- Email: [email protected]
Characterization of different tumor antigens present in cells transformed by simian virus 40
Cell, Vol. 16, 335-346,
October
1979,
Copyright
0 1979
by MIT
Characterization of Different Tumor Antigens in Cells Transformed by Simian Virus 40
Alan E. Smith, Ros Smith and Eva Paucha’ Translation Laboratory Imperial Cancer Research Fund London WC2A 3PX, England
Summary In addition to large T and small t antigens, cells transformed by simian virus 40 (SV40) commonly contain other proteins which specifically immunoprecipitate with SV40 anti-T serum and which are not detected in untransformed cells. The additional tumor antigens (T-Ags) fall into two groups: those having a close structural relationship with normal SV40 T-Ags, and those unrelated to large T and small t. The latter are probably nonviral T-Ags (NVTAgs). The NVT-Ags comprise a family of proteins of molecular weight 50,000-55,000. Fingerprint analysis shows that NVT-Ags have few if any peptides in common with large T or small t, and that they lack the amino terminal tryptic peptide and the peptides unique to small t. NVT-Ags from different species have different fingerprints, but those isolated from different transformants of the same cell line are identical. The size of NVT is unaltered in cells transformed by mutants of SV40 with deletions in the region 0.50-0.55 map units. The mRNA for NVT does not hybridize to SV40 DNA. The other forms of T-Ag isolated from transformed cells fall into three classes: shortened forms of large T (truncated large T); multiple species of T-Ag with molecular weights very similar to, but distinct from, those of normal large T (large T doublets and triplets); and elongated forms of large T (super T). These proteins all contain the normal amino terminus of SV40 T-Ags, and the truncated forms of large T lack peptides from the carboxy terminal half of large T. One species of super T (molecular weight 130,000) contains only those methionine tryptic peptides present in normal large T, although it may contain some peptides in more than one copy. Introduction Monkey cells productively infected with simian virus 40 (SV40) contain at least two major virus-coded early proteins (Prives et al., 1977; Rundell et al., 1977; Crawford et al., 1978; Paucha et al., 1978b). These proteins, or tumor antigens (T-ASS), are referred to as large T and small t and have molecular weights of 90,000-l 00,000 and 15,000-20,000, respectively. Fingerprint analysis has shown that large T and small t share some peptide sequences, although both proteins also have unique regions (Simmons and Martin, l Present address: Department of Biochemistry, Imperial Science and Technology, London SW7. England.
Cottege
of
Present
1978; Smith, Smith and Paucha, 1978). The shared region has been located at the amino terminal end of the two proteins and is coded for by DNA sequences which extend between 0.65 and 0.60 map units (mu) on SV40 DNA (Paucha et al., 1978b; Volckaert, Van de Voorde and Fiers, 1978). The unique region of small t has been mapped between 0.60 and 0.55 mu and the unique region of large T between 0.54 and 0.17 mu (Paucha and Smith, 1978; Crawford et al., 1978; Fiers et al., 1978; Reddy et al., 1978). Mouse cells productively infected with polyoma virus (Py) contain early proteins which appear to correspond approximately in molecular weight and sequence arrangement to SV40 large T and small t (Feunteun et al., 1976; Ito, Spurr and Dulbecco, 1977a; Schaffhausen, Silver and Benjamin 1978) and also contain at least one other early protein which is present in major amounts. This is referred to as middle T and has a molecular weight in the range 50,000-60,000 (Ito, Brocklehurst and Dulbecco, 1977b). Py middle T shares amino terminal sequences with Py large and small T-Ags, but because of an alternative splicing pathway the sequences at its carboxy terminus are unique (Smart and Ito, 1978; Hutchinson, Hunter and Eckhart 1978; Soeda et al., 1979). Studies on the T-Ags present in cells transformed by wild-type SV40 indicate that they usually contain SV40 large T and small t. In some cell lines other forms of T-Ag have also been detected. These include proteins considerably larger than large T (Lichaa and Niesar, 1977; Kress, De Vaux Saint Cyr and Girard, 1977; Kress et al., 1979) multiple species of proteins with mobility similar to that of large T (Prives, Gluzman and Winocour, 1978) and proteins having an apparent molecular weight intermediate between those of large T and small t, with a prominent component of 5055,000 daltons (Chang et al., 1977; Schmidt-Ullrich et al., 1977; Gaudray, Rassoulzadegan and Cuzin, 1978; Lane and Crawford, 1979; Linzer and Levine, 1979; Edwards, Khoury and Martin, 1979; Chang et al., 1979; Kress et al., 1979; Melero et al., 1979). In the experiments reported here we have analyzed the different forms of T-Ag present in a number of transformed cell lines by tryptic peptide fingerprinting and by cell-free synthesis. We have asked three questions: are the different forms of T-Ag coded for by SV40 DNA? What is their relationship to SV40 large T and small t? Do SV40-transformed cells contain a protein equivalent to the polyoma virus early protein middle T? We show that a protein of molecular weight 5055,000, which is present in major amounts in immunoprecipitates from SV40-transformed mouse, rat and some hamster cell lines, is not analogous to Py middle T and is probably not coded for by SV40 DNA. We also show that the various other forms of T-Ag detected are related to large T or small t in that they
Cell 336
appear to be shortened proteins.
or elongated
T and small t (Smith et al., 1978). We have now examined the T-Ags present in cells transformed by SV40. Figure 1 shows that in addition to large T and small t, the SV40-transformed cell lines examined contain other polypeptides which are immunoprecipitated by anti-T serum and not by control serum. Of particular interest is a protein(s) which migrates close to the 53,000 dalton marker (‘%glutamate dehydrogenase) and is immunoprecipitated from SV40-transformed rat and mouse cells and from some SV40-transformed hamster cells. The protein, which in these extracts is labeled with 35S-methionine to approximately the same extent as large T, is not detected using anti-T serum in immunoprecipitates of the untransformed parents of the transformed cells. Its mobility varies slightly when the protein is extracted from transformed cells of different species, but it is unaltered (compare tracks E and F) when the cells are transformed by mutants of SV40 with deletions in the region between 0.60 and 0.55 mu (Sleigh et al., 1978). Furthermore, the mobility of the protein seems to be independent of the size or number of other TAgs present in immunoprecipitates of cells of a given species. Here we present evidence suggesting that
forms of the two
Results Extraction of SV40 Tumor Antigens from SV40Transformed Cells Depending upon the conditions used, various proteins can be detected in cells productively infected with SV40 by immunoprecipitation with serum from animals bearing SV40-induced tumors (anti-T serum). Some of these proteins (such as SV40 VPl, actin, myosin and fibronectin; Smith et al., 1978) result from nonspecific precipitation because they are also present in immunoprecipitates using control serum. In addition to large T and small t, other proteins which react only with anti-T serum are sometimes detected (Carroll and Smith, 1976). At least some of these proteins (previously referred to as the 84,000 and 89,000 dalton forms of T-AS) are proteolytic degradation products of large T (94,000 daltons). When extraction and immunoprecipitation conditions are optimized the only proteins specifically immunoprecipitated in major amounts from productively infected CVI cells are large
A M
CT
B Ci
C CT
D
F T
‘c
F . T
Super-T -
3094 81605340-
-
Large-T -
-
NVT-
-
VP1
-Truncated 1arQe.T
t
Figure
1. lmmunoprecipitation
of T-Ags
from Different
SV40-Transformed
Small- t
Cell Lines
Cells were labeled with ?S.-methionine and proteins were extracted as described (Smith et al., 1976). The extracts were preabsorbed with sheep serum and immunoprecipitated using hamster control (C) or anti-T (T) serum and analyzed on 15% polyacrylamide gels. The following cells were used: (A) SVA31; (6) SWSV/3T3; (C) H65; (D) productively infected CVI cells: (E) REF 2006; (F) FIEF wt6A. Tracks (AD) were run on the same polyacrylamide gel. Autoradiography was for 3 days. The numbers refer to the molecular weight (X lo-? of the “Clabeled marker proteins run in an adjacent track CM).
T-Ags 337
from SV40-Transformed
Cells
the 53,000 dalton T-Ag is not coded for by SV40, and thus we refer to this protein as nonviral T-Ag (NVTAg). Figure 1 shows that in addition to NVT-Ag, some transformed cells contain other proteins which react specifically with anti-T serum and which are not detected in untransformed cells. The number and amount of these additional forms of T-Ag varies, even between different clones of cells that were transformed at the same time by wild-type virus. These additional forms of T-Ag fall into three subgroups: multiple species of large T with very similar although distinct mobilities (doublet and triplet forms of large T) (such as SWSV/3T3 cells); species of T-Ag with a mobility considerably slower than that of large T (such as REF 2006); and proteins other than NVT-Ag with a mobility intermediate between those of large T and small t (such as SWSV/3T3). Some examples of the different forms of T-Ag from transformed cells are shown in Figure 1; these and additional data are summarized in Table 1. Fingerprint Analysis of T-Ags from SV40Transformed Cells To establish the relationship between the various TAgs from transformed cells, we fingerprinted %-methionine-labeled proteins from different SV40-transformed cell lines, using methods previously developed to examine T-Ags from productively infected cells (Mellor and Smith, 1978; Smith et al., 1978). Rat Cells Several lines of rat embryo fibroblast cells which were transformed using both wild-type SV40 and mutants with deletions in the region from 0.60-0.55 mu (Sleigh et al., 1978) were screened for the presence of TAgs. All the lines contain large T of normal mobility and NVT-Ag, but only those transformed by wild-type virus contain detectable small t. In addition to these bands, several lines contain a large T doublet. Line Table
1. Major
Cell Line
Tumor
Antigens
lmmunoprecipitated
Truncated Small t
from SV40-Transformed
REF 2006, which was transformed by mutant virus, has a form of T-Ag which migrates particularly slowly and has an apparent molecular weight of about 130,000 (Figure 1). This line was studied further by fingerprint analysis and compared with the wild-typetransformed REF wt6A line. Figure 2 shows the fingerprints of normal sized large T and NVT-Ag isolated from REF wt6A cells and electrophoresed at either pH 2.1 (Figure 2A and 28) or PH 6.5 (Figures 2D and 2E) in the first dimension. The fingerprints of large T are indistinguishable from those of large T isolated from productively infected cells and analyzed under the same conditions (see Figure 5 of Smith et al., 1978). The fingerprints of NVT-Ag are quite different from those of large T with very few peptides of similar mobility. Experiments in which digests of large T and NVT-Ag were mixed prior to fractionation (Figures 2C and 2F) showed that only two or three of the major peptides from the two proteins co-migrate after separation at pH 2.1 in the first dimension, and none of the major peptides co-migrate after separation at pH 6.5. To examine further the possible relationship between some of the minor peptides whose mobilities are similar after two-dimensional fractionation, the peptides in each spot of the pH 2.1 fingerprints of large T, NVT-Ag and a mixture of the two proteins (Figures 2A, 26 and 2C, respectively) were eluted from the cellulose plates and reelectrophoresed at either pH 6.5 or 3.5. Figure 21 shows the reanalysis of peptides X (from large T), Y (from NVT-Ag) and Z (from a mixture of the two). After electrophoresis at pH 3.5, X separated into three peptides (a, b and c), Y separated into two peptides (d and e) and Z separated into four peptides. These data show that most of the labeled material from large T spot X separates (that is, peptides a and c) from that present in NVTAg. However, peptide b from large T and peptide e from NVT-Ag still co-migrate (albeit in the neutral Cell Lines
Small t
Truncated Larae T
NVT-Aa
-
-
-
Larae
T
Large T Doublets or Triolets
Sunnr
Rat Rat 1 REF wt6A REF 2006
+
+ -
+ +
+ +
+/-
+
Mouse A31 BALB/c3T3 SVA31 E7 Swiss/3T3 SWSV/3T3
-
+ -
-
+ -
+ -
-
+
+
+
+
+
-
-
-
-
-
-
+ +/-
+ + +
+
Hamster Cl3 SV26 H65 FLSV
-
+ + +
-
T
Cell 338
Figure
2. Tryptic
Peptide
Fingerprints
of T-Ags
from SV40-Transformed
REF Cells
(A-C, G and H) used electrophoresis at pH 2.1 in the first dimension and (D-F) at pH 6.5. (A and D) Large T (wt6A); (6 and E) NVT-Ag (wt6A): (C and F) large T + NVT-Ag (wt6A); (G) super T (2006); (H) NVT-Ag (2006). (I) shows reanalysis of the peptides from fingerprints of large T (X), NVTAg (Y) and a mixture of the two (2) at pH 3.5. (a, b and c) indicate the positions of the peptides from large T SPOt X: (d and e) indicate those from NVT-Ag spot Y.
region of the electrophoretogram) after three fractionation procedures. Only one other spot from the fingerprints of large T and NVT-Ag from REF cells contained material which failed to separate after electrophoresis in three dimensions. The finding that these peptides co-migrate does not necessarily mean that they are identical. Nevertheless, although the data presented here do establish that the great majority of methionine tryptic peptides from NVT-Ag are not present in large T, we cannot
exclude the possibility that the proteins share some very limited amino acid sequence homology. The fingerprint of the very large T-Ag from REF 2006 cells (Figure 2G) is strikingly similar to that of normal sized large T isolated from REF wt6A cells. Thus, although the protein is considerably larger than large T, no additional methionine-containing tryptic peptides are apparent. The relative intensity of some of the spots is altered, however, suggesting that some peptides may be present in more than one copy.
T-Ags 339
from SV40-Transformed
Cells
Whatever the explanation of this result, it is clear that the 130,000 dalton protein is related to large T. Kress et al. (1979) have also characterized similar T-Ag species from transformed cells; they refer to such forms as super T. The fingerprint at pH 2.1 of NVT-Ag from FIEF 2006 is indistinguishable from that of FIEF wt6A cells (Figure 2H). Analysis at pH 6.5 (data not shown) also suggested that the proteins are identical. We conclude that NVT-Ag isolated from different lines of SV40transformed FIEF cells is the same, and is affected neither by deletions in SV40 DNA in the 0.60-0.55 mu region nor by whatever factors are responsible for the appearance of different forms of large T (such as the 130,000 dalton super T). Mouse Cell Lines The T-Ags from two lines of SV40-transformed mouse cells were analyzed by fingerprinting methods. The SV40-transformed Swiss/BT3 (SWSV/3T3) line was chosen because it appeared to have a large number of different forms of T-AS. The cells contain three TAgs with approximately the same mobility as normal large T (a large T triplet), a 53,000 dalton NVT-Ag (which has a slightly faster mobility than the equivalent protein from rat cells) and several proteins with apparent molecular weights in the range 25,00040,000. For comparison we used an SV40-transformed BALB/c3T3 line (SVA31 E7). Figure 3 shows the pH 2.1 fingerprints of the normal sized large T from SWSV/3T3 cells (A) and of a larger form with an apparent molecular weight of about 110,000 (D). The large T fingerprint is indistinguishable from that of large T from productively infected cells or SV40-transformed rat cells. The larger 110,000 dalton form of T-Ag has a fingerprint similar to that of the normal sized large T. It contains all the peptides found in the normal protein and also may contain two or three other peptides not normally present. We do not know the origin of these additional peptides; they could originate from extra peptide sequences present in the 110,000 dalton protein, or they may arise from contaminating background proteins. Irrespective of the origin of the additional peptides, the data establish that the 110,000 dalton protein is another example of super T. The second form of T-Ag with a size similar to but distinguishable from that of normal large T was also fingerprinted (data not shown). This T-AS, which has an apparent molecular weight of about 100,000, has a fingerprint virtually identical to that of normal large T with no detectable additional methionine tryptic peptides or any prominent peptides missing. Figure 3E shows the pH 2.1 fingerprint of a 33,000 dalton form of T-Ag isolated from the SWSV/3T3 cells. The fingerprint contains many but not all of the peptides found in large T. In particular, it contains the N terminal peptide and the peptides shared between small t and large T. It does not contain the peptides
unique to small t or some of the peptides unique to large T. These data establish the 33,000 dalton protein to be a shortened form of large T; we refer to such proteins as truncated large T. Figure 36 shows the pH 2.1 fingerprint of NVT-Ag from SWSV/3T3 cells. This fingerprint is quite different from that of large T, and mixing experiments (Figure 3C) confirmed that the two proteins contain only a few peptides which co-migrate after two-dimensional fractionation. Among the peptides which do comigrate is one which migrates with peptide X from large T. Fingerprint analysis at pH 6.5 in the first dimension also showed that SWSV/3T3 NVT-Ag is different from SV40 large T, but again one peptide from both proteins did co-migrate (data not shown). The fingerprint of NVT-Ag from mouse cells is different from that of rat REF cells, although the fingerprints do show some similarity. Mixing experiments showed that a few peptides from the two proteins do co-migrate after two dimensions (Figure 3F), including one with a mobility similar to that of peptide Y from NVT-Ag from REF cells. Analysis at pH 6.5 in the first dimension also suggests that NVT-Ags from rat and mouse cells have a limited number of peptides in common (data not shown). We conclude from these analyses that mouse NVT-Ag has few (if any) peptides in common with large T, but that it may share a limited number of peptides with the equivalent protein from transformed rat cells. The SVA31E7 line of SV40-transformed BALB/ c3T3 mouse cells contains large T, NVT-Ag and small t (Figure 1). Figures 3G and 3H show the pH 2.1 fingerprints of large T and NVT-Ag from these cells. They show that the large T is indistinguishable from large T from productively infected cells, and that NVTAg has only a few peptides which co-migrate with peptides from large T. These include one which migrates with large T peptide X. The fingerprint of SVA31 E7 NVT-Ag is similar but not identical to that of NVT-Ag from SWSV/3T3 cells. We do not know the significance of this apparent difference in NVT-Ag isolated from SV40-transformed cells of different strains of mouse. Recent analysis of NVT-Ag from three other SV40-transformed BALB/c3T3 lines indicates that these cells yielded an NVT-Ag fingerprint similar to that of SWSV/3T3 NVT-Ag (F. Chaudry and A. E. Smith, unpublished results). Hamster Cells SV40-transformed hamster cells show some variation in the appearance of NVT-Ag, although they all contain normal large T. For example, FLSV cells and SV40transformed CHL cells yield little detectable NVT-Ag, H65 cells yield relatively small amounts of a protein in this molecular weight range and SV28 cells yield relatively large amounts. Edwards et al. (1979) have also reported low yields of NVT-Ag from some SV40transformed hamster cells. Figures 4A and 4D show the pH 2.1 fingerprints of
Cell 340
Figure
3. “SMethionine-Labeled
Tryptic
Peptide
Fingerprints
of T-Ags
from SV40-Transformed
Mouse
(SWSV/3T3); (F) NVT-Ag
Cells
All separations used pH 2.1 in the first dimension. (A) Large T (SWSV/3T3): 110,000 dalton large T (SWSV/3T3); (E) 33,000 dalton truncated large fSVA31); (H) NVT-Ag (SVA31); (1) large T + NVT-Ag (SVA31).
(B)‘NVT-Ag T (SWSV/3T3);
(C) large T + NVT-Ag (SWSV/3T3); (D) (REF wt6A + SWSV/3T3); (G) large T
large T isolated from SV28 and H65 cell lines, respectively. Both fingerprints are indistinguishable from one another and from normal large T. Figures 48 and 4E show the fingerprints of the NVT-Ag from these lines. The fingerprints are totally different from one another and from that of large T and do not show any obvious relationship to NVT-Ag from rat and mouse cells. Fingerprints prepared at pH 6.5 in the first dimension confirmed that large T and hamster NVT-Ag share few if any peptides. Mixing experiments at pH 2.1 (Figure 40 showed that although some peptides from large
T and SV28 NVT-Ag co-migrate under these conditions, the hamster NVT-Ag does not contain material which co-migrates with peptide X. Cell-Free Synthesis of NVT-Ag We have previously studied the cell-free synthesis of SV40 T-Ags using mRNA isolated from productively infected CVI cells and shown that the mRNAs coding for SV40 large T and small t hybridize to SV40 DNA immobilized on cellulose. In such experiments we did not detect a protein of mobility similar to that of NVT-
T-Ags 341
from SV40-Transformed
Figure
4. %-Methionine
Cells
Tryptic
Peptide
Fingerprints
of T-Ags
Electrophoresis in the first dimension was at pH 2.1. (A) Large T NVT-Ag (H65); (F) line diagram of methionine-containing tryptic the peptides. Dotted peptides are shared by large T and small unique peptide which maps at 0.585 mu; (6) is a peptide which in the legend to Figure 2.
from SV40-Transformed
Hamster
Cells
(SV28); (6) NVT-Ag (SV28); (C) large T + NVT-Ag (SV28); (D) large T (H65); (E) peptides present in SV40 T-Ags showing the origin and/or sequence of some of t; cross-hatched peptides are unique to small t. (A) is the cysteine-rich. small t maps toward the carboxy terminal half of large T (see text). Peptide X is defined
Ag made in response to mRNA which binds to viral DNA. This result was not unexpected, however, since we do not detect NVT-Ag in extracts of productively infected cells (Paucha and Smith; 1978; Paucha, Harvey and Smith, 1978a). In contrast, the fraction of mRNA from cells productively infected with Py, which binds to Py DNA, directs the synthesis of the three major Py early proteins (Hunter, Hutchinson and Eckhart, 1978; A. Mellor and A. E. Smith, unpublished results). To determine whether the mRNA coding for the 5055,000 dalton NVT-Ag found in some S/40-transformed cells is coded for by SV40 DNA we have used the same experimental procedures to select mRNA from SV40-transformed cells. The proteins synthesized in vitro in response to the purified mRNAs were examined both before and after immunoprecipitation with serum from hamsters bearing SV40-induced tumors. Figure 5 shows an example of the proteins detected in such experiments. Although large T and small t were both present in the products synthesized in response to mRNA which bound to SV40 DNA, a protein with mobility similar to that of the 53,000
dalton marker was not detected in major yield. This result was obtained with several batches of mRNA from each of the transformed cell lines tested, including batches made by different methods. NVT-Ag was not present in the products made by the SV40 DNAbound mRNA fraction even when less stringent hybridization conditions were used to select the viralcoded mRNA8. Similarly, the appearance of NVT-Ag was not sensitive to changes in the conditions used for cell-free synthesis or to the addition of membrane preparations (Elder et al., 1979). In contrast to the results on NVT-Ag synthesis using purified SV40 mRNA, proteins similar in size to the super T of REF 2006 and the shortened forms of large T of SWSV/3T3 were synthesized in response to the fraction of mRNA from the respective transformed cell lines which bound to SV40 DNA (Figure 5). Furthermore, proteins with a mobility greater than that of small t were synthesized in response to the bound fraction when mRNA from cells transformed by mutants of SV40 with deletions in the region at 0.60-55 mu was translated in vitro (Figure 5). These proteins are probably truncated forms of small t. Taken to-
Cell 342
a A
B
C
D
E
F
G
b
H
ABCDEFGM
C-i-
M
-
130 94 81
Super-T Large-T--’
- 130 -94 -81 -60 -53
VP1 VP2 ~Trur?cated law-T
-40
L
VP3 r
-
Small-t
‘\,
-Truncated small-t Figure 5. Translation of mRNA from SV40-Transformed with Hamster Anti-T Serum
Cells after Selection
on SV40 DNA Cellulose,
14 before(b)
and after (a) lmmunoprecipitation
(a) Purified mRNA was translated in nucleased L cell extracts and 10 ~1 aliquots were immunoprecipitated with either control (C) or hamster T(T) serum. The mRNA was from the following transformed cells: (A) REF wt6A; (B) REF 884; (C) REF 890; (D) REF 2001; (E) REF 2003; (F) 2006; (G) SVA31; (H) SWSV/3T3. (b) Purified mRNA was translated in nucleased L cell extracts and 10 pl of the total reaction mixture were applied to a 15% polyacrylamide Track (A) No added RNA: mRNA purified from (6) productively infected CVI cells: (C and D) SWSV/3T3 cells; (E and F) BALB/cSV3T3 cells: SVA31 cells.
gether, these results show that the mRNAs coding for SV40 large T and small t and for the abnormal forms of both of these proteins (found in certain transformed cells) can be selected by hybridization to SV40 DNA. The finding that NVT-Ag was not detected in major amounts either before or after immunoprecipitation of the products made in response to mRNA purified by hybridization to SV40 DNA suggests that the mRNA coding for NVT-Ag does not contain extensive sequences complementary to the viral DNA. If the mRNA coding for NVT-Ag is not present in the fraction of mRNA bound to SV40 DNA then it should be present in the unbound fraction. Detection of the mRNA for NVT-Ag in this fraction, however, requires an immunoprecipitation step to separate NVT-Ag from all the other proteins coded for by mRNAs from the host cell. The RNA in the unbound fraction was therefore translated and the synthesized products immunoprecipitated with hamster anti-T serum. In such experiments, however, we could not detect NVT-Ag in the synthesized products either in response to total mRNA from transformed cells or in response to the non-SV40 DNA-bound fraction. There are several explanations for this result, but the most probable is that nascent NVT-Ag synthesized in vitro is not immuno-
antiREF gel. (G)
precipitable using our anti-T serum. Other experiments suggest that this result can be explained if SV40 large T and NVT-Ag in extracts from transformed cells are associated in some way such that the hamster anti-T serum used immunoprecipitates NVTAg because it is present in a complex with large T (Lane and Crawford, 1979; Linzer and Levine, 1979). Other investigators have found that some anti-T serum appears to react directly with NVT-Ag (Lane and Crawford, 1979; Linzer and Levine, 1979; Kress et al., 1979). Figure 6 shows the result of an experiment in which mRNA from SVA31 mouse cells was fractionated by hybridization to SV40 DNA cellulose, the bound and unbound fractions were translated in vitro and the products were immunoprecipitated using a mouse anti-T serum reported to react directly with NVT-Ag (Lane and Crawford, 1979). Material which co-migrates with NVT-Ag is present in the products made in response to total transformed cell mRNA and to the mRNA in the unbound fraction, whereas SV40 large T and small t are made by mRNAs in both the total mRNA and SV40 DNA-bound fraction. Although we have not proved by peptide fingerprinting that the 53,000 dalton protein detected is identical to NVT-Ag isolated from cells, these data suggest that the mRNA
T-Ags 343
from SV40-Transformed
Cells
AB ---mm
C
DEF
CTCTCTCTCTTM
La1‘ge e60 -53
NV
-40
8 nail
Figure
-
6. Cell-Free
Synthesis of
NVT-Ag
mRNA was isolated from SVA31 cells and translated in vitro both before (tracks A and 6) and after hybridization to SV40 DNA cellulose (tracks C-E). The translation products of mRNA which bound to SV40 DNA are shown in track C. and that which did not bind in tracks D and E. The cell-free products were immunoprecipitated with either hamster control (C) or anti-T Cr) serum (track A) or mouse control (0 or anti-T (T) serum (tracks B-E). Track F is an immunoprecipitate using hamster anti-T serum of an extract of 35S-methionine-labeled WA31 cells. The immunoprecipitates were all analyzed on the same 15% polyacrylamide gel. Tracks A, B and F were exposed for 2 days, the remainder for 4 days.
for NVT-Ag does not bind to SV40 DNA under conditions where the mRNAs for SV40-coded T-Ags do bind. Discussion NVT-Ag Is Not Equivalent to Polyoma Middle T and Is Probably Not Coded for by SV4D DNA From our previous work on the amino terminal sequence and tryptic peptide fingerprint analysis of SV40 T-Ags we have been able to identify several methionine-containing peptides and to map the peptides on SV40 DNA. Of particular interest are the amino terminal tryptic peptide N-Acetyl-Met-Asp-Lys (Mellor and Smith, 1978; Paucha et al., 1978b); the peptides shared between large T and small t (Smith et al., 1978); two methionine peptides unique to small t, including one cysteine-rich peptide which is acidic at pH 2.1 and maps at position 0.585 mu (Paucha and Smith, 1978); the peptides unique to large T, including one which we have mapped to the carboxy terminal half of large T by analysis of different size forms of large T synthesized in vitro in response to SV40 complementary RNA (Paucha et al., 1978a); and MetLys, which is believed to be present in the shared region of large and small T-Ags and in both unique regions (our unpublished results). The position of these peptides in two-dimensional fingerprints (at pH 2.1 in the first dimension) is indicated in Figure 4F.
Using these peptides as markers we have examined the fingerprints of different forms of SV40 T-Ag from transformed cells. We were particularly interested in a protein, referred to here as NVT-Ag, which is specifically immunoprecipitated from SV40-transformed rat, mouse and some hamster cells and which has a molecular weight of about 53,000, similar to that of Py middle T. Examination of the fingerprints of NVTAg isolated from rat, mouse or hamster cells shows that it does not contain the amino terminal peptide of the SV40 T-Ags or the small t unique methionine peptides. These results, together with the findings that the size of NVT-Ag is unaltered in cells transformed by SV40 mutants with deletions in the region between 0.60 and 0.55 mu and that the mRNA for NVT-Ag does not hybridize to SV40 DNA, show that NVT-Ag from SV40-transformed cells is not related to SV40 large T or small t in the same way as the three Py early proteins are related to one another. Further analysis of the fingerprint data on NVT-Ag showed that it contains few if any of the methionine peptides shared between large and small T-Ags or those unique to large T. In some fingerprints major methionine peptides from NVT and large T co-migrated, although when these were reanalyzed using a different pH for separation the major peptides could all be separated. Nevertheless, some minor peptides from NVT-Ag which co-migrate with large T peptides were detected, and one of these (peptide Ye) may be present in NVT-Ag from both rat and mouse. While this result is intriguing, we do not at present believe it to be significant. We have also compared fingerprints of “P-labeled large T and NVT-Ag from mouse cells, and these experiments showed the phosphopeptides of the two proteins to be unrelated (our unpublished results). Chang et al. (1979) have also concluded that SV40 large T and NVT-Ag have very few sequences in common, and Kress et al. (1979) have concluded that the two proteins have no methionine peptides in common. Taken together all these data suggest that the bulk, probably all, of NVT-Ag is not coded for by RNA sequences translated in the same reading frame as SV40 large and small T-Ags. Thus if NVT-Ag is virus-coded it must be synthesized on an mRNA read entirely or at least predominantly in a reading frame different from that which codes for large T and small t. The sequence of the E strand of SV40 DNA indicates that there are few extensive open reading frames present apart from those used for the synthesis of large T and small t. To construct from these sequences an mRNA capable of directing the synthesis of a 50-55,000 dalton protein would require many splicing events. Furthermore, the finding that NVT-Ag has a different fingerprint when isolated from transformed cells from different species indicates that such splicing, were it to occur, would be species-specific. Although neither of these possi-
Cell 344
bilities has been ruled out, we believe that the simplest interpretation of our results is that the protein is not virus-coded. This conclusion is also supported by experiments on the cell-free synthesis of T-Ags from transformed cells. Although we found that NVT-Ag synthesized in vitro was not immunoprecipitated with our hamster anti-T serum, this complication did not prevent us from asking whether the mRNA for NVT-Ag would bind to SV40 DNA. Our earlier experiments had shown that mRNA purified by preparative scale hybridization to SV40 DNA is sufficiently pure to identify products made by the virus-specified mRNA without immunoprecipitation of the cell-free products. Indeed, the experiments shown in Figure 5 show that the mRNAs coding for several abnormal forms of T-Ag present in some transformed cells were readily purified using SV40 DNA cellulose. The abnormal forms of T-AS, including some that were not detected in extracts of labeled cells (such as the shortened forms of small t), were readily detected in the products synthesized by the purified mRNAs without immunoprecipitation. In contrast, all our efforts to detect the synthesis of NVTAg in the fraction of mRNA from transformed cells which binds to SV40 DNA were unsuccessful. In parallel experiments, mRNA isolated from cells productively infected with Py and purified by Py DNA cellulose directed the synthesis of all three Py early proteins-large T, middle T and small t (A. Mellor and A. E. Smith, unpublished results; see also Hunter et al., 1978). Experiments using mouse anti-T serum, which is reported to immunoprecipitate NVT-Ag directly (Lane and Crawford, 1979), indicated that mRNA coding for a protein with mobility similar to that of NVT-Ag could be detected in both total mRNA from SVA31 E7 mouse cells and the fraction of mRNA which does not bind to SV40 DNA cellulose. We have not isolated enough of the protein made in vitro to prove that it is identical to NVT-Ag isolated from cells, but at this stage it seems a reasonable assumption. If this is the case, then we can conclude that the mRNA for NVT-Ag does not hybridize to SV40 DNA under any of the conditions we have tested. Although these results suggest that NVT-Ag is not coded for by SV40 DNA, we cannot exclude the possibility that the mRNA which codes for NVT-Ag is virus-coded but spliced to such an extent that it can no longer form a stable hybrid with immobilized SV40 DNA, even under conditions in which we know that a 200 bp hybrid is stable (our unpublished results). At present, however, we believe the most probable explanation of our results is that NVT-Ag is not coded for by SV40 DNA. The finding of Linzer and Levine (1979) that NVT-Ag is present in uninfected embryonal carcinoma cells strongly suggests that the protein is host cell-coded.
Other Forms of T-Ag from Transformed Cells The abnormal virus-coded T-Ags detected in the experiments reported here and in a large number of other SV40-transformed cells that we have examined (F. Chaudry, E. Paucha, P. J. G. Rigby and A. E. Smith, unpublished results) fall into three subgroups: super T, large T doublets and triplets and truncated large T and small t. We do not know the origin of these different species, but they are all synthesized in vitro when mRNA from the appropriate cell line is translated (Figure 5). This suggests that the proteins arise from different mRNAs rather than by post-translational modification. Fingerprint analysis of super T shows that in one case (REF 20061, although the protein is considerably larger than large T, it contains no detectable additional methionine tryptic peptides. However, some peptides may be present in more than one copy, suggesting that part of the large T molecule may be duplicated. It could be that the mRNA coding for such a protein is generated by some abnormal splicing of SV40 high molecular weight nuclear RNA transcribed from genomes integrated in tandem. Truncated forms of large T could be generated as a result of a mutation in the integrated DNA, or alternatively they could arise from an integrated sequence which lacks a complete copy of the early region. Our earlier results together with the results of the experiments described here clearly show that SV40 does not code for major amounts of a protein analogous to Py middle T in either productively infected or transformed cells. Comparison of the DNA sequences of SV40 and Py indicates that the middle T coding region may be unique to Py (Soeda et al., 1979). We emphasize, however, that a related SV40-coded protein could be generated assuming the mRNA coding for it were spliced in a different manner. Our experiments have so far examined only the major proteins and mRNAs from SV40-infected and -transformed cells. The existence of additional minor SV40 early proteins remains possible. We are continuing our attempts to detect such minor T-Ag species. Experimental
Procedures
Cells The origin of most of the cell lines used has been reported in previous publications (Carroll and Smith, 1976; Crawford et al., 1978; Smith et al., 1978). The SV40-transformed REF cell lines were obtained from W. C. Topp (Sleigh et al., 1978). Antiserum The antiserum used in most of the experiments described was raised by injection of H65-90b cells into golden hamsters (Smith et al., 1978). The mouse serum with anti-NVT-Ag activity was a gift from D. Lane and described by Lane and Crawford (1979). Extraction and lmmunopracipitation of T-Ags Cells were grown in 30 or 90 mm plastic dishes in Dulbecco’s modified E4 containing 510% fetal calf serum. The cells were labeled with ‘?S-methionine (Radiochemical Centre. Amersham) as
T-Ags 345
from SV40-Transformed
Cells
they approached confluence. Conditions for labeling and extraction of the cells have been described (Smith et al., 1978). Conditions for immunoprecipitation were as described (Smith et al., lQ78), except that in some experiments we omitted the preabsorption step using sheep serum. Immunoprecipitates were collected using protein A-bearing S. aureus bacteria and separated on 10 or 15% discontinuous pH polyacrylamide gels (Smith et al., 1978). Peptide Fingerprinting Conditions for preparative scale immunoprecipitation, autoradiography. elution. performic oxidation and trypsin digestion have been described (Smith et al., 1978). The fingerprints were prepared using Kodak plastic-backed cellulose thin-layer sheets, and electrophoresis in the first dimension was at pH 2.1 or 6.5 as described. Autoradiography of the dried sheets was for between 2 and 25 days. To elute peptides from the thin-layer sheets for further analysis, the appropriate region of the cellulose was scraped from the plastic and eluted using 2 times 0.5 ml of 1 M formic acid. The eluted peptides were washed with water and reanalyzed on further cellulose sheets or on paper. Purification of mRNA Using SV40 DNA Cellulose Preparation of SV40 DNA and mRNA from cells grown in culture has been described (Paucha et al., 1978a). The DNA was linearized with restriction enzymes and covalently attached to cellulose using the method of Noyes and Stark (1975) as described by Paucha and Smith (1978). About 5 mg of total RNA isolated from different transformed cell lines were hybridized to cellulose containing about 200 pg of bound SV40 DNA. After washing. the hybridized mRNA was eluted. reprecipitated and resuspended in about 25 pl of water. Between 1 and 3 pl were subsequently translated in 25 pl cell-free reactions (Paucha et al., 1978a). Cell-Free Synthesis Conditions for cell-free synthesis using extracts from L cells which were sometimes pretreated with micrococcal nuclease to reduce endOQenOUS mRNA activity and immunoprecipitation of the cell-free products have been described (Paucha and Smith, 1978; Paucha et al., 1 Q78a). The immunoprecipitates were fractionated on 15% polyacrylamide gels and autoradiography was for l-l 0 days. Acknowledgments We thank Kit Osborn for help with growing cells, Dr. W. C. Topp for SV40-transformed REF cells, Paul Nicklin for help in raising antisera and growing protein A-bearing S. aureus and Dr. David Lane for mouse antiserum with anti-NVT-Ag activity. We are grateful to Drs. T. Chang and M. Kress for helpful discussions and to Drs. Kress and Levine for Sending us manuscripts prior to publication. We thank Miss Anne Roberts for help with the preparation of the manuscript and Drs. F. McCormick and M. Hayman for critical readinQS. E. P. was supported by a fellowship from the Canadian MRC. Received
May 10, 1979;
revised
July 23, 1979
References Carroll, R. B and Smith, A. E. (1976). T antigen from simian virus 40-infected Nat. Acad. Sci. USA 73, 2254-2258. Chang. C.. Pancake, S. J.. Luborsky. Detergent solubilization and partial surface and transplantation antigens cells. Int. J. Cancer 19, 258-266.
Monomer molecular Weight of and transformed cells. Proc. S. W. and Mora. P. T. (1977). purification of tumor Specific from SV40 transformed mouse
Chang. C., Simmons, 0. T.. Martin, M. A. and Mora. P. T. (1979). Identification and partial characterization of tumor antigens from SV40 transformed mouse cells. J. Virol. 31. 463-471. Crawford, L. V., Cole, C. N.. Smith, A. E.. Paucha. E.. Tegtmeyer. P.. Rundell. K. and Berg, P. (1978). Organization and expression of early
genes
of simian
virus 40. Proc.
Nat. Acad.
Sci. USA 75, 117-l
21,
Edwards, C. A., Khoury, G. and Martin, R. G. (1979). Phosphorylation of T-antigen expression in cells transformed by wild type and tsA mutants of SV40. J. Virol. 29, 753-762. Elder. K. T.. Bye. J. M.. Waterfield. M. D., Skehel, J. J. and Smith. A. E. (1979). In vitro synthesis, glycosylation, and membrane insertion of influenza virus haemagglutinin. Virology 95, 343-350. Feunteun, J., Sompayrac. L., Fluck, M. and Benjamin, T. (1976). Localization of gene functions in polyoma virus DNA. Proc. Nat. Acad. Sci. USA 73, 4169-4173. Fiers, W.. Contreras, R., Haegeman. G.. Rogiers. R.. Van de Voorde, A.. Van Heuverswyn. H.. Van Herreweghe. J., Volckaert, G. and Ysabaert. M. (1978). Complete nucleotide sequence of SV40 DNA. Nature 273, 113-l 20. Gaudray. P.. Rassoulzadegan, M. and Cuzin. F. (1978). Expression of simian virus 40 early genes in transformed rat cells is correlated with maintenance of the transformed phenotype. Proc. Nat. Acad. Sci. USA 75, 4987-4991. Hunter, T., Hutchinson. M. A. and Eckhart. polyoma virus T antigens in vitro. Proc. 5917-5921.
W. (1978). Nat. Acad.
Hutchinson. M. A., Hunter, T. and Eckhart, W. (1978). tion of T antigens in polyoma-infected and transformed 65-77. Ito. T.. Spurr. N. and polyoma virus T-antigen.
Translation Sci. USA
Characterizacells. Cell 15,
Dulbecco, R. (1977a). Characterization Proc. Nat. Acad. Sci. USA 74, 1259-l
Ito. Y.. Brocklehurst. J. R. and Dulbecco. proteins in the plasma membrane of cells formed by polyoma virus. Proc. Nat. Acad.
of 75,
of 263.
R. (1977b). Virus-specific lytically infected or transSci. USA 74, 4666-4670.
Kress. M., De Vaux Saint Cyr. C. and Girard. M. (1977). weight of SV40 T-antigens. INSERM 69, 79-92.
The molecular
Kress, M., May, E., Cassingena, R. and May, P. (1979). SV40 transformed cells express new species of proteins precipitable by SV40 anti-tumor serum. J. Virol. 31, 472-483. Lane, D. P. and Crawford, L. V. (1979). T-antigen is bound protein in SV40 transformed cells. Nature 2 78, 261-283. Lichaa. present
to a host
M. and Niesar, E. (1977). Characterization of SV40 T-antigen in SV40 transformed cells. INSERM 69, 21 l-222.
Linzer. D. I. H. and Levine, A. J. (1979). Characterization of a 54,OOOK dalton cellular SV40 tumor antigen present in SV40-transformed cells and uninfected embryonal carcinoma cells. Cell 17, 4352. Melero, J. A., Stitt. D. T., Mangel. W. F. and Carrel, R. B. (1979). Identification of a new polypeptide species (48-55K) immunoprecipitable by antiserum to purified large-T antigen in SV40 infected and transformed cells. Virology 93, 466-480. Mellor. A. and Smith, A. E. (1978). Characterization of the aminoterminal tryptic peptide of simian virus 40 small-t and large-T antigens. J. Virol. 28, 992-996. Noyes. B. E. and Stark, G. R. (1975). DNA covalently coupled to cellulose.
Nucleic acid hybridization Cell 5, 301-310.
Paucha. E. and Smith, A. E. (1978). The sequences and 0.54 map units on simian virus 40 DNA code region of small t antigen. Cell 15, 101 l-l 020. Paucha, E., Harvey, R. and Smith, A. E. (1978a). of simian virus 40 T-antigens. J. Virol. 28, 154-l
using
between 0.59 for the unique
Cell free synthesis 70.
Paucha, E.. Mellor, A., Harvey, R., Smith, A. E.. Hewick. R. M. and Waterfield. M. D. (1 Q78b). Large and small tumor antigens from SV40 have identical amino termini mapping at 0.65 map units. Proc. Nat. Acad. Sci. USA 75, 2165-2189. Prives. C.. Gluzman, Y. and Winocour, E. (1978). Cellular and cellfree synthesis of simian virus 40 T-antigens in permissive and transformed cells. J. Virol. 25, 587-595. Prives,
C., Gilboa.
E., Revel,
M. and Winocour.
E. (1977).
Cell-free
Cell 346
translation of simian virus 40 early messenger RNA coding antigen. Proc. Nat. Acad. Sci. USA 74, 457-461. Reddy, V. 6.. Thimmappaya, 6.. Dhar. Ft., Subramanian. 8. S.. Pan, J.. Celma, M. L. and Weissman, S. M. (1978). of SV40. Science 200, 494-502. Rundell. K., Collins, Nathans, D. (1977). Virol. 2 7, 636-646.
for viral TK. N.. Zain, The genome
J. K., Tegtmeyer, P., Ozer, H. L., Lai, C. and Identification of simian virus 40 protein A. J.
Schaffhausen, 6. S., Silver, J. E. and Benjamin, T. L. (1978). Tumor antigen(s) in cells productively infected by wild-type polyoma virus and mutant NG-10. Proc. Nat. Acad. Sci. USA 75, 79-83. Schmidt-Ullrich, R., Scott-Thompson, W., Peck, S. L. and Wallack, D. F. H. (1977). Simian virus 40-specific proteins in the membranes of simian virus 40-transformed hamster and mouse cells. Proc. Nat. Acad. Sci. USA 74, 5069-5072. Simmons, D. T. and Martin, M. A. (1978). Common methionine peptides near the amino terminal end of primate papovavirus antigens. Proc. Nat. Acad. Sci. USA 75, 1131-l 135.
tryptic tumor
Sleigh, M. J., Topp. W. C.. Hanich, R. and Sambrook, J. F. (1978). Mutants of SV40 with an altered small t protein are reduced in their ability to transform cells. Cell 14, 79-88. Smart, J. E. and Ito, Y. (1978). Three species of polyoma virus tumor antigens share common peptides probably near the amino termini of the proteins. Cell 15, 1427-l 438. Smith, A. E., Smith, R. and Paucha, E. (1978). Extraction and fingerprint analysis of SV40 large and small T-antigens. J. Virol. 28, 140153. Soeda. E.. Arrand, J. R., Smolar, N. and Griffin, B. E. (1979). Sequence from early region of polyoma virus DNA containing viral replication origin and encoding small, middle and (part of) large T antigens. Cell 17, 357-370. Volckaert. G., Van de Voorde. A. and Fiers, W. (1978). Nucleotide sequence of SV40 small-t gene. Proc. Nat. Acad. Sci. USA 75, 216021 64.
October
1979,
Copyright
0 1979
by MIT
Characterization of Different Tumor Antigens in Cells Transformed by Simian Virus 40
Alan E. Smith, Ros Smith and Eva Paucha’ Translation Laboratory Imperial Cancer Research Fund London WC2A 3PX, England
Summary In addition to large T and small t antigens, cells transformed by simian virus 40 (SV40) commonly contain other proteins which specifically immunoprecipitate with SV40 anti-T serum and which are not detected in untransformed cells. The additional tumor antigens (T-Ags) fall into two groups: those having a close structural relationship with normal SV40 T-Ags, and those unrelated to large T and small t. The latter are probably nonviral T-Ags (NVTAgs). The NVT-Ags comprise a family of proteins of molecular weight 50,000-55,000. Fingerprint analysis shows that NVT-Ags have few if any peptides in common with large T or small t, and that they lack the amino terminal tryptic peptide and the peptides unique to small t. NVT-Ags from different species have different fingerprints, but those isolated from different transformants of the same cell line are identical. The size of NVT is unaltered in cells transformed by mutants of SV40 with deletions in the region 0.50-0.55 map units. The mRNA for NVT does not hybridize to SV40 DNA. The other forms of T-Ag isolated from transformed cells fall into three classes: shortened forms of large T (truncated large T); multiple species of T-Ag with molecular weights very similar to, but distinct from, those of normal large T (large T doublets and triplets); and elongated forms of large T (super T). These proteins all contain the normal amino terminus of SV40 T-Ags, and the truncated forms of large T lack peptides from the carboxy terminal half of large T. One species of super T (molecular weight 130,000) contains only those methionine tryptic peptides present in normal large T, although it may contain some peptides in more than one copy. Introduction Monkey cells productively infected with simian virus 40 (SV40) contain at least two major virus-coded early proteins (Prives et al., 1977; Rundell et al., 1977; Crawford et al., 1978; Paucha et al., 1978b). These proteins, or tumor antigens (T-ASS), are referred to as large T and small t and have molecular weights of 90,000-l 00,000 and 15,000-20,000, respectively. Fingerprint analysis has shown that large T and small t share some peptide sequences, although both proteins also have unique regions (Simmons and Martin, l Present address: Department of Biochemistry, Imperial Science and Technology, London SW7. England.
Cottege
of
Present
1978; Smith, Smith and Paucha, 1978). The shared region has been located at the amino terminal end of the two proteins and is coded for by DNA sequences which extend between 0.65 and 0.60 map units (mu) on SV40 DNA (Paucha et al., 1978b; Volckaert, Van de Voorde and Fiers, 1978). The unique region of small t has been mapped between 0.60 and 0.55 mu and the unique region of large T between 0.54 and 0.17 mu (Paucha and Smith, 1978; Crawford et al., 1978; Fiers et al., 1978; Reddy et al., 1978). Mouse cells productively infected with polyoma virus (Py) contain early proteins which appear to correspond approximately in molecular weight and sequence arrangement to SV40 large T and small t (Feunteun et al., 1976; Ito, Spurr and Dulbecco, 1977a; Schaffhausen, Silver and Benjamin 1978) and also contain at least one other early protein which is present in major amounts. This is referred to as middle T and has a molecular weight in the range 50,000-60,000 (Ito, Brocklehurst and Dulbecco, 1977b). Py middle T shares amino terminal sequences with Py large and small T-Ags, but because of an alternative splicing pathway the sequences at its carboxy terminus are unique (Smart and Ito, 1978; Hutchinson, Hunter and Eckhart 1978; Soeda et al., 1979). Studies on the T-Ags present in cells transformed by wild-type SV40 indicate that they usually contain SV40 large T and small t. In some cell lines other forms of T-Ag have also been detected. These include proteins considerably larger than large T (Lichaa and Niesar, 1977; Kress, De Vaux Saint Cyr and Girard, 1977; Kress et al., 1979) multiple species of proteins with mobility similar to that of large T (Prives, Gluzman and Winocour, 1978) and proteins having an apparent molecular weight intermediate between those of large T and small t, with a prominent component of 5055,000 daltons (Chang et al., 1977; Schmidt-Ullrich et al., 1977; Gaudray, Rassoulzadegan and Cuzin, 1978; Lane and Crawford, 1979; Linzer and Levine, 1979; Edwards, Khoury and Martin, 1979; Chang et al., 1979; Kress et al., 1979; Melero et al., 1979). In the experiments reported here we have analyzed the different forms of T-Ag present in a number of transformed cell lines by tryptic peptide fingerprinting and by cell-free synthesis. We have asked three questions: are the different forms of T-Ag coded for by SV40 DNA? What is their relationship to SV40 large T and small t? Do SV40-transformed cells contain a protein equivalent to the polyoma virus early protein middle T? We show that a protein of molecular weight 5055,000, which is present in major amounts in immunoprecipitates from SV40-transformed mouse, rat and some hamster cell lines, is not analogous to Py middle T and is probably not coded for by SV40 DNA. We also show that the various other forms of T-Ag detected are related to large T or small t in that they
Cell 336
appear to be shortened proteins.
or elongated
T and small t (Smith et al., 1978). We have now examined the T-Ags present in cells transformed by SV40. Figure 1 shows that in addition to large T and small t, the SV40-transformed cell lines examined contain other polypeptides which are immunoprecipitated by anti-T serum and not by control serum. Of particular interest is a protein(s) which migrates close to the 53,000 dalton marker (‘%glutamate dehydrogenase) and is immunoprecipitated from SV40-transformed rat and mouse cells and from some SV40-transformed hamster cells. The protein, which in these extracts is labeled with 35S-methionine to approximately the same extent as large T, is not detected using anti-T serum in immunoprecipitates of the untransformed parents of the transformed cells. Its mobility varies slightly when the protein is extracted from transformed cells of different species, but it is unaltered (compare tracks E and F) when the cells are transformed by mutants of SV40 with deletions in the region between 0.60 and 0.55 mu (Sleigh et al., 1978). Furthermore, the mobility of the protein seems to be independent of the size or number of other TAgs present in immunoprecipitates of cells of a given species. Here we present evidence suggesting that
forms of the two
Results Extraction of SV40 Tumor Antigens from SV40Transformed Cells Depending upon the conditions used, various proteins can be detected in cells productively infected with SV40 by immunoprecipitation with serum from animals bearing SV40-induced tumors (anti-T serum). Some of these proteins (such as SV40 VPl, actin, myosin and fibronectin; Smith et al., 1978) result from nonspecific precipitation because they are also present in immunoprecipitates using control serum. In addition to large T and small t, other proteins which react only with anti-T serum are sometimes detected (Carroll and Smith, 1976). At least some of these proteins (previously referred to as the 84,000 and 89,000 dalton forms of T-AS) are proteolytic degradation products of large T (94,000 daltons). When extraction and immunoprecipitation conditions are optimized the only proteins specifically immunoprecipitated in major amounts from productively infected CVI cells are large
A M
CT
B Ci
C CT
D
F T
‘c
F . T
Super-T -
3094 81605340-
-
Large-T -
-
NVT-
-
VP1
-Truncated 1arQe.T
t
Figure
1. lmmunoprecipitation
of T-Ags
from Different
SV40-Transformed
Small- t
Cell Lines
Cells were labeled with ?S.-methionine and proteins were extracted as described (Smith et al., 1976). The extracts were preabsorbed with sheep serum and immunoprecipitated using hamster control (C) or anti-T (T) serum and analyzed on 15% polyacrylamide gels. The following cells were used: (A) SVA31; (6) SWSV/3T3; (C) H65; (D) productively infected CVI cells: (E) REF 2006; (F) FIEF wt6A. Tracks (AD) were run on the same polyacrylamide gel. Autoradiography was for 3 days. The numbers refer to the molecular weight (X lo-? of the “Clabeled marker proteins run in an adjacent track CM).
T-Ags 337
from SV40-Transformed
Cells
the 53,000 dalton T-Ag is not coded for by SV40, and thus we refer to this protein as nonviral T-Ag (NVTAg). Figure 1 shows that in addition to NVT-Ag, some transformed cells contain other proteins which react specifically with anti-T serum and which are not detected in untransformed cells. The number and amount of these additional forms of T-Ag varies, even between different clones of cells that were transformed at the same time by wild-type virus. These additional forms of T-Ag fall into three subgroups: multiple species of large T with very similar although distinct mobilities (doublet and triplet forms of large T) (such as SWSV/3T3 cells); species of T-Ag with a mobility considerably slower than that of large T (such as REF 2006); and proteins other than NVT-Ag with a mobility intermediate between those of large T and small t (such as SWSV/3T3). Some examples of the different forms of T-Ag from transformed cells are shown in Figure 1; these and additional data are summarized in Table 1. Fingerprint Analysis of T-Ags from SV40Transformed Cells To establish the relationship between the various TAgs from transformed cells, we fingerprinted %-methionine-labeled proteins from different SV40-transformed cell lines, using methods previously developed to examine T-Ags from productively infected cells (Mellor and Smith, 1978; Smith et al., 1978). Rat Cells Several lines of rat embryo fibroblast cells which were transformed using both wild-type SV40 and mutants with deletions in the region from 0.60-0.55 mu (Sleigh et al., 1978) were screened for the presence of TAgs. All the lines contain large T of normal mobility and NVT-Ag, but only those transformed by wild-type virus contain detectable small t. In addition to these bands, several lines contain a large T doublet. Line Table
1. Major
Cell Line
Tumor
Antigens
lmmunoprecipitated
Truncated Small t
from SV40-Transformed
REF 2006, which was transformed by mutant virus, has a form of T-Ag which migrates particularly slowly and has an apparent molecular weight of about 130,000 (Figure 1). This line was studied further by fingerprint analysis and compared with the wild-typetransformed REF wt6A line. Figure 2 shows the fingerprints of normal sized large T and NVT-Ag isolated from REF wt6A cells and electrophoresed at either pH 2.1 (Figure 2A and 28) or PH 6.5 (Figures 2D and 2E) in the first dimension. The fingerprints of large T are indistinguishable from those of large T isolated from productively infected cells and analyzed under the same conditions (see Figure 5 of Smith et al., 1978). The fingerprints of NVT-Ag are quite different from those of large T with very few peptides of similar mobility. Experiments in which digests of large T and NVT-Ag were mixed prior to fractionation (Figures 2C and 2F) showed that only two or three of the major peptides from the two proteins co-migrate after separation at pH 2.1 in the first dimension, and none of the major peptides co-migrate after separation at pH 6.5. To examine further the possible relationship between some of the minor peptides whose mobilities are similar after two-dimensional fractionation, the peptides in each spot of the pH 2.1 fingerprints of large T, NVT-Ag and a mixture of the two proteins (Figures 2A, 26 and 2C, respectively) were eluted from the cellulose plates and reelectrophoresed at either pH 6.5 or 3.5. Figure 21 shows the reanalysis of peptides X (from large T), Y (from NVT-Ag) and Z (from a mixture of the two). After electrophoresis at pH 3.5, X separated into three peptides (a, b and c), Y separated into two peptides (d and e) and Z separated into four peptides. These data show that most of the labeled material from large T spot X separates (that is, peptides a and c) from that present in NVTAg. However, peptide b from large T and peptide e from NVT-Ag still co-migrate (albeit in the neutral Cell Lines
Small t
Truncated Larae T
NVT-Aa
-
-
-
Larae
T
Large T Doublets or Triolets
Sunnr
Rat Rat 1 REF wt6A REF 2006
+
+ -
+ +
+ +
+/-
+
Mouse A31 BALB/c3T3 SVA31 E7 Swiss/3T3 SWSV/3T3
-
+ -
-
+ -
+ -
-
+
+
+
+
+
-
-
-
-
-
-
+ +/-
+ + +
+
Hamster Cl3 SV26 H65 FLSV
-
+ + +
-
T
Cell 338
Figure
2. Tryptic
Peptide
Fingerprints
of T-Ags
from SV40-Transformed
REF Cells
(A-C, G and H) used electrophoresis at pH 2.1 in the first dimension and (D-F) at pH 6.5. (A and D) Large T (wt6A); (6 and E) NVT-Ag (wt6A): (C and F) large T + NVT-Ag (wt6A); (G) super T (2006); (H) NVT-Ag (2006). (I) shows reanalysis of the peptides from fingerprints of large T (X), NVTAg (Y) and a mixture of the two (2) at pH 3.5. (a, b and c) indicate the positions of the peptides from large T SPOt X: (d and e) indicate those from NVT-Ag spot Y.
region of the electrophoretogram) after three fractionation procedures. Only one other spot from the fingerprints of large T and NVT-Ag from REF cells contained material which failed to separate after electrophoresis in three dimensions. The finding that these peptides co-migrate does not necessarily mean that they are identical. Nevertheless, although the data presented here do establish that the great majority of methionine tryptic peptides from NVT-Ag are not present in large T, we cannot
exclude the possibility that the proteins share some very limited amino acid sequence homology. The fingerprint of the very large T-Ag from REF 2006 cells (Figure 2G) is strikingly similar to that of normal sized large T isolated from REF wt6A cells. Thus, although the protein is considerably larger than large T, no additional methionine-containing tryptic peptides are apparent. The relative intensity of some of the spots is altered, however, suggesting that some peptides may be present in more than one copy.
T-Ags 339
from SV40-Transformed
Cells
Whatever the explanation of this result, it is clear that the 130,000 dalton protein is related to large T. Kress et al. (1979) have also characterized similar T-Ag species from transformed cells; they refer to such forms as super T. The fingerprint at pH 2.1 of NVT-Ag from FIEF 2006 is indistinguishable from that of FIEF wt6A cells (Figure 2H). Analysis at pH 6.5 (data not shown) also suggested that the proteins are identical. We conclude that NVT-Ag isolated from different lines of SV40transformed FIEF cells is the same, and is affected neither by deletions in SV40 DNA in the 0.60-0.55 mu region nor by whatever factors are responsible for the appearance of different forms of large T (such as the 130,000 dalton super T). Mouse Cell Lines The T-Ags from two lines of SV40-transformed mouse cells were analyzed by fingerprinting methods. The SV40-transformed Swiss/BT3 (SWSV/3T3) line was chosen because it appeared to have a large number of different forms of T-AS. The cells contain three TAgs with approximately the same mobility as normal large T (a large T triplet), a 53,000 dalton NVT-Ag (which has a slightly faster mobility than the equivalent protein from rat cells) and several proteins with apparent molecular weights in the range 25,00040,000. For comparison we used an SV40-transformed BALB/c3T3 line (SVA31 E7). Figure 3 shows the pH 2.1 fingerprints of the normal sized large T from SWSV/3T3 cells (A) and of a larger form with an apparent molecular weight of about 110,000 (D). The large T fingerprint is indistinguishable from that of large T from productively infected cells or SV40-transformed rat cells. The larger 110,000 dalton form of T-Ag has a fingerprint similar to that of the normal sized large T. It contains all the peptides found in the normal protein and also may contain two or three other peptides not normally present. We do not know the origin of these additional peptides; they could originate from extra peptide sequences present in the 110,000 dalton protein, or they may arise from contaminating background proteins. Irrespective of the origin of the additional peptides, the data establish that the 110,000 dalton protein is another example of super T. The second form of T-Ag with a size similar to but distinguishable from that of normal large T was also fingerprinted (data not shown). This T-AS, which has an apparent molecular weight of about 100,000, has a fingerprint virtually identical to that of normal large T with no detectable additional methionine tryptic peptides or any prominent peptides missing. Figure 3E shows the pH 2.1 fingerprint of a 33,000 dalton form of T-Ag isolated from the SWSV/3T3 cells. The fingerprint contains many but not all of the peptides found in large T. In particular, it contains the N terminal peptide and the peptides shared between small t and large T. It does not contain the peptides
unique to small t or some of the peptides unique to large T. These data establish the 33,000 dalton protein to be a shortened form of large T; we refer to such proteins as truncated large T. Figure 36 shows the pH 2.1 fingerprint of NVT-Ag from SWSV/3T3 cells. This fingerprint is quite different from that of large T, and mixing experiments (Figure 3C) confirmed that the two proteins contain only a few peptides which co-migrate after two-dimensional fractionation. Among the peptides which do comigrate is one which migrates with peptide X from large T. Fingerprint analysis at pH 6.5 in the first dimension also showed that SWSV/3T3 NVT-Ag is different from SV40 large T, but again one peptide from both proteins did co-migrate (data not shown). The fingerprint of NVT-Ag from mouse cells is different from that of rat REF cells, although the fingerprints do show some similarity. Mixing experiments showed that a few peptides from the two proteins do co-migrate after two dimensions (Figure 3F), including one with a mobility similar to that of peptide Y from NVT-Ag from REF cells. Analysis at pH 6.5 in the first dimension also suggests that NVT-Ags from rat and mouse cells have a limited number of peptides in common (data not shown). We conclude from these analyses that mouse NVT-Ag has few (if any) peptides in common with large T, but that it may share a limited number of peptides with the equivalent protein from transformed rat cells. The SVA31E7 line of SV40-transformed BALB/ c3T3 mouse cells contains large T, NVT-Ag and small t (Figure 1). Figures 3G and 3H show the pH 2.1 fingerprints of large T and NVT-Ag from these cells. They show that the large T is indistinguishable from large T from productively infected cells, and that NVTAg has only a few peptides which co-migrate with peptides from large T. These include one which migrates with large T peptide X. The fingerprint of SVA31 E7 NVT-Ag is similar but not identical to that of NVT-Ag from SWSV/3T3 cells. We do not know the significance of this apparent difference in NVT-Ag isolated from SV40-transformed cells of different strains of mouse. Recent analysis of NVT-Ag from three other SV40-transformed BALB/c3T3 lines indicates that these cells yielded an NVT-Ag fingerprint similar to that of SWSV/3T3 NVT-Ag (F. Chaudry and A. E. Smith, unpublished results). Hamster Cells SV40-transformed hamster cells show some variation in the appearance of NVT-Ag, although they all contain normal large T. For example, FLSV cells and SV40transformed CHL cells yield little detectable NVT-Ag, H65 cells yield relatively small amounts of a protein in this molecular weight range and SV28 cells yield relatively large amounts. Edwards et al. (1979) have also reported low yields of NVT-Ag from some SV40transformed hamster cells. Figures 4A and 4D show the pH 2.1 fingerprints of
Cell 340
Figure
3. “SMethionine-Labeled
Tryptic
Peptide
Fingerprints
of T-Ags
from SV40-Transformed
Mouse
(SWSV/3T3); (F) NVT-Ag
Cells
All separations used pH 2.1 in the first dimension. (A) Large T (SWSV/3T3): 110,000 dalton large T (SWSV/3T3); (E) 33,000 dalton truncated large fSVA31); (H) NVT-Ag (SVA31); (1) large T + NVT-Ag (SVA31).
(B)‘NVT-Ag T (SWSV/3T3);
(C) large T + NVT-Ag (SWSV/3T3); (D) (REF wt6A + SWSV/3T3); (G) large T
large T isolated from SV28 and H65 cell lines, respectively. Both fingerprints are indistinguishable from one another and from normal large T. Figures 48 and 4E show the fingerprints of the NVT-Ag from these lines. The fingerprints are totally different from one another and from that of large T and do not show any obvious relationship to NVT-Ag from rat and mouse cells. Fingerprints prepared at pH 6.5 in the first dimension confirmed that large T and hamster NVT-Ag share few if any peptides. Mixing experiments at pH 2.1 (Figure 40 showed that although some peptides from large
T and SV28 NVT-Ag co-migrate under these conditions, the hamster NVT-Ag does not contain material which co-migrates with peptide X. Cell-Free Synthesis of NVT-Ag We have previously studied the cell-free synthesis of SV40 T-Ags using mRNA isolated from productively infected CVI cells and shown that the mRNAs coding for SV40 large T and small t hybridize to SV40 DNA immobilized on cellulose. In such experiments we did not detect a protein of mobility similar to that of NVT-
T-Ags 341
from SV40-Transformed
Figure
4. %-Methionine
Cells
Tryptic
Peptide
Fingerprints
of T-Ags
Electrophoresis in the first dimension was at pH 2.1. (A) Large T NVT-Ag (H65); (F) line diagram of methionine-containing tryptic the peptides. Dotted peptides are shared by large T and small unique peptide which maps at 0.585 mu; (6) is a peptide which in the legend to Figure 2.
from SV40-Transformed
Hamster
Cells
(SV28); (6) NVT-Ag (SV28); (C) large T + NVT-Ag (SV28); (D) large T (H65); (E) peptides present in SV40 T-Ags showing the origin and/or sequence of some of t; cross-hatched peptides are unique to small t. (A) is the cysteine-rich. small t maps toward the carboxy terminal half of large T (see text). Peptide X is defined
Ag made in response to mRNA which binds to viral DNA. This result was not unexpected, however, since we do not detect NVT-Ag in extracts of productively infected cells (Paucha and Smith; 1978; Paucha, Harvey and Smith, 1978a). In contrast, the fraction of mRNA from cells productively infected with Py, which binds to Py DNA, directs the synthesis of the three major Py early proteins (Hunter, Hutchinson and Eckhart, 1978; A. Mellor and A. E. Smith, unpublished results). To determine whether the mRNA coding for the 5055,000 dalton NVT-Ag found in some S/40-transformed cells is coded for by SV40 DNA we have used the same experimental procedures to select mRNA from SV40-transformed cells. The proteins synthesized in vitro in response to the purified mRNAs were examined both before and after immunoprecipitation with serum from hamsters bearing SV40-induced tumors. Figure 5 shows an example of the proteins detected in such experiments. Although large T and small t were both present in the products synthesized in response to mRNA which bound to SV40 DNA, a protein with mobility similar to that of the 53,000
dalton marker was not detected in major yield. This result was obtained with several batches of mRNA from each of the transformed cell lines tested, including batches made by different methods. NVT-Ag was not present in the products made by the SV40 DNAbound mRNA fraction even when less stringent hybridization conditions were used to select the viralcoded mRNA8. Similarly, the appearance of NVT-Ag was not sensitive to changes in the conditions used for cell-free synthesis or to the addition of membrane preparations (Elder et al., 1979). In contrast to the results on NVT-Ag synthesis using purified SV40 mRNA, proteins similar in size to the super T of REF 2006 and the shortened forms of large T of SWSV/3T3 were synthesized in response to the fraction of mRNA from the respective transformed cell lines which bound to SV40 DNA (Figure 5). Furthermore, proteins with a mobility greater than that of small t were synthesized in response to the bound fraction when mRNA from cells transformed by mutants of SV40 with deletions in the region at 0.60-55 mu was translated in vitro (Figure 5). These proteins are probably truncated forms of small t. Taken to-
Cell 342
a A
B
C
D
E
F
G
b
H
ABCDEFGM
C-i-
M
-
130 94 81
Super-T Large-T--’
- 130 -94 -81 -60 -53
VP1 VP2 ~Trur?cated law-T
-40
L
VP3 r
-
Small-t
‘\,
-Truncated small-t Figure 5. Translation of mRNA from SV40-Transformed with Hamster Anti-T Serum
Cells after Selection
on SV40 DNA Cellulose,
14 before(b)
and after (a) lmmunoprecipitation
(a) Purified mRNA was translated in nucleased L cell extracts and 10 ~1 aliquots were immunoprecipitated with either control (C) or hamster T(T) serum. The mRNA was from the following transformed cells: (A) REF wt6A; (B) REF 884; (C) REF 890; (D) REF 2001; (E) REF 2003; (F) 2006; (G) SVA31; (H) SWSV/3T3. (b) Purified mRNA was translated in nucleased L cell extracts and 10 pl of the total reaction mixture were applied to a 15% polyacrylamide Track (A) No added RNA: mRNA purified from (6) productively infected CVI cells: (C and D) SWSV/3T3 cells; (E and F) BALB/cSV3T3 cells: SVA31 cells.
gether, these results show that the mRNAs coding for SV40 large T and small t and for the abnormal forms of both of these proteins (found in certain transformed cells) can be selected by hybridization to SV40 DNA. The finding that NVT-Ag was not detected in major amounts either before or after immunoprecipitation of the products made in response to mRNA purified by hybridization to SV40 DNA suggests that the mRNA coding for NVT-Ag does not contain extensive sequences complementary to the viral DNA. If the mRNA coding for NVT-Ag is not present in the fraction of mRNA bound to SV40 DNA then it should be present in the unbound fraction. Detection of the mRNA for NVT-Ag in this fraction, however, requires an immunoprecipitation step to separate NVT-Ag from all the other proteins coded for by mRNAs from the host cell. The RNA in the unbound fraction was therefore translated and the synthesized products immunoprecipitated with hamster anti-T serum. In such experiments, however, we could not detect NVT-Ag in the synthesized products either in response to total mRNA from transformed cells or in response to the non-SV40 DNA-bound fraction. There are several explanations for this result, but the most probable is that nascent NVT-Ag synthesized in vitro is not immuno-
antiREF gel. (G)
precipitable using our anti-T serum. Other experiments suggest that this result can be explained if SV40 large T and NVT-Ag in extracts from transformed cells are associated in some way such that the hamster anti-T serum used immunoprecipitates NVTAg because it is present in a complex with large T (Lane and Crawford, 1979; Linzer and Levine, 1979). Other investigators have found that some anti-T serum appears to react directly with NVT-Ag (Lane and Crawford, 1979; Linzer and Levine, 1979; Kress et al., 1979). Figure 6 shows the result of an experiment in which mRNA from SVA31 mouse cells was fractionated by hybridization to SV40 DNA cellulose, the bound and unbound fractions were translated in vitro and the products were immunoprecipitated using a mouse anti-T serum reported to react directly with NVT-Ag (Lane and Crawford, 1979). Material which co-migrates with NVT-Ag is present in the products made in response to total transformed cell mRNA and to the mRNA in the unbound fraction, whereas SV40 large T and small t are made by mRNAs in both the total mRNA and SV40 DNA-bound fraction. Although we have not proved by peptide fingerprinting that the 53,000 dalton protein detected is identical to NVT-Ag isolated from cells, these data suggest that the mRNA
T-Ags 343
from SV40-Transformed
Cells
AB ---mm
C
DEF
CTCTCTCTCTTM
La1‘ge e60 -53
NV
-40
8 nail
Figure
-
6. Cell-Free
Synthesis of
NVT-Ag
mRNA was isolated from SVA31 cells and translated in vitro both before (tracks A and 6) and after hybridization to SV40 DNA cellulose (tracks C-E). The translation products of mRNA which bound to SV40 DNA are shown in track C. and that which did not bind in tracks D and E. The cell-free products were immunoprecipitated with either hamster control (C) or anti-T Cr) serum (track A) or mouse control (0 or anti-T (T) serum (tracks B-E). Track F is an immunoprecipitate using hamster anti-T serum of an extract of 35S-methionine-labeled WA31 cells. The immunoprecipitates were all analyzed on the same 15% polyacrylamide gel. Tracks A, B and F were exposed for 2 days, the remainder for 4 days.
for NVT-Ag does not bind to SV40 DNA under conditions where the mRNAs for SV40-coded T-Ags do bind. Discussion NVT-Ag Is Not Equivalent to Polyoma Middle T and Is Probably Not Coded for by SV4D DNA From our previous work on the amino terminal sequence and tryptic peptide fingerprint analysis of SV40 T-Ags we have been able to identify several methionine-containing peptides and to map the peptides on SV40 DNA. Of particular interest are the amino terminal tryptic peptide N-Acetyl-Met-Asp-Lys (Mellor and Smith, 1978; Paucha et al., 1978b); the peptides shared between large T and small t (Smith et al., 1978); two methionine peptides unique to small t, including one cysteine-rich peptide which is acidic at pH 2.1 and maps at position 0.585 mu (Paucha and Smith, 1978); the peptides unique to large T, including one which we have mapped to the carboxy terminal half of large T by analysis of different size forms of large T synthesized in vitro in response to SV40 complementary RNA (Paucha et al., 1978a); and MetLys, which is believed to be present in the shared region of large and small T-Ags and in both unique regions (our unpublished results). The position of these peptides in two-dimensional fingerprints (at pH 2.1 in the first dimension) is indicated in Figure 4F.
Using these peptides as markers we have examined the fingerprints of different forms of SV40 T-Ag from transformed cells. We were particularly interested in a protein, referred to here as NVT-Ag, which is specifically immunoprecipitated from SV40-transformed rat, mouse and some hamster cells and which has a molecular weight of about 53,000, similar to that of Py middle T. Examination of the fingerprints of NVTAg isolated from rat, mouse or hamster cells shows that it does not contain the amino terminal peptide of the SV40 T-Ags or the small t unique methionine peptides. These results, together with the findings that the size of NVT-Ag is unaltered in cells transformed by SV40 mutants with deletions in the region between 0.60 and 0.55 mu and that the mRNA for NVT-Ag does not hybridize to SV40 DNA, show that NVT-Ag from SV40-transformed cells is not related to SV40 large T or small t in the same way as the three Py early proteins are related to one another. Further analysis of the fingerprint data on NVT-Ag showed that it contains few if any of the methionine peptides shared between large and small T-Ags or those unique to large T. In some fingerprints major methionine peptides from NVT and large T co-migrated, although when these were reanalyzed using a different pH for separation the major peptides could all be separated. Nevertheless, some minor peptides from NVT-Ag which co-migrate with large T peptides were detected, and one of these (peptide Ye) may be present in NVT-Ag from both rat and mouse. While this result is intriguing, we do not at present believe it to be significant. We have also compared fingerprints of “P-labeled large T and NVT-Ag from mouse cells, and these experiments showed the phosphopeptides of the two proteins to be unrelated (our unpublished results). Chang et al. (1979) have also concluded that SV40 large T and NVT-Ag have very few sequences in common, and Kress et al. (1979) have concluded that the two proteins have no methionine peptides in common. Taken together all these data suggest that the bulk, probably all, of NVT-Ag is not coded for by RNA sequences translated in the same reading frame as SV40 large and small T-Ags. Thus if NVT-Ag is virus-coded it must be synthesized on an mRNA read entirely or at least predominantly in a reading frame different from that which codes for large T and small t. The sequence of the E strand of SV40 DNA indicates that there are few extensive open reading frames present apart from those used for the synthesis of large T and small t. To construct from these sequences an mRNA capable of directing the synthesis of a 50-55,000 dalton protein would require many splicing events. Furthermore, the finding that NVT-Ag has a different fingerprint when isolated from transformed cells from different species indicates that such splicing, were it to occur, would be species-specific. Although neither of these possi-
Cell 344
bilities has been ruled out, we believe that the simplest interpretation of our results is that the protein is not virus-coded. This conclusion is also supported by experiments on the cell-free synthesis of T-Ags from transformed cells. Although we found that NVT-Ag synthesized in vitro was not immunoprecipitated with our hamster anti-T serum, this complication did not prevent us from asking whether the mRNA for NVT-Ag would bind to SV40 DNA. Our earlier experiments had shown that mRNA purified by preparative scale hybridization to SV40 DNA is sufficiently pure to identify products made by the virus-specified mRNA without immunoprecipitation of the cell-free products. Indeed, the experiments shown in Figure 5 show that the mRNAs coding for several abnormal forms of T-Ag present in some transformed cells were readily purified using SV40 DNA cellulose. The abnormal forms of T-AS, including some that were not detected in extracts of labeled cells (such as the shortened forms of small t), were readily detected in the products synthesized by the purified mRNAs without immunoprecipitation. In contrast, all our efforts to detect the synthesis of NVTAg in the fraction of mRNA from transformed cells which binds to SV40 DNA were unsuccessful. In parallel experiments, mRNA isolated from cells productively infected with Py and purified by Py DNA cellulose directed the synthesis of all three Py early proteins-large T, middle T and small t (A. Mellor and A. E. Smith, unpublished results; see also Hunter et al., 1978). Experiments using mouse anti-T serum, which is reported to immunoprecipitate NVT-Ag directly (Lane and Crawford, 1979), indicated that mRNA coding for a protein with mobility similar to that of NVT-Ag could be detected in both total mRNA from SVA31 E7 mouse cells and the fraction of mRNA which does not bind to SV40 DNA cellulose. We have not isolated enough of the protein made in vitro to prove that it is identical to NVT-Ag isolated from cells, but at this stage it seems a reasonable assumption. If this is the case, then we can conclude that the mRNA for NVT-Ag does not hybridize to SV40 DNA under any of the conditions we have tested. Although these results suggest that NVT-Ag is not coded for by SV40 DNA, we cannot exclude the possibility that the mRNA which codes for NVT-Ag is virus-coded but spliced to such an extent that it can no longer form a stable hybrid with immobilized SV40 DNA, even under conditions in which we know that a 200 bp hybrid is stable (our unpublished results). At present, however, we believe the most probable explanation of our results is that NVT-Ag is not coded for by SV40 DNA. The finding of Linzer and Levine (1979) that NVT-Ag is present in uninfected embryonal carcinoma cells strongly suggests that the protein is host cell-coded.
Other Forms of T-Ag from Transformed Cells The abnormal virus-coded T-Ags detected in the experiments reported here and in a large number of other SV40-transformed cells that we have examined (F. Chaudry, E. Paucha, P. J. G. Rigby and A. E. Smith, unpublished results) fall into three subgroups: super T, large T doublets and triplets and truncated large T and small t. We do not know the origin of these different species, but they are all synthesized in vitro when mRNA from the appropriate cell line is translated (Figure 5). This suggests that the proteins arise from different mRNAs rather than by post-translational modification. Fingerprint analysis of super T shows that in one case (REF 20061, although the protein is considerably larger than large T, it contains no detectable additional methionine tryptic peptides. However, some peptides may be present in more than one copy, suggesting that part of the large T molecule may be duplicated. It could be that the mRNA coding for such a protein is generated by some abnormal splicing of SV40 high molecular weight nuclear RNA transcribed from genomes integrated in tandem. Truncated forms of large T could be generated as a result of a mutation in the integrated DNA, or alternatively they could arise from an integrated sequence which lacks a complete copy of the early region. Our earlier results together with the results of the experiments described here clearly show that SV40 does not code for major amounts of a protein analogous to Py middle T in either productively infected or transformed cells. Comparison of the DNA sequences of SV40 and Py indicates that the middle T coding region may be unique to Py (Soeda et al., 1979). We emphasize, however, that a related SV40-coded protein could be generated assuming the mRNA coding for it were spliced in a different manner. Our experiments have so far examined only the major proteins and mRNAs from SV40-infected and -transformed cells. The existence of additional minor SV40 early proteins remains possible. We are continuing our attempts to detect such minor T-Ag species. Experimental
Procedures
Cells The origin of most of the cell lines used has been reported in previous publications (Carroll and Smith, 1976; Crawford et al., 1978; Smith et al., 1978). The SV40-transformed REF cell lines were obtained from W. C. Topp (Sleigh et al., 1978). Antiserum The antiserum used in most of the experiments described was raised by injection of H65-90b cells into golden hamsters (Smith et al., 1978). The mouse serum with anti-NVT-Ag activity was a gift from D. Lane and described by Lane and Crawford (1979). Extraction and lmmunopracipitation of T-Ags Cells were grown in 30 or 90 mm plastic dishes in Dulbecco’s modified E4 containing 510% fetal calf serum. The cells were labeled with ‘?S-methionine (Radiochemical Centre. Amersham) as
T-Ags 345
from SV40-Transformed
Cells
they approached confluence. Conditions for labeling and extraction of the cells have been described (Smith et al., 1978). Conditions for immunoprecipitation were as described (Smith et al., lQ78), except that in some experiments we omitted the preabsorption step using sheep serum. Immunoprecipitates were collected using protein A-bearing S. aureus bacteria and separated on 10 or 15% discontinuous pH polyacrylamide gels (Smith et al., 1978). Peptide Fingerprinting Conditions for preparative scale immunoprecipitation, autoradiography. elution. performic oxidation and trypsin digestion have been described (Smith et al., 1978). The fingerprints were prepared using Kodak plastic-backed cellulose thin-layer sheets, and electrophoresis in the first dimension was at pH 2.1 or 6.5 as described. Autoradiography of the dried sheets was for between 2 and 25 days. To elute peptides from the thin-layer sheets for further analysis, the appropriate region of the cellulose was scraped from the plastic and eluted using 2 times 0.5 ml of 1 M formic acid. The eluted peptides were washed with water and reanalyzed on further cellulose sheets or on paper. Purification of mRNA Using SV40 DNA Cellulose Preparation of SV40 DNA and mRNA from cells grown in culture has been described (Paucha et al., 1978a). The DNA was linearized with restriction enzymes and covalently attached to cellulose using the method of Noyes and Stark (1975) as described by Paucha and Smith (1978). About 5 mg of total RNA isolated from different transformed cell lines were hybridized to cellulose containing about 200 pg of bound SV40 DNA. After washing. the hybridized mRNA was eluted. reprecipitated and resuspended in about 25 pl of water. Between 1 and 3 pl were subsequently translated in 25 pl cell-free reactions (Paucha et al., 1978a). Cell-Free Synthesis Conditions for cell-free synthesis using extracts from L cells which were sometimes pretreated with micrococcal nuclease to reduce endOQenOUS mRNA activity and immunoprecipitation of the cell-free products have been described (Paucha and Smith, 1978; Paucha et al., 1 Q78a). The immunoprecipitates were fractionated on 15% polyacrylamide gels and autoradiography was for l-l 0 days. Acknowledgments We thank Kit Osborn for help with growing cells, Dr. W. C. Topp for SV40-transformed REF cells, Paul Nicklin for help in raising antisera and growing protein A-bearing S. aureus and Dr. David Lane for mouse antiserum with anti-NVT-Ag activity. We are grateful to Drs. T. Chang and M. Kress for helpful discussions and to Drs. Kress and Levine for Sending us manuscripts prior to publication. We thank Miss Anne Roberts for help with the preparation of the manuscript and Drs. F. McCormick and M. Hayman for critical readinQS. E. P. was supported by a fellowship from the Canadian MRC. Received
May 10, 1979;
revised
July 23, 1979
References Carroll, R. B and Smith, A. E. (1976). T antigen from simian virus 40-infected Nat. Acad. Sci. USA 73, 2254-2258. Chang. C.. Pancake, S. J.. Luborsky. Detergent solubilization and partial surface and transplantation antigens cells. Int. J. Cancer 19, 258-266.
Monomer molecular Weight of and transformed cells. Proc. S. W. and Mora. P. T. (1977). purification of tumor Specific from SV40 transformed mouse
Chang. C., Simmons, 0. T.. Martin, M. A. and Mora. P. T. (1979). Identification and partial characterization of tumor antigens from SV40 transformed mouse cells. J. Virol. 31. 463-471. Crawford, L. V., Cole, C. N.. Smith, A. E.. Paucha. E.. Tegtmeyer. P.. Rundell. K. and Berg, P. (1978). Organization and expression of early
genes
of simian
virus 40. Proc.
Nat. Acad.
Sci. USA 75, 117-l
21,
Edwards, C. A., Khoury, G. and Martin, R. G. (1979). Phosphorylation of T-antigen expression in cells transformed by wild type and tsA mutants of SV40. J. Virol. 29, 753-762. Elder. K. T.. Bye. J. M.. Waterfield. M. D., Skehel, J. J. and Smith. A. E. (1979). In vitro synthesis, glycosylation, and membrane insertion of influenza virus haemagglutinin. Virology 95, 343-350. Feunteun, J., Sompayrac. L., Fluck, M. and Benjamin, T. (1976). Localization of gene functions in polyoma virus DNA. Proc. Nat. Acad. Sci. USA 73, 4169-4173. Fiers, W.. Contreras, R., Haegeman. G.. Rogiers. R.. Van de Voorde, A.. Van Heuverswyn. H.. Van Herreweghe. J., Volckaert, G. and Ysabaert. M. (1978). Complete nucleotide sequence of SV40 DNA. Nature 273, 113-l 20. Gaudray. P.. Rassoulzadegan, M. and Cuzin. F. (1978). Expression of simian virus 40 early genes in transformed rat cells is correlated with maintenance of the transformed phenotype. Proc. Nat. Acad. Sci. USA 75, 4987-4991. Hunter, T., Hutchinson. M. A. and Eckhart. polyoma virus T antigens in vitro. Proc. 5917-5921.
W. (1978). Nat. Acad.
Hutchinson. M. A., Hunter, T. and Eckhart, W. (1978). tion of T antigens in polyoma-infected and transformed 65-77. Ito. T.. Spurr. N. and polyoma virus T-antigen.
Translation Sci. USA
Characterizacells. Cell 15,
Dulbecco, R. (1977a). Characterization Proc. Nat. Acad. Sci. USA 74, 1259-l
Ito. Y.. Brocklehurst. J. R. and Dulbecco. proteins in the plasma membrane of cells formed by polyoma virus. Proc. Nat. Acad.
of 75,
of 263.
R. (1977b). Virus-specific lytically infected or transSci. USA 74, 4666-4670.
Kress. M., De Vaux Saint Cyr. C. and Girard. M. (1977). weight of SV40 T-antigens. INSERM 69, 79-92.
The molecular
Kress, M., May, E., Cassingena, R. and May, P. (1979). SV40 transformed cells express new species of proteins precipitable by SV40 anti-tumor serum. J. Virol. 31, 472-483. Lane, D. P. and Crawford, L. V. (1979). T-antigen is bound protein in SV40 transformed cells. Nature 2 78, 261-283. Lichaa. present
to a host
M. and Niesar, E. (1977). Characterization of SV40 T-antigen in SV40 transformed cells. INSERM 69, 21 l-222.
Linzer. D. I. H. and Levine, A. J. (1979). Characterization of a 54,OOOK dalton cellular SV40 tumor antigen present in SV40-transformed cells and uninfected embryonal carcinoma cells. Cell 17, 4352. Melero, J. A., Stitt. D. T., Mangel. W. F. and Carrel, R. B. (1979). Identification of a new polypeptide species (48-55K) immunoprecipitable by antiserum to purified large-T antigen in SV40 infected and transformed cells. Virology 93, 466-480. Mellor. A. and Smith, A. E. (1978). Characterization of the aminoterminal tryptic peptide of simian virus 40 small-t and large-T antigens. J. Virol. 28, 992-996. Noyes. B. E. and Stark, G. R. (1975). DNA covalently coupled to cellulose.
Nucleic acid hybridization Cell 5, 301-310.
Paucha. E. and Smith, A. E. (1978). The sequences and 0.54 map units on simian virus 40 DNA code region of small t antigen. Cell 15, 101 l-l 020. Paucha, E., Harvey, R. and Smith, A. E. (1978a). of simian virus 40 T-antigens. J. Virol. 28, 154-l
using
between 0.59 for the unique
Cell free synthesis 70.
Paucha, E.. Mellor, A., Harvey, R., Smith, A. E.. Hewick. R. M. and Waterfield. M. D. (1 Q78b). Large and small tumor antigens from SV40 have identical amino termini mapping at 0.65 map units. Proc. Nat. Acad. Sci. USA 75, 2165-2189. Prives. C.. Gluzman, Y. and Winocour, E. (1978). Cellular and cellfree synthesis of simian virus 40 T-antigens in permissive and transformed cells. J. Virol. 25, 587-595. Prives,
C., Gilboa.
E., Revel,
M. and Winocour.
E. (1977).
Cell-free
Cell 346
translation of simian virus 40 early messenger RNA coding antigen. Proc. Nat. Acad. Sci. USA 74, 457-461. Reddy, V. 6.. Thimmappaya, 6.. Dhar. Ft., Subramanian. 8. S.. Pan, J.. Celma, M. L. and Weissman, S. M. (1978). of SV40. Science 200, 494-502. Rundell. K., Collins, Nathans, D. (1977). Virol. 2 7, 636-646.
for viral TK. N.. Zain, The genome
J. K., Tegtmeyer, P., Ozer, H. L., Lai, C. and Identification of simian virus 40 protein A. J.
Schaffhausen, 6. S., Silver, J. E. and Benjamin, T. L. (1978). Tumor antigen(s) in cells productively infected by wild-type polyoma virus and mutant NG-10. Proc. Nat. Acad. Sci. USA 75, 79-83. Schmidt-Ullrich, R., Scott-Thompson, W., Peck, S. L. and Wallack, D. F. H. (1977). Simian virus 40-specific proteins in the membranes of simian virus 40-transformed hamster and mouse cells. Proc. Nat. Acad. Sci. USA 74, 5069-5072. Simmons, D. T. and Martin, M. A. (1978). Common methionine peptides near the amino terminal end of primate papovavirus antigens. Proc. Nat. Acad. Sci. USA 75, 1131-l 135.
tryptic tumor
Sleigh, M. J., Topp. W. C.. Hanich, R. and Sambrook, J. F. (1978). Mutants of SV40 with an altered small t protein are reduced in their ability to transform cells. Cell 14, 79-88. Smart, J. E. and Ito, Y. (1978). Three species of polyoma virus tumor antigens share common peptides probably near the amino termini of the proteins. Cell 15, 1427-l 438. Smith, A. E., Smith, R. and Paucha, E. (1978). Extraction and fingerprint analysis of SV40 large and small T-antigens. J. Virol. 28, 140153. Soeda. E.. Arrand, J. R., Smolar, N. and Griffin, B. E. (1979). Sequence from early region of polyoma virus DNA containing viral replication origin and encoding small, middle and (part of) large T antigens. Cell 17, 357-370. Volckaert. G., Van de Voorde. A. and Fiers, W. (1978). Nucleotide sequence of SV40 small-t gene. Proc. Nat. Acad. Sci. USA 75, 216021 64.