- Email: [email protected]
Expression of novel alternatively spliced isoforms of the oct-1 transcription factor1
Biochimica et Biophysica Acta 1444 (1999) 295^298
Short sequence-paper
Expression of novel alternatively spliced isoforms of the oct-1 transcription factor1 Joseph Riss, Reuven Laskov * The Hubert Humphrey Center for Experimental Medicine and Cancer Research, The Hebrew University^Hadassah Medical School, Jerusalem 91120, Israel Received 19 October 1998; received in revised form 9 December 1998; accepted 16 December 1998
Abstract Analysis of the alternatively spliced isoforms of the human and mouse oct-1 genes, combined with their exon^intron structure, show a high level of evolutionary conservation between these two species. The differential expression of several oct1 isoforms was examined by reverse transcription^polymerase chain reaction performed on the 3P region of the murine oct-1 cDNA. Variations in the relative levels and patterns of expression of the isoforms were found among different tissues. Three novel isoforms originating from the 3P-distal region of oct-1, were isolated and sequenced: Two were derived from testis, and one from myeloma cells. Splicing out of different exons as revealed in the structure of these isoforms results in reading frameshifts that presumably lead to the expression of shortened Oct-1 proteins, with distinct C-terminal tails. Altogether, six out of the eight known murine oct-1 isoforms may have distinct C-termini, implying that these multiple tails have different functional roles in cellular differentiation and physiology. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Transcription; Di¡erential splicing; cDNA isoform; POU gene
In eukaryotes, oct-1 is a ubiquitously expressed regulatory gene of the POU homeo domain family. The Oct-1 protein binds to the octamer motif (ATGCAAAT) and related DNA sequences, that are present in the control regions of a variety of ubiquitously expressed genes, such as snRNA and histone H2B, and in tissue-speci¢c genes, such as the immunoglobulin (Ig) genes [1^3]. An additional intriguing function of Oct-1 is its ability to bind to the origin of DNA replication and to a¡ect the initiation of replication of adenoviruses [4]. Oct-1 was found to in* Corresponding author. Fax: +972-2-641-4583; E-mail: [email protected] 1 The sequence data reported in this paper have been submitted to the EMBL/GenBank under accession numbers AFO95458, AFO95459, AFO95460.
teract with a variety of tissue speci¢c transcription factors and viral proteins to form transcription and/ or DNA replication complexes [3,4]. The Oct-1 protein contains, in addition to the central POU-speci¢c and POU-homeo DNA binding domains, a variety of activation domains which include two 5P glutaminerich domains, a serine^threonine-rich domain and a hydrophobic C-terminal region [1,5]. In all cell lines and tissues tested, oct-1 is expressed as a very large ( s 10 kb) transcript [6,7], which partially overlaps (in its 3P region) with the antisense CD3 gene locus [6]. Multiple alternatively spliced isoforms were identi¢ed in human and murine cells (Fig. 1). The human and murine oct-1 cDNAs were found to be very homologous at their nucleotide and deduced amino acid sequences [1,6^9]. Isolation of the
0167-4781 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 9 8 ) 0 0 2 8 0 - 2
BBAEXP 91240 9-2-99
296
J. Riss, R. Laskov / Biochimica et Biophysica Acta 1444 (1999) 295^298
Fig. 1. Exon composition of the oct-1 gene and its alternatively spliced isoforms. The upper scheme shows the exon structure of oct-1 as deduced from published data of the human [10] and murine genomic clones [6], and from the sequences of various cDNA clones [1,7^9,11]. The deduced exon structure of the various isoforms is shown below (isoforms 1^10). The dashed-line rectangles represent the POU-speci¢c and POU-homeo domains. Solid-line rectangles represent exons with lengths in nucleotide bases. Spliced exons are represented by solid or dashed oblique lines for proven or presumed exons, respectively. Straight solid lines represent cDNA sequences with no corresponding genomic clones that prove their exon^intron structure. Isoforms 7^9 were isolated by RT^PCR of the 3P region of oct-1. Their upstream sequences are therefore not known and are represented by straight dotted lines. Isoform 10 starts with a block of 204 bp of unknown origin and terminates in a stop codon (asterisk) in the intronic sequence between exons 16 and 17 [6]. The cellular origin of the human and murine cDNA clones are given in parentheses: Ntera2D1 and F9, embryonal carcinomas; 70Z/ 3, Pre-B cell line; 4T00.1 and NS/0, myelomas.
human oct-1 gene and determination of its exon^intron boundaries [10] allows the delineation of the exon composition of two alternatively spliced isoforms [11] and comparison with the exon organization of the 3P region of the murine oct-1 gene [6]. A very high level of conservation was found at the 3P region, beyond the most conserved POU speci¢c and POU homeodomains. Thus, the murine exons V, W, X and Y [6] are very similar in sequences and identical in size to the corresponding human exons 13,
14, 16 and 17 (Fig. 1). The 3P-ultimate human exon was longer than 500 nt and its 3P boundary was not de¢ned [10]. The latter is homologous to the murine exon Z [6]. Isolation of the murine isoforms 5 and 6 reveals that the Z-exon is most likely subdivided into at least four exons (exons 18^21 in Fig. 1). In the present work, we have con¢rmed the presence of exons 18^21 by isolation of three additional murine isoforms (#7^9, Fig. 1). Each of these exons is de¢ned by speci¢c deletions occurring in the di¡er-
BBAEXP 91240 9-2-99
J. Riss, R. Laskov / Biochimica et Biophysica Acta 1444 (1999) 295^298
Fig. 2. Analysis of the expression of the 3P region of the oct-1 gene (exons 17^20) in S107 myeloma and in various murine tissues. cDNA was synthesized by RT and oligo dT from total cellular RNA. PCR was performed using primers from the neighboring exons (16, 21) as depicted in the scheme (arrows). The PCR products of S107 cells were electrophoresed, stained with ethidium bromide and hybridized to an oct-1 3P probe spanning exons 13^21 (A, lanes 1 and 2, respectively). Multiple bands are apparent in the stained gel and upon hybridization. Similar analysis was performed on RNA from various tissues (B,C). The blotted ¢lter was exposed overnight (C), and for a shorter period (4.5 h), so as to demonstrate the weaker hybridization bands (B, lanes 1^3). The fragments of 680, V600 and 295 ¢t in size to isoforms 3,4, 5,7 and 6, respectively (Fig. 1). Three bands of V500, V400 and 360 bp do not correspond to any of the known isoforms. The overall level of oct-1 expression was highest in testis, and lowest in adult and newborn (NB) liver and spleen.
ent isoforms. These results indicate that the murine and human oct-1 genes are composed of at least 21 exons, including exon 15 (72 nt) detected in murine cDNA [7,9]. The occurrence of multiple oct-1 isoforms raises the possibility that they are expressed in a tissue-speci¢c manner. This is supported by the ¢ndings that testis di¡ers from other tissues in its relative level of expression of certain oct-1 isoforms [7]. In the present work, we have searched for additional evidence of `tissue speci¢city' of the oct-1 isoforms. The screening was done by reverse transcription^polymerase chain reaction (RT^PCR) per-
297
formed on the 3P region of the murine oct-1 [7]. Two oct-1 primers, sense (5P-GCCTCATGGCACCCTCA-3P) and antisense (5P-TAGAAAGTTCTCCAATCC-3P), were used to screen the expression of exons 16^21 in the 3P-region of oct-1 in di¡erent tissues. Fig. 2 shows that multiple PCR bands were obtained from RNA of S107 myeloma cells and from the tissues examined. The 680, V600 and 295 bp PCR bands conform in size with isoforms 3,4, 5,7 and 6, respectively. Bands V500, V400 and 360 bp correspond to as yet unidenti¢ed isoforms. It can be seen that di¡erent global levels of expression of the oct-1 products were obtained: the highest level was found in testis, and the lowest in liver and spleen. One or more of the V600, V500, V360 and 295 bp fragments were weakly expressed or absent in brain, liver and spleen (Fig. 2). From similar RT^PCRs performed on RNA of testis and a myeloma cell line (4T00.1), we have isolated and sequenced eight oct-1 cDNA clones, six of which constitute distinct isoforms. Three of the clones conform in size to the 680, and one to the 295 bp main PCR bands. They proved to be identical by sequence to the 3P regions of murine isoforms 3, 4 and 6, respectively. Another clone (609 bp) was identical to isoform 5 (Fig. 1). Three of the eight clones were novel (#7, 8, 9 in Fig. 1). All conform with alternative spliced products of the four consecutive 17^20 oct-1 exons. Two of the novel isoforms (#7, 9) could not have been due to PCR artifacts generated by recombination of the already known cDNA clones. Fig. 3 shows the deduced amino acid sequences of the three novel isoforms. Isoform 7 isolated from myeloma cells lacks exon 17. Its translational product exhibits a frameshift followed by an early stop codon. Thus, its predicted protein has a distinct, much shorter C-terminal tail of three amino acids (SCF). Isoforms 8 and 9 were isolated from testis and are smaller in size than the lowest PCR band of 295 bp (Fig. 2). However, in contrast to the other tissues, testis contains a heterogeneous hybridization signal in the region below 295 bp that conforms in size with these molecular species. Since isoform 8 lacks exons 18 and 20, its deduced protein is shorter due to a frameshift, and has a novel C-terminal tail of 9 ami-
BBAEXP 91240 9-2-99
298
J. Riss, R. Laskov / Biochimica et Biophysica Acta 1444 (1999) 295^298
Fig. 3. Deduced amino acid sequences of the carboxy-terminal regions of the three novel oct-1 isoforms (#7^9). The amino acid sequence of exons 16^20 of the murine isoform 3 is taken as the prototype carboxy-terminal tail (upper sequence). Each of the isoforms 7, 8 and 9 have a distinct amino acid tail due to frameshifts resulting from splicing of internal exon(s). An asterisk indicates the presence of a stop codon.
no acids. Isoform 9 lacks a very large region that includes exons 17^20 and has an insertion of G at the splice junction between exon 16 and 21. This results in a protein with a shorter C-terminal region and a novel tail of 13 amino acids. The deduced proteins of two of the previously reported oct-1 isoforms 1 and 6 (Fig. 1) may also have novel C-terminal tails due to frameshifts resulting from deletions of exons 14, 15 and 20, respectively. In addition, isoform 10 has a unique tail [8], due to the presence of a stop codon at the start of an unspliced intron [6]. The combined data show that ¢ve out of the eight known murine isoforms may have distinct amino acid tails, and suggest that structural variations at the C-termini of the Oct-1 proteins are of functional importance.
References [1] R.A. Sturm, G. Das, W. Herr, Genes Dev. 2 (1988) 1582^ 1599. [2] M. Wegner, D.W. Drolet, M.G. Rosenfeld, Curr. Opin. Cell Biol. 5 (1993) 488^498. [3] C.P. Verrijzer, P.C. Van der Vliet, Biochim. Biophys. Acta 1173 (1993) 1^21. [4] F.E.J. Coenjaerts, J.A.W.M. Van Osterhout, P.C. Vand der Vliet, EMBO J. 13 (1994) 5401^5409. [5] M. Tanaka, J.S. Lai, W. Herr, Cell 68 (1992) 755^767. [6] A. Lerner, L. D'Adamio, A.C. Diener, L.K. Clayton, E.L. Reinherz, J. Immunol. 151 (1993) 3152^3162. [7] J. Ja¡e, M. Hochberg, J. Riss, T. Hasin, L. Reich, R. Laskov, Biochim. Biophys. Acta 1261 (1995) 201^209. [8] A.G. Stepchenko, Nucleic Acids Res. 20 (1992) 1419. [9] N. Suzuki, W. Peter, T. Ciesiolka, P. Gruss, H.R. Scho«ler, Nucleic Acids Res. 21 (1993) 245^252. [10] R.A. Sturm, J.L. Cassady, G. Das, A. Romo, G.A. Evans, Genomics 16 (1993) 333^341. [11] G. Das, W. Herr, J. Biol. Chem. 268 (1993) 25026^25032.
BBAEXP 91240 9-2-99
Short sequence-paper
Expression of novel alternatively spliced isoforms of the oct-1 transcription factor1 Joseph Riss, Reuven Laskov * The Hubert Humphrey Center for Experimental Medicine and Cancer Research, The Hebrew University^Hadassah Medical School, Jerusalem 91120, Israel Received 19 October 1998; received in revised form 9 December 1998; accepted 16 December 1998
Abstract Analysis of the alternatively spliced isoforms of the human and mouse oct-1 genes, combined with their exon^intron structure, show a high level of evolutionary conservation between these two species. The differential expression of several oct1 isoforms was examined by reverse transcription^polymerase chain reaction performed on the 3P region of the murine oct-1 cDNA. Variations in the relative levels and patterns of expression of the isoforms were found among different tissues. Three novel isoforms originating from the 3P-distal region of oct-1, were isolated and sequenced: Two were derived from testis, and one from myeloma cells. Splicing out of different exons as revealed in the structure of these isoforms results in reading frameshifts that presumably lead to the expression of shortened Oct-1 proteins, with distinct C-terminal tails. Altogether, six out of the eight known murine oct-1 isoforms may have distinct C-termini, implying that these multiple tails have different functional roles in cellular differentiation and physiology. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Transcription; Di¡erential splicing; cDNA isoform; POU gene
In eukaryotes, oct-1 is a ubiquitously expressed regulatory gene of the POU homeo domain family. The Oct-1 protein binds to the octamer motif (ATGCAAAT) and related DNA sequences, that are present in the control regions of a variety of ubiquitously expressed genes, such as snRNA and histone H2B, and in tissue-speci¢c genes, such as the immunoglobulin (Ig) genes [1^3]. An additional intriguing function of Oct-1 is its ability to bind to the origin of DNA replication and to a¡ect the initiation of replication of adenoviruses [4]. Oct-1 was found to in* Corresponding author. Fax: +972-2-641-4583; E-mail: [email protected] 1 The sequence data reported in this paper have been submitted to the EMBL/GenBank under accession numbers AFO95458, AFO95459, AFO95460.
teract with a variety of tissue speci¢c transcription factors and viral proteins to form transcription and/ or DNA replication complexes [3,4]. The Oct-1 protein contains, in addition to the central POU-speci¢c and POU-homeo DNA binding domains, a variety of activation domains which include two 5P glutaminerich domains, a serine^threonine-rich domain and a hydrophobic C-terminal region [1,5]. In all cell lines and tissues tested, oct-1 is expressed as a very large ( s 10 kb) transcript [6,7], which partially overlaps (in its 3P region) with the antisense CD3 gene locus [6]. Multiple alternatively spliced isoforms were identi¢ed in human and murine cells (Fig. 1). The human and murine oct-1 cDNAs were found to be very homologous at their nucleotide and deduced amino acid sequences [1,6^9]. Isolation of the
0167-4781 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 7 8 1 ( 9 8 ) 0 0 2 8 0 - 2
BBAEXP 91240 9-2-99
296
J. Riss, R. Laskov / Biochimica et Biophysica Acta 1444 (1999) 295^298
Fig. 1. Exon composition of the oct-1 gene and its alternatively spliced isoforms. The upper scheme shows the exon structure of oct-1 as deduced from published data of the human [10] and murine genomic clones [6], and from the sequences of various cDNA clones [1,7^9,11]. The deduced exon structure of the various isoforms is shown below (isoforms 1^10). The dashed-line rectangles represent the POU-speci¢c and POU-homeo domains. Solid-line rectangles represent exons with lengths in nucleotide bases. Spliced exons are represented by solid or dashed oblique lines for proven or presumed exons, respectively. Straight solid lines represent cDNA sequences with no corresponding genomic clones that prove their exon^intron structure. Isoforms 7^9 were isolated by RT^PCR of the 3P region of oct-1. Their upstream sequences are therefore not known and are represented by straight dotted lines. Isoform 10 starts with a block of 204 bp of unknown origin and terminates in a stop codon (asterisk) in the intronic sequence between exons 16 and 17 [6]. The cellular origin of the human and murine cDNA clones are given in parentheses: Ntera2D1 and F9, embryonal carcinomas; 70Z/ 3, Pre-B cell line; 4T00.1 and NS/0, myelomas.
human oct-1 gene and determination of its exon^intron boundaries [10] allows the delineation of the exon composition of two alternatively spliced isoforms [11] and comparison with the exon organization of the 3P region of the murine oct-1 gene [6]. A very high level of conservation was found at the 3P region, beyond the most conserved POU speci¢c and POU homeodomains. Thus, the murine exons V, W, X and Y [6] are very similar in sequences and identical in size to the corresponding human exons 13,
14, 16 and 17 (Fig. 1). The 3P-ultimate human exon was longer than 500 nt and its 3P boundary was not de¢ned [10]. The latter is homologous to the murine exon Z [6]. Isolation of the murine isoforms 5 and 6 reveals that the Z-exon is most likely subdivided into at least four exons (exons 18^21 in Fig. 1). In the present work, we have con¢rmed the presence of exons 18^21 by isolation of three additional murine isoforms (#7^9, Fig. 1). Each of these exons is de¢ned by speci¢c deletions occurring in the di¡er-
BBAEXP 91240 9-2-99
J. Riss, R. Laskov / Biochimica et Biophysica Acta 1444 (1999) 295^298
Fig. 2. Analysis of the expression of the 3P region of the oct-1 gene (exons 17^20) in S107 myeloma and in various murine tissues. cDNA was synthesized by RT and oligo dT from total cellular RNA. PCR was performed using primers from the neighboring exons (16, 21) as depicted in the scheme (arrows). The PCR products of S107 cells were electrophoresed, stained with ethidium bromide and hybridized to an oct-1 3P probe spanning exons 13^21 (A, lanes 1 and 2, respectively). Multiple bands are apparent in the stained gel and upon hybridization. Similar analysis was performed on RNA from various tissues (B,C). The blotted ¢lter was exposed overnight (C), and for a shorter period (4.5 h), so as to demonstrate the weaker hybridization bands (B, lanes 1^3). The fragments of 680, V600 and 295 ¢t in size to isoforms 3,4, 5,7 and 6, respectively (Fig. 1). Three bands of V500, V400 and 360 bp do not correspond to any of the known isoforms. The overall level of oct-1 expression was highest in testis, and lowest in adult and newborn (NB) liver and spleen.
ent isoforms. These results indicate that the murine and human oct-1 genes are composed of at least 21 exons, including exon 15 (72 nt) detected in murine cDNA [7,9]. The occurrence of multiple oct-1 isoforms raises the possibility that they are expressed in a tissue-speci¢c manner. This is supported by the ¢ndings that testis di¡ers from other tissues in its relative level of expression of certain oct-1 isoforms [7]. In the present work, we have searched for additional evidence of `tissue speci¢city' of the oct-1 isoforms. The screening was done by reverse transcription^polymerase chain reaction (RT^PCR) per-
297
formed on the 3P region of the murine oct-1 [7]. Two oct-1 primers, sense (5P-GCCTCATGGCACCCTCA-3P) and antisense (5P-TAGAAAGTTCTCCAATCC-3P), were used to screen the expression of exons 16^21 in the 3P-region of oct-1 in di¡erent tissues. Fig. 2 shows that multiple PCR bands were obtained from RNA of S107 myeloma cells and from the tissues examined. The 680, V600 and 295 bp PCR bands conform in size with isoforms 3,4, 5,7 and 6, respectively. Bands V500, V400 and 360 bp correspond to as yet unidenti¢ed isoforms. It can be seen that di¡erent global levels of expression of the oct-1 products were obtained: the highest level was found in testis, and the lowest in liver and spleen. One or more of the V600, V500, V360 and 295 bp fragments were weakly expressed or absent in brain, liver and spleen (Fig. 2). From similar RT^PCRs performed on RNA of testis and a myeloma cell line (4T00.1), we have isolated and sequenced eight oct-1 cDNA clones, six of which constitute distinct isoforms. Three of the clones conform in size to the 680, and one to the 295 bp main PCR bands. They proved to be identical by sequence to the 3P regions of murine isoforms 3, 4 and 6, respectively. Another clone (609 bp) was identical to isoform 5 (Fig. 1). Three of the eight clones were novel (#7, 8, 9 in Fig. 1). All conform with alternative spliced products of the four consecutive 17^20 oct-1 exons. Two of the novel isoforms (#7, 9) could not have been due to PCR artifacts generated by recombination of the already known cDNA clones. Fig. 3 shows the deduced amino acid sequences of the three novel isoforms. Isoform 7 isolated from myeloma cells lacks exon 17. Its translational product exhibits a frameshift followed by an early stop codon. Thus, its predicted protein has a distinct, much shorter C-terminal tail of three amino acids (SCF). Isoforms 8 and 9 were isolated from testis and are smaller in size than the lowest PCR band of 295 bp (Fig. 2). However, in contrast to the other tissues, testis contains a heterogeneous hybridization signal in the region below 295 bp that conforms in size with these molecular species. Since isoform 8 lacks exons 18 and 20, its deduced protein is shorter due to a frameshift, and has a novel C-terminal tail of 9 ami-
BBAEXP 91240 9-2-99
298
J. Riss, R. Laskov / Biochimica et Biophysica Acta 1444 (1999) 295^298
Fig. 3. Deduced amino acid sequences of the carboxy-terminal regions of the three novel oct-1 isoforms (#7^9). The amino acid sequence of exons 16^20 of the murine isoform 3 is taken as the prototype carboxy-terminal tail (upper sequence). Each of the isoforms 7, 8 and 9 have a distinct amino acid tail due to frameshifts resulting from splicing of internal exon(s). An asterisk indicates the presence of a stop codon.
no acids. Isoform 9 lacks a very large region that includes exons 17^20 and has an insertion of G at the splice junction between exon 16 and 21. This results in a protein with a shorter C-terminal region and a novel tail of 13 amino acids. The deduced proteins of two of the previously reported oct-1 isoforms 1 and 6 (Fig. 1) may also have novel C-terminal tails due to frameshifts resulting from deletions of exons 14, 15 and 20, respectively. In addition, isoform 10 has a unique tail [8], due to the presence of a stop codon at the start of an unspliced intron [6]. The combined data show that ¢ve out of the eight known murine isoforms may have distinct amino acid tails, and suggest that structural variations at the C-termini of the Oct-1 proteins are of functional importance.
References [1] R.A. Sturm, G. Das, W. Herr, Genes Dev. 2 (1988) 1582^ 1599. [2] M. Wegner, D.W. Drolet, M.G. Rosenfeld, Curr. Opin. Cell Biol. 5 (1993) 488^498. [3] C.P. Verrijzer, P.C. Van der Vliet, Biochim. Biophys. Acta 1173 (1993) 1^21. [4] F.E.J. Coenjaerts, J.A.W.M. Van Osterhout, P.C. Vand der Vliet, EMBO J. 13 (1994) 5401^5409. [5] M. Tanaka, J.S. Lai, W. Herr, Cell 68 (1992) 755^767. [6] A. Lerner, L. D'Adamio, A.C. Diener, L.K. Clayton, E.L. Reinherz, J. Immunol. 151 (1993) 3152^3162. [7] J. Ja¡e, M. Hochberg, J. Riss, T. Hasin, L. Reich, R. Laskov, Biochim. Biophys. Acta 1261 (1995) 201^209. [8] A.G. Stepchenko, Nucleic Acids Res. 20 (1992) 1419. [9] N. Suzuki, W. Peter, T. Ciesiolka, P. Gruss, H.R. Scho«ler, Nucleic Acids Res. 21 (1993) 245^252. [10] R.A. Sturm, J.L. Cassady, G. Das, A. Romo, G.A. Evans, Genomics 16 (1993) 333^341. [11] G. Das, W. Herr, J. Biol. Chem. 268 (1993) 25026^25032.
BBAEXP 91240 9-2-99