Journal of General Virology (1991), 72, 2981-2988. Printed in Great Britain
2981
Baculovirus expression of the human papillomavirus type 16 capsid proteins: detection of LI-L2 protein complexes Shang-Zhong Xi and Lawrence M. Banks* International Centre for Genetic Engineering and Biotechnology, Padriciano 99, 1-34012 Trieste, Italy
The human papillomavirus (HPV) type 16 major capsid proteins L1 and L2 have been produced in a baculovirus expression system. Both proteins are expressed at a high level and can be readily solubilized. The L1 capsid protein migrates close to its expected Mr of 60K. On the other hand L2 exhibits a much higher Mr migrating at 73K, which is considerably greater than its predicted M~ of 50K. The identity of both proteins has been confirmed also by Western blot analysis. Both proteins are produced in drastically reduced amounts in the presence of tunicamycin. In addition both LI and L2 show interesting patterns of
phosphorylation. LI is phosphorylated only weakly and this appears to be quite labile, whereas L2 is very heavily phosphorylated and this, in contrast, appears to be very stable. We have also made use of a dual expression vector for co-expressing the L1 and L2 proteins within the same baculovirus-infected cell. The results obtained from this system demonstrate the presence of protein complexes forming between the two capsid proteins. These studies indicate that at least the initial events in capsid assembly of HPVs can occur in the absence of viral DNA.
Introduction
sion has been used to good effect with the HPV-16 L1 protein where a glycosylated form of L1 was produced (Browne et al., 1988). However, vaccinia virus expression has limitations on the amount of protein produced and we were keen to be able to produce substantial quantities of protein for further biochemical and functional analysis, including studies on protein-protein interactions. Thus it was decided to express the HPV-16 Ll and L2 genes in a baculovirus system which has been used successfully to express a number of viral capsid proteins including late genes from HPV-11 and HPV-6 (Rose et al,, 1990). This report describes the expression of L1 and L2 by recombinant baculoviruses, either separately or together within the same infected cell. The results demonstrate the presence of protein complexes between these capsid proteins and show a complex series of post-translational modifications to the L1 and L2 proteins, including glycosylation and phosphorylation.
For some time certain human papillomaviruses (HPVs), particularly HPV-16 and HPV-18, have been implicated as causative agents in cervical cancer (for reviews see Vousden, 1989; Matlashewski, 1989). Considerable progress has been made in identifying the genes of the virus responsible for cell transformation. Thus much emphasis has been placed on E7 which is active in human (Hawley-Nelson et al., 1989; Munger et al., 1989) and rodent cells (Matlashewski et al., 1987; Phelps et al., 1988; Storey et al., 1988) and E6 which is primarily active in human cells (Hawley-Nelson et al., 1989; Munger et al., 1989). However studies on the viral late gene products, specifically the viral capsid proteins, are much less advanced. The main reason for this is lack of suitable in vitro systems for propagating the virus which thus precludes analysis of late gene expression. These late proteins have considerable potential as targets for preventing either virus infection or assembly. A number of laboratories have used prokaryotic systems to express these late viral proteins (Komly et al., 1986; Banks et al., 1987) but clearly there are limitations to these systems for studying protein function in eukaryotic cells. Alternative approaches are to use viral expression systems such as vaccinia virus or baculovirus as a means of obtaining substantial amounts of eukaryotically expressed and processed proteins. Vaccinia virus expres0001-0417 © 1991 SGM
Methods Plasmids. The HPV-16L1 and L2 openreadingframes(ORFs)were cloned into the baculovirus transfer vector p36C (Page, 1989) as follows. For L1, an Avail fragmentincludingnucleotides(nt) 5471 to 112 was excisedfromthe HPV-16 genome,blunt-endedwith Klenow polymeraseand BgllI linkerswere ligated. This was then clonedinto the BamHl site of p36C, For L2, oligonucleotideswere synthesized
2982
S.-Z. Xi and L. M. Banks
either side of the ORF containing BamHl restriction sites and-the L2 ORF was amplified from the HPV-16 genome by polymerase chain reaction (PCR). The resulting DNA encompassing nt 4213 to 5686 was cloned into the BamHl site of p36C and the resulting clone was verified by DNA sequence analysis. The L1 and L 2 0 R F s were similarly cloned into the dual transfer vector pAcUW3 (l~indly provided by Dr Ulrike Weyer), with L1 in the BgllI site under the control of the pl0 promoter and L2 in the BamHI site under the control of the polyhedrin promoter.
5471 ~
~
Cell and virus culture. Spodoptera frugiperda clone 9 (Sf9) cells were grown in TC-100 medium (Flow Laboratories) supplemented with 10 foetal calf serum at 22 °C. Autographa californica nuclear polyhedrosis virus (AcNPV) was cultured in Sf9 cells at 27 °C. Generation of L1 and L2 recombinant AcNPV. AcNPV DNA was cotransfected with either p36C-L1, p36C-L2 and pAcUW3-L1/2 plasmids into Sf9 cells by calcium phosphate precipitation. Virus was allowed to develop over 4 days, harvested from the tissue culture supernatant and then titrated Jn six-well dishes..Recd~binant viruses were screened for by the absence of polyhedgin particles in the viral plaques. Recombinant virus plaques were picked and subjected to two further rounds of plaque purification. Protein profiles were determined by harvesting recombinant virus-infected cells 72 h post-infection (p.i.), followed by analysis by SDS-PAGE. Western blot analysis. Sf9 cells were infected with recombinant virus at 10 p.f.u, per cell and infection was allowed to proceed for various times as indicated in the text. After this time the cells were pelleted and resuspended in EB buffer (50 mM-Tris-HC1 pH 7.5, 0.5 mM-DTT). The cells were sonicated for 1 min on ice and debris was removed by centrifugation at 14000 r.p.m, for 10 min. The extracted proteins were the~subjected to SDS-PAGE and Western blot analysis as described previously (Banks et al., 1987).
~
~
~
-
t
112
pol~:htr'~dr i n promoter
p36C Amp ~
Fig. 1. Construction of recombinant transfer vectors containing HPV16 L1 and L2 sequences. Filled-in boxes indicate the L1 and L 2 0 R F s and numbers are coordinates on the HPV-16 genome. Both LI and L2 were cloned into the BamHI site as described in the text in front of the polyhedrin promoter as indicated. The ampicillin resistance marker is also shown. The diaeram is not to scale.
(a) 1
2
(b) 1
2
(c) 1
2
(d) 1
2 •
?
Analysis o f protein complexes. Cells were infected with recombinant virus containing HPV-16 L1 and L2 as described above. After 72 h the cells were harvested and proteins extracted in EB. Immunoprecipitations were then performed on the clarified extracts using either anti-L 1 (Banks et al., 1987; MacLean et al., 1990), L2 (Firzlaff et al., 1987), cmyc (Evan et al., 1985) or preimmune antibodies. The complexes were collected on Protein A Sepharose and then washed extensively in 250 mM-NaC1, 0.1 ~ NP40 and 50 m~-HEPES. The precipitates were then separated on SDS-PAGE and proteins analysed by Western blot analysis as described previously (Banks et al., 1987).
Results Expression of HPV-16 L1 and L2 in baculovirus HPV-16 L1 and L2 ORFs were cloned into the baculovirus transfer vector p36C as described (Fig. 1). For HPV-16 L1 two separate constructs were made. One was produced from an AvalI fragment including nt 5471 to 112 and is shown in Fig. 1; this construct contains additional base pairs before the initiation codon. In addition we cloned a PCR-amplified L1 fragment and attempted to insert a cloning site just 5' of the initiation codon. Unfortunately, sequence analysis of this construct revealed a PCR mutation at the initiating A T G resulting in a truncated L 1 0 R F ; thus no further use was made of this construct. However throughout the sequence analysis additional base changes were detected in
Fig. 2. Expression of HPV-16 L1 and L2 proteins in recombinant baculovirus-infec~ed cell~ Cell~ were infected with recombinant virus and harvested after 72h. The protein profile from HPV-16 L1 recombinant virus is shown in (a) lane 1 with wild-type virus in lane 2. Western blot analysis with a MAb against HPV-16 L1 is shown in (b), with wild-type virus in lane 1 and LI recombinant in lane 2. The protein profile from HPV-16 L2 recombinant virus is shown in (c) lane 1, with wild-type virus in lane 2. Western blot analysis with a polyclonal anit-L2 antibody is shown in (d), with wild-type virus in lane 1 and L2 recombinant in lane 2. Mr markers arC earbonic anhydrase (29K), ovalbumin (43K), BSA (68K) andphosphorylase b (97K).
the L1 sequence which were also found to be present in the prototype HPV-16 DNA. A summary of the base changes detected which differ from published sequences are shown in Table 1. Both of these changes, a CAT
BacuIovirus expression of HPV-16 proteins (a) 1
2983
(b) 2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
9
--200K
--97K --68K LI --43K
Fig. 3. Effect of tunicamycin on HPV-16 L1 expression. Cells were infected with HPV-16 L1 recombinant virus and harvested at different times p.i. Protein profiles were monitored and are shown in (a). Times of harvesting were 48 h (lanes 1 and 4), 72 h (lanes 2 and 5) and 96 h (lanes 3, 6, 7 and 8). Cells in the absence of tunicamycin are in lanes 1 to 3 and 7. Cells in the presence of 10 ~tg/ml tunicamycin are in lanes 4 to 6 and 8. Wild-type virus is shown in lanes 7 and 8. Western blot analysis is shown in (b). Lanes are as described in (a) with lane 9 showing the Mr markers as in Fig. 2 plus myosin (200K).
(a)
(b) 1
2
3
4
5
6
6
5
4
3
2
1
L2
Fig. 4. Effect of tunicamycin on HPV-16 L2 expression. Cells were infected with HPV-16 L2 recombinant virus and harvested at different times p.i. Protein profiles were monitored and are shown in (a). Times of harvesting were 48 h (lanes 1 and 4), 72 h (lanes 2 and 5) and 96 h (lanes 3 and 6). Cells in the absence of tunicamycin are in lanes 1 to 3. Cells in the presence of 10/.tg/ml tunicamycin are in lanes 4 to 6. Western blot analysis is shown in (b). Lanes are as described in (a).
addition at position 6900 and a GAT deletion at position 6950 have also been reported recently (Parton, 1990). For cloning the HPV-16 L2 into p36C, PCR was performed to give a BamHI fragment with the 5' end just
upstream of the initiation codon. Sequence analysis of this clone revealed two base pair changes in the PCR product when compared with the published sequence. One was A to T at residue 4363 which we also found to be
2984
S.-Z. Xi and L. M. Banks
Table 1. Changes in L1 and L2 nucleotide sequence ORF
Base change
Resulting amino acid change
L1
nt 6900 insert CAT nt 6950 delete GAT nt 4363 A ~ T nt 4623 T ,,* G*
Insert serine Deleteaspartic acid Glutamic acid ~ Aspartic acid Valine ,~ Glycine*
L2
* This change is the result of a PCR artefact. present in the prototype HPV-16 L2 sequence (Table 1), and this change results in Asp in place of Glu. The second which was purely PCR-generated at residue 4623 resulted in T to G giving Gly in place of Val. Considering the heterogeneous nature of the different HPV L2 proteins, we felt that this change would not result in an alteration in the tertiary structure of the protein, so continued to use this for expression purposes. The p36C-L1 and p36C-L2 plasmids were then cotransfected with A c N P V D N A into Sf9 cells and recombinant viruses were screened and analysed by S D S - P A G E as described in Methods. The results for two recombinant viruses containing L1 and L2 are shown in Fig. 2. It is apparent that the Ll-containing virus produced a recombinant protein migrating at approximately 60K Mr (Fig. 2a, lane 1). This compares favourably with the predicted Mr of 54K. The identity of this protein was confirmed by Western blot analysis using an anti-HPV-16 L1 monoclonal antibody (MAb) raised against an L1-/~-galactosidase fusion protein (MacLean et al., 1990). As shown (Fig. 2b, lane 2) this antibody reacted strongly with the expressed 60K protein. The protein profile from the HPV-16 L2-expressing recombinant virus is shown in Fig. 2 (c), lane 1. It is clear that the protein expressed migrates with an apparent Mr of 73K. This is particularly surprising since the predicted size is approximately 50K, although this result is in agreement with previous reports for the Mr of L2 from HPV-1 and HPV-11 (Doorbar & Gallimore, 1987; Rose et al., 1990). The identity of this L2 protein was confirmed by Western blot analysis using a polyclonal rabbit antibody raised against the whole HPV-16 L2 protein (Firzlaff et al., 1987). As shown (Fig. 2d, lane 2) this antibody reacts strongly with the expressed 73K protein, with a weak reaction against a protein of approximately 40K Mr. L I and L2 accumulation is inhibited by tunicamycin Having obtained good expression of the HPV-16 capsid proteins it was interesting to determine whether or not they were glycosylated in this system. Glycosylation of HPV-16 L1 has already been reported in a vaccinia virus expression system (Browne et al., 1988) but comparison
Fig. 5. Analysis of phosphorylation state of the HPV-16 L1 and L2 recombinant proteins. Cells were infected with recombinant baculovirus and after 48 h were labelled with 300 ~tCi/ml organic 32p for a further 2 h. Cells were then harvested and protein profilesmonitored by autoradiography. Lanes are as follows:wild-type virus, lane 1; L2 recombinant, lane 2; L1 recombinant, lane 3. The L1 and L2 proteins are indicated.
between the two systems would be useful. Thus cells were infected with recombinant viruses expressing L1 and L2 and grown in the absence or presence of 10~tg/ml tunicamycin over a number of days. The cells were harvested at differept time points and the proteins analysed by S D S - P A G E and Western blotting. The results obtained for L1 are shown in Fig. 3. In the absence of tunicamycin maximal L1 expression was obtained at 72 to 96 h (lanes 2 and 3 respectively). However, in the presence of tunicamycin there was a marked reduction in the amount of L1 protein expressed (lanes 4 to 6). Western blot analysis (Fig. 3b) confirmed this and also revealed a slight reduction in the Mr of the protein. Interestingly, this reduction in size seemed to be primarily mediated by the absence of a higher Mr form of the L1 protein. Thus in this system it seems that HPV-16 L1 is glycosylated but at a much lower level than that observed for L1 expressed in vaccinia virus. The results obtained for L2 are shown in Fig. 4. As observed with L1 we see maximal protein expression at 72 to 96 h p.i. (lanes 2 and 3 respectively). In addition, in the presence of tunicamycin there was a dramatic reduction in the level of L2 expression (lanes 4 to 6). There was however no significant shift in the Mr of the protein indicating no major level of glycosylation,
Baculovirus expression of HPV-16 proteins although a minor decrease would be difficult to detect in a protein of this size.
(a)
2985
(b) 1
2
3
4
1
2
3
4
5
6
LI and L2 are phosphoproteins To characterize the HPV-16 capsid proteins further it was interesting to determine whether they were phosphorylated in this expression system. Cells were infected with recombinant virus for 72 h and then labelled for 2 h with 300 ~tCi/ml organic 3zp. The cells were harvested and the proteins analysed by S D S - P A G E and autoradiography. The results obtained are shown in Fig. 5. The L2 protein was found to be heavily phosphorylated (lane 2). Interestingly the L1 protein was also phosphorylated (lane 3) but obviously at a much lower level than that observed for L2. The identity of both the labelled proteins as L1 and L2 was also confirmed by Western blot analysis (data not shown). The rate of phosphate turnover on these proteins was then investigated. The cells were infected and labelled as above but chased with cold phosphate for different time points. The results for L2 are shown in Fig. 6(a). Constant amounts of L2 protein were loaded throughout as ensured by Western blot analysis (data not shown). The phosphorylation of L2 was found to be quite stable, with label still present 24 h into the chase, and a half-life of approximately 3 h. In contrast the results with L1 (Fig. 6b) indicate that the phosphate label on this protein was more labile, with a half-life of approximately 60 min.
:L1
Fig. 6. Rate of phosphate turnover of the L1 and L2 proteins. Cells were infected with recombinant baculovirus and after 48 h were labelled with 3"00~Ci/ml organic 32p for a further 2 h. The labelled medium was then removed, the cells were then pulsed with cold medium and harvested at different time points, and the protein profiles were monitored by autoradiography. The pattern observed with L2 is shown in (a). Lanes are as follows:0 h pulse (lane 1), 3 h pulse (lane 2), 6 h pulse (lane 3) and 24 h pulse (lane 4). The pattern observed with LI is shown in (b). Lanes are as follows: wild-type virus (lane l), 0 h pulse (lane 2), 30 rain pulse (lane 3), 60 min pulse (lane 4), 90 min pulse (lane 5) and 120 min pulse (lane 6). L1 and L2 proteins are indicated.
L1 and L2 form protein complexes Having shown that the HPV-16 capsid proteins could be expressed and post-translationally modified in the baculovirus expression system it was of interest to determine whether these proteins were capable of forming protein complexes. Since these proteins are the major components of the viral capsid it is likely that they will have a very close interaction in vivo. In addition there have been reports of using baculovirus to express other viral capsid proteins to obtain partial capsid assembly (French & Roy, 1990). For these studies the L1 and L2 genes were cloned into the dual expression transfer vector p A c U W 3 (Fig. 7). This was cotransfected with A c N P V D N A and recombinant clones were screened and selected as described above. One recombinant virus was obtained expressing good levels of both L1 and L2 proteins as determined by Western blot analysis (Fig, 8). In this expression system L1 appears to be produced in larger amounts than L2, which probably reflects the relative strength of the polyhedrin and p l 0 promoters. To investigate possible protein interactions between the L1 and L2 proteins, cells were infected with the double recombinant virus and extracts were made after
Ll
}~
~-L
L2
}7~
Fig. 7. Construction of recombinant transfer vector containing both HPV-16 L1 and L20RFs in the same vector. The fragments described in Fig. 1 were cloned into pAcUW3 as shown. L1 was cloned into the BgllI site in front of the baculovirus p 10 promoter. L2 was cloned into the BamHI site in front of the baculovirus polyhedrin promoter. The ampicillin resistance marker is also shown. The diagram is not to scale.
72 h. Firstly immunoprecipitations were done using a series of antibodies which have been shown previously to react specifically with their target antigen. Thus the antipeptide antibody directed to the last 14 amino acids of the HPV-16 L1 protein (Banks et aL, 1987), the anti cmyc antibody (Evan et al., 1985), the rabbit preimmune
2986
S . - Z . X i and L. M . B a n k s
(a) I
(b) 2
(c) 1
2
(a)
(d) 1
2
1
1
2
(b) 2
3
4
1
5
2
P7K
-97K
i8K -,68K L2
L1--~ -43K -43K
L1 -29K -29K Fig. 8. Co-expression of the HPV-16 LI and L2 proteins in recombinant baculovirus-infected cells. Cells were infected with recombinant virus and harvested after 72 h. The protein profile from recombinant virus is shown in (a) lane 2 with wild-type virus in lane 1. Western blot analyses are as follows: anti-L1 MAb in (b); polyclonal anti-L2 antibody in (c); preimmune antibody in (d). Lanes are as described for (a) and the positions of the L1 and L2 proteins are indicated.
antibody or the anti-HPV-16 L2 a n t i b o d y (Firzlaff et al., 1987) were used. The resultant i m m u n e complexes were then subjected to S D S - P A G E and proteins were analysed by Western blotting using an anti-HPV-16 L1 M A b . T h e results o b t a i n e d are shown in Fig. 9(a). Clearly the L 1 protein was coprecipitated by the anti-L2 antibody (Fig. 9a, lane 3). N o such result was obtained with either the c - m y c or p r e i m m u n e antibodies (Fig. 9a, lanes 1 and 2 respectively). T h e converse experiment was then performed using anti-L 1 antibody to precipitate the protein complex and then the Western blot was p r o b e d with anti-L2 polyclonal a n t i b o d y (Fig. 9b). T h e L2 protein was detected only after precipitation with the anti-L1 a n t i b o d y (Fig. 9b, lane 2); no L2 was precipitated with the p r e i m m u n e antibody (lane 1). There was no case in w h i c h we observed anti-L1 a n t i b o d y precipitating L2 protein or vice versa, f r o m baculovirusinfected cells expressing L2 or L1 alone (data not shown), nor did we see any cross-reaction on W e s t e r n blots (Fig. 8), thus confirming that the above results were not due to cross-reactivity of the antibodies. These results d e m o n strate that i m m u n o p r e c i p i t a t i o n with either anti-L1 or -L2 antibodies results in coprecipitation o f the other
Fig. 9. Coprecipitation of the HPV-16 L1 and L2 proteins. Cells were infected with recombinant baculovirus expressing both the L1 and L2 proteins. Cells were harvested after 72 h and immunoprecipitation followed by Western blot analysis was performed. (a) Recombinant virus-infected cells were immunoprecipitated with anti c-myc antibody (lane 1), preimmune rabbit serum (lane 2), anti-HPV-16 L2 antibody (lane 3) and anti-HPV-16 LI peptide antibody (lane 4). Total infected cell extract was run on lane 5. The imrnunoprecipitates were separated by SOS-PAGE, Western-blotted and probed with an anti-HPV-16 L1 MAb. '(b) Infected cells were immunoprecipitated with mouse preimmune serum (lane 1) or anti-HPV-16 LI MAb (lane 2). The immunoprecipitates were separated by SOS-PAGE, Western-blotted and probed with an anti-HPV-16 L2 polyclonal antibody.
capsid protein, a result which is mediated t h r o u g h a p r o t e i n - p r o t e i n interaction between the capsid proteins.
Discussion This paper describes the expression and characterization of the H P V - 1 6 capsid proteins, L I and L2, in a baculovirus expression system. This provides i m p o r t a n t biochemical information regarding potential processing and control o f these proteins which is currently unavailable from prokaryotic expression systems. Most importantly the results presented here clearly show the presence of protein complexes forming between t h e L1 and L2 proteins. Both proteins are m a d e in substantial quantities in the Sf9 cells, although L2 is clearly the most a b u n d a n t
Baculovirus expression of HP V-16 proteins species when both proteins are expressed from the same promoter. It is possible that L2 is more stable than the L1 protein and this is suggested to some degree by the phosphate labelling data. However, it is also possible that lower L1 expression is a result of the larger noncoding region prior to the ATG in the p36C-L1 construct. In the dual expression system L1 is produced in larger amounts and it seems likely that this is due to differences between the two promoters used to express L1 and L2. The expression of both L1 and L2 is adversely affected to a high degree by tunicamycin. This agent blocks Nlinked glycosylation and in this system LI appears to be weakly glycosylated, which is in agreement with previous reports (Browne et al., 1988). With L2, glycosylation is not as apparent, but there is a dramatic reduction in the amount of protein expressed. Clearly, however, glycosylation does not account for the aberrant high Mr of the L2 protein. This high Mr for the papillomavirus L2 protein has been observed for bovine papillomavirus type 1 (BPV-1) (Jin et al., 1989), BPV-2 (Rippe & Meinke, 1989), HPV-1 (Doorbar & Gallimore, 1987), HPV-6 (Rose et al., 1990) and HPV-16 shown here. Both the HPV-16 L1 and L2 proteins are phosphorylated. L2 has by far the highest number of phosphate residues of the two proteins, making L2 a very major phosphoprotein. The phosphorylation of L2 appears to be quite stable, as would be expected for a structural protein. Interestingly tunicamycin did not adversely affect the half-life of L2 (data not shown) implying that it simply inhibits L2 accumulation. This high level of phosphorylation implies an important function with possible interaction with viral DNA and a role in viral packaging and assembly. The L1 protein on the other hand is only weakly phosphorylated implying only a few phosphate groups per molecule. In addition, this phosphorylation appears to be more unstable with a half-life of around 60 min. It is possible that this is purely due to the expression system but it may also represent some mechanism for controlling the packaging of the viral DNA. The observation that the HPV-16 capsid proteins can physically interact in this system is particularly interesting. This demonstrates that we may expect to see at least some partial capsid assembly of the virus and further studies are now in progress. In addition this result is of value for our understanding of the mechanisms of capsid assembly. Clearly we can obtain interaction between the capsid proteins in the absence of any other additional HPV factors including either viral proteins and more importantly viral DNA. It thus seems likely that we may expect to see L1-L2 protein complexes forming within the HPV-16-infected cell prior to association with the viral DNA for packaging. As well as studies aimed at
2987
identifying any partial capsids which may be produced it will be interesting to determine the effects of the various post-translational modifications on the ability of these proteins to form complexes. In conclusion, we have been able to demonstrate that the HPV-16 L1 and L2 proteins expressed in a baculovirus system will interact to form protein complexes. In addition these proteins are subjected to a series of post-translational modifications including glycosylation and phosphorylation. Current studies are now aimed at further elucidating the mechanism of interaction between the L1 and L2 proteins. We are most grateful to Peter Ertl of the Wellcome Research Laboratories for invaluable help and advice in setting up the baculovirus expression system. We are also very grateful to Martin Page for his gift of the plasmid p36C, Ulrike Weyer for the generous gift of plasmid pAcUW3, to Denise Galloway for the anti-L2 antibody and to Lionel Crawford for the anti-L1 MAb.
References BANKS, L., MATLASHEWSKI,G , PIM, D., CHURCHER, M., ROBERTS, C. & CRAWFORD, L. (1987). Expression of human papillomavirus type 6 and 16 capsid proteins in bacteria and their antigenic characterization. Journal of General Virology 68, 3081-3089. BROWNE, H. M., CHURCHER, M. J., STANLEY, M. A., SMITH, G. L. & MINSON, A. C. (1988). Analysis of the L1 gene product of human papillomavirus type 16 by expression in a vaccinia virus recombinant. Journal of General Virology 69, 1263-1273. DOORBAR, J. & GALLXMORE, P. H. (1987). Identification of proteins encoded by the L1 and L2 open reading frames of human papillomavirus la. Journal of Virology 61, 2793-2799. EVAr~, G. I., LEWIS, G. K., RAMSEY, G. & BISHOP, J. M. (1985). Isolation of monoclonal antibodies specific for human c-myc proto oncogene product. Molecular and Cellular Biology 5, 3610-3616. FIRZLAFE, J. M., KIVIAT, N. B., BECKMANN,A. M., JENISON, S. A. & GALLOWAY,n . A. (1987). Detection of human papillomavirus capsid antigens in various squamous epithelial lesions using antibodies directed against the L1 and L2 open reading frames. Virology 164, 467-477. FRENCH, T. J. & ROY, P. (1990). Synthesis of bluetongue virus (BTV) corelike particles by a recombinant baculovirus expressing the two major structural core proteins of BTV. Journalof Virology 64, 15301536. HAWLEY-NELSON, P., VOUSDEN,K. H., HUBBERT, N. L., LowY, D. R. & SCmLLER, J. T. (1989). HPV E6 and E7 proteins cooperate to immortalise human foreskin keratinocytes. EMBO Journal 8, 39053910. JIN, X. W., COWSERT, L. M., PILACINSKI,W. P. & JENSON, A. B. (1989). Identification of L2 open reading frame gene products of bovine papillomavirus type 1 using monoclonal antibodies. Journal of General Virology 70, 1133-1140. KOMLY, C. A., BREITBURD,F., CROISSANT,O. & STREECK, R. E. (1986). The L2 open reading frame of human papillomavirus type la encodes a minor structural protein carrying type specific antigens. Journal of Virology 60, 813-816. MACLEAN, C. F., CHURCHER, M. J., MEINKE, J., SMITH, G. L., HIGGINS, G., STANLEY, M. & MINSON, A. C. (1990). Production and characterisation of a monoclonal antibody to human papillomavirus type 16 using a recombinant vaccinia virus. Journal of Clinical Pathology 43, 488-492. MATLASHEWSKI, G. (1989). The cell biology of human papillomavirus transformed cells. Anticancer Research 9, 1447-1556.
2988
S . - Z . X i and L. M . B a n k s
MATLASHEWSKI,G., SCHNEIDER,J., BANKS,L., JONES,N., MURRAY,A. CRAWFORD, t. (1987). Human papillomavirus type 16 DNA cooperates with activated ras in transforming cells. EMBO Journal 6, 1741-1746. MUNGER, K., PHELPS, W. C., BUBB,V., HOWLEY,P. M. & SCHLEGEL, R. (1989). The E6 and E7 genes of the human papillomavirus type 16 are together necessary and sufficient for transformation of primary human keratinocytes. Journal of Virology 63, 4417M421. PAGE, M. J. (1989). p36C: an improved baculovirus vector for producing high levels of mature recombinant proteins. Nucleic Acids Research 17, 454. PAR'I'ON, A. (1990). Nucleotide sequence of the HPV-16 L1 open reading frame. Nucleic Acids Research 15, 3631. PHELPS, W. C., YEE, C. L., MUNGER,K. & HOWLEY,P. M. (1988). The human papillomavirus type 16 E7 gene encodes transactivation and transformation functions similar to those of adenovirus E 1A. Cell 53, 539-547.
RIPPE, R. & MEINKE,W. (1989). Identification and characterisation of the BPV-2 L2 protein. Virology 171, 298-301. ROSE, R. C., BONNEZ, W., STRIKE, D. G. & REICHMAN,R. C. (1990). Expression of the full-length products of the human papillomavirus type 6b (HPV-6b) and HPV-II L2 open reading frames by recombinant baculovirus, and antigenic comparisons with HPV-11 whole virus particles. Journal of General Virology 71, 2725-2729. STOREY, A., PIM, D., MURRAY, A., OSBORN, K., BANKS, L. & CRAWFORD, L. (1988). Comparison of the in vitro transforming activities of human papillomavirus types. EMBO Journal 7, 18151820. VOUSDEN,K. (1989). Human papillomaviruses and cervical carcinoma. Cancer Cells 1, 43-50.
(Received 10 June 1991 ; Accepted 2 September 1991)
2981
Baculovirus expression of the human papillomavirus type 16 capsid proteins: detection of LI-L2 protein complexes Shang-Zhong Xi and Lawrence M. Banks* International Centre for Genetic Engineering and Biotechnology, Padriciano 99, 1-34012 Trieste, Italy
The human papillomavirus (HPV) type 16 major capsid proteins L1 and L2 have been produced in a baculovirus expression system. Both proteins are expressed at a high level and can be readily solubilized. The L1 capsid protein migrates close to its expected Mr of 60K. On the other hand L2 exhibits a much higher Mr migrating at 73K, which is considerably greater than its predicted M~ of 50K. The identity of both proteins has been confirmed also by Western blot analysis. Both proteins are produced in drastically reduced amounts in the presence of tunicamycin. In addition both LI and L2 show interesting patterns of
phosphorylation. LI is phosphorylated only weakly and this appears to be quite labile, whereas L2 is very heavily phosphorylated and this, in contrast, appears to be very stable. We have also made use of a dual expression vector for co-expressing the L1 and L2 proteins within the same baculovirus-infected cell. The results obtained from this system demonstrate the presence of protein complexes forming between the two capsid proteins. These studies indicate that at least the initial events in capsid assembly of HPVs can occur in the absence of viral DNA.
Introduction
sion has been used to good effect with the HPV-16 L1 protein where a glycosylated form of L1 was produced (Browne et al., 1988). However, vaccinia virus expression has limitations on the amount of protein produced and we were keen to be able to produce substantial quantities of protein for further biochemical and functional analysis, including studies on protein-protein interactions. Thus it was decided to express the HPV-16 Ll and L2 genes in a baculovirus system which has been used successfully to express a number of viral capsid proteins including late genes from HPV-11 and HPV-6 (Rose et al,, 1990). This report describes the expression of L1 and L2 by recombinant baculoviruses, either separately or together within the same infected cell. The results demonstrate the presence of protein complexes between these capsid proteins and show a complex series of post-translational modifications to the L1 and L2 proteins, including glycosylation and phosphorylation.
For some time certain human papillomaviruses (HPVs), particularly HPV-16 and HPV-18, have been implicated as causative agents in cervical cancer (for reviews see Vousden, 1989; Matlashewski, 1989). Considerable progress has been made in identifying the genes of the virus responsible for cell transformation. Thus much emphasis has been placed on E7 which is active in human (Hawley-Nelson et al., 1989; Munger et al., 1989) and rodent cells (Matlashewski et al., 1987; Phelps et al., 1988; Storey et al., 1988) and E6 which is primarily active in human cells (Hawley-Nelson et al., 1989; Munger et al., 1989). However studies on the viral late gene products, specifically the viral capsid proteins, are much less advanced. The main reason for this is lack of suitable in vitro systems for propagating the virus which thus precludes analysis of late gene expression. These late proteins have considerable potential as targets for preventing either virus infection or assembly. A number of laboratories have used prokaryotic systems to express these late viral proteins (Komly et al., 1986; Banks et al., 1987) but clearly there are limitations to these systems for studying protein function in eukaryotic cells. Alternative approaches are to use viral expression systems such as vaccinia virus or baculovirus as a means of obtaining substantial amounts of eukaryotically expressed and processed proteins. Vaccinia virus expres0001-0417 © 1991 SGM
Methods Plasmids. The HPV-16L1 and L2 openreadingframes(ORFs)were cloned into the baculovirus transfer vector p36C (Page, 1989) as follows. For L1, an Avail fragmentincludingnucleotides(nt) 5471 to 112 was excisedfromthe HPV-16 genome,blunt-endedwith Klenow polymeraseand BgllI linkerswere ligated. This was then clonedinto the BamHl site of p36C, For L2, oligonucleotideswere synthesized
2982
S.-Z. Xi and L. M. Banks
either side of the ORF containing BamHl restriction sites and-the L2 ORF was amplified from the HPV-16 genome by polymerase chain reaction (PCR). The resulting DNA encompassing nt 4213 to 5686 was cloned into the BamHl site of p36C and the resulting clone was verified by DNA sequence analysis. The L1 and L 2 0 R F s were similarly cloned into the dual transfer vector pAcUW3 (l~indly provided by Dr Ulrike Weyer), with L1 in the BgllI site under the control of the pl0 promoter and L2 in the BamHI site under the control of the polyhedrin promoter.
5471 ~
~
Cell and virus culture. Spodoptera frugiperda clone 9 (Sf9) cells were grown in TC-100 medium (Flow Laboratories) supplemented with 10 foetal calf serum at 22 °C. Autographa californica nuclear polyhedrosis virus (AcNPV) was cultured in Sf9 cells at 27 °C. Generation of L1 and L2 recombinant AcNPV. AcNPV DNA was cotransfected with either p36C-L1, p36C-L2 and pAcUW3-L1/2 plasmids into Sf9 cells by calcium phosphate precipitation. Virus was allowed to develop over 4 days, harvested from the tissue culture supernatant and then titrated Jn six-well dishes..Recd~binant viruses were screened for by the absence of polyhedgin particles in the viral plaques. Recombinant virus plaques were picked and subjected to two further rounds of plaque purification. Protein profiles were determined by harvesting recombinant virus-infected cells 72 h post-infection (p.i.), followed by analysis by SDS-PAGE. Western blot analysis. Sf9 cells were infected with recombinant virus at 10 p.f.u, per cell and infection was allowed to proceed for various times as indicated in the text. After this time the cells were pelleted and resuspended in EB buffer (50 mM-Tris-HC1 pH 7.5, 0.5 mM-DTT). The cells were sonicated for 1 min on ice and debris was removed by centrifugation at 14000 r.p.m, for 10 min. The extracted proteins were the~subjected to SDS-PAGE and Western blot analysis as described previously (Banks et al., 1987).
~
~
~
-
t
112
pol~:htr'~dr i n promoter
p36C Amp ~
Fig. 1. Construction of recombinant transfer vectors containing HPV16 L1 and L2 sequences. Filled-in boxes indicate the L1 and L 2 0 R F s and numbers are coordinates on the HPV-16 genome. Both LI and L2 were cloned into the BamHI site as described in the text in front of the polyhedrin promoter as indicated. The ampicillin resistance marker is also shown. The diaeram is not to scale.
(a) 1
2
(b) 1
2
(c) 1
2
(d) 1
2 •
?
Analysis o f protein complexes. Cells were infected with recombinant virus containing HPV-16 L1 and L2 as described above. After 72 h the cells were harvested and proteins extracted in EB. Immunoprecipitations were then performed on the clarified extracts using either anti-L 1 (Banks et al., 1987; MacLean et al., 1990), L2 (Firzlaff et al., 1987), cmyc (Evan et al., 1985) or preimmune antibodies. The complexes were collected on Protein A Sepharose and then washed extensively in 250 mM-NaC1, 0.1 ~ NP40 and 50 m~-HEPES. The precipitates were then separated on SDS-PAGE and proteins analysed by Western blot analysis as described previously (Banks et al., 1987).
Results Expression of HPV-16 L1 and L2 in baculovirus HPV-16 L1 and L2 ORFs were cloned into the baculovirus transfer vector p36C as described (Fig. 1). For HPV-16 L1 two separate constructs were made. One was produced from an AvalI fragment including nt 5471 to 112 and is shown in Fig. 1; this construct contains additional base pairs before the initiation codon. In addition we cloned a PCR-amplified L1 fragment and attempted to insert a cloning site just 5' of the initiation codon. Unfortunately, sequence analysis of this construct revealed a PCR mutation at the initiating A T G resulting in a truncated L 1 0 R F ; thus no further use was made of this construct. However throughout the sequence analysis additional base changes were detected in
Fig. 2. Expression of HPV-16 L1 and L2 proteins in recombinant baculovirus-infec~ed cell~ Cell~ were infected with recombinant virus and harvested after 72h. The protein profile from HPV-16 L1 recombinant virus is shown in (a) lane 1 with wild-type virus in lane 2. Western blot analysis with a MAb against HPV-16 L1 is shown in (b), with wild-type virus in lane 1 and LI recombinant in lane 2. The protein profile from HPV-16 L2 recombinant virus is shown in (c) lane 1, with wild-type virus in lane 2. Western blot analysis with a polyclonal anit-L2 antibody is shown in (d), with wild-type virus in lane 1 and L2 recombinant in lane 2. Mr markers arC earbonic anhydrase (29K), ovalbumin (43K), BSA (68K) andphosphorylase b (97K).
the L1 sequence which were also found to be present in the prototype HPV-16 DNA. A summary of the base changes detected which differ from published sequences are shown in Table 1. Both of these changes, a CAT
BacuIovirus expression of HPV-16 proteins (a) 1
2983
(b) 2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
9
--200K
--97K --68K LI --43K
Fig. 3. Effect of tunicamycin on HPV-16 L1 expression. Cells were infected with HPV-16 L1 recombinant virus and harvested at different times p.i. Protein profiles were monitored and are shown in (a). Times of harvesting were 48 h (lanes 1 and 4), 72 h (lanes 2 and 5) and 96 h (lanes 3, 6, 7 and 8). Cells in the absence of tunicamycin are in lanes 1 to 3 and 7. Cells in the presence of 10 ~tg/ml tunicamycin are in lanes 4 to 6 and 8. Wild-type virus is shown in lanes 7 and 8. Western blot analysis is shown in (b). Lanes are as described in (a) with lane 9 showing the Mr markers as in Fig. 2 plus myosin (200K).
(a)
(b) 1
2
3
4
5
6
6
5
4
3
2
1
L2
Fig. 4. Effect of tunicamycin on HPV-16 L2 expression. Cells were infected with HPV-16 L2 recombinant virus and harvested at different times p.i. Protein profiles were monitored and are shown in (a). Times of harvesting were 48 h (lanes 1 and 4), 72 h (lanes 2 and 5) and 96 h (lanes 3 and 6). Cells in the absence of tunicamycin are in lanes 1 to 3. Cells in the presence of 10/.tg/ml tunicamycin are in lanes 4 to 6. Western blot analysis is shown in (b). Lanes are as described in (a).
addition at position 6900 and a GAT deletion at position 6950 have also been reported recently (Parton, 1990). For cloning the HPV-16 L2 into p36C, PCR was performed to give a BamHI fragment with the 5' end just
upstream of the initiation codon. Sequence analysis of this clone revealed two base pair changes in the PCR product when compared with the published sequence. One was A to T at residue 4363 which we also found to be
2984
S.-Z. Xi and L. M. Banks
Table 1. Changes in L1 and L2 nucleotide sequence ORF
Base change
Resulting amino acid change
L1
nt 6900 insert CAT nt 6950 delete GAT nt 4363 A ~ T nt 4623 T ,,* G*
Insert serine Deleteaspartic acid Glutamic acid ~ Aspartic acid Valine ,~ Glycine*
L2
* This change is the result of a PCR artefact. present in the prototype HPV-16 L2 sequence (Table 1), and this change results in Asp in place of Glu. The second which was purely PCR-generated at residue 4623 resulted in T to G giving Gly in place of Val. Considering the heterogeneous nature of the different HPV L2 proteins, we felt that this change would not result in an alteration in the tertiary structure of the protein, so continued to use this for expression purposes. The p36C-L1 and p36C-L2 plasmids were then cotransfected with A c N P V D N A into Sf9 cells and recombinant viruses were screened and analysed by S D S - P A G E as described in Methods. The results for two recombinant viruses containing L1 and L2 are shown in Fig. 2. It is apparent that the Ll-containing virus produced a recombinant protein migrating at approximately 60K Mr (Fig. 2a, lane 1). This compares favourably with the predicted Mr of 54K. The identity of this protein was confirmed by Western blot analysis using an anti-HPV-16 L1 monoclonal antibody (MAb) raised against an L1-/~-galactosidase fusion protein (MacLean et al., 1990). As shown (Fig. 2b, lane 2) this antibody reacted strongly with the expressed 60K protein. The protein profile from the HPV-16 L2-expressing recombinant virus is shown in Fig. 2 (c), lane 1. It is clear that the protein expressed migrates with an apparent Mr of 73K. This is particularly surprising since the predicted size is approximately 50K, although this result is in agreement with previous reports for the Mr of L2 from HPV-1 and HPV-11 (Doorbar & Gallimore, 1987; Rose et al., 1990). The identity of this L2 protein was confirmed by Western blot analysis using a polyclonal rabbit antibody raised against the whole HPV-16 L2 protein (Firzlaff et al., 1987). As shown (Fig. 2d, lane 2) this antibody reacts strongly with the expressed 73K protein, with a weak reaction against a protein of approximately 40K Mr. L I and L2 accumulation is inhibited by tunicamycin Having obtained good expression of the HPV-16 capsid proteins it was interesting to determine whether or not they were glycosylated in this system. Glycosylation of HPV-16 L1 has already been reported in a vaccinia virus expression system (Browne et al., 1988) but comparison
Fig. 5. Analysis of phosphorylation state of the HPV-16 L1 and L2 recombinant proteins. Cells were infected with recombinant baculovirus and after 48 h were labelled with 300 ~tCi/ml organic 32p for a further 2 h. Cells were then harvested and protein profilesmonitored by autoradiography. Lanes are as follows:wild-type virus, lane 1; L2 recombinant, lane 2; L1 recombinant, lane 3. The L1 and L2 proteins are indicated.
between the two systems would be useful. Thus cells were infected with recombinant viruses expressing L1 and L2 and grown in the absence or presence of 10~tg/ml tunicamycin over a number of days. The cells were harvested at differept time points and the proteins analysed by S D S - P A G E and Western blotting. The results obtained for L1 are shown in Fig. 3. In the absence of tunicamycin maximal L1 expression was obtained at 72 to 96 h (lanes 2 and 3 respectively). However, in the presence of tunicamycin there was a marked reduction in the amount of L1 protein expressed (lanes 4 to 6). Western blot analysis (Fig. 3b) confirmed this and also revealed a slight reduction in the Mr of the protein. Interestingly, this reduction in size seemed to be primarily mediated by the absence of a higher Mr form of the L1 protein. Thus in this system it seems that HPV-16 L1 is glycosylated but at a much lower level than that observed for L1 expressed in vaccinia virus. The results obtained for L2 are shown in Fig. 4. As observed with L1 we see maximal protein expression at 72 to 96 h p.i. (lanes 2 and 3 respectively). In addition, in the presence of tunicamycin there was a dramatic reduction in the level of L2 expression (lanes 4 to 6). There was however no significant shift in the Mr of the protein indicating no major level of glycosylation,
Baculovirus expression of HPV-16 proteins although a minor decrease would be difficult to detect in a protein of this size.
(a)
2985
(b) 1
2
3
4
1
2
3
4
5
6
LI and L2 are phosphoproteins To characterize the HPV-16 capsid proteins further it was interesting to determine whether they were phosphorylated in this expression system. Cells were infected with recombinant virus for 72 h and then labelled for 2 h with 300 ~tCi/ml organic 3zp. The cells were harvested and the proteins analysed by S D S - P A G E and autoradiography. The results obtained are shown in Fig. 5. The L2 protein was found to be heavily phosphorylated (lane 2). Interestingly the L1 protein was also phosphorylated (lane 3) but obviously at a much lower level than that observed for L2. The identity of both the labelled proteins as L1 and L2 was also confirmed by Western blot analysis (data not shown). The rate of phosphate turnover on these proteins was then investigated. The cells were infected and labelled as above but chased with cold phosphate for different time points. The results for L2 are shown in Fig. 6(a). Constant amounts of L2 protein were loaded throughout as ensured by Western blot analysis (data not shown). The phosphorylation of L2 was found to be quite stable, with label still present 24 h into the chase, and a half-life of approximately 3 h. In contrast the results with L1 (Fig. 6b) indicate that the phosphate label on this protein was more labile, with a half-life of approximately 60 min.
:L1
Fig. 6. Rate of phosphate turnover of the L1 and L2 proteins. Cells were infected with recombinant baculovirus and after 48 h were labelled with 3"00~Ci/ml organic 32p for a further 2 h. The labelled medium was then removed, the cells were then pulsed with cold medium and harvested at different time points, and the protein profiles were monitored by autoradiography. The pattern observed with L2 is shown in (a). Lanes are as follows:0 h pulse (lane 1), 3 h pulse (lane 2), 6 h pulse (lane 3) and 24 h pulse (lane 4). The pattern observed with LI is shown in (b). Lanes are as follows: wild-type virus (lane l), 0 h pulse (lane 2), 30 rain pulse (lane 3), 60 min pulse (lane 4), 90 min pulse (lane 5) and 120 min pulse (lane 6). L1 and L2 proteins are indicated.
L1 and L2 form protein complexes Having shown that the HPV-16 capsid proteins could be expressed and post-translationally modified in the baculovirus expression system it was of interest to determine whether these proteins were capable of forming protein complexes. Since these proteins are the major components of the viral capsid it is likely that they will have a very close interaction in vivo. In addition there have been reports of using baculovirus to express other viral capsid proteins to obtain partial capsid assembly (French & Roy, 1990). For these studies the L1 and L2 genes were cloned into the dual expression transfer vector p A c U W 3 (Fig. 7). This was cotransfected with A c N P V D N A and recombinant clones were screened and selected as described above. One recombinant virus was obtained expressing good levels of both L1 and L2 proteins as determined by Western blot analysis (Fig, 8). In this expression system L1 appears to be produced in larger amounts than L2, which probably reflects the relative strength of the polyhedrin and p l 0 promoters. To investigate possible protein interactions between the L1 and L2 proteins, cells were infected with the double recombinant virus and extracts were made after
Ll
}~
~-L
L2
}7~
Fig. 7. Construction of recombinant transfer vector containing both HPV-16 L1 and L20RFs in the same vector. The fragments described in Fig. 1 were cloned into pAcUW3 as shown. L1 was cloned into the BgllI site in front of the baculovirus p 10 promoter. L2 was cloned into the BamHI site in front of the baculovirus polyhedrin promoter. The ampicillin resistance marker is also shown. The diagram is not to scale.
72 h. Firstly immunoprecipitations were done using a series of antibodies which have been shown previously to react specifically with their target antigen. Thus the antipeptide antibody directed to the last 14 amino acids of the HPV-16 L1 protein (Banks et aL, 1987), the anti cmyc antibody (Evan et al., 1985), the rabbit preimmune
2986
S . - Z . X i and L. M . B a n k s
(a) I
(b) 2
(c) 1
2
(a)
(d) 1
2
1
1
2
(b) 2
3
4
1
5
2
P7K
-97K
i8K -,68K L2
L1--~ -43K -43K
L1 -29K -29K Fig. 8. Co-expression of the HPV-16 LI and L2 proteins in recombinant baculovirus-infected cells. Cells were infected with recombinant virus and harvested after 72 h. The protein profile from recombinant virus is shown in (a) lane 2 with wild-type virus in lane 1. Western blot analyses are as follows: anti-L1 MAb in (b); polyclonal anti-L2 antibody in (c); preimmune antibody in (d). Lanes are as described for (a) and the positions of the L1 and L2 proteins are indicated.
antibody or the anti-HPV-16 L2 a n t i b o d y (Firzlaff et al., 1987) were used. The resultant i m m u n e complexes were then subjected to S D S - P A G E and proteins were analysed by Western blotting using an anti-HPV-16 L1 M A b . T h e results o b t a i n e d are shown in Fig. 9(a). Clearly the L 1 protein was coprecipitated by the anti-L2 antibody (Fig. 9a, lane 3). N o such result was obtained with either the c - m y c or p r e i m m u n e antibodies (Fig. 9a, lanes 1 and 2 respectively). T h e converse experiment was then performed using anti-L 1 antibody to precipitate the protein complex and then the Western blot was p r o b e d with anti-L2 polyclonal a n t i b o d y (Fig. 9b). T h e L2 protein was detected only after precipitation with the anti-L1 a n t i b o d y (Fig. 9b, lane 2); no L2 was precipitated with the p r e i m m u n e antibody (lane 1). There was no case in w h i c h we observed anti-L1 a n t i b o d y precipitating L2 protein or vice versa, f r o m baculovirusinfected cells expressing L2 or L1 alone (data not shown), nor did we see any cross-reaction on W e s t e r n blots (Fig. 8), thus confirming that the above results were not due to cross-reactivity of the antibodies. These results d e m o n strate that i m m u n o p r e c i p i t a t i o n with either anti-L1 or -L2 antibodies results in coprecipitation o f the other
Fig. 9. Coprecipitation of the HPV-16 L1 and L2 proteins. Cells were infected with recombinant baculovirus expressing both the L1 and L2 proteins. Cells were harvested after 72 h and immunoprecipitation followed by Western blot analysis was performed. (a) Recombinant virus-infected cells were immunoprecipitated with anti c-myc antibody (lane 1), preimmune rabbit serum (lane 2), anti-HPV-16 L2 antibody (lane 3) and anti-HPV-16 LI peptide antibody (lane 4). Total infected cell extract was run on lane 5. The imrnunoprecipitates were separated by SOS-PAGE, Western-blotted and probed with an anti-HPV-16 L1 MAb. '(b) Infected cells were immunoprecipitated with mouse preimmune serum (lane 1) or anti-HPV-16 LI MAb (lane 2). The immunoprecipitates were separated by SOS-PAGE, Western-blotted and probed with an anti-HPV-16 L2 polyclonal antibody.
capsid protein, a result which is mediated t h r o u g h a p r o t e i n - p r o t e i n interaction between the capsid proteins.
Discussion This paper describes the expression and characterization of the H P V - 1 6 capsid proteins, L I and L2, in a baculovirus expression system. This provides i m p o r t a n t biochemical information regarding potential processing and control o f these proteins which is currently unavailable from prokaryotic expression systems. Most importantly the results presented here clearly show the presence of protein complexes forming between t h e L1 and L2 proteins. Both proteins are m a d e in substantial quantities in the Sf9 cells, although L2 is clearly the most a b u n d a n t
Baculovirus expression of HP V-16 proteins species when both proteins are expressed from the same promoter. It is possible that L2 is more stable than the L1 protein and this is suggested to some degree by the phosphate labelling data. However, it is also possible that lower L1 expression is a result of the larger noncoding region prior to the ATG in the p36C-L1 construct. In the dual expression system L1 is produced in larger amounts and it seems likely that this is due to differences between the two promoters used to express L1 and L2. The expression of both L1 and L2 is adversely affected to a high degree by tunicamycin. This agent blocks Nlinked glycosylation and in this system LI appears to be weakly glycosylated, which is in agreement with previous reports (Browne et al., 1988). With L2, glycosylation is not as apparent, but there is a dramatic reduction in the amount of protein expressed. Clearly, however, glycosylation does not account for the aberrant high Mr of the L2 protein. This high Mr for the papillomavirus L2 protein has been observed for bovine papillomavirus type 1 (BPV-1) (Jin et al., 1989), BPV-2 (Rippe & Meinke, 1989), HPV-1 (Doorbar & Gallimore, 1987), HPV-6 (Rose et al., 1990) and HPV-16 shown here. Both the HPV-16 L1 and L2 proteins are phosphorylated. L2 has by far the highest number of phosphate residues of the two proteins, making L2 a very major phosphoprotein. The phosphorylation of L2 appears to be quite stable, as would be expected for a structural protein. Interestingly tunicamycin did not adversely affect the half-life of L2 (data not shown) implying that it simply inhibits L2 accumulation. This high level of phosphorylation implies an important function with possible interaction with viral DNA and a role in viral packaging and assembly. The L1 protein on the other hand is only weakly phosphorylated implying only a few phosphate groups per molecule. In addition, this phosphorylation appears to be more unstable with a half-life of around 60 min. It is possible that this is purely due to the expression system but it may also represent some mechanism for controlling the packaging of the viral DNA. The observation that the HPV-16 capsid proteins can physically interact in this system is particularly interesting. This demonstrates that we may expect to see at least some partial capsid assembly of the virus and further studies are now in progress. In addition this result is of value for our understanding of the mechanisms of capsid assembly. Clearly we can obtain interaction between the capsid proteins in the absence of any other additional HPV factors including either viral proteins and more importantly viral DNA. It thus seems likely that we may expect to see L1-L2 protein complexes forming within the HPV-16-infected cell prior to association with the viral DNA for packaging. As well as studies aimed at
2987
identifying any partial capsids which may be produced it will be interesting to determine the effects of the various post-translational modifications on the ability of these proteins to form complexes. In conclusion, we have been able to demonstrate that the HPV-16 L1 and L2 proteins expressed in a baculovirus system will interact to form protein complexes. In addition these proteins are subjected to a series of post-translational modifications including glycosylation and phosphorylation. Current studies are now aimed at further elucidating the mechanism of interaction between the L1 and L2 proteins. We are most grateful to Peter Ertl of the Wellcome Research Laboratories for invaluable help and advice in setting up the baculovirus expression system. We are also very grateful to Martin Page for his gift of the plasmid p36C, Ulrike Weyer for the generous gift of plasmid pAcUW3, to Denise Galloway for the anti-L2 antibody and to Lionel Crawford for the anti-L1 MAb.
References BANKS, L., MATLASHEWSKI,G , PIM, D., CHURCHER, M., ROBERTS, C. & CRAWFORD, L. (1987). Expression of human papillomavirus type 6 and 16 capsid proteins in bacteria and their antigenic characterization. Journal of General Virology 68, 3081-3089. BROWNE, H. M., CHURCHER, M. J., STANLEY, M. A., SMITH, G. L. & MINSON, A. C. (1988). Analysis of the L1 gene product of human papillomavirus type 16 by expression in a vaccinia virus recombinant. Journal of General Virology 69, 1263-1273. DOORBAR, J. & GALLXMORE, P. H. (1987). Identification of proteins encoded by the L1 and L2 open reading frames of human papillomavirus la. Journal of Virology 61, 2793-2799. EVAr~, G. I., LEWIS, G. K., RAMSEY, G. & BISHOP, J. M. (1985). Isolation of monoclonal antibodies specific for human c-myc proto oncogene product. Molecular and Cellular Biology 5, 3610-3616. FIRZLAFE, J. M., KIVIAT, N. B., BECKMANN,A. M., JENISON, S. A. & GALLOWAY,n . A. (1987). Detection of human papillomavirus capsid antigens in various squamous epithelial lesions using antibodies directed against the L1 and L2 open reading frames. Virology 164, 467-477. FRENCH, T. J. & ROY, P. (1990). Synthesis of bluetongue virus (BTV) corelike particles by a recombinant baculovirus expressing the two major structural core proteins of BTV. Journalof Virology 64, 15301536. HAWLEY-NELSON, P., VOUSDEN,K. H., HUBBERT, N. L., LowY, D. R. & SCmLLER, J. T. (1989). HPV E6 and E7 proteins cooperate to immortalise human foreskin keratinocytes. EMBO Journal 8, 39053910. JIN, X. W., COWSERT, L. M., PILACINSKI,W. P. & JENSON, A. B. (1989). Identification of L2 open reading frame gene products of bovine papillomavirus type 1 using monoclonal antibodies. Journal of General Virology 70, 1133-1140. KOMLY, C. A., BREITBURD,F., CROISSANT,O. & STREECK, R. E. (1986). The L2 open reading frame of human papillomavirus type la encodes a minor structural protein carrying type specific antigens. Journal of Virology 60, 813-816. MACLEAN, C. F., CHURCHER, M. J., MEINKE, J., SMITH, G. L., HIGGINS, G., STANLEY, M. & MINSON, A. C. (1990). Production and characterisation of a monoclonal antibody to human papillomavirus type 16 using a recombinant vaccinia virus. Journal of Clinical Pathology 43, 488-492. MATLASHEWSKI, G. (1989). The cell biology of human papillomavirus transformed cells. Anticancer Research 9, 1447-1556.
2988
S . - Z . X i and L. M . B a n k s
MATLASHEWSKI,G., SCHNEIDER,J., BANKS,L., JONES,N., MURRAY,A. CRAWFORD, t. (1987). Human papillomavirus type 16 DNA cooperates with activated ras in transforming cells. EMBO Journal 6, 1741-1746. MUNGER, K., PHELPS, W. C., BUBB,V., HOWLEY,P. M. & SCHLEGEL, R. (1989). The E6 and E7 genes of the human papillomavirus type 16 are together necessary and sufficient for transformation of primary human keratinocytes. Journal of Virology 63, 4417M421. PAGE, M. J. (1989). p36C: an improved baculovirus vector for producing high levels of mature recombinant proteins. Nucleic Acids Research 17, 454. PAR'I'ON, A. (1990). Nucleotide sequence of the HPV-16 L1 open reading frame. Nucleic Acids Research 15, 3631. PHELPS, W. C., YEE, C. L., MUNGER,K. & HOWLEY,P. M. (1988). The human papillomavirus type 16 E7 gene encodes transactivation and transformation functions similar to those of adenovirus E 1A. Cell 53, 539-547.
RIPPE, R. & MEINKE,W. (1989). Identification and characterisation of the BPV-2 L2 protein. Virology 171, 298-301. ROSE, R. C., BONNEZ, W., STRIKE, D. G. & REICHMAN,R. C. (1990). Expression of the full-length products of the human papillomavirus type 6b (HPV-6b) and HPV-II L2 open reading frames by recombinant baculovirus, and antigenic comparisons with HPV-11 whole virus particles. Journal of General Virology 71, 2725-2729. STOREY, A., PIM, D., MURRAY, A., OSBORN, K., BANKS, L. & CRAWFORD, L. (1988). Comparison of the in vitro transforming activities of human papillomavirus types. EMBO Journal 7, 18151820. VOUSDEN,K. (1989). Human papillomaviruses and cervical carcinoma. Cancer Cells 1, 43-50.
(Received 10 June 1991 ; Accepted 2 September 1991)
