Pat1: a topoisomerase II-associated protein required for faithful chromosome transmission in
Saccharomyces cerevisiae
Pat1: a topoisomerase II-associated protein required for faithful chromosome transmission in Saccharomyces cerevisiae
Xiaoqi
Wang
,
Paul M.
Watt
+
,
Edward J.
Louis
1
,
Rhona H.
Borts
1
and
Ian D.
Hickson*
Imperial Cancer Research Fund and
1
Yeast Genetics, Institute of Molecular Medicine, John Radcliffe Hospital,
Oxford
OX3 9DU,
UK
Received July 23, 1996;
Revised and Accepted October 25, 1996
ABSTRACT
Saccharomyces cerevisiae
top2
mutants deficient in topoisomerase II activity are defective in chromosome
segregation during both mitotic and meiotic cell divisions. To identify proteins that act in concert with topoisomerase II during
chromosome segregation in
S.cerevisiae
, we have used a two-hybrid cloning approach. We report the isolation of the
PAT1
gene (for protein associated with topoisomerase II), which encodes a novel 90
kDa proline- and glutamine-rich protein that interacts with a highly conserved, leucine-rich region of topoisomerase II
in vivo
. Strains lacking Pat1p exhibit a slow growth rate and a phenotype reminiscent
of conditional
top2
mutants grown at the semi-permissive temperature; most notably, a reduced fidelity of chromosome
segregation during both mitosis and meiosis. These findings indicate that the
PAT1
gene is necessary for accurate chromosome transmission during cell division in
eukaryotic cells and suggest that the interaction of Pat1p and topoisomerase II
is an important component of this function.
INTRODUCTION
Each cell division cycle in eukaryotes consists of two key phases during which
the genome is first precisely replicated and then evenly segregated between the
daughter cells. There are several threats to the integrity of the genome from
exogenous damaging agents, such as ionizing and ultraviolet radiation, but one
of the major sources of genomic instability are the processes of DNA
metabolism. Errors incurred either during DNA replication and recombination, or
during the segregation of chromosomes at mitosis, can lead to many different
forms of chromosomal aberrations, including deletions and unequal exchanges
between homologous sequences (for reviews see
1
-
4
).
Topoisomerase II is an important component of the cellular armoury that
counteracts threats to the structural integrity of chromosomes. Conditional
mutants of budding and fission yeast lacking topoisomerase II activity at the
restrictive temperature die in the course of an abortive mitosis in which
chromosomal breakage results from an attempt to segregate sister-chromatids that have not been disentangled following DNA replication (
5
-
8
). Mutants lacking a functional
TOP2
gene also fail to proceed through meiosis, although in this case the cells
arrest at the pachytene stage of meiosis I without a loss of viability (
9
,
10
). In addition to this essential function in cell division, topoisomerase II is
required both for the relief of torsional stress that can build up during DNA
transcription and replication, and for the structural maintenance of
chromosomal loci comprising repetitive DNA elements (reviewed in
11
-
13
). This latter role is particularly highlighted in the case of the tandemly
duplicated rDNA locus, which is highly unstable in
top2
mutants (
14
). This `suppression' of DNA repeat instability by topoisomerase II is
apparently a function shared with topoisomerase I, since impairment of the
function of both of these classes of topoisomerases causes a synergistic effect
on rDNA locus recombination, and leads to the excision of rDNA elements in the
form of extrachromosomal DNA rings (
15
).
In order to identify gene products that either cooperate with, or regulate the
functions of, topoisomerase II during different aspects of DNA metabolism in
Saccharomyces cerevisiae,
we have used a two-hybrid cloning approach. We have previously described the isolation of the
SGS1
gene encoding a putative DNA helicase homologous to the
Escherichia coli
RecQ, and human BLM (Bloom's syndrome) and WRN (Werner's syndrome) proteins, which interacts with topoisomerase II
in vivo
(
16
). Strains lacking functional Sgs1p show a defect in chromosome segregation and
an elevated frequency of recombination (
16
,
17
). Here, we report the identification of the
S.cerevisiae
PAT1
gene product via its ability to interact with a leucine-rich region of the topoisomerase II protein. We show that the phenotype of
[Delta]
pat1
strains is very similar to that of
top2
ts
mutants grown at the semi-permissive temperature (
11
-
13
); most notably a reduction in cell viability and a decrease in the fidelity of
chromosome segregation during mitosis and meiosis.
MATERIALS AND METHODS
Saccharomyces cerevisiae
strains
The genotypes of strains used in this work are listed in Table
1
. Strain EGY48 (
23
) was obtained from Dr R. Brent. Strain YPH277 (
18
) was obtained from Dr P. Heiter. All of the remaining strains are based on the
YP1 (
19
) and Y55 (
20
) backgrounds. Strains PW30, PW40 and PW50 have been described previously (
16
,
17
). Strain XQ2 was generated by crossing strains Y55-2372 and PW30 [Delta]
pat1
and selecting segregants after tetrad dissection that were either Leu
+
Trp
+
can
r
(for a [Delta]
pat1
segregant) or Leu
-
Trp
+
can
r
(for a wild-type control). To generate the homozygous [Delta]
pat1
diploid (XQD20) for meiosis missegregation studies, XQ2 [Delta]
pat1
was crossed with PW40 [Delta]
pat1
selecting for Trp
+
Ura
+
. The appropriate wild-type control diploid was generated by crossing XQ2 with PW40. XQD30 was
generated by crossing strains YPH277 and PW50. The XQD40 diploid, and a
derivative containing a homozygous deletion of
PAT1
, were generated by crossing strains EJL374-8A and PW50, or their respective [Delta]
pat1
derivatives.
Isogenic deletions of the
PAT1
gene in the above strains were made by transforming (
21
) with pXQ101 (containing
LEU2
inserted into a partially deleted
PAT1
coding region; see below),
which had been digested with
Xba
I and
Bgl
II. Diploid strains were constructed by mating, followed by appropriate
selection, as well as by confirmation of a non-mating phenotype.
Growth of micro-organisms/DNA manipulation
Growth of
E.coli
and yeast and standard recombinant DNA techniques were as described by Ausubel
et al
. (
21
) and Sherman (
22
). Nucleotide sequencing was performed using the dideoxy chain termination
method and Sequenase (US Biochemicals).
Plasmids
The plasmid pJG45 (
23
), in which the activation library was constructed (
16
), together with the LexA-fusion DNA binding domain vector (pEG202) and the lacZ reporter plasmid
(pSH1834), were kindly provided by Dr R. Brent. pLexTopDT was constructed by
cloning an
Ssp
I fragment of
TOP2
(representing residues 1118-1429), in frame with
lexA
into pEG202. pLexTopD and pLexTopT constructs (which contain residues 1109-1163 and 1168-1429, respectively) were made similarly, except that in the
construction of pLexTopT, an octameric
Eco
RI linker was inserted into the
Eco
RV site of
TOP2
, to create the in-frame 5' cloning junction. The non-specific LexA fusion plasmids; pLexA-Max (
24
) and pHM12 (a kind gift from Drs R. Finley and R. Brent), contain the entire
coding region of the human Max protein and 295 residues of
Drosophila
Cdc2 kinase, respectively.
To generate pXQ101 for targeted disruption of
PAT1
, the full 2391 bp coding region of
PAT1
was amplified from genomic
S.cerevisiae
DNA (S288C, ATCC#26108) by PCR using the following primers (both written 5' -> 3') and the product was cloned into pCR-Script
TM
SK(+), using a kit supplied by the manufacturer (Stratagene).
5' primer: GAAGCACTAGCAATGTCCTTCTTTGGG
3' primer: CTTTAGTTCTGATATTTCAGC
The resulting construct was then digested with
Xba
I and
Bgl
II (deleting 1068 bp of the
PAT1
coding
sequence), prior to the insertion of a 2230 bp fragment containing the
LEU2
gene to generate pXQ101.
A
PAT1
clone was also isolated using standard techniques (
21
) from an
S.cerevisiae
genomic DNA library (kindly supplied by Dr J. Diffley), using the pActPat1
insert as the probe. To generate pXQ102 for expression of Pat1p in
S.cerevisiae
, the full length
PAT1
coding region was excised from pCR-Script (see above) and inserted into the
Xho
I site of pYES (Invitrogen).
Activator fusion library construction
A library of
S.cerevisiae
genomic DNA comprising 3-5 * 10
6
primary transformants was constructed as described by Watt
et al
. (
16
). Plasmid DNA was prepared directly from the pooled, unamplified
E.coli
transformants.
Two-hybrid screen
The two-hybrid screen was performed essentially as described by Gyuris
et al
. (
23
) and Zervos
et al
. (
24
). The yeast activation domain library described above was transformed into the
reporter strain EGY48, which had previously been transformed with pLexTopDT and
pSH1834. Plasmids were isolated from yeast which survived selection for leucine prototrophy and were blue on plates
containing X-gal only when the medium was supplemented with galactose (to induce expression from the
Gal1
promoter in pActPat1).
Immunoprecipitations
Strain EGY48 was transformed with plasmids pActPat1 (encoding HA-tagged Pat1p) and pTopoII-myc (encoding myc-tagged topo II; ref.
25
). Yeast were lysed and extracts were separated on SDS-polyacrylamide gels using the discontinuous system of Laemmli (
26
). Immunoprecipitations were carried out using either anti-myc (9E10) antibody, anti-HA (12CA5) antibody or an anti-ICAM1 control antibody of the same subclass (IgG1). Western
blots were probed with either the 9E10 anti-myc antibody or the 12CA5 anti-HA antibody, using standard techniques (
21
).
Mitotic missegregation
Control strain YPH277, and a derivative with a deletion of the
PAT1
gene, contain a chromosomal fragment marked with the
URA3
gene. Following selection of colonies on SC medium lacking uracil, to ensure
maintenance of the chromosomal fragment, the cells were streaked onto
nonselective medium. Entire single colonies from these plates were then picked
and re-plated on both SC medium containing 5-fluoroorotic acid (5-FOA) at 1 mg/ml, and on SC medium alone. Fluctuation analyses
were performed as described by Lea and Coulson (
27
) using the method of the median. CHEF gel analysis of DNA from selected 5-FOA resistant colonies was performed using standard techniques (
21
).
Meiosis I missegregation
Diploids were selected on SC medium (-Trp -Ura) and single colonies were used to prepare patches on a non-selective YPD plate. Yeast were replica-plated onto sporulation medium and the plates were
incubated at 20oC for 5 days. Gluculase (NEN) treated spores were sonicated briefly, before
plating dilutions both on SC medium and on SC (-Ura -Trp), in each case containing canavanine at 0.004%. The frequency
of meiosis I missegregation was calculated as the ratio of total (Trp
+
Ura
+
can
r
) spores to total can
r
spores.
MAT
[alpha],
ura3-52, leu2
[Delta]
, ade2-101, lys2, cyh2
YPH277
MAT
a
,
ura3-52, lys2-801, ade2-101, trp1
[Delta]
1, leu2
[Delta]
1, CRVII
(RAD2.d.YPH277) URA3 SUP11
PW30
MAT
[alpha],
ura3-n, leu2
[Delta]
, met13-2, cyh2
PW40
MAT
a
,
ura3-n, leu2
[Delta]
, met14-1, his6-1, lys9-1, trp1::URA3
Y55-2372
MAT
a
,
ura3-n, leu2
[Delta]
, trp1-b, can1
EJL374-8A
MAT
a
,
ura3-52, ade2-101, lys2, leu2
[Delta]
, his4-12, can1
XQ2
MAT
[alpha],
ura3-n, leu2
[Delta]
, can1
XQD30
MAT
a
,
ura3-52, lys2-801, ade2-101, trp1
[Delta]
1, leu2
[Delta]
1, CRVII
(
RAD2.d.YPH277) URA3 SUP11
MAT
[alpha],
ura3-52, leu2
[Delta]
, ade2-101, lys2, cyh2
XQD20
MAT
[alpha]
ura3-n, leu2
[Delta]
,
can1
MAT
a
,
ura3-n, leu2
[Delta]
, met14-1, his6-1, lys9-1, trp1::URA3
XQD40
MAT
a
,
ura3-52, ade2-101, lys2, leu2
[Delta]
,
his4-12, can1
MAT
[alpha],
ura3-52, leu2
[Delta]
, ade2-101, lys2, cyh2
RESULTS
Cloning of the
PAT1
gene
To identify proteins that interact with topoisomerase II in
S.cerevisiae
, we used the two-hybrid cloning system developed by Fields and Song (
28
) and adapted by Zervos
et al
. (
24
). The DNA binding domain fusion (`bait') comprised residues 1118-1429 of yeast topoisomerase II fused to
E.coli
LexA protein (construct pLexTopDT; ref.
16
), while a library of yeast proteins fused to the B42 transcriptional activator
represented the `prey'.
Strain EGY48 (Table
1
), harbouring the `bait' construct was transformed with the transcriptional
activation domain library, yielding 3 * 10
6
independent colonies. Of these, 10 displayed both leucine prototrophy and a
blue coloration on plates containing X-gal. Five clones represented the
SGS1
gene and have been discussed elsewhere (
16
). Two identical clones, which were distinct from
SGS1
, contained a single long open reading frame identical to a sequence present on
chromosome III that had been identified as part of the yeast genome sequencing
project and designated YCR077c. Based upon the data presented in this paper, we
propose that this gene be renamed
PAT1
(for protein associated with topoisomerase II).
The nucleotide sequence of the full length
PAT1
gene was generated using clones derived both by PCR amplification of yeast
genomic DNA and by screening a yeast genomic DNA library. Our nucleotide
sequence for
PAT1
did not differ from that available in the DDBJ/EMBL/GenBank database (accession
no. S53590). The region of Pat1p expressed by the transcriptional activation
domain construct (designated pActPat1) is between amino acids 84 and 494.
The predicted Pat1 protein sequence is unusually rich in proline and glutamine
residues (overall 6.8 and 7.5%, respectively), and in a 154 amino acid stretch
beginning at residue 90, the Pat1 protein contains 34 proline and 20 glutamine
residues (22 and 13%, respectively). However, the Pat1 amino acid sequence
lacks any readily-identifiable motifs that could give a direct clue to function.
Comparison of the predicted Pat1p amino acid sequence with protein sequences in
the Swissprot database revealed some similarity with several proline and/or
glutamine-rich proteins, including the
Xenopus
oocyte-specific p100 protein (
29
), and a domain within CBP, the CREB-binding protein from human and mouse cells (
30
-
32
) (Fig.
2
). However, the significance of these sequence similarities is as yet unclear.
Functional yeast topoisomerase II interacts with Pat1p via a leucine-rich domain
To assess the specificity of the interaction between Pat1p and topoisomerase II,
and to identify the region of yeast topoisomerase II with which Pat1p
interacts, we tested whether constructs comprising residues 1109-1429, 1109-1163 or 1168-1429 (pLexTopDT, pLexTopD and pLexTopT, respectively; see
ref.
16
) of topoisomerase II could activate transcription in the presence of the
pActPat1 construct. As controls, the non-specific LexA fusion proteins pLexMAX and pHAP1 were tested similarly.
Yeast expressing the pLexTopDT construct (Fig.
1
) and the pLexTopD construct (data not shown) showed leucine prototrophy upon
induction of the
Gal1
promoter in pActPat1, while no significant activation (lack of leucine
prototrophy) was seen in the case of either pLexTopT or the non-specific fusion proteins, pLexMAX and pHAP1 (Fig.
1
). These results indicate both that Pat1p interacts with the highly conserved,
leucine-rich domain of topoisomerase II between residues 1109 and 1163, and that
this interaction is apparently specific.
Deletion of the
PAT1
gene affects cell growth, viability and morphology
To study the function(s) of Pat1p, we generated targeted deletions of the
PAT1
gene in a variety of genetic backgrounds. This was achieved by insertion of the
LEU2
gene after codon 100 of
PAT1
with the concomitant deletion of 1068 bp of the
PAT1
coding region. The presence of termination codons in the
LEU2
gene fragment precludes the possibility that a Leu2-Pat1 fusion protein could be synthesized
in vivo
. The truncated polypeptide that might be produced in the [Delta]
pat1
strains would contain only 100 residues of Pat1p (and would lack the
topoisomerase II-interaction domain) and it is highly unlikely, therefore, that this
truncated Pat1p would be functional.
Although some minor degree of inter-strain variability was observed, all haploid [Delta]
pat1
strains grew much more slowly than their isogenic parental controls. To confirm
that deletion of
PAT1
conferred a slow growth phenotype, a diploid heterozygous for the
PAT1
gene was constructed from the isogenic haploid strains, EJL374-8A and PW50, the latter of which contained a deletion of
PAT1
. Figure
3
shows that tetrads obtained from sporulation of this heterozygous diploid
strain generated two fast growing and two slow growing segregants when
dissected onto YPD medium. The
LEU2
marker used to disrupt
PAT1
in all cases segregated with the slow growth phenotype.
Homozygous
[Delta]
pat1
diploid strains have a reduced ability to generate viable spores
Dissection of >500 tetrads derived from sporulation of the diploid XQD30 strain
carrying a homozygous deletion of
PAT1
revealed that both sporulation frequency and spore viability were decreased
compared with the isogenic wild-type control diploid (Table
2
). The overall spore viability for the [Delta]
pat1
diploid was reduced to 62% (compared with 97% for the control strain), and a
large excess of tetrads was observed that contained two or fewer viable spores
(Table
2
).
The slow growth phenotype of
[Delta]
pat1
strains can be reversed by expression of recombinant Pat1p
To confirm that the slow growth and poor viability of [Delta]
pat1
strains was a direct result of a loss of
PAT1
gene function, we cloned the full length
PAT1
coding region downstream of the
GAL1
promoter in the yeast expression vector pYES2 (to generate pXQ102; see
Materials and Methods) and expressed Pat1p in the PW50 [Delta]
pat1
strain. The presence of pXQ102 harbouring the
PAT1
gene had no influence on the growth rate of PW50 [Delta]
pat1
cells on YPD plates, where the activity of the
GAL1
promoter would be repressed by the presence of glucose in the medium. However,
near wild-type levels of growth were seen on plates containing 2% galactose, which
would lead to an induction of Pat1p expression (data not shown). Consistent
with this result, the viability and rate of growth of [Delta]
pat1
spore colonies (scored as Leu
+
) that retained the pXQ102 (scored as Ura
+
) was enhanced, in comparison with vector-only controls, after dissection of tetrads derived from sporulation of the
XQD40 diploid containing a homozygous deletion of
PAT1
onto plates containing 2% galactose (Fig.
5
).
[Delta]
pat1
strains show an increased incidence of chromosome missegregation during both
mitotic and meiotic cell divisions
Since [Delta]
pat1
strains showed nuclear abnormalities and poor sporulation efficiency, and Pat1p
had been shown to interact with a protein required to effect faithful
chromosome transmission during cell division, we analyzed whether [Delta]
pat1
strains showed an altered fidelity of chromosome segregation at mitosis and/or
meiosis. To quantify the frequency of mitotic missegregation in [Delta]
pat1
strains, the
PAT1
gene was deleted in YPH277, a strain that carries a chromosomal fragment marked
with the
URA3
gene (
18
). Cells lacking this fragment, as a result of either chromosomal loss or
nondisjunction during mitotic division, become resistant to 5-FOA. Table
3
shows that deletion of
PAT1
caused an average 5.8-fold increase in the rate of chromosome missegregation per cell generation
compared with that seen in the wild-type YPH277 strain. CHEF gel analysis was used to confirm that 5-FOA resistance had arisen due to loss of the chromosomal fragment
and not via a gene conversion event involving the chromosomal
ura3-52
allele (data not shown).
Missegregation (rate per cell generation) (* 10
4
)
YPH277
YPH277 [Delta]
pat1
1
0.13
1.23 (9.5
a
)
2
0.14
1.13 (8.1)
3
0.55
1.31 (2.4)
4
1.40
4.18 (3.0)
a
Figures in parentheses represent the fold increase in the rate of mitotic
missegregation in the YPH277 [Delta]
pat1
strain for each experimental determination. Mean rates for YPH277 and YPH277 [Delta]
pat1
in each case are significantly different (
P
< 0.01).
To analyse meiotic chromosome missegregation, a homozygous deletion of
PAT1
was generated in strain XQD20, which is genetically marked for monitoring
missegregation that occurs specifically during meiosis I. This strain is
isogenic with PWD70 used previously for such analyses (
16
), with the exception that the
can1
allele replaced the
cyh2
allele. One of the haploid parents of XQD20 has a
URA3
disruption of the
TRP1
gene near the centromere on chromosome IV, while the other has an intact
TRP1
gene at the same chromosomal location. Both parents carry a
ura3n
mutation elsewhere with genome. Missegregation of chromosome IV in XQD20
(either via chromosome nondisjunction or precocious sister segregation) can be
monitored by measuring the frequency of Ura
+
Trp
+
progeny that have successfully sporulated (assessed by following the recessive
can1
mutation, conferring resistance to canavanine). In contrast, normal Mendelian
segregation would give rise to Ura
+
Trp
-
and Ura
-
Trp
+
progeny in equal numbers, but not to Ura
+
Trp
+
progeny. Table
4
shows that deletion of
PAT1
caused an ~4-fold increase in the frequency of meiosis I missegregation of
chromosome IV. The frequency of missegregation seen in the XQD20 wild-type control was very similar to that reported previously for the near
isogenic PWD70 strain (
16
).
a
Mean of four independent experiments.
b
Mean of five independent experiments.
Mean frequencies for XQD20 and XQD20 [Delta]
pat1
are significantly different (
P
< 0.01).
DISCUSSION
We have shown that the
PAT1
gene of
S.cerevisiae
encodes a topoisomerase II-associated protein that is required for the maintenance of genome
integrity. Yeast lacking functional Pat1p exhibit a number of phenotypic
abnormalities, including a reduced growth rate, reduced viability and reduced
ability to undergo productive nuclear division. Based upon the observations
that Pat1p interacts with topoisomerase II, and that
pat1
mutants display a reduced fidelity of chromosome segregation, we suggest that
Pat1p is a component of the cellular machinery that ensures error-free chromosome segregation occurs during cell division. However, Pat1p is
not an essential gene and it would appear, therefore, that Pat1p either
performs a non-essential role during chromosome segregation, or that other proteins can
substitute in the absence of Pat1p.
Since the failed segregation of essential chromosomes cannot be studied
directly, assays of mitotic missegregation generally utilise non-essential chromosomal fragments (
18
). Similarly, the assay of meiosis I missegregation described here relies on an
ability to monitor viable progeny from tetrads that may contain up to three non-viable spores. Moreover, these assays monitor the segregation of only one
of the
S.cerevisiae
chromosomes. Thus, the frequency of missegregation revealed by the available
assays may not reflect the true extent to which the segregation process is
aberrant in a particular mutant, since it is clearly impossible to `score' non-viable progeny. In those cases where chromosome segregation still occurs
in a mutant strain, albeit with reduced efficiency, but is associated with an
increase in the rate of cell death, the available missegregation assays may
give a substantial underestimate of the true extent to which the segregation
process is defective. Based upon the high frequency of apparently aploid cells
seen in cultures of strains lacking functional
PAT1
, and the high percentage of dissected [Delta]
pat1
tetrads that contained two or fewer viable spores, we suggest that this is
likely to be the case for [Delta]
pat1
strains.
The predicted Pat1 protein sequence shows some similarity with the sequences of
proline- and glutamine-rich proteins from other eukaryotic species. The
Xenopus
oocyte-specific p100 protein has not been characterized in detail, but has been
shown to be a single-stranded DNA binding protein (
29
). In contrast, the CBP factor from human and mouse cells has been well
characterised biochemically as a CREB binding protein, and is known to mediate
cAMP-dependent transcriptional activation of those genes that contain the cAMP
response element in their promoters (
30
,
31
). The region of greatest homology between Pat1p and CBP is within a glutamine-rich domain of CBP that could be involved in directing protein-protein interactions (
30
). Together with data suggesting that glutamine-rich domains in other proteins are important for directing interactions
with protein partners, it is tempting to speculate that Pat1p acts (perhaps not
exclusively) as an adapter protein. Further two-hybrid screens using Pat1p as the `bait', might serve as a means to
identify other factors important for chromosome segregation and/or maintenance
of the stability of the rDNA locus. It may also be of significance that one of
the two human topoisomerase II enzymes, the 170 kDa [alpha] isoform, has been shown to bind to CREB (
33
). Whether topoisomerase II, CREB, CBP and the homologue of Pat1p form a
multienzyme complex in human cells will require further analyses.
The
PAT1
gene has previously been assigned the name YCR077c as a result of its
identification as an open reading frame in the fully sequenced chromosome III
of
S.cerevisiae
. As part of the yeast genome analysis, Rodriguez-Cousino
et al
. (
34
) also created a strain in which YCR077c was disrupted by insertion of the
URA3
gene. They reported that little, if any, gross phenotypic changes were evident
in the disruptant, apart from a small effect on the rate of cell growth.
Indeed, they indicated that a homozygous mutant diploid strain sporulated
normally. We have generated targeted deletions of
PAT1
in six different genetic backgrounds (this and unpublished studies) and in each
case a moderate to severe slow growth phenotype was observed. The difference in
growth rate of [Delta]
pat1
spore colonies compared with that of
PAT1
spore colonies from the same tetrads reported here, is typical of the effects
that we have observed consistently. The reason for the apparent discrepancy
between our data and those of Rodriguez-Cousino
et al
. (
34
) is not apparent at this stage, but might possibly relate to the fact that
Rodriguez-Cousino
et al.
(
34
) generated a very small deletion within the
PAT1
gene, which might have caused retention of some residual Pat1p protein
function. Alternatively, it could be that a slow growth suppressor mutation had
developed in the
pat1
strain studied by Rodriguez-Cousino
et al
. (
34
), which masked the strong phenotypic effects of the
pat1
mutation. Indeed, it is clear that such slow growth suppressors arise at a high
frequency, since we have identifed two independent suppressors which restore
near wild-type growth to the LS375 [Delta]
pat1
strain (unpublished data).
In summary, we have identified the Pat1 protein as a cellular partner for
topoisomerase II and we have shown that strains lacking functional Pat1p
display phenotypic similarities to conditional topoisomerase II-deficient mutants. Although several genes have been identified in
eukaryotes that affect chromosome segregation and general genome stability,
very few of them are known to be involved in the pathway that deals with the
resolution of topological problems that occur during different aspects of DNA
metabolism. Further analysis of [Delta]
pat1
strains should shed more light on the precise reasons why the functions of
topoisomerase II and Pat1p are apparently coordinated in DNA metabolism.
ACKNOWLEDGEMENTS
We thank Drs G. Fink, P. Hieter, R. Brent and R. Finley for yeast strains and
plasmids, and Dr J. Diffley for the yeast genomic DNA library. We also thank Dr
C. Norbury for critical reading of the manuscript and Mrs E. Clemson for typing
the manuscript. X.W., P.M.W. and I.D.H. are supported by the Imperial Cancer
Research Fund. E.J.L. and R.H.B. are supported by the Wellcome Trust.
REFERENCES
1 German,J. (1974) In Chromosomes and Cancer. John Wiley and Sons, NY.
