To investigate the response of S. solfataricus and S. acidocaldarius to UV light induced DNA damage, cultures were exposed to 200 J/m² single wavelength (254 nm) UV light during early exponential growth (OD₆₀₀ about 0.2, Figure 1a). At each time point indicated, samples were taken for mRNA and protein extraction, and for flow cytometry. Preliminary studies showed that higher UV fluency levels resulted in the death of the majority of the cells (data not shown). UV treatment was carried out at room temperature and the cultures were kept in the dark throughout the procedure to prevent photoreactivation. To compensate for effects related to temperature changes, control cultures were subjected to the same procedure except for the UV irradiation. The temperature shock and handling effects caused growth retardation in both the control and irradiated cultures, particularly in the first hour after treatment (Figure 1b). The appearance and clearance of DNA damage following UV irradiation was quantified using an antibody specific for cyclobutane pyrimidine dimers (CPDs). We observed biphasic kinetics where a majority of CPDs (50-60%) was repaired within 30 minutes and the rest were repaired at a slower pace, with 95% repair after 4 h (Figure 1c).

Overall, 163 S. solfataricus and 157 S. acidocaldarius genes showed a >2-fold change in transcript abundance (up or down) after UV treatment compared to controls. For both species more genes were up-regulated than down-regulated, although this difference was more pronounced in S. acidocaldarius than in S. solfataricus. This is contrary to findings for Halobacterium NRC-1, where most changes directly after UV treatment were in the form of down-regulation, mainly of metabolic genes 6. The greatest changes in transcript levels occurred at 90 and 120 minutes. Expression ratios for selected genes following UV irradiation are summarized in Table 1, and Additional data files 1-6 provide comprehensive data for all S. solfataricus and S. acidocaldarius genes on the arrays. The transcriptional responses of selected genes were analyzed further by quantitative RT-PCR to provide an independent confirmation of the microarray data (Table 2). We observed a good correlation between the two methods (Figure 2), providing confidence that the microarray data reflected real changes in the relative abundance of mRNA levels in UV treated and control cells. In general, we observed a good agreement between the two species for both induced and repressed genes in all the gene categories analyzed, including genes encoding hypothetical proteins (Table 1). This reflects the shared evolutionary heritage and lifestyle of these two Sulfolobus species, which grow in very similar environments. It also suggests that the transcriptional effects observed are not artifactual. Where there are differences, for example, in the expression changes observed with Cdc6-2 and RadA described below, these may reflect real differences in the UV damage response of the two species.

The initial response to UV irradiation must involve the generation of signals within the cell, most obviously DNA damage due to photoproducts and strand breaks, and increased concentrations of ROS. Signal transduction by as-yet unknown mechanisms led to a transcriptional response. In these experiments our earliest time point was taken at 30 minutes post-UV irradiation. It is clear that a transcriptional response is already well underway at this early time point. A number of genes show high levels of induction at 30 minutes. Single-stranded DNA binding proteins (SSBs) in bacteria and eukarya are known to play an important role in the early detection of DNA damage, and the S. solfataricus SSB has been shown to detect UV photoproducts in vitro 11. Our microarray data show two-fold up-regulation of the ssb gene (sso2364, saci0975) at 30 minutes after UV treatment for S. solfataricus and at 30 and 60 minutes for S. acidocaldarius, but this mild induction was not apparent from quantitative RT-PCR (Table 2) and not obviously reflected at the protein level (see later). Protein kinases play an essential role in the mediation of the DNA damage response in eukarya. The gene encoding the atypical Ser-/Thr-RIO protein kinase-1 is induced at early time points in both S. solfataricus (sso2374) and S. acidocaldarius (saci_0965). The biological targets of RIO kinases are not yet known, but they have been implicated in cell cycle progression and chromosomal maintenance in yeast 12. It will be interesting to determine whether the archaeal RIO kinase mediates protein phosphorylation in response to DNA damage, analogous to the ATM/ATR damage response pathway in eukarya. The genes for beta carotene biosynthesis crtB and crtZ (sso2905 and sso2906; saci_1733 and saci_1734) are strongly induced at the 30 minutes time point. Production of beta carotene has previously been observed in response to UV irradiation in S. acidocaldarius 13, and the pigment has a protective effect against UV light. Thus, Sulfolobus cells apparently synthesize protective pigments in response to damaging levels of UV light.

There are also early transcriptional responses in genes encoding a variety of metabolic pathways. Five genes encoding subunits of a putative heterodisulfide reductase (Hdr), arranged in an operon in both Sulfolobus species (e.g. sso1127-35), are strongly induced following UV irradiation. The function of this multienzyme complex is important for methanogenesis in methane-producing archaea, but is not understood in organisms such as Archaeoglobus fulgidus and Sulfolobus species, which do not generate methane. One suggestion is that these enzymes function in an electron transport chain that oxidizes menaquinol and passes electrons to a sulfate reduction pathway 14. Another induced enzyme with a related function is the flavin adenine dinucleotide (FAD)-dependent sulfide:quinone oxidoreductase (SQR), which oxidizes sulfide and passes electrons into the respiratory chain, and is essential for chemolithoautotrophic growth using hydrogen sulfide as an electron donor 15. The gene encoding this protein is induced in both S. solfataricus (sso2261) and S. acidocaldarius (saci_0331), and it is notable that the latter is located on the genome next to the S. acidocaldarius hdr genes (saci_0334-0339). Two genes in the riboflavin biosynthetic pathway (sso0401, sso0402; saci_0821, saci_0822) are also induced at early time points. Riboflavin is a precursor of FAD and flavin mononucleotide (FMN) 16. Given the observation that many FAD-dependent oxidoreductases are also induced, this is consistent with an increased requirement for the cofactor following UV irradiation.

Both eukarya and bacteria have mechanisms to inhibit DNA replication and cell division following DNA damage to allow DNA repair to take place, and a similar response has been observed in the euryarchaeon Pyrococcus abyssi following ionizing irradiation 7. Genes encoding DNA replication proteins were amongst the most highly repressed following UV irradiation in both S. solfataricus and S. acidocaldarius. We observed significant down-regulation of the genes for the replication initiation proteins Cdc6-1 and Cdc6-3 17, MCM helicase, and Gins proteins 18 (Table 1).

In contrast with the repression of cdc6-1 and cdc6-3, strong up-regulation of cdc6-2 was evident in S. solfataricus. cdc6-1 and cdc6-3 expression levels were also reduced in S. acidocaldarius, whilst cdc6-2 levels were unchanged. Thus, in both organisms the levels of Cdc-6 protein are likely to be increased compared to Cdc6-1 and Cdc6-3 following UV irradiation. The increase in S. solfataricus cdc6-2 transcription was confirmed by quantitative RT-PCR (Table 2), and western blotting with antibodies against all three Cdc6 proteins showed that the transcriptional changes were reflected in the expression levels of each of the proteins (Figure 3). At 1 h after UV treatment there was a clear increase in the amount of Cdc6-2, which persisted until 4 h post-treatment before declining again. Cdc6 proteins are AAA+ type ATPases that bind to replication origins and interact with the replicative helicase MCM 19. The role of Cdc6-2 in DNA replication is not completely clear, but elevated levels are found in stationary phase cultures and during the post-replicative cell cycle stage, that is, in non-replicating cells 17. Also, Cdc6-2 can bind at origins at sites overlapping the binding sites for Cdc6-1 and Cdc6-3. These observations have prompted the suggestion that Cdc6-1 and Cdc6-3 act in a positive fashion to promote replication initiation, while Cdc6-2 may act as a repressor 17. The differential transcriptional control of cdc6-2 from cdc6-1 and cdc6-3 following UV DNA damage therefore suggests a mechanism by which initiation of DNA replication could be repressed, giving cells time to carry out DNA repair.

Sulfolobus species encode two to three paralogues of the 7 kDa chromatin protein Sul7 (Sso7 or Sac7) 20. Previous reports showed a considerable decrease of sso7 gene transcripts in UV-damaged S. solfataricus cultures 8. In agreement with these studies, the two sso7 genes (sso9180, sso9535) on the microarrays showed substantial down-regulation in the microarray analyses, placing them amongst the most strongly repressed genes observed in this study. Sso7 makes up about 5% of the total soluble cell protein and is one of the 2 main chromatin proteins present in Sulfolobus species 20. A reduction of sso7 transcript levels to one-third compared to those in control cells may reflect the repression of DNA replication, as chromatin proteins are needed in proportion to the amount of newly synthesized DNA. In contrast, the genes encoding the chromatin proteins Alba1 and Alba2 and Alba1 acetylase Pat (sso2813) 21 showed little change in expression following UV irradiation; however, the transcript abundance of the Alba1 deacetylase Sir2 (sso2478) did increase in both organisms, raising the possibility that Alba1 acetylation levels are reduced in response to DNA damage.

We also assessed DNA content and cell size of control and UV-treated cultures by flow cytometry. The Sulfolobus cell cycle is dominated by the post-replicative phase 22 and, during exponential growth, a majority of the population contains two chromosomes, while most of the remaining cells are in the replicative stage, and only a small proportion (≤5%) contain a single chromosome. The control cultures displayed typical DNA content distributions corresponding to this cell cycle organization at all time points, as did the UV-irradiated cultures at early time points after treatment (Figure 4). In the UV-treated cultures, the number of cells that contained a single chromosome equivalent gradually increased in relative abundance with time. In parallel, cells with a DNA content of more than one chromosome equivalent started to appear, eventually forming a continuous distribution that merged with the leftmost peak, which corresponds to DNA-less cells and background debris. Also, an increased proportion of cells with a DNA content between one and two chromosome equivalents became apparent between 1 h and 5 h after UV irradiation. The increased frequency of cells containing a single chromosome could be due to inhibition of replication initiation, in accordance with the changes in the expression patterns of the cdc6 genes (above), although the extensive DNA degradation indicates that chromosome breakdown also contributed to the relative increase in one-chromosome cells.

At 5-6 h after treatment, the DNA content distributions began to resemble those of the untreated control cultures, in addition to a large peak corresponding to DNA-less cells. Thus, the UV irradiation resulted in chromosome breakdown and generation of DNA-less cells in part of the cell cultures, while the remaining population recovered and eventually restored the typical DNA content distribution of an exponentially growing culture, in accordance with the continued increase in optical density (Figure 1b). No change in the flow cytometry light scatter distributions was observed in either culture during the time period analyzed (data not shown), indicating that any effects on cell size or cell integrity were below the detection limit.

In addition to the major replicative DNA polymerase DpoI 23, the S. solfataricus genome encodes three other DNA polymerases: DpoII, DpoIII and DpoIV 2425. The latter is a member of the DinB family of lesion bypass polymerases, and is specialized for the bypass of UV photoproducts 26, whilst DpoII and III have not been studied biochemically and their functions in vivo are unknown. Of the four polymerase genes, only dpoII was induced on the microarray following UV damage (Table 1), a result confirmed by quantitative RT-PCR (Table 2). The gene is annotated as two open reading frames (ORFs; sso8124 and sso1459) in the S. solfataricus genome due to a frameshift at position 75 in the gene, and both are up-regulated significantly. This may be a sequencing error, as the intact gene has been sequenced previously from the same strain of S. solfataricus 27. The most closely related polymerase in the crenarchaeon Pyrobaculum aerophilum, PolB3, has been shown to possess DNA polymerase activity 28, but the two proteins share only 27% sequence identity and the Sulfolobus enzyme lacks the clear amino-terminal exonuclease domain of PolB3. The observation that the Sulfolobus dpoII gene is induced after UV irradiation whilst the replication machinery is repressed is suggestive of a role for the enzyme in the repair of DNA damage. Although most DNA repair enzymes are expressed constitutively in Sulfolobus (see below), it may be advantageous to control the expression of DpoII so that it does not interfere with DNA replication during undisturbed growth. Clearly the role of this enzyme deserves further study.

The archaeal basal transcription machinery, consisting in essence of a 12-subunit RNA polymerase holoenzyme and the transcription initiation proteins TBP (TATA binding protein) and TFB (transcription factor 2B homolog), represents a minimal version of the eukaryal transcription apparatus 29. TBP and TFB bind at archaeal promoters, recruiting RNA polymerase to initiate transcription, and no other factors are required in vitro 30. Following UV irradiation, we observed no significant changes in the transcript levels of TBP, nor any of the RNA polymerase subunits. Interestingly, however, there was a big change in transcription levels for the homologues of the basal transcription factor TFB in both species after UV-treatment. TFB comprises three domains; the carboxy-terminal core domain interacts with TBP and binds the BRE (the TFB responsive element) next to the TATA box using a helix-turn-helix (HTH) motif. The amino-terminal Zn ribbon domain, conserved in all eukaryal and archaeal TFB proteins, binds to RNA polymerase. A construct consisting of the Zn ribbon domain alone can inhibit transcription initiation in vitro by competing with TFB for recruitment of RNA polymerase. The B-finger domain, an extended region of the protein between the Zn ribbon and core domains, interacts with RNA polymerase and stimulates transcription initiation 31.

The genomes of all three Sulfolobus species sequenced contain three homologues of the tfb gene (Figure 5a). Two of them (tfb1 and tfb2) encode full length TFB proteins, and TFB1 has been shown to support transcription initiation in vitro 30. The third homologue (sso0280 in S. solfataricus) is significantly shorter than the other two. The sequence includes the Zn ribbon but lacks the B-finger domain and has an abbreviated core domain (Figure 5a). Under normal growth conditions in the absence of UV irradiation, the CP (crossing point: the cycle number at which the PCR product reaches a set threshold, a number that is lower for more abundant transcripts) values obtained from quantitative RT-PCR indicated that tfb1 is the most abundant transcript of the three, with tfb2 expressed at a low level (Table 2). Following UV irradiation, however, the relative abundances of these three transcripts changed significantly (Tables 1 and 2). Whilst there was a significant down-regulation of tfb2 (sso0946) and no significant change in the transcription levels of tfb1 (sso0446), tfb3 was one of the most highly up-regulated transcripts in response to UV irradiation in both Sulfolobus species (Figure 5b), and this was reflected in the increased concentrations of the TFB3 protein in UV-irradiated cells (Figure 5c). The radically altered organization of TFB3 suggests a role other than transcription initiation in Sulfolobus species. As it retains the Zn ribbon domain for interaction with RNA polymerase, TFB3 may act as a competitive inhibitor of transcription initiation, as has been noted for the Zn ribbon domain in isolation 32. This was confirmed in vitro using the purified native RNA polymerase and recombinant TFB1 and TFB3 proteins from S. solfataricus (Figure 5d). In the absence of TFB3, TFB1 interacts with RNA polymerase, and can be pulled-down with antibodies raised against the B' subunit of RNA polymerase. However, increasing concentrations of TFB3 abolish the interaction between RNA polymerase and TFB1, consistent with the model proposed above. TFB3 is unlikely to interact with the BRE, as the conserved carboxy-terminal HTH motif that makes this contact is absent. The dramatic increase in tfb3 transcript abundance following DNA damage may provide a mechanism to modulate transcription from certain promoters.

Transcription and DNA repair are linked intimately, both through transcriptional responses to DNA damage and by the fact that many DNA lesions are repaired following an encounter with a transcribing RNA polymerase molecule (transcription coupled repair). In most bacteria, many DNA repair proteins are under the control of the LexA repressor 33. Archaea lack LexA homologues and, therefore, a classical SOS regulon, but an SOS-type response is still possible, even if not mediated by LexA. RadA (the archaeal RecA homolog) is up-regulated in response to UV damage in Halobacterium NRC-1; moderately in one study 6 and by up to seven-fold in another 34, and moderately up-regulated in response to ionizing radiation in the hyperthermophile Pyrococcus furiosus 35. Our data showed no increase in radA transcript (Table 1) or protein levels (Figure 6) in S. solfataricus. In S. acidocaldarius, however, we observed a two-fold up-regulation of radA at 2 h.

UV irradiation causes a broad spectrum of DNA lesions, including photoproducts (predominantly CPDs and 6-4 photoproducts), oxidative damage to bases and single-strand DNA breaks. Despite the clear indications that Sulfolobus species have a transcriptional response to DNA damage, including induction of cdc6-2 and tfb3, we found no significant changes in transcription of known DNA repair proteins (Table 1). The experiments reported here were carried out in the absence of light, and no up-regulation of the photolyase gene was observed. Likewise, the lesion bypass DNA polymerase DpoIV, specialized for replication past UV photoproducts 26, was not induced.S. acidocaldarius possesses a gene encoding a UV damage endonuclease (saci1096) of the UVDE family 36, which is UV inducible in Schizosaccharomyces pombe following UV irradiation 37. The S. acidocaldarius gene was only mildly induced after UV irradiation (Table 1).

In addition, none of the putative archaeal nucleotide excision repair proteins (XPB1, XPB2, XPD, XPF, Fen1, SSB) showed significant induction following UV irradiation. The data for xpb1 contrasts with an earlier study that reported a significant increase in xpb1 transcription following UV irradiation 8. We therefore checked transcription of the xpb1 gene using quantitative PCR, which confirmed that no induction occurred (Table 2). However, the experimental conditions were not identical in our experiments and those of Salerno et al. 8, as in the latter a higher UV flux was used and transcript levels were monitored over a longer time period. Genes encoding enzymes involved in the pathway of base excision repair, which are required for the removal of bases damaged by UV light, were also not up-regulated significantly, and the same was true for the double-strand break repair proteins required for homologous recombination and the rescue of stalled replication forks (Table 1).

In addition to the generation of photoproducts, UV irradiation results in the generation of ROS that cause a broad spectrum of DNA damage. We observed significant up-regulation of a cluster of ORFs in both species (sso2078-2080, saci1820-1823) with probable roles in the detoxification of ROS. One of the ORFs (sso2079, saci1821) codes for the Sulfolobus Dps protein. Dps is expressed under conditions of oxidative stress, when it forms a cage-like structure, binds to DNA and catalyzes the oxidation of reactive Fe(II) to Fe(III) ions, preventing the generation of damaging ROS via Fenton chemistry 38. The dps gene is also up-regulated following UV irradiation in Halobacterium NRC1 6 and gamma irradiation of P. furiosus 35. The other members of this cluster are a homologue for a NRAMP protein 39 with a potential role in the transport of Fe^2+/3+ and Mn²⁺ (sso2078, saci1820) and a ferritin homologue implicated in the storage of Fe^2+/3+ (sso2080, saci1822). The gene encoding peptide methionine sulfoxide reductase (MsrA) is also induced at early time points in both species (sso1503, saci_1170). MsrA reverses oxidative damage to methionine side chains in proteins and, therefore, protects the cell from oxidative stress and ROS. These gene products are thus likely to provide a front-line defense against DNA and protein damage caused by ROS generated as a consequence of UV irradiation.

S. solfataricus P2 (DSMZ 1617) and S. acidocaldarius (DSMZ 639) were grown aerobically at 80°C to early exponential phase (OD₆₀₀ approx. 0.18). Exposure to 200 J/m² UV light (254 nm) was carried out using a Stratalinker^® UV crosslinker 1800 (Stratagene, La Jolla, CA, USA). Cultures were irradiated in 25 ml volumes in standard pre-warmed Petri dishes to ensure UV light exposure of all cells, and returned to the incubator immediately afterwards. Total handling and UV exposure time totaled less than 5 minutes, and culture temperatures did not drop below 55°C. An earlier study showed that this dose of UV results in an average of 1 CPD per kilobase of DNA for S. solfataricus 8. Control cultures were treated in the same way except for the UV irradiation. Samples for flow cytometry and western blots were taken at various times 1 h before and up to 6 h after UV exposure. Samples for RNA extraction were taken at 30, 60, 90 and 120 minutes after UV treatment, and RNA was extracted immediately.

For RNA extraction, samples were centrifuged at 4,000 rpm for 10 minutes at 4°C and the cell pellet was re-suspended in 1 ml of lysis buffer (4 M guanidine isothiocyanate, 25 mM sodium citrate, pH 7, 0.5% sarcosyl, 100 mM b-mercaptoethanol) and incubated on ice for 15 minutes. Na-acetate was added to a final concentration of 0.2 M. The samples were extracted once with an equal volume of water saturated phenol and once with 24:1 (v/v) chloroform:isoamylalcolhol. RNA was precipitated overnight with an equal volume of isopropanol and washed with 75% ethanol. The pellet was resuspended in 174 μl H₂O and 20 μl of 10× DNase buffer (100 mM Tris, pH 7.5, 25 mM MgCl₂, 5 mM CaCl₂) and treated with 45 U DNase I for 30 minutes at 37°C. Sodium acetate (0.2 M final concentration) was added and RNA was extracted once with an equal volume of water saturated phenol and once with chloroform:isoamylalcohol. RNA was precipitated overnight with an equal volume of isopropanol and washed with 75% ice cold ethanol. Pellets were dried under vacuum and re-suspended in H₂O. The quality of the RNA was checked on 2% agarose gels and the concentration determined by measuring the UV absorbance at 260 nm.

cDNA was prepared using random hexamer primers (GE Healthcare, Little Chalfont, Bucks, UK). Each cDNA reaction (20 μl) contained 15 μg of RNA, 5 μg primers, 0.5 mM dATP, 0.5 mM dGTP, 0.5 mM dCTP, 0.1 mM dTTP, 0.4 mM aminoallyl-dUTP (Sigma-Aldrich Co Ltd, Gillingham, UK), 42 mM DTT and 400 units Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA). The RNA was incubated with the random hexamer primers at 70°C for 10 minutes and cooled rapidly on ice. Nucleotides and reverse transcriptase were added to the primed RNA and cDNA synthesis was carried out at 42°C for 2 h. RNA was degraded by addition of 2 μl 200 mM EDTA and 3 μl 1 M NaOH and the reaction was incubated at 70°C for 15 minutes, after which 3 μl of 1 M HCl was added to neutralize the solution. The cDNA was purified using a MinElute PCR purification kit (Qiagen Ltd, Crawley, W. Sussex, UK). The cDNA was eluted in 20 μl 0.1 M NaHCO₃ and labeled with either Cy3 or Cy5 monoreactive dye (Amersham) in the dark for 90 minutes. Samples to be co-hybridized were mixed and purified using Qiagen MinElute columns. The labeled cDNA was eluted in 20 μl elution buffer (10 mM Tris-HCl pH 8.5).

DNA microarrays containing 1,914 and 2,488 gene-specific tags for S. acidocaldarius and S. solfataricus, respectively, were produced as described previously 47.

Microarray slides were pre-hybridized for 40 minutes at 42°C in 10 mg/ml BSA, 5× SSC and 0.1% SDS, washed three times in dH₂O, followed by one wash in isopropanol and dried by centrifugation. The labeled cDNA was mixed with 57 μl hybridization mix (6.5× SSC, 0.16% SDS, 66% formamide) followed by addition of 10 μg tRNA and 10 μg herring sperm DNA. Samples were denatured for 2 minutes at 95°C and rapidly cooled on ice. The array slide was placed in a hybridization chamber (TeleChem International, Sunnyvale, CA, USA) and a cover slip (number 2 LifterSlip; Erie Scientific Company, Portsmouth, NH, USA) applied over the printed area. The hybridization mixture was added and the chamber was incubated at 42°C for 16 h. Slides were washed for 5 minutes in 2× SSC, 0.1% SDS at 42°C, then for 10 minutes in 0.1× SSC, 0.1%SDS at room temperature and 5 times in 0.1× SSC at room temperature and dried by centrifugation. Slides were scanned using a GenePix Personal 4100A microarray scanner (Molecular Devices Corp, Sunnyvale, CA, USA).

All data analysis was carried out using GenePix Pro 5.1 (Molecular Devices) and the free web-based BioArray Software Environment (BASE) 48. Normalization within and between arrays was performed to remove dye specific effects and to compensate for differences between arrays 49. Data from duplicate spots were merged, as were data from dye-swap arrays and the triplicate experiments, and each value thus represented data from 12 hybridizations.

RNA from S. solfataricus P2 was extracted at time points 30, 60, 90 and 120 minutes following UV irradiation, and purified using an RNeasy Kit (Qiagen) according to the manufacturer's directions, with the substitution of Proteinase K in place of lysozyme, allowing more efficient digestion of the protein S layer surrounding Sulfolobus cells. Quantitative PCR was carried out using the Biorad iScript One-Step RT-PCR kit with SYBR Green kit according to the manufacturer's directions using a BioRad iQ5 RT-PCR system. The amplification efficiency for each primer set was determined by calibration with genomic DNA. Reactions for UV-irradiated and control samples were analyzed in triplicate by quantitative PCR. Crossing-points (CPs) for each amplification were measured, and the ratio of gene expression in the UV-treated versus control cultures at each time point was quantified using the method of 50. The gene-specific forward and reverse oligos used for amplification were (5' to 3'): tfb1for, GGCCAGAACTTTGGATGAGA' tfb1rev, CCAGCAGTTAACCCAGAACC; tfb2for, CGCGTTGAAAAGAGTCCAAT; tfb2rev, GGAAGCTGCGCTCAAAGATA; tfb3for, TTAGATTCGCGTTAAATAATGG; tfb3rev, CAAATACGATCGCTTTCTTCG; cdc6-2for, CCACATAGAGAAGAGAAGATTAAGG; cdc6-2rev, GTAGCTGTTTTCCCAGTACC; xpb1for, TGAATGCAGGGGTTCTTGTT; xpb1rev, AGTTTTGCTTGCTTGCCATT; dpoIIfor, CCGCCTAGGGATAAAACCAT; dpoIIrev, CCTCAACTTCAGGCTTTTCG; ssbfor, AGTTTTGGAAGCAAGCGAAG; ssbrev, GTGGTCCACGCGTTTTCTAT.

Samples (0.1 ml of cell culture) were fixed in 70% ethanol (final concentration) and stored at 4°C. The fixed cells were precipitated at 13,000 rpm for 30 minutes and re-suspended in 1 ml Tris-MgCl₂ (10 mM Tris pH 7.4, 10 mM MgCl₂) and precipitated again for 30 minutes, 13,000 rpm at 4°C. The pellets were re-suspended in 70 μl of Tris-MgCl₂ and mixed with an equal volume of staining solution containing ethidium bromide (20 μg/ml) and mithramycin A (100 μg/ml). Sample analysis was performed as described previously 22, except that an A40 Analyzer instrument (Apogee Flow Systems, Hemel Hempstead, Herts, UK) was used.

For western blots, whole cell protein was separated on NuPage 4-12% Bis-Tris SDS gels (Invitrogen). Protein content was measured by Bradford assay and the protein content of each sample was normalized. Western blots were carried out following standard procedures. For the production of TFB3 antibodies, 1 mg of purified TFB3 protein was used to raise polyclonal antibodies in sheep (Scottish National Blood Transfusion Service).

DNA was extracted using the DNeasy Tissue kit (Qiagen) according to the manufacturer's protocol. DNA was quantified spectrophotometrically and DNA concentrations were adjusted to 10 ng/μl in H₂O. DNA was added at 300 ng per well to microtitre plates (Greiner Bio-One Ltd, Stroudwater, Stonehouse, UK) and allowed to dry at 37°C for 48 h. Plates were washed 4 times with 100 μl washing buffer (PBS, 0.1% Tween 20) and 100 μl blocking buffer (PBS, 0.1% Tween 20, 5% (w/v) milk powder) were added to each well and left at room temperature for 90 minutes. Blocking buffer was removed and wells were washed 4 times with 100 μl washing buffer. The CPD-specific antibody (Kamiya Biomedical Company, Seattle, WA, USA) was added at a concentration of 1:250 (in washing buffer) and plates were incubated at room temperature for 90 minutes. Each well was washed five times with 100 μl washing buffer and 100 μl of a 1:5,000 diluted HRP-conjugated anti-mouse antibody (Pierce, Cramlington, Northumberland, UK) was added. Plates were incubated for 90 minutes at room temperature, after which each well was washed five times with 100 μl washing buffer. A 150 μl 1-Step™ Turbo TMB Elisa (Pierce) was added to each well and incubated for 5-30 minutes at room temperature. Reactions were stopped by the addition of 150 μl 2 M sulfuric acid and absorbance was quantified at 450 nm.

tfb1 and tfb3 genes from S. solfataricus genomic DNA were amplified by PCR and cloned into a modified Gateway vector for expression with a cleavable amino-terminal his-tag. Full details of this methodology will be published elsewhere, and are available from the corresponding author on request.

pDEST-TFB1 and pDEST-TFB3 were transformed to Escherichia coli Rosetta (DE3) or C43 (DE3) cells, respectively. Cells were grown in LB medium at 37°C to an OD₆₀₀ of 0.6. At this point, induction of His₆-tagged TFB1 and TFB3 was carried out by 200 μM isopropyl β-D-thiogalactopyranoside IPTG overnight at room temperature. Cells were harvested and re-suspended in lysis buffer (20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 0.1% Triton X-100, 1 mM MgCl₂ and Complete EDTA-free protease inhibitors (Roche Basel, Switzerland)), lysed by sonication and clarified by centrifugation. The supernatant was heated to 70°C for 10 minutes and re-centrifuged. The resultant supernatant was diluted two-fold in buffer A (20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 30 mM NaH₂PO₄) plus 30 mM imidazole and was then applied to a column containing Ni-NTA-Agarose (HiTrap 5 ml chelating HP; GE Healthcare defined above) pre-equilibrated with buffer A + 30 mM imidazole. The proteins were eluted with a linear gradient of imidazole (buffer A + 500 mM imidazole). Fractions containing His-TFB1 and His-TFB3 were identified by SDS-PAGE and pooled. For His-TFB3 purification, all the buffers were supplemented with 2 mM DTT. His-TFB1 was dialyzed against Tev cleavage buffer (20 mM Tris-HCl (pH 7.0), 500 mM NaCl, 1 mM DTT, 30 mM NaH₂PO₄ and 10% glycerol) overnight at 4°C. The day after, the protein was Tev cleaved overnight at room temperature by adding a final concentration of 200 ng/μl of Tev protease. Cleaved TFB1 was re-purified by loading onto the same column pre-equilibrated with buffer A + 30 mM imidazole and collecting the flow through. Positive fractions were pooled and dialyzed extensively against freezing buffer (50 mM Tris-HCl (pH 7.5), 200 mM KCl, 1 mM DTT, 1 mM EDTA, 0.01% Triton X-100 and 50% glycerol). Further purification of His-TFB3 was achieved by loading the protein onto a HiLoad 26/60 Superdex 200 pgr size exclusion column (GE Healthcare) and following the protocol as shown below with buffer supplemented with 1 mM DTT.

Ten liters of S. solfataricus P2 cells were grown up to mid-log phase. At this point, cells were harvested and re-suspended in lysis buffer (50 mM MES (pH 6.0), 100 mM NaCl, 1 mM EDTA, 1 mM DTT and Complete protease inhibitors (Roche)), lysed by sonication and clarified by centrifugation. The resultant supernatant was two times diluted in buffer 1 (20 mM MES (pH 6.0), 1 mM EDTA and 0.5 mM DTT) plus 10 mM NaCl and was then applied to a pre-equilibrated Hi-Trap 5 ml Heparin column (Amersham). The proteins were eluted with a linear gradient of NaCl (buffer 1 + 1 M NaCl). Fractions containing the RNA polymerase were identified by dot-blot immunodetection with RNA polymerase subunit B' specific antibodies and pooled. RNA polymerase was purified to homogeneity by using a HiLoad 26/60 Superdex 200 pgr size exclusion column (GE Healthcare) equilibrated with gel filtration buffer (20 mM MES (pH 6.0), 200 mM NaCl, 1 mM EDTA and 0.5 mM DTT). Fractions containing the RNA polymerase were identified by SDS-PAGE and confirmed by mass spectrometry. Positive fractions were pooled and dialyzed extensively against freezing buffer.

For anti-RNA polymerase immunoprecipitations, 2 μl of antibody was added to 20 μl of Dynabeads Protein G (Invitrogen) in antibody binding buffer (20 mM MES (pH 5.0), 0.1% Tween-20) at room temperature. After 40 minutes, beads were washed twice in the same buffer plus 0.1% Tween-20 continued by an extra wash in binding buffer 'BB' (PBS, 150 mM NaCl, 0.1% Tween-20). A mix of RNA polymerase, TFB1 and His-TFB3 in amounts indicated in the figure legends were mixed in 'BB' in a total volume of 100 μl. This mix was combined with antibody-coated beads for 2 h at room temperature. Then, the beads were washed twice with buffer 'BB' and PBS and re-suspended in 2× SDS-PAGE loading buffer.

The raw data have been submitted to the EBI ArrayExpress database with accession code E-MEXP-1252.