BIOINFORMATICS OF GENE REGULATION, GENE NETWORKS AND METHABOLIC PATHWAYS

222

THE BSP-REPEATS FROM CANIDAE CONTAIN A BIDIRECTIONAL PROMOTER FOR THE RNA POLYMERASE III POTENTIALLY CAPABLE OF ENCODING DOUBLE-STRANDED RNA *Yudin N.S., Naykova T.M., Kondrakhin Yu.V., Kobzev V.F., Romaschenko A.G.

Institute of Cytology and Genetics SB RAS, Novosibirsk, Russia e-mail: yudin@bionet.nsc.ru *Corresponding author

Keywords: RNA polymerase III, intragenic type 2 promoter, double-stranded RNA

Resume

Motivation: Using the standard method of contextual analysis, sequences highly homologous to the RNA polymerase III transcribed tRNA and SINE promoter elements were identified within the Bsp repeat DNAs from the canids silver fox and racoon dog. The aim was to clarify the functional activity of the computer identified potential RNA polymerase III regulatory elements.

Results: Of the four analyzed DNA fragments, only rsV3 variant was transcribable in vitro by RNA polymerase III in nuclear extracts derived from HeLa cells. Assays of the transcriptional products demonstrated that a functionally active bidirectional RNA polymerase III promoter is present within rsV3 DNA. The direct and reverse strands were found to differ in transcriptional efficiency. The specific DNA structure construct is potentially capable of coding for short dsRNA performing different regulatory functions in the eukaryotic genomes.

Availability: Detailed information about use of algorithms of contextual search for the RNA polymerase III transcribed genes are available at request. The algorithm of context search is described by Kondrakhin Yu.V. et al., (see this volume).

Introduction

Table 1. Sequence homologous to the RNA polymerase III promoter elements identified within the the Bsp repeats. А box В box Frag-

ment Orientation

Sequence Localization SA

Spacer size, bp. Sequence Localization SB

Transcriptio-nal efficiencyin vitro

Direct GAGGTAAGTGG 34-43 6.41 32 AGCTCAACCCA 75-85 5.74 + Reverse TAGTCTTGTGG 64-55 6.54 41 GCATCGGGGCT 14-4 5.00 +++ rsV3 Reverse TGGTGCAGATGT 242-253 6.69 51 AGCTCAACTCT 183-193 6.22 - Direct GAGCTAAGTGG 322-332 7.34 32 AGTTCAATCCA 364-374 7.00 - Direct CAGTGGAGTGG 610-620 7.49 - - - - - Direct GAACTAAGTGG 1055-1065 5.47 - - - - - rsV51

Direct CAGTGGAGTGG 1344-1354 7.49 - - - - - Reverse TAGTCTTGTGG 19-10 7.05 - - - - - rsV5 Direct - - - - AGCTCAACCCA 31-39 5.74 - Direct - - - - CGTTCAAGTCA 625-636 7.72 - Direct - - - - TGTTCGAAACT 741-749 7.40 - rsN2 Direct AAAGGGTGTTGG 1130-1141 5.77 30 AGTTCAGGCCT 1171-1179 5.88 -

Designations: SA и SB are the measures of similarity of Bsp repeat DNAs to the canonical A and B boxes of RNA polymerase III transcribed gene promoters [17].

The highly repeated sequences in the genomes of different Canidae species, we have designated as the Bsp repeats (1), show the characteristic features of regulatory type DNA (2). The major evolutionary steps of these complex hierarchical DNA molecules have most likely long preceded speciation within the Canidae family (3,4). The increasing abundance of their structures containing DNA elements of the regulatory type has suggested that the Bsp repeats might have been involved in the cladogenetic reorganization of the genomic material in ancestral Canidae (2,5). The association of satellite-like, along with interspersed-like repeats (SINE), is a feature providing indirect support for this possibility (5,6). In fact, intragenic promoter elements, characteristic of the RNA polymerase III transcribed genes (tRNA, 5S RNA, U6 RNA genes) and SINE from the genomes of various eukaryotic species, have been identified within the Bsp-repeats by computer approaches. Thus, nucleotide sequences homologous to the A and B boxes of the type 2 promoter tRNA genes have been identified (1).

BGRS� 2000

223

Table 1 presents fragments of the Bsp repeats from the silver fox and racoon dog genomes that are homologous to the A and B boxes (3,8). The measure of similarity of each of the Bsp fragments to the canonical site DNA from the tRNA promoter regions is given in a separate column in Table 1. Fig. 1 represents schematically the position of the potential A and B boxes with respect to each other in a surrounding of functional motifs capable of, for example, modulating RNA polymerase II transcription through binding of regulatory proteins that are irrelevant to RNA polymerase III (1). The SINE-like repeat DNAs are known to associate the structural features of the RNA polymerases III and II regulatory regions, abundantly interspersed among the RNA polymerase II transcribed genes (10,11,12). Fig.1 shows the unusual combinations of regulatory RNA polymerase III elements in the Bsp repeats that make the A and B of Bsp repeats different in orientation, structure, sequence composition and spacer size from the expected canonical. This raised the question: Can the specific transcriptional factors bind to these unusual A and B boxes, and, possibly, interact to provide appropriate assembly of the RNA polymerase III transcriptional machinery and to start RNA synthesis ? Our aim was to analyze the in vitro functional activity of the A and B boxes identified within the Bsp repeats by computer search.

Methods and algorithms The potential transcriptional capacity of the Bsp repeats was tested in nuclear extracts derived from HeLa cells (13). To exclude the possible transcription of DNA sequences by RNA polymerase II, RNA was synthesized in the presence of alpha-amanitin concentration (0.5 µg/ml) that did not affect RNA polymerase III activity, yet inhibited that of RNA polymerase II (14).

Results Using a standard system in vitro, in conditions optimized for RNA polymerase III transcription, DNA rsV3 from the silver fox genome within the pUV3 and pBV3 plasmids was the only one selectively transcribed. As a result, there were two transcripts of discrete size, 4.5 and 3.2 kb in pBV3. When pUV3 was used as template, the synthesized product was of different sizes. At a high concentration (200 µg/ml), alpha-amanitin inhibited the activity of both RNA polymerases III and II so that transcription of the Bsp repeats was completely suppressed. This meant that a functionally active promoter was present within the rsV3 Bsp repeat fragment. In an attempt to determine transcript sizes, the experiments were repeated on a linear rsV3 DNA fragment. The synthesized RNA was analyzed in 1.8% agarose gel electrophoresis. As a result, transcripts of different sizes were detected. Transcripts of 60 bp were severalfold more abundant than those of 300 bp (Fig. 2). To identify the initiating transcription strand, the synthesized products were analyzed with the primer extension assay, using reverse transcriptase. Each of the primers was complementary to one of the strands. The transcription probability for both rsV3 strands (Fig.1) could not be excluded. When oligonucleotide 5'-GCTTGTCTGAAGTGTTGTGA-3' complementary to RNA that was homologous to the reverse rsV3 strand served as primer (Fig.1), the synthesized DNA was 30 bp long. The DNAs synthesized by extension of primer 5'-GTGCATCTCTGAGGGACTTA-3' complementary to the RNA, was homologous to the direct rsV3 strand, were of different lengths 167, 150, 128, 119, 106 and shorter. This indicated that the direct and reverse strands differed in their expression levels, with one of the promoters initiating the transcription of products of different sizes.

Discussion In vitro tests of the computer detected promoter elements within the Bsp repeat demonstrated that some of the Bsp repeat variants are transcribable by RNA polymerase III. A contextual analysis of the A and B box structure, which considered the module organization and the numerous disposition interactions within and between the modules, was performed (9,15). From the results it was concluded that Bsp repeat transcription was provided

Figure 1. Scheme of position of the potential A and B boxes with respect to each other and some functional motifs capable of modulating RNA polymerase II transcription at both strands of four Bsp repeat fragments.

BGRS� 2000

224

not only by the structure of the RNA polymerase III promoter elements as such, but also by the surrounding motifs with other than RNA polymerases III regulatory functions. Thus, combinations of modules in A and B boxes (at positions 322-332 and 364-374, respectively) (Table 1) are more similar (SA=7.34; SB=7.00) to the canonical variants of the tRNA (9) in rsV51 as compared with rsV3 (SA=6.41; SB=5.74). rsV3 variant is highly homologous to rsV51 in this particular DNA region (1,3). However, rsV3 contains substitutions not occurring within the known tRNA genes in the A and B box regions. Despite this, a fully efficient transcriptional machinery is accurately assembled in this region of rsV3 DNA and not in that of rsV51 DNA. This may be accounted for by the presence of a potential site for the "ubiquitous" SP1 transcription factor between the A and B boxes in rsV51 (1). The selective SP1 binding may hinder the assembly of the active RNA transcription machinery. The absence of rsV5 transcription was expected because the boxes A and B were misoriented with respect to each other (Table 1). One plausible reason why there was no transcription in rsN2 was the strong deviation of the A and B boxes from the canonical sequences (Sa=5.77, SB=5.88). In the A box, highly conserved adenine (A) was substituted by T at positions 7, and the slightly lesser conserved nucleotides were substituted at positions 1 (T→A) and 3 (G→A), too. Module 2 -GGG- variant (at positions 4,5,6, the A box) never occurs in combination with GTTGG sequence at the 3'-termini of the A box in certain tRNA genes. As noted above, transcription is initiated on both RNA strands, with the promoter strength of the A and B boxes (at positions 64-55 and 14-4, respectively) of the reverse strand exceeding that of the direct strand. It is noteworthy that, of the two box pairs on the reverse strand, the one less similar to the canonical variants of the tRNA A and B boxes (at position 64-55 and 14-4) were more selective for the assembly of the initiation RNA polymerase transcription (Table 1). It could not be ruled out that substitutions of highly conserved guanine G→T at position 12 of the A box (see Table 1, positions 242-253), as well as T→C (at positon 3 of the B box), negatively affected the functional activity of the promoter elements. The structural specificity of the other A and B boxes on the rsV3 reverse strand (Fig.1, positions 64-55 and 14-4 respectively) probably made this box pair the highest transcription initiating. The B box is the least similar (SB=5.00) to the canonical of the tRNA genes with respect to structure (9). In the B box, the conserved nucleotides are substituted at positions 2 and 7 (G→C, A→G, respectively, see Table 1). The A box consists of the structural variants of modules that abundantly occur in the tRNA genes. These are module 1 (-TAG-; at positions 1,2,3, the A box), module 2 (-TCT-; at positions 4,5,6, the A box), and module 3 (-GTGG-; at positions 8,9,10,11 of the short variant of the A box) (9). The A box is unique in that -TAG- is combined with -TCT-, not with TGG-, as expected for the tRNA genes. Moreover, conserved adenine at position 7 of the A box is replaced by T in it. It is pertinent to note that the substitution of the conserved A→G at position 7 has been reported for the VAI promoter in adenovirus 2 showing the high transcriptional activity characteristic of rsV3 (16). Consequently, the distribution of specific nucleotides at each of the positions of the A and B boxes are accurately patterned to provide the appropriate transcription level. To our knowledge, the bidirectionality of RNA synthesis for rsV3 in vitro (Fig.2) has not been observed for the RNA polymerase III transcribed genes (17). However, bidirectional promoters were detected in the regulatory region of the RNA polymerase II transcribed genes (18). There is a slight similarity between the rsV3 promoter regions and composed of two oppositely oriented and quite far apart tRNA genes of the trypanosome promoter (19). The DNA region identified within rsV3 with overlapping A and B box pairs on complementary strands appears to be potentially capable of coding for short double stranded RNA (dsRNA). It is known that dsRNAs regulate eukaryotic genome functions by activating specific PKR family protein kinases that inhibit translation process (20).

A B C

pBV

5

pUV

3

pBV

3pU

V51

pUN

2

pBV

5

pUV

3

pBV

3

pUV

51

pUN

2

M

alpha-amanitin0.5 µg/ml 200 µg/ml

45003200

16311419

517

M

517

396

214

75

Dire

ctpr

imer

Rev

erse

prim

er M1419

517396

214

7565

Figure 2. Selective transcription of Bsp repeats in vitro. (А). In vitro transcriptional products of different clones of Bsp repeats (pUV3, pBV3, pUV51, pUV5, pUN2) in nuclear extracts derived from HeLa cells. RNA was synthesized in the presence of the alpha-amanitin concentration (0.5 µg/ml) that did not affect RNA polymerase III activity, yet inhibited that of RNA polymerase II, and at a concentration of 200 µg/ml, that inhibited the activity of both RNA polymerases III and II. (B). Run-off transcription of rsV3 DNA fragment in vitro. (C). Analysis of rsV3 DNA fragment in vitro transcripts using primer extension.

BGRS� 2000

225

Screening for sequences homologous to the bidirectionally promoter construct we found within rsV3 in the GeneBank revealed no such sequences. The potentially significant regulatory role of the DNA constructs as sources of genes coding for the nontranslated dsRNAs should be kept in mind. This warrants further screening rounds of the GeneBank that would include sequence data for full size eukaryotic genomes. The absence of homology of rsV3 with the known tRNA genes, the ancestors of many SINE sequences, is evidence that the rsV3 RNA polymerase III promoter boxes might have arisen otherwise than those in SINE sequences. Taken together, the hierarchical structure of the Bsp repeats, the abundance of direct and inverted short repeats and numerous unusually combined RNA polymerase III elements (Table 1, Fig. 1) strongly suggest that a convergent mechanism underlies the evolutionary emergence of the promoter elements within the Bsp repeats.

Acknowledgments This work was supported partly by the Russian Foundation for Basic Research grant 99-04-49722. We are grateful to Prof. K. Seifart for providing us reagents and interest in the work.

References 1. Potapov V.A., Solov'ev V.V., Romaschenko A.G., Sosnovtsev S.V., Ivanov S.V. (1991) Features of the structure and

evolution of the complex tandemly organized Bsp repeats in the fox genome. II. Tissue-specific and recombination sites of the BamHI dimer. Mol.Biol. (Msk), 25, 116-132.

2. Romaschenko A.G., Yudin N.S. (2000) Molecular evolution of Bsp repeats from Canidae: structural development and mechanisms of DNA informational enrichment. In "Modern concepts of evolutionary genetics" V.K. Shumny and A.L. Markel eds., Novosibirsk, ICG SD RAS, 177-188 (in Russian).

3. Potapov V.A., Solov'ev V.V., Romaschenko A.G., Sosnovtsev S.V., Ivanov S.V. (1990) Structure and evolution of complex tandem Bsp repeats in fox genome. I. Structure and internal organization of a BamHI dimer. Mol.Biol. (Msk), 24, 1649-1665.

4. Sosnovtsev S.V., Ivanov S.V., Solov'ev V.V., Potapov V.A., Romaschenko A.G. (1993) Recombination events in evolution of satellite-like Bsp repeats: formation of subrepeat and monomer units predates the divergence of Canidae lineages. Mol.Biol. (Msk), 27, 992-1013.

5. Belyaeva T.A., Vishnivetsky P.N., Potapov V.A., Zhelezova A.I., Romaschenko A.G. (1992) Species- and tissue-specific transcription of complex, highly repeated satellite-like Bsp elements in the fox genome. Mammalian Genome, 3, 233-236.

6. Mikhailova S.V., Babenko V.N., Romashchenko A.G., Beliaeva T.A., Melidi N.N., Lavrinenko V.A., Guvakova T.V., Ivanova L.N. (1995) Distribution of Canidae Bsp-repeat transcripts in arctic fox kidney: structural similarity of Bsp-repeats with SI NEs. Dokl Akad Nauk, 343, No. 2, 260-264.

7. Potapov V.A., Sosnovtsev S.V., Solov'ev V.V., Ivanov S.V., Romaschenko A.G., Kolchanov N.A. (1988) Structure of complex repeats from fox C-heterochromatin: regulatory elements of replication, recombination and gene expression control. Dokl Akad Nauk, 299, No. 5, 1250-1255.

8. Ivanov S.V., Potapov V.A., Filipenko E.A., Romaschenko A.G. (1991) Heterogeneity of the Canidae Bsp repeat family: discovery of the EcoRI subfamily. Genetika, 27, No.6, 973-982.

9. Rogozin I.B., Kondrakhin Yu.V., Naykova T.A., Yudin N.S., Voevoda M.I., Romaschenko A.G. This issue.. 10. Weiner A.M., Deininger P.L., Efstratiadis A. (1986) Nonviral retroposones: genes, pseudogenes and transposable

elements generated by the reverse flow of genetic information. Ann. Rev. Biochem., 55, 631-661. 11. Britten R.J. (1997) Mobile elements inserted in the distant past have taken an important functions. Gene, 205. 177-182. 12. Britten R.J. (1996) DNA sequence insertion and evolutionary variation in gene regulation. Proc. Natl. Acad. Sci. USA,

93, 9374-9377. 13. Weil A.P., Segall J., Harris B., Ng S.Y., Roeder R.G. (1979) Faithful transcription of eukaryotic genes by RNA

polymerase III in systems reconstituted with purified DNA templates. J. Biol. Chem.., 254, No. 13, 6163-6173. 14. Transcription and translation. A practical approach. (1987) B.D. Hames and S.J. Higgins eds., IRL Press, Washington. 15. Kondrakhin Yu.V., Rogozin I.B., Romaschenko A.G. This issue. 16. Schneider H.R., Waldschmidt R., Seifart K.H. (1990) Human transcription factor IIIC contains a polypeptide of 55 kDa

specially binding to Pol III genes. Nucleic Acids Res., 18, No.16, 4743-4750. 17. Geiduschek P.E., Tocchini-Valentini G.P. (1988) Transcription by RNA polymerase III. Ann. Rev. Biochem., 57, 873-

914. 18. Romaschenko A.G., Potapov V.A., Solov'ev V.V. (1989) Distribution, structure and functional significance of complex

repeated DNA sequences from animal genomes. In "Structural-functional organization of genome" Ed. by V.K. Shumny, Novosibirsk, Nauka, 80-114.

19. Nakaar V; Dare AO; Hong D; Ullu E; Tschudi C (1994) Upstream tRNA genes are essential for expression of small nuclear and cytoplasmic RNA genes in trypanosomes. Mol Cell Bio., 14, No.10, 6736-6742.

20. Proud C.G. (1995) PKR: a new name and new roles. Trends Biochem. Sci., 20, 241-246. 21. Smit A.F., Riggs A.D. (1995) MIRs are classic, tRNA-derived SINEs that amplified before the mammalian radiation. .

Nucleic Acids Res., 23, No. 1, 98-102.