DEVELOPMENTAL
BIOLOGY
14!$197-205
(1992)
Developmental Regulation of Alternative Splicing in the mRNA Encoding Xenopus laevis Neural Cell Adhesion Molecule (NCAM) AARON M. ZORN AND PAUL A. KRIEG Center
for
Developmental
Biology,
Department Accepted
of Zoology,
University
September
of Texas at Austin,
Austin,
Texas
78712
19, 1991
The neural cell adhesion molecule (NCAM) is thought to play a role in the formation of the vertebrate nervous system. In mammals and chicken, it is known that more than 100 different forms of the NCAM protein can be generated by alternative splicing of one primary transcript and it is possible that these different forms have distinct biological functions. A large part of the diversity is generated by alternative mRNA splicing in two regions, called the ?r and the muscle specific domain (MSD), that encode portions of the extracellular domain of the NCAM protein. In this report, we describe the tissue and developmental expression of the ?r and MSD sequences in the amphibian, Xenopu.s Lewis Our experiments show that NCAM transcripts are present in all tissues examined including muscle, heart, liver, kidney, and brain. We have identified a 30-base exon, similar to the r domain observed in mammals, that is not present in maternal NCAM RNA but appears in a subset of the NCAM mRNA population shortly after neural induction. At the predicted location of the MSD we have detected only two alternatively spliced exons, 3 bases and 15 bases in length. In no X laevis tissue examined did we detect the two additional alternatively spliced exons which are present in the MSD region of mammalian and chicken NCAM RNAs. Finally, the analysis has revealed a dynamic and complex pattern of expression of alternatively spliced NCAM mRNAs during embryogenesis. High levels of expression of specific forms of NCAM RNA correlate with major morphogenic events such as neural tube formation and metamorphosis. o 1992 Academic PWSS, IDC.
tains no transmembrane or membrane attachment sequences (Gower et aZ., 1988). These four forms of the NCAM protein are encoded by a single gene and generated by alternative splicing and the use of alternative polyadenylation signals (for review see Walsh and Dickson, 1989). The relative amounts of the different forms of NCAM show variation during development and from tissue to tissue, raising the possibility that each protein form has a distinct biological function. NCAM shows one of the most extensive patterns of alternative splicing of any gene yet characterized. Together with the major RNA forms, small alternatively spliced exons in the region encoding the extracellular domain of NCAM protein permit the synthesis of more than 100 different mRNAs from a single primary transcript. A 30-base sequence designated the a (Santoni et al, 1989) or VASE domain (Small et al, 1988) is present in rat and mouse in a subset of NCAM RNAs from the brain and heart (Small and Akeson, 1990). So far, r domain sequences have not been detected in chicken (Small et aL, 1988). The second site of alternative exon usage, designated as the muscle specific domain (MSD), alters the sequence of NCAM protein in the membrane proximal region (Dickson et aL, 1987). The MSD in mammals is encoded by four small exons 15, 48, 42, and 3 bases in length (Santoni et al, 1989; Thompson et al, 1989; Reyes et uL, 1991), while chicken has exons 15,33, 42, and 3 bases in length (Owens et aL, 1987; Prediger et aL, 1988). It appears that all combinations of these minor exons may be used in different NCAM mRNAs
INTRODUCTION
Neural cell adhesion molecule (NCAM) is a cell surface glycoprotein that mediates cell-cell interactions through homophilic binding. NCAM is expressed predominantly in neural tissue, but lower levels of NCAM are also present in a variety of embryonic and adult tissues (Edelman, 1986). NCAM plays a role in the cell interaction events that occur during development of the vertebrate nervous system. In particular it has been implicated in axonal migration, myoblast fusion, and the formation of neuromuscular junctions. NCAM may also play a part in the early cell segregation events that occur during the specification of presumptive neural tissue (for review see Edelman, 1988; Rutishauser and Jessell, 1988). NCAM proteins exist in four major forms, designated 180, 140,120, and 105 kDa, according to their apparent molecular weight when analyzed on SDS-acrylamide gels (Cunningham et aL, 1987; Gower et cd., 1988). The differences between these proteins reside in the cytoplasmic and transmembrane domains, while all forms share a common extracellular domain. The 180- and 140kDa forms are transmembrane proteins possessing a large cytoplasmic domain or a small cytoplasmic domain, respectively (Gennarini et aL, 1984). The 120-kDa form lacks transmembrane and cytoplasmic domains and is attached to the outer side of the membrane by a phosphatidylinositol linkage (He et aZ., 1986; Hemperly et aL, 1986; He et aL, 198’7). Finally, the 105kDa molecule appears to be a secreted form of NCAM because it con197
0012-1606192 $3.00 Copyright All rights
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
198
DEVELOPMENTAL
BIOLOGY
(Santoni et al., 1989; Small and Akeson, 1990; Reyes et aZ., 1991). At present, it is not known if the different forms of NCAM protein resulting from these alternatively spliced transcripts have distinct functions. It is of interest, therefore, to determine if any of the NCAM molecules are restricted to particular tissues or to particular stages of development, in the hope that this information will provide insights into the function of the different proteins. While RNA blot analysis may be used to study the expression of mRNAs encoding the major size classes of NCAM proteins, the resolution and sensitivity of blot analysis are not sufficient to reveal the patterns of expression of NCAM mRNAs that contain the small alternatively spliced exons. In this report we describe the use of the polymerase chain reaction (PCR) to investigate the expression of 7~and MSD sequences in Xenopus. Our first objective was to determine if the alternatively spliced exons were present in frog NCAM mRNAs and, if so, to determine the extent to which the sequences encoded by the exons had been evolutionarily conserved. Second, we wished to characterize the use of the T and MSD exons during formation of the embryonic nervous system. To achieve these goals we have analyzed NCAM RNA populations from adult Xenow tissues and throughout embryonic development. Our experiments have identified a 30-base exon similar to the 7r domain exon present in mammals. Transcripts containing the a domain are not present in the maternal NCAM mRNA, but appear as a subset of the total mRNA population during neurogenesis. At the predicted location of the MSD we detect two alternatively spliced exons 3 and 15 bases in length. Within the limits of sensitivity of PCR analysis, Xenopus NCAM RNA does not contain the two additional alternatively spliced MSD exons present in mammals and chickens. Analysis of NCAM transcripts expressed during Xenopus embryogenesis reveals that a dynamic and complex pattern of alternative splicing occurs during formation of the embryonic nervous system. The high levels of expression of specific alternatively spliced forms of NCAM RNA correlate with major morphogenic events such as neural tube formation and metamorphosis. MATERIALS
AND
METHODS
Animals X laevis adults were purchased from Nasco or Xenopus I. Females were induced to ovulate by injection of 1000 units of human chorionic gonadotropin (Sigma) and eggs were fertilized in vitro with minced testis. Embryos were maintained in 20% Steinberg’s solution and staged according to Nieuwkoop and Faber (1967). RNA Extractions and cDNA Synthesis Total RNA was isolated from eggs and embryos as described by Melton and Cortese (1979) and purified fur-
VOLUME
149,1992
ther by precipitation with 4 MLiCl. RNA was extracted from adult Xenopus tissues using a variation of the LiCl/urea procedure. First strand cDNA was synthesized from 1 pg of total RNA from adult tissues or one embryo equivalent of total RNA (about 3 pg), as previously described (Maniatis et al, 1982). Priming of NCAM-specific cDNA synthesis utilized oligonucleotide E13a (5’-CCTGCTTGAT/,TATGTCCACTTTTAT-3’), which is complementary to NCAM nucleotides 18551880, within exon 13 (Krieg et al, 1989). Identical exon 13 sequences are shared by all NCAM transcripts except the mRNA encoding the secreted 105-kDa form of the protein. This transcript contains an alternative exon 13 that encodes the carboxyl terminal of the secreted protein. The absence of 105-kDa mRNA sequences from the cDNA sample is not a significant obstacle to these studies because this form of NCAM mRNA is undetectable in frog embryos during embryogenesis (Krieg et al., 1989). PCR Ampli&ations In all experiments, oligonucleotide primers were designed to be complementary to sequences of both X laevis NCAM genes (Krieg et al, 1989; Tonissen and Krieg, unpublished data). Figure 1 shows the position of the oligonucleotides used for the PCR amplifications of the T and MSD regions of the NCAM cDNA. Oligonucleotides E7 (5’-CCTAGTATCACCTGGAGAA-3’) and E8 (5’-TTCAGAGTAAGAGCTGA-3’) flank the exon 7/8 junction. These oligos are complementary to NCAM cDNA at nucleotide positions 1013-1034 and 1108-1124, respectively. Oligonucleotides flanking the exon 12/13 junction, El2 (5’-GAGACAGCACACCTTCACAG-3’) and E13b (5’-ATCTTCACTCAAATGCCCTA-3’), are complementary to nucleotide positions 1760-1780 and 1826-1846, respectively (Krieg et ak, 1989). PCR amplification mixes contained 10% of the cDNA reaction mix plus 300 ng of each primer, 200 PM dNTPs, 50 mM TrisHCl (pH 9.0), 1.5 mlM MgC&, 0.1% gelatin, 1% Triton X-100, and 2.5 units of Taq DNA polymerase (Promega) in a loo-111 volume. Reactions were taken through 25 or 30 amplification cycles of 1 min at 94”C, 1.5 min at 5O”C, and 2 min at 72°C. Amplification through 25 cycles was in the linear range for quantitating PCR products of this size and abundance. This was demonstrated by amplifying samples containing different quantities of NCAM template and showing that the amount of product generated was proportional to the quantity of input template (data not shown). In some instances 30 amplification cycles were required to detect rare sequences. It should be noted that products of this size and abundance plateau after 30 cycles and in these cases quantitation between different samples is not accurate (data not shown).
ZORN AND KRIEG
El2
E13b
Alternative
E13a
x domain
FIG. 1. PCR analysis of NCAM alternative splicing. The solid line represents the extracellular domain of NCAM with the 5 Ig loops. The double vertical lines represents the cell membrane. The predicted location of the ?r and MSD splice sites, at the exon ‘7/8 and exon 12/13 junctions, respectively, are indicated. The arrows show the location of oligonucleotides used to prime cDNA synthesis and PCR amplification. Oligo E13a is complementary to exon 13 sequences in NCAM RNA and was used to prime the synthesis of NCAM-specific cDNA. This cDNA was subsequently used as a template for PCR amplifications. Oligos E7 and ES are complementary to sequences in exons 7 and 8, respectively. These were used to PCR amplify the region spanning the exon 7/8 junction, which contains the predicted location of the * domain splice site. Similarly, oligos El2 and E13b were used to amplify the region spanning the exon 12/13 junction which is the location of the MSD splice site.
Analysis
of PCR Products
PCR products were 32P end-labeled with polynucleotide kinase and fractionated on 8 or 12% acrylamide gels containing 8.3 M urea and the position of the labeled bands was detected by autoradiography. Alternatively, unlabeled PCR products were fractionated on denaturing acrylamide gels, electroblotted to nylon membranes, and then probed with “P-labeled NCAM oligonucleotide E13c (5’-TTTTGGAGCACTAGGTTCCC-3’). Denaturing gels were used for all analyses in order to avoid artifacts generated by heteroduplex formation (Zorn and Krieg, 1991). RESULTS
Tissue and Developmental Sequences
Expression
of ?r Domain
Until recently a comprehensive investigation of alternative splicing in NCAM RNA was difficult, because analysis required the isolation of individual clones from cDNA libraries prepared from different tissues and from embryos at different stages of development. If a particular alternatively spliced product is rare in the mRNA population, a very large number of cDNA clones must be examined before the alternatively spliced product can be identified. The sensitivity of PCR theoretically permits the detection of all alternatively spliced products in any given mRNA population. Novel splicing forms identified by PCR may then be cloned and analyzed further. The general approach used to investigate alternative splicing in NCAM mRNA is presented in Fig. 1. Synthesis of cDNA from different adult tissues or embryonic
splicing of NCAM mRNA
199
stages was primed using an NCAM-specific oligonucleotide complementary to sequences located in exon 13. The cDNA preparation was used as a substrate for all subsequent PCR experiments. An internal primer (rather than oligo(dT)) was used to prime cDNA synthesis because some forms of NCAM mRNA in Xenopus are up to 9.4 kb long (Kintner and Melton, 1987). Use of an oligo(dT) primer could result in a significant underrepresentation of extracellular domain sequences derived from the longest mRNA molecules due to incomplete first-strand synthesis by reverse transcriptase. Since the genome of X laevis is pseudotetraploid (Kobel and DuPasquier, 1986), it contains two copies of the NCAM gene (Krieg et ak, 1989). The oligonucleotide primers used in PCR amplifications were designed to be complementary to both gene sequences in the regions flanking the predicted exon 7/B and exon 12/13 junctions, i.e., flanking the ?r and MSD alternative splicing regions, respectively. Alternative splicing of NCAM mRNA between exons 7 and 8 has been observed in rat (Small and Akeson, 1990) and mouse (Santoni et aL, 1989). Our first experiments were designed to determine if Xenopus NCAM mRNA exhibits alternative splicing at the predicted location of the r domain. Since NCAM is expressed at the highest levels in neural tissues, cDNA derived from adult brain RNA was used as the substrate for the initial PCR amplifications. In Fig. 2A, lane B shows the presence of two fragments with lengths of 112 and 142 bases. These sizes are consistent with the predictions for NCAM mRNA containing no insertion at the exon ‘7/8 junction and a 30-base insertion, respectively. PCR products from adult brain that contained the 30-base ?r domain insertion were cloned and sequenced. Sequences of 15 independent clones containing the r domain indicated that both Xenopus NCAM genes were represented and that the r domain sequences of the two genes differed from each other by only one nucleotide. As shown in Fig. 2B, the nucleotide sequence of the Xenopus ?r domain is very similar (83% identical) to the mammalian sequence (Small et al, 1988; Santoni et a& 1989). The r domain exon encodes 10 amino acids. The deduced amino acid sequence of the r domain of Xenopus NCAM gene 2 is identical at 9/10 positions with the mammalian sequence, while the Xenopus NCAM gene 1 is identical at B/10 positions. Further PCR experiments (Fig. 2A) show that all adult tissues examined, heart, muscle, kidney, and liver, also express NCAM RNA containing the ?r domain, although the relative amount of this RNA varies greatly between tissues. For example, about 1% of total NCAM mRNA in liver contains the 7r domain, while in kidney the proportion is nearly 50%. Having demonstrated the presence of alternative splicing at the exon 7/B junction in adult tissues we ex-
DEVELOPMENTAL BIOLOGY
200
VOLUME 149,1992
B
2
A
BHSK
Xenopus
L;M m
30 base 7r mseriion
-190
-
-
Rat 147
,,1. )’
t,s &,‘a
30 base x insertion
-
No insertion
-.
C
.:,I)
0
“.
-
190
-
147
-
110
‘
BIOLOGY
14!$197-205
(1992)
Developmental Regulation of Alternative Splicing in the mRNA Encoding Xenopus laevis Neural Cell Adhesion Molecule (NCAM) AARON M. ZORN AND PAUL A. KRIEG Center
for
Developmental
Biology,
Department Accepted
of Zoology,
University
September
of Texas at Austin,
Austin,
Texas
78712
19, 1991
The neural cell adhesion molecule (NCAM) is thought to play a role in the formation of the vertebrate nervous system. In mammals and chicken, it is known that more than 100 different forms of the NCAM protein can be generated by alternative splicing of one primary transcript and it is possible that these different forms have distinct biological functions. A large part of the diversity is generated by alternative mRNA splicing in two regions, called the ?r and the muscle specific domain (MSD), that encode portions of the extracellular domain of the NCAM protein. In this report, we describe the tissue and developmental expression of the ?r and MSD sequences in the amphibian, Xenopu.s Lewis Our experiments show that NCAM transcripts are present in all tissues examined including muscle, heart, liver, kidney, and brain. We have identified a 30-base exon, similar to the r domain observed in mammals, that is not present in maternal NCAM RNA but appears in a subset of the NCAM mRNA population shortly after neural induction. At the predicted location of the MSD we have detected only two alternatively spliced exons, 3 bases and 15 bases in length. In no X laevis tissue examined did we detect the two additional alternatively spliced exons which are present in the MSD region of mammalian and chicken NCAM RNAs. Finally, the analysis has revealed a dynamic and complex pattern of expression of alternatively spliced NCAM mRNAs during embryogenesis. High levels of expression of specific forms of NCAM RNA correlate with major morphogenic events such as neural tube formation and metamorphosis. o 1992 Academic PWSS, IDC.
tains no transmembrane or membrane attachment sequences (Gower et aZ., 1988). These four forms of the NCAM protein are encoded by a single gene and generated by alternative splicing and the use of alternative polyadenylation signals (for review see Walsh and Dickson, 1989). The relative amounts of the different forms of NCAM show variation during development and from tissue to tissue, raising the possibility that each protein form has a distinct biological function. NCAM shows one of the most extensive patterns of alternative splicing of any gene yet characterized. Together with the major RNA forms, small alternatively spliced exons in the region encoding the extracellular domain of NCAM protein permit the synthesis of more than 100 different mRNAs from a single primary transcript. A 30-base sequence designated the a (Santoni et al, 1989) or VASE domain (Small et al, 1988) is present in rat and mouse in a subset of NCAM RNAs from the brain and heart (Small and Akeson, 1990). So far, r domain sequences have not been detected in chicken (Small et aL, 1988). The second site of alternative exon usage, designated as the muscle specific domain (MSD), alters the sequence of NCAM protein in the membrane proximal region (Dickson et aL, 1987). The MSD in mammals is encoded by four small exons 15, 48, 42, and 3 bases in length (Santoni et al, 1989; Thompson et al, 1989; Reyes et uL, 1991), while chicken has exons 15,33, 42, and 3 bases in length (Owens et aL, 1987; Prediger et aL, 1988). It appears that all combinations of these minor exons may be used in different NCAM mRNAs
INTRODUCTION
Neural cell adhesion molecule (NCAM) is a cell surface glycoprotein that mediates cell-cell interactions through homophilic binding. NCAM is expressed predominantly in neural tissue, but lower levels of NCAM are also present in a variety of embryonic and adult tissues (Edelman, 1986). NCAM plays a role in the cell interaction events that occur during development of the vertebrate nervous system. In particular it has been implicated in axonal migration, myoblast fusion, and the formation of neuromuscular junctions. NCAM may also play a part in the early cell segregation events that occur during the specification of presumptive neural tissue (for review see Edelman, 1988; Rutishauser and Jessell, 1988). NCAM proteins exist in four major forms, designated 180, 140,120, and 105 kDa, according to their apparent molecular weight when analyzed on SDS-acrylamide gels (Cunningham et aL, 1987; Gower et cd., 1988). The differences between these proteins reside in the cytoplasmic and transmembrane domains, while all forms share a common extracellular domain. The 180- and 140kDa forms are transmembrane proteins possessing a large cytoplasmic domain or a small cytoplasmic domain, respectively (Gennarini et aL, 1984). The 120-kDa form lacks transmembrane and cytoplasmic domains and is attached to the outer side of the membrane by a phosphatidylinositol linkage (He et aZ., 1986; Hemperly et aL, 1986; He et aL, 198’7). Finally, the 105kDa molecule appears to be a secreted form of NCAM because it con197
0012-1606192 $3.00 Copyright All rights
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
198
DEVELOPMENTAL
BIOLOGY
(Santoni et al., 1989; Small and Akeson, 1990; Reyes et aZ., 1991). At present, it is not known if the different forms of NCAM protein resulting from these alternatively spliced transcripts have distinct functions. It is of interest, therefore, to determine if any of the NCAM molecules are restricted to particular tissues or to particular stages of development, in the hope that this information will provide insights into the function of the different proteins. While RNA blot analysis may be used to study the expression of mRNAs encoding the major size classes of NCAM proteins, the resolution and sensitivity of blot analysis are not sufficient to reveal the patterns of expression of NCAM mRNAs that contain the small alternatively spliced exons. In this report we describe the use of the polymerase chain reaction (PCR) to investigate the expression of 7~and MSD sequences in Xenopus. Our first objective was to determine if the alternatively spliced exons were present in frog NCAM mRNAs and, if so, to determine the extent to which the sequences encoded by the exons had been evolutionarily conserved. Second, we wished to characterize the use of the T and MSD exons during formation of the embryonic nervous system. To achieve these goals we have analyzed NCAM RNA populations from adult Xenow tissues and throughout embryonic development. Our experiments have identified a 30-base exon similar to the 7r domain exon present in mammals. Transcripts containing the a domain are not present in the maternal NCAM mRNA, but appear as a subset of the total mRNA population during neurogenesis. At the predicted location of the MSD we detect two alternatively spliced exons 3 and 15 bases in length. Within the limits of sensitivity of PCR analysis, Xenopus NCAM RNA does not contain the two additional alternatively spliced MSD exons present in mammals and chickens. Analysis of NCAM transcripts expressed during Xenopus embryogenesis reveals that a dynamic and complex pattern of alternative splicing occurs during formation of the embryonic nervous system. The high levels of expression of specific alternatively spliced forms of NCAM RNA correlate with major morphogenic events such as neural tube formation and metamorphosis. MATERIALS
AND
METHODS
Animals X laevis adults were purchased from Nasco or Xenopus I. Females were induced to ovulate by injection of 1000 units of human chorionic gonadotropin (Sigma) and eggs were fertilized in vitro with minced testis. Embryos were maintained in 20% Steinberg’s solution and staged according to Nieuwkoop and Faber (1967). RNA Extractions and cDNA Synthesis Total RNA was isolated from eggs and embryos as described by Melton and Cortese (1979) and purified fur-
VOLUME
149,1992
ther by precipitation with 4 MLiCl. RNA was extracted from adult Xenopus tissues using a variation of the LiCl/urea procedure. First strand cDNA was synthesized from 1 pg of total RNA from adult tissues or one embryo equivalent of total RNA (about 3 pg), as previously described (Maniatis et al, 1982). Priming of NCAM-specific cDNA synthesis utilized oligonucleotide E13a (5’-CCTGCTTGAT/,TATGTCCACTTTTAT-3’), which is complementary to NCAM nucleotides 18551880, within exon 13 (Krieg et al, 1989). Identical exon 13 sequences are shared by all NCAM transcripts except the mRNA encoding the secreted 105-kDa form of the protein. This transcript contains an alternative exon 13 that encodes the carboxyl terminal of the secreted protein. The absence of 105-kDa mRNA sequences from the cDNA sample is not a significant obstacle to these studies because this form of NCAM mRNA is undetectable in frog embryos during embryogenesis (Krieg et al., 1989). PCR Ampli&ations In all experiments, oligonucleotide primers were designed to be complementary to sequences of both X laevis NCAM genes (Krieg et al, 1989; Tonissen and Krieg, unpublished data). Figure 1 shows the position of the oligonucleotides used for the PCR amplifications of the T and MSD regions of the NCAM cDNA. Oligonucleotides E7 (5’-CCTAGTATCACCTGGAGAA-3’) and E8 (5’-TTCAGAGTAAGAGCTGA-3’) flank the exon 7/8 junction. These oligos are complementary to NCAM cDNA at nucleotide positions 1013-1034 and 1108-1124, respectively. Oligonucleotides flanking the exon 12/13 junction, El2 (5’-GAGACAGCACACCTTCACAG-3’) and E13b (5’-ATCTTCACTCAAATGCCCTA-3’), are complementary to nucleotide positions 1760-1780 and 1826-1846, respectively (Krieg et ak, 1989). PCR amplification mixes contained 10% of the cDNA reaction mix plus 300 ng of each primer, 200 PM dNTPs, 50 mM TrisHCl (pH 9.0), 1.5 mlM MgC&, 0.1% gelatin, 1% Triton X-100, and 2.5 units of Taq DNA polymerase (Promega) in a loo-111 volume. Reactions were taken through 25 or 30 amplification cycles of 1 min at 94”C, 1.5 min at 5O”C, and 2 min at 72°C. Amplification through 25 cycles was in the linear range for quantitating PCR products of this size and abundance. This was demonstrated by amplifying samples containing different quantities of NCAM template and showing that the amount of product generated was proportional to the quantity of input template (data not shown). In some instances 30 amplification cycles were required to detect rare sequences. It should be noted that products of this size and abundance plateau after 30 cycles and in these cases quantitation between different samples is not accurate (data not shown).
ZORN AND KRIEG
El2
E13b
Alternative
E13a
x domain
FIG. 1. PCR analysis of NCAM alternative splicing. The solid line represents the extracellular domain of NCAM with the 5 Ig loops. The double vertical lines represents the cell membrane. The predicted location of the ?r and MSD splice sites, at the exon ‘7/8 and exon 12/13 junctions, respectively, are indicated. The arrows show the location of oligonucleotides used to prime cDNA synthesis and PCR amplification. Oligo E13a is complementary to exon 13 sequences in NCAM RNA and was used to prime the synthesis of NCAM-specific cDNA. This cDNA was subsequently used as a template for PCR amplifications. Oligos E7 and ES are complementary to sequences in exons 7 and 8, respectively. These were used to PCR amplify the region spanning the exon 7/8 junction, which contains the predicted location of the * domain splice site. Similarly, oligos El2 and E13b were used to amplify the region spanning the exon 12/13 junction which is the location of the MSD splice site.
Analysis
of PCR Products
PCR products were 32P end-labeled with polynucleotide kinase and fractionated on 8 or 12% acrylamide gels containing 8.3 M urea and the position of the labeled bands was detected by autoradiography. Alternatively, unlabeled PCR products were fractionated on denaturing acrylamide gels, electroblotted to nylon membranes, and then probed with “P-labeled NCAM oligonucleotide E13c (5’-TTTTGGAGCACTAGGTTCCC-3’). Denaturing gels were used for all analyses in order to avoid artifacts generated by heteroduplex formation (Zorn and Krieg, 1991). RESULTS
Tissue and Developmental Sequences
Expression
of ?r Domain
Until recently a comprehensive investigation of alternative splicing in NCAM RNA was difficult, because analysis required the isolation of individual clones from cDNA libraries prepared from different tissues and from embryos at different stages of development. If a particular alternatively spliced product is rare in the mRNA population, a very large number of cDNA clones must be examined before the alternatively spliced product can be identified. The sensitivity of PCR theoretically permits the detection of all alternatively spliced products in any given mRNA population. Novel splicing forms identified by PCR may then be cloned and analyzed further. The general approach used to investigate alternative splicing in NCAM mRNA is presented in Fig. 1. Synthesis of cDNA from different adult tissues or embryonic
splicing of NCAM mRNA
199
stages was primed using an NCAM-specific oligonucleotide complementary to sequences located in exon 13. The cDNA preparation was used as a substrate for all subsequent PCR experiments. An internal primer (rather than oligo(dT)) was used to prime cDNA synthesis because some forms of NCAM mRNA in Xenopus are up to 9.4 kb long (Kintner and Melton, 1987). Use of an oligo(dT) primer could result in a significant underrepresentation of extracellular domain sequences derived from the longest mRNA molecules due to incomplete first-strand synthesis by reverse transcriptase. Since the genome of X laevis is pseudotetraploid (Kobel and DuPasquier, 1986), it contains two copies of the NCAM gene (Krieg et ak, 1989). The oligonucleotide primers used in PCR amplifications were designed to be complementary to both gene sequences in the regions flanking the predicted exon 7/B and exon 12/13 junctions, i.e., flanking the ?r and MSD alternative splicing regions, respectively. Alternative splicing of NCAM mRNA between exons 7 and 8 has been observed in rat (Small and Akeson, 1990) and mouse (Santoni et aL, 1989). Our first experiments were designed to determine if Xenopus NCAM mRNA exhibits alternative splicing at the predicted location of the r domain. Since NCAM is expressed at the highest levels in neural tissues, cDNA derived from adult brain RNA was used as the substrate for the initial PCR amplifications. In Fig. 2A, lane B shows the presence of two fragments with lengths of 112 and 142 bases. These sizes are consistent with the predictions for NCAM mRNA containing no insertion at the exon ‘7/8 junction and a 30-base insertion, respectively. PCR products from adult brain that contained the 30-base ?r domain insertion were cloned and sequenced. Sequences of 15 independent clones containing the r domain indicated that both Xenopus NCAM genes were represented and that the r domain sequences of the two genes differed from each other by only one nucleotide. As shown in Fig. 2B, the nucleotide sequence of the Xenopus ?r domain is very similar (83% identical) to the mammalian sequence (Small et al, 1988; Santoni et a& 1989). The r domain exon encodes 10 amino acids. The deduced amino acid sequence of the r domain of Xenopus NCAM gene 2 is identical at 9/10 positions with the mammalian sequence, while the Xenopus NCAM gene 1 is identical at B/10 positions. Further PCR experiments (Fig. 2A) show that all adult tissues examined, heart, muscle, kidney, and liver, also express NCAM RNA containing the ?r domain, although the relative amount of this RNA varies greatly between tissues. For example, about 1% of total NCAM mRNA in liver contains the 7r domain, while in kidney the proportion is nearly 50%. Having demonstrated the presence of alternative splicing at the exon 7/B junction in adult tissues we ex-
DEVELOPMENTAL BIOLOGY
200
VOLUME 149,1992
B
2
A
BHSK
Xenopus
L;M m
30 base 7r mseriion
-190
-
-
Rat 147
,,1. )’
t,s &,‘a
30 base x insertion
-
No insertion
-.
C
.:,I)
0
“.
-
190
-
147
-
110
‘
