VIROLOGY
179,926-930
(1990)
Tumorigenic
Epidermal
WEN CHANG,*~’
COLIN
Poxviruses: Characterization
of the Expression of an
Growth Factor Related Gene in Shope Fibroma Virus
MACAULAY,t
SHIU-LOK Hu,* JAMES P. TAM,+ AND GRANT MCFADDEN+”
“Oncogene. 3005 First Avenue, Seattle, Washington 98 12 1; tDepartment of Biochemistry, University of Alberta,
Edmonton, Canada, T6G 2H7; and *The Rockefeller University, 1230 York Avenue, New York, 1002 1
Received May 17, 1990; accepted August 1, 1990
The transcription and translation of an epidermal growth factor (EGF) related gene in the Leporipoxvirus Shope
fibroma virus (SFV), termed the Shope fibroma growth factor (SFGF), have been characterized. Three early RNA
transcripts complimentary to an anti-SFGF oligonucleotide
were detected by Northern blot analysis, while no late
transcripts were expressed. The activity of the SFGF early promoter was measured using a transient gene expression
assay in SFV-infected cells using the bacterial chloramphenicol acetyltransferase as a reporter gene. Deletion analysis
showed that the functional SFGF promoter domain is an AT-rich sequence contained within 30 bp of the major transcriptional initiation site as is typical of early poxvirus promoters. An intracellular form of the SFGF gene product was
immunoprecipitated
from infected lysates using rabbit antisera raised against a synthetic SFGF (amino acids 26-80). A
16-kDa product was detected, while in cells infected in the presence of tunicamycin, the immunoprecipitated
product
had a mobility on SDS-polyacrylamide
gels of approximately 6 kDa, indicating that the SFGF gene product is extensively post-transcriptionally
modified. The intracellular 16-kDa form can be pulse-chased to a 14-kDa form but the
secreted form of SFGF could not be detected in the medium using this anti-peptide antiserum.
o 1990 Academic
PWSS, IIIC.
proliferate. This proposal was further strengthened
when it was found that both the Orthopo~virus vaccinia
and the Leporipoxvirus SFV contain genes which encode an epidermal growth factor (EGF) related protein
(4-9). The vaccinia growth factor gene (VGF) is transcribed at early times (IO), and the gene product has
been purified from the medium of infected cells. Amino
acid sequence analysis of the secreted VGF and immunoprecipitation
using anti-VGF peptide-specific rabbit antisera has demonstrated that VGF is initially synthesized as an N-linked glycosylated membrane-associated precursor which is cleaved and finally released
as a secreted molecule. (8, 1 I). Although vaccinia virus
does not induce tumors, Buller eta/. (12, 13) found that
when the VGF gene was inactivated by targeted insertional mutagenesis,
the VGF- virus was less pathogenic as measured by intracranial injections in mice.
Furthermore, when inoculated onto the chorioallantoic
membrane of fertilized chicken eggs, the deletion variant induced less proliferation of the surrounding uninfected ectodermal and endodermal tissues than did
the wild-type virus (12, 13).
As a first step in investigating the function of the
Shope fibroma virus growth factor (SFGF) in a biology
of a prototype tumorigenic poxvirus, we have examined the expression of this gene at the transcriptional
and translational levels.
Northern blots of total cellular RNA from SFV-infected RK-13 cells were hybridized with a 32P-end-labeled oligonucleotide
probe complimentary
to the
The Poxviridae family comprises a very large and diverse group of viruses that have been isolated from
most vertebrates and some insects (for review see Ref.
(1)). The vertebrate poxviruses which induce a hyperproliferative response are not restricted to one genus
but are a collection of viruses from various genera
whose common feature is to induce a benign proliferation of tissue at the site of infection in their natural
hosts. Members of this group include the Yatapoxvirus
Yaba monkey tumor virus, the unclassified human poxvirus Molluscum contagiosum, and the Leporipoxvirus
Shope fibroma virus (SFV). SFV, the most well-characterized member of the tumorigenic poxviruses, causes
benign proliferation of connective tissue at the site of
infection when inoculated into the skin of an adult rabbit (reviewed in Ref. (2)). Subsequent induction of both
the humoral and cell-mediated
immune responses
against viral antigens allows the rabbit to overcome the
infection, and the associated tumor completely
regresses in an immunocompetent
host.
Although the mechanism of tumor formation in SFVinfected animals is not yet clarified, Kato et al. (3) proposed that a growth promoting factor of either cellular
or viral origin is released from the SFV-infected cells
and induces the neighboring uninfected fibroblasts to
’ Present address:
Center for Cancer Research
and Department
of Biology,
Massachusetts
Institute
for Technology,
77 Massachusetts Avenue,
Cambridge,
MA 02139.
’ To whom requests
for reprints
should be addressed.
0042-6822190
$3.00
CopyrIght 8 1990 by Academic
Press. Inc.
All rights of reproduction
in any form reserved.
926
SHORT
COMMUNICATIONS
FIG. 1. Northern
blot analysis
of SFGF transcripts
from SFV-infected
RK-13 cells. RK-13 cells were infected
with SFV and total
cellular
RNA was extracted
at various
times postinfection
(p.i.) as
indicated.
Total RNA (20 pg) was electrophoresed
on a 1% formaldehyde agarose
gel, transferred
to a nylon membrane
(N-Hybond).
and
probed with an antisense
SFGF 32P-end-labeled
oligonucleotide
(5’.
TTGTAAACGAGAAGAGTTATTATGCGGAATATACGTATTATCAATGGCATGGTTATGCACTTAATA-3’).
The sizes of the RNA
markers
are indicated
at the left of the figure.
SFGF gene, and three major RNA species ranging in
size from 2.4 to 3.2 kbp were detected, with the smallest species being the most abundant (Fig. 1). All three
species were detected at 2,4, and 6hr, but not at 24 hr
postinfection,
consistent with these being early mRNA
transcripts.
Although the function of early promoter sequence has been conserved
between the lepori- and
the ortho-poxviruses
(14) the timing of the switch between early and late gene expression
has not. Early
transcription
occurs prior to DNA synthesis,
which in
vaccinia is at 2-3 hr postinfection (I!?), whereas in SFV
it is at 8-9 hr postinfection
(76). Thus, at 6 hr postinfection early transcripts
are abundant in SFV-infected ceils
and absent in vaccinia-infected
cells (14). It is possible
that either all three species are 5’-coterminal
SFGF
mRNA, with the larger two species occurring as a result of read-through
of early transcriptional
termination
signals as is common among early poxvirus genes, or
they may represent RNAs initiating from the promoter
of the upstream and overlapping ORF Tl l-L. In either
case, it is clear that all the species are expressed
as
early RNAs only.
A second antisense SFGF oligonucleotide
was hybridized to total early RNA from SFV-infected cells, and
the hybrids were extended with reverse transcriptase,
treated with RNase A, and then electrophoresed
on an
8% sequencing gel (Fig. 2). Of the two extended products observed, the major product occurs at a thymine
927
in the sense sequence 75 nucleotides upstream of the
initiating AUG codon for the ORF and a second minor
extended product corresponds
to a C residue three
nucleotides upstream of the major start site. Relatively
few early poxvirus transcripts
have 5’ ends which map
to pyrimidines.
Given the fact that Boone and Moss
(17) found that the initiating position of bulk early RNA
consisted
of 55% G and 459/o A residues, while Davison and Moss (19) have shown that substituting
a pyrimidine for the initiating purine residue results in a
change of initiation sites to the next purine, it is possible that these designated pyrimidine residues are not
the true sites of initiation of transcription
for these RNA
species. For example, the structure
of the cap of the
RNA itself could affect the precise mapping of the 5’
end of the transcript.
If the reverse transcriptase
did
not fully extend the primers and failed to copy the last
nucleotide then the initiating nucleotide for both transcripts would be a purine (A).
In the Sl analysis of Macaulay et al. (18) this SFGF
transcript was detected only at late times. Further work
has shown that the SFV infections used in that work
were not of sufficiently high m.o.i. to ensure that a second round of infections could not occur at late times of
infection. For example, Sl analysis of the Tl-ORF in
SFV originally indicated that the identical start site was
1
g$
‘&
CTACGTAG
i
C
C
FIG. 2. Determination
of the 5’ transcriptional
start of the SFGF
gene by primer extension
analysis.
A 32P-end-labeled
oligonucleotrde (YTAGAGAGGCCACTAGGTT3’)
was hybridized
to total RNA
(10 pg), from mock- or SFV-infected
RK-13 cells (4 hr postinfection).
The products
of the reverse transcriptase
reaction were analyzed
on
an 8% urea-acrylamide
gel beside a dideoxy
sequencing
reaction
using the same oligonucleotide
primer. The sense sequence
on the
right illustrates
the position
to whrch the major (T) and minor (C)
extended
products
(arrows)
correspond.
928
SHORT
COMMUNICATIONS
used at both early and late times. However, we later
showed that at higher multiplicites the Tl promoter is
clearly shut off at late times (14). In the work presented
here the m.o.i. was 10 and no late transcripts
were
detected by Northern blot analysis at late times (Fig. 1)
or by primer extension analysis (not shown), thus supporting the contention that SFGF is a bona fide early
gene.
To characterize
the 5’ sequence boundaries of the
SFGF early promoter the minimal upstream sequences
which allow the promoter to function in a transient CAT
gene expression
assay were determined (20). The putative SFGF early promoter was first isolated on a 372bp Bgll-Hincll
DNA fragment
spanning
the region
-336 to +36 relative to the transcriptional
initiation site
and cloned into a CAT expression
vector to create
pSFV-CAT. This clone was then used to generate a
series of 5’ deletions that span the putative promoter
sequence and several of these were then used as substrates
for oligonucleotide-directed
mutagenesis.
These various constructs
and their activities in the
transient CAT expression
assays are shown in Fig. 3.
Deletion of the sequences
between -336 and -43
did not significantly alter the amount of CAT activity
(Fig. 3, upper panel), and further deletion from -43 to
-25 reduced CAT activity to 77% that of the full-length
clone (Fig.; 3, Del. -25). However, deletion of the ATrich region between -25 and -15 completely eliminated CAT activity (Fig. 3, Del. -15). When an 18-bp
oligonucleotide
containing
the sequences
between
-43 and -25 was inserted into the Del-l 5 construct to
give Mut-1 , only 12% of the CAT activity was regenerated, showing that sequences
between -25 and -15
define the 5’ boundary of the SFGF promoter. The importance of this region is demonstrated
by the Mut-2
construct
in which the sequence
between
-25 and
-15 is inserted adjacent to the 5’ initiation site. This
clone is able to restore 659/o of the wild-type activity
indicating that the sequences
deleted by Del-l 5 are
sufficient to provide substantial promoter function. The
sequences
between -14 and +1 were able to augment activity only by a further 13 to 77% (Fig. 4, Del25). Within the SFGF promoter is an A-rich octanucleotide that is present as two partially overlapping direct
repeats between -23 and -10 (underlined in Fig. 3B).
The importance of these A residues is demonstrated
by the Mut-3 construct,
where two A residues are inserted into Del-l 5 to reconstruct
the octanucleotide,
in
which CAT expression
is increased from 2 to 72% of
wild-type
levels. These results are consistent
with
previous findings that A-rich regions are important
early promoter elements (74, 79, 21).
To analyze the SFGF gene product we utilized a rabbit antiserum
raised against
a synthetic
peptide
. . . . . . . . . . . . . . . . ..RR~ATRTRATRCGRCATRCTRTRRTA
-------_
. . . . . . . . . . . . . . . . . . . . . . . . . . ..ATRTRRTR~GR~~T~~TRT~~~~
77
Del
-25
2
Del
-15
TGRTRGGRTTTRTTCRGG..........RTRTARTAEGRCRTRCTATRATR
12
nut-1
. . . . . . . . . . . . . . . . ..Gfl~...............CTRTRATR
65
nut-2
. . . . . . . . . . . . . . . . . . . . . . . . ..RRRTRTARTRCGRCATRCTRTRRTR
- - - - ----
63
nut-3
TR
0.7
Dal
..,..,,...,...............................,......
l6
FIG. 3. Mutational
analysis
of the SFGF early promoter.
(Upper
panel) Transient
CAT-gene
expression
assays of the various deletion
mutants of the SFGF promoter.
The Control lane denotes uninfected
RK-13 cells that have been transfected
with pSFV-CAT.
The position
of each deletion
mutation
is represented
on the line drawing.
+l is
defined as the “C” residue corresponding
to the first RNA start site,
and t33 is the initiating
ATG for the CAT gene. (Lower panel) The
sequences
of various
promoter
mutants
and their corresponding
CAT activities.
The major and minor RNA start sites are indicated
by
asterisks
(*), and an AT-rich
octanucleotide
sequence
that is repeated twice in the wild-type
sequence
is underlined.
The numbers
immediately
to the right of the sequences
indicate the percentage
CAT activity of the mutant
compared
to pSFV-CAT.
The name of
each mutant construct
is given on the far right.
that contains the primary translated
SFGF
amino acid sequences
from isoleucine (26) to tyrosine
(80) (22). When this peptide is correctly folded it is biologically active in growth factor assays and it competes
with EGF for binding to the EGF receptor (22, 23).
Western blot analysis was used to successfully
demonstrate the reactivity of the rabbit antiserum toward
both the synthetic peptide and a bacterial fusion protein consisting of the 37-kDa N-terminal portion of the
Escherichia co/i trp E protein fused to the entire 80-aa
coding sequence of SFGF in a PATH expression
vector. However, all attempts using Western blot analysis
to detect SFGF in lysates or in the conditioned medium
of SFV-infected
cells with this anti-peptide
antibody
have been negative (not shown).
To increase the sensitivity of detection we attempted
SFGFzs-so
SHORT
[I
~2~31415~6i7~6~9~l0~1I~12~13~14
1 - 1 + 1 - 1+ 1 - 1 + 1 -
CJGF(&5-8,,)
1 - 1t
I + I -
1+ 1-
COMMUNICATIONS
1 +
FIG. 4. Pulse-chase
analysis of SFGF. Total cell lysates were prepared from RK-13 cells that were infected
with SFV (m.o.i.
20),
pulsed with [%Jcysteine
for 15 min, and chased with normal medium for the various times indicated.
Tunrcamycin
(10 pglml) was
used to block N-terminal
glycoslyation
In lanes 1 and 2. Antiserum
preadsorbed
with the SFGF,,,O
peptide was used in the even-numbered lanes, while untreated
antiserum
was used to immunopreciprtate SFGF. Molecular
weight markers
are indicated
to the left of the
gel. The Infected
cells were homogenized
in RIPA buffer supplemented wrth sodium dodecyl sulfate (SDS) to 5% (150 mM NaCI, 10
mlLl Tris, pH 7.5, 1% Triton X-100,
1% Na-deoxycholate.
and 5%
SDS). This lysate was boiled for 5 min to denature
the protein and
then diluted with RIPA (minus SDS) to give a final SDS concentration
of 1%. AntI-SFGF,,~,,
antiserum
was added to a final dilution
of
1: 100 and Incubated
at 4°C for 1 hr, protein
A-Sepharose
was
added and incubated
on ice for 30 min, and the Sepharose
beads
were pelleted
In a microfuge.
The beads were washed
with RIPA
minus SDS and then dissolved
In Laemmli gel loading buffer (27).
to immunoprecipitate
SFGF from SFV-infected
cells.
No polypeptide was specifically immunopecipitated
by
the antiserum using standard conditions (not shown).
In view of the fact that the antiserum
recognizes
both the denatured fusion protein and the synthetic
blot analysis, we modified the
SFGF,a-a, by Western
immunoprecipitation
procedure by first boiling the cell
lysate in 5% SDS and then adding the antiserum in the
presence of 1 O/oSDS to see if the denatured form of
SFGF would be recognized.
A 16-kDa polypeptide
could now be specifically
immunoprecipitated
from
SFV-infected cells (Fig. 4, lane 3) and this immunoprecipitation was blocked by preincubating
the antiserum
with the synthetic SFGF,,.*, (Fig. 4, lane 4).
The calculated molecular weight of the full-length
SFGF is 10,974 Da and, with the removal of a putative
1 g-amino-acid
hydrophobic
signal sequence,
would
be reduced to 8575 Da. Given that there is a difference
between this calculated molecular weight and the observed mobility of immunoprecipitated
SFGF, we investigated whether SFGF might be glycosylated
similarly to VGF (8, 1 I), by preincubating the cells in tunicamycin, which inhibits N-linked glycosylation (Fig. 4,
929
lanes 1 and 3). A broad 6-kDa band is specifically immunoprecipitated
in the presence of the inhibitor,
which is much closer to the calculated molecular
weight of the SFGF reading frame.
Pulse-chase analysis (Fig. 4, lanes 3-14) shows
that a 16- and an 18-kDa band are recognized by the
SFGF,,-,, antibody. Although the 16-kDa band is
chased to a 14-kDa form we have been unable to detect secreted SFGF by immunoprecipitation with this
synthetic peptide antibody. We interpret these results
to indicate that, like VGF, SFGF is an early viral gene
product that is post-translationally modified byglycosylation with the possible removal of a putative hydrophobic N-terminal signal sequence to give a final 16-kDa
form as the most abundant form within the infected
cell. This 16-kDa form is further processed to give a
14-kDa intracellular form; however, which of these
forms is actually secreted will require analysis by antibody reagents which recognize the native conformation of the mature growth factor (in progress).
Inspection of the SFGF sequence (1 I) shows that,
unlike other members of the EGF family, including VGF
(8), EGF (24) and TGF-a (25, 26) SFGF is not first made
in a pre-pro-form, and there is no obvious C-terminal
putative membrane spanning hydrophobic domain.
Therefore, our inability to detect a secreted form of
SFGF with the synthetic peptide antibody may be due
to the fact that it is relatively cell associated or, alternatively, that this antiserum does not recognize the final
peptide product because post-translational processing masks the major antigenic site. An activity which
competes with EGF for binding to its receptor has been
detected in both SFV-infected cell lysates and conditioned medium (W. Chang, unpublished results), but
because the levels of production of SFGF protein are
so low it has not been practical to purify and sequence
this activity directly as was done with VGF (8).
One way to address the question of the role of SFGF
in infected host tissues would be to inactivate the
SFGF gene by the targeted insertion of a marker gene.
Preliminary studies on the function of such deletion
mutations of EGF-related genes in several Leporipoxviruses indicates that this gene family is not solely responsible for the proliferative response at the site of
infection but their deletion profoundly affects the virulence of the viruses in the infected animal (Opgenorth
et a/., in preparation). The analysis of the role of these
growth factors in the pathogenesis of the tumorigenic
poxviruses should shed light on the biological functions of other members of the EGF receptor ligand family as well.
ACKNOWLEDGMENTS
We thank Scott Austin for preparing
man for the PATH expression
vector,
oligonucleotides.
Paula Traktand S. Kasinec for help with the
930
SHORT
COMMUNICATIONS
manuscript
preparation.
G.M. and CM. are supported
by the Alberta
Heritage
Foundation
for Medical
Research
and the NCI of Canada.
REFERENCES
1. Moss,
B., In “Fields
Virology”
(B. N. Fields, Ed.). 2nd ed. pp
2074-2111.
Raven Press, NY, 1990.
2 MCFADDEN,
G., ln “Virus
Diseases
in Laboratory
and Captive
Animals”
(G. Darai, Ed.), pp. 37-62.
Martinus
Nijhoff, Boston,
1988.
3. KATO, S., ONO, K., MIYAMOTO,
H., and MANTANI, M., Biken J. 9,
51-61
(1966).
4. BLOOMQUIST, M. C., HUNT. L. T., and BARKER, W. C.E, Proc. Nat/.
Acad. Sci. USA 81, 7363-7367
(1984).
5. BROWN, J. P., TWARDZIK, D. R., MARQUARDT, H., and TODARO, G. J.,
Nature (London)
313, 49 l-492
(1985).
6. CHANG, W., LIM, J. G., HELLSTROM, I., and GENTRY, L. E., J. Viral.
62, 1080-1083
(1988).
7. REISNER, A. H., Nature (London)
313, 801-803
(1985).
8. STROOBANT, P., RICE, A. P., GULLICK, J. W., CHENG, D. J., KERR,
I. M., and WATERFIELD, M. D., Ce//42,
383-393
(1985).
9. TWARDZIK, D. R., BROWN, J. P., RANSHALIS, J. E., TODARO, G. J., and
Moss, B., Proc. Nat/. Acad. Sci. USA 82 5300-5304
(1985).
10. VENKATESAN, S., BAROUDY, B. M., and Moss, B., Cell 25, 80%
813 (1981).
Il. CHANG, W., UPTON, C., Hu, S. L., PRUCHIO, A. F., and MCFADDEN,
G., Mol. Cell. Biol. 7, 535-540
(1988).
72. BULLER, R. M. L.. CHAKRABARTI, S., COOPER, J. A., TWARDZIK, D. R.,
and Moss, B., J. Viral. 62, 866-874
(1988).
73. BULLER, R. M. L., CHAKRABARTI, S., Moss, B., and FREDRICKSON,
T., virology
164, 182-192
(1988).
14. MACAULAY,
C., and MCFADDEN,
G., Virology
721, 237-246
(1989).
15. TRAKTMAN, P., In “Curr.
Topics in Microbiology
and Immunology” (R. W. Moyer,
Ed.), Vol 163, pp. 93-123.
Springer-Verlag, Berlin, 1990.
16. DERANGE, A. M., MACAUIAY,
C., BLOCK, W., MUELLER, T., and
MCFADDEN,
G., J. Viral. 50, 408-416
(1984).
17. BOONE, R. F.. and Moss, B., J. MOL. 26, 554-569
(1977).
18. MACAULAY,
C., UPTON, C., and MCFADDEN,
G., L/iro/ogy
158,
381-393
(1987).
19. DAVISON, A. J., and Moss,
B., 1. Biol. Chem. 210, 749-769
(1989).
20. COCHRAN, M. A., MACK~,
M., and Moss, B., Proc. Nat/. Acad.
SC;. USA 82, 1 S-23 (1985).
21. YUEN. L., DAVISON, A. J., and Moss, B., Proc. Nat/. Acad. Sci.
USA 84, 6069-6073
(1987).
22. LIN, Y. H., CAPORASO, G.. CHANG, P. Y., KE, X. H., and TAM, J. P.,
Biochemistry
27, 5640-5645
(1988).
23. YE, Y., LIN, Y., and TAM, J. P., Biochem.
Biophys. Res. Commun.
154,497-501
(1988).
24. SCOT,
J., URDEA, M., QUIROGA, M., SANCHEZ-PESCADOR,
R.,
FONG, N., SELBY. M., RU‘TTER, W. J., and BELL, G. I., Science
221, 236-240
(1983).
25. LEE, D. C.. ROSE, T. M., WEBB, N. R., and TODARO, G. J., Nature
(London)
313,489-491
(1985).
26. TODARDO, G. J., LEE, D. C., WEBB, N. R., and BROWN, J. P., ln
“Cancer
Cells,” Vol. 3, pp. 51-58.
Cold Spring Harbor Laboratory, Cold Spring Harbor,”
NY.
27. LAEMMLI, U. K.. Nature (London)
227, 680-685
(1970).