| Journal of General Virology |
| SUMMARY | INTRO | METHODS | RESULTS | DISCUSSION | FOOTNOTES | REFS |
| First posted online 14 September 2000 | FULL-LENGTH ARTICLE |
| Rec 27 July 2000; Acc 6 September 2000 | DOI: 10.1099/vir.0.17314-0 |
Adrian P. Abbotts, Valerie G. Preston, Michelle Hughes, Arvind H. Patel and Nigel D. Stow
MRC Virology Unit, Institute of Virology,
Church Street, Glasgow G11 5JR, UK
The UL15 and UL28 proteins of herpes simplex virus type 1 are both required for the packaging of replicated viral DNA into the viral capsid. We have expressed UL28 and a functional epitope-tagged form of UL15 in mammalian and insect cells. Immunoprecipitation experiments confirmed that the two proteins can interact. In agreement with previous results, UL15, when expressed alone, entered the nucleus but UL28 remained cytoplasmic. When co-expressed the two proteins co-localized in the nucleus. Six UL28 deletion mutants were constructed and similarly analysed. The results obtained by immunoprecipitation and immunofluorescence were consistent and demonstrate that at least two separate regions of the UL28 polypeptide chain have the ability to interact with UL15. Surprisingly, three of the mutants prevented the UL15 protein from localizing to the cell nucleus, and these were not functional in a transient DNA packaging assay. Of the three UL28 mutant proteins that entered the nucleus with UL15, one containing an internal deletion of 13 amino acids was able to complement a UL28 null mutant in both DNA packaging and virus yield assays, demonstrating that this region of the protein is not essential for function. In addition to interacting with the UL28 protein we also demonstrated that UL15 molecules can interact with each other, and that sequences within the second exon contribute to this interaction.
Introduction |
The genomes of herpesviruses are linear
double-stranded DNAs of 125245 kbp which replicate in the nuclei of
infected cells generating tandem head-to-tail concatemers. During the
assembly of progeny particles, the excision of unit length molecules from
these concatemers is tightly coupled to their packaging into preformed
capsids. A single cis-acting DNA sequence element and at least six
proteins, which are well conserved throughout the virus family, appear to
play direct roles in the cleavage/packaging process (for a review see Homa
& Brown, 1997
). In herpes simplex
virus type 1 (HSV-1), an alphaherpesvirus, the packaging proteins are
encoded by genes UL6, UL15, UL17, UL28, UL32 and UL33. Mutants with
lesions in these genes are defective in both cleavage of concatemeric DNA
and DNA encapsidation (Homa & Brown, 1997
; Salmon et al., 1998
; Lamberti & Weller; 1998
). A seventh gene product, encoded by UL25, is not required
for cleavage but is apparently necessary for retention of DNA by the
capsid (McNab et al., 1998
), as well as functioning at an early stage during virus
entry (Addison et al., 1984
).
The functions of the individual proteins in HSV-1
DNA packaging remain poorly understood, but it is anticipated that their
roles will be similar to those of analogous proteins of double-stranded
DNA bacteriophage (for reviews see Catalano et al., 1995
; Fujisawa & Morita, 1997
; Catalano, 2000
). The latter proteins recognize the DNA to be packaged,
assemble a 'packaging complex' at a specific DNA entry site on a preformed
capsid, cleave the DNA at an appropriate site to initiate encapsidation,
inject it into the capsid and cleave again to terminate the process. The
packaging proteins of bacteriophage may be structural components of the
particle (e.g. portal proteins at the capsid vertex used for DNA entry),
transiently associated during encapsidation (e.g. the terminase enzyme) or
may perform catalytic roles in forming the various complexes. A pivotal
role is played by the terminase, generally comprised of two proteins,
which interacts with both the DNA substrate and the portal vertex,
functions as an ATP-driven pump to translocate the genome into the capsid
shell and carries out the cleavage reactions.
The precursor capsid for HSV-1 DNA packaging is
termed a procapsid and consists of an icosahedral assembly of capsid
proteins around a proteinaceous scaffold. B capsids of similar composition
and lacking viral DNA accumulate in HSV-1-infected cells. Packaging of DNA
into the procapsid results in the release of the scaffold and the
formation of a C capsid, which subsequently acquires tegument and envelope
to generate the virion (for reviews see Rixon, 1993
; Homa & Brown, 1997
). To date no portal vertex has been identified but the UL6
and UL25 proteins have been shown to be associated with B capsids, C
capsids and virions (Patel & Maclean, 1995
; Ali et al., 1996
; McNab et al., 1998
). In contrast the UL15 and UL28 proteins are found
predominantly in B capsids, although a modified form of UL15 may also be
present in C capsids and virions (Yu & Weller, 1998 a
; Salmon & Baines, 1998
; Taus & Baines, 1998
). Neither the UL32 nor the UL33 protein was detected in
any of the capsid forms (Lamberti & Weller, 1998
; Reynolds et al., 2000
). UL17 has been described as a virion protein located in
the tegument layer surrounding the capsid, but is also found in B and C
capsids (Salmon et al., 1998
; Goshima et al., 2000
).
The involvement of the UL15 and UL28 proteins in
HSV-1 DNA packaging was first demonstrated with temperature-sensitive
mutant viruses and confirmed through the characterization of null mutants
(Addison et al., 1990
; Poon & Roizman, 1993
; Cavalcoli et al., 1993
; Tengelsen et al., 1993
; Yu et al., 1997
; Baines et al., 1997
). Subsequently, several lines of evidence have suggested a
direct interaction between the two proteins, possibly analogous to that
between the subunits of bacteriophage terminase. Immunofluorescence
studies indicated that when expressed alone UL15 exhibited a nuclear
localization. In contrast, UL28, or the homologous protein of another
alphaherpesvirus, pseudorabies virus (PRV), remained cytoplasmic.
Co-expression of HSV-1 UL15 with the HSV-1 or PRV UL28 protein, however,
facilitated entry of the latter into the nucleus (Koslowski et al.,
1997
, 1999
). UL15 and UL28 were additionally demonstrated to
co-purify from HSV-1-infected cells, apparently as a heterodimer
(Koslowski et al., 1999
).
Further support is provided by studies utilizing a
betaherpesvirus, human cytomegalovirus (HCMV). Mutants resistant to
benzimidazole compounds that selectively inhibit processing and packaging
of HCMV DNA have been isolated and the increased resistance was shown to
result from alterations within the HCMV UL89 and UL56 proteins, the
homologues of HSV-1 UL15 and UL28, respectively (Underwood et al.,
1998
; Krosky et al., 1998
). Moreover, HCMV UL56 has been reported to bind
to the viral DNA packaging signal, a property in common with the small
subunits of several terminases (Bogner et al., 1998
). However, this observation awaits confirmation
and no similar activity has yet been attributed to HSV-1 UL28. Finally,
HSV-1 UL15 and its herpesvirus homologues show limited sequence similarity
to gp17, the large subunit of the terminase complex of bacteriophage T4
(Davison, 1992
). The similarity includes a
consensus ATP-binding site which has been demonstrated by site-directed
mutagenesis to be essential for UL15 function (Yu & Weller, 1998
b
).
Only one report has attempted to identify regions of
UL15 or UL28 important for interaction. These studies demonstrated that
the C-terminal 155 amino acids of the PRV UL28 protein were necessary for
its nuclear localization in cells superinfected with a PRV UL28 null
mutant or co-transfected with HSV-1 UL15 (Koslowski et al., 1997
). It is not known, however, whether the complex
formed between the heterologous proteins is functional in DNA packaging.
In this manuscript we extend these findings by using immunoprecipitation
assays to confirm the interaction between the HSV-1 UL15 and UL28
proteins, and examining the ability of a series of UL28 deletion mutants
to support DNA packaging and to interact with UL15.
Methods |
Cells and viruses. Baby hamster kidney 21
clone 13 (BHK) cells were grown in Glasgow MEM supplemented with 10 %
tryptose phosphate broth, 10 % newborn calf serum, 100 U/ml penicillin and
100 µg/ml streptomycin. Vero cells were grown in Dulbecco's MEM
containing 5 % foetal calf serum and the same antibiotics. Spodoptera
frugiperda (Sf) cells (strain IPLB-SF-21; Kitts et al., 1990
) were maintained in TC100 medium supplemented
with 5 % foetal calf serum and the same antibiotics. The HSV-1 UL28 null
mutant gCB was propagated on the Vero cell-derived complementing cell
line, C1 (Tengelsen et al., 1993
), and S648, a UL15 null mutant, was propagated in clone 17
cells (Baines et al., 1997
). The parental virus used for the construction of
recombinant baculoviruses was AcPAK6 (Bishop, 1992
). DNA fragments for expression (see below) were cloned
into the transfer vector pAcCL29-1 (Livingstone & Jones, 1989
) and recombinants were isolated through
recombination with Bsu36I-cleaved AcPAK6 DNA essentially as
described by Kitts et al. (1990
).
Expression of UL28 proteins. Plasmid pUL28
contains the HSV-1 DNA fragment spanning nucleotides 58182 (EagI
site; positions from McGeoch et al., 1988
) to 55761 (SgrAI site) inserted into the
SmaI site of the expression vector pCMV10 (Stow et al.,
1993
) such that full-length UL28 protein
is expressed under the control of the HCMV major immediate early (IE)
promoter. A panel of six UL28 gene deletions was made by utilizing
convenient restriction endonuclease sites within pUL28 (Table 1). The full-length UL28 fragment was cloned into
the SmaI site of the transfer vector pAcCL29-1 in the correct
orientation downstream of the polyhedrin gene promoter and the
corresponding
2,
3,
4,
5 and
6 deletions were introduced. Recombinant
baculoviruses were constructed using the resulting plasmids.
Table 1. UL28 mutants used in these studies
|
pCMV10 construct |
Recombinant baculovirus |
Region deleted* |
Amino acid sequence |
Predicted Mr |
|
pUL28 |
AcUL28 |
|
1785 (wild-type) |
85577 |
|
pUL28 |
Not constructed |
AgeI (56419)AgeI (56056) |
1579 and 701785 |
71903 |
|
pUL28 |
AcUL28 |
NcoI (57814)NcoI (57181) |
1115 and 327785 |
62560 |
|
pUL28 |
AcUL28 |
SmaI (56767)SmaI (56728) |
1464 and 478785 |
84241 |
|
pUL28 |
AcUL28 |
PstI (56026)PstI (ds) |
1712+LV |
77846 |
|
pUL28 |
AcUL28 |
SmaI (56767)PstI (ds) |
1464+LV |
50528 |
|
pUL28 |
AcUL28 |
EagI (us)SmaI (56728) |
MDTAD+478785 |
34474 |
* ds indicates a site in the poly-linker downstream of the UL28 ORF; us indicates a site upstream of the UL28 ORF.
In the
4 and
5
constructs an XbaI linker (New England Biolabs) with termination
codons in all three frames was inserted at the site of the deletion and
specifies the C-terminal amino acids LV. In the
6
constructs sequences from the N terminus of the HSV-1 UL8 gene derived
from plasmid pE8 (Stow et al., 1993
) and specifying a 14 bp upstream untranslated region and
the first five amino acids (MDTAD) were inserted at the site of the
deletion to provide an in-frame initiation codon.
Expression of UL15 proteins. To express
full-length UL15 protein, fragments specifying the first and second exons
were first assembled with a synthetic oligonucleotide so as to produce a
single fragment equivalent to a cDNA copy of the spliced mRNA (Dolan et
al., 1991
; A. Davison, unpublished data).
This fragment (nucleotides 2901030048 linked to 3363535094)
was cloned using BglII linkers into a derivative of pCMV10
(pCMV10-BglII) in which the SmaI site had been converted to
a BglII site. A plasmid containing the insert in the correct
orientation was designated pJM9. Plasmid pJM19, encoding the full-length
UL15 protein linked to a C-terminal epitope tag derived from the HCMV UL83
gene product (pp65) was constructed by inserting a synthetic
oligonucleotide duplex between the MluI site at position 34801 and
the XbaI site in the poly-linker of pJM9. The oligonucleotide was
designed so that an epitope tag sequence (ERKTPRVTGG) was added to the
authentic C terminus of UL15. To generate plasmid pMH20, which encodes a
tagged version of UL15 exon II, a fragment encoding the epitope tag and
most of exon II was first excised from JM19. This fragment was inserted
into pCMV10-BglII together with an oligonucleotide which specified
the remaining sequence and an in-frame ATG initiation codon. The above
UL15 gene fragments were also transferred to the baculovirus transfer
vector pAcCL29-1 and used to construct recombinant viruses. AcUL15,
AcUL15-pp65 and AcUL15E2-pp65 contain the inserts from pJM9, pJM19 and
pMH20, respectively.
Antibodies. Purified mouse monoclonal
antibody reactive with the HCMV pp65 epitope tag (anti-pp65) was purchased
from Capricorn Products (AntiCMV late nuclear protein). Rabbit polyclonal
antibody R123 was obtained following immunization with a bacterially
expressed protein representing amino acids 138785 of UL28. Mouse
monoclonal antibody 13924, reactive with the HSV-1 UL9 protein, was
described previously (Stow et al., 1998
).
Transient complementation yield assay.
Monolayers of BHK cells in 35 mm Petri dishes (2x106 cells per
plate) were transfected with pCMV10-derived expression plasmids by the
calcium phosphate procedure followed by treatment with DMSO at 4 h (Stow
& Wilkie, 1976
). Each monolayer
received 1 µg of the indicated plasmid and 12 µg calf thymus
carrier DNA. The transfected cells were infected with 5 p.f.u. per cell of
the appropriate HSV-1 null mutant in a volume of 200 µl. One hour
after virus addition the inoculum was removed and the infectivity of
residual virus was inactivated with an acidglycine wash (Rosenthal
et al., 1984
). The plates were washed
once with 0.14 M NaCl, exposed to 0.1 M glycine, 0.14 M NaCl pH 3.0 for 1
min, washed once with Eagle's medium containing 5 % newborn calf serum,
100 U/ml penicillin and 100 µg/ml streptomycin (EC5), and incubation
was continued for 18 h at 37 °C in 2 ml EC5. The cells were scraped
into the growth medium, sonicated and the yield of virus was titrated at
37 °C on both Vero cells and the appropriate complementing cell
line.
Transient complementation packaging assay.
Plasmid pSA1 was constructed by inserting a 200 bp fragment spanning the
junction between two tandem 'a' sequences between the HindIII and
EcoRI sites of pS1 (Stow & McMonagle, 1983
) which contains a copy of the HSV-1 oriS DNA
replication origin. In the presence of wild-type HSV-1 the pSA1 amplicon
is both efficiently replicated and packaged into virus particles,
confirming that the inserted fragment contains a functional packaging
signal (Nasseri & Mocarski, 1988
; P. D. Hodge, unpublished data). To examine whether
pCMV10-derived expression plasmids could complement the packaging defects
of S648 or gCB, monolayers of BHK cells in 35 mm dishes were transfected
with 1.0 µg expression plasmid, 0.5 µg pSA1 and 12 µg calf
thymus DNA, and infected with null mutant virus as described above, except
that the acidglycine wash was omitted. At 20 h post-infection (p.i.)
the cells from each monolayer were resuspended in TBS and divided into two
equal samples which were used to prepare total and DNase-resistant (i.e.
encapsidated) DNA as described previously (Stow et al., 1983
; Stow, 1998
). DNA samples were cleaved with EcoRI and
DpnI, fractionated by agarose gel electrophoresis, transferred to a
Hybond-N membrane (Amersham) and replicated (DpnI-resistant) pSA1
DNA was detected by hybridization to a probe prepared from the plasmid
vector pAT153. Phosphorimages of the Southern blots were acquired using
the Personal Molecular Imager and Quantity One software
(Bio-Rad).
Immunoprecipitation assays.
Immunoprecipitation assays were performed as described by McLean et
al. (1994
). Monolayers of Sf cells
(1.2x106 cells per 22 mm diameter tissue culture well) were
infected with 5 p.f.u. per cell recombinant baculoviruses and labelled
with [35S]L -methionine from 24 to 40 h
p.i. Soluble extracts (150 µl per well) were prepared and 130 µl
was incubated with 1 µl undiluted R123 or anti-pp65 antibody as
indicated. The immune complexes were collected on protein ASepharose
beads, washed and the proteins separated by SDSPAGE. To detect
labelled proteins, gels were either dried and subjected to phosphorimage
analysis or fixed, treated with En3Hance (Du Pont) and exposed to
autoradiographic film. Western blots were performed as described by Towbin
et al. (1979
). The membranes were
blocked at room temperature for 90 min using 5 % dried milk in TBS and
incubated with anti-UL28 rabbit serum R123 at a dilution of 1/200 in TBS
containing 0.1 % Tween-20 and 5 % dried milk (TBSTM). After 90 min, the
membrane was washed extensively with TBSTM, incubated for 30 min with
alkaline phosphatase-conjugated goat anti-rabbit IgG (Promega, 1/7500 in
TBSTM), washed again and bound antibody was detected using a BCIP/NBT
liquid substrate system (Sigma).
Immunofluorescence assays. Vero cells were seeded onto glass coverslips in Linbro wells (1.5x105 cells per 13 mm diameter coverslip) 1 day prior to lipofection. Each well received the indicated plasmids (total of 1 µg DNA) and 6 µl lipofectamine (Life Sciences) in 200 µl unsupplemented Dulbecco's MEM. At 16 h post-transfection the cells were fixed with 5 % formaldehyde in PBS containing 2 % sucrose, and permeabilized with 0.5 % NP-40 in PBS with 10 % sucrose. The primary anti-pp65 and R123 antibodies were diluted 1/500 and 1/200, respectively, in PBS containing 1 % foetal calf serum (PBSF). After incubation at room temperature for 1 h, the coverslips were washed at least six times with PBSF, then treated with both fluorescein isothiocyanate (FITC)-conjugated sheep anti-rabbit IgG (Sigma) and Cy5-conjugated goat anti-mouse IgG (Amersham), each diluted 1/200 in PBSF. After 30 min the coverslips were again washed with PBSF and mounted with AF1 (Citifluor). The coverslips were examined using a Zeiss LSM 510 confocal microscope system in conjunction with a Zeiss Axioplan 63x oil immersion objective lens (NA 1.4) and lasers with excitation lines at 488 and 633 nm. The two channels were scanned separately and the same settings maintained throughout. Captured images were exported and compiled using Adobe Photoshop.
Results |
Expression of functional UL15 and UL28 proteins
Virus yield complementation
assays were performed in order to determine whether the parental plasmids
used in these studies were capable of expressing functional UL15 and UL28
protein. Table 2 shows that in BHK cells transfected
with pJM9 or pJM19 (which express untagged and pp65-tagged versions of
full-length UL15, respectively) the yield of the UL15 null mutant S648
(titrated on clone 17 cells) was at least 500-fold greater than in control
cells that received pUL28. Although a low level of recombinant virus
capable of growth on Vero cells was detected (present in the S648 stock
and/or generated by recombination with pJM9 or pJM19), the increased
yields comprised predominantly mutant viruses, demonstrating that
complementation had occurred in the presence of the two plasmids.
Similarly, replication of the UL28 null mutant, gCB, occurred in cells
transfected with pUL28 or pUL28
3, but not in control cells transfected
with the vector pCMV10. These data indicate that the wild-type proteins
expressed by both pJM9 and pUL28 are able to support virus replication and
that the presence of the epitope tag on the protein encoded by pJM19 does
not significantly alter the activity of the UL15 protein.
Table 2. Complementation of null mutants by HSV-1 UL15 and UL28 expression plasmids
BHK cells were transfected and superinfected with the indicated combination of plasmid and virus and the progeny were titrated on the cell lines shown (p.f.u./ml). ND, Not determined.
|
Plasmid |
Virus |
Titre on clone 17 cells |
Titre on C1 cells |
Titre on Vero cells |
|
pUL28 |
S648 |
2.8x102 |
ND |
2.8x102 |
|
pJM9 (UL15) |
S648 |
1.7x105 |
ND |
1.4x103 |
|
pJM19 (UL15-pp65) |
S648 |
1.4x105 |
ND |
5.8x102 |
|
pCMV10 |
gCB |
ND |
<10 |
<10 |
|
pUL28 |
gCB |
ND |
3.7x104 |
<10 |
|
pUL28 |
gCB |
ND |
3.1x104 |
<10 |
Expression of mutated UL28 polypeptides in mammalian cells
Plasmids containing
deleted copies of the UL28 gene downstream of the HCMV major IE promoter
were constructed as described in Methods and Table 1.
The ability of the encoded proteins to function in DNA packaging was
compared to wild-type UL28 encoded by pUL28 in a transient DNA packaging
assay. BHK cells were co-transfected with the amplicon plasmid pSA1 and
one of the plasmids encoding wild-type or mutated UL28, and superinfected
with gCB. Total and packaged (DNase-resistant) DNA was prepared 20 h p.i.
and the presence of replicated pSA1 was detected by Southern blotting and
hybridization. Fig. 1 (b) shows that in each
instance pSA1 replicated to a similar extent. Packaging of the replicated
pSA1 DNA was undetectable in cells which received the control plasmid
pCMV10 or mutant plasmids pUL28
1, pUL28
2,
pUL28
4, pUL28
5 or pUL28
6. In contrast pSA1 was detectable in
the DNase-resistant DNA fraction from cells transfected with pUL28 or
pUL28
3. Both the wild-type protein and the pUL28
3
product lacking amino acids 465477 support DNA encapsidation to a
similar extent in this transient assay. In agreement with this pUL28
3 was
also able to support the replication of gCB in the yield complementation
assay (Table 2), while the other UL28 mutants did not
(data not shown).
Fig. 1. Ability of mutated UL28 proteins to
support DNA packaging. BHK cells were co-transfected with pSA1 and either
a UL28-expressing plasmid or the vector pCMV10 as indicated and
superinfected with gCB. At 20 h p.i. DNase-resistant DNA (a) and
total DNA (b) were prepared and analysed as described in Methods.
The positions of linear pSA1 molecules are indicated.
Co-immunoprecipitation of UL15 and UL28
Labelled extracts were prepared
from Sf cells infected with AcUL28, AcUL15-pp65 or the two viruses in
combination and precipitated with anti-pp65 antibody. Fig.
2 shows that UL15 was precipitated from both extracts prepared from
cells which received AcUL15-pp65. In contrast, precipitation of UL28 was
specifically dependent upon the presence of UL15. Since both UL15 and UL28
may be involved in binding to DNA, the experiment shown was performed in
the presence of 50 µg/ml ethidium bromide throughout the
precipitation and washing stages, which has been shown to selectively
inhibit DNA-dependent protein associations (Lai & Herr, 1992
). Subsequent experiments demonstrated that
essentially identical results were obtained in the absence of ethidium
bromide. Polyclonal antibody R123 against UL28 similarly specifically
co-precipitated UL15 and UL28 from extracts of cells co-infected with
AcUL28 and AcUL15, indicating that the presence of the pp65 epitope tag
was not responsible for the observed interaction (data not shown). These
results therefore support the earlier conclusion (Koslowski et al.,
1999
) that HSV-1 UL15 and UL28
interact.
Fig. 2. Co-immunoprecipitation of UL28 and UL15
proteins. Sf cells were infected with AcUL15-pp65 (15), AcUL28 (28) or the
two viruses together and labelled with [35S]methionine.
Extracts were reacted with anti-pp65 antibody and the immunoprecipitates
were analysed by SDSPAGE alongside samples of total cell proteins
and the starting extracts. Radioactivity in the dried gel was detected
with a phosphorimager. The positions of the UL15 and UL28 proteins are
indicated. The apparent faster migration of UL15 when precipitated in the
absence of UL28 than in its presence is a feature of this particular gel
and was not seen in other experiments.
Co-precipitation of mutant UL28 proteins with UL15
We next determined whether the
mutated UL28 proteins could interact with UL15. The UL28
2,
UL28
3, UL28
4, UL28
5 and UL28
6
proteins were expressed by recombinant baculoviruses and examined in
co-immunoprecipitation experiments. Extracts prepared from Sf cells singly
infected with these viruses or co-infected with AcUL15-pp65 were
precipitated with anti-pp65 antibody and analysed by SDSPAGE. Since
the protein encoded by AcUL28
4 co-migrates with the tagged UL15
protein, the immunoprecipitated proteins were in some instances detected
by Western blotting with antibody R123 against UL28. Fig.
3 (a) shows that wild-type UL28 and the UL28
2,
UL28
3 and UL28
4 proteins were not detected by Western blotting in
the control precipitates from singly infected cells but that all were
specifically precipitated in the presence of UL15. Similarly, Fig. 3 (b) demonstrates that
35S-labelled UL28
5 and UL28
6 were
detected by phosphorimage analysis in the immunoprecipitates of extracts
from cells mixedly infected with AcUL15-pp65 but not in the single
infection controls. Specific co-precipitation of the UL28
2 and
UL28
3 proteins with UL15 was also readily demonstrated by
phosphorimage analysis of 35S-labelled polypeptides (data not
shown).
Fig. 3. Immunoprecipitation of mutant UL28
proteins with UL15. Sf cells were infected with the indicated recombinant
baculoviruses [15, AcUL15-pp65; 28, AcUL28; 28
2,
AcUL28
2; 28
3, AcUL28
3; 28
4,
AcUL28
4; 28
5, AcUL28
5; 28
6,
AcUL28
6] and labelled extracts were immunoprecipitated with
anti-pp65 antibody. Total cellular proteins and the immunoprecipitates
were resolved by SDSPAGE. (a) The proteins were transferred
to nylon membranes, UL28-related proteins were detected with antibody R123
and the membranes were photographed. (b) The gels were dried and
radioactive proteins were detected with a phosphorimager.
These data therefore indicate that all five mutated
proteins retain the ability to interact with UL15. The observation that
both the UL28
5 and UL28
6 proteins interact suggests that at least two
separate regions of UL28, located within amino acids 1464 and
478785 may independently contribute to binding. The less efficient
co-precipitation of UL28
4 with UL15 (Fig. 3 a)
was probably due to lower solubility, rather than a weaker interaction per
se, although the actual contribution of amino acids 713785 to
binding remains to be determined.
UL15 molecules can interact with each other
A recombinant baculovirus expressing exon II of UL15 tagged with the pp65 epitope was used to test the possibility that UL15 molecules might interact with each other. Sf cells were infected either singly or in combination with AcUL15 and AcUL15E2-pp65, and extracts were prepared and immunoprecipitated with anti-pp65 antibody. Fig. 4 shows that the antibody precipitates the tagged exon II but not wild-type UL15 from singly infected cells. However, both the truncated and full-length proteins were detected following immunoprecipitation of the extract from mixedly infected cells. This demonstrates that UL15 molecules can interact not only with UL28 but also with each other, possibly enabling the formation of higher-order protein complexes.
Fig. 4. UL15 protein can interact with itself.
Sf cells were infected with AcUL15 (15), AcUL15E2-pp65 (E2) or the two
viruses together and labelled with [35S]methionine. Extracts
were reacted with anti-pp65 antibody and the immunoprecipitates were
analysed by SDSPAGE alongside samples of total cell proteins.
Radioactivity in the dried gel was detected with a phosphorimager. The
positions of the UL15 and exon II proteins are indicated.
Intracellular localization of UL15 and wild-type and mutant UL28 proteins
Initial
experiments were performed in order to confirm that the co-expression of
HSV-1 UL15 was necessary for the nuclear uptake of UL28 (Koslowski et
al., 1999
). Vero cells were transfected with
pJM19 (encoding UL15-pp65) and pUL28 either separately or in combination
using cationic liposomes. The cells were fixed and stained with a mixture
of anti-pp65 and R123 primary antibodies followed by mixed Cy5- and
FITC-conjugated secondary antibodies. Fig. 5
(ac) shows that when expressed alone UL15-pp65 was specifically
detected by excitation of the Cy5 fluor and localized to the nucleus. In
contrast, UL28 alone was specifically detected by excitation of the
FITC-conjugated antibody and exhibited a cytoplasmic localization (Fig. 5 df). In co-transfected cells both proteins
localized to the nucleus (Fig. 5 g, h)
where they exhibited a high degree of co-localization (merged image; Fig. 5 i). These data show that both proteins are
readily detected in co-transfected cells and confirm that UL28 is only
translocated to the nucleus in the presence of UL15.
Fig. 5. Intracellular localization of UL15 and
UL28 proteins. Vero cells were transfected with pJM19 encoding UL15-pp65
(ac), pUL28 (df) or both plasmids
together (gi). The cells were fixed, reacted with
mixtures of the primary and secondary antibodies as described and examined
by confocal microscopy. The three panels in each row (from left to right)
show for the same field of cells Cy5 fluorescence (red, detecting
UL15-pp65), FITC fluorescence (green, detecting UL28) and a merged image
of the two.
Similar experiments were performed to investigate
whether UL15 could translocate the mutated UL28 proteins to the nucleus
(Fig. 6). In cells which received pUL28
1,
pUL28
2, pUL28
3, pUL28
4, pUL28
5 or
pUL28
6 alone the mutated UL28 protein in each instance exhibited a
cytoplasmic localization similar to wild-type UL28 (Fig.
6 a, e, i, m, q, u).
Co-expression with UL15 resulted in nuclear uptake of the pUL28
3,
pUL28
5 and pUL28
6 proteins but not the other mutated UL28
polypeptides (Fig. 6 b, f, j,
n, r, v). As expected, the nuclear UL28 proteins
showed a high degree of co-localization with UL15 (Fig.
6 k, l, s, t, w, x).
Surprisingly, in the presence of the other three UL28 mutants (UL28
1,
UL28
2 and UL28
4) UL15 was retained in the cytoplasm, but again
extensively co-localized with the UL28 protein (Fig. 6
c, d, g, h, o, p).
Fig. 6.
Intracellular localization of UL15 and mutated UL28 proteins. Vero cells
were transfected with pUL28
1, pUL28
2,
pUL28
3, pUL28
4, pUL28
5 or pUL28
6
either alone (a, e, i, m, q, u)
or together with pJM19 encoding UL15-pp65 (bd,
fh, jl, np,
rt, vx), and the expressed
proteins were detected by confocal microscopy as described in the legend
to Fig. 5. The left panel of each row shows
representative staining of UL28 proteins with FITC in the singly
transfected cells. The next three panels show FITC staining (UL28
proteins), Cy5 staining (UL15-pp65) and a merged image of the two for the
same field of co-transfected cells. (y, z) Merged images of
FITC staining (UL28 proteins) and Cy5 staining (UL9) in Vero cells
co-transfected with pUL28 and pE9 (y) or pUL28
1 and
pE9 (z).
In order to exclude the possibility that these
mutated UL28 proteins might be having a non-specific inhibitory effect on
the nuclear uptake of proteins, they were co-expressed with HSV-1 UL9
which, like UL15, is known to be efficiently translocated into the nucleus
when expressed alone (Malik et al., 1996
). UL9 protein was detected with monoclonal
antibody 13924 and Cy5-conjugated secondary antibody, and UL28 was
detected with R123 and FITC-conjugated secondary antibody. Fig. 6 (y) and (z) indicate that in cells
co-transfected with pE9, expressing UL9 protein (Stow et al.,
1993
), and either pUL28 or pUL28
1,
respectively, localization of UL9 to the nucleus was unaffected, whilst
the UL28 protein remained in the cytoplasm. Identical results were
obtained in co-transfections with pUL28
2 and
pUL28
4 (data not shown), indicating that UL15 is specifically
retained in the cytoplasm of cells co-expressing these three mutated UL28
proteins. Each of the six mutated UL28 proteins therefore retains the
ability to interact with UL15, as indicated either by an ability to retain
UL15 in the cytoplasm or to be translocated with it into the
nucleus.
Discussion |
Koslowski et al. (1999
) previously demonstrated that HSV-1 UL28 required the
presence of UL15 in order to be translocated into the nucleus and that the
two proteins co-purified from HSV-1-infected cells, apparently as a
heterodimer. These data provided strong evidence for an interaction
between the two proteins. In this paper we have presented two additional
independent lines of evidence to support this hypothesis. First, UL28
specifically co-precipitated with UL15 from extracts of Sf cells infected
with recombinant baculoviruses (Fig. 2). Second, we
have extended the previous immunofluorescence data by demonstrating a high
degree of co-localization of the two proteins in cells in which they are
both expressed (Fig. 5). The significance of these
observations is enhanced by the demonstration that the cloned UL15 and
UL28 genes used in these experiments can support DNA packaging and virus
growth (Fig. 1 and Table
2).
Taken together with the earlier demonstration of an
interaction between PRV UL28 and HSV-1 UL15 (Koslowski et al.,
1997
), and the genetic support for an
interaction between the homologous proteins of HCMV (Underwood et
al., 1998
; Krosky et al., 1998
) there now seems little doubt that the two
proteins establish a functional interaction and that this is likely to be
conserved throughout the herpesvirus family. Circumstantial evidence,
outlined in Introduction, suggests that the two proteins may function in a
similar way to the bacteriophage terminases during packaging, but this
remains to be verified experimentally. Our demonstration that UL15
molecules have the ability to interact with one another (Fig. 4) is not inconsistent with this suggestion. In fact
the heterodimeric bacteriophage terminases probably form multimers as
complexes containing the DNA to be packaged, terminase enzyme and
preformed capsid are assembled (Fujisawa & Morita, 1997
; Catalano, 2000
). It will therefore be of interest to determine whether
UL15 and UL28 can generate larger complexes during packaging or if the
UL15UL15 and UL15UL28 interactions are mutually
exclusive.
This paper reports the first examination of the
ability of HSV-1 UL28 mutants to interact with HSV-1 UL15. The
intracellular localization of PRV UL28 mutants was previously examined by
immunofluorescence in cells co-infected with a PRV UL28 null mutant or
co-transfected with an HSV-1 UL28-expressing plasmid (Koslowski et
al., 1997
). These two approaches yielded
consistent results, although it should be noted that evidence for the
formation of a functional complex between the PRV and HSV-1 proteins is
lacking. The results of our immunoprecipitation and immunofluorescence
experiments (Figs 3 and 6) provide complementary lines of evidence that
five of the UL28 mutants (UL28
2, UL28
3,
UL28
4, UL28
5 and UL28
6) remain able to interact with HSV-1
UL15. A similar conclusion was reached for the sixth mutant, UL28
1,
based only on the fluorescence study. Since the regions of UL28 contained
within the UL28
5 and UL28
6 proteins are non-overlapping (amino acids
1464 and 478785, respectively) it would appear that at least
two separate regions of HSV-1 UL28 must be able to interact independently
with UL15. This result differs from the study with PRV UL28 mutants in
which the C-terminal 155 amino acids were required for binding to UL15 of
either PRV or HSV-1 (Koslowski et al., 1997
). A possible explanation is that the PRV protein contains
interacting sequences corresponding to those in the C-terminal but not the
N-terminal portion of the HSV-1 protein. However, the ability of several
of our UL28 mutants to retain UL15 in the cytoplasm raises the alternative
possibility that the PRV protein containing amino acids 1569 formed
a complex with UL15 which remained in the cytoplasm and was not detected
because the cells were not co-stained with an antibody that would
recognize the latter protein. Interestingly, the presence of separate
regions of UL28 able to interact with UL15 might possibly contribute to
the two proteins forming multimeric assemblies, as proposed
above.
Inspection of alignments of the amino acid sequences
of homologues of the UL28 protein encoded by alpha-, beta- and
gammaherpesviruses reveals that the regions of highest conservation are
confined to amino acids 1427 and 496785 of the HSV-1 sequence.
The intervening region not only shows poor sequence conservation but also
differs significantly in length between different viruses, suggesting that
it might possibly serve as a 'spacer' between separate domains of the
protein. Of the six mutants tested only UL28
3,
which contains the smallest deletion (13 amino acids), was able to support
virus growth and DNA packaging (Table 1 and Fig. 1) and interestingly its lesion is entirely within
this poorly conserved region.
Although the ability of the UL28
1,
UL28
2 and UL28
4 proteins to retain UL15 in the cytoplasm
constitutes strong evidence for an interaction, the mechanism by which
this occurs is not clear. Since these proteins did not inhibit UL9 nuclear
localization it is unlikely that they cause a non-specific inhibition of
nuclear transport. It is possible that the failure of their complexes with
UL15 to enter the nucleus results from misfolded regions masking a UL15
nuclear localization signal or causing decreased solubility of the complex
in the cytoplasm. The presence of distinct domains within UL28 probably
contributes to the mechanism by which these three proteins retain UL15 in
the cytoplasm, since in each instance one of the two postulated domains
remains intact and potentially able to fold correctly to provide an
interface for the proteinprotein interaction, even though the
remaining portions may be misfolded.
The inability of these three proteins to complement
the null mutant gCB for growth or DNA packaging can also be explained by
their failure to enter the nucleus, but it remains likely that the regions
deleted also contain residues directly involved in the packaging process.
The phenotypes of the UL28
5 and UL28
6 proteins indicate that sequences from
within both the putative domains contribute to the nuclear function of
UL28. In this regard it is interesting to note that residues 197225
of the HSV-1 protein contain a motif
(CX2CX8NXGX11CXH) which is conserved
throughout the mammalian and avian herpesviruses and may represent a metal
ion-binding region. It is hoped that site-directed mutagenesis of this and
other regions of HSV-1 UL28 will shed further light upon the functions and
interactions of this protein during DNA packaging.
We thank Fred Homa, Joel Baines and Andrew Davison for providing virus null mutants, complementing cell lines and the UL15 'cDNA' fragment. James McVicar, the recipient of a ROPA award from the Medical Research Council, provided excellent technical assistance and we are grateful to Duncan McGeoch for helpful comments on the manuscript.
References |
Bishop, D. H. L. (1992). Baculovirus expression vectors. Seminars in Virology 3, 253264.
Davison, A. J. (1992). Channel catfish virus: a new type of herpesvirus. Virology 186, 914.
Fujisawa, H. & Morita, M. (1997). Phage DNA packaging. Genes to Cells 2, 537545.
Rixon, F. J. (1993). Structure and assembly of herpesviruses. Seminars in Virology 4, 135144.
© 2000 SGM
This article is now available in the December 2000 print issue of JGV (vol. 81, 29993009). The complete issue of the journal may be seen in electronic form on JGV Online.