A cDNA encoding the DNA ligase I homologue has been isolated from Arabidopsis thaliana using a degenerate PCR approach. The ORF of this cDNA encodes an amino acid sequence of 790 residues, representing a protein with a theoretical molecular mass of 87.8 kDa and an isoelectric point (pI) of 8.20. Alignment of the A. thaliana DNA ligase protein sequence with the sequence of DNA ligases from human (Homo sapiens), murine (Mus musculus), clawed toad (Xenopus laevis) and the yeasts Schizosaccharomyces pombe and Saccharomyces cerevisiae showed good sequence homology (42–45% identity, 61–66% similarity), particularly around the active site. Sequence data indicate that the Arabidopsis DNA ligase is the homologue of the animal DNA ligase I species. Functional analysis of the cDNA clone demonstrated its ability to complement the conditional lethal phenotype of an S. cerevisiae cdc9 mutant defective in DNA ligase activity, confirming that the cloned sequence encodes an active DNA ligase. The level of the DNA ligase transcript was not increased in A. thaliana seedlings in response to DNA damage induced by a period of enhanced UV-B irradiation. However, the cellular level of the DNA ligase mRNA transcript does correlate with the replicative state of plant cells.
DNA ligases (polydeoxyribonucleotide synthases) play a crucial role in several aspects of DNA metabolism, including DNA replication, DNA excision repair and genetic recombination. These enzymes catalyse the formation of a phosphodiester bond at single- and double-stranded breaks in double-stranded DNA, and are divided into two broad classes on the basis of co-factor requirement for either NAD+ or ATP (Lindahl and Barnes 1992). All eukaryotic ligases identified to date show a requirement for ATP.
In contrast to mammalian cells there is relatively little information concerning either the number or roles of DNA ligase activities in plant cells. This is due, at least in part, to the fact that plant cells contain active deoxyribonucleases, making the development of reliable and sensitive assays for DNA ligase activities difficult (Jenns and Bryant 1978). However, a DNA ligase activity has been purified 600-fold from pea (Pisum sativum L.) roots (Kessler 1971), from carrot (Daucus carota L.) cells (Tsukada and Nishi 1971) and Lilium microspores (Howell and Hecht 1971), but no detailed characterization of the ligase activity has been performed. The partial purification and the characterization of DNA ligase activity has also been reported from P. sativum seedlings (Daniel et al. 1985), where two DNA ligase populations were isolated from different subcellular locations, one form of the ligase activity being found bound to chromatin, while the second activity was localized to the soluble fraction. Both of these ligase activities from pea seedlings required Mg2+ and ATP. Additional evidence supporting the possibility that there may be two forms of DNA ligase in pea seedlings concerned differences in the response of these ligase activities to the presence of spermidine. In contrast, when crude extracts of rye (Secale cereale L.) seedlings were separated on a sucrose density gradient only a single broad peak of DNA ligase activity was detected, corresponding to a protein species having a molecular weight of about 68 kDa (Elder et al. 1987).
As a consequence of the lack of detailed biochemical characterization of the enzymes involved and the availability of few cloned genes, our understanding of DNA repair mechanisms in plants (reviewed by Britt 1996) lags some way behind that in other systems. What is assumed to be excision repair has been demonstrated in a number of plant systems where DNA lesions are removed in the absence of photoreactivating light (thereby preventing direct reversal of UV-B induced DNA damage by the activity of the light-dependent enzyme DNA photolyase). In Arabidopsis thaliana in particular, light-independent repair of UV-induced DNA lesions has been observed (Britt et al. 1993;Pang and Hays 1991). Thus plants, of necessity, require a DNA ligase activity to participate in the final step of excision repair, in addition to the requirement for DNA ligase activity in both DNA replication and recombination. Here we report the cloning, functional analysis and expression patterns of a DNA ligase from A. thaliana, the first plant homologue described for this class of enzymes.
Results and discussion
PCR amplification of a DNA ligase sequence from A. thaliana using degenerate primers
A cDNA encoding part of the A. thaliana DNA ligase I was isolated using the PCR strategy illustrated schematically in Fig. 1. Initially, the degenerate oligonucleotide primers L1 and L2 were designed on the basis of conserved amino acid motifs in an alignment of eukaryotic DNA ligases, whose sequences were already present in public databases (human DNA ligase I, Barnes et al. 1990; mouse, Savini et al. 1994; and the two distantly related yeast sequences, Saccharomyces cerevisiae, Barker and Johnston 1983, and S.pombe, Johnston et al. 1986). These primers were used to amplify a fragment of 550 bp from an A. thaliana cDNA library in the vector λYES (Elledge et al. 1991). The amplified product was cloned into pCRII (Invitrogen, Leek, the Netherlands) to generate a clone designated pLig10, which was sequenced to confirm the identity of the amplified DNA as a putative DNA ligase. The remainder of the DNA ligase sequence was isolated using the strategy illustrated in Fig. 1, and the assembled sequence for fragments I–IV confirmed the identity of the cDNA product as a putative A. thaliana DNA ligase I homologue.
Characterization of the A. thaliana DNA ligase I homolgue
To isolate a full-length cDNA for the A. thaliana DNA ligase I, the cDNA insert from pLig10 was purified, labelled by oligopriming with [α-32P]dATP, and used to probe an A. thaliana cDNA library in λZAP2 (Kieber et al. 1993) by plaque hybridization. Using subsequent rounds of hybridization with radiolabelled fragments II–IV (Fig. 1), we were able to isolate a full-length cDNA for DNA ligase I. This clone was designated pAtlig1A and contains the entire coding sequence for the A. thaliana DNA ligase, as demonstrated by sequence homology with previously isolated DNA ligase sequences (Fig. 2). The A. thaliana DNA ligase protein contains 790 amino acid residues and has a predicted molecular mass of 87.8 kDa.
The A. thaliana DNA ligase protein sequence was aligned with the sequences of four other eukaryotic DNA ligase I homologues present in the EMBL database (Fig. 2). The mouse DNA ligase sequence was not included in these alignments due to its high identity (84%) to the human DNA ligase I sequence. The A. thaliana protein sequence is typical of other eukaryotic DNA ligases in that it contains two domains, a C-terminal region that is highly conserved and that contains the active site of the enzyme (Tomkinson et al. 1991b), and a less conserved N-terminal portion not necessary for the catalytic activity but that has been suggested to be involved in the regulation of activity of the mammalian enzyme (Kodama et al. 1991;Savini et al. 1994). The N-terminal portion also contains a putative nuclear localization sequence (Montecucco et al. 1995). Within the C-terminal region there are obvious localized homologies to the mammalian, yeast and amphibian sequences. Two of the conserved amino acid motifs within the C-terminal region of the DNA ligase I homologues are indicated in Fig. 2 (L1 and L2), and these regions of homology were utilized in designing the degenerate primers used to isolate the initial DNA ligase PCR fragment. The active site comprises a 17-amino acid peptide that is conserved in multiple DNA and RNA ligases. The invariant lysine (K) residue (at position 444 in the A. thaliana protein) within the active site motif is thought to participate in the adenylation reaction (Lindahl and Barnes 1992).
Sequence similarity and copy number
The amino acid identities and similarities for pair-wise comparisons of the different DNA ligase I homologues and the human DNA ligase III and IV homologues are presented in Fig. 3. The A. thaliana protein exhibited a similar level of identity to each of the DNA ligases from yeasts, mammals and amphibians (42–46% identity) with no obvious bias in relatedness to any one particular ligase. The results also suggest that the A. thaliana DNA ligase belongs to the DNA ligase I family, since the A. thaliana protein shares 45% identity with the human DNA ligase I homologue but only 24% identity with human DNA ligases III and IV. A Southern blot of genomic A. thaliana DNA digested with the restriction enzymes EcoR1, HindIII and BamH1 (Fig. 4a) gave either two or three discrete bands, suggesting that the DNA ligase gene is present as a single or low copy number gene within the A. thaliana genome.
Functional complementation of the DNA ligase mutation in S. cerevisiae
The budding yeast S. cerevisiae contains two distinct DNA ligase activities, of which the Cdc9 DNA ligase is the major joining activity (Ramos et al. 1997), and cells carrying a thermosensitive mutation in the CDC9 gene fail to grow at temperatures above 30°C (Johnston and Nasmyth 1978). To demonstrate that the A. thaliana cDNA clone pAtlig1A encodes a functional DNA ligase, we investigated the ability of this cDNA to complement a cdc9 mutation and restore growth at the restrictive temperature to a yeast strain carrying this mutation. For this, the 2.4kb cDNA insert in pAtlig1A was amplified by PCR using primers L11 and L12 (see the Experimental procedures), and the product cloned into the yeast expression vector pYEUra3 such that the expression of the A. thaliana cDNA was under the control of the inducible GAL1 promoter. This construct was used to transform S. cerevisiae strain SB799 to uracil prototrophy at 23°C, and the transformants assayed for viability at 37°C on media supplemented with either glucose or galactose/raffinose. The A. thaliana DNA ligase I polypeptide was able to complement the cdc9 temperature-sensitive DNA ligase mutant during growth at the non-permissive temperature (37°C) only when expression of the DNA ligase cDNA was induced by the presence of galactose and raffinose in the growth medium (Fig. 5). Neither the SB799 strain alone nor the SB799 strain transformed with the pYEUra3 plasmid will grow at 37°C. The exchange of glucose for galactose and raffinose in the growth medium does not result in functional complementation of the cdc9 mutation by the A. thaliana cDNA clone, as expected for a galactose inducible construct. The rescue of cdc9 by the cDNA cloned in the antisense orientation with respect to the GAL1 promoter is surprising, but is probably due to expression mediated via cryptic promoter elements in adjacent vector sequences providing low-level constitutive expression. This phenomenon has been observed and described previously for the rescue of other temperature-sensitive mutations in S. cerevisiae (Patterson et al. 1986).
Tissue-specific expression of DNA ligase and response to UV-B-induced DNA damage
Northern analysis of total RNA extracted from different tissues from 35-day light grown plants demonstrated that DNA ligase I transcripts could be detected in all plant tissues analysed (roots, leaves and flowering structures;Fig. 4b), with the expression of the DNA ligase transcript being highest in the flowering structures. The flowering structures in A. thaliana contain a series of floral meristems undergoing cell division and differentiation to produce the floral organs. Consequently this tissue contains a relatively high proportion of cells undergoing mitosis and meiosis and these processes appear correlated with a high level of expression of the DNA ligase. In addition preliminary studies using polyclonal antibodies raised to a conserved sequence at the C-terminus of eukaryotic DNA ligases (Tomkinson et al. 1990) indicate that a nuclear protein, whose antigenic properties and chromatographic behaviour are similar to those of the human DNA ligase I protein (Lindahl and Barnes 1992), is present at high levels in nuclei from cells at the base of the primary wheat leaf within the meristematic region (R.M. Taylor et al. unpublished results). The level of this protein declines in cells from wheat leaf sections distal to the meristem, reflecting the reduction in the number of cells undergoing cell division in these sections. This association of higher levels of DNA ligase with replicating cells is in agreement with many of the studies on mammalian cells, where increased ligase activity (Elder and Rossignol 1990) and transcript levels (Montecucco et al. 1992) are seen when quiescent cells are induced to proliferate. The steady state level of the DNA ligase transcript is severely reduced when seedlings are grown for 9 days in the absence of light or light-grown seedlings are subjected to a 48-h dark treatment (Fig. 4b). However, transcript levels are restored to levels found in light-grown tissue when the 48-h dark-treated tissue is subjected to a subsequent 12-h light treatment, suggesting that light may play a role in the regulation of transcription of this gene (Fig. 4b).
Northern analysis was performed on total RNA extracted from A. thaliana seedlings that had been exposed to an enhanced UV-B treatment known to induce significant levels of thymine dimers and (6–4)photoproducts in DNA (UV-B incident dose rate 1.25 W m–2 in the absence of photoreactivating wavelengths of light;Taylor et al. 1996). No change in the steady state levels of the DNA ligase mRNA was observed in response to irradiation (Fig. 4c). A similar result was seen when A. thaliana cell cultures were irradiated with UV-B (G.I. Jenkins and J. Christie, personal communication). In yeast, DNA ligase message levels have been shown to increase within 2–4 h in response to relatively harsh DNA damaging conditions (UV-C radiation, 254 nm;Johnson et al. 1986), and mammalian primary fibroblast cells exposed to UV-C radiation show a threefold increased level of the DNA ligase I message 24 h post-irradiation (Montecucco et al. 1992). However, under the much lower biologically relevant doses of UV-B irradiation used here, no increase could be detected in DNA ligase I transcription levels in A. thaliana seedlings, suggesting that increased DNA ligase I transcription is not part of a stress response to elevated levels of DNA damage or elevated levels of UV-B irradiation, at least under the conditions employed in this study.
From the sequence alignment data presented here (Fig. 3), the A. thaliana cDNA clone encodes a protein that is a homologue of the mammalian DNA ligase I protein, and the DNA ligase properties of the A. thaliana protein have been confirmed by its ability to functionally complement a temperature-sensitive mutation in the S. cerevisiae DNA ligase I homologue. Over recent years three distinct ligases have been cloned from mammalian cells and a fourth DNA ligase activity characterized. There is also evidence that other eukaryotes, such as Drosophila melanogaster (Takahashi, and Senshu 1987) and the yeast S. cerevisiae (Ramos et al. 1997;Wilson et al. 1997), contain at least two distinct DNA ligase species. It is not known if there is more than one form of DNA ligase in plant cells, either nuclear encoded or plastid encoded. It will now be of interest to elucidate the precise in vivo role of the plant DNA ligase I homologue and also to determine the number of other distinct DNA ligase species and their specific roles in DNA replication, repair and recombination in higher plants. This knowledge is necessary for an understanding of DNA metabolism in higher plants during the critical phases of the cell cycle involving mitosis and meiosis, and in responses to DNA damage elicited via environmental agents and stress.
Escherichia coli DH5α (supE44 ΔlacU169 [80lacZΔM15]hsdR17 recA1 endA1 gyrA96 thi-1 relA1) was used for the routine preparation and storage of recombinant DNA, while strain XL-1 Blue (recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac–F’[ProAB lacIq lacZΔM15 Tn10 (tet)r])was used as host for plaque hybridization. Saccharomyces cerevisiae strain SB799 (cdc9 ade ura3) was used to test for functional complementation. DNA procedures, yeast and bacterial manipulations were performed using established protocols (Sambrook et al. 1989) unless otherwise stated. Plasmid DNA was prepared on an analytical scale by alkaline lysis (Birnboim and Doly 1979) or using QIAGEN columns according to the manufacturer's instructions (Qiagen, UK). DNA fragments were labelled by random hexanucleotide priming with [α-32P]dATP and the Klenow fragment of Escherichia coli DNA polymerase I (Boehringer Mannheim, Lewes, UK). Yeasts were transformed using alkaline cations as described previously (Soni et al. 1993).
Oligonucleotides were synthesized on an Applied Biosystems A381 Synthesizer using phosphoramidite chemistry. The oligonucleotides used to amplify fragments of A. thaliana DNA ligase I were: L1 5′GA (G/A)TA (T/C)AA (A/G)TA (T/C)GA (T/C) GG, L2 5′ C (T/C)TT (T/C)TTIA (AG/C) (T/C) TTIA (A/G)CCA, L3 5′ TCCGGGT ACTTCCCAGTGTT, L4 5′ ACTTTAACGTCAAGGAG, L5 5′ TCAGATGCTACC TATGAG, L6 5′ CGTGAATGTAAGCGTGAC, L7 5′ TGAGAACCT AA CC GC AA, L8 5′ CACAAGAAGTGCCTTCATTCG, L9 5′ GCGGCAATAACAGTTCTCAA, L10 5′ GACC ACGCGTATCGATGTCGAC. Oligonucleotides L11 (5′ ATGTTAGCGATTCGATCGTCGAATTAC) and L12 (5′ GAGGCCAGATTACTGTTGTTC) were used to generate the full length DNA ligase cDNA.
Typically, PCR reactions (50 μl) contained 10 μl 5× reaction buffer, 10 ng template DNA, 10 pmol of each primer, 0.25 μm each dNTP, 1 unit enzyme (Taq DNA polymerase; Boehringer Mannheim, UK) and MgCl2 at concentrations up to 3.5 mm. For reactions involving primers L1–L10, conditions were typically 95°C 30 sec (1 cycle); 95°C 30 sec, 55°C 30 sec, 72°C 1min (30 cycles); 72°C 5 min (1 cycle). For amplification of the full-length cDNA using primers L11 and L12, the reaction conditions were: 95°C 30 sec (1 cycle); 95°C 30 sec, 60°C 1 min, 72°C 2.5 min (30 cycles); 72°C 5 min, 25°C 10 min (1 cycle).
Plant material and growth environments
Light grown seedlings were raised in an environmental growth chamber (SGC970/C/HQI, Sanyo-Gallenkamp, Loughborough, UK) under controlled conditions of constant humidity (70%), with 16 h light (20°C), 8 h dark (15°C) cycles for the period stated. Incident radiation was supplied by a combination of high intensity discharge lamps (Osram, Light Source Supplies, Bishops Stortford, UK) and tungsten lamps. The quantum flux density was measured daily 2 h into the photoperiod and was equivalent to 100 μmol m–2 s–1 PAR for light grown tissue. Arabidopsis thaliana (c.v. Landsberg erecta) seeds were sown directly onto damp compost and germinated under the conditions described above; etiolated tissue was raised under the same conditions in the absence of light. Details concerning the exact age and the stage during the day/night cycle at which the tissue was used are presented in the legends to each figure.
RNA isolation and Northern analysis
Total RNA was isolated from A. thaliana seedlings using a modified version of the method of Knight and Gray (1994). The extraction buffer contained 100 mm Tris HCl, pH 8.5, 100 mm NaCl, 20 mm EDTA, 1% Sarcosyl and 1.7% diethyldithiocarbamate (4 ml per g of tissue). The RNA was selectively precipitated from the total nucleic acids by the addition of 4 m LiCl. Northern analysis was performed essentially as described by Taylor et al. (1996).
Genomic DNA was isolated from whole A. thaliana seedlings by the method of Doyle and Doyle (1990). DNA (12 μg) was digested (16 h, 37°C) with 400 units of enzyme in a volume of 400 μl. Samples were ethanol precipitated and resuspended in a volume of 20 μl TE and electrophoresed on a 0.7% agarose gel. The DNA fragments were transferred to positively charged nylon membranes and hybridization carried out at 65°C according to the manufacturer's instructions (Dupont, Brussels, Belgium), using the radiolabelled fragment I (Fig. 1) as a probe. Filters were washed to a final stringency of 0.1 × SSC at 65°C and subjected to autoradiography.
The nucleotide sequence data reported will appear in the EMBL, Genbank and DDBJ nucleotide sequence databases under the accession number X97924.
We thank Stephen Elledge and Rachael Duncan (Zeneca Agrochemicals, Jealott's Hill, Bracknell, UK) for providing the A. thaliana cDNA libraries, and Tom Lindahl (ICRF) for the DNA ligase I antibody. This work was supported by a research grant from the Biotechnology & Biological Sciences Research Council.
EMBL Data Library accession number X97924 (Arabidopsis thaliana mRNA for DNA ligase).