Cell, Vol. 71, 267-276,
October
16, 1992, Copyright
0 1992 by Cell Press
The Arrest of Replication Forks in the rDNA of Yeast Occurs Independently of Transcription Bonita J. Brewer, Daniel Lockshon, and Walton L. Fangman Department of Genetics, SK-50 University of Washington Seattle, Washington 98195
Summary Replication forks, moving opposite to the direction of transcription, are arrested at the 3’ ends of the 3% transcription units in the rDNA locus of S. cerevisiae. Because of its position and polarity, we tested the hypothesis that this replication fork barrier (RFB) results from the act of transcrlption. Three results contradict this hypothesis. First, the RFB persists in a strain containing a disruption of the gene for the 135 kd subunit of RNA polymerase I. Second, the RFB causes a polar arrest of repliccltion forks when transplanted to a plasmid. Third, transcrlption by RNA polymerase II of a piasmid copy of the 35s transcription unit lacking the RFB does not generate a barrier. We propose that repiication forks are arrested in a directional manner through the binding of one or more proteins to two closely spaced sites in the RFB. Introduction The rDNA locus of yeast is an approximately 2 Mb region of chromosome XII that consists of tandemly repeated 9 kb units, each of which encodes the four structural RNAs of the ribosome (reviewed in Warner, 1989; Figure 1A). The 8800 base 3% rRNA that is a precursor to the 18,25, and 5.8s rRNAs is transcribed by RNA poiymerase I in a direction that is opposite to the transcription of the 150 base 5s gene by RNA poiymerase III. Transcribed sequences are interrupted by two nontranscribed spacers (Figure lA), one between the 3’ ends of the transcripts (NTSl) and a second, containing a potential origin of repiication (rDNA ARS), between the 5’ends of the transcripts (NTS2). The replication of the rDNA locus has been studied by analyzing replication intermediates on two-dimensional (2D) gels (Linskens and Huberman, 1988; Brewer and Fangman, 1988). Replication intermediates that contain an origin of replication have two diverging replication forks and consist of a series of branched molecules referred to as bubbles. Under the appropriate conditions of eiectrophoresis, a series of bubble intermediates migrates on 2D gels in a characteristic manner and produces the arc of hybridization illustrated in Figure 16 (Brewer and Fangman, 1987). Fragments that lie between origins may be replicated by a single fork, moving through the fragment from one end to the other, and thus during their replication a series of branched molecules called simple Ys is generated. The migration of simple Ys on 2D gels is clearly distinguishable from the migration of bubbles (Figure 1C).
By analyzing restriction fragments from different portions of the rDNA repeating unit, it has been shown that repiication begins in the NTSP, but in only a subset of the repeats (Linskens and Huberman, 1988; Brewer and Fangman, 1988; Fangman and Brewer, 1991). Repeats that do not contain an active origin are replicated passively by a replication fork that originates in a nearby repeat. A unique feature of the rDNA locus is that the two forks initiating at an rDNA origin experience unequal fates. The fork moving through the 35s transcription unit in the same direction as RNA poiymerase I transcription proceeds unimpeded through multiple repeats, while the opposite fork proceeds through the adjacent 5s gene and NTSl and then arrests at a site (the RFB) near the transcription terminator for the adjacent, upstream 3% transcription unit. The arrest of forks at the RFB is deduced from the accumulation of replication intermediates of a specific size. When DNA is harvested from an asynchronous culture, all possible stages of replication intermediates are represented in the sample. Therefore, the amount of any particular intermediate reflects the amount of time the replication complex idles at a specific point in the restriction fragment. Replication intermediates within the 35s region produce a uniform distribution of simple Y intermediates on a 2D gel. in contrast, restriction fragments that contain the RF6 accumulate a class of replication intermediates that generate an intense hybridization signal at a unique spot along the otherwise uniform arc of simple Y replication intermediates (Figure 1D). The location of the spot along the arc of simple Ys varies with the choice of restriction enzyme, and, therefore, the position of the spot can be used to map the site of fork arrest and also can be used to determine the direction the replication forks were moving before their arrest. From this type of analysis it was deduced that the RFB arrests forks in a polar manner (Brewer and Fangman, 1988). Using an alkaline second dimension in their2D gels, Linskensand Huberman (1988) came to similar conclusions. Because of its polarity and its location near the transcription terminator for RNA poiymerase I, it was suggested that transcription itself might be responsible for the arrest of replication forks. In this report we provide further characterization of the RFB. In particular, we have investigated the possibility that the barrier isdue toacoliision between RNA poiymerase molecules and the replication complex. Results Location of the RFB The RFB previously had been mapped to a subregion of the nontranscribed spacer between the 3’ends of the 35s and 5s transcription units, at position 7020 f 80 bp (Figure 2A; Brewer and Fangman, 1988). To localize the pOSition of fork arrest more precisely, total cellular DNA was digested with restriction enzymes, and fragments of approximately 2 kb were analyzed by 2D gel electrophoresis and hybridization with a probe from NTSl For fragments
Cdl 266
A.
rDNA RFFI 4
NT.7
d-“”
*
3 kb
““T ARS
B.
BUBBLE
C.
Figure 1, Cartoons of Yeast Intermediates on 2D Gels
SIMPLE
rDNA
Y
D.
and the Migration
BARRIER
of Replication
(A) Illustrated are two repeats of the 9.06 kb rDNA unit from chromosome XII showing the positions of transcription units (35S and 5s) and the positions of the rDNA ARS (in NTSP) and the RFB (in NTSl). (B)The expected hybridization pattern for the replication of a restriction fragment that contains an origin of replication at its center. The replication intermediates depart from the arc of linears (broken line) at the position of unbranched, unit-length linears (spot in lower right corner). The first dimension is from left to right; the second dimension is from top to bottom. (See Experimental Procedures, and Brewer and Fangman [I9671 for details.) (C) The expected hybridization pattern for a restriction fragment that is replicated passively by a fork entering from one end. The replication intermediates depart from the arc of linears at the position of unbranched molecules and then return to it, as the fragment nears the completion of its replication, at a position of twice the length of the fragment. (D) The expected hybridization pattern for a fragment containing an impediment to replication fork movement. The accumulation of replication intermediatesof a specific size results in an increase in the hybridization signal at a specific place along the simple Y arc.
6 (Hindlll-SnaBI) and C (Hpal-Bglll), the rightward end of the fragment is nearthe origin and the leftward end isclose to the site where the RFB was mapped previously (Figure 2A). If the barrier were within the fragment, intermediates
would accumulate that were nearly fully replicated; however, if the RFB were to lie outside the fragment, only a uniform arc of simple Ys would be observed. The results of these two digests are shown in Figures 28 and 2C. The Hindlll-SnaBI fragment replicates as a series of simple Ys with a prominent accumulation of intermediates near the position of 4.2 kb linears (arrow), indicating that the RFB is present on this fragment and that the barrier is very near the Hindlll site. The replication intermediates of the HpalBglll fragment produce a continuous arc of simple Y intermediates with no apparent pause site. We therefore conclude that forks are arresting in the 129 bp between the Hindlll and Hpal sites that define the left ends of fragments B and C (Figure 2A). Closer inspection of the barrier region of the HindlllSnaBl gel (Figure 28) reveals that the hybridization signal from the RFB is not spherical but somewhat elongated. To analyze the RFB at higher resolution, the smaller fragment D (Nhel-Pvull; Figure 2A) was examined. We observed what appeared to be two discrete points along the simple Y arc at which forks arrest (Figure 20). The origin-proximal spot is more intense than the distal one. We cannot determine whether any given fork is arrested twice or is arrested at only one site or the other. The RFB in an RNA Polymerase I Mutant Strain Transcribing rDNAchromatin is readily recognizable in the electron microscope by the “Christmas tree” appearance of the densely packed, nascent transcripts associated with the DNA fiber. Saffer and Miller (1988) found examples of actively transcribed rDNA repeats within regions of replicating chromatin, providing indisputable evidence that the two potentially conflicting processes can occur simultaneously on a given DNA molecule. Perhaps the simplest model to account for both the position and polarity of the RFB is that collision with RNA polymerase I prevents repliFigure
2. Mapping
of the RFB
(A) Detailed restriction map of the NTS region I I of the rDNA repeat. Arrows correspond to the 9 9.08/O kb 6 I 8 35s precursor and the 5S gene. Within the 3% region, the 3’ end of the mature 26s rRNA seI IB quences is indicated by the dark bar. RestricI IC I ID tion sites: A. Accl; 8. Balll: E, EcoRI: HI. Hpal: H3, HindIll;. M, Smal;N, Nhel; P, Pvullf S, D NM-PVUII B HindIII-SnaBI c HpaI-BglII SnaBI. The bars labeled B, C, and D, refer to restriction fragments analyzed in panels (B), (C), and (D), respectively. (B) A 2D gel analyzing the Hindlll-SnaBl frag ment (fragment B) from the rDNA NTS region. The fragment was detected by using the HpalPvull fragment as a probe. The simple Y arc , begins in the 2.1 kb dark spot in the lower right corner and returns to the arc of linears at a size of approximately 4.2 kb. The accumulated replication intermediates are indicated by the arrow. The additional spot on the arc of linears is most likely the result of a partial digestion. (C)A 2D gel analyzing the Hpal-Bglll fragment (fragment C) from the rDNA NTS region. The same probe as in (B) was used. Again, the additional spots on the arc of linears are the result of partial digestion. (D) A 2D gel of the 1.2 kb Nhel-Pvull fragment (fragment 0) from NTSl. A light exposure is shown to reveal the appearance of two arrest points at the RFB. The dashed line marks the position of the arc of linears; the dotted line marks the position of the simple Y arc. Both lines were traced from a longer exposure.
Replication 269
Fork Barrier
in Yeast
rDNA
cation forks from entering the 3’ end of the 35s transcription unit (Brewer and Fangman, 1988). Since the initiation of transcription of 35s rRNA occurs at a rate sufficient to keep the template entirely covered by RNA polymerase molecules, if a replication fork were to approach such an actively transcribed region from its 3’end, the enzymes at the fork would repeatedly encounter RNA polymerases at the transcription terminator. We previously proposed that such repeated encounters may be sufficient to block the progress of this replication fork. The 35s region would then be replicated by a fork moving in from the C/end, i.e., moving in the same direction as RNA polymerase I. To determine the role of RNA polymerase I in the arrest of forks in the rDNA locus, we analyzed a yeast strain that has a disruption in the gene for a catalytic subunit of RNA polymerase I (RPA735; Nogi et al., 1991). This disruption is lethal, but the strain can be maintained by the presence of a multicopy plasmid containing the 35s transcription unit under the control of the RNA polymerase II promoter, GAL7(for a plasmid map see Figure 4A). Viability is absolutely dependent on the presence of galactose. The addition of glucose decreases the rate of 18s and 25s transcription more than 1Bfold (Nogi et al, 1991). The residual transcription that remains after glucose repression is probably derived from the plasmid, since the GAL7 promoter is not efficiently repressed when present in high copy numbers (Baker et al., 1987). We analyzed the chromosomal copies of the rDNA repeat in the rpa735::LEU2 strain. First, we discovered through Southern hybridization that in the absence of RNA polymerase I transcription, the rDNA locus had undergone a deletion of more than 80% of its repeats; however, the repeats that remained appeared unrearranged (data not shown). Using a probe to the nontranscribed spacer to study only the chromosomal copies of the rDNA, we found that the replication fork arrest at the RFB was indistinguishable from that found in the isogenic wild-type rDNA locus (Figure 3). All aspects of replication assayed were indistinguishable from the wild-type isogenic strain: the efficiency of initiation (data not shown), the intensity of the RFB relative to simple Y intermediates and the unbranched monomer spot (Figure 3) and the appearance of the double barrier in the RFB fragment (data not shown). The simplest interpretation of these data is that the act of transcribing the 35s rRNA by RNA polymerase I is not responsible for blocking the movement of replication forks into the 3’ end of the 35s transcription unit. Replication of the GAL7- 3% RNA Polymerese II Transcription Unit To determine whether transcription is capable of arresting replication forks, we examined the replication of the 35s transcription unit on the GAL7-35s plasmid in the rpa735::LfU2 strain. On this plasmid, the 35s rDNA is located between the promoter and transcription terminator of the GAL7 gene (Figure 4A). The rDNA insert, which extends only to the Hindlll site at 8855 bp (Figure 2A), includes the processing site to generate the mature end of the 25s rRNA, but lacks the signals for RNA polymerase I transcription initiation and termination, and does not in-
C. rpal3kUXL2
Figure
3. The RFB in the Absence
of Polymerase
I Transcription
(A) Bglll map of the rDNA. (E) 2D gel analysis of the 4.6 kb Bglll rDNA fragment from wild-type cells. The RFB region is at the center of this fragment. The second spot of hybridization on the linear arc at 9.1 kb is the result of an incomplete digestion by Bglil. (C) 2D gel analysis of the 4.6 kb Bglll rDNA fragment from cells carrying the rpa735::LEU2 allele. Since the repeat number of rDNA units is approximately 6-fold lower in this strain as compared with the wild-type strain (data not shown), a longer exposure was required to achieve a similar intensity of hybridization to the replication intermediates. The additional spot on the arc of linears is probably the junction fragment with non-rDNA sequences.
elude the Hindlll-Hpal RFB sequence (Figure 2A). Based on past studies, replication from the 2 pm origin is expected to be bidirectional (Brewer and Fangman, 1987; Hubermanet al., 1987). Given thestructureof theplasmid, the counterclockwise replication fork would be expected to travel through vector sequences and arrive at the 3’end of the 35s transcription unit before the clockwise fork, which is moving in the direction of 35s transcription, arrives there. If transcription by RNA polymerase II occurs
Figure
4. Replication
of the GAL7-35s
Plasmid
(A) Partial restriction map of pNOYl02. Only the restriction sites relevant to this study are shown. (B)A2Dgelofthe5.3 kb Hindlllfragment. Usingaprobefrom pBR322, a simple Y arc with no apparent replication pause site is detected. The shape of the simple Y arc is slightly distorted as a consequence of the larger size of the fragment.
Cell 270
in S phase, then collisions between RNA polymerase II and DNA polymerase would be anticipated in the region of the GAL7 terminator. The replication of the Hindlll restriction fragment containing the GAL7 terminator from this plasmid was examined. No evidence for a new RFB in the region of the GAL7 terminator was found (Figure 4B). The absence of an RFB on the GAL7-3% plasmid might be the consequence of unequal fork rates. As an extreme example, if the origin were not bidirectional in this construct, the entire plasmid might be replicated by a clockwise fork. To measure the direction of fork movement in the vicinity of the 3’ end of the plasmid copy of the 3% transcription unit, we analyzed fragments by a modified 2D gel procedure that includes a second restriction digest between the two dimensions of electrophoresis (Fangman and Brewer, 1991). After the restriction fragments are separated by mass in the first dimension, the entire gel lane is equilibrated with restriction enzyme buffer and then the DNA is digested in situ with a second restriction enzyme. The second enzyme is chosen by the position of its target site within the fragment of interest. In general, it is best if the second enzyme removes 250%40% of one end of the restriction fragment. The remaining shortened fragment is then analyzed by Southern hybridization after the second dimension of electrophoresis. Directional information is deduced by observing the location of the simple Y arc relative to the position of the unbranched fragment. If replication forks first enter the end of the fragment detected by the probe, the simple Y arc would arise from the spot of unbranched linears (Figure 5A). However, if the replication forks first enter the end of the fragment nearest the restriction site used for the second digestion, the simple Y arc detected by the probe is displaced horizontally from the spot of linears (Figure 58). This procedure has been tested previously with the native 2 pm plasmid (Fangman and Brewer, 1991).
A
B HiidIIl
fragmen
-
C 1dl-l
We used in situ cleavage to examine the direction of replication fork movement through the Hindlll fragment located between the 2 pm origin and the 3’end of the 35s transcription unit in plasmid pNOY102 (Figure 4A). The data in Figure 5C show the arc of large simple Y molecules that results from the failure of Aval to cleave some of the Hindlll fragments after the first dimension. The smaller simple Ys created by Aval cleavage produce a single arc that arises from the spot of unbranched molecules (similar to thesituation diagrammed in Figure 5A). Therefore, replication forks first enter this fragment at the end farthest from the Aval site, indicating that this fragment is replicated in the counterclockwise direction. Noforks moving in the opposite direction were detected. Furthermore, similar analysis of the Aval-Ncol fragment that contains the GAL transcription terminator (Figure 4A) indicates that in more than 90% of the plasmid molecules, the counterclockwise moving forks proceed into the 3’ end of the 35s transcription unit at least as far as the Ncol site (data not shown). We conclude that either galactose-induced transcription by RNA polymerase II and replication do not occur at the same time on any given molecule or collisions between RNA polymerase II and the proteins at the replication fork are resolved in favor of replication. Analysis of the RFB on Plasmids Two observations suggest that the rDNA RFB is not a simple consequence of the act of transcription: replication forks arrest at the 3’ end of the chromosomal rDNA 355 transcription units in the absence of RNA polymerase I transcription, and they are not arrested at the 3’ end of a highly active RNA polymerase II 35s transcription unit contained on a plasmid. These results strongly imply that the barrier is independent of transcription and instead is due to a specific sequence found in the rDNA chromosomal locus but not in the plasmid rDNA copies. If this hypothesis were true, the RFB should be able to arrest
2w
origin
GAL-T
35s
Figure 5. Direction mid pNOY 102
of Fork Movement
in Plas-
(A and B) Cartoons of fork direction 2D gels. A Hindlll fragment replicated from either end is illustrated at the top of the figure and molecules that differ in their extent of replication are separated in the first dimension. The position of the Aval site within the Hindlll fragment is shown by the dotted line. After cleavage with Aval in the first dimension gel slice, the second dimension of electrophoreis is performed. The probe used is indicated bythe hatched rectangle. The 2D gel patterns for the Hindlll fragment (thin line), the arc of Hindill linears of various sizes (dashed line), and the Aval-Hindlll replication intermediates (thick line) detected by the probe are illustrated. See text for interpretations. (C) The direction of fork movement in the Hindlll fraament between the 2 urn oriain and the 3’ end-of the 355 transcription u&. The 5.3 kb Hindlll fragment was subjected to first dimension electrophoresis and then digested in the gel with Aval. After the second dimension of electrophoresis, replication intermediates were detected using a probe to the pBR322 sequences. Fewer than 50% of the molecules were cleaved with Aval. The upper arc contains replication intermediates of the entire Hindlll fragment and is comparable to the simple Y arc shown in Figure 4. The lower simple Y arc corresponds to the replication intermediates in the 3.04 kb Hindlll-Aval fragment adjacent to the 2 pm origin. The dashed lines connect the simple Y arcs with their corresponding linear spots.
Replication 271
A
Fork Barrier
pBB6 RFB
Figure
in Yeast
rDNA
B
(+)
pBB6
forks when transplanted to a new context and the polarity of arrest should be maintained. To test this hypothesis, a 950 bp fragment of rDNA containing the site of fork arrest was inserted in both orientations into the plasmid pBB6 adjacent to the AR3 origin (Figures 6A and 6B). The plasmids were introduced into yeast and replication intermediates were then analyzed on 2D gels to determine whether the RFB was capable of arresting forks in an orientationdependent manner when removed from the rDNA locus. As shown in Figure 6C, only one orientation of the insert was effective in arresting fork movement, and that orientation relative to ARS7 is the same as is found in the rDNA between the RFB and the nearby rDNA ARS. In the other orientation of the insert (Figure 6D), there was no accumulation of forks in the rDNA fragment. Recently, Kobayashi et al. (1992) have also demonstrated that the RFB blocks forks in an orientation-dependent manner on a 2 urn plasmid. The ability to detect the RFB on plasmids makes it possible to map easily the sequences essential for the fork arrest. Whileforksarrest in the 129 bp Hindlll-Hpalfragment of the rDNA, the sequences that actually cause the arrest may be some distance away. Deletion analysis of the RFB was carried out on the pBB6-RFB(+) plasmid using convenient restriction sites. The fragments of rDNA remaining on three of the plasmids extend from the EcoRl site to the Hindlll site, the Hpal site, or the Accl site, respectively (Figure 7A). The barrier was present if the EcoRI-Hpal or EcoRI-Accl fragment was present on the plasmid (Figures 78 and 7D) but missing if only the EcoRI-Hindlll fragment was present (Figure 7C). Thus, it would appear that the sequences required to cause the fork barrier lie within the 129 bp between the Hindlll and Hpal sites. The complementary set of deletions was also examined: they included the rDNA sequences between the Hindlll, Hpal, or Accl sites and the Pvull site. The Pvull-Accl and the PvullHpal fragments showed no significant barrier (data not shown). If the 129 bp sequence between Hindlll and Hpal were sufficient to cause fork arrest, we would expect the Hindlll-Pvull fragment to cause forks to arrest. Only avery weak barrier was detected (data not shown). We conclude that the Hindlll-Hpal fragment is necessary but not sufficient for fork arrest (at least in the context of this particular
RFB (-)
6. The RFB on a Plasmid
(A) A map of the plasmid pBB6-RFB(+) is shown. The 950 bp rDNA sequence from EcoRl to Pvull was inserted in the unique EcoRl site of pBB6. (In the process of manipulating the fragment for insertion, the natural Pvull site was destroyed). Forks moving clockwise from ARSI would encounter the RFB with the same polarity encountered by leftward moving forks in the rDNA. (B) A map of the plasmid pBB6-RFB(-) is shown. Forks moving clockwise from AR.9 would encounter the RFB with the same polarity encountered by rightward moving forks in the rDNA. (C)The 2D gel contains a Pvull-EcoRV digest of pBB6-RFB(+). Replication intermediates were detected with an AM7 probe. On this light exposure, bubbles, which are only faintly visible, convert to large simple Ys as the counterclockwise fork passes the EcoRV site. The clockwise fork pauses at the RFB (arrow). (D) The 2D gel of Pvull-EcoRV digested pBB6-RFB(-) plasmid was hybridized with an ARS7 probe. The bubbles and large simple Ys are identical to those found in (C) except that forks do not pause at the RFB. In both autoradiograms, spots along the arcs of linears result from partial digestion of the plasmid and cross-hybridization to the chromosome IV AR.97 sequences. The compact dark spot on the bubble arc in (C) is spurious background hybridization.
C
E
H3HI
A
P
D
Figure 7. Deletion Analysis Plasmid pBB6-RFB(+)
of the
RFB
on
(A)A mapof plasmid pBBGRFB(+)showing the extents of 3 deletions. The horizontal dashed lines labeled 8, C and D indicate the sequences that have been deleted from the plasmids. A, Accl; E, EcoRI; HI, Hpal; H3, Hindlll; P, Pvull. (ED) 2D gel analysis of the three deletions illustrated in (A). The Pvull-EcoRV digested plasmids were analyzed on 2D gels with a probe to ARS7. The deletion shown in (8) contains rDNA from the EcoRl to Hpal sites; the deletion shown in (C) contains rDNA from the EcoRl to Hindlll sites; the deletion shown in (D) contains rDNAfrom the EcoRl to Accl sites. An RFB is evident in (B) and (D) (arrows).
Cell 272
plasmid) and that there appears to be an essential function present in the adjacent EcoRI-Hindlll fragment. Kobayashi et al. (1992) performing a similar analysis on RFB constructs in a 2 pm vector, found that the necessary and sufficient sequences reside in a 109 bp region centrally located in the 129 bp Hindlll-Hpal fragment. This difference in the requirement for flanking sequences may be due to differences in the contexts in the 2 pm and ARSl plasmids. While the barrier apparently functions on the pBB6RFB plasmid, the relative number of stalled intermediates is not as great as in the chromosomal rDNA locus. There are several possible explanations for the apparent reduction in signal from the plasmid copy of the RFB. The 950 bp fragment might not contain all of the DNA sequences essential for full RFB activity. Alternatively, nucleolar localization might be required for full RFB activity, and the plasmid may not be properly localized in the nucleolus. Finally, replication fork rates on the plasmid might not be the same as those in the rDNA locus, leading to a different ratio of simple Ys to stalled forks. To distinguish among these three possibilities, a derivative of the RFB plasmid was integrated into the rDNA. We deleted the ARS7 origin from the plasmid pBC4, which contains a tandem duplication of the RFB fragment (see Experimental Procedures). Partial digestion of the new plasmid by Hpal generated linears that were transformed into yeast, and Ura+ transformants were selected. Integrants were characterized by Southern hybridization using vector sequences as a probe and two classes of transformants were recovered (Figure 8A) depending upon which of the tandem repeats in the pBC4 derivative had been cleaved with Hpal. Both classes of transformants were analyzed for barrier activity at the three RFBs. In the integrant that contains two RFBs adjacent to the 3’ end of the 35s transcription unit, the two RFBs appear to have different efficiencies of fork arrest (Figure 88): the RFB closest to the 35s transcription terminator is a strong barrier, while the RFB adjacent to the vector sequences is equivalent to the barrier on the RFB plasmid. The single RFB adjacent to the 5s gene also appears to be an inefficient barrier (data not shown). However, in the integrant containing two RFBs adjacent to the 5s gene, the two RFBs are equivalent (Figure 8C). In comparison with the single RFB at the 3’end of the 35s transcription unit, these two RFBs appear to be inefficient at stopping forks (data not shown). The efficiency of fork arrest at the chromosomal RFBs in the native rDNA locus and in the RFB-integrant strains was quantitated by measuring the percentage of forks that move leftward after passing the RFB. In unmodified rDNA the portion of the 35s transcription unit immediately upstream of the RFB is replicated predominantly (>90%) by forks moving rightward (Figures 9A and 9D). Thus, in its native context, the RFB is judged to be eff icient at arresting forks: in fewer than 10% of the rDNA repeats the leftwardmoving fork is able to continue past the RFB. However, in the RF6 integrants, the RFBs to the right of the vector sequences are diminished in their ability to arrest forks. In the integrant in which there is a single RFB rightward of
A
RFB NarI t
RFB
URA3 -v ____-,______
PSI
B w Ph -q
NarI
BglI
BglII
I
-‘--,-‘---Pal
BglI I
C
BglII I
probe
B
Figure
9. Analysis
of Integrated
Tandem
Copies
of the RFB
(A) Maps of two transformants recovered after integration of a ptasmid containing a tandem repeat of the EcoRI-Pvull RFB fragment. Because the integration event with a chromosomat copy of the RFB fragment could occur at either of the two RFB copies on the plasmid, two different arrangements of the inserted sequences were found (top and bottom). The striped bar represents the Hindlll-Hpal fragment containing essential RFB sequences. The gray bar represents the 95Obp EcoRI-Pvull fragment. Vector sequences are indicated by the dashed line. (B) The 4.1 kb Narl-Pstl fragment of the transformant shown at the top of (A) was analyzed on a 2D gel by using a probe to the URAI gene. There are two points of arrest that correspond to the positions of the two RFBs. The RFB adjacent to the 3’end of the 3% transcription unit is the more efficient barrier. (C) The 4.9 kb Bgll-Bglll fragment of the transformant shown at the bottom of (A) was analyzed on a 2D gel by using a probe to a portion of the pUC vector sequences. There are two points of fork arrest that correspond to the positions of the two RF&.
the vector, most forks (700%80%) are capable of traversing the barrier (Figures 9B and 9E). When a second RFB is located rightward of the vector, approximately 50% of the forks pass the two barriers (Figures 9C and 9F). These results suggest that neither poor nucleolar localization nor differences in fork rate are responsible for the differences in the barrier seen between the plasmid copy of the RFB and the native chromosomal RFB. Instead, the results indicate that there may be some additional sequences required for an efficient barrier that are not present on the 950 bp fragment, and, further, that the sequences required for an efficient barrier map to the left of this fragment. RFB fragments that extend an additional 250 bp into the 35s transcription unit do not show an increase in barrier strength in plasmid constructs (K. Friedman and W. L. F., unpublished data); however, plasmids with an entire 35s transcription unit display a wild-type barrier (Y. Lim and B. J. B., unpublished data). We are currently searching for the additional sequences needed for a fully functional RFB.
Replication 273
Fork Barrier
in Yeast
rDNA
A Nhel
B
,,,w
. . .. . . . . . . . . . . . . . . . RI Need
October
16, 1992, Copyright
0 1992 by Cell Press
The Arrest of Replication Forks in the rDNA of Yeast Occurs Independently of Transcription Bonita J. Brewer, Daniel Lockshon, and Walton L. Fangman Department of Genetics, SK-50 University of Washington Seattle, Washington 98195
Summary Replication forks, moving opposite to the direction of transcription, are arrested at the 3’ ends of the 3% transcription units in the rDNA locus of S. cerevisiae. Because of its position and polarity, we tested the hypothesis that this replication fork barrier (RFB) results from the act of transcrlption. Three results contradict this hypothesis. First, the RFB persists in a strain containing a disruption of the gene for the 135 kd subunit of RNA polymerase I. Second, the RFB causes a polar arrest of repliccltion forks when transplanted to a plasmid. Third, transcrlption by RNA polymerase II of a piasmid copy of the 35s transcription unit lacking the RFB does not generate a barrier. We propose that repiication forks are arrested in a directional manner through the binding of one or more proteins to two closely spaced sites in the RFB. Introduction The rDNA locus of yeast is an approximately 2 Mb region of chromosome XII that consists of tandemly repeated 9 kb units, each of which encodes the four structural RNAs of the ribosome (reviewed in Warner, 1989; Figure 1A). The 8800 base 3% rRNA that is a precursor to the 18,25, and 5.8s rRNAs is transcribed by RNA poiymerase I in a direction that is opposite to the transcription of the 150 base 5s gene by RNA poiymerase III. Transcribed sequences are interrupted by two nontranscribed spacers (Figure lA), one between the 3’ ends of the transcripts (NTSl) and a second, containing a potential origin of repiication (rDNA ARS), between the 5’ends of the transcripts (NTS2). The replication of the rDNA locus has been studied by analyzing replication intermediates on two-dimensional (2D) gels (Linskens and Huberman, 1988; Brewer and Fangman, 1988). Replication intermediates that contain an origin of replication have two diverging replication forks and consist of a series of branched molecules referred to as bubbles. Under the appropriate conditions of eiectrophoresis, a series of bubble intermediates migrates on 2D gels in a characteristic manner and produces the arc of hybridization illustrated in Figure 16 (Brewer and Fangman, 1987). Fragments that lie between origins may be replicated by a single fork, moving through the fragment from one end to the other, and thus during their replication a series of branched molecules called simple Ys is generated. The migration of simple Ys on 2D gels is clearly distinguishable from the migration of bubbles (Figure 1C).
By analyzing restriction fragments from different portions of the rDNA repeating unit, it has been shown that repiication begins in the NTSP, but in only a subset of the repeats (Linskens and Huberman, 1988; Brewer and Fangman, 1988; Fangman and Brewer, 1991). Repeats that do not contain an active origin are replicated passively by a replication fork that originates in a nearby repeat. A unique feature of the rDNA locus is that the two forks initiating at an rDNA origin experience unequal fates. The fork moving through the 35s transcription unit in the same direction as RNA poiymerase I transcription proceeds unimpeded through multiple repeats, while the opposite fork proceeds through the adjacent 5s gene and NTSl and then arrests at a site (the RFB) near the transcription terminator for the adjacent, upstream 3% transcription unit. The arrest of forks at the RFB is deduced from the accumulation of replication intermediates of a specific size. When DNA is harvested from an asynchronous culture, all possible stages of replication intermediates are represented in the sample. Therefore, the amount of any particular intermediate reflects the amount of time the replication complex idles at a specific point in the restriction fragment. Replication intermediates within the 35s region produce a uniform distribution of simple Y intermediates on a 2D gel. in contrast, restriction fragments that contain the RF6 accumulate a class of replication intermediates that generate an intense hybridization signal at a unique spot along the otherwise uniform arc of simple Y replication intermediates (Figure 1D). The location of the spot along the arc of simple Ys varies with the choice of restriction enzyme, and, therefore, the position of the spot can be used to map the site of fork arrest and also can be used to determine the direction the replication forks were moving before their arrest. From this type of analysis it was deduced that the RFB arrests forks in a polar manner (Brewer and Fangman, 1988). Using an alkaline second dimension in their2D gels, Linskensand Huberman (1988) came to similar conclusions. Because of its polarity and its location near the transcription terminator for RNA poiymerase I, it was suggested that transcription itself might be responsible for the arrest of replication forks. In this report we provide further characterization of the RFB. In particular, we have investigated the possibility that the barrier isdue toacoliision between RNA poiymerase molecules and the replication complex. Results Location of the RFB The RFB previously had been mapped to a subregion of the nontranscribed spacer between the 3’ends of the 35s and 5s transcription units, at position 7020 f 80 bp (Figure 2A; Brewer and Fangman, 1988). To localize the pOSition of fork arrest more precisely, total cellular DNA was digested with restriction enzymes, and fragments of approximately 2 kb were analyzed by 2D gel electrophoresis and hybridization with a probe from NTSl For fragments
Cdl 266
A.
rDNA RFFI 4
NT.7
d-“”
*
3 kb
““T ARS
B.
BUBBLE
C.
Figure 1, Cartoons of Yeast Intermediates on 2D Gels
SIMPLE
rDNA
Y
D.
and the Migration
BARRIER
of Replication
(A) Illustrated are two repeats of the 9.06 kb rDNA unit from chromosome XII showing the positions of transcription units (35S and 5s) and the positions of the rDNA ARS (in NTSP) and the RFB (in NTSl). (B)The expected hybridization pattern for the replication of a restriction fragment that contains an origin of replication at its center. The replication intermediates depart from the arc of linears (broken line) at the position of unbranched, unit-length linears (spot in lower right corner). The first dimension is from left to right; the second dimension is from top to bottom. (See Experimental Procedures, and Brewer and Fangman [I9671 for details.) (C) The expected hybridization pattern for a restriction fragment that is replicated passively by a fork entering from one end. The replication intermediates depart from the arc of linears at the position of unbranched molecules and then return to it, as the fragment nears the completion of its replication, at a position of twice the length of the fragment. (D) The expected hybridization pattern for a fragment containing an impediment to replication fork movement. The accumulation of replication intermediatesof a specific size results in an increase in the hybridization signal at a specific place along the simple Y arc.
6 (Hindlll-SnaBI) and C (Hpal-Bglll), the rightward end of the fragment is nearthe origin and the leftward end isclose to the site where the RFB was mapped previously (Figure 2A). If the barrier were within the fragment, intermediates
would accumulate that were nearly fully replicated; however, if the RFB were to lie outside the fragment, only a uniform arc of simple Ys would be observed. The results of these two digests are shown in Figures 28 and 2C. The Hindlll-SnaBI fragment replicates as a series of simple Ys with a prominent accumulation of intermediates near the position of 4.2 kb linears (arrow), indicating that the RFB is present on this fragment and that the barrier is very near the Hindlll site. The replication intermediates of the HpalBglll fragment produce a continuous arc of simple Y intermediates with no apparent pause site. We therefore conclude that forks are arresting in the 129 bp between the Hindlll and Hpal sites that define the left ends of fragments B and C (Figure 2A). Closer inspection of the barrier region of the HindlllSnaBl gel (Figure 28) reveals that the hybridization signal from the RFB is not spherical but somewhat elongated. To analyze the RFB at higher resolution, the smaller fragment D (Nhel-Pvull; Figure 2A) was examined. We observed what appeared to be two discrete points along the simple Y arc at which forks arrest (Figure 20). The origin-proximal spot is more intense than the distal one. We cannot determine whether any given fork is arrested twice or is arrested at only one site or the other. The RFB in an RNA Polymerase I Mutant Strain Transcribing rDNAchromatin is readily recognizable in the electron microscope by the “Christmas tree” appearance of the densely packed, nascent transcripts associated with the DNA fiber. Saffer and Miller (1988) found examples of actively transcribed rDNA repeats within regions of replicating chromatin, providing indisputable evidence that the two potentially conflicting processes can occur simultaneously on a given DNA molecule. Perhaps the simplest model to account for both the position and polarity of the RFB is that collision with RNA polymerase I prevents repliFigure
2. Mapping
of the RFB
(A) Detailed restriction map of the NTS region I I of the rDNA repeat. Arrows correspond to the 9 9.08/O kb 6 I 8 35s precursor and the 5S gene. Within the 3% region, the 3’ end of the mature 26s rRNA seI IB quences is indicated by the dark bar. RestricI IC I ID tion sites: A. Accl; 8. Balll: E, EcoRI: HI. Hpal: H3, HindIll;. M, Smal;N, Nhel; P, Pvullf S, D NM-PVUII B HindIII-SnaBI c HpaI-BglII SnaBI. The bars labeled B, C, and D, refer to restriction fragments analyzed in panels (B), (C), and (D), respectively. (B) A 2D gel analyzing the Hindlll-SnaBl frag ment (fragment B) from the rDNA NTS region. The fragment was detected by using the HpalPvull fragment as a probe. The simple Y arc , begins in the 2.1 kb dark spot in the lower right corner and returns to the arc of linears at a size of approximately 4.2 kb. The accumulated replication intermediates are indicated by the arrow. The additional spot on the arc of linears is most likely the result of a partial digestion. (C)A 2D gel analyzing the Hpal-Bglll fragment (fragment C) from the rDNA NTS region. The same probe as in (B) was used. Again, the additional spots on the arc of linears are the result of partial digestion. (D) A 2D gel of the 1.2 kb Nhel-Pvull fragment (fragment 0) from NTSl. A light exposure is shown to reveal the appearance of two arrest points at the RFB. The dashed line marks the position of the arc of linears; the dotted line marks the position of the simple Y arc. Both lines were traced from a longer exposure.
Replication 269
Fork Barrier
in Yeast
rDNA
cation forks from entering the 3’ end of the 35s transcription unit (Brewer and Fangman, 1988). Since the initiation of transcription of 35s rRNA occurs at a rate sufficient to keep the template entirely covered by RNA polymerase molecules, if a replication fork were to approach such an actively transcribed region from its 3’end, the enzymes at the fork would repeatedly encounter RNA polymerases at the transcription terminator. We previously proposed that such repeated encounters may be sufficient to block the progress of this replication fork. The 35s region would then be replicated by a fork moving in from the C/end, i.e., moving in the same direction as RNA polymerase I. To determine the role of RNA polymerase I in the arrest of forks in the rDNA locus, we analyzed a yeast strain that has a disruption in the gene for a catalytic subunit of RNA polymerase I (RPA735; Nogi et al., 1991). This disruption is lethal, but the strain can be maintained by the presence of a multicopy plasmid containing the 35s transcription unit under the control of the RNA polymerase II promoter, GAL7(for a plasmid map see Figure 4A). Viability is absolutely dependent on the presence of galactose. The addition of glucose decreases the rate of 18s and 25s transcription more than 1Bfold (Nogi et al, 1991). The residual transcription that remains after glucose repression is probably derived from the plasmid, since the GAL7 promoter is not efficiently repressed when present in high copy numbers (Baker et al., 1987). We analyzed the chromosomal copies of the rDNA repeat in the rpa735::LEU2 strain. First, we discovered through Southern hybridization that in the absence of RNA polymerase I transcription, the rDNA locus had undergone a deletion of more than 80% of its repeats; however, the repeats that remained appeared unrearranged (data not shown). Using a probe to the nontranscribed spacer to study only the chromosomal copies of the rDNA, we found that the replication fork arrest at the RFB was indistinguishable from that found in the isogenic wild-type rDNA locus (Figure 3). All aspects of replication assayed were indistinguishable from the wild-type isogenic strain: the efficiency of initiation (data not shown), the intensity of the RFB relative to simple Y intermediates and the unbranched monomer spot (Figure 3) and the appearance of the double barrier in the RFB fragment (data not shown). The simplest interpretation of these data is that the act of transcribing the 35s rRNA by RNA polymerase I is not responsible for blocking the movement of replication forks into the 3’ end of the 35s transcription unit. Replication of the GAL7- 3% RNA Polymerese II Transcription Unit To determine whether transcription is capable of arresting replication forks, we examined the replication of the 35s transcription unit on the GAL7-35s plasmid in the rpa735::LfU2 strain. On this plasmid, the 35s rDNA is located between the promoter and transcription terminator of the GAL7 gene (Figure 4A). The rDNA insert, which extends only to the Hindlll site at 8855 bp (Figure 2A), includes the processing site to generate the mature end of the 25s rRNA, but lacks the signals for RNA polymerase I transcription initiation and termination, and does not in-
C. rpal3kUXL2
Figure
3. The RFB in the Absence
of Polymerase
I Transcription
(A) Bglll map of the rDNA. (E) 2D gel analysis of the 4.6 kb Bglll rDNA fragment from wild-type cells. The RFB region is at the center of this fragment. The second spot of hybridization on the linear arc at 9.1 kb is the result of an incomplete digestion by Bglil. (C) 2D gel analysis of the 4.6 kb Bglll rDNA fragment from cells carrying the rpa735::LEU2 allele. Since the repeat number of rDNA units is approximately 6-fold lower in this strain as compared with the wild-type strain (data not shown), a longer exposure was required to achieve a similar intensity of hybridization to the replication intermediates. The additional spot on the arc of linears is probably the junction fragment with non-rDNA sequences.
elude the Hindlll-Hpal RFB sequence (Figure 2A). Based on past studies, replication from the 2 pm origin is expected to be bidirectional (Brewer and Fangman, 1987; Hubermanet al., 1987). Given thestructureof theplasmid, the counterclockwise replication fork would be expected to travel through vector sequences and arrive at the 3’end of the 35s transcription unit before the clockwise fork, which is moving in the direction of 35s transcription, arrives there. If transcription by RNA polymerase II occurs
Figure
4. Replication
of the GAL7-35s
Plasmid
(A) Partial restriction map of pNOYl02. Only the restriction sites relevant to this study are shown. (B)A2Dgelofthe5.3 kb Hindlllfragment. Usingaprobefrom pBR322, a simple Y arc with no apparent replication pause site is detected. The shape of the simple Y arc is slightly distorted as a consequence of the larger size of the fragment.
Cell 270
in S phase, then collisions between RNA polymerase II and DNA polymerase would be anticipated in the region of the GAL7 terminator. The replication of the Hindlll restriction fragment containing the GAL7 terminator from this plasmid was examined. No evidence for a new RFB in the region of the GAL7 terminator was found (Figure 4B). The absence of an RFB on the GAL7-3% plasmid might be the consequence of unequal fork rates. As an extreme example, if the origin were not bidirectional in this construct, the entire plasmid might be replicated by a clockwise fork. To measure the direction of fork movement in the vicinity of the 3’ end of the plasmid copy of the 3% transcription unit, we analyzed fragments by a modified 2D gel procedure that includes a second restriction digest between the two dimensions of electrophoresis (Fangman and Brewer, 1991). After the restriction fragments are separated by mass in the first dimension, the entire gel lane is equilibrated with restriction enzyme buffer and then the DNA is digested in situ with a second restriction enzyme. The second enzyme is chosen by the position of its target site within the fragment of interest. In general, it is best if the second enzyme removes 250%40% of one end of the restriction fragment. The remaining shortened fragment is then analyzed by Southern hybridization after the second dimension of electrophoresis. Directional information is deduced by observing the location of the simple Y arc relative to the position of the unbranched fragment. If replication forks first enter the end of the fragment detected by the probe, the simple Y arc would arise from the spot of unbranched linears (Figure 5A). However, if the replication forks first enter the end of the fragment nearest the restriction site used for the second digestion, the simple Y arc detected by the probe is displaced horizontally from the spot of linears (Figure 58). This procedure has been tested previously with the native 2 pm plasmid (Fangman and Brewer, 1991).
A
B HiidIIl
fragmen
-
C 1dl-l
We used in situ cleavage to examine the direction of replication fork movement through the Hindlll fragment located between the 2 pm origin and the 3’end of the 35s transcription unit in plasmid pNOY102 (Figure 4A). The data in Figure 5C show the arc of large simple Y molecules that results from the failure of Aval to cleave some of the Hindlll fragments after the first dimension. The smaller simple Ys created by Aval cleavage produce a single arc that arises from the spot of unbranched molecules (similar to thesituation diagrammed in Figure 5A). Therefore, replication forks first enter this fragment at the end farthest from the Aval site, indicating that this fragment is replicated in the counterclockwise direction. Noforks moving in the opposite direction were detected. Furthermore, similar analysis of the Aval-Ncol fragment that contains the GAL transcription terminator (Figure 4A) indicates that in more than 90% of the plasmid molecules, the counterclockwise moving forks proceed into the 3’ end of the 35s transcription unit at least as far as the Ncol site (data not shown). We conclude that either galactose-induced transcription by RNA polymerase II and replication do not occur at the same time on any given molecule or collisions between RNA polymerase II and the proteins at the replication fork are resolved in favor of replication. Analysis of the RFB on Plasmids Two observations suggest that the rDNA RFB is not a simple consequence of the act of transcription: replication forks arrest at the 3’ end of the chromosomal rDNA 355 transcription units in the absence of RNA polymerase I transcription, and they are not arrested at the 3’ end of a highly active RNA polymerase II 35s transcription unit contained on a plasmid. These results strongly imply that the barrier is independent of transcription and instead is due to a specific sequence found in the rDNA chromosomal locus but not in the plasmid rDNA copies. If this hypothesis were true, the RFB should be able to arrest
2w
origin
GAL-T
35s
Figure 5. Direction mid pNOY 102
of Fork Movement
in Plas-
(A and B) Cartoons of fork direction 2D gels. A Hindlll fragment replicated from either end is illustrated at the top of the figure and molecules that differ in their extent of replication are separated in the first dimension. The position of the Aval site within the Hindlll fragment is shown by the dotted line. After cleavage with Aval in the first dimension gel slice, the second dimension of electrophoreis is performed. The probe used is indicated bythe hatched rectangle. The 2D gel patterns for the Hindlll fragment (thin line), the arc of Hindill linears of various sizes (dashed line), and the Aval-Hindlll replication intermediates (thick line) detected by the probe are illustrated. See text for interpretations. (C) The direction of fork movement in the Hindlll fraament between the 2 urn oriain and the 3’ end-of the 355 transcription u&. The 5.3 kb Hindlll fragment was subjected to first dimension electrophoresis and then digested in the gel with Aval. After the second dimension of electrophoresis, replication intermediates were detected using a probe to the pBR322 sequences. Fewer than 50% of the molecules were cleaved with Aval. The upper arc contains replication intermediates of the entire Hindlll fragment and is comparable to the simple Y arc shown in Figure 4. The lower simple Y arc corresponds to the replication intermediates in the 3.04 kb Hindlll-Aval fragment adjacent to the 2 pm origin. The dashed lines connect the simple Y arcs with their corresponding linear spots.
Replication 271
A
Fork Barrier
pBB6 RFB
Figure
in Yeast
rDNA
B
(+)
pBB6
forks when transplanted to a new context and the polarity of arrest should be maintained. To test this hypothesis, a 950 bp fragment of rDNA containing the site of fork arrest was inserted in both orientations into the plasmid pBB6 adjacent to the AR3 origin (Figures 6A and 6B). The plasmids were introduced into yeast and replication intermediates were then analyzed on 2D gels to determine whether the RFB was capable of arresting forks in an orientationdependent manner when removed from the rDNA locus. As shown in Figure 6C, only one orientation of the insert was effective in arresting fork movement, and that orientation relative to ARS7 is the same as is found in the rDNA between the RFB and the nearby rDNA ARS. In the other orientation of the insert (Figure 6D), there was no accumulation of forks in the rDNA fragment. Recently, Kobayashi et al. (1992) have also demonstrated that the RFB blocks forks in an orientation-dependent manner on a 2 urn plasmid. The ability to detect the RFB on plasmids makes it possible to map easily the sequences essential for the fork arrest. Whileforksarrest in the 129 bp Hindlll-Hpalfragment of the rDNA, the sequences that actually cause the arrest may be some distance away. Deletion analysis of the RFB was carried out on the pBB6-RFB(+) plasmid using convenient restriction sites. The fragments of rDNA remaining on three of the plasmids extend from the EcoRl site to the Hindlll site, the Hpal site, or the Accl site, respectively (Figure 7A). The barrier was present if the EcoRI-Hpal or EcoRI-Accl fragment was present on the plasmid (Figures 78 and 7D) but missing if only the EcoRI-Hindlll fragment was present (Figure 7C). Thus, it would appear that the sequences required to cause the fork barrier lie within the 129 bp between the Hindlll and Hpal sites. The complementary set of deletions was also examined: they included the rDNA sequences between the Hindlll, Hpal, or Accl sites and the Pvull site. The Pvull-Accl and the PvullHpal fragments showed no significant barrier (data not shown). If the 129 bp sequence between Hindlll and Hpal were sufficient to cause fork arrest, we would expect the Hindlll-Pvull fragment to cause forks to arrest. Only avery weak barrier was detected (data not shown). We conclude that the Hindlll-Hpal fragment is necessary but not sufficient for fork arrest (at least in the context of this particular
RFB (-)
6. The RFB on a Plasmid
(A) A map of the plasmid pBB6-RFB(+) is shown. The 950 bp rDNA sequence from EcoRl to Pvull was inserted in the unique EcoRl site of pBB6. (In the process of manipulating the fragment for insertion, the natural Pvull site was destroyed). Forks moving clockwise from ARSI would encounter the RFB with the same polarity encountered by leftward moving forks in the rDNA. (B) A map of the plasmid pBB6-RFB(-) is shown. Forks moving clockwise from AR.9 would encounter the RFB with the same polarity encountered by rightward moving forks in the rDNA. (C)The 2D gel contains a Pvull-EcoRV digest of pBB6-RFB(+). Replication intermediates were detected with an AM7 probe. On this light exposure, bubbles, which are only faintly visible, convert to large simple Ys as the counterclockwise fork passes the EcoRV site. The clockwise fork pauses at the RFB (arrow). (D) The 2D gel of Pvull-EcoRV digested pBB6-RFB(-) plasmid was hybridized with an ARS7 probe. The bubbles and large simple Ys are identical to those found in (C) except that forks do not pause at the RFB. In both autoradiograms, spots along the arcs of linears result from partial digestion of the plasmid and cross-hybridization to the chromosome IV AR.97 sequences. The compact dark spot on the bubble arc in (C) is spurious background hybridization.
C
E
H3HI
A
P
D
Figure 7. Deletion Analysis Plasmid pBB6-RFB(+)
of the
RFB
on
(A)A mapof plasmid pBBGRFB(+)showing the extents of 3 deletions. The horizontal dashed lines labeled 8, C and D indicate the sequences that have been deleted from the plasmids. A, Accl; E, EcoRI; HI, Hpal; H3, Hindlll; P, Pvull. (ED) 2D gel analysis of the three deletions illustrated in (A). The Pvull-EcoRV digested plasmids were analyzed on 2D gels with a probe to ARS7. The deletion shown in (8) contains rDNA from the EcoRl to Hpal sites; the deletion shown in (C) contains rDNA from the EcoRl to Hindlll sites; the deletion shown in (D) contains rDNAfrom the EcoRl to Accl sites. An RFB is evident in (B) and (D) (arrows).
Cell 272
plasmid) and that there appears to be an essential function present in the adjacent EcoRI-Hindlll fragment. Kobayashi et al. (1992) performing a similar analysis on RFB constructs in a 2 pm vector, found that the necessary and sufficient sequences reside in a 109 bp region centrally located in the 129 bp Hindlll-Hpal fragment. This difference in the requirement for flanking sequences may be due to differences in the contexts in the 2 pm and ARSl plasmids. While the barrier apparently functions on the pBB6RFB plasmid, the relative number of stalled intermediates is not as great as in the chromosomal rDNA locus. There are several possible explanations for the apparent reduction in signal from the plasmid copy of the RFB. The 950 bp fragment might not contain all of the DNA sequences essential for full RFB activity. Alternatively, nucleolar localization might be required for full RFB activity, and the plasmid may not be properly localized in the nucleolus. Finally, replication fork rates on the plasmid might not be the same as those in the rDNA locus, leading to a different ratio of simple Ys to stalled forks. To distinguish among these three possibilities, a derivative of the RFB plasmid was integrated into the rDNA. We deleted the ARS7 origin from the plasmid pBC4, which contains a tandem duplication of the RFB fragment (see Experimental Procedures). Partial digestion of the new plasmid by Hpal generated linears that were transformed into yeast, and Ura+ transformants were selected. Integrants were characterized by Southern hybridization using vector sequences as a probe and two classes of transformants were recovered (Figure 8A) depending upon which of the tandem repeats in the pBC4 derivative had been cleaved with Hpal. Both classes of transformants were analyzed for barrier activity at the three RFBs. In the integrant that contains two RFBs adjacent to the 3’ end of the 35s transcription unit, the two RFBs appear to have different efficiencies of fork arrest (Figure 88): the RFB closest to the 35s transcription terminator is a strong barrier, while the RFB adjacent to the vector sequences is equivalent to the barrier on the RFB plasmid. The single RFB adjacent to the 5s gene also appears to be an inefficient barrier (data not shown). However, in the integrant containing two RFBs adjacent to the 5s gene, the two RFBs are equivalent (Figure 8C). In comparison with the single RFB at the 3’end of the 35s transcription unit, these two RFBs appear to be inefficient at stopping forks (data not shown). The efficiency of fork arrest at the chromosomal RFBs in the native rDNA locus and in the RFB-integrant strains was quantitated by measuring the percentage of forks that move leftward after passing the RFB. In unmodified rDNA the portion of the 35s transcription unit immediately upstream of the RFB is replicated predominantly (>90%) by forks moving rightward (Figures 9A and 9D). Thus, in its native context, the RFB is judged to be eff icient at arresting forks: in fewer than 10% of the rDNA repeats the leftwardmoving fork is able to continue past the RFB. However, in the RF6 integrants, the RFBs to the right of the vector sequences are diminished in their ability to arrest forks. In the integrant in which there is a single RFB rightward of
A
RFB NarI t
RFB
URA3 -v ____-,______
PSI
B w Ph -q
NarI
BglI
BglII
I
-‘--,-‘---Pal
BglI I
C
BglII I
probe
B
Figure
9. Analysis
of Integrated
Tandem
Copies
of the RFB
(A) Maps of two transformants recovered after integration of a ptasmid containing a tandem repeat of the EcoRI-Pvull RFB fragment. Because the integration event with a chromosomat copy of the RFB fragment could occur at either of the two RFB copies on the plasmid, two different arrangements of the inserted sequences were found (top and bottom). The striped bar represents the Hindlll-Hpal fragment containing essential RFB sequences. The gray bar represents the 95Obp EcoRI-Pvull fragment. Vector sequences are indicated by the dashed line. (B) The 4.1 kb Narl-Pstl fragment of the transformant shown at the top of (A) was analyzed on a 2D gel by using a probe to the URAI gene. There are two points of arrest that correspond to the positions of the two RFBs. The RFB adjacent to the 3’end of the 3% transcription unit is the more efficient barrier. (C) The 4.9 kb Bgll-Bglll fragment of the transformant shown at the bottom of (A) was analyzed on a 2D gel by using a probe to a portion of the pUC vector sequences. There are two points of fork arrest that correspond to the positions of the two RF&.
the vector, most forks (700%80%) are capable of traversing the barrier (Figures 9B and 9E). When a second RFB is located rightward of the vector, approximately 50% of the forks pass the two barriers (Figures 9C and 9F). These results suggest that neither poor nucleolar localization nor differences in fork rate are responsible for the differences in the barrier seen between the plasmid copy of the RFB and the native chromosomal RFB. Instead, the results indicate that there may be some additional sequences required for an efficient barrier that are not present on the 950 bp fragment, and, further, that the sequences required for an efficient barrier map to the left of this fragment. RFB fragments that extend an additional 250 bp into the 35s transcription unit do not show an increase in barrier strength in plasmid constructs (K. Friedman and W. L. F., unpublished data); however, plasmids with an entire 35s transcription unit display a wild-type barrier (Y. Lim and B. J. B., unpublished data). We are currently searching for the additional sequences needed for a fully functional RFB.
Replication 273
Fork Barrier
in Yeast
rDNA
A Nhel
B
,,,w
. . .. . . . . . . . . . . . . . . . RI Need
