nmtl of Fission Yeast A HIGHLY TRANSCRIBED GENE COMPLETELY REPRESSED BY THIAMINE*

The first fully repressible gene in fission yeast is described. In minimal medium it is highly transcribed producing a mRNA which is 50-100 times more abun- dant than the cycl mRNA. By contrast, in minimal medium supplemented with thiamine at a concentration of 0.5 PM or greater the transcript is undetectable. The gene has been called nmtl (for no message in thiamine). The 5’ and 3’ ends of the transcript have been mapped and indicate an unspliced mRNA of 1.3 kilobases with no evidence of heterogeneity at either end. The single major open reading frame encodes a protein of 39 kDa. The gene product is most likely involved in thiamine biosynthesis and consistent with this is the observation that the nmtl::uru4 disruption strain is a thiamine auxotroph. The kinetics of tran- scriptional repression and induction have been studied. Addition of thiamine to log phase cells growing minimal medium results in complete disappearance of the nmtl message within 3 h. Removal of thiamine from the medium produces the first detectable message after 10 h and maximal steady-state levels after 16 h. Nu- clear “run on” experiments demonstrate that control is exerted at the level of has been subcloned, and thiamine-me-diated transcriptional control has been transferred to the bacterial reporter gene chloramphenicol acetyl- transferase. Transcription initiation

The first fully repressible gene in fission yeast is described.
In minimal medium it is highly transcribed producing a mRNA which is 50-100 times more abundant than the cycl mRNA.
By contrast, in minimal medium supplemented with thiamine at a concentration of 0.5 PM or greater the transcript is undetectable. The gene has been called nmtl (for no message in thiamine).
The 5 Transcription initiation is a major control point in the expression of eukaryotic genes. Loss of transcriptional control can have profound consequences for the cell including uncontrolled proliferation and oncogenic transformation. Not surprisingly this process has been the subject of intensive investigation in recent years.
The yeast Saccharomyces cereuisiae has proved to be a valuable experimental model. In general, the organization of yeast promotors is similar to those of higher eukaryotes. A TATA box is usually found upstream of the mRNA initiation site albeit at a variable distance of between 40 and 120 base pairs, and upstream regulatory elements, analogous to eukaryotic enhancers, have been shown to be the specific targets for positive and negative regulatory factors (for a review see Guarente, 1987). The availability of a variety of inducible genes, and the ease with which the genetic and molecular genetic approaches can be applied in yeast has led to the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 USC. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 505493.
identification of cis-and trans-acting elements in a number of instances.
Particularly revealing has been the analysis of genes involved in amino acid biosynthesis which are co-ordinately controlled in response to amino acid starvation (Delforge et al., 1975). The upstream region of each of these genes contains a g-base pair palindromic sequence Sd-ATGA(C/G)TCAT-3' (Donahue et ai, 1983) which subsequent analysis has shown to be a target for the positive tram-activating protein GCN4 (Hill et al., 1986). The GCN4 protein contains two functionally distinct domains, the C-terminal region involved in DNA binding and an internal acidic region which activates the transcriptional machinery . Curiously, the DNA-binding domain shows unexpectedly high amino acid sequence homology with the DNA-binding domain of the jun oncoprotein (Vogt et al., 1987), a component of the transcription activation complex of HeLa cells, AP-1. Moreover, AP-1 recognizes a O-base pair target sequence almost identical to the sequence recognized by GCN4 (Angel et al., 1987;Lee et al., 1987), and in an appropriate protein chimera the jun protein is capable of regulating the transcriptional activity of the amino acid biosynthetic genes in yeast (Struhl, 1987). The human oncogene, fos, is another transcriptional activator in human cells which binds DNA by associating with AP-1 complex (Chiu et al., 1988). fos is also capable of stimulating transcription of yeast genes when linked to an appropriate DNA-binding motif (Lech et ai., 1988). From these data it seems probable that much of the basic machinery of transcriptional control has been highly conserved throughout eukaryotic evolution.
In Schizosaccharomyces pombe, similarity with the transcriptional machinery of higher eukaryotes is if anything even more pronounced. The TATA element is located a more uniform 25-30 base pairs upstream of the start of transcription (Russell, 1983) exactly as in mammalian cells (Benoist et al., 1980); spliced genes are more frequent and the splice site consensus CTNAC (Hindley and Phear, 1984) located near the 3' end of the intron is a degenerate version of the TACTAAC box found in S. cerevisiae (Langford and Gallwitz, 1983) and similar to the consensus found in higher eukaryotes (Mount, 1982). Indeed the SV40 large T precursor is correctly spliced and makes functional protein in fission yeast (Kaufer et al., 1985). S. pombe also has an AP-l-like factor which gives an identical footprint on SV40 DNA to human AP-1 (Jones et al., 1988). Up to now, however, a more detailed analysis of transcriptional control in fission yeast has not been possible because of a lack of known transcriptionally regulateable genes in this organism. Thus far only phol and p/w4 which code for two acid phosphatases have been isolated, however, neither has an abundant transcript and neither is completely repressed in the uninduced state (Maundrell et al., 1985a;Scweingruber et al., 1986). In the present study I report the isolation of nmtl, the first fully regulateable gene in S. pombe.

nmtl of Fission Yeast
It is among the very highly expressed genes in this organism and is totally repressed in medium containing thiamine.

Media and General
Techniques-Yeast media and transformation were as described by Beach and Nurse (1981). For bacteria, standard media were used throughout. CAT' activity was assayed as Jones et al. (1988). General procedures used in cloning and blotting are described in Maniatis et al. (1982).

Preparation of RNA and Northern
Blotting-Total RNA was prepared as described by Maundrell et al. (1985a) and resuspended in TE at 2 mg/ml. For Northern blotting, 10 Fg of RNA was denatured in 0.5 M glyoxal and 50% MeZSO and run in 1.2% agarose containing 25 mM phosphate buffer, pH 6.5. RNA was transferred to Biodyne B in 20 X SSC (1 X SSC contains 0.15 M NaCl, 0.015 M sodium citrate) by capillary blotting overnight. The membrane was prehybridized in 50% formamide, 5 x Denhardt's (1 x Denhardt's contains 0.02% bovine serum albumin, 0.02% Ficoll, 0.02% polyethylene glycol), 1 M NaCl, 0.1% SDS,' 5 mM EDTA, 100 pg/ml calf thymus DNA at 42 "C for 2 h and hybridized overnight in the same solution containing '*Plabeled probe made by random priming gel-purified restriction fragments (Feinberg and Vogelstein, 1984) as described in the legends to figures. cDNA Probes and Differential Screening of the Genomic Library-Wild-type S. pombe 972h was grown to mid log phase (8 X lo6 cells/ ml) in 200 ml of minimal medium or 200 ml of minimal medium containing 2 pM thiamine and RNA was prepared. Polyadenylated mRNA from each preparation was selected on oligo(dT) cellulose (Maniatis et al., 1982) and used as template for synthesis of 32Plabeled cDNA (Gubler and Hoffman, 1983). A genomic library made from a partial Hind111 digest of S. pombe DNA cloned into DB262 (Wright et al., 1986) was plated and transferred to nitrocellulose filters. The filters were incubated at 65 "C overnight with lo6 cpm/ ml cDNA (+ thiamine) in 6 x SSC, 2 x Denhardts, 0.5% SDS, 100 pg/ml herring sperm DNA. After hybridization the filters were washed in 2 X SSC, 0.5% SDS, 65 "C (2 washes, 30 min), 0.1 X SSC, 0.5% SDS 50 "C (1 wash, 30 min) and exposed to autoradiographic film. The tirst probe was then removed in 50% formamide, 1 X SSC at 70 "C and the filters reprobed with cDNA (-thiamine) as above.
From the results of differential hybridization, several colonies were selected and one of these contained nmtl, the subject of the present report. To obtain the flanking sequence upstream of nmtl, the same genebank was hybridized to the 32P-labeled nmtl probe and a 4.1-kb fragment was recovered which contains the original Hind111 fragment of 2.4 kb and a second Hind111 fragment of 1.7 kb which was shown by Southern blotting to be contiguous with and upstream of the 2.4kb fragment in the genome and not the result of coligation during cloning.
DNA Sequencing and in Vitro Mutagen&+-DNA fragments were cloned into either M13mp19 or pUCI19 and sequenced in both directions from deletions made using Exonuclease III and Sl nuclease (Henikoff, 1984). In some cases custom-made sequencing primers were also used. Reactions were performed using Sequinase according to the suppliers instructions. Sequence data handling was performed using the UWGCG software (Devereux et al., 1983). Site-directed mutagenesis was performed as described  and in more detail elsewhere.' 5' and 3' Mapping of the nmtl Transcript-The 5' end of nmtl mRNA was mapped by two methods: (a) by primer extension using a synthetic oligonucleotide 5'-CCTCAATCCCTTCACGCTCA-3' complementary to the coding sequence between positions +90 and +109. The oligonucleotide was 3ZP-labeled with polynucleotide kinase and annealed-and extended with reverse transcriptase as described (Grimm et al.. 1988); and (b) bv Sl mapping using a uniformly labeled probe. In this case the probe was made byannealing the oligonucleotide above to single strand Ml3 DNA carrying the nmtl-coding 1 The abbreviations used are: CAT, chloramphenicol acetyltransferase; kb, kilobase; SDS, sodium dodecyl sulfate.
strand, extending the primer using the "Klenow" enzyme in the presence of 3ZP-labeled dCTP and digesting with HindIII. The reaction mixture was denatured and radioactive 293 nucleotide fragment was purified on a preparative sequencing gel and used for Sl mapping essentially as described (Berk and Sharp, 1977). The 3' end of the mRNA was mapped by Sl mapping. The probe in this case was made from an Exonuclease III-derived mutant, carrying a 3' deletion which extends into the clone as far as position +1310. Single strand Ml3 phage DNA containing this deletion was annealed with the 17-mer universal primer, extended with Klenow enzyme, and digested with NcoI. The 371-nucleotide fragment was purified and used for SI-mapping as above.
Transcriptional "Run On"-The procedure was based on that described by Elion and Warner (1986). Wild-type cells were grown to mid-log phase in minimal medium or minimal medium supplemented with 2 PM thiamine. From each culture, 3 X 10' cells were harvested by centrifugation, washed in TMN (10 mM Tris-HCl, pH 7.4, 5 mM MgCl2, 100 mM NaCl) and resuspended in 950 ~1 of HzO. 50 ~1 of 10% sodium lauryl sarcosine was added and the suspension incubated on ice for 30 min. Cells were pelleted for 2 min in an Eppendorf centrifuge, and the pellet was recentrifuged to remove all detergent. The final pellet was resuspended in 100 pl of reaction mixture (50 mM Tris-HCl, pH 7.9, 100 mM KCI, 5 mM MgCh, 1 mM MnClz, 2 mM dithiothreitol, 0.5 mM ATP, 0.25 mM GTP, 0.25 mM CTP, 10 mM phosphocreatine, 10 @g/ml phosphocreatine kinase) containing 10 al (100 IrCi) of lc~-~'PluTP and incubated at 25 "C. After 5 min 1 ml of'ice-cold TMN containing 1 mM UTP was added. Cells were pelleted, and RNA was extracted as described (Maundrell et al. 1985a). Identical slot blots containing 5 pg/slot of various immobilized plasmid DNAs (see Fig. 3) were hybridized overnight to 5 X lo6 cpm of each RNA probe using the same hybridization conditions employed for Northern blotting. Washing conditions were the same as described above for differential screening of the genomic library and filters were exposed to Kodak X-omatAR for 2 days.
Disruption of nmtl-The procedure was essentially that described by Grimm et al. (1988). The 2.4-kb Hind111 fragment containing the nmtl coding region was cloned into pUC19 and the 0.8-kb XhoI-NcoI fragment was replaced by a 1.8-kb segment containing the ura4 gene. The plasmid was digested with PvuII which cuts on either side of the polylinker and approximately 10 pg of the linear fragment containing the disrupted gene was isolated by preparative electrophoresis and used to transform the diploid strain hjh ade6-704/ade6-704 leul-32/ leuI-32 ura4-dlB/ura4-dl8 to uracil prototrophy. Of 10 stable Ura+ transformants analyzed by Southern blotting, two were the result of replacing one nmtl allele of the diploid with a single copy of the disrupted allele. Sporulatine h/h* revertants were obtained from both homologous integrantsanh tetrads were analyzed by microdissection (Gutz et al., 1974). Identical results were obtained with both strains.

Isolation of Thiamine Regulateable Genes by Differential
Hybridization-As part of a search for inducible genes in fission yeast, I have screened a genomic library for sequences whose expression is sensitive to the presence of exogenous thiamine.
In S. pombe, thiamine is not an essential requirement for growth although it is maintained at high intracellular concentration which suggests that as in S. cerevisiae, active biosynthesis of the vitamin can occur. The weakly thiaminerepressible genepho4 has been previously described (Schweingruber et aZ., 1986). For a more extensive search for thiamineregulated genes, cells were grown in either minimal medium or minimal medium supplemented with 2 PM thiamine, and the poly(A)+ mRNA was prepared from each culture and used to prime the synthesis of corresponding 32P-labeled cDNA probes. A genomic library consisting of partially digested HindIII fragments cloned into pDB262 (Wright et al., 1986) was then screened sequentially with each of the two cDNA probes. Several different clones were isolated all of which hybridized more strongly to cDNA derived from cells grown in minimal medium alone. One of these clones contained a single 2.4-kb Hind111 fragment, and hybridized as strongly as any other clone in the gene bank, suggesting that in the nmtl of Fission Yeast 10859 induced state it produces a highly represented cellular mRNA. The gene encoded by this fragment was subsequently referred to as nmtl (see next section) and is the subject of the present report.
Thiamine Is a Potent Inhibitor of nmtl Expression-Northern blotting analysis of total S. pombe RNA using the 2.4-kb Hind111 fragment as probe revealed a single transcript with an estimated molecular weight of approximately 1.3 kb (see legend to Fig. 1). Expression of this transcript was highly sensitive to the presence of thiamine. As shown in Fig. 1, thiamine concentrations in excess of 0.05 pM resulted in a progressive decrease in the steady-state level of the nmtl mRNA while at a concentration of 0.5 pM thiamine or greater, no hybridization at all could be detected even after long exposures of the blot. For this reason the gene was designated nmtl. As a control the same RNA samples were run in parallel and probed to reveal the cycl mRNA ( Fig. 1, lower panel).
The probe in this case was the 0.54-kb EcoRI-Sal1 fragment from pPoCYC(O.54) (Russell and Hall, 1982). The probes for both blots were labeled to the same specific activity, and the same number of cpm were used for each hybridization (see legend to Fig. 1). The results using the cycl probe confirm that approximately equal amounts of RNA were loaded in each lane. Moreover, quantitation based on densitometric scanning (see legend to Fig. 1) indicates that nmtl mRNA when fully expressed is 50-100 times more abundant than cycl mRNA.
To study the kinetics of thiamine repression, a cell culture was grown to early log phase in minimal medium and supplemented with 2 FM thiamine at TO. Samples were taken at l-h intervals thereafter for RNA preparation. A Northern blot probed to reveal the nmtl mRNA is shown in Fig. 2A. on nmtl expression. 972h-cells were grown to mid-log phase in minimal medium or minimal medium supplemented with increasing concentrations of thiamine from 0.01 to 2 pM as indicated. Total RNA was extracted and analyzed as described under "Materials and Methods." Specific activity of the 2.4-kb nmtl probe was 8 x 10' cpm/Fg. 10' cpm were used for hybridization. The blot shown was exposed for 9 h. Re-exposure of the same blot for 2 weeks showed a trace of mRNA in 0.2 j.iM thiamine but no detectable mRNA in 0.5 pM thiamine or greater. Molecular weight determination was based on the migration of HeLa cell rRNAs, yeast rRNAs, E. coli rRNAs, globin mRNA, and E. coli tRNA run in parallel and stained in 1 fig/ml ethidium bromide. The estimated size of the nmtl transcript is 1.3 kb. In the lower panel, identical RNA samples were run and processed in parallel. The probe in this case was the 0.54-kb EcoRI-Sal1 insert of pPoCYC(O54) (Russell and Hall, 1982). The specific activity was 6 X 10s and as above, lo7 cpm were used for hybridization. Densitometric scanning was performed on blots exposed for shorter periods using a LKB Ultroscan XL. Integrated peak heights were normalized for time of exposure, length of sequence complementarity (I.3 and 0.5 kb for nmtl and cycl, respectively), and the small variation in specific activity. Estimates indicate that the fully expressed nmtl mRNA is 50-IOO-fold more abundant than cycl mRNA.  2. A, Northern blot analysis showing the kinetics of repression of nmtl transcription by thiamine. 972h-cells were grown to early log phase (2 x lo6 cells/ml) in minimal medium. At 7'0 10s cells were withdrawn for RNA preparation, and the remainder of the culture was supplemented with 2 pM thiamine. Further aliquots of 10s cells were taken at intervals as indicated. RNA was extracted and analyzed as described under "Materials and Methods." B, Northern blot analysis showing the kinetics of induction of nmtl expression. 972h-cells were grown to mid-log phase in minimal medium containing 2 pM thiamine and an aliquot removed for RNA preparation (TO). The cells were washed in minimal medium and resuspended at lo6 cells/ml in fresh minimal medium. Further aliquots were withdrawn at intervals as indicated. After 10 h, as cells approached the end of log phase they were diluted in fresh prewarmed minimal medium for the analysis of later time points. RNA was extracted and analyzed as described under "Materials and Methods." message completely disappears from the cell within 3 h or about one cell cycle under these conditions.
The kinetics of nmtl induction are revealed in the reciprocal experiment in which cells were grown in minimal medium containing 2 pM thiamine, washed in minimal medium alone, and resuspended in fresh minimal medium at TO. RNA prepared at intervals thereafter shows the first appearance of nmtl mRNA after approximately 10 h in culture and full transcriptional activity by about 16 h. The longer time interval required for induction presumably reflects the high intracellular concentration of thiamine, and the fact that several generations are required to deplete this internal pool. From these experiments I conclude that in the inducing conditions nmtl is highly transcribed and that in the presence of thiamine at 0.5 jtM or greater, expression is totally repressed.

Thiamine Acts as a Repressor of Transcription
Initiation-Disappearance of mRNA from the cell as determined by Northern blotting can be due either to inhibition of transcription initiation or to an increase in the rate of mRNA turnover. To investigate this in the case of nmtl, transcriptional run on experiments were performed. Cells were harvested in midlog phase after growth in minimal medium or minimal medium containing 2 pM thiamine, and cells from each culture were permeabilized and incubated with [cx-"~P]UTP for 5 min essentially as described by Elion and Warner, (1986; see "Materials and Methods"). Radiolabeled RNAs were then extracted and hybridized to identical slot blots carrying various immobilized plasmid DNAs. As can be seen from the results shown in Fig. 3 the level of hybridization of radioactive RNAs to the two control genes ura4 and cycl is independent of the presence of thiamine in the growth medium, while the pUC119 vector alone does not hybridize to any cellular message after the high stringency washing conditions employed. On the other hand, the nmtl containing plasmid reacts very strongly with RNA from cells grown in minimal medium alone and hardly at all with RNA from cells grown in the presence of thiamine. These results indicate that the main effect of DNAs. I, pUC119 vector; 2, pCG1 containing urd (Grimm et al., 1988); 3, pPoCYC(O.54) (Russell and Hall, 1982). 4, runt1 in pUC119. 5 x lo6 cpm were used for each hybridization.
exogenous thiamine on nmtl expression is to drastically reduce the rate of transcription initiation. Structure of the nmtl Gene-The 2.4-kb Hind111 fragment containing the nmtl gene was completely sequenced in both directions from a series of deletion mutants prepared by digestion with Exonuclease III and nuclease Sl (Henikoff, 1984).
A major unspliced open reading frame of 346 amino acids was found which potentially encodes a 39-kDa polypeptide (Fig. 4). A search of the EMBL and Genbank databases showed no obvious homologies with other known proteins.
In order to analyze the 5' regulatory region of nmtl, a more extensive upstream region was obtained by reprobing the same partial Hind111 gene bank with the radiolabeled 2.4-kb insert. Among the positive colonies, one was found which contained the original 2.4-kb fragment together with an additional fragment of 1.7 kb. Southern blots of total S. pombe genomic DNA digested with a variety of diagnostic enzymes and probed with each of the two fragments separately, showed that the 1.7-kb fragment was contiguous with and upstream of the 2.4kb fragment (data not shown). Most of the 1.7-kb fragment was also sequenced in both directions. The nucleotide sequence and restriction map of the nmtl locus is shown in Fig.  4.
The 5' end of nmtl mRNA was determined by primer extension using an end-labeled oligonucleotide (Fig. 5A) and by Sl mapping using a uniformly labeled probe (see "Materials and Methods"). Both sets of data revealed that transcription initiates on the A residue at position -69 and confirm that no splicing occurs in this region. The sequence TATATAAA occurs from -34 to -27 relative to the start of transcription. The 3' end of the transcript was determined by Sl mapping using a continuously labeled probe of 371 nucleotides extending downstream from the NcoI site within the coding region (see "Materials and Methods"). A single protected fragment of 244 nucleotides (Fig. 5B) locates the 3' end of the message 142 nucleotides after the stop codon. The predicted length of the mRNA based on mapping of the ends is therefore 1252 nucleotides, close to the value determined by Northern blotting. From these data I conclude that transcription of the nmtl gene produces a simple unspliced mRNA and that there is no detectable heterogeneity at either end (see Fig. 4B).
The Cloned Fragment Contains the Transcriptional Regulatory Elements-To ascertain whether the 4.1-kb cloned fragment was capable of regulated transcription of nmtl, the complete region in the original vector DB262, was transformed into fission yeast and grown in the presence or absence of thiamine. RNA was prepared and analyzed by Northern blotting (Fig. 6). Densitometric scanning of the blot shown in Fig. 6 shows that cells grown in inducing conditions (lane c) contain approximately twice as much nmtl mRNA as control cells transformed with DB262 alone (lane a). In repressing conditions nmtl mRNA is undetectable both in the control (lane b) and in the nmtl transformant (lane d).
To be certain that the increased transcript level comes from the plasmid, a derivative of the original clone was made in which the 185-base pair BamHI-BamHI fragment was deleted from the nmtl coding region (see Fig. 4). Analysis of the RNA from cells transformed with the deletion clone is shown in Fig. 6 (lanes e and f). As expected, two RNAs are detected, a full-length chromosomal transcript and a shorter transcript derived from the plasmid. Estimates based on densitometric scanning show that in inducing conditions (lane e), slightly more than half of the message (55%) is transcribed from the plasmid. When cells are grown in thiamine, transcription from both the chromosome and the plasmid is totally repressed.
From these data I conclude that the cloned nmtl gene and its flanking regions include all the regulatory elements necessary for thiamine-mediated control of transcription, and second, that normal regulation can occur even in the presence of multiple copies of the promotor.
A nmtl Disruption Strain Is a Thiamine Auxotroph-As a means of investigating the function of nmtl, the chromosomal gene was disrupted essentially as described by Grimm et al. (1988). The region between the XhoI and NcoI sites of the coding sequence was replaced with the ura4 gene and the linear fragment was isolated and used to transform the diploid strain h'/h+, ura4-dlB/ura4-d18, leul-32/leul-32, ade6-7041 ade6-704 to uracil prototrophy. Restriction analysis of 10 stable Ura+ diploid transformants showed two which had the predicted pattern for homologous integration at the nmtl locus (data not shown). A Hind111 digest of such a diploid (Fig. 7, lane D) produced equimolar amounts of a 2.4 and a 3.4-kb band corresponding to the wild type and disrupted fragments, respectively. A sporulating h+/hgO diploid revertant was selected by exposure to iodine vapor and the products of meiosis analyzed by tetrad dissection. In all cases all four spores were viable on rich medium. DNA from each of the spores was analyzed by Southern blotting and as expected showed two wild-type and two disrupted alleles, the two disrupted alleles segregating with the Ura+ phenotype (Fig.  7A). The growth characteristics of each of the four spores from a single tetrad cultured in minimal medium (supplemented with adenine, leucine, and uracil) with or without exogenous thiamine is shown in Fig. 7B and Table I. The results show that for nmtl cells (upper panel) addition of thiamine to the culture medium produces a stimulation in the rate of cell doubling of approximately lo%, showing indeed that in minimal medium the availability of thiamine is rate limiting for cell growth. Cells carrying the nmtlzural disrupted allele (lower panel) also grow in minimal medium containing thiamine and show the same accelerated pattern of growth shown by wild-type cells. However, in the absence of exogenous thiamine, cells which lack a functional nmtl allele are not capable of sustained growth.
Thiamine-mediated Control of Transcription Can be Transferred to Heterologous Genes-As a first step toward identifying the elements responsible for control of nmtl transcription, the flanking sequences upstream and downstream of the nmtl coding region were isolated and used to construct the expression vector pREP1 shown in Fig. 8. For this construct the 1.2-kb fragment which extends from the upstream BclI site to the initiator ATG was judged sufficient to contain the complete promotor. To facilitate fusion of reporter or other heterologous genes, the sequence around the initiator codon was modified by in vitro mutagenesis to incorporate recogni- tion sites for restriction enzymes Me1 and Sal1 (see Fig. 8). transcriptional stop element. By means of the reconstructed In addition, a second element, the 1-kb sequence extending polylinker, it is thus possible to replace precisely the nmtl downstream of the nmtl stop codon was copied by polymerase coding region with any other heterologous sequence of interchain reaction and inserted behind the promotor to provide a est. The complete expression cassette was then inserted be-  tween the PstI and SstI sites of a S. pombe shuttle vector containing the LEU2 gene of S. cerevisiae and arsl of S. pombe cloned into the Hind111 and EcoRI sites, respectively, of the pUC119 polylinker (see Hindley et al., 1987). This construct was designated pREP1. An integrating plasmid, pRIP1, was derived from pREP1 by deleting arsl. A more One copy of nmtl in the diploid strain h+/h+ ade6-704/ade6-704 leul-32/leul-32 ura4-d18/ura4-d18 was disrupted by replacing the 0.8-kb XhoI-NcoI fragment with the 1%kb S. pombe Ural gene (see "Materials and Methods"). DNA from the diploid digested with Hind111 is shown (D). The two hybridizing fragments of 2.4 and 3.4 kb correspond to the wild type and disrupted genes, respectively. The diploid was sporulated and DNA isolated from each of the four spores (Maundrell et al., 198513). The wild-type (spores 1 and 3) and disrupted (spores 2 and 4) genes segregate with the Ura-(-) and Ura' (+) phenotypes, respectively. B, growth curves showing the thiamine dependence of the nmtlxra4 disruption strain.
Each of the four spores in (A) above was cultured in minimal medium supplemented with adenine leucine and uracil either with (solid symbols) or without (open symbols) thiamine. Cells were counted at intervals using a haemocytometer.
Upper panel shows nmtl cells derived from spore 1 (diamonds) and spore 3 (triangles). Lower panel shows nmtlxrd cells derived from spores 2 (squares) and 4 (circles). detailed description of this and other vectors based on the nmtl control elements will be described elsewhere.' To test the activity of the nmtl promotor, the bacterial gene chloramphenicol acetyltransferase was used as a convenient reporter gene (Gorman et al., 1982). An MeI site was introduced at the initiator methionine codon by in vitro mutagenesis and the 0.8-kb NdeI-BamHI containing the entire coding region was subcloned between the nmtl promotor and transcriptional stop elements of pREP1 and pRIP1. S. pombe transformants were obtained which carried either multiple extrachromosomal copies of pREPl-CAT or a single A, schematic representation of the 4.1.kb genomic fragment containing the nmtl gene. The area shown in black is the nmtl coding region. The extent of the nmtl transcript is indicated by the arrow. The &pled areas represent the regions used for vector construction. B, structure of pREP1. The nmtl promotor was modified in two ways: first the internal iWe site at position -768 was destroyed as shown; second the sequence around the ATG was altered by site-directed mutagenesis to incorporate restriction sites for NdeI and SalI. The transcriptional stop element extending from the nmtl stop codon TAA to a site approximately 1 kb downstream was amplified by polymerase chain reaction and BanHI, SmaI, and SstI sites added as indicated. Both elements were assembled into the polylinker of pUC119 which contained in addition the LEII2 gene of S. cereuisiae and arsl of S. pombe cloned into the Hind111 and EcoRI sites, respectively. The resulting vector was termed pREP1. An integrating version of this vector, called pRIP1, was obtained by deleting arsl. leul-32 transformants containing either pREPl-CAT (lanes a and b) and pRIPl-CAT (lanes c and d) were grown in minima1 medium (lanes a and c) or minimal medium containing 2 pM thiamine (lanes b and d). 10' cells from each culture were processed for CAT assay as described by Jones et al. (1988). integrated copy of pRIPl-CAT. Parallel cultures of each transformant were grown in minimal medium either with or without thiamine and CAT activities in the respective extracts were assayed (Fig. 9) A systematic analysis of the thiamine responsive elements in the nmtl promotor is in progress.

DISCUSSION
In this report I describe the first fully repressible gene to be isolated from fission yeast. In minimal medium, it is among the very highly transcribed genes in the cell; in minimal medium containing thiamine at a concentration of 0.5 pm or greater, gene expression is repressed and to judge from long overexposure of the Northern blot shown in Fig. 1, the transcript completely disappears from the cell. This gene has been designated nmtl. The kinetics of induction and repression are shown in Fig. 2. Addition of thiamine to cells growing in minimal medium results in complete disappearance of nmtl mRNA within less than 1 cell cycle ( Fig. 2A) indicating a rapid transcriptional response and also a high rate of mRNA turnover. In the reciprocal experiment in which thiamine is washed out from the medium and the cells shifted to minimal medium alone, induction of nmtl transcription occurs after about 10 h in culture and maximal mRNA levels are obtained by 16 h. Transcriptional run on experiments show that thiamine control operates by inhibiting transcription initiation. The cloned portion of the genome which spans nmtl is 4.1 kb and contains all the sequences required for the thiamine mediated control of expression. This conclusion is based on the results in Fig. 6 which demonstrate that nmtl on a high copy number plasmid is repressed in the presence of thiamine as completely as the endogenous chromosomal gene. Scanning the nmtl promotor for repetitive elements, which in other eukaryotic genes have often been identified as targets for trans-acting molecules, revealed the presence of the 9-mer 5'-AAATAATAC-3' present at position -684 to -676 and an exact repeat at position -577 to -569 (underlined in Fig. 4B). Experiments are in progress to test the functional significance of this sequence. The identical 9-mer also occurs at position -122 to -114 in the promotor of the only other thiamine sensitive gene sequenced to date, the PH03 gene of S. cereuisiae (Bawja et al., 1984). As discussed in the Introduction, &-acting regulatory elements are often highly conserved among eukaryotes, even between yeast and man.
The function of the nmtl gene product is not known and a search of the EMBL and Genbank databases showed no obvious homologies with other known proteins. From the pattern of expression, a likely function would be an involvement in thiamine biosynthesis, and support for this prediction comes from the finding that the nmtl::ura4 disruption strain is a thiamine auxotroph. Furthermore, overexpression of nmtl results in approximately 10% stimulation of growth rate," similar to that obtained by adding thiamine to minimal me-dium (Table I). The effect in both cases is probably to increase the concentration of intracellular thiamine and this further strengthens the notion that nmtl is involved in thiamine production.
Chiu, R., Boyle, W. J., Meek, J., Smeal, T., Hunter, T., and Karin, M. (1988) Cell 64, 541-552 Why a gene involved in thiamine biosynthesis should be so highly expressed is not immediately clear. However, the observed stimulation of growth in the presence of exogenous thiamine indicates that the availability of the vitamin is rate limiting in minimal medium and therefore that selection has favored high level expression of the thiamine biosynthetic genes. The finding that overexpression of nmtl also increases growth rate in minimal medium suggests that the nmtl gene product performs a rate-limiting step in this pathway. Further analysis of this system should enable a detailed investigation of transcription activation in fission yeast analogous to studies previously performed in budding yeast. In addition the regulateable promotor of nmtl offers for the first time the possibility of developing an inducible expression system for this organism. In this paper I describe the construction pREP1 and RIPl, extrachromosomal and integrating vectors, respectively, which have been shown to confer thiamine-regulated expression on the bacterial chloramphenicol acetyltransferase gene. A more detailed description of these expression vectors will be presented elsewhere.'