D. (1997). Liquefaction of

Thomas, C. J., Gooday, G. W., King, L. A. & Possee, R. D. (2000).

Autographa californica nucleopolyhedrovirus-infected insects is dependent on the integrity of virus-encoded chitinase and cathepsin genes. Virology 238, 243–253. Hunter, F. R., Crook, N. E. & Entwistle, P. F. (1984). Viruses as pathogens for the control of insects. In Microbial Methods for Environmental Biotechnology, pp. 323–347. Edited by J. M. Grainger & J. M. Lynch. New York & London : Academic Press.

Mutagenesis of the active coding region of the Autographa californica nucleopolyhedrovirus chiA gene. Journal of General Virology 81, 1403–1411. Toyooka, K., Okamoto, T. & Minamikawa, T. (2000). Mass transport of Proform of a KDEL-tailed cysteine proteinase (SH-EP) to protein storage vacuoles by endoplasmic reticulum-derived vesicle is involved in protein mobilisation in germinating seeds. Journal of Cell Biology 148, 453–464.

Kang, W., Tristem, M., Maeda, S., Crook, N. E. & O’Reilly, D. R. (1998).

Identification and characterization of the Cydia pomonella granulovirus cathepsin and chitinase genes. Journal of General Virology 79, 2283–2292.

GJE

Received 15 August 2001 ; Accepted 9 November 2001