Cell, Vol. 18. 1299-l
Splicing
302.
December
1979,
Copyright
0 1979
by MIT
and the Formation
Dean H. Hamer* and Philip Ledert *Recombinant DNA Research Unit National Institute of Allergy and Infectious and tLaboratory of Molecular Genetics National Institute of Child Health and Human Development Bethesda, Maryland 20205
Diseases
Summary To determine whether RNA splicing plays an obligatory role in gene expression, we have constructed a series of SV4CMransducing viruses carrying various combinations of splice junctions derived from the viral genome and a mouse globin gene. All of the viruses that retain at least one functional splice junction, derived from either the viral or the mouse genome, encode stable hybrid RNAs. In contrast, a virus from which all the splice junctions have been removed fails to produce any detectable stable RNA. These results suggest that splicing is a prerequisite for stable RNA formation. Introduction Many eucaryotic genes are discontinuous-that is, the information in the DNA is not colinear with the information found in the corresponding messenger RNA or protein. Instead, these genes are interrupted by intervening sequences of noninformational DNA (Brack and Tonegawa, 1977; Jeffreys and Flavell, 1977; Berget, Moore and Sharp, 1977; Breathnach, Mandel and Chambon, 1977; Kitchingman, Lai and Westphal, 1977; Aloni et al., 1977; Chow et al., 1977; Doe1 et al., 1977; Leder et al., 1977; Tilghman et al., 1977; Berk and Sharp, 1978; Valenzuela et al., 1978; Weinstock et al., 1978). The formation of functional mRNAs from such discontinuous genes has been studied in detail in several systems (Smith and Lingrel, 1978; Kinniburgh, Mertz and Ross, 1978; Blanchard et al., 1978; Tilghman et al., 1978a). In each case, the primary transcript contains the intervening as well as the structural sequences. This precursor RNA is then processed so as to remove the intervening sequences and rejoin the structural sequences in the appropriate order. This processing reaction has been termed RNA splicing. Several hypotheses as to the biological significance of discontinuous genes and RNA splicing have already been advanced. These include models in which discontinuous genes might, in certain instances, accelerate evolution by facilitating the formation of new proteins by combining parts of old ones (Gilbert, 19781, or, in other instances, stabilize duplicate genes by acting as barriers to recombination (Tiemeier et al., 1978). It has also been suggested that interrupted
of Stable RNA
genes might represent a primitive genetic structure inherited from the common ancester of procaryotes and eucaryotes (Darnell, 1978; Doolittle, 1978). Although there is some evidence to support each of these hypotheses, they are all based upon evolutionary arguments and are therefore not readily susceptible to a direct experimental test. On the other hand, the physiological role that RNA splicing plays in gene expression can be tested experimentally. To discover whether such a role (or roles) exists and what it might be, we have characterized the RNAs encoded by a series of SV40 recombinants carrying a splice junction from the 3’ portion of the chromosomal mouse pmai globin gene (Konkel, Tilghman and Leder, 1978; Tilghman et al., 1978b). By varying the structure of the SV40 vector and the orientation of the globin fragment, it was possible to create all possible combinations of viral and chromosomal splice junctions, and hence to determine whether splicing plays an obligatory role in gene expression. We find that stable RNAs are produced by all of the viruses that retain at least one functional splice junction, regardless of its viral or chromosomal origin. The virus from which all of the splice junctions have been eliminated, however, does not accumulate stable RNA. It therefore seems that splicing is a prerequisite for stable RNA formation. Results
and Discussion
Figure 1 shows the structures of the recombinant viruses (designated l-4) used in this study. These recombinants were constructed using two different SV40 vectors derived by excision of viral late region sequences. Both types of vector retain the complete viral early gene region, the origin of DNA replication, the extreme 3’ late mRNA sequences, including the polyadenylation site, and the extreme 5’ late mRNA sequences, including the late region promoter (Fiers et al., 1978; Reddy et al., 1978). The two vectors differ in that one retains the splice junction for the 19s late mRNA, whereas the second lacks this region and contains no known late mRNA splice junctions (Ghosh et al., 1979). These two vectors were used to clone a fragment from the 3’ portion of the chromosomal mouse pma’ globin gene. As we have shown previously, this fragment contains a splice junction that is utilized when it is transcribed in the sense but not the antisense orientation (Hamer and Leder, 1979). It was therefore possible to control the presence of a functional globin splice junction simply by inserting the globin fragment in both possible orientations relative to the SV40 late promoter. As shown in Figure 1, the resulting four viruses contain all possible combinations of SV40 and globin splice junctions. All these viruses have been propagated in African green monkey kidney cells, the permissive host for SV40. by
Cell 1300
VIRUS
ORIENTATION OF GLOBIN FRAGMENT
STRUCTURES OF RECOMBINANT VIRUSES AND THEIR TRANSCRIPTS
SPLICE JUNCTIONS --SV40 GLOBIN
STABLE RNA
1
SENSE
2
ANTI-SENSE
t
3
SENSE
+
4
ANTI-SENSE
= Figure
1.
Structures and
Tentative
SV40 structural sequence SV40 flanking sequence SV40 intervening sequence Transcription
Maps
of Recombinant
m 0
globin structural sequence globin flanking sequence
a
globin intervening sequence (sense)
m
globin intervening sequence (anti-sense)
Viruses
The bar diagrams show the structures of the late regions of the four SV4O-globin recombinant viruses extending (from left to right) from the origin of DNA replication (0.67 map units) to the late region termination signal (0.17 map units). Although not shown in this figure, each of the viruses retains the complete SV40 early gene region. Viruses 1 and 2 also retain the SV40 sequences between 0.72 and 0.82 map units. a region that includes the intervening sequence and splice junction for the viral late 19s mRNA. whereas viruses 3 and 4 lack this region. The mouse chromosomal j? globin DNA fragment inserted in these viruses contains an intervening sequence and splice junction derived from the 3’ portion of the gene. This fragment is inserted in the sense orientation relative to the SV40 late promoter in viruses 1 and 3. and consequently the globin splice junction is utilized. In contrast, viruses 2 and 4 carry the globin fragment in the anti-sense orientation relative to the SV40 promoter and hence the globin splice junction is not functional. The stabilities and tentative structures of the RNAs encoded by these viruses were determined by gel transfer hybridization experiments and electron microscopic observations (Hamer and Leder, 1979; Figure 2).
mixed infection with a complementing SV40 temperature-sensitive mutant as helper. Because each of the recombinants retains the SV40 late promoter, whereas the globin fragment contains no known promoter, we anticipated that in each case the globin sequences would be transcribed into a hybrid transcript initiated within the viral DNA. Our previous analysis of the RNAs encoded by the two viruses that retain the SV40 splice junction (viruses 1 and 2) confirmed the production of such hybrid RNAs with stabilities comparable to that of SV40 late mRNA (Hamer and Leder, 1979). To characterize the globin RNAs produced in monkey cells infected with the two viruses lacking the SV40 splice junction (viruses 3 and 4), we performed the gel transfer hybridization experiment shown in Figure 2. In this experiment, total poly(A)-containing RNA from infected cells was denatured, electrophoresed through an agarose gel, transferred and covalently linked to diazobenzyloxymethyl paper, then hybridized to various radioactive DNA probes to reveal the separated RNA species (M&taster and Carmichael, 1977; Alwine, Kemp and Stark, 1977). When SV40 DNA was used as probe (lanes 1 and 2), both preparations showed the characteristic 16s (1500 bases) and 19s (2300 bases)
late mRNA species derived from transcription of the helper virus. When a fragment containing globin-coding and 3’ flanking sequences was used as probe (lanes 3 and 4), we could not detect any stable globin transcripts produced by virus 4, which carries the globin fragment in the anti-sense orientation and thus contains no functional RNA splice junctions. In contrast, virus 3, which carries the globin fragment in the sense orientation and thus contains a functional splice junction, produced two stable globin RNAs with lengths of approximately 1400 and 500 bases. Neither of these globin RNAs hybridized to a globin intervening sequence probe (lanes 5 and 6), indicating that they are spliced. Based on our previous experience with viruses 1 and 2, we believe that the 500 base globin RNA is terminated at the globin gene polyadenylation site, whereas the less abundant 1400 base species is terminated at the downstream SV40 polyadenylation site. These observations complete the table shown in Figure 1. Stable RNAs are produced by the viruses that contain functional splice junctions derived from the virus, the globin insert or both. Only the virus that contains no functional splice junctions fails to produce a stable globin RNA. We conclude that the presence
Splicing 1301
and Formation
of Stable
RNA
5
1 2
Figure 2. Gel Transfer by Viruses 3 and 4
Hybridization
Analysis
of the RNAs
6
Encode
Splicing
302.
December
1979,
Copyright
0 1979
by MIT
and the Formation
Dean H. Hamer* and Philip Ledert *Recombinant DNA Research Unit National Institute of Allergy and Infectious and tLaboratory of Molecular Genetics National Institute of Child Health and Human Development Bethesda, Maryland 20205
Diseases
Summary To determine whether RNA splicing plays an obligatory role in gene expression, we have constructed a series of SV4CMransducing viruses carrying various combinations of splice junctions derived from the viral genome and a mouse globin gene. All of the viruses that retain at least one functional splice junction, derived from either the viral or the mouse genome, encode stable hybrid RNAs. In contrast, a virus from which all the splice junctions have been removed fails to produce any detectable stable RNA. These results suggest that splicing is a prerequisite for stable RNA formation. Introduction Many eucaryotic genes are discontinuous-that is, the information in the DNA is not colinear with the information found in the corresponding messenger RNA or protein. Instead, these genes are interrupted by intervening sequences of noninformational DNA (Brack and Tonegawa, 1977; Jeffreys and Flavell, 1977; Berget, Moore and Sharp, 1977; Breathnach, Mandel and Chambon, 1977; Kitchingman, Lai and Westphal, 1977; Aloni et al., 1977; Chow et al., 1977; Doe1 et al., 1977; Leder et al., 1977; Tilghman et al., 1977; Berk and Sharp, 1978; Valenzuela et al., 1978; Weinstock et al., 1978). The formation of functional mRNAs from such discontinuous genes has been studied in detail in several systems (Smith and Lingrel, 1978; Kinniburgh, Mertz and Ross, 1978; Blanchard et al., 1978; Tilghman et al., 1978a). In each case, the primary transcript contains the intervening as well as the structural sequences. This precursor RNA is then processed so as to remove the intervening sequences and rejoin the structural sequences in the appropriate order. This processing reaction has been termed RNA splicing. Several hypotheses as to the biological significance of discontinuous genes and RNA splicing have already been advanced. These include models in which discontinuous genes might, in certain instances, accelerate evolution by facilitating the formation of new proteins by combining parts of old ones (Gilbert, 19781, or, in other instances, stabilize duplicate genes by acting as barriers to recombination (Tiemeier et al., 1978). It has also been suggested that interrupted
of Stable RNA
genes might represent a primitive genetic structure inherited from the common ancester of procaryotes and eucaryotes (Darnell, 1978; Doolittle, 1978). Although there is some evidence to support each of these hypotheses, they are all based upon evolutionary arguments and are therefore not readily susceptible to a direct experimental test. On the other hand, the physiological role that RNA splicing plays in gene expression can be tested experimentally. To discover whether such a role (or roles) exists and what it might be, we have characterized the RNAs encoded by a series of SV40 recombinants carrying a splice junction from the 3’ portion of the chromosomal mouse pmai globin gene (Konkel, Tilghman and Leder, 1978; Tilghman et al., 1978b). By varying the structure of the SV40 vector and the orientation of the globin fragment, it was possible to create all possible combinations of viral and chromosomal splice junctions, and hence to determine whether splicing plays an obligatory role in gene expression. We find that stable RNAs are produced by all of the viruses that retain at least one functional splice junction, regardless of its viral or chromosomal origin. The virus from which all of the splice junctions have been eliminated, however, does not accumulate stable RNA. It therefore seems that splicing is a prerequisite for stable RNA formation. Results
and Discussion
Figure 1 shows the structures of the recombinant viruses (designated l-4) used in this study. These recombinants were constructed using two different SV40 vectors derived by excision of viral late region sequences. Both types of vector retain the complete viral early gene region, the origin of DNA replication, the extreme 3’ late mRNA sequences, including the polyadenylation site, and the extreme 5’ late mRNA sequences, including the late region promoter (Fiers et al., 1978; Reddy et al., 1978). The two vectors differ in that one retains the splice junction for the 19s late mRNA, whereas the second lacks this region and contains no known late mRNA splice junctions (Ghosh et al., 1979). These two vectors were used to clone a fragment from the 3’ portion of the chromosomal mouse pma’ globin gene. As we have shown previously, this fragment contains a splice junction that is utilized when it is transcribed in the sense but not the antisense orientation (Hamer and Leder, 1979). It was therefore possible to control the presence of a functional globin splice junction simply by inserting the globin fragment in both possible orientations relative to the SV40 late promoter. As shown in Figure 1, the resulting four viruses contain all possible combinations of SV40 and globin splice junctions. All these viruses have been propagated in African green monkey kidney cells, the permissive host for SV40. by
Cell 1300
VIRUS
ORIENTATION OF GLOBIN FRAGMENT
STRUCTURES OF RECOMBINANT VIRUSES AND THEIR TRANSCRIPTS
SPLICE JUNCTIONS --SV40 GLOBIN
STABLE RNA
1
SENSE
2
ANTI-SENSE
t
3
SENSE
+
4
ANTI-SENSE
= Figure
1.
Structures and
Tentative
SV40 structural sequence SV40 flanking sequence SV40 intervening sequence Transcription
Maps
of Recombinant
m 0
globin structural sequence globin flanking sequence
a
globin intervening sequence (sense)
m
globin intervening sequence (anti-sense)
Viruses
The bar diagrams show the structures of the late regions of the four SV4O-globin recombinant viruses extending (from left to right) from the origin of DNA replication (0.67 map units) to the late region termination signal (0.17 map units). Although not shown in this figure, each of the viruses retains the complete SV40 early gene region. Viruses 1 and 2 also retain the SV40 sequences between 0.72 and 0.82 map units. a region that includes the intervening sequence and splice junction for the viral late 19s mRNA. whereas viruses 3 and 4 lack this region. The mouse chromosomal j? globin DNA fragment inserted in these viruses contains an intervening sequence and splice junction derived from the 3’ portion of the gene. This fragment is inserted in the sense orientation relative to the SV40 late promoter in viruses 1 and 3. and consequently the globin splice junction is utilized. In contrast, viruses 2 and 4 carry the globin fragment in the anti-sense orientation relative to the SV40 promoter and hence the globin splice junction is not functional. The stabilities and tentative structures of the RNAs encoded by these viruses were determined by gel transfer hybridization experiments and electron microscopic observations (Hamer and Leder, 1979; Figure 2).
mixed infection with a complementing SV40 temperature-sensitive mutant as helper. Because each of the recombinants retains the SV40 late promoter, whereas the globin fragment contains no known promoter, we anticipated that in each case the globin sequences would be transcribed into a hybrid transcript initiated within the viral DNA. Our previous analysis of the RNAs encoded by the two viruses that retain the SV40 splice junction (viruses 1 and 2) confirmed the production of such hybrid RNAs with stabilities comparable to that of SV40 late mRNA (Hamer and Leder, 1979). To characterize the globin RNAs produced in monkey cells infected with the two viruses lacking the SV40 splice junction (viruses 3 and 4), we performed the gel transfer hybridization experiment shown in Figure 2. In this experiment, total poly(A)-containing RNA from infected cells was denatured, electrophoresed through an agarose gel, transferred and covalently linked to diazobenzyloxymethyl paper, then hybridized to various radioactive DNA probes to reveal the separated RNA species (M&taster and Carmichael, 1977; Alwine, Kemp and Stark, 1977). When SV40 DNA was used as probe (lanes 1 and 2), both preparations showed the characteristic 16s (1500 bases) and 19s (2300 bases)
late mRNA species derived from transcription of the helper virus. When a fragment containing globin-coding and 3’ flanking sequences was used as probe (lanes 3 and 4), we could not detect any stable globin transcripts produced by virus 4, which carries the globin fragment in the anti-sense orientation and thus contains no functional RNA splice junctions. In contrast, virus 3, which carries the globin fragment in the sense orientation and thus contains a functional splice junction, produced two stable globin RNAs with lengths of approximately 1400 and 500 bases. Neither of these globin RNAs hybridized to a globin intervening sequence probe (lanes 5 and 6), indicating that they are spliced. Based on our previous experience with viruses 1 and 2, we believe that the 500 base globin RNA is terminated at the globin gene polyadenylation site, whereas the less abundant 1400 base species is terminated at the downstream SV40 polyadenylation site. These observations complete the table shown in Figure 1. Stable RNAs are produced by the viruses that contain functional splice junctions derived from the virus, the globin insert or both. Only the virus that contains no functional splice junctions fails to produce a stable globin RNA. We conclude that the presence
Splicing 1301
and Formation
of Stable
RNA
5
1 2
Figure 2. Gel Transfer by Viruses 3 and 4
Hybridization
Analysis
of the RNAs
6
Encode
