Transcription, Nucleosome Stability, and DNA Repair in a Yeast ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No.9,Issue of March 25, pp. 599-6005, 1992 Printed in U.S.A.

Transcription, Nucleosome Stability, and DNA Repair ina Yeast Minichromosome* (Received for publication, August 8, 1991)

Jirair Bedoyan, Ranjan Gupta, Fritz ThomaS, and Michael J. SmerdonQ From the Department of Biochemistry and Biophysics, Washington State University, Pullman, Washington 99164-4660 and the Slnstitut fur Zellbiologie, ETH-Honggerberg, CH-8093 Zurich, Switzerland

The template strand of the URA3 gene in the mini- ing the mechanisms of DNA repair (reviewed in Refs. 1 and chromosome YRpTRURAP is repaired of UV-induced 2). A number of reports have appeared on repair of UVcyclobutyl pyrimidine dimers (PD) much more effi- induced damage in yeast plasmids (3-11).Some of these cientlythanthe nontemplate strandingrowth-arreports examine the repair of plasmid DNA (damaged in vitro) rested Saccharomyces cerevisiae cells (Smerdon, M. after being introduced into repair-proficient or -deficient J., and Thoma,F. (1990) CelZ61,675). However, other yeast strains (3, 4,9-11), while others examine the repair of regions of the plasmid are also efficiently repaired.We plasmid DNA as chromatin (damaged in the cell) in both have examined the transcription and chromatin struc-repair proficient and deficient yeast cells (5-8). Taken toture of these regions in growth-arrested cells to allow gether, these studies indicate that specific gene products in a more detailed comparison of transcription, nucleosome stability, and excision repair efficiency. North- yeast may be necessary for the recognition and/or repair of damaged DNA folded into chromatin (as compared with naern analysis, using strand-specific probes, indicates ked DNA). To date, no such gene product has been identified. that four different transcripts are made from YRpIt has also been established that mammalian cells (e.g. TRURAP in addition to the URA3 mRNA in both growing and growth-arrestedcells. The templates for Chinese hamster ovary and human) preferentially repair a these transcripts encompass all of the efficiently re- number of DNA lesions in (at least some) transcriptionally paired regions outside of the URA3 gene. Nucleosome active genes relative to the bulk of the genome and inactive mapping indicates that the structureof the minichro- genes (12-16). Furthermore, it was reported that the yeast mosome in growth-arrested cells is indistinguishable Saccharomyces cerevisiae preferentially repairs UV-induced from thatof growing cells except for two nucleosomes cyclobutane pyrimidine dimers (PD)’ in the transcriptionally in theregion 500-800 base pairs 6‘ of the URA3 gene active a-mating typelocus (MATa) ascompared to the inacwhich becomemuch less stable in growth-arrested tive locus (HMLa) (17). We have measured rate constants of cells. Comparison of the distributionof YRpTRURAP excision repair at specific PD sites ineach strand of the 2619topoisomers in the two states,however, indicates that bp minichromosome TRURAP in growth-arrested S. cerevisthese nucleosomes are not lost from the majority of iue cells (7). This plasmid containsa single intact gene plasmid molecules. One of the four transcripts initiates(URA3) and 14 nucleosomes of known positions (18). Prefin this region and increases by more than &fold in erential repair of PDsites (>5-fold) was observed in the growth-arrested cells. Another transcript extends into template strand of the active URA3 gene in TRURAP, coma “slowly repaired” region which contains a very stable pared to the non-template strand (7). In addition, regions nucleosome. By determining the stability and relative outside of the URA3 gene werealso efficiently repaired. Under amounts (at equilibrium) of the RNAs madefrom normal growth conditions, some of these regions contain less YRpTRURAP, transcription rates were determined and compared with the averagePD repair rate for the stable nucleosomes, while the three nucleosomes in one of different template regions. The results indicate that: these regions (called UNF for “unknown function”) are very 1) the rate of excision repair increases once a low, stable (18).Therefore, we have performed a detailed analysis of the entire basal rate of transcription is achieved; 2) beyond this of transcriptionandchromatinstructure TRURAP minichromosome in growth-arrested cells to more rate of transcription there is no simple correlation between the rates of transcription and PD repair; 3) closely examine the correlation of these features with repair nucleosome stability may “override” the coupling be- rates at specific sites. Wewished to determine: ( a ) if any tween transcription and repair if the transcription ratechanges occur in the nucleosome structure during incubation is low; and 4) at higher transcriptionrates, repair may of cells in water as compared with cells maintained in growth be insensitive to nucleosome stability. medium (i.e. the conditions of previous nucleosome mapping studies (18)); ( b ) if the efficient removal of PD from regions of TRURAP outside of URA3 indicates that transcription also occurs in these regions, or whether thisenhanced repair Yeast is now widely used as a model eucaryote for examin- correlates with an instability of nucleosomes in these regions * This work wassupported by National Institutes of Health Grants during incubation of cells in water; (c) if transcription occurs ES04106 and ES02614 (to M. J. S.)and by the Swiss National Science in a region shown to be repaired inefficiently (7) and which Foundation and the Prof. Dr. Max Cloetta Foundation (grant to F. contains stable nucleosomes (e.g. the UNF region) (18); and T.) The costs of publication of this article were defrayed in part by ( d ) whether a correlation exists between transcription rates the payment of page charges. This article must therefore be hereby and PD repair rates in TRURAP. marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § To whom correspondence should be addressed. Tel.: 509-3356853; Fax: 509-335-9688.

The abbreviations used are: PD, pyrimidine dimer(s); bp, base pair(s); UNF, unknown function; kb, kilobase pair(s); T, top strand; B, bottom strand.

5996

Transcription, Structure,and Repair of a Yeast Minichromosome

5997

Chromatin Structure of TRURAP in Growth-arrested Cells-To investigate whether the chromatinstructurein TRURAP changes upon incubation of cells in water (ie. “liquid holding”), FTY23 cells were grownin selective medium and then incubated in water at 30 “C for 2.5 h. Plasmid chromatin was partially purified, and the chromatin structures were analyzed by micrococcal nuclease digestion. To determine the location of nucleosomes and nuclease-sensitive regions, the cutting sites for micrococcal nuclease in chromatin were compared with those in protein-free DNA using indirect end-labeling (19, 20). Protection of140-200 bp against cutting by micrococcal nuclease operationally defines a positioned nucleosome. Using this criteria, mapping from the EcoRI site (in the clockwise direction) and XbaI site (in the counterclockwise direction) yields six precisely positioned nucleosomes flanked by nuclease-sensitive regions at the 5‘ and 3’ ends on URA3, a nuclease-sensitive region around ARS1, and three positioned nucleosomes (I, 11, 111) in the UNF region (Fig. 1, A and B ) . In theTRPl region, footprints of nucleosomes 3 and 4 were detected at low and high levels of digestion (Fig. 1, A and B ) . These structures are indistinguishable from the chromatin structures mapped in growing cells incubated in selective medium (18). Mapping from the Hind111 site reveals a clear difference between chromatin isolated from cells grown in selective medium (SD-chromatin) and cells growth-arrested in water (water-chromatin; Fig. IC).In SD-chromatin, the different cutting patternsof chromatin lanes andDNA lanes show the footprints of three precisely positioned nucleosomes in the TRPl region (boxes 1-3). However, multiple cutting sites appear in the region of nucleosomes 1 and 2 of water-chro-

matin, and much less protection is observed. These results cannot distinguish between a complete loss of nucleosomes 1 and 2 of TRPl upon incubation of FTY23 cells in water and increased nucleosome in~tability.~ Because the presence or absence of nucleosomes in plasmid chromatin is reflected in the distribution of negatively supercoiled topoisomers (21), we compared the topoisomer distribution of TRURAP DNA under the two conditions of incubation by employing chloroquine-agarose gel electrophoresis (22). The distribution of supercoiled topoisomers migrating at the position of form I DNA did not show any difference under the two incubation conditions (compare the densitometer scans shown in Fig. 2). This was the case for gels run at different chloroquine concentrations where the migration of topoisomers is shifted relative to thepositions of form I1 and form I11 DNA. There isapproximately one negative supercoil per nucleosome (21). Therefore, absence of any change in the topoisomer distribution suggests that the number of nucleosomes in TRURAP remain unchanged in growth-arrested cells compared with cells maintained in selective medium. This indicates that theincreased instability in water of TRPl nucleosomes 1 and 2 does not representan actual loss of these nucleosomes followed by a relaxation of the released supercoils. Overall Transcription of TRURAP-To assess the overall transcription of TRURAP, a double-strandDNA probe of the entire plasmid (TRURAP/RP) was made by random primer labeling. Total RNAwas isolated from both FTY23 cells (containingTRURAP),andSc3 cells (parentalstrain of FTY23) grown in growth medium or incubated in water for 2 or 4 h prior to harvest. The water-incubation period arrests cell growth (data not shown) and plasmid replication (8) and was the condition used for the DNA repair studies of Smerdon and Thoma (7). After electrophoresis on formamide/formaldehyde-agarose gels, the RNA was transferred to nylon membranes and hybridized with the TRURAP/RP probe (Northern blot analysis). Surprisingly, in addition to the 1.0-kb URA3 mRNA, four different transcripts are detected in both sets of cells (Fig. 4, B, lanes 3-6 and H, lane 3). These transcripts are derived exclusively from the TRURAP plasmid since no RNA from the parental strain(Sc3) is detected with this probe (Fig. 4B,lanes 1 and 2) even though similar amounts of total RNA were loaded on the gel (Fig. 4A).The amounts of each transcript, relative to URA3 mRNA, are shown in Table I (discussed in more detail below). It is clear from these data that considerably more transcription occurs on this 2.6-kb minichromosome than in the selectable URA3 gene. Indeed, approximately 64%of the TRURAP plasmid is used as template for transcription (see Table I1 and Fig. 8). Transcripts Made from the Top Strand of TRURAP-To determine the location of the templates for the different transcripts, several different strand-specific probes were used (Fig. 3). The RNA probe TRP5’/T hybridizes to the EcoRIXbaI DNA fragment of the “bottom strand” of the TRP-5’ region of TRURAP (Fig. 3). (Throughout this manuscript, “topstrand” (T) refers tothe DNA strand of TRURAP containing the nontranscribed strand of the URA3 gene; “bottom strand” (B) refers to theDNA strand containing the template used for URA3 mRNA synthesis.) This probe will, therefore, hybridize to RNA transcripts made with the top strand of the TRPl-5’ region serving as template. Northern

Portions of this paper (“Materials and Methods”) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass, Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

Throughout this paper, the term “nucleosome instability” refers to either the disappearance of precise positioning of nucleosomes during extensive digestion with micrococcal nuclease, or to the alteration of the position of nucleosomes on a given sequence in different plasmid constructs in uiuo (18).

The stability of nucleosomes in TRURAP for cells maintained in normal growth medium versus cells transferred to water was examined using micrococcal nuclease and indirect end-labeling (e.g. see Ref. 18). These studies were complemented by experiments using chloroquine-agarose gels to compare the average degree of supercoiling (reflecting the average number of nucleosomes) in TRURAP under the two conditions. In addition, Northern analysis, using strand-specific probes, was used to carry out a detailed analysis of the transcripts made from TRURAP under the two conditions. These latter studiesinclude measurements of RNA stability, using 1,lO-phenanthroline to inhibit transcription, and S1 nuclease mapping of the 3’ end of two of the transcriptsmade. We find that transcription does indeed occur in all regions of TRURAP that are efficiently repaired, even though some of thesetranscripts (presumably) have no function. Furthermore, the DNA in the very stable nucleosome UNFI serves as a templatefor one of the transcripts, although this template is very weaklytranscribed. By analyzing the kinetics of repair and transcription in specific regions (domains) of TRURAP, we find examples where the average rate of repair varies by 4.2-fold while the rate of transcription varies by >lO-fold. We, therefore, conclude that a “tightcoupling” between transcription rates and PD repair rates does not occur in this minichromosome once a low level of transcription is established. MATERIALS AND METHOD$

RESULTS

5998

of a Yeast Minichromosome

Transcription, Structure, Repair and

C

D

CC DD CC 2 50

2 .4 1 0

D

C 50 2

10.4

2

D 2

EcoRl cl

A

Xbal ccl

B

c c(c 50250 5 0

Hindlll

C D C C 5 0 2

1 0 2

ccl

C

FIG. 1. Chromatin structure of TRURAP and the stability of nucleosomes in T R P l after incubation of cells in water. Chromatin was prepared from FTY23 cells grown in selective medium (SO) (panels A and B and lanes 1-3 in panel C) or after incubation in water for 2.5 h(lanes 4-8 inpanel C). Partially purified minichromosomes (lanes C) and deproteinized DNA (lanes D ) were digested with micrococcal nuclease (0-250 units) and the cutting sites were mapped using indirect endlabeling (e.g. Ref. 18).Mapping was clockwise (cl) from EcoRI ( A ) ,counterclockwise (ccl) from XbaI (R), and counterclockwise from Hind111 (C). The TRPI, ARS1, UNF, and URA3 regions are indicated. RI defines the EcoRI site on TRURAP. Protected regions of 140-200 bp are assumed to represent the footprints of positioned histone octamers and are, therefore, interpreted as positioned nucleosomes (bores). The chromatin structure inferred by this method always reflects the average structure of a minichromosome population; individual minichromosomes might have different structures. Also, note that nucleosomes 1 and 2 of TRPI-5‘ (stippled rectangles) are observed in SD chromatin and not in water chromatin.

analysis of total RNA from FTY23 cells, incubated in either growth medium or water, yields a 1.2- and a 1.3-kbtranscript with this probe (Fig. 4C, lunes 3-6). (These two transcripts are resolved on longer gels.) The 1.2-kb transcript constitutes 74 k 13% ( n = 4) of these two transcripts and, as observed with the TRURAP/RP probe (above), these transcripts are derived only from TRURAP (Fig. 4C, lunes 1 and 2). Furthermore, quantitativeanalysis of these bands (see “Materials and Methods”) for cells grown in SD medium and water indicates that thelevel of the 1.2-kb/1.3-kb doublet decreases by -58% after 2.5 h of incubation in water (Fig. 5B). Use of a second probe (UNF/T), which hybridizes to the UNF region of the bottom strand of TRURAP (Fig. 3), demonstrates that the 1.2- and 1.3-kb transcripts extend beyond the EcoRI site into the UNFregion (Fig. 4H, lune 2). Therefore, at least two RNA transcripts are made from the top strand and extend into the region containing three stable nucleosomes (18). A third strand-specific RNA probe (URA3’/T) hybridizes to transcripts made from the top strandof the StuI-Hind111 fragment, which is at the 3’ end of the URA3 gene (Fig. 3). In thiscase, no RNA transcripts from either theSc3 or FTY23 strains were detected (Fig. 4G).This result indicates that the 1.2- and 1.3-kb transcriptsterminate before reaching the URA3 domain on the top strand. The template strand for these transcripts must, therefore, lie within 1250-1 bp and 2619-1781 bp region of TRURAP going from the right to left orientation (i.e. upstream of URA3) (see Figs. 3 and 8). (By using a longer probe from the URA3 gene region (called URA/ RP; see below) (Fig. 3), we were able to eliminate the possibility of positions 1250-615 bp on TRURAP (from right to

left orientation) serving as templates for either the 1.2- or 1.3-kb transcripts.) Since the two transcripts detected by the TRP5’/T and UNF/T probes extend into a region of stable nucleosomes, we wished to examine their correlation with the drop inrepair rate that occurs near the very stable nucleosome UNFI (7) (see also Fig. 8). This required a precise determination of the 3‘ termini of these transcripts using S1 nuclease protection analysis. A fifth DNA fragment (called UNF/Sl) encompassing all of the UNF region (Fig. 3) was used for these studies, and only the 3’ end of the top strand of this fragment was 32P-labeled(see “Materials and Methods”). After RNA-DNA hybridization and S1 nuclease digestion, two bands at 430 and 544 bp were obtained (Fig. 6, lunes 8 and 9),corresponding to the 3’ end of the 1.2- and the1.3-kb transcripts, respectively. (These sizes are accurate to within 10 bp of the actual termination site, which is within the limits of resolution of the nucleosome positionson TRURAP (Ref. 18 and Fig. l).) Quantitation of these bands from two different gels indicates that the 430-bp SI-protected fragment constitutes 68 ? 8% (n = 4) of the two bands. This value is close to thatobtained from Northern analysis (see above) and indicates that the 1.2- and 1.3-kb transcripts terminate at approximately 2190 and 2075 bp on the TRURAP plasmid, respectively, or at each end of nucleosome UNFI (Fig. 8). Transcripts made from the Bottom Strand of TRURAPTo assess transcription of the bottom strand (B) in the URA3 region, we first performed Northern analysis with the URA/ RP DNA probe, which was generated by random primer extension of the 1166-bp HindIII-Hind111fragment of URA3.

Transcription, Structure,

and Repair of a Yeast Minichromosome

5999

This probe will hybridize to RNA transcripts from both the minor 1.7-kb transcript (Fig. 4F, lanes 1 and 2), and no top and bottom strands (Fig. 3). Using this probe, a major transcripts were detected from Sc3 cells (Fig. 4F, lane 3). RNA transcript (1.0 kb) and a minor RNA transcript (1.7 kb) These transcripts are identical in length to those detected was detected in total RNA from strain FTY23 (Fig. 40, lanes with the URA/RP DNA probe, and changes in the levels 3-6). Once again, no RNA from strain Sc3 hybridized with between the different samples, when cells are incubated in the URA/RP DNA probe (Fig. 40, lanes 1and2). The amount water, are similar to the corresponding transcripts detected of the major transcript (1.0 kb; URA3) decreases by -54% using the URA/RP DNA probe. after 2.5 h of incubation inwater, while the level of the minor Using the strand-specific TRP5’/B RNA probe (Fig. 3), transcript (1.7 kb) increases by >5-fold within the sameperiod two transcripts were also detected (Fig. 4E, lanes 3-6). The (Fig. 5A). Changes inthe level of the 1.7-kb transcript canbe major transcript is 0.9 kb and the minor transcript is 1.7 kb. clearly seen by comparing lanes 3-6 in Fig. 4, B and 0. As with the 1.7-kb transcript detected by the URA/RP and We then performed Northern analyses with strand-specific URA3’/B probes, the level of the 1.7-kb transcript increases RNA probes that hybridize to transcripts made from the by -5-fold after 2.5 h in water, while the relative level of the bottom strand astemplate. As expected, the URA3’/B RNA 0.9-kb (TRP5’/B) transcript increases by -&fold over the probe (Fig. 3), also detected the 1.0-kb URA3 mRNA and the same timeperiod (Fig. 5 B ) . Both thesize and relative increase in water-incubatedcells indicate that the minor 1.7-kb transcript, detected using the TRP5’/B probe, is the same1.7-kb transcript detected using the URA/RP and the URA3‘/B probes. These results indicate that the template strandfor this transcript includes part (or all) of the EcoRI-XbaI fragment (in the TRPl-5‘region) and extends to the HindIII-Hind111 region of URA3 on the bottom strand. Transcription of this RNA may be initiated by the promotersretained intheTRPl-5’ sequence (see below). Since no0.9-kb transcript was detected usingeither theURA/ R P or the URA3’/B probes, the template strand for this transcript excludes the URA3 gene domain but includes part of the TRPl-5‘ region. Based on the size of this transcript, we estimate that the transcription start site is between 1900 and 2330 bp (discussed below). Transcription Consensus Sequences in TRURAP-A computer search of various consensussequences in TRURAPwas performed to help predict initiation and termination sitesof thetranscripts.The various symbols in Fig. 8 show the midpoint position of these sequences (see legend of Fig. 8). In addition, two elements, described by Bajwa et al. (23) to be similar totheupstream activator sequence (UASG,~) Migration CGGA(G/C)GAC(A/T)GTC(G/C)TCCG, are found between FIG. 2. Distribution of TRURAP topoisomers in cells incu- 2182 and 2196 bp (69% similar) and between 2573 and 2589 bated in water or growth medium. Supercoiled TRURAP DNA bp (78% similar), and are referred to as UASc.l-like element (Form I, FI) was extracted from FTY23 cells incubated in growth or in water for 2 h ( W ) .The different topoisomers were 2 and 1, respectively. One of these elements may serve to medium ( M ) resolved in 1.5% agarose gels containing either 40 pg/ml chloroquine initiate expression of the 0.9-kb transcript (see “Discussion”). ( A ) or 20 pg/ml chloroquine ( B ) (see “Methods and Materials”).The Furthermore, Yarger et al. (24) report thata 110-bp transcripdistribution of Form I DNA is shiftedaway from the Form 111 ( F I I I ) tion terminator exists upstream ofURA3 and could be inDNA when achloroquine concentration of20 pg/ml is used (B). volved in termination of this transcript (Fig. 8; small open Densitometer scansof the topoisomer distribution from lune M (solid line) have been superimposed with the scan from lane W (dashed arrowhead). Finally, since TRURAP was constructed from line) in each panel todemonstratethesimilarityin topoisomer the 1.45-kb TRPlARSl circle, an expression signal believed to be present between positions 1 and 170 bp, is still present distribution under thetwo conditions of incubation. TABLE I Features of the various transcrhts made from TRURAP Relative RNA size

Probes used for detection

abundance

Half-life (at equilibrium)

kb 1.7

URA/RP, URA3’/B, TRURAP/ RP, TRP5’/B 1.0 URA/RP, URA3’/B, TRURAP/RP 0.9 TRURAP/RP, TRP5’/B 1.2 TRURAP/RP UNF/T, TRP5’/T, 1.3’ TRURAP/RP ND UNF/T. TRP5’/T.

min“

min

10

Relative rate constant of RNA synthesis”

0.16

X 10’

1.1

6.9 1.4 43 0.6 ND “Valuesrepresent $/[RNA],.O for cells incubatedfor 2-4 h in water,where [RNA],., is the equilibrium concentration of URA3 mRNA. At equilibrium, ($),[RNA], = (k& (discussed in Ref. 30), where k$ = zero-order rate constant of RNA synthesis, k d = first-order rate constant of RNA degradation = In2/half-life, and [RNA], = equilibrium concentration of RNA “x”. * ND, not determined. The half-life of this RNA is estimated to be >150 min, and the relative rate of RNA synthesis is estimated to be cO.1 min” X 10’ (see “Results”). 10 13

1.00 0.26 0.38 0.14

6000

Transcription, Structure, and Repair

of a Yeast Minichromosome

in TRURAP (Fig. 8; large open arrowhead) and may serve to initiate expression of the 1.7-kb transcript. Using all of the above information, we were able to assign putative transcription initiation and termination signals to the various transcripts made from TRURAP (see Fig. 8). Stability of the RNA Transcripts from TRURAP-We also wished to measure the half-life of the individual transcripts t o determine whether active transcription was occurring during the 4-h water-incubation period. This period includes the

DNA repair period (1-2.5 h) used by Smerdon and Thoma (7) to measurerepair inthese regions. In S. cereuisiae, two different populations of mRNA are observed (25, 26): one of short half-life (3-20 min) and anotherof long half-life (50 to >lo0 min). One method for measuring mRNA stability in yeast hasbeen through theuse of various transcription inhibitors (25-28). In the present study,we have used 1,lO-phenanthroline, aneffective zinc chelator used previously to block RNA synthesis (25), to measure the stability of transcripts made from TRURAP. For these experiments, FTY23 cells were maintained inwater for 1 h before the addition of ARS 1 inhibitor. TRPl-S URA3 TRPl-3’ UNF T 5’ 3’ A half-life of 43 min was obtained for the 1.2-kb transcript TRPS’IT URAB’IT UNFIT 0 (Fig. 7). Therefore, this (presumably) “nonsense transcript,” URAlRP Probes falls into the category of “stable m R N A in yeast. From this value, we estimate that over 70% of the 1.2-kb transcript 4 l2kb S z E 13kb 5’ * detected after 1 h of incubation of cells in water is due to . . . .... . . ... ..... .... . . ............ .......... ................. ... ........... transcription of this RNA after the start of the water-incu1 0 kb 0 . 9 h b P “5 bation period. Our longest time point in these experiments c 1 7 kb 5’ was 80 min. We, therefore, could not accurately calculate the half-life of the 1.3-kb transcript since it decreased so little in this time period. However, we estimate this value to be >150 B 3’A A A . I A A 5 min, because the half-life obtained for the 1.2-kb/1.3-kb doubE X k H P E 1 ” ” 1 ” ” i ” ” 1 ” ” 1 ” ” 1 let was 76 min (Fig. 7). 1 500 1000 1500 2000 2500 The half-life for boththe 1.0-kb transcript (i.e. URA3 2619 FIG. 3. Location of probes usedin these studies. The strand- mRNA) and the 1.7-kb transcript is 10 min (Fig. 7). This specific RNA probes (lightly stippled bars) hybridize to the corre- value is very close to the value reported for URA3 mRNA sponding adjacent RNA strand(s) made from TRURAP (filled lines made from the yeast genome (25, 29). Therefore, during the with arrows). The double-stranded URA/RP probe (filled bar) hy- 1-4-h water-incubation period, over 90% of the URA3 mRNA bridizes to both strands of TRURAP. The single-stranded UNF/S1 DNA probe (very lightly stippled bars) was used in the S1 nuclease (1.0 kb) and the 1.7-kb transcripts detected are due to tranprotection analysis (see Fig. 6). “Top strand” ( T )refers to the non- scription during thatperiod. template strand for URA3 mRNA, and “bottom strand”( B )refers to A half-life of 13 min was measured for the 0.9-kb RNA the template strand for URA3 mRNAsynthesis. The top strand (Fig. 7). It is obvious from the 7-fold increase in the level of shows thetranslation domain of URA3 as well as the disrupted this transcript (relative to URA3 mRNA) after about 2 h of translation domain of TRPl. The large vertical open arrow denotes incubation in water (Fig. 5), that thereis active transcription the location of the ARSlconsensus sequence. The region downstream of ARSl is termed the “unknown function” (UNF) region. The UNF during the water-incubation period. We estimate that over contains part of the 5’-regulatory region of GAW (see “Results”). 95% of the 0.9-kb RNA detected after 1 h in water is due to The scale (in bp) is from the unique EcoRI site of TRURAP. Closed transcription during this time. Thus, each of the templates arrowheads represent restriction sites on TRURAP used for subclon- on the TRURAPplasmid is being actively transcribed during ing the probes ( E , EcoRI; X,XbaI; H, HindIII;S, StuI; and P, PstI). The open arrowhead represents a restriction site found uniquely in the period in which repair measurements were made (7). Analysis of RNA Synthesis Rates from TRURAP--In order the UNF/T , urobe . ( i x . is not D a r t of TRURAP but is in the vector to calculate the rates of mRNA synthesis in each region, used to subclone these probes):

- -

Strain: S S F GrowthMedium: M W M Time (hr): 0 4 0 1 2 3

FIG.4. Northern analysis of total RNA from strain Sc3 and FTY23 using various probes. S and F represent strain Sc3 and FTY23, respectively. Lanes contain total RNA from late-log phase cells maintained in minimal medium supplemented with the necessary amino acids ( M )or switched to water ( W ) ,respectively. Cells were suspended in water and incubated at 30 “C for the times shown. Panels arefor blots probed with TRURAP/RP ( i x . a doublestranded probe of theentire plasmid DNA) ( B ) ,TRP5’/T ( C ) ,URA/RP (D), TRPB’/B ( E ) , URA3‘/B ( F ) , and URA3’/T ( G ) (see Fig. 3). Panel A representsthe agarose gel, stained with ethidium bromide to identify the two major rRNAs in yeast, used for blot B. The panels show only regions of the blots containing detectablebands. Panel H represents Northern blots probed with TRURAP/RP (lane 3) or UNF/T (lane 2 ) . Lane 1 in panel H shows some of the molecular size markers used for these gels.

25s rRNA

*

F M 4 4

F W 2 5

F W

Strain:

F

F S

GrowthMedium: W W M Time (hr): 0 4 0

4

6

7

-1.7 -1.0 (URA3)

A 10s rRNA t

Probe

B

TRURAPIRP

-

-

G URA3’1T

-

kb 1.7

1.3 1.2 -1.0 (URA3) 4

-0.9

c

TRP5‘IT

E

URAIRP TRP5‘1B

1 2 3

*

R(l=

lo-fold. By comparison with the results of Smerdon and Thoma(7), each of the approximate template regions for these transcripts contains PD sites that are efficiently repaired (Fig. 8, burs). It isalso clear from our data thatthe 1.2-kb transcript terminates just before the stable UNFI nucleosome (18) (see Fig. 6). Because the 1.3-kb transcript constitutes approximately 25% of the two transcripts detected and because the half-life of this transcriptis relatively high compared with the 43 min measured for the 1.2-kb transcript, we estimate that the relative rate of 1.3-kb RNA synthesis is 5-fold. Therefore, this region may indeed play a role in bothtypes of activity, as suggested by Yarger et al. (24). It is interesting to note that both the URA3 mRNA and the 1.2-kb/1.3-kb RNA doublet, which may originate from the same TATA element region and aretranscribed in opposite directions, decrease in their relative levels in growth-arrested cells. Concurrently, the relative levels of two other transcripts (0.9- and 1.7-kb RNAs), originating upstreamof URA3, increase. Whether the process which increases the level of one of these pairs of transcripts directly affects the decrease of the other pair, remains to be determined. It is possible that such variations

I 1 FIG.8. Comparison of the predicted RNA template regions and repair rate constants at specific P D sites on each strand of TRURAP. The first-order rate constants for repair at PD sites in TRURAP (bars) are taken from Smerdon and Thoma (7). Bars with an asterisk represent average values determined from mapping the repair rates from two different restriction sites (7). The translation domains of URA3 and TRPI, as well as the position of nucleosomes (circles), are shown aboue the figure. Shaded nucleosomes 1 and 2 of the TRP1-5’ region become unstable after incubation of cells in water (see Fig. 1 and “Results”). Large arrowheads represent sites of TATAelements found from a computer search forthe TATA(T/A)A(T/A) sequence (solid), a TATA elementfound in higher eucaryotes and yeast (38-40), or from published reports (open) (41). Small arrowheads represent sites of transcription termination signalsfoundfroma computersearch for theTAG...TAGT(or TATGT). . .(AT > 50%). . .TTT sequence (solid), a proposed transcription termination signal in yeast (42), or from published reports (open) (24). Arrows represent sitesof TTTTTATA proposed to signal both transcription termination and polyadenylation in yeast(43). Open diamonds represent sites ofAATAAA and TAAATAA(A/G) believed to represent polyadenylation signals for higher eucaryotes (44) and yeast (45), respectively. Transcription symbols aboue the scale correspond to transcripts from the B strand (and vice versa for the T strand). Transcripts are shown next to their respective template strand. The dashed portion of each transcript represents the uncertainty in initiation/termination positions for the predicted sequence signals. The 3‘ endsof the 1.2- and 1.3-kb RNA were determined by S1 nuclease protection analysis (see Fig. 6).

would make some previously nonfunctional transcription regulatory regions available to transcriptionfactors. Transcription and PD Repair-Comparison of the results of Smerdon and Thoma (7) with those of the present study indicates that transcriptionoccurs in allregions of TRURAP that are efficiently repaired after UV-induced damage (Fig. 8). We have used unirradiated cells to monitor transcription of TRURAP because it is known that the transcriptioncomplex is blocked at PD sites (32,33) and, therefore, UV damage could interfere with determination of the “unimpeded transcription rate” for the different template regions. Thus, the use of unirradiated cells allowed us toexamine the correlation between the average transcription frequency of a specific domain priorto UV damage and the ratewhich at that domain is repaired following introduction of transcription-blocking photoproducts. This was done by first determining the average first-order rate constant for repair of PD sites and thezeroorder rate constant for transcription in each template region of TRURAP (including template regions used for several transcripts), aswell as thefraction of total PD atzero repair time for each template (Table 11). This comparison indicates

Transcription, Structure, and Repair of a Yeast Minichromosome

6003

TABLEI1 Average repair rateand rate of transcription in specific domainsof TRURAP Average repairb Strand ~~

B B B T T

Transcription‘

Domain” ~

~

~

Rate constant

Rate

Relative rate

min” X 10’ 1.58 (R= 0.98) 1.16 (R = 0.97) 0.58 (R = 0.82) 1.60 85 (R = 0.99) 20 (R = 0.90) 0.30

PD/min X IO’

%

0.40

100 15 12

Rate constant

~~

800-1800bp 1-650 bp 2300-2619 bp 1-750 bp, 2200-2619 bp 800-1800 bp

~

0.06

0.05 0.34 0.08

Relative rate

~~~~~

min”

X 10’

743

(k’d1.0

+ (k’J1.7 = 8.0

(k’s11.2

+ (k’A1.3

(k’,)o.s + ( k ’ s h . 7 = 2.5 (k’J0.9 18 = 1.4

0.6

k’, = 0.0

100 31

8

0

Repairrate constants were not measuredfor PD sites between positions 1-190 bp and 2550-2619 bp on TRURAP (see Fig. 8) due to using the EcoRI/XbaI fragment (1-186 bp) as probe to measure repairrate constants at specific PD sites (7). * Values are calculated from the data used in Smerdon and Thoma (7). Average first-order rate constants of PD repair for a specific domain represent the slope of a linear-fit of -ln[(A,=, B,=,)/(A,=o B,=o)]versus repairtime t ; where A, = sum of all intensities at T4 endoV-sensitive sites, over a specific domain, for a specific repair time and probe,and B, is the same as A, except for a different probe. Values in brackets represent the linear correlation coefficients for these fits. The average rates of PD repair were calculated from the fraction of PD at zero repair time for a specific domain and the average PD/plasmid-strand at the specific UV dose (i.e. 0.6 PD/plasmid-strand see Smerdon et al. (8)). See Table I for description of notations and values used, and ( k ’ s ) x= (k,),/[RNA]l.o. A value of k, = 0 was taken for the non-template strand of URAJ (see “Results”).Zero-orderrate constants of transcription for domains used for more than one RNA transcript were obtained from the sum of the relative rate constants of the respective transcripts. a

+

there is not a simple (e.g. linear) correlation between these rates in the TRURAP minichromosome. For example, the difference between the rates of transcription of URA3 mRNA and the 1.2-kb transcript is >lo-fold, while the average rate of repair of these two templates varies by menslonal chloroquine-agarose gels (22). TRURAP DNA was solated from FTY23 cells (grown as belore and mubated ln Water for 2 h) uslng a glass.bead l y s ~method (J Mueiler and M. J Smerdon. unpublished resuils) Topojsomers were resolved on 1 5% agarose gels (84 x 20 cm)contalnlng chloroqucne dlphasphate (Stgma) at elther 20 pglml or 40pglml. m both the gel and the electrophores6 buffer Samples were sublected to electrophoresls at room temperatureat 4-5 Vlcm for 1 4 h.Aftereleclrophores1s gels weretranslerred to Zetabmdmembranes.hybrldlzed wrth 3+Iabeled probes.exposed to fuli-RX film and autoradiographs werescanned as descrlbed earlier Maootno NuciwpSome Pos~lrons~ For -SD-chromatln", yeast cells were grown On selective medlum ( t o an absorbance 01 about 1 to 2 at 600 nm) and harvested by centrifugatjon. for ilqufd holdlng and growth arrest (.water-chromatln"), cells were g r ~ w nIn 3 Lselect!ve medlum l o an absorbance of 1 , harvestedby centrllugat,on. washed once m 800 ml delonlzed water and resuspended I" one L delonlzed wafer The cells were shaken at 300C for and h 5 2 harvested by centrlfugatlon Prepwallon of plasmld chromaltn and control DNA, dlgestlon wlth micmcoccal nuclease and mapplng of the culttng sltes were doneasdescrlbed previously (18. 50). The cells were converted to spheroplasts uslng 25 mg Zymalyase(100T. Kmn Brewerles. Takasakt. Japan) m 80 ml spheroplasting buffer [40 mM K z H P 0 4 . pH 7.5. 20 mM p-mercaptoethanol, 1mM phenylmethanesulfonylfluorldeIPMSFI. 1M sarbltol] at 30% for 40 mln The spheroplasts I" 120 ml spheroplastlng buller(w,thout werecollectedbycentrllugalion,washedonce p-mercaptoethanol).and lysed in 20 ml coldbufferA (20 mM Tris. pH 8. 150 mM NaCI. 5 mM KCI, 1 mM EDTA, I m MP M S f ) contalning 0.2% Trlton The genamlc chromatm was pelleted (18,OOOxg. at 4OC) and conlamlnatlng malettal was removed from the supernatant by gel flltratlon on Sephacryl S300 ln buflerA The plasmld chromatin eluted ~n the voidvolumeandwaspooled Hall 01 the pool was used 10 extract ~ o n t l o lDNA Mciococcal nuclease dlgesttons of chromatin and DNA were done buffer in A, Supplemented wlth 5 mM CaClz lor 5 mln at 37OC uslng differentenzymeconcentrations To map the cutltng ales. DNA was extracted. CUI wllh a restrlctton enzyme (EcoRI, Xbal, Hlndlll, asindicated I" legend to Flg l ) , separated an 1 % agarose gels ~n tm-borate buffer containlng elhidrum bromlde blotted 10 nylon membranes. hybrldlzed to radioactivelylabelledprobes(derwed from EcoRI-Xbal and EcoRV~Hrodlllof the T R P l regton) The bandsweredetected w n g Fuy-RX Fllms and enhancer screens. Protected reglans of 140 to 200 bpareassumed to represent the lootprlnts of positioned hlstone octamers and are. therelore. inlerpreted as posmoned nucleosomes Note. that the chromatin StrUCtule lnlerredby this methodalwaysreflects the averagestructure of a minichromosomepopulaflan,lndwdualmlmchramosomesmlghthavedlllerent StrUCtures.