Download Chromosomes in Saccharomyces cerevisiae

MOLECULAR AND CELLULAR BIOLOGY, Sept. 1986, p. 3166-3172
0270-7306/86/093166-07$02.00/0
Copyright © 1986, American Society for Microbiology
Vol. 6, No. 9
Construction and Behavior of Circularly Permuted and Telocentric
Chromosomes in Saccharomyces cerevisiae
ANDREW W. MURRAY AND JACK W. SZOSTAK*
Department of Molecular Biology, Massachusetts General Hospital, Boston, Massachusetts 02114
Received 6 March 1986/Accepted 20 May 1986
We are attempting to define the structural requirements
for accurate chromosome segregation in mitosis and meiosis
(9a, 10). Examination of the Saccharomyces cerevisiae
genetic map (9) reveals that most yeast chromosomes are
metacentric, that is, their centromeres lie roughly halfway
between their telomeres. The known correlation between the
physical and genetic maps (19) suggests that all the telomeres
lie at least 100 kilobases (kb) from a centromere. These
natural chromosomes are lost at a frequency of about 10'
per cell division (5). In contrast, a 55-kb metacentric artificial
chromosome, whose telomeres are about 25 kb from the
centromeres, is lost at a frequency of about 10-2 per cell
division, and 10- to 15-kb metacentric artificial chromosomes, whose telomeres are about 5 kb from the centromere,
are lost at a frequency of 10-1 per cell division (2, 10). We
wished to determine whether the difference in mitotic stability between natural and artificial chromosomes was related
to the difference in centromere-telomere separation. To this
end, we developed techniques that allow us to produce a
wide variety of novel derivatives of yeast chromosome III.
We produced both metacentric and telocentric chromosomes, all of which carry telomeres derived from those of
the Tetrahymena ribosomal DNA plasmid (23, 24) rather
than those of natural yeast chromosomes.
We demonstrate that telocentric chromosomes are slightly
less mitotically stable than metacentric natural chromosomes, but are much more stable than metacentric artificial
chromosomes which have similar separations between their
centromeres and the closest telomere. Telocentric chromosomes also segregate normally in meiosis. Finally, we show
that the telomere-associated X and Y' sequences (1) are
completely dispensable for normal chromosome segregation.
Mo.), and media were prepared by the method of Sherman et
al. (17).
Strains. The yeast strains used in this work are listed in
Table 1. Escherichia coli JA300 (25) and BAl (leuB6
trpCJ116 hisB thyA thi Strr Tcr hsdR hsdM; A. W. Murray,
unpublished data) were used for plasmid constructions.
Bacterial plasmids used in this study were constructed by
standard recombinant DNA techniques. Standard techniques were used for the construction of diploids, sporulation, and tetrad analysis (17).
Yeast transformation and Southern blots. Yeast transformations and Southern blots (18) were performed as described by Orr-Weaver et al. (13).
Mitotic stability measurements. The mitotic stability of
derivatives of chromosome III was measured by a quantitative mating assay as described previously (Murray et al., in
press). Cells were grown in leucine-free medium (to prevent
early loss of the marked chromosome III derivative, which
carries the only functional LEU2 gene) and then mated to a
haploid tester strain before plating on medium on which only
the products of the mating could grow. Because cells which
have lost the marked chromosome III derivative can divide
a small number of times on leucine-free medium, the results
are expressed as the frequency of chromosome loss (chromosome loss events per cell) rather than the rate of chromosome loss (loss events per cell division). In all cases true
chromosome loss events were distinguished from a mitotic
recombination or gene conversion event by monitoring the
retention of a genetic marker on the opposite side of the
centromere from MA T. In each experiment at least 100
products of each mating were tested for chromosome loss in
this way. The results shown in Table 2 are the average of at
least three independent experiments.
Construction of YLp16. The linear plasmid YLpl6 was
constructed from the bacterial plasmid A9Op2 and the terminal fragment of the Tetrahymena ribosomal DNA plasmid
(hereafter called the Tr end). A9Op2 is a progenitor of the
plasmid A164p2 (Fig. lb), but lacks the inverted repeat of the
Tr ends flanking the yeast HIS3 gene present in A164p2.
A9Op2 was cut with SalI and ligated to a gel-purified XhoIHhaI fragment containing the Tr end and derived from
pSZ221 (23). This ligation mix was used to transform the leu2
MATERIALS AND METHODS
Enzymes, chemicals, and media. Restriction enzymes and
T4 DNA ligase were obtained from New England BioLabs,
Inc. (Beverly, Mass.) and used as recommended by the
manufacturer. Medium components were from Difco Laboratories (Detroit, Mich.) and Sigma Chemical Co. (St. Louis,
*
Corresponding author.
3166
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
We developed techniques that allow us to construct novel variants of Saccharomyces cerevisiae chromosomes.
These modified chromosomes have precisely determined structures. A metacentric derivative of chromosome
III which lacks the telomere-associated X and Y' elements, which are found at the telomeres of most yeast
chromosomes, behaves normally in both mitosis and meiosis. We made a circularly permuted telocentric
version of yeast chromosome III whose closest telomere was 33 kilobases from the centromere. This telocentric
chromosome was lost at a frequency of 1.6 x 10-5 per cell compared with a frequency of 4.0 x 10-6 for the
natural metacentric version of chromosome III. An extremely telocentric chromosome whose closest telomere
was only 3.5 kilobases from the centromere was lost at a frequency of 6.0 x 10-5. The mitotic stability of
telocentric chromosomes shows that the very high frequency of nondisjunction observed for short linear
artificial chromosomes is not due to inadequate centromere-telomere separation.
3167
MODIFIED YEAST CHROMOSOMES
VOL. 6, 1986
TABLE 1. Yeast strains
J. Haber
This work
This work
This work
This work
This work
This work
This work
This work
This work
adel Ieu2 ura3 can) cir+ (circular chromosome III)
adel Ieu2 ura3 can) cir° (circular chromosome III)
a adel Ieu2 ura3 can) cir° (chromosome III-Tr: URA3 LEU2)
a adel leu2 ura3 canl cir+ (chromosome III-TcA: URA3 LEU2)
a leu2-3,112 his3-11,15 trpl ura3 canl cir+
a trpl can) cir+
a leu2-3,112 his3-11,15 trpl ura3 can) cir+
a adel leu2-3,112 his3-11,15 trp) ura3 can) cir+ (chromosome III-Tr: URA3 LEU2)
a adel leu2-3,112 trp) ura3 can] cir+ (chromosome III-Tr: URA3 LEU2)
a leu2-3,112 his3-11,15 trpl ura3 canl cir+
a leu2-3,112 his3-11,15 trpl ura3 canl cir+ (chromosome III-TcA: URA3 LEU2)
a adel leu2-3,112 his3-11,15 trpl ura3 canl cir+ (chromosome lII-Tr: URA3 LEU2)
DA164.9A x DA1145-I (normal III + III-TcA)
DA188.1B x DA188.1C (III-TcA + III-TcA)
DA188.1A x DA188.26B (normal III + III-TcA)
DA188.39D x DA168.17A (normal III + normal III)
DA188.39D x DA235.24D (normal III + III-Tr)
Derived from DA192 by transformation (normal III + III-TcB)
Derived from DA192 by transformation (normal III + III-TcC)
a leu2-3,112 his3-11,15 arg4 ura3 cir+ (quantitative mating tester)
a leu2-3,112 his3-11,15 arg4 ura3 cir+ (quantitative mating tester)
a leu2-3,112 trpl can) cir+ (quantitative mating tester)
a leu2-3,112 trpl can) cir+ (quantitative mating tester)
a
a
ura3 strain A288, and Leu+ Ura+ transformants were selected. These were restriction mapped on Southern blots to
verify that they contained a plasmid of the correct structure.
The transformant TA1145 was used to isolate TcIII-A as
described below. The version of YLpl6 in TA1145 contains
a tandem head-to-tail duplication of the Tr end at its left end
(Fig. lc) which arose because the XhoI-HhaI Tr fragment
was contaminated with an HhaI Tr fragment owing to
incomplete XhoI digestion. This contaminating HhaI fragment then ligated onto the HhaI end of an XhoI-HhaI
fragment which had already been ligated onto the Sall site of
A9Op2.
RESULTS
Directed chromosome breakage. One possible explanation
for the failure of artificial chromosomes to segregate as
faithfully as their natural counterparts is that they carry
telomeres derived from those of the Tetrahymena ribosomal
DNA plasmid (Tr ends [24]) rather than those found on
natural yeast chromosomes. Although the Tr ends in yeasts
are modified by the addition of several hundred base pairs of
the yeast telomeric repeat sequence C1_3A (16), they lack the
telomere-associated sequences X and Y', which lie adjacent
to the telomeric repeat sequences on many yeast chromosomes (1). To determine whether the X and Y' elements play
any role in normal chromosome behavior, we sought to
replace the ends of chromosome III with Tr ends.
We took advantage of the existence of a circular derivative
of chromosome III (Fig. la) created by recombination between the two silent mating-type cassettes, HML and HMR
(4, 7). HML and HMR lie close to the left end right ends of
chromosome III, respectively, and contain homologous sequences in direct orientation. Recombination between the
silent cassettes creates a circular chromosome and a short
linear molecule, which carries the chromosomal telomeres.
This linear fragment lacks a centromere and is rapidly lost,
but since none of the genetic information which it carries is
essential, the strains carrying the circular chromosome III
are viable as haploids. We relinearized chromosome III by
This work
This work
This work
This work
This work
This work
This work
This work
This work
M. Whiteway
M. Whiteway
This work
This work
directing a plasmid carrying an inverted repeat of the Tr ends
to integrate at the HML-HMR fusion generated by the
original recombination event (Fig. lb). After the plasmid had
integrated, the inverted repeat of the Tr ends resolved to
give a pair of telomeres (23), generating a new version of
chromosome III which we call chromosome III-Tr. This
novel linear chromosome lacks the telomere-associated sequences which normally lie centromere distal to HML and
HMR.
The plasmid A164p2 (Fig. lb) carries the yeast HMRa
locus, the selectable markers LEU2 and URA3, and an
inverted repeat of the Tr ends separated by the yeast HIS3
gene. A164p2 DNA was cut with BamHI, which separates
the HIS3 gene from the rest of the molecule, and then treated
with DNA ligase to generate circular molecules which carried an uninterrupted inverted repeat of the Tr ends (we used
this strategy because uninterrupted inverted repeats cannot
be maintained in E. coli [8]). The circular DNA molecules
were recut with XbaI, which cleaves the plasmid only in
HMRa and therefore directs the plasmid to integrate at the
HML-HMR fusion (12). The XbaI-cut DNA was transformed into the leu2 ura3 yeast strain TA1540 which carries
the circular version of chromosome III, and Ura+ Leu+
transformants were selected.
TABLE 2. Loss frequencies of chromosome III derivativeSa
TelomereLoss frequency
centromere
Length
( 0s
Strain
events/cell
Derivative
(kb)
separation
(kb)b
(x lo-5)
Normal III
III-Tr
I11-TcA
III-TcB
III-TcC
DA226
DA247
DA192
TA1545
TA1425
530C
500
500
480
470
250
250
33
12
3.5
0.4
0.3
1.6
2.9
6.0
±
±
±
±
±
0.2
0.2
1.3
1.6
3.4
aThe frequency of chromosome loss was measured by a quantitative
mating assay as described in Materials and Methods.
b
Separation between the centromere and the closest telomere. Figures for
the metacentric chromosomes are estimates.
cFrom Schwartz and Cantor (15).
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
A288
TA1540
TA1593, -94
TA1145-I
DA164.9A
DA168.17A
DA188.1A
DA188.1B
DA188.1C
DA188.39D
DA188.26B
DA235.24D
DA188
DA190
DA192
DA226
DA247
TA1545
TA1424, -25
DM64.2A
DM64.5B
DA168.9B
DA168.9D
Source
Genotype
Strain
3168
MURRAY AND SZOSTAK
b)
Qi)
/11-11
.,Ot- ..
.. I
",pi
-
A -(,4a2
r.
lij!l
~~~~-,
*
9
0
4
:.,
.-Xlp
n
F4o
,X
<
,:k
*
C)
bRT,
';?
.
4--- - .....V._ p;
-
r.p
%
....
.;
....
.
d)
-B.;ir-
-xRAt3
3
-M. R
-
c
P.
r,
C
ci
J
go'.e
X
l
T
N
e
T.r
UJA3
QfAR/NML.
rP,P2P E
--&
Pv!zu PFvulXbl
ct B
X
Bam
ml, 2-i41l4 2341L 234I1 234lM
¶33
Bg4B~ F3rpF.
E
.2
4MR
"
qg
xo
ir
_9 --go
_
-.4
-~~
4
eo-/4",
.,
6 6--S
-
&_
,etS
i-4~~~~~4-
LEJ2
Tr
.
.
....,3g
-_
,3~~
...
FIG. 1. Circularly permuted version of chromosome III. (a) Construction of the linear chromosome III-Tr. From top to bottom the figure
shows: the normal version of chromosome III; the circular derivative of chromosome III (4); chromosome III-Tr which results from the
resolution of atn inverted repeat of the Tr ends into a pair of telomeres. Symbols: , yeast chromosomal telomeres and telomere-associated
sequences; 4. Tr ends (Tetrahymena telomeres as modified in yeast cells). (b) Restriction enzyme maps of A164p2, the HML-HMR fusion,
and the chromosome ends produced by the integration and resolution of A164p2 (after removal of the HIS3 gene by BamHI digestion and
religation and subsequent digestion with XbaI to target integration) at the HML-HMR fusion. The circular map of A164p2 is not to scale.
Symbols: E', LEU2; EJ, URA3; L=I, HMR sequences; LF, HML sequences; _, Tr ends. Restriction enzyme sites: B, BamHI, Bg,
BglII; E, EcoRI; H, Hindlll; P1, PvuI; P2, Pvull; Xb, XbaI. (c) Construction of the telocentric chromosome III-TcA. The integration event
of the linear plasmid YLp16 at the leu2 gene of the circular chromosome III of strain A288 (kindly provided by J. Haber) is shown. See text
for further details. The restriction maps of YLp16, the region around the chromosomal Ieu2 gene, and the products of integration are shown.
YLp16 contains a BamHI and an XhoI site which are absent from A164p2. Symbols are as in panel b. (d) Restriction enzyme mapping of the
chromosome ends in chromosomes 1lI-Tr and III-TcA. Lanes: M, molecular weight markers (sizes [in kilobases] indicated in the left margin);
1, TA1540 (circle lIl); 2, TA1145-I (III-TcA); 3, TA1593 (III-Tr); 4, TA1594 (lII-Tr). DNA was digested with the indicated restriction
enzymes, run on a 0.5% agarose gel, Southern blotted, and probed with pSZ57 (12) which contains pBR322 plus the LEU2 gene. Fragments
are marked as deriving from the telomere marked by the URA3 gene (A), the telomere marked by the LEU2 gene (O), or the endogenous leu2
gene (0). Note that in chromosome III-TcA the bands derived from the endogenous Ieu2 gene are absent, showing that integration of the Tr
ends has occurred at this locus.
The restriction map of the telomeres of chromosome
III-Tr can be predicted from the restriction map of A164p2
(Fig. lb). Figure ld shows the autoradiogram of a Southern
blot which verifies that transformants TA1593 and TA1594
carry chromosome III-Tr. The LEU2 and URA3 genes are
separated from each other by the resolution of the inverted
repeat of the Tr ends which lies between them. Thus, the
telomeres of chromosome III-Tr are genetically as well as
physically marked: the left one by URA3 and the right by
LEU2.
We used a quantitative mating assay (3) to investigate the
mitotic stability of chromosome III-Tr. We made diploid
strains by mating a MATot haploid carrying chromosome
III-Tr to a MATa haploid which carried the normal version of
chromosome III. Such diploid, MATa/MATa cells are incapable of mating with haploids. Loss of the MATa allele
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
s-
MOL. CELL. BIOL.
VOL. 6, 1986
MODIFIED YEAST CHROMOSOMES
3169
TABLE 3. Behavior of telocentric chromosomes in meiosiSa
Cross
DA235
Chromosome III
DA190
Linkage of LEU2
Marker segregation
and URA3
Marker
4
3
2
1
0
80
14
2
0
0
5
22
14
8
0
LEU2
URA3
0
MAT
MAT
4+:0-
3+:1-
2+:2-
1+:3-
0+:4-
0
00
0
0
0
1
00
0
0
0
0
77
1
00
0
0
0
0
0
PD
NPD
TT
00
9
11
55
0
0
5
0
0
III + III-Tr
+
DA236
DA188
No. of tetrads with the following
no. of viable spores:
111 + III-Tr J
III + III-TcA
III-TcA + III-TcA
LEU2
URA3
MAT
33
3
2
1
0
79
79
5
5
5
32
0
0
a Crosses were performed, sporulated, and subjected to tetrad analysis as described by Sherman et al. (17). In both cross DA235 and cross DA190 one tetrad
with four viable spores was not scored for marker segregation.
b Abbreviations: PD, parental ditype; NPD, nonparental ditype; Tr, tetratype.
quently than a natural chromosome (10). This discrepancy
could reflect the shorter overall size of the artificial chromosome, the smaller separation between its centromere and
telomeres, or the presence on natural chromosomes of
previously unidentified specialized sequences that are required for accurate chromosome segregation. We examined
the role of telomere-centromere separation by analyzing the
behavior of telocentric derivatives of chromosome III.
We made a full-length, telocentric chromosome by directing a linear plasmid to integrate close to the centromere of
the circular version of chromosome III. The construction of
the acentric linear yeast plasmid YLp16 is described in
Materials and Methods. YLp16 was used to transform the
leu2 ura3 strain A288, which carries the circular version of
chromosome III. Because YLp16 carries ARS elements
associated with the HMR locus and the Tr ends, it can
replicate extrachromosomally, but is mitotically unstable.
Integration of YLp16 into the circular version of chromosome III relinearizes the chromosome and makes the LEU2
and URA3 markers, originally carried by the plasmid, mitotically stable. Therefore, continuous growth in leucine-free
medium enriches for integrants. We determined that the
integrant TA1145-I arose from recombination between the
LEU2 sequences on YLp16 and the nonfunctional leu2 gene
on chromosome III by showing that the restriction map of
the chromosome ends agreed with the map predicted from
the known restriction maps of YLp16 and the leu2 region
(Fig. lc and d). Because the leu2 gene is only 21 kb to the left
of CEN3, integration of YLp16 at leu2 creates a full-length,
circularly permuted, telocentric chromosome, chromosome
III-TcA, whose left arm is only 34 kb long. The sequences
from HML to LEU2 which are normally found on the left
arm of chromosome III have been transferred by this series
of manipulations to the right arm of chromosome III-TcA.
Genetic tests indicated that the integration event had occurred such that the ends of chromosome III-TcA are
uniquely marked: the left end carries URA3 and a
nonfunctional leu2 gene, while the right end carries the
wild-type LEU2 gene (data not shown).
The loss frequency of the telocentric chromosome was
measured by performing quantitative mating assays on diploids that contained one normal and one telocentric version
of chromosome III (Table 2). The loss frequency of chromosome III-TcA is 1.6 x 10-5, about four times that of the
normal metacentric version of chromosome III. The separation between the centromere and the nearest telomere in
chromosome III-TcA (34 kb) is similar to that in the 55-kb
artificial chromosome YLp22 (23 kb; loss frequency, 1.5 x
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
carried on chromosome III-Tr generates an a-mating cell,
which can be detected by its ability to mate to an a-mating
tester strain. Because chromosome III-Tr carries the only
functional alleles of the LEU2 and URA3 genes on its ends,
genuine chromosome loss events can be distinguished from
other events which produce cells which lack MATa information. Only in loss events will both the LEU2 and URA3
genes also be lost from the diploid. We compared the loss
frequency of chromosome III-Tr with that of the natural
version of chromosome III (Table 2). The loss frequency of
chromosome III-Tr is nearly identical to that of the normal
version of chromosome III, demonstrating that the absence
of the telomere-associated X and Y' sequences does not
destabilize a yeast chromosome in mitosis.
We also examined the meiotic behavior of chromosome
III-Tr by crossing the haploid transformants TA1593 and
TA1594 to strain DA164.9A which carries the natural version of chromosome III. The resulting diploids, DA235 and
DA236, were sporulated and subjected to tetrad analysis to
monitor the segregation of chromosome III. Because the two
diploids behaved identically, the data from the two crosses
were combined (Table 3). Several lines of evidence suggest
that chromosome III segregation in these crosses is normal.
First, 83% of the tetrads had four viable spores, a value
similar to that of control crosses with two normal versions of
chromosome III (data not shown). Second, among the tetrads with only two or three viable spores, all the viable
spores were capable of mating. If spore death was due to
nondisjunction of chromosome III, some of the viable spores
would contain two copies of chromosome III. A large
fraction of such disomic spores would carry both MATa and
MATa information and be incapable of mating. Third, we
could monitor the segregation of three genetic markers on
chromosome III: MAT, and the terminal markers of chromosome III-Tr, LEU2 and URA3. Both MAT and URA3
showed classical Mendelian 2:2 segregation in 79 of 79
tetrads, while LEU2 showed 2:2 segregation in 77 tetrads
and one case each of 3 LEU2:1 leu2 and 1 LEU2:3 leu2
segregation. Finally, we could monitor recombination between the telomeres of chromosome III-Tr. As expected
from the genetic length of chromosome III (150
centimorgans [9]), terminal LEU2 and URA3 markers were
not linked (the ratio of parental ditype:nonparental
ditype:tetratype segregation was 9:11:55, which is not significantly different from the 1:1:4 ratio expected for unlinked
genes).
Mitotic stability of telocentric chromosomes. A 55-kb linear
artificial chromosome is lost a thousand times more fre-
3170
MURRAY AND SZOSTAK
a))"'.
MOL. CELL. BIOL.
.-
_
J
i
.~-.f=
A233pl
e.
.xI'.-
o 7\
C)
-ic
2
K-
1'.
e
Telr:
r
r
(b) Plasmid A233pl, which was used to construct chromosome
III-TcC. Symbols in addition to those used in panel a:
pBR322
sequences; Apr, beta-lactamase gene. Restriction enzyme sites: B,
BamHI; P2, PvuII; E, EcoRI. The plasmid is shown cut with BamHI
which exposes the homolgy to the Tr end and CEN3. (c) Southern
blot of BamHI digests of DNA from six diploid strains (TA1424 to
TA1429) containing chromosome III-TcC and the strain carrying
chromosome II-TcA (DA192) from which they were derived by
one-step transplacement. The probe was pSZ64 (12), which contains
pBR322 and the 1.2-kb BamHI-XhoI HIS3 fragment. In all six
transformants the 12-kb band which carries the left telomere of
chromosome IIl-TcA (and hybridizes to pBR322) has been replaced
by a new band at 3.5 to 4.0 kb which shows the length heterogeneity
typical of a terminal restriction fragment. The sizes of molecular
weight standards in lanes M (in kilobases) are indicated at the right
,
margin.
10-2 [10]). We conclude that the location of the centromere
along the length of the chromosome is not a major determinant of mitotic chromosome behavior.
To reduce the distance between the centromere and
telomere still further, we constructed deletions which removed the DNA between LEU2 and CEN3 to yield the
extremely telocentric chromosome, III-TcC. Figure 2a
shows the strategy for making chromosome III-TcC by the
one-step transplacement technique of Rothstein (14). The
diploid strain DA192 contains one copy of chromosome III
and one copy of chromosome III-TcA and is homozygous for
his3 and trpl. To create chromosome III-TcC, this strain
was transformed with a 3.0-kb BamHI fragment derived
from plasmid A233pl (Fig. 2b), and His' transformants were
selected. One end of the fragment is homologous to sequences immediately to the right of CEN3, while the other is
homolgous to the Tetrahymena sequences at the telomeres
of chromosome III-TcA. A double crossover (or gene conversion) between these regions of homology deletes all the
sequences between CEN3 and the telomere of chromosome
III-TcA including the URA3 gene. As a result, the distance
from the left telomere to the functional centromere is only
3.2 kb plus the length of C1_3A repeats added to the Tr end
(Fig. 2a). We screened 16 His' transformants and found 6
that had become Ura-. DNA from these was prepared,
digested with BamHI, run on an agarose gel, Southern
blotted, and probed with the plasmid pSZ64 (12) which
contains pBR322 and the yeast HIS3 gene. Figure 2c shows
that the 12-kb band derived from the left end of chromosome
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
FIG. 2. Construction of an extremely telocentric chromosome.
(a) Construction of the extremely telocentric chromosome III-TcC
by one-step transplacement. The interaction of a linear restriction
fragment derived from the plasmid A233pl with the left arm of
chromosome III-TcA is shown. A double crossover, utilizing the
homology provided by the ends of the restriction fragment, replaces
the sequences that lie between the Tr end and CEN3 of chromosome
III-TcA with the HIS3 gene. Symbols: _., Tr ends;
HIS3
gene; CL1, CEN3; EA, URA3 gene. For further details, see the text.
III-TcA disappeared and was replaced by a new band which
ranged in size from 3.5 to 4.0 kb in different transformants.
This length heterogeneity is typical of the terminal fragments
of linear molecules (24). The length of the C1_3A repeats at
these chromosomal telomeres ranged from 0.3 to 0.7 kb,
which is somewhat larger than the values reported for small
linear plasmids (16, 24). For two transformants, TA1424 and
TA1425, the structure of the left end of the chromosome was
confirmed by restriction mapping on Southern blots with
HIS3 and CEN3 sequences as probes (data not shown).
The loss frequency of chromosome III-TcC was measured
at 6.0 x 10-5 (Table 2), about 15 times higher than that of the
normal version of chromosome III. A less telocentric chromosome, III-TcB (also derived from chromosome III-TcA
by one-step transplacement), whose left end is 12 kb from
the centromere, had a loss frequency of 2.9 x 1i-5. We
conclude that as chromosomes are made progressively more
telocentric, they segregate less faithfully in mitosis, although
this effect is much less dramatic than that due to changes in
overall chromosome length (6, 10, 21; Murray et al., in
press). Zakian et al. (26) and Surosky and Tye (22) have also
shown that telocentric yeast chromosomes are only slightly
less stable than natural metacentric chromosomes. The
mitotic stability of telocentric chromosomes shows that a
nearby telomere does not prevent the centromere from
interacting normally with the mitotic spindle and implies that
the instability of short artificial chromosomes is not due to
their inability to attach to the spindle. We cannot rule out the
possibility that one long chromosome arm is needed for
attachment to the spindle at every division, although experiments with dicentric linear plasmids demonstrate that molecules with two very short arms attach in at least some
fraction of mitoses (Murray et al., in press).
Meiotic behavior of telocentric chromosomes. Having established that telocentric chromosomes segregated normally
in mitosis, we wished to examine their behavior in meiosis.
As a first step, we crossed the haploid TA1145-I, which
carries the telocentric chromosome III-TcA, to the haploid
DA164.9A, which carries the natural version of chromosome
III, to yield the diploid DA188. This diploid was sporulated
and subjected to tetrad analysis (Table 3). Unlike the
isogenic cross involving the metacentric chromosome III-Tr,
the spore viability was poor: only 5 of 49 tetrads had four
viable spores. We believe that this inviability is due to the
generation of abnormal chromosomes, including duplications, deletions, dicentrics, and acentrics, by recombination
between circularly permuted chromosomes. This interpretation is strongly supported by the observation that the terminal markers of chromosome III-TcA, LEU2 and URA3, are
tightly linked in the tetrads with four viable spores, but not
in tetrads with less than four viable spores (data not shown).
The simplest interpretation of these data is that four viable
spores can only be recovered from tetrads in which there is
an even number of exchanges between any two chromatids
(which leaves the terminal markers linked to each other). A
similar situation is encountered in a diploid which contains
one normal and one circular version of chromosome III (4).
When this diploid was put through meiosis, only 46 of 216
tetrads had four viable spores. Fifteen of the four-spored
tetrads had double crossovers on chromosome III, and all of
these involved only two of the four chromatids.
We created a diploid with two copies of chromosome
III-TcA by crossing the two Leu+ Ura+ spores of one of the
tetrads with four viable spores from DA188. The resulting
diploid, DA190, showed a high level of spore viability (85%
of the tetrads had four viable spores), and MAT segregated
VOL. 6, 1986
2:2 in all of the tetrads with four viable spores (Table 2).
Thus, telocentric chromosomes segregate normally in meiosis as well as mitosis. Zakian et al. (26) have shown that
when chromosome IV is fragmented into two telocentric
chromosome halves, the two chromosomes halves both
segregate away from an intact full-length chromosome IV in
meiosis
I.
3171
the fact that circular molecules are topologically different
from linear ones, rather than the fact that linear chromosomes have telomeres while circular ones do not. Specifically, we propose that the catenation of sister chromatids
directs their accurate segregation from each other and that
this catenation can be maintained on topologically closed
circular molecules but is resolved by rotation of linear
molecules about each other (11). As the size of linear
chromosomes (either artificial chromosomes [6, 10; Murray
et al., in press] or fragments of natural chromosomes [21;
Murray et al., in press]) increases, their copy number
declines, and they begin to show the ordered segregation
characteristic of natural chromosomes. We propose that this
size dependence of chromosome segregation reflects the
need to form topologically closed domains which can retain
the catenation between sister chromatids that is generated
during DNA replication (11, 20). Other lines of evidence that
support the catenation-dependent model of chromosome
segregation and its application to meiotic chromosome segregation are discussed elsewhere (11).
ACKNOWLEDGMENTS
We are grateful to G. Tschumper, D. Dawson, and M. Whiteway
for supplying strains and to V. Zakian, R. Surosky, C. Newlon, and
B. Tye for communicating results before publication. We especially
thank D. Dawson and M. Whiteway for many stimulating discussions and helpful comments on the manuscript. We thank members
of the Szostak laboratory and Harvard Gardens for moral support.
The initial stages of this work were performed at the Dana-Farber
Cancer Institute and were supported by Public Health Service grant
GM-32039 from the National Institutes of Health. Work done at
Massachusetts General Hospital was supported by a grant from
Hoechst AG.
LITERATURE CITED
1. Chan, C. S. M., and B.-K. Tye. 1983. Organization of DNA
sequences and replication origins at yeast telomeres. Cell
33:563-573.
2. Dani, G. M., and V. A. Zakian. 1983. Mitotic and meiotic
stability of linear plasmids in yeast. Proc. NatI. Acad. Sci. USA
80:3406-3410.
3. Dutcher, S. K., and L. H. Hartwell. 1982. The role of S.
cerevisiae cell division cycle genes in nuclear fusion. Genetics
100:175-184.
4. Haber, J. E., P. C. Thorburn, and D. Rogers. 1984. Meiotic and
mitotic behavior of dicentric chromosomes in Saccharomyces
cerevisiae. Genetics 106:185-205.
5. Hartwell, L. H., S. K. Dutcher, J. S. Wood, and B. Garvik. 1982.
The fidelity of mitotic chromosome reproduction in S. cereivisiae. Recent Adv. Yeast Mol. Biol. 1:28-38.
6. Heiter, P., C. Mann, M. Snyder, and R. W. Davis. 1985. Mitotic
stability of yeast chromosomes: a colony assay that measures
non-disjunction and chromosome loss. Cell 40:381-392.
7. Klar, A. J. J., J. N. Strathern, J. B. Hicks, and D. Prudente.
1983. Efficient production of a ring derivative of chromosome
III by the mating-type switching mechanism in Saccharomyces
cerevisiae. Mol. Cell. Biol. 3:803-810.
8. Lilley, D. M. J. 1981. In vivo consequences of plasmid topology.
Nature (London) 292:380-382.
9. Mortimer, R. K., and D. Schild. 1980. Genetic map of Saccharomyces cerevisiae. Microbiol. Rev. 44:519-571.
9a.Murray, A. W., N. P. Schultes, and J. W. Szostak. 1986.
Chromosome length controls mitotic chromosome segregation
in yeast. Cell 45:529-536.
10. Murray, A. W., and J. W. Szostak. 1983. Construction of
artificial chromosomes in yeast. Nature (London) 305:189-193.
11. Murray, A. W., and J. W. Szostak. 1985. Chromosome segregation in mitosis and meiosis. Annu. Rev. Cell Biol. 1:289-315.
12. Orr-Weaver, T. L., J. W. Szostak, and R. J. Rothstein. 1981.
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
DISCUSSION
This paper describes the development of new methods for
making predetermined alterations in chromosome structure.
We studied the mitotic properties of the manipulated chromosomes to explore the relationship between chromosome
structure and behavior. Table 2 lists the mitotic loss frequencies of the molecules we constructed.
The telomeres of normal yeast chromosomes terminate in
tandem copies of the telomere repeat sequence C1 3A (1, 16,
24). Many chromosomal telomeres have one to four tandem
repeats of the 7-kb Y' sequence adjacent to the telomere
repeat sequence, and these in turn are adjacent to a single
copy of the less well conserved X sequence (1). We found
that these sequences are not required for normal chromosome function: a novel version of chromosome III whose
telomeres are composed of the Tr ends plus the yeast
telomeric repeat sequence and lack associated X or Y'
elements shows normal behavior in mitosis and meiosis. The
ability to introduce unique genetic and physical markers into
the telomeres of particular chromosomes is likely to be
generally useful in studies of chromosome structure and
behavior.
In our previous analysis of the structural requirements for
faithful chromosome segregation we suggested two explanations for the extreme mitotic instability of short (10 to 15 kb)
linear artificial chromosomes (10): (i) that the telomeres of
these molecules bind to some hypothetical telomere-binding
site, which prevents centromeres which are physically close
to a telomere from attaching to the mitotic spindle, and (ii)
that catenation of sister chromatids provides a physical
linkage which ensures their accurate segregation. The latter
idea was first suggested by Sundin and Varshavsky (20) and
is the basis of a detailed model of mitotic and meiotic
chromosome segregation (11). The accurate segregation of
telocentric chromosomes (22; this paper) argues strongly
against the telomere-binding model. In addition, we have
shown that a 16-kb dicentric artificial chromosome is rapidly
rearranged to give monocentric derivatives (Murray et al., in
press). This strongly suggests that both centromeres of the
dicentric chromosome are capable of attaching to the mitotic
spindle in a high fraction of cell divisions. If telomere binding
prevented attachment of adjacent centromeres, the linear
dicentric should rearrange slowly or not at all.
The effects of physical length on the mitotic segregation of
circular and linear centromeric plasmids are distinct. Although the stability of circular centromeric plasmids increases with size, molecules as small as 3 kb still show
proper disjunction in more than 95% of cell divisions (6; K.
Bloom, personal communication). Short linear artificial
chromosomes show a completely different behavior: they
are present in many copies per cell and appear to segregate
randomly at mitosis (2, 10). The accurate segregation of
telocentric chromosomes and the rapid rearrangement of a
short dicentric artificial chromosome imply that nearby
telomeres do not affect the ability of the centromere to
interact with the mitotic spindle. We suggest that the poor
segregation of short linear artificial chromosomes is due to
MODIFIED YEAST CHROMOSOMES
3172
13.
14.
15.
16.
17.
18.
19.
MURRAY AND SZOSTAK
Cell 18:309-319.
20. Sundin, O., and A. Varshavsky. 1981. Arrest of segregation
leads to accumulation of highly intercatenated dimers: dissection of the final stages of SV40 DNA replication. Cell
25:659-669.
21. Surosky, R. T., C. S. Newlon, and B.-K. Tye. 1986. The mitotic
stability of deletion derivatives of chromosome Ill in yeast.
Proc. Natl. Acad. Sci. USA 83:414-418.
22. Surosky, R. T., and B.-K. Tye. 1985. Construction of telocentric
chromosomes in Saccharomyces cerevisiae. Proc. Natl. Acad.
Sci. USA 82:2106-2110.
23. Szostak, J. W. 1982. Structural requirements for telomere resolution. Cold Spring Harbor Symp. Quant. Biol. 47:1187-1194.
24. Szostak, J. W., and E. G. Blackburn. 1982. Cloning yeast
telomeres on linear plasmid vectors. Cell 29:245-255.
25. Tschumper, G., and J. Carbon. 1980. Sequence of a yeast DNA
fragment containing a chromosomal replicator and the TRPI
gene. Gene 10:157-166.
26. Zakian, V. A., H. M. Blanton, L. Wetzel, and G. M. Dani. 1986.
Size threshold for yeast chromosomes: generation of telocentric
chromosomes from an unstable minichromosome. Mol. Cell.
Biol. 6:925-932.
Downloaded from mcb.asm.org at Harvard Libraries on July 16, 2009
Yeast transformation: a model system for the study of recombination. Proc. Natl. Acad. Sci. USA 78:6354-6358.
Orr-Weaver, T. L., J. W. Szostak, and R. J. Rothstein. 1983.
Genetic application of yeast transformation with linear and
gapped plasmids. Methods Enzymol. 101:228-245.
Rothstein, R. J. 1982. One step gene disruption in yeast.
Methods Enzymol. 101:202-211.
Schwartz, D. C., and C. R. Cantor. 1984. Separation of chromosome sized DNAs by pulsed field gradient gel electrophoresis. Cell 37:67-75.
Shampay, J., J. W. Szostak, and E. Blackburn. 1984. DNA
sequence of telomeres maintained in yeast. Nature (London)
310:154-157.
Sherman, F., C. Fink, and C. Lawrence. 1979. Methods in yeast
genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor,
N.Y.
Southern, E. M. 1975. Detection of specific sequences among
DNA fragments separated by gel electrophoresis. J. Mol. Biol.
98:503-517.
Strathern, J. N., C. S. Newlon, I. Herskowitz, and J. B. Hicks.
1979. Isolation of a circular derivative of yeast chromosome III:
implications for the mechanism of mating type interconversion.
MOL. CELL. BIOL.