The stress-activated MAP kinase Sty1/Spc1 and a 3′-regulatory element mediate UV-induced expression of the uvi15⁺ gene at the post-transcriptional level

Abstract

Exposure of Schizosaccharomyces pombe cells to UV light results in increased uvi15⁺ gene expression at both the mRNA and protein levels, leading to elevated cell survival. This UV-induced expression of the uvi15⁺ gene was reduced in Δsty1 and Δwis1 cells lacking the stress-activated protein kinase pathway, but not in DNA damage checkpoint mutants. To further understand the cellular mechanisms responsible for this UV-induced expression, the transcription rate and mRNA half-life were investigated. Transcription run-on assays revealed that the rate of uvi15⁺ transcription was increased 1.8-fold regardless of Sty1 when cells were UV irradiated. The half-life of uvi15⁺ mRNA was also increased 1.5-fold after UV irradiation, but it was decreased in the Δsty1 background for both basal and UV-induced mRNAs, indicating that the stress-activated MAPK cascade can mediate UV-induced gene expression by increasing mRNA half-life. Deletion analyses identified a 54 nt element downstream of the distal poly(A) site, which was involved in the increased half-life of uvi15⁺ mRNA. These results suggest that both Sty1 and the 3′-regulatory element regulate UV-induced expression of the uvi15⁺ gene at the post-transcriptional level.

INTRODUCTION

Eukaryotes have developed complex but structurally and functionally conserved systems responding to UV light during evolution. One of the well-known systems is the mitogen-activated protein (MAP) kinase cascade which involves sequential activation of three distinct kinases through phosphorylation: MAPK kinase kinase (MAPKKK) activated by an extracellular stimulus phosphorylates MAPK kinase (MAPKK), which then activates MAPK by dually phosphorylating two residues in a closely located motif, a threonine and a tyrosine (1,2). Studies on the MAPKs have revealed that the MAPK cascade is highly conserved from yeast to mammalian cells, including human (3,4). The mammalian c-Jun N-terminal kinases (JNKs) and p38 kinases are the most well-studied MAPKs and are known to be activated in response to a variety of stress conditions, including osmotic stress, heat shock, oxidative stress, UV light and some DNA-damaging agents, and hence are termed stress-activated MAPKs (SAPKs).

In the fission yeast Schizosaccharomyces pombe the SAPK pathway consists of Win1, Wak1–Wis1–Sty1 kinase cascade (5). Like p38 kinase, Sty1 (also known as Spc1) is activated by a similar range of environmental damages and activated Sty1 regulates two downstream transcriptional activators, Atf1 (6,7) and Pap1 (8), resulting in expression of the stress-induced genes. Based on studies on Δatf1 and Δpap1 mutants in combination with Δsty1 cells, Atf1 and Pap1 are presumed to have their own target genes regulated in a stress-specific and Sty1-dependent manner (9).

UV light also produces DNA lesions, mainly in the form of 6–4 photoproducts and cyclobutane dimers. Damaged DNA triggers a variety of cellular responses, including damage tolerance, cell cycle arrest and transcriptional activation, which are collectively known as DNA damage checkpoint control (10–12). Genetic studies on budding and fission yeasts have identified many proteins involved in the DNA damage checkpoint pathway. In S.pombe, a set of checkpoint Rad proteins (Rad1, Rad3, Rad9, Rad17, Rad26 and Hus1) and two downstream kinases, Chk1 and Cds1, are essential for proper damage signal transduction and checkpoint controls (13).

Though both the SAPK cascade and the DNA damage checkpoint pathway are activated by UV light, their sensing and triggering mechanisms seem to be quite different. Mammalian JNK is highly activated by UV light in the absence of nuclear DNA, indicating that damaged DNA is not required for the SAPK-mediated UV response (14). Rather, it is suggested that SAPK activation may be induced by a cell membrane-bound sensor or modification of membrane lipids. In contrast, DNA lesions are recognized by a number of checkpoint proteins and these protein–DNA complexes or processed intermediates may trigger activation of the DNA damage checkpoint pathway (12).

In addition to the changes at the level of transcription initiation, alterations in post-transcriptional processes are also known to mediate expression of some stress-inducible genes. For example, expression of the gadd and p21^Waf1 genes was increased by UV light through mRNA stabilization (15,16). Recently, it has been shown that the p38 and JNK SAPKs can induce stabilization of several mRNAs, including interleukin (IL)-6 and IL-8 mRNAs by cytokines (17) and IL-2 mRNA in activated T cells (18). However, the mechanisms by which UV-induced expression occurs post-transcriptionally remain largely unknown and whether SAPKs can contribute to this process is currently one of the major questions in delineation of the cellular responses to UV light.

The uvi15⁺ gene of S.pombe was originally cloned by subtraction hybridization on the basis of its inducibility by UV irradiation (19). However, its transcript level was also increased by other DNA-damaging agents, heat shock and nutrient starvation, which suggested that uvi15⁺ was a stress-inducible gene. Δuvi15 mutants show UV sensitivity and sporulation defects and rapidly lose viability in the stationary phase or under starvation conditions, suggesting that Uvi15 might be involved in cellular protection against UV light and nutrient starvation, as well as in sporulation (20).

In this study, we have focused on elucidating the mechanisms responsible for UV-induced expression of the uvi15⁺ gene in order to better understand the cellular UV response and underlying signaling pathways in eukaryotes. We demonstrate that the SAPK pathway is involved in UV-induced expression of the uvi15⁺ gene at the post-transcriptional level and that an increase in uvi15⁺ mRNA half-life is mediated through an unusual 3′-regulatory element downstream of the poly(A) site. This provides the first evidence that the SAPK Sty1 contributes to UV-induced gene expression by increasing the mRNA half-life.

MATERIALS AND METHODS

Strains, cell culture and transformation

Schizosaccharomyces pombe strain JY741 (h^– ade6-M210 ura4-D18 leu1-32) obtained from Dr M. Yamamoto (University of Tokyo, Tokyo, Japan) was used for integrative transformation. Strains JM1160 (h^– ade6-M216 ura4-D18 leu1-32 sty1::ura4), JM504 (h⁺ ura4-D18 leu1-32 wis1::ura4), JM1529 (h⁺ his7-366 ura4-D18 leu1-32 atf1::ura4) and JM1166 (h^– ura4-D18 leu1-32 pap1::ura4) were gifts of Dr J. Millar (National Institute for Medical Research, London, UK). Strains 1451 (ade6-604 ura4-D18 leu1-32 cds1::ura4 chk1::ura4), 1324 (ade6-704 ura4-D18 leu1-32 rad1::ura4), 1378 (h^– ade6-704 ura4-D18 leu1-32 rad3::ura4), 1161 (ade6-M210/ade6-M216 on Ch16 ura4-D18 leu1-32 rad9::ura4), 941 (h^– ade6-704 ura4-D18 leu1-32 rad17::ura4), 1123 (h^– ade6-704 ura4-D18 leu1-32 rad26::ura4) and Δcds1 (ura4-D18 leu1-32 cds1::ura4) were from Dr A. M. Carr (Sussex University, Falmer, UK). Strain TE484 (h^– ura4-D18 leu1-32 hus1::LEU2) was obtained from Dr T. Enoch (Harvard Medical School, Boston, MA). All of the above strains were used for northern blot analysis. Escherichia coli strain DH5α was used as the host for construction and amplification of plasmid DNAs.

Schizosaccharomyces pombe cells were grown in YE (3% glucose, 0.5% yeast extract), Edinburgh minimal medium (EMM), phosphate-free minimal medium (EMMP) or MB medium supplemented with appropriate amino acids (21,22). Transformation of S.pombe cells was performed as described (21).

Construction of plasmids and deletions

The 4.5 kb EcoRI fragment containing the uvi15⁺ open reading frame (ORF) and its 5′- and 3′-flanking sequences was cloned into the integrative plasmid pJK148 (purchased from American Type Culture Collection, Rockville, MD). The EcoRI–ScaI CAT segment (417 bp) from pCAT-Basic (Promega, Madison, WI) was inserted into the SauI site (blunt ended prior to ligation) in the second exon of uvi15⁺ to produce pJK15CAT (construct I in Fig. 4). This segment was introduced to discriminate between the integrated uvi15⁺ and the original endogenous copy of the uvi15⁺ gene after chromosomal integration at the leu1-32 locus.

Figure 4.

Requirement for the 3′-flanking region beyond the distal poly(A) site for UV-induced expression. (A) uvi15⁺ genomic DNA was trimmed from both the 5′- and 3′-ends and integrated at the leu1-32 locus on the chromosome to develop a single copy background. A 417 bp CAT segment was inserted into the second exon of uvi15⁺ to discriminate between the integrated uvi15⁺ and the original endogenous copy of uvi15⁺ (left). Cells bearing each construct on their chromosome were exposed to UV light and post-incubated for 1 h. Total RNA was extracted and analyzed by northern blot using the CAT segment as probe. The transcript levels from each construct were normalized to act1⁺ levels and relative fold induction by UV light was calculated from three independent experiments (right). (B) A representative blot is shown. The Roman numerals correspond to each construct illustrated in (A). C, control; U, UV irradiated.

Within the 5′-flanking region of the uvi15⁺ gene, deletions to –2112 and –1244 (relative to the translation initiation site) were made by XbaI or BamHI–BglII digestion and self-ligation to produce pJK15CATX and pJK15CATB (constructs II and III in Fig. 4, respectively). For deletions in the 3′-flanking region of the uvi15⁺ gene, restriction endonucleases BstXI, EcoRV, NdeI and EcoRI were used to obtain deletions to +1747, +1590, +1423 and +1186, respectively. To create fine deletions between +1423 and +1186, primer 15BN2 (5′-AGCTTCATTTTCGTGAAC-3′) or primer 15BN1 (5′-CGTAAGATATATGCCTCA-3′) was utilized with primer 15BGL2 (5′-GGTCGAATAATTATAGATCTTG-3′) to generate 621 and 512 bp PCR products using pJK15CATB as template. After TA cloning into pCR2.1 (Invitrogen, San Diego, CA), the BglII–HincII fragments from these clones were replaced with the BglII–HincII fragment of pJK15CATB to produce 3′-deletions to +1348 and +1239, respectively.

To generate an internal deletion from +1186 to +1239, the overlap extension technique was employed using the pJK15CATB template as described by Ho et al. (23). PCR products from two initial sets [produced with primers 15BGL2 + 15B-5BT (5′-AA-CTTTTTAGAATTCTTATACGTATCGATCAAAAACCG-3′) and primers T7 + 15BN1-5UP (5′-TAAGAATTCTAAAAAGTTTAAAGTTTCTAAGA-3′)] were combined and subjected to a second primer extension reaction using the external primers 15BGL2 and T7. After TA cloning, the BglII–HincII fragment devoid of the segment from +1186 to +1239 was replaced with the original BglII–HincII fragment of pJK15CATB. An internal deletion from +1239 to +1348 was achieved by the same strategy with primers Δ3948-BT (5′-AACCACGCATCGTAAGATATATGCCTCAAT-3′) and Δ3948-UP (5′-ATATCTTACGATGCGTGCTTTTTCTATGGT-3′). All deletion constructs were verified by DNA sequencing (24) and integrated into the leu1-32 locus of the S.pombe chromosome using the integrative plasmid pJK148. Single copy integration was confirmed by Southern blot analysis.

To generate the rhp51::ura4 hybrid, a BamHI–NheI fragment of the rhp51⁺ gene was ligated with the StuI–HindIII fragment of the ura4⁺ gene (construct 514 in Fig. 8). Deletion of sequences downstream of the poly(A) site was performed by PCR with primers 51BAMH1 (5′-CCCGGGGATCCTTTACCAGTA-3′) and ura43D (5′-CAAATTACTTTGAATTCCCAAG-3′) to create construct 5143D. Insertion of the 54 bp element into the EcoRI site of the 3′-end of 5143D produced construct 5143D-54.

Figure 8.

UV-induced increase in the half-life of rhp51::ura4 hybrid mRNA by the 54 nt element. (A) Structures of the rhp51::ura4 hybrid constructs. Half of the ORF and the 3′-flanking region of the rhp51⁺ gene were replaced with those of the ura4⁺ gene. Sequence elements directing 3′-end formation of ura4⁺ mRNA are symbolized by small stippled boxes. The 54 nt element of uvi15⁺, denoted by the hatched box, was inserted as a single copy in the sense orientation immediately downstream of the first poly(A) site in construct 5143D-54. (B) Decay of hybrid mRNAs in wild-type cells. Cells bearing each construct were either untreated (CON) or UV irradiated (UV). After post-incubation for 50 min, 1,10-phenanthroline was added to a final concentration of 100 µg/ml and aliquots of cells were removed at the indicated time points for total RNA isolation. Equal amounts of RNA were subjected to northern blot analysis and probed for rhp51::ura4 hybrid mRNA. The asterisk denotes the correctly processed mRNA transcripts, whereas the arrowhead indicates the read-through transcript. Construct 514 produces two different mRNA species because it has two poly(A) sites in the 3′-flanking region. The signals on the northern blot were quantified and normalized to the loading control and mRNA half-lives of each hybrid mRNA were determined by plotting the percentage of relative mRNA amounts remaining versus time of 1,10-phenanthroline addition. (C) Plot under normal conditions. (D) Plot under UV-irradiated conditions. (E) Efficient mRNA 3′-end formation by the 54 nt element. Percentage read-through generation was calculated as the amount of read-through relative to the total amount of RNA specific for the probe.

UV irradiation and northern blot analysis

Schizosaccharomyces pombe cells grown to early exponential phase (OD₅₉₅ ≈ 0.5, ∼1 × 10⁷ cells/ml) were harvested, washed and resuspended in distilled water to a final density of 2 × 10⁸ cells/ml. The cell suspension was spread onto a glass Petri dish (150 mm diameter) and then exposed to 254 nm UV light from a Stratalinker1800 (Stratagene, La Jolla, CA) at 240 J/m². After UV irradiation, the cells were collected, resuspended in fresh medium at the original culture density and incubated at 30°C in the dark for 1 h. Aliquots of cells were withdrawn from which total RNA was isolated by extraction with phenol/chloroform/SDS as described by Jang et al. (25). About 10–25 µg total RNA was separated by electrophoresis through a 1.5% agarose gel containing 0.67 M formaldehyde, blotted onto nylon membrane and hybridized with radiolabeled probes. After stringent washes, the blot was exposed to X-ray film or a phosphorimager (BAS1500; Fuji, Tokyo, Japan) and the relative fold UV induction of uvi15⁺ transcripts was calculated and normalized to the loading control.

Transcription run-on assay

One-hundred milliliter cultures were grown overnight in YES to an OD₅₉₅ of 0.5 and subjected to UV irradiation as described above. After post-incubation for 50 min at 30°C in the dark, transcription run-on assays were performed as described by Hansen et al. (26). About 2 × 10⁶ c.p.m. of labeled RNA was partially hydrolyzed in 0.2 N NaOH for 6 min on ice, neutralized (0.2 M Tris–HCl, pH 7.2) and used for hybridization.

Analysis of mRNA half-life by steady-state labeling with ³²PO₄

Cells were grown in EMMP (remove Na₂HPO₄ and replace potassium hydrogen phthalate with 1.64 g/l sodium acetate, adjusted to pH 5.5 with acetic acid) with 1 mM phosphate (added from a 0.5 M NaH₂PO₄ stock solution) at 30°C to an OD₅₉₅ of 0.5. After mock or UV irradiation, the cells were resuspended at 2 × 10⁷ cells/ml in fresh EMMP containing 50 µM phosphate and 100 µCi/ml carrier-free [³²P]phosphoric acid (NEX054; NEN, Boston, MA) and incubated at 30°C in the dark. Aliquots of the culture were removed after 10–50 min labeling. Total RNA was extracted from frozen cell pellets and poly(A)⁺ RNA was purified with an Oligotex mRNA isolation kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. To measure the decay rates of uvi15⁺ mRNA, 0.2 µg poly(A)⁺ RNA from each time point was heated to 65°C for 5 min and hybridized with denatured plasmid DNA (10 µg/slot) containing uvi15⁺ cDNA or the EcoRI–ScaI CAT segment (417 bp) as an insert. RNA decay curves were generated by plotting 1 – A/A^_∞ versus time of ³²PO₄ labeling (where A is the specific activity of an RNA at time t and A^_∞ is the specific activity at time ∞; 27). As an approximation, the specific activity of the 40 min sample was used as A^_∞.

DNA slot blots

Plasmid DNAs were denatured by incubation with 0.1 vol 1 N NaOH for 30 min at 37°C and neutralized by adding 4 vol 6× SSC. Excess amounts of DNA (10 µg denatured DNA/slot) were filtered onto nylon membranes through a multi-slot filtration manifold and immobilized by UV cross-linking. Hybridization was carried out as described (26).

Reverse transcription and PCR analysis (RT–PCR)

About 5 µg total RNA from S.pombe was reverse transcribed with 200 U Superscript II reverse transcriptase (Life Technologies, Gaithersburg, MD) at 42°C for 50 min in the presence of 10 pmol of the following antisense primers which span the 3′-downstream region of the uvi15⁺ gene: primer 1 (5′-GTAGAAAGAACAAGCCAAATAG-3′), primer 2 (5′-CGAGTCATTTCAACTAATCC-3′), primer 3 (5′-CTAAAGAGTTCATCTACAC-3′), primer 4 (5′-AAACAAATAAGGAATTCTTATAC-3′), primer 5 (5′-CGTAAGATATATGCCTCA-3′) and primer 6 (5′-CTATCTACCTGTTTGAAAT-3′). The positions of primers 1–6 are illustrated in Figure 7A. One-tenth of the reverse transcribed cDNA products was then amplified by PCR (30 amplification cycles) with 10 pmol of a uvi15⁺-specific 5′-primer (5′-CAATTATTTAAGGATCCAAATGTCTGCTC-3′) and the corresponding antisense 3′-primer using 2.5 U ExTaq (TaKaRa, Shiga, Japan). Various RT–PCR products were electrophoresed on a 1.2% agarose gel and visualized by ethidium bromide staining.

Figure 7.

Detection of uvi15⁺ transcript downstream of the poly(A) site. (A) The relative position of each primer used for RT–PCR analysis is illustrated with the deduced size of their products. The hatched box denotes the 54 nt element. Note that the figure is not drawn to scale. (B) RT–PCR products generated by 30 amplification cycles were separated on a 1.2% agarose gel. The lane number corresponds to each antisense 3′-primer shown in (A). No product was observed when RT was omitted during cDNA synthesis (lane C), which means that the RT–PCR products are not due to DNA contamination. The upper small band in lane 5 is a PCR artifact. M, 1 kb plus DNA ladder marker (Life Technologies).

Measurement of mRNA decay using 1,10-phenanthroline

Schizosaccharomyces pombe cells incubated at 30°C in EMM to OD₅₉₅ ≈ 0.5 were subjected to UV irradiation as described above. After post-incubation at 30°C in the dark for 50 min, 1,10-phenanthroline was added at 100 µg/ml to the post-incubation medium and aliquots of cells were removed at the 0, 10, 20, 30, 40, 50 and 60 min time points and quick frozen in liquid nitrogen. Total RNA was isolated and analyzed by northern blot. RNA decay curves were generated by plotting the percentage of relative mRNA amounts remaining versus the time of 1,10-phenanthroline treatment on a semi-log scale.

RESULTS

UV-induced expression of uvi15⁺ mRNA is dependent on the Wis1–Sty1 SAPK pathway

As described in the previous study, the transcript level of the uvi15⁺ gene was rapidly increased within 1 h after UV irradiation, reaching its maximum at a UV dose of 240 J/m² (19). Accordingly, 240 J/m² and 1 h post-incubation were employed throughout this study as the conditions to induce uvi15⁺ gene expression. To gain an insight into the mechanism governing UV inducibility, we analyzed the UV-responsiveness of the uvi15⁺ gene in cells lacking either the stress-activated protein kinase cascade or DNA damage checkpoint control, which was previously shown to invoke cellular responses to UV light leading to altered patterns of gene expression in S.pombe (6,8,13). Cells were UV irradiated or left untreated and total RNA was extracted to detect the uvi15⁺ transcript by northern blot analysis. As shown in Figure 1A, the degree of UV induction was significantly decreased in Δsty1 and Δwis1 cells, which are defective in the SAPK cascade. Two different species of uvi15⁺ transcript were observed which stemmed from different polyadenylation sites in its 3′-end (20). In contrast, UV inducibility was not reduced in several DNA damage checkpoint mutants (Δrad1, Δrad3, Δrad9, Δrad17, Δrad26, Δhus1, Δcds1 and Δcds1Δchk1), as shown in Figure 1B. These results imply that UV-inducible expression of the uvi15⁺ gene is regulated by the SAPK cascade rather than the DNA damage checkpoint pathway. Interestingly, Δatf1 and Δpap1 cells showed similar uvi15⁺ induction levels to wild-type cells (Fig. 1A), indicating that UV induction of uvi15⁺ transcripts is not mediated through Atf1 and/or Pap1 in the SAPK cascade.

Figure 1.

uvi15⁺ mRNA expression following UV irradiation. (A) Exponentially growing cells with a defect in the SAPK cascade were exposed to 240 J/m² UV light and post-incubated for 1 h in the dark. Total RNAs were extracted and uvi15⁺ mRNA expression was assessed by northern blot analysis. The relative fold UV induction was calculated and normalized to act1⁺ levels. (B) Cells with a defect in the DNA damage checkpoint control were subjected to the same procedure as in (A).

uvi15⁺ transcription rate is increased by UV light in a Sty1-independent manner

To directly determine whether the rate of transcription was increased following UV irradiation, transcription run-on assays were performed. Prior to the assays, northern blot analysis confirmed that detergent permeabilization did not induce uvi15⁺ expression or disrupt UV-inducible characteristics (data not shown). As shown in Figure 2, UV light increased the transcription rate of uvi15⁺ by ~1.8-fold in the wild-type (normalized to the level of act1⁺). In Δsty1 cells, a similar fold induction was maintained, although the signals were weak in both control and UV-irradiated samples. These findings indicate that UV light elevates the rate of uvi15⁺ transcription independently of the SAPK Sty1.

Figure 2.

Transcription run-on assay of wild-type or Δsty1 cells. (A) Wild-type (WT) or Δsty1 cells were either untreated (CON) or UV irradiated (UV). After post-incubation for 50 min, cells were permeabilized and used for run-on assay as described in Materials and Methods. About 2 × 10⁶ c.p.m. of labeled RNA was used for hybridization to detect uvi15⁺ and act1⁺ signals. (B) Increases in uvi15⁺ transcription rate induced by UV light were quantified in wild-type and Δsty1 cells. After removing non-specific hybridization signals (pBluescript), uvi15⁺ signals were normalized to act1⁺ signals and relative fold UV induction was calculated.

The half-life of uvi15⁺ mRNA is enhanced by UV light but reduced in Δsty1 cells

Since the uvi15⁺ gene showed >3-fold UV induction at the transcript level, a 1.8-fold increase in the transcription rate did not fully account for the increased steady-state level of uvi15⁺ mRNA. Hence, we determined whether an increase in mRNA half-life might be associated with uvi15⁺ induction by UV light. We employed the steady-state labeling method using ³²PO₄, because 1,10-phenanthroline, a widely used transcription inhibitor in yeast, was found to induce uvi15⁺ expression, as expected from a previous study (28).

Wild-type or Δsty1 cells grown in EMMP were labeled with ³²PO₄ after mock or UV irradiation and total and poly(A)⁺ RNAs were purified from cells after 10–50 min labeling. Approximately 0.2 µg labeled poly(A)⁺ RNA from each time point was hybridized with denatured uvi15⁺ cDNA on slot blots and the signals were visualized and quantified (Fig. 3A and C). Decay rates of uvi15⁺ mRNA were measured by plotting 1 – A/A_^∞ versus time of labeling on a semi-log scale, where A and A_∞ are the specific activities of RNA at times t and ∞, respectively (27). As shown in Figure 3B, in the wild-type the half-life of uvi15⁺ mRNA was enhanced by UV light. In control and UV-irradiated cells, it was ∼18 and 27 min (1.5-fold increase), respectively. However, this phenomenon was not found in Δsty1 cells (Fig. 3D). Rather, the half-lives of both basal and UV-induced transcripts were decreased to ∼13 min. These results suggested that the SAPK Sty1 might play a role in the UV-induced increase in uvi15⁺ mRNA half-life.

Figure 3.

Measurement of uvi15⁺ mRNA half-life by steady-state labeling. (A) Wild-type cells grown in EMMP with 1 mM phosphate were either left untreated (CON) or exposed to UV light (UV). After resuspension in fresh EMMP containing ³²PO₄ (100 µCi/ml), aliquots of the culture were removed after 10–50 min. An equal amount (0.2 µg) of poly(A)⁺ RNA purified from each time point was hybridized with denatured uvi15⁺ cDNA (10 µg/slot) immobilized on nylon membrane. A representative example of two independent experiments is shown. (B) The mRNA half-life of uvi15⁺ in wild-type cells with or without UV irradiation was determined by plotting 1 – A/A_^∞ versus time of ³²PO₄ labeling, where A is the specific activity of an RNA at time t and A_^∞ is the specific activity at time ∞. (C and D) Effect of deletion Δsty1 on uvi15⁺ mRNA half-life. The same experimental procedure was employed with Δsty1 cells in two independent experiments.

A 3′-flanking region beyond the distal poly(A) site is required for UV-induced expression of the uvi15⁺ gene

In an attempt to identify the regulatory element(s) involved in UV-induced uvi15⁺ gene expression, it was necessary to identify the minimal genomic boundaries required for induction. A cloned genomic DNA covering the uvi15⁺ ORF was tagged with a 417 bp segment from the chloramphenicol acetyltransferase (CAT) gene at the SauI site in the second exon to distinguish it from the endogenous uvi15⁺ transcript. After trimming with suitable restriction endonucleases from both the 5′- and 3′-ends (Fig. 4A), the uvi15⁺ genomic DNA constructs were integrated at the leu1-32 locus of the S.pombe chromosome in order to develop a stable single copy background using plasmid pJK148 as described by Keeney and Boeke (29). The single copy integrants were confirmed by Southern blot analysis (data not shown). When the integrants were assayed by northern blot analysis using the CAT segment as probe, it was surprisingly found that the 3′-flanking region from +1186 to +1883 bp beyond the distal poly(A) site was required for UV-induced expression in addition to the 5′-upstream region spanning up to –1244 bp from the translation initiation site (+1) (Fig. 4A and B). Two uvi15⁺ transcripts nearly co-migrated due to the increase in size with the 417 bp CAT segment.

To define the cis-acting element(s), a series of deletions was made in the 3′-flanking region by suitable restriction digestion and PCR and were chromosomally integrated as before. When the deletion reached +1186 bp (construct X in Fig. 5), UV induction was barely observed (Fig. 5A and B). Further evidence for the presence of a regulatory element between +1186 and +1239 bp was provided by PCR-based internal deletion analysis. The construct devoid of this region (construct XI in Fig. 5) showed markedly reduced UV induction (Fig. 5A and B). These results reveal that the 3′-flanking region located immediately downstream of the distal poly(A) site serves as a major regulatory element for UV induction of the uvi15⁺ gene. Interestingly, sequences within this 54 bp region showed nearly perfect dyad symmetry (Fig. 5C).

Figure 5.

Deletion analysis of the 3′-flanking region. (A) A series of deletions was generated in the 3′-flanking region of the uvi15⁺ gene by restriction digestion and PCR. After chromosomal integration, each integrant was subjected to UV irradiation and northern blot analysis. The intact 5′-portion of each construct (up to –1244 bp from the first ATG) was omitted in the schematic diagrams for convenience (left). The transcript levels from each construct were normalized to act1⁺ levels and relative fold induction by UV light was calculated (right). (B) A representative blot is shown from three to six independent experiments. The Roman numerals correspond to each construct illustrated in (A). C, control; U, UV irradiated. (C) Sequences between +1186 and +1226 bp show nearly perfect dyad symmetry.

The 54 nt element increases the half-life of uvi15⁺ mRNA at the primary RNA level

To address the role of this element in UV-induced expression, transcription run-on assays and steady-state labeling with ³²PO₄ were performed using cells bearing construct III or XI in Figure 5, where the 54 bp between +1186 and +1239 bp was intact or internally deleted, respectively. Normal increases in transcription rate induced by UV light were observed despite the internal deletion (Fig. 6A). However, a UV-induced increase in uvi15⁺ mRNA half-life was not detected in this internally deleted construct, while the intact construct produced an increased half-life comparable to that found in the original endogenous copy of uvi15⁺ (Fig. 6B and C; see also Fig. 3B). These results indicate that this element contributes to UV-induced expression of the uvi15⁺ gene by a post-transcriptional mechanism.

Figure 6.

Effect of internal deletion on the transcription rate and mRNA half-life of uvi15⁺. Transcription run-on assays and steady-state labeling were performed on cells in which the 54 nt element was intact or internally deleted (constructs III and XI in Fig. 5A, respectively). (A) About 2 × 10⁶ c.p.m. of labeled RNA from the transcription run-on assay was hybridized with denatured CAT and act1⁺ probes. CAT signals (i.e. integrated uvi15⁺ signals) were normalized to act1⁺ signals and relative fold UV induction was calculated. (B) The decay rate of the mRNA derived from the intact construct was plotted as described in Materials and Methods. The CAT probe was used to detect signals from the chromosomally integrated construct. (C) The decay rate of the mRNA derived from the internally deleted construct was plotted.

Unlike other regulatory elements involved in post-transcriptional control, this 54 nt element is not included in the mature uvi15⁺ mRNA, since it resides outside the distal poly(A) site (as confirmed by the 3′-RACE PCR technique; data not shown). However, recent studies in S.pombe show that RNA polymerase II transcription proceeds beyond the poly(A) site and that the sequences therein provide information required for transcription termination and RNA 3′-end formation, which are closely coupled to efficient gene expression (30–32). Therefore, we analyzed whether this 54 nt element could be transcribed into primary RNA transcripts by RT–PCR. Total RNA was hybridized to antisense 3′-downstream primers 1–6 and was reverse transcribed (Fig. 7A). The converted cDNA products were subsequently amplified by PCR. As shown in Figure 7B, transcripts were detected at least 112 nt downstream of the distal poly(A) site where the 54 nt element resides. The same degree of 3′-extension of the transcript was found irrespective of UV irradiation (data not shown). This raised the possibility that the 54 nt element might exert an effect on UV-induced expression at the level of the primary RNA transcript.

The 54 nt element confers a UV-induced increase in half-life of a heterologous mRNA

To analyze whether the 54 nt element could manifest its ability within other genetic contexts, a rhp51::ura4 hybrid was utilized as the reporter mRNA, where half of the ORF and the 3′-flanking region of the S.pombe rhp51⁺ gene were replaced with those of the ura4⁺ gene to produce an in-frame fusion (Fig. 8A). This was chosen for the following reasons. First, the rhp51⁺ promoter is UV-inducible, so that the level of transcripts derived from it was not down-regulated by UV light, leading to easy monitoring of the decay kinetics. Second, the rhp51⁺ promoter is not induced by 1,10-phenanthroline, a transcription inhibitor in yeast. This feature enabled us to use this drug rather than ³²PO₄ during the RNA decay study. Finally, the 3′-flanking region of the ura4⁺ gene has been intensively studied and the locations of poly(A) sites and mRNA 3′-end-forming signals have been exactly identified (30). Thus, using the rhp51::ura4 hybrid with a modified 3′-flanking region, it was possible to confirm whether the 54 nt element really could function at the level of primary RNA transcripts.

The 54 nt element was inserted as a single copy in the sense orientation immediately downstream of the poly(A) site of the rhp51::ura4 hybrid to produce a similar situation as in the uvi15⁺ gene context (construct 5143D-54 in Fig. 8A). Constructs 514, 5143D-54 and 5143D were introduced into wild-type cells. After mock or UV irradiation, cells were post-incubated for 50 min to induce transcription of hybrid transcripts. 1,10-Phenanthroline was then added to the medium (EMM) to inhibit further RNA synthesis. Total RNAs were extracted at the indicated time points and analyzed by northern blot. The half-life of mRNA derived from 5143D-54 was greatly increased by UV irradiation (Fig. 8B–D). The level of transcripts at 60 min after 1,10-phenanthroline treatment remained over 90% of the level at 0 min. The increase in the half-life of 514 hybrid transcripts induced by UV irradiation seems to derive from the rhp51⁺ part of the construct, because 514 transcripts show nearly identical decay kinetics to rhp51⁺ transcripts (Fig. 8D). However, the UV-induced increase in mRNA half-life was further augmented by introduction of the 54 nt element outside the poly(A) site, demonstrating that the 54 nt element of the uvi15⁺ gene really increases mRNA half-life at the primary RNA level.

RNAs derived from construct 5143D were shown to contain significant numbers of read-through transcripts due to lack of proper information for utilization of poly(A) sites, as mRNA 3′-end formation in S.pombe seems to be directed by downstream sequences beyond the poly(A) site (30). As shown in Figure 8E, these formed ∼40% of total 5143D RNA transcripts. Insertion of the 54 nt element downstream of the poly(A) site greatly reduced the occurrence of read-through transcripts (to below 10%), which means that the 54 nt element has proper 3′-end-forming signals for S.pombe mRNAs.

DISCUSSION

The major aim of our study was to elucidate the mechanism(s) underlying UV-induced expression of the uvi15⁺ gene in the fission yeast S.pombe whose core properties with respect to UV response are known to be similar to those of mammals. From this approach we hoped to ultimately gain a more comprehensive view of the eukaryotic UV response.

The SAPK pathway and UV-induced expression of the uvi15⁺ gene

Recent studies showed that the mammalian UV response is not triggered by DNA damage (14,33). Instead, sensors in the cell membrane transduce signals to downstream kinases, which, in turn, activate further downstream targets to induce the UV response (34). SAPKs are at the core of this signaling and among them are p38 kinase and the JNKs, which have been studied in great detail. Schizosaccharomyces pombe Sty1 resembles mammalian p38 kinase and is known to be involved in a variety of stress responses. Our observation that the UV-induced increase in uvi15⁺ mRNA level was affected by the Wis1–Sty1 cascade but not by the DNA damage checkpoint controls presents another example of similarity between mammalian and fission yeast UV responses.

Although Sty1-dependent stress responses (including the UV response) mostly result via Atf1 and Pap1 through enhanced transcription, Δatf1 and Δpap1 cells show normal uvi15⁺ induction by UV light. Sty1 kinase appears to regulate uvi15⁺ expression post-transcriptionally (Fig. 3C and D) instead of via Atf1 and Pap1. The rapid kinetics of induction of uvi15⁺ mRNA and the lack of a requirement for de novo protein synthesis (assayed with cycloheximide; data not shown) imply that when cells are exposed to UV light Sty1 rapidly activates a pre-existing protein factor(s), which then increases the half-life of uvi15⁺ mRNA. However, this strategy does not seem to be universal for other types of inducers of uvi15⁺ mRNA, since Sty1 regulates methyl methanesulfonate-induced expression without affecting mRNA half-life (unpublished results).

Many researchers have recently reported that pro-inflammatory cytokines, bacterial lipopolysaccharide and a range of cellular stresses can enhance mRNA levels post-transcriptionally via SAPK pathways in mammalian cells (17,18,35,36). However, to the best of our knowledge, this is the first study that provides evidence that SAPK also post-transcriptionally contributes to UV-induced expression in eukaryotic cells. This finding is important in that studies on UV-inducible gene expression in eukaryotes thus far have focused on enhanced transcription mediated by transcription factors. Moreover, our results may provide clues to UV-induced expression of other eukaryotic genes whose mechanisms remain largely unknown.

The role of the 54 nt element in UV-induced expression

Deletion analyses of the 3′-flanking region of the uvi15⁺ gene revealed a sequence element necessary for UV-induced expression (Figs 5 and 6). This 54 nt element is unique in its ability to function post-transcriptionally downstream of the distal poly(A) site. As this element was transcribed into the primary RNA transcripts irrespective of UV irradiation (Fig. 7), the 54 nt element is thought to exert its effect at the level of primary RNA transcripts. This idea was confirmed by studies using a rhp51::ura4 hybrid reporter mRNA. Insertion of the 54 nt element downstream of the poly(A) site greatly increased the mRNA half-life upon UV irradiation (Fig. 8). This transferable nature suggests the possibility that the strategy shown here may not be restricted to the uvi15⁺ gene. Probably, it may be involved in post-transcriptional regulation of other UV-inducible genes, though some examples should be found in future studies. The notion that the sequences within primary RNA transcripts regulate gene expression post-transcriptionally is supported by recent studies on the human N-myc gene (37). Although the exact mechanism is not yet fully understood, it was suggested that a 116 nt element within the first intron directed tissue-specific expression by selective (de)stabilization of the primary N-myc transcripts.

Sequences within the 54 nt element bear no significant homology to known post-transcriptional regulatory elements, although the AU-rich content is similar to that of AU-rich elements (AREs) found in the 3′-untranslated regions of many labile mRNAs encoding proto-oncogenes and cytokines (38–40). The 54 nt element also shows nearly perfect dyad symmetry which could form a RNA hairpin loop structure to which trans-acting proteins may bind. In fact, a stem–loop structure and its binding protein have been shown to be required for nucleocytoplasmic transport and regulated mRNA degradation of replication-dependent histone genes (41). Whether a binding factor(s) really exists and whether the 54 nt element is able to form the proposed secondary structure remain to be elucidated in a future study.

The proximity of the 54 nt element to the poly(A) site suggests another possible role in transcriptional pausing and/or 3′-end formation of mRNA. According to studies by Proudfoot and colleagues, sequences downstream of the poly(A) site provide information for proper 3′-end formation of S.pombe mRNA (30–32). Consistent with their studies, read-through transcripts were found in a construct whose entire sequences downstream of the poly(A) site were deleted (construct 5143D in Fig. 8E). However, the percentage of read-through transcripts greatly decreased to below 10% in 5143D-54 with the 54 nt element at the 3′-end of the poly(A) site. A similar phenomenon was also reproduced in the uvi15⁺ context (data not shown). These findings indicate that the 54 nt element also acts as a 3′-end-forming signal for S.pombe mRNA. It is quite interesting that a single 54 nt element is able to increase mRNA half-life in a UV-inducible manner and to direct mRNA 3′-end formation.

Possible mechanisms of action of the 54 nt element

It is possible to postulate two models for the effect of the 54 nt element on mRNA half-life: (i) regulation of primary RNA stability and its processing; (ii) production of more stable mRNA in the course of pre-mRNA processing. In the first model, the role of the 54 nt element would be to stabilize the primary transcript and to regulate the efficiency of its processing to mRNA in a UV-inducible manner during the extensive post-transcriptional processing, leading to enhanced accumulation of mature mRNA in the steady-state. It may be achieved through the formation of putative secondary structure and/or efficient RNA 3′-end formation by the 54 nt element. If the primary RNA transcripts exist more stably by this mechanism and if they are more efficiently converted to mature mRNAs, the increase in half-life observed in Figures 3B, 6B and 8B might derive from the accumulation of mRNA. However, if the changes in half-life are caused by structural differences between basal and UV-induced uvi15⁺ mRNAs, the first model is inadequate to explain this possibility. Alternatively, the 54 nt element, possibly with its binding counterpart(s), may modulate some of the post-transcriptional processing steps, resulting in mature mRNA with a longer half-life in UV-irradiated cells. Conceivably, regulation of the length of the poly(A) tail would be the simplest way to achieve this goal (42).

It seems likely that the 54 nt element may be the most downstream effector of the Sty1 signaling cascade leading to an increase in mRNA half-life for the following reasons. Firstly, the positive effect of the 54 nt element was completely abolished in a Δsty1 background. Secondly, in the absence of this element, Sty1 could not increase mRNA half-life upon UV exposure. The role of the 54 nt element, however, appears to be limited to the UV-induced increase via Sty1 signaling, because only the UV-induced uvi15⁺ mRNA half-life was affected by removal of the 54 nt (Fig. 6C).

Future studies are underway to elucidate the exact mechanism of action of the 54 nt element and its integration with Sty1-mediated signaling. In addition, identification of a putative binding counterpart(s) for the 54 nt element will allow us to determine new downstream targets of Sty1 kinase involved in the UV response.