Index INTRODUCTION MATERIAL AND METHODS cDNA library construction and DNA sequencing Bioinformatics analysis RT-PCR RESULTS The CDK5RAP1_v3 and CDK5RAP1_v4 cDNA sequences Sequence characterization Expression patterns of CDK5RAP1_v3 and CDK5-RAP1_v4 DISCUSSION References

INTRODUCTION

Neuronal Cdc2-like kinase (Nclk) is involved in the regulation of neuronal differentiation and neuro-cytoskeleton dynamics. The active kinase consists of a catalytic subunit, Cdk5, and a 25 kDa activator protein (p25^nck5a) derived from a 35-kDa neuron-specific protein (p35^nck5a) (Ching et al., 2000). Previous experiments suggested that Cdk5 exists in three distinct states: a free monomer, a form associated with p25^nck5a and a form complexed with p35^nck5a (Lee et al., 1996). CDK5RAP1, CDK1 regulatory subunit associated protein 1 (Aliases: C20orf34, C42, CGI-05, HSPC167, Loc 51654) is one of the p35^nck5a-associated proteins. It was expressed in Escherichia coli strain BL21 and was shown to undergo high affinity association with both the free and the Cdk5 complex forms of p35^nck5a (Ching et al., 2000).

Two novel splice variants of CDK5RAP1, which we named CDK5RAP1_v3 and CDK5RAP1_v4, were isolated through the large-scale sequencing analysis (for detecting new full-length human genes) of a human fetal brain cDNA library. We termed the original two splice variants CDK5RAP1_v1 (NM_016408) and CDK5RAP1_v2 (NM_016082), Here we report cloning and characterization of CDK5RAP1_v3 and CDK5RAP1_v4, together with data of the CDK5RAP1 gene structure and mRNA tissue distribution.

MATERIAL AND METHODS

cDNA library construction and DNA sequencing

A cDNA library was constructed in a modified pBluescript II SK (+) vector with human fetal brain mRNA purchased from Clontech. A 0.5-kb DNA fragment containing SfiA (5’-GGCCATTATGGCC-3’) and SfiB (5’-GGCCGCCTCGGCC-3’) recognition sites was introduced into the EcoR and Not sites of pBluescript II SK (+) (Stratagene); the modified vector was then digested with Sfi and the large fragment was excised and purified for library construction. After SfiA disgestion and cDNA size fractionation, cDNAs longer than 500bp were ligated into the SfiA and SfiB sites of the modified vector and the resultant vector was transformed into DH5 α Escherichia coli strain. A 96-well R.E.A.L plasmid ki (Qiagen) was used to synthesize the double-stranded plasmid. Sequencing reactions were performed using BigDye Terminator Cycle Sequencing Kit and BigDye Primer Cycle Sequencing Kit (Perkin-Elmer). The complete sequence was determined and confirmed by primer walking strategy. Subsequent editing and assembly of all the sequences from one clone was performed using Acembly (Sanger’s Center). In total, approximately 20000 human cDNAs were sequenced during 1999–2001.

Bioinformatics analysis

DNA and protein sequence comparisons were carried out using BLAST2.0 at NCBI (http://www.ncbi.nlm.nih.gov/blast). Motif analysis was done with RPS-BLAST searching, PROFILESCAN (http://www.isrec.isb-sib.ch/software/PFSCAN_form.html) and SMART searching (http://smart.embl-heidelberg.de/) were also used. Moreover, relative analysis software programs include GCG (Wisconsin Package, Version 10.0), Gene Runner (http://www.generunner.com/), and GeneDoc (http://www.psc.edu/biomed/genedoc/) were used.

RT-PCR

Two human Multiple Tissue cDNA (MTC) panels (CLONTECH) were used as PCR templates according to the manufacturer’s protocol. The sense primer sequence for CDK5RAP1_v3 and CDK5RAP1_v4 was 5’- gcagctgattcatgagagagataac -3’, and the antisense primer for CDK5RAP1_v3 and CDK5-RAP1_v4 was 5’- ACCAACACTCATTTCACAGTGCAAG-3’. G3PDH-specific primers were 5’-TGAAGGTCGGAG-TCAACGGATTTGGT-3’ (G3PDHF) and 5’-CATGTGGGCCATGAGGTCCACCAC-3’ (G3PDHR). Twenty-four cycles (for G3PDH) or 36 cycles (for CDK5RAP1_v3 and CDK5-RAP1_v4) of amplification (30 s at 94°C, 45 s at 67°C and 1.0 min at 72°C) were performed using Taq plus DNA polymerase (Sangon). The PCR products of CDK5-RAP1_v3, CDK5RAP1_v4 and G3PDH were then electrophoresed on a 1.5% agarose gel. The products generated from human CDK5RAP1_v3 and CDK5RAP1_v4 were 621bp and 489bp, respectively.

RESULTS

The CDK5RAP1_v3 and CDK5RAP1_v4 cDNA sequences

During the large-scale sequencing analysis of a human fetal brain cDNA library constructed in our laboratory, we isolated two novel splice variants of the human CDK5RAP1 gene, named CDK5RAP1_v3 and CDK5RAP1_v4. We submitted the cDNA sequences of CDK5RAP1_v3 and CDK5RAP1_v4 to Genbank under the accession numbers AY462283 and AY462284.

Sequence characterization

The CDK5RAP1_v3 and CDK5RAP1_v4 cDNAs are 1923bp and 1792bp in length, respectively (Fig. 1). The BLAST analysis of these two sequences revealed that CDK5RAP1_v4 lacked 1 exon (exon k) that was present in CDK5RAP1_v3, and as a result these cDNAs encoded different putative proteins (Fig. 2). The deduced proteins were 574 amino acids (CDK5RAP1_v3) and 426 amino acids (CDK5RAP1_v4) in length, and shared the 420 N-terminal amino acids. An upstream in-frame stop codon (tga) was found at position 27–29, and one possible polyadenylation signal (AATAAA) in CDK5RAP1_v3 and two possible polyadenylation signals (AATAAA) in CDK5RAP1_v4 were found after the stop codons utilized by these mRNAs (Fig. 1). In addition, a potential signal peptide was found at residues 1 to 33 of both CDK5RAP1_v3 and CDK5RAP1_v4 (Fig. 1). RPS-BLAST and SMART searching revealed that CDK5RAP1_v3 had an Elp3 domain encompassing residues 248 to 487, a radical_SAM domain (residues 252 to 447), and a TRAM domain (residues 501 to 574). In contrast, CDK5RAP1_v4 had a partial Elp3 domain encompassing residues 248 to 426, and a radical_SAM domain (residues 252 to 424). Moreover, an Elp3 domain and a radical_SAM domain were found in each of CDK5-RAP1_v1, CDK5RAP1_v2, AAD34147.1 and AAH01215; A TRAM domain was found in these proteins except AAD34147.1. Alignment of human CDK5RAP1_v3, CDK5RAP1_v4 and other CDK5RAP1 members was performed, and revealed similarities and identities among these proteins (Fig. 3).

View Details

Fig. 1 The cDNA and deduced amino acid sequences of CDK5RAP1_v3 and CDK5RAP1_v4. The nucleotides are numbered at the left. The in-frame stop codon (tag) is shown in bold; the asterisks denote stop codons; three possible polyadenylation signals (aataaa or attaaa) are boxed. The potential signal peptide is shown in bold; the conserved Elp3 domain is underlined; the radical_SAM domain is underlined and shaded; and the TRAM domain is shaded.

View Details

Fig. 2 Genome structure and alternative splicing of human CDK5RAP1 was predicted by alignment of cDNAs with genomic DNA. P1 and P2 show the positions of primers for RT-PCR. The start codon ATG and stop codon TGA/TAG are marked. The same exon is shown by the same letter in various cDNAs; if the length of the exon is different, “2” is added after the letter to discriminate between the 2 types of exons; exon b2 is the right part of exon b; exon m2 is the left part of exon m. Gray shading of a box indicates a common sequence. The black shading indicates the possible gap of nucleotides 975-1016 of AF152097 cDNA. The accession numbers of these splice variants are CDK5RAP1_v3 (AY462283), CDK5RAP1_v4 (AY462284), CDK5RAP1_v1 (NM_016408), CDK5RAP1_v2 (NM_016082), AF152097 and BC001215.

View Details

Fig. 3 Alignment of human CDK5RAP1_v3, CDK5RAP1_v4 and other CDK5RAP1 members. The aligned protein sequences are registered under Genbank accession numbers: CDK5RAP1_v3 (AY462283), CDK5RAP1_v4 (AY462284), CDK5RAP1_v1 (NP_057492), CDK5RAP1_v2 (NP_057166), AAD34147.1 (AAD34147.1 is encoded by AF152097 cDNA) and AAH01215.1 (AAH01215.1 is encoded by BC001215 cDNA). The alignment was performed using GeneDoc program (http://www.psc.edu/biomed/genedoc/). Residues with dark shading are identical amino acids. Gray shading indicates 80–90% similarity and light gray indicates 60–70% similarity.

To determine the chromosome locations of the CDK5-RAP1_v3 and CDK5RAP1_v4, the cDNA sequences were searched against the human genome database with the BLAST program (http://www.ncbi.nlm.nih.gov/BLAST). The gene was thereby mapped to contig NT_028392.4 from chromosome 20q11.2. Moreover, human CDK5-RAP1_v3 was found to consist of 13 exons and CDK5-RAP1_v4 to consist of 12 exons spanning about 41.3kb.

Expression patterns of CDK5RAP1_v3 and CDK5-RAP1_v4

In order to investigate the tissue expression patterns of the two novel splice variants, we performed RT-PCR. The sense and antisense primer sequences for CDK5RAP1_v3 and CDK5RAP1_v4 were located in exon i and exon m, respectively (Fig. 2). The RT-PCR results showed that human CDK5RAP1_v3 was widely expressed in human tissues. The expression level of CDK5-RAP1_v3 was relatively high in placenta and lung; low levels of expression were detected in heart, brain, liver, skeletal muscle, pancreas, spleen, thymus, small intestine and peripheral blood leukocytes. In contrast, human CDK5RAP1_v4 was mainly expressed in brain, placenta and testis (Fig. 4).

View Details

Fig. 4 Expression patterns of CDK5RAP1_v3 and CDK5RAP1_v4 in 16 human tissues. RT-PCR analysis of CDK5RAP1_v3, CDK5RAP1_v4 and G3PDH (as a control) was performed as described in the text.

DISCUSSION

Here we report two novel splice variants of the human CDK5RAP1 gene named CDK5RAP1_v3 and CDK5-RAP1_v4. The alternative splicing caused the putative Elp3 domain and TRAM domain of CDK5RAP1_v3 and CDK5RAP1_v4 to be different from each other. Elp3 has been isolated as part of a larger six-subunit complex specifically associated with elongating RNA polymerase II (Winkler et al., 2001). It was shown to have histone acetyltransferase (HAT) activity in vitro, and was shown genetically to interact with histone tails and other chromatin-modifying factors (Wittschieben et al., 2000), suggesting that it enhances transcription by modifying chromatin structure, which might ease progression of the polymerase through nucleosomes (Wittschieben et al., 1999). In vivo, ELP3 gene deletion confers phenotypes such as slow growth adaptation, slow gene activation, and temperature sensitivity (Wittschieben et al., 1999). In addition to HAT activity, Elp3-related proteins might possess a second enzyme activity (histone demethylase activity) (Chinenov, 2002). Furthermore, the Elp3 superfamily includes the MiaB family (Esberg et al., 1999) and coproporphyrinogen III oxidase (Lash et al., 2002). All members of the Elp3 superfamily contain a region similar to the catalytic domain of the radical SAM family (Chinenov, 2002). Radical SAM proteins catalyze diverse reactions, including unusual methylations, iso-merization, sulfur insertion, ring formation, anaerobic oxidation and protein radical formation. They function in DNA precursor, vitamin, cofactor, antibiotic and herbicide biosynthesis and in biodegradation pathways (Sofia et al., 2001). Evidence exists that these proteins generate a radical species by reductive cleavage of S:-adenosylmethionine (SAM) through an unusual Fe-S center (Sofia et al., 2001). The TRAM domain is found in two distinct classes of tRNA-modifying enzymes, namely uridine methylases of the TRM2 family and enzymes of the MiaB family, which are involved in 2-methylthioadenine formation (Anantharaman et al., 2001). This domain is predicted to bind tRNA and deliver the RNA-modifying enzymatic domains to their targets. In addition, the TRAM domain is present in several other proteins associated with the translation machinery and in a family of small, uncharacterized archaeal proteins that are predicted to play a role in the regulation of tRNA modification or translation (Anantha-raman et al., 2001). CDK5RAP1_v3 has an integrated Elp3 domain, a radical_SAM domain and a TRAM domain. In contrast, as result of alternative splicing a partial Elp3 domain and an intergrated radical_SAM domain were found in CDK5RAP1_v4, but no TRAM domain was present, and this may affect the enzymatic activity and the ability of CDK5RAP1_v4 to bind and modify tRNA.

Moreover, AF152097 cDNA (CGI-05) was assembled by a comparative proteomic approach (Comparative gene identification, CGI) (Chunhung et al., 2000). Through careful bioinformatics analyses, we found that nucleotides 975-1016 of AF152097 cDNA do not exist in human genomic contig NT_028392.4 from chromosome 20 (Fig. 3). However, nucleotides 975-1016 of AF152097 cDNA could be found in the human DNA sequence from clone RP5-1187J4 on chromosome 20q11.1-11.23 (GenBank acc. NO: AL355392.7), so AF152097 cDNA may also be a splice variant of CDK5RAP1, which would mean that genomic contig NT_028392.4 from chromosome 20 is still incomplete. On the other hand, because no human EST data support the notion that this region is expressed and no region homologous to nucleotides 975-1016 of AF152097 cDNA could be found in the mouse DNA sequence from clone RP23-154J12 on chromosome 2 (GenBank acc. NO: AL732601.16), nucleotides 975-1016 of AF152097 cDNA may be a gap as well in contrast to human genomic contig NT_028392.4 from chromosome 20 (Fig. 2).

Although both CDK5RAP1_v3 and CDK5RAP1_v4 were expressed in brain and placenta, the abundance of CDK5RAP1_v3 in brain and placenta was distinctly higher than that of CDK5RAP1_v4. In contrast, CDK5RAP1_v4 could be detected whereas CDK5RAP1_v3 could not in testis. In addition, BLAST N Searching in the non-redundant and human ESTs databases of NCBI also roughly revealed the number of clones obtained for each splice variant, which could be taken to indicate the relative abundance. These relative abundances were 82%, 6%, 6%, 2%, 0%, 2% for CDK5RAP1_v1, CDK5-RAP1_v2, CDK5RAP1_v3, CDK5RAP1_v4, AF152097 (assembled by comparative proteomic approach) and BC001215, respectively (data not shown). The different expression patterns of the splice variants and comparison of the domains in putative proteins imply that these variants might play different roles in certain tissues. Further studies will be necessary to define the precise roles of splice variants of CDK5RAP1.

This research was a part of Project 30070383 supported by the National Natural Science Foundation of China.