Current Genetics
Curr Genet (1992) 22:37-40
9 Springer-Verlag 1992
Alternative modes of m R N A processing in a 3' splice site mutant of Neurospora crassa Jane H. Kinnaird 3, Dean F. Revell 2, Ian E Connerton 2, Isobelle Hasleham l, and John R. S. Fincham 1 1 Department of Genetics, University of Cambridge, England, UK 2 AFRC Institute of Food Research, Shinfield, Reading, England, UK 3 Wellcome Unit of Molecular Parasitology, University of Glasgow, Scotland, UK Received December 18, 1991
Summary. The
m u t a n t of N e u r o s p o r a c r a s s a is shown to have a double base-pair change, G T G for the normal TAG, at the 3' end of the second intron of the a m (NADP-specific glutamate dehydrogenase, G D H ) gene. The greater part of the m u t a n t a m transcript accumulates as two fragments hybridising to probes for sequences respectively upstream and downstream of the 5' e~d of the intron. Two processed transcripts approximating to normal full length m R N A were identified. In one the second intron was intact; in the other the second intron was spliced out through the use of an A A G sequence, 20 base-pairs into the third exon, as a 3' acceptor site. A G A G sequence, only four base-pairs downstream from the normal acceptor site, does not appear to be used.
Key words:
am s
Neurospora
crassa -
Intron - Alternative
splicing
Materials and methods Neurospora mutant. The mutant, labelled in our stock collection as am4-6-2, was one of those used for fine-structure mapping of the am
(NADP-specific glutamate dehydrogenase) gene (Fincham 1967). It was supposed to have been derived, through six successive backcrosses to a standard wild-type, from the am 4 mutant, reported by Roberts (1971) to produce enzymically inactive GDH protein, identified as serologically cross-reacting material (CRM). Recent unpublished results in our laboratories point to a mistake in strain designations. The CRM-positive status of the original am 4 mutant has been confirmed and its mutational change defined as GGC to GAC (glycine to aspartate) in codon 112; an identical change has been found in two CRM-positive backcrossed strains supposedly derived from am s, a CRM-negative mutant induced by ultraviolet light. At the same time, "am4-6-2'' turns out to be CRM-negative. The original am s has been lost, but we assume that the am'* and am 8 labels were switched at some stage of backcrossing and that "am4-6-2'' is a am ~ derivative (the only one extant). In this paper the " a m 4 - 6 - 2 '' strain is simply referred to as am s. Cloning and sequencing procedures were similar to those of Kinnaird and Fincham (1983). RNA was prepared by the procedure of SokoloWski et al. (1990) and freed of DNA by treatment with RNase-free DNase. Polyadenylated RNA was prepared from total DNA-free RNA by use of an oligo-dT cellulose column (ManiatiS et al. 1982). Polymerase chain reactions (PCRs) from either genomic DNA or cDNA followed the protocols of Sambrook et al. (1989). Molecular methods.
Introduction Nuclear introns o f fungi, like those o f other eukaryotes, invariably have the dinucleotide sequence A G at their 3' termini, almost always preceded by a pyrimidine residue~ There seems, however, to be no general and absolute requirement for a pyrimidine in this position. In S a c c h a r o m y c e s c e r e v i s i a e the general rule is that 3' splicing takes place at the first A G sequence downstream f r o m the "lariat" branch point ( R u b y and Abelson 1991). In N e u r o s p o r a c r a s s a , Kuiper et al. (1988) showed that, in a m u t a n t with T A G changed to TAA at the 3' end of an intron, default splicing occurred efficiently at a 3' T A G sequence five bases further downstream. To our knowledge there has been, no test in N e u r o s p o r a of the use of a n A A G or G A G sequence in default of PyAG.
Offprint requests to:
J. R. S. Fincham
Results and discussion Sequencing of the a m 8 allele revealed a double base-pair mutation at the 3' end o f the second intron. The wild-type TAG 3'-splice Sequence is changed to G T G (Fig. 1). Northern blots of total wild-type R N A , probed with the 2.64 kb B a m H 1 fragment covering the entire a m gene (Kinnaird and Fincham 1983), showed a single band of m R N A , k n o w n to be approximately 1.7 kb. A similar analysis of a m s reveals a fainter band of approximately the same size, together with two smaller and more abundant fragments. To test the hypothesis that these smaller fragments corresponded to unligated splicing intermediates, replicate strips of Northern blots were hybridised to
38
1. Sequences of wild-type and of the am 8 allele in the region of the 3' junction of the second intron. The sequences shown are those of the antisense DNA strands, and correspond to CTCTAGTG and CTCGTGTG in the sense-strands of wild-type and mutant respectively.
Fig.
, 200b ,
Bglll
3'
5'
ANNI
m
2
B [
4
I
T
(
Fig. 2A. Structure of the am primary transcript, with introns black and 5' and 3' untranslated sequences shaded; the arrow indicates the position of the unique BglII restriction site. B the 5' 40% of the transcript expanded, with thefilled triangle showing the position of the am 8 mutation, and the horizontal arrows showing the positions and orientations of the primers used for PCR on eDNA. Primer 1 was a 20-mer corresponding to bases 244-253 of the B a m H l fragment encompassing the am gene (Kinnaird and Fincham 1983); primer 2 was a 24-met corresponding to bases 579-602; primer 3 was a 24-mer corresponding to the complement of bases 949-972. The black triangle indicates the position of the am 8 mutation Key to probes:
separated fragments of the a m probe obtained by cutting at the unique B g l I I site located 21 nucleotides upstream of the second intron (Fig. 2). The smaller of the two transcript fragments hybridised exclusively to the upstream probe (Fig. 3A, track 2; Fig. 3 B, track 1), the larger fragment hybridised predominantly to the downstream probe (Fig. 3 A, track 4; Fig. 3 B, track 2), and the approximately normal-sized a m a transcripts hybridised strongly to both probes. The upstream transcript fragment, presumably lacking a polyadenylation signal, was not detected in R N A fractionated by binding to oligo-dT cellulose (Fig. 3 B, track 4). In order to determine the composition of the approximately 1.7 kb polyadenylated a m 8 R N A , the polyadenylated R N A fraction was reverse-transcribed using an antisense primer (primer 3) complementary to sequence downstream from the second intron, and products spanning the second intron were amplified as double-stranded
U = Upstream D = Downstream
Fig. 3A, B. Lanes from two Northern blots (A, B) of RNA from
wild-type and am 8 RNA probed in sections with the total am gene (2.64 kb BamH1 fragment) or with upstream or downstream BglII subfragments (see Fig. 2), A I , wild-type and A2, am 8 total RNA, both probed with the upstream fragment; A3, wild-type and A4, am 8 total RNA, both probed with the downstream fragment; B1, B2 and B3 all am a total RNA, probed respectively with upstream, downstream and total am fragments; B4 polyadenylated am 8 RNA probed with the total am fragment. Different am 8 RNA preparations were used for the A and B blots. The amounts of RNA loaded were approximately 1.5 gg (A1, A3), 3.5 p,g (A2, A4), 2.5 gg ( B I - 3 ) and 0.5 btg (B4)
D N A by P C R after addition of a second primer corresponding to sequence either upstream of the first intron (primer 1) or between the two introns (primer 2 - see Fig. 2). Analysis of the reaction products on 0.75% agarose revealed in each case, a strong, apparently dou-
39 Ke y: C = cDNA template G = Genomic DNA template M = C + G products mixed
Fig. 4. Size analysis of PCR products on a 4% polyacrylamide electrophoresis gel. 1, primer 1 + 3 product from genomic DNA including both introns - predicted size 731 bp; 2, primer 1 +3 products from a m 8 cDNA; 3, mixture of 1 and 2; 4, primer 2 + 3 products from a m 8 cDNA; 5 a n d 6, primer 2 + 3 product from genomic DNA - predicted size 394 bp. The flanking tracks are size markers; sizes are indicated in bp
OM ...
CCTACTGGCC~GTAAT
G G..4 6b .,GTCTCT 7~GIT G A G C A T G G G T G G T G G C A A G G G T G G T . . .
[B ) .., C C T A C T G G C C G T A A G T G , . 4 6 b . . G T C T C G T G T G A G C A T G G G T G G T G G C A A G G G T G G T . . .
jJ
(C)
.
.
.
.
.
.
Fig. 5. A shows the wild-type DNA sequences spanning the ends of the a m second intron, with the normal intron termini shown by arrows. B, C are the sequences of cloned cDNA PCR products, generated by primers 2+3 from a m 8 cDNA. Of four clones sequences, three had sequence B and one sequence C
ble, D N A band in the ranges 650 700 bp and 3 0 0 400 bp for the 1 + 3 and 2 + 3 primer combinations respectively. The doublets were each resolved in 4% polyacrylamide gels into two discrete bands of approximate sizes 320 bp and 400 bp for primers 2 + 3, and 590 bp and 670 bp for primers 1 + 3 (Fig. 4). The more abundant, approximately 400 bp, c D N A fragment was indistinguishable in size from the PCR fragment (of predicted size 394 bp) obtained with the same primers from genomic DNA, and seemed likely to be due to failure to splice out the second intron. The estimated size o f the less abundant fragment, 320 bp, was slightly smaller than expected from normal splicing of the second intron (expected size 333 bp). O f the two primer 1 + 3 products, the larger was clearly smaller than the fragment generated by PCR with primers 1 + 3 from genomic D N A and was consistent in
size with removal of the first intron only. The size of the smaller product was consistent with removal of both introns. The polyacrylamide gel also revealed the presence of other much larger and much smaller PCR products in lower yield (Fig. 4); these are presumed to be PCR artefacts and have not been further analysed. For sequence analysis, an additional supply of the approximately 320+400 bp c D N A fragment mixture was generated from the primer 1 + 3 PCR products by a second PCR using primers 2 + 3. The two main products appeared identical in size (data not shown) to those obtained with primers 2 + 3 directly from the c D N A (Fig. 4, track 4). Several fragments were recovered as clones in the 'Bluescript' ptasmid and sequenced, with the results summarized in Fig. 5. The larger product retains the second intron sequence in its entirety. The sequence of the smaller product shows the second exon joined to the 5' end of a glycine codon 20 nucleotides into the third exon, indicating that an A A G lysine codon is being used as an alternative 3' splice site. This mode of splicing will generate a + 1 frameshift in the coding sequence of the third exon, and is expected to yield a 317 bp PCR product with primers 2 + 3, in good agreement with observation. If the G A G glutamate codon situated only four bases downstream from the normal intron boundary were also used as a surrogate splice site one would expect a 333 bp P C R product, which should have been clearly separable from 317 bp on 4% polyacrylamide. No hint of doubleness in the smaller PCR product has been observed in several different polyacrylamide electrophoresis gels, and we therefore conclude that the G A G sequence is recognized as a 3' splice signal in N e u r o s p o r a only with very
40 low efficiency, if at all. Our results do show that the Neurospora spliceosome recognizes A A G to an appreciable extent. It may be that, in our example, the A A G sequence would have been used with greater efficiency had its distance (33 bases) from the putative intron branch-point C T G A C approximated more closely to the 1 0 - 2 0 bp typical of Neurospora introns. It is worth noting that at least one example of a naturally occurring A A G 3' intron splice signal has been reported in a filamentous fungus - in intron 6 of the facA (acetyl-coenzyme A synthetase) gene of Aspergillus nidulans (Connerton et al. 1990). In comparable studies on Saccharomyces (Fouser and Friesen 1987), the lack of an acceptable 3' acceptor site was not found to inhibit cleavage at the 5' end of the intron. Our results, on the other hand, show a degree o f dependence of 5' cleavage on the 3' splice sequence, since in a substantial proportion (perhaps about 20 per cent) of am 8 transcripts the 5' cut is not made and the intron survives intact. This suggests that fully efficient 5' cleavage depends on the entire spliceosome complex and not just on the consensus 5' slice signal.
Acknowledgements, We acknowledge helpful discussion with Dr. J. Umen. References Connerton IE Fincham JRS, Sandeman RA, Hynes M (1990) Mol Microbiol 4:451 460 Fincham JRS (1967) Genet Res 9:49-62 Fouser LA, Friesen JD (1987) Mol Cell Biol 7:225-230 Kinnaird JH, Fincham JRS (1983) Gene 26:253-260 Kuiper MTR, Holtrop M, Vennema H, Lambowitz AM, de Vries H (1988) J Biol Chem 263:2848-2852 Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning: laboratory manual, 1st edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Roberts DB (1971) J Gen Microbiol 69:143-144 Ruby SW, Abelson J (1991) Trends Genet 7:79-85 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual. 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Sokolowski V, Kaldenhoff R, Ricci M, Russo WEA (1990) Fungal Genetics. Newslett 37:41-43 Communicated by B. S. Cox
Curr Genet (1992) 22:37-40
9 Springer-Verlag 1992
Alternative modes of m R N A processing in a 3' splice site mutant of Neurospora crassa Jane H. Kinnaird 3, Dean F. Revell 2, Ian E Connerton 2, Isobelle Hasleham l, and John R. S. Fincham 1 1 Department of Genetics, University of Cambridge, England, UK 2 AFRC Institute of Food Research, Shinfield, Reading, England, UK 3 Wellcome Unit of Molecular Parasitology, University of Glasgow, Scotland, UK Received December 18, 1991
Summary. The
m u t a n t of N e u r o s p o r a c r a s s a is shown to have a double base-pair change, G T G for the normal TAG, at the 3' end of the second intron of the a m (NADP-specific glutamate dehydrogenase, G D H ) gene. The greater part of the m u t a n t a m transcript accumulates as two fragments hybridising to probes for sequences respectively upstream and downstream of the 5' e~d of the intron. Two processed transcripts approximating to normal full length m R N A were identified. In one the second intron was intact; in the other the second intron was spliced out through the use of an A A G sequence, 20 base-pairs into the third exon, as a 3' acceptor site. A G A G sequence, only four base-pairs downstream from the normal acceptor site, does not appear to be used.
Key words:
am s
Neurospora
crassa -
Intron - Alternative
splicing
Materials and methods Neurospora mutant. The mutant, labelled in our stock collection as am4-6-2, was one of those used for fine-structure mapping of the am
(NADP-specific glutamate dehydrogenase) gene (Fincham 1967). It was supposed to have been derived, through six successive backcrosses to a standard wild-type, from the am 4 mutant, reported by Roberts (1971) to produce enzymically inactive GDH protein, identified as serologically cross-reacting material (CRM). Recent unpublished results in our laboratories point to a mistake in strain designations. The CRM-positive status of the original am 4 mutant has been confirmed and its mutational change defined as GGC to GAC (glycine to aspartate) in codon 112; an identical change has been found in two CRM-positive backcrossed strains supposedly derived from am s, a CRM-negative mutant induced by ultraviolet light. At the same time, "am4-6-2'' turns out to be CRM-negative. The original am s has been lost, but we assume that the am'* and am 8 labels were switched at some stage of backcrossing and that "am4-6-2'' is a am ~ derivative (the only one extant). In this paper the " a m 4 - 6 - 2 '' strain is simply referred to as am s. Cloning and sequencing procedures were similar to those of Kinnaird and Fincham (1983). RNA was prepared by the procedure of SokoloWski et al. (1990) and freed of DNA by treatment with RNase-free DNase. Polyadenylated RNA was prepared from total DNA-free RNA by use of an oligo-dT cellulose column (ManiatiS et al. 1982). Polymerase chain reactions (PCRs) from either genomic DNA or cDNA followed the protocols of Sambrook et al. (1989). Molecular methods.
Introduction Nuclear introns o f fungi, like those o f other eukaryotes, invariably have the dinucleotide sequence A G at their 3' termini, almost always preceded by a pyrimidine residue~ There seems, however, to be no general and absolute requirement for a pyrimidine in this position. In S a c c h a r o m y c e s c e r e v i s i a e the general rule is that 3' splicing takes place at the first A G sequence downstream f r o m the "lariat" branch point ( R u b y and Abelson 1991). In N e u r o s p o r a c r a s s a , Kuiper et al. (1988) showed that, in a m u t a n t with T A G changed to TAA at the 3' end of an intron, default splicing occurred efficiently at a 3' T A G sequence five bases further downstream. To our knowledge there has been, no test in N e u r o s p o r a of the use of a n A A G or G A G sequence in default of PyAG.
Offprint requests to:
J. R. S. Fincham
Results and discussion Sequencing of the a m 8 allele revealed a double base-pair mutation at the 3' end o f the second intron. The wild-type TAG 3'-splice Sequence is changed to G T G (Fig. 1). Northern blots of total wild-type R N A , probed with the 2.64 kb B a m H 1 fragment covering the entire a m gene (Kinnaird and Fincham 1983), showed a single band of m R N A , k n o w n to be approximately 1.7 kb. A similar analysis of a m s reveals a fainter band of approximately the same size, together with two smaller and more abundant fragments. To test the hypothesis that these smaller fragments corresponded to unligated splicing intermediates, replicate strips of Northern blots were hybridised to
38
1. Sequences of wild-type and of the am 8 allele in the region of the 3' junction of the second intron. The sequences shown are those of the antisense DNA strands, and correspond to CTCTAGTG and CTCGTGTG in the sense-strands of wild-type and mutant respectively.
Fig.
, 200b ,
Bglll
3'
5'
ANNI
m
2
B [
4
I
T
(
Fig. 2A. Structure of the am primary transcript, with introns black and 5' and 3' untranslated sequences shaded; the arrow indicates the position of the unique BglII restriction site. B the 5' 40% of the transcript expanded, with thefilled triangle showing the position of the am 8 mutation, and the horizontal arrows showing the positions and orientations of the primers used for PCR on eDNA. Primer 1 was a 20-mer corresponding to bases 244-253 of the B a m H l fragment encompassing the am gene (Kinnaird and Fincham 1983); primer 2 was a 24-met corresponding to bases 579-602; primer 3 was a 24-mer corresponding to the complement of bases 949-972. The black triangle indicates the position of the am 8 mutation Key to probes:
separated fragments of the a m probe obtained by cutting at the unique B g l I I site located 21 nucleotides upstream of the second intron (Fig. 2). The smaller of the two transcript fragments hybridised exclusively to the upstream probe (Fig. 3A, track 2; Fig. 3 B, track 1), the larger fragment hybridised predominantly to the downstream probe (Fig. 3 A, track 4; Fig. 3 B, track 2), and the approximately normal-sized a m a transcripts hybridised strongly to both probes. The upstream transcript fragment, presumably lacking a polyadenylation signal, was not detected in R N A fractionated by binding to oligo-dT cellulose (Fig. 3 B, track 4). In order to determine the composition of the approximately 1.7 kb polyadenylated a m 8 R N A , the polyadenylated R N A fraction was reverse-transcribed using an antisense primer (primer 3) complementary to sequence downstream from the second intron, and products spanning the second intron were amplified as double-stranded
U = Upstream D = Downstream
Fig. 3A, B. Lanes from two Northern blots (A, B) of RNA from
wild-type and am 8 RNA probed in sections with the total am gene (2.64 kb BamH1 fragment) or with upstream or downstream BglII subfragments (see Fig. 2), A I , wild-type and A2, am 8 total RNA, both probed with the upstream fragment; A3, wild-type and A4, am 8 total RNA, both probed with the downstream fragment; B1, B2 and B3 all am a total RNA, probed respectively with upstream, downstream and total am fragments; B4 polyadenylated am 8 RNA probed with the total am fragment. Different am 8 RNA preparations were used for the A and B blots. The amounts of RNA loaded were approximately 1.5 gg (A1, A3), 3.5 p,g (A2, A4), 2.5 gg ( B I - 3 ) and 0.5 btg (B4)
D N A by P C R after addition of a second primer corresponding to sequence either upstream of the first intron (primer 1) or between the two introns (primer 2 - see Fig. 2). Analysis of the reaction products on 0.75% agarose revealed in each case, a strong, apparently dou-
39 Ke y: C = cDNA template G = Genomic DNA template M = C + G products mixed
Fig. 4. Size analysis of PCR products on a 4% polyacrylamide electrophoresis gel. 1, primer 1 + 3 product from genomic DNA including both introns - predicted size 731 bp; 2, primer 1 +3 products from a m 8 cDNA; 3, mixture of 1 and 2; 4, primer 2 + 3 products from a m 8 cDNA; 5 a n d 6, primer 2 + 3 product from genomic DNA - predicted size 394 bp. The flanking tracks are size markers; sizes are indicated in bp
OM ...
CCTACTGGCC~GTAAT
G G..4 6b .,GTCTCT 7~GIT G A G C A T G G G T G G T G G C A A G G G T G G T . . .
[B ) .., C C T A C T G G C C G T A A G T G , . 4 6 b . . G T C T C G T G T G A G C A T G G G T G G T G G C A A G G G T G G T . . .
jJ
(C)
.
.
.
.
.
.
Fig. 5. A shows the wild-type DNA sequences spanning the ends of the a m second intron, with the normal intron termini shown by arrows. B, C are the sequences of cloned cDNA PCR products, generated by primers 2+3 from a m 8 cDNA. Of four clones sequences, three had sequence B and one sequence C
ble, D N A band in the ranges 650 700 bp and 3 0 0 400 bp for the 1 + 3 and 2 + 3 primer combinations respectively. The doublets were each resolved in 4% polyacrylamide gels into two discrete bands of approximate sizes 320 bp and 400 bp for primers 2 + 3, and 590 bp and 670 bp for primers 1 + 3 (Fig. 4). The more abundant, approximately 400 bp, c D N A fragment was indistinguishable in size from the PCR fragment (of predicted size 394 bp) obtained with the same primers from genomic DNA, and seemed likely to be due to failure to splice out the second intron. The estimated size o f the less abundant fragment, 320 bp, was slightly smaller than expected from normal splicing of the second intron (expected size 333 bp). O f the two primer 1 + 3 products, the larger was clearly smaller than the fragment generated by PCR with primers 1 + 3 from genomic D N A and was consistent in
size with removal of the first intron only. The size of the smaller product was consistent with removal of both introns. The polyacrylamide gel also revealed the presence of other much larger and much smaller PCR products in lower yield (Fig. 4); these are presumed to be PCR artefacts and have not been further analysed. For sequence analysis, an additional supply of the approximately 320+400 bp c D N A fragment mixture was generated from the primer 1 + 3 PCR products by a second PCR using primers 2 + 3. The two main products appeared identical in size (data not shown) to those obtained with primers 2 + 3 directly from the c D N A (Fig. 4, track 4). Several fragments were recovered as clones in the 'Bluescript' ptasmid and sequenced, with the results summarized in Fig. 5. The larger product retains the second intron sequence in its entirety. The sequence of the smaller product shows the second exon joined to the 5' end of a glycine codon 20 nucleotides into the third exon, indicating that an A A G lysine codon is being used as an alternative 3' splice site. This mode of splicing will generate a + 1 frameshift in the coding sequence of the third exon, and is expected to yield a 317 bp PCR product with primers 2 + 3, in good agreement with observation. If the G A G glutamate codon situated only four bases downstream from the normal intron boundary were also used as a surrogate splice site one would expect a 333 bp P C R product, which should have been clearly separable from 317 bp on 4% polyacrylamide. No hint of doubleness in the smaller PCR product has been observed in several different polyacrylamide electrophoresis gels, and we therefore conclude that the G A G sequence is recognized as a 3' splice signal in N e u r o s p o r a only with very
40 low efficiency, if at all. Our results do show that the Neurospora spliceosome recognizes A A G to an appreciable extent. It may be that, in our example, the A A G sequence would have been used with greater efficiency had its distance (33 bases) from the putative intron branch-point C T G A C approximated more closely to the 1 0 - 2 0 bp typical of Neurospora introns. It is worth noting that at least one example of a naturally occurring A A G 3' intron splice signal has been reported in a filamentous fungus - in intron 6 of the facA (acetyl-coenzyme A synthetase) gene of Aspergillus nidulans (Connerton et al. 1990). In comparable studies on Saccharomyces (Fouser and Friesen 1987), the lack of an acceptable 3' acceptor site was not found to inhibit cleavage at the 5' end of the intron. Our results, on the other hand, show a degree o f dependence of 5' cleavage on the 3' splice sequence, since in a substantial proportion (perhaps about 20 per cent) of am 8 transcripts the 5' cut is not made and the intron survives intact. This suggests that fully efficient 5' cleavage depends on the entire spliceosome complex and not just on the consensus 5' slice signal.
Acknowledgements, We acknowledge helpful discussion with Dr. J. Umen. References Connerton IE Fincham JRS, Sandeman RA, Hynes M (1990) Mol Microbiol 4:451 460 Fincham JRS (1967) Genet Res 9:49-62 Fouser LA, Friesen JD (1987) Mol Cell Biol 7:225-230 Kinnaird JH, Fincham JRS (1983) Gene 26:253-260 Kuiper MTR, Holtrop M, Vennema H, Lambowitz AM, de Vries H (1988) J Biol Chem 263:2848-2852 Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning: laboratory manual, 1st edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Roberts DB (1971) J Gen Microbiol 69:143-144 Ruby SW, Abelson J (1991) Trends Genet 7:79-85 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual. 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Sokolowski V, Kaldenhoff R, Ricci M, Russo WEA (1990) Fungal Genetics. Newslett 37:41-43 Communicated by B. S. Cox
