Cop)light 0 1996 by the Genetics Society of America

The rad9 Gene of Coprinus cinereus Encodes a Proline-Rich Protein Required

for Meiotic Chromosome Condensation and Synapsis

Lisa C. Seitz, Keliang Tang, W. Jason Cummings and Miriam E. Zolan

Department of Biology, Indiana University, Bloomington, Indiana 47405
Manuscript received July 19, 1995
Accepted for publication December 22, 1995

ABSTRACT
The rad9 gene of Copnnus cinereus is essential for the normal completion of meiosis. We examined
surface-spread preparationsof wild-type andrad91 nuclei from the meiotic stages of karyogamy through
metaphase I, and we determined the primary sequence, structure, and meiotic expression of the rad9
gene. Inwild-type C. cinereus, karyogamy is followed by condensation and alignment ofhomologous
chromosomes. Condensation and axial core development largely precede synapsis, which often initiates
at telomeres. A diffuse diplotene phase coincides with dissolution of the synaptonemal complex, and
subsequently chromosomesfurther condenseas the cells progressinto metaphaseI. In contrast, although
karyogamy and nucleolar fusion are apparently normal in rad91 basidia, only short stretchesof synapto-
nemal complex form. These correlate with stretches of condensed chromatin, mostly atapparent chromo-
some ends, and regions of presumptive triple synapsis are numerous. rad9-1 basidia enter the diffuse
stage of early diplotene, and then 50% of these cells enter metaphase I by the criteria of nucleolar
elimination and at least some chromatin condensation. rad9 gene expression is induced after gamma
irradiation and duringmeiosis. The gene has 27 exons and encodesa predicted protein of 2157 amino
acids, with a proline-rich amino terminus.

T HE evolution ofmeiosis has likely included the

refinement of mechanisms that ensure the proper
segregation of homologous chromosomes during the
first meiotic division. Two prominent, and related,pro-
cesses are recombination and the formation of a tripar-
tite proteinaceousstructure called the synaptonemal
complex (SC; MOSES 1968). Analysis of the respective
roles of recombination and the SC in chromosome seg-
regation has led to the idea that recombination-based
processes are important both for initial homolog recog-
nition (CARPENTER 1987) and for the maintenance, in
the form of chiasmata, of connections between h o m e
logs until anaphase I (BAKERet al. 1976; ROEDER1990).
The role of the SC in most organisms is not homolog
recognition per se; SC formation occurs readily between
nonhomologous chromosomes (LOIDLet al. 1991) and
nonhomologous regions of chromosomes (MCCLIN-
TOCK 1933; MAGUIRE and 1994). Instead, the SC
may be essential for the conversion of cross-over events
into chiasmata (ENGEBRECHT et al. 1990;SYMet al. 1993).
The formation and dissolution of the SC during mei-
osis and the condensation and synapsis of homologs
have been examined in detail using electron micros
copy (VON WETTSTEINet al. 1984). Components and
processes of SC formation have recently been identified
and characterized in several organisms (for example,
HOLMet al. 1981; DRESSERand GIROUX 1988; DIETRICH
Caespondzngauthw: Miriam Zolan,Department of Biology, Indiana
University, Bloomington,IN 47405.
E-mail: mzolan@bio.indiana.edu
Genetics 1 4 2 1105-1117 (April, 1996)
et al. 1992; S m et al. 1993; DOBSONet al. 1994).In
addition to proteins that function directly in SC struc-
ture and formation, other gene products with roles in
general DNA metabolism are indirectly involved in suc-
cessful SC formation and are directly or indirectly a
part of meiotic recombination and the progression of
meiosis. A prominent classof these genes is the rad
genes which are required both for DNA repair and
for the successful completion of meiosis (GAME1983,
1993). It has been proposed that mitotic DNA repair
processeshave been recruited during evolution into
roles that facilitate the synapsis and segregation of h e
mologs ( A L A N 1 et al. 1990; KLECKNERet al. 1991). It
has also been suggested that DNA repair per se is an
importantcomponent of meiotic DNA metabolism
(BERNSTEIN and BERNSTEIN1991); however, it is diffi-
cult to understand how defects in DNA repair would
aflect meiotic progression and chromosome behavior
if the role of DNA repair in meiosis were restricted to
base sequence error correction.The available evidence
points to a more direct function of certain rad genes
in the behavior of meiotic chromosomes, because muta-
tions in these genes result in defects in chromosome
condensation, synapsis, and segregation (ZOLAN et al.
1988; ALAN1 et al. 1990; PUMU et al. 1992; VALENTINE
et al. 1995; RAMESH and ZOLAN 1996).
We are studying rad genes in the basidiomycete Copri-
nus cinereus,in which meiosis is naturally synchronous.
Each mushroom cap contains 10 million meiotic cells,
which go through meiosis in one wave, tightly coordi-
nated with fruiting body development. At the midpoint
1106 C. L. Seitz et al.
of pachytene, which lasts for -5 hr, essentially all mei-
otic cells of the mushroom cap are at this stage ( ~ J U
and LU 1970). Thus, C. ciweus is an excellent experi-
mental organism to use for examining the processes of
meiotic chromosome metabolism, the underlying ge-
netic basis for these processes, and the relationship be-
tween meiotic chromosome behavior and other aspects
of DNA metabolism.
The rad9 gene was initially identified in a screen for
radiation-sensitive, meiotic mutants (ZOLAN et al. 1988).
The rad9-l mutant is characterized by reduced basidio-
sporeproduction, low spore viability, and failure to
complete meiotic divisions (ZOLAN et al. 1988; VALEN-
TINE et al. 1995). ZOLAN et al. (1988) observed stretches
of SC in thin-sectioned embedded nuclei of rad9-1 cells,
and these were not discernably different, in kind or
in frequency, from those observed in wild-typecells.
However, because single sections of nuclei were exam-
ined, the lengths of the observed SCs and the time
course of SC formation and dissolution were not deter-
mined.
In a separate study, the rad9 gene was isolated from
a chromosome-specific cosmid library (ZOLAN et al.
1992). Thecloned gene was shown to complement the
rad9-l mutation for both radiation resistance and meio-
sis, and it mapped to the genetically determined rad9
locus.
To further investigate the function of the rad9 gene
in meiosis, we examined the time course of SC forma-
tion and dissolution, and the concomitant condensa-
tion cycles, of wild-type and rad9-l cells. We found that
the rad9-l mutant has a novel meiotic defect, and that
the two stages of meiotic chromatin condensation are
uncoupled in this mutant. In addition, we determined
the sequence and structure of the rad9gene, and found
two striking features: the gene is unusually large, with
an open reading frame of nearly 6.5 kb; and the amino
terminus of the predicted Rad9 protein has a proline
content thatis more than fourtimes that of the remain-
der of the protein.
MATERIALSAND METHODS
Strainsandcultureconditions: The wild-type dikaryon
used in this study has been previously described (VALENTINE
et al. 1995). Two rad91 dikaryons, both congenic with the
wild-type control strain, were examined. Each was formed by
crossing sibling rad91 isolates from the fourth generation of
backcross into the Rad+ strain Okayama-7 (Wu et al. 1983).
Culture conditions, matings, and fruiting conditions were as
previously described (ZOLAN et al. 1988).
Microscopy: Surface spreads of C. cinereus chromosomes
were prepared as described by PUKKILA et al. (1992). Spreads
were viewed and photographed using a Philips EM301 trans-
mission electron microscope. For the determination ofSC
length, photographs were scanned using Photoshop (Adobe
Systems, Inc.) and lengths were measured using NIH Image,
version 1.56 (RASBAND and BRIGHT1995) on a Macintosh
Quadra 800. To observe chromosome condensation,samples
were fixed and spread in the same manner as for electron
microscopy, but were spread on ProbeOn Plus microscope
slides (Fisher Scientific). Slides were stained with 0.1 mg/ml
acridine orange (Sigma A4921) in 70 mM phosphate buffer
(22.4 mM Na2HP04,47.6 mM KH2P04, pH6.5) for 8-12 min
at 4" (DRESSER and GIROUX1988). Slides were rinsed with
cold 70 mM phosphate buffer, andthen samples were
mounted with the same buffer. Spreads were viewed and pho-
tographed using a Nikon Microphot FXA and Ektachrome
slide film (Kodak).
Nucleic acid isolation, electrophoresis, and hybridization:
RNA was isolated from cap and stipe tissue and analyzed by
Northern analysis as described by YEACER STASSENet al.
(1996). DNA was isolated as described by ZOLANand PUKK~LA
(1986), and gel electrophoresis and Southern hybridization
were done as described in ZOLAN et al. (1992).
Isolation of rad9 cDNk A C. cinereus meiotic cDNA library
(YEAGER STASSEN et al. 1996) was screened using clones that
spanned the genomic sequence of rad9. Three cDNA clones,
which included 2.3 kb of the 3' end of the gene, 1 kb of the
5' end, and0.7 kb of the middle of the gene,were recovered.
After these were sequenced, reverse transcription and PCR
(RT-PCR) were used (asdescribed below) to isolate cDNA
representing the remainder of the gene.
DNAsequencing: Clones for DNA sequencing were pre-
pared by either the Wizard miniprep procedure (Promega)
or by a Qiagen plasmid kit. DNA sequencing was performed
by the dideoxynucleotide method (SANCERet al. 1977) with
35Slabeled nucleotides and Sequenase (US Biochemical), by
automated sequencing at the University of North Carolina-
Chapel Hill automated DNA sequencing facility on a model
373ADNA sequencer (Applied Biosystems) using the Taq
DyeDeoxy Terminator Cycle sequencing kit (Applied Biosys-
terns), and by automated sequencing at the IndianaInstitute
for Molecular and Cellular Biology using a Li-Cor model 4000
DNA sequencer following reaction preparation with SequiTh-
erm thermostable DNA polymerase (Epicentre Technolo-
gies). Sequences were assembled using the DNAsis (Hitachi
Software Engineering Co.) program. Subclones representing
the entire lO.&kb Smd-SpeI fragment of cosmid l B l l (Figure
7 ) were sequenced, using manual sequencing. Several s u b
clones (into Bluescript, Stratagene) were used for this pur-
pose. In addition, oligonucleotide primers (19 or 20 nucleo-
tides) were used as internalpriming sites forsequencing
reactions. All primers, for both DNA sequencing and PCR
(see below) were designed using the primer analysis program
Oligo (National Biosciences, Inc.). A single strand of DNA
sequence was determined for the 10.8-kb SmaI-SpeI genomic
fragment. Then, cDNA clones were sequenced such that the
second strand of coding region sequence was determined,
and the ends of PCR clone Bwere sequenced. Finally, primers
and subclones were used to complete the second strand of
all intron sequences. Therefore, double strand sequence in-
formation was obtained for the complete region spanned by
PCR clone B of Figure 7 , and this information was submitted
to GenBank (accession no. U34998). The cDNA and genomic
sequences were compared to determine the locations and
sizes of introns.
C. cinereus transformations: Transformations into C. n'ner-
eus protoplasts were performed as described in ZOLAN et al.
(1992). For the transformations using clones of the PCR prod-
ucts shown in Figure 7, 0.5 p g of cosmid Llc5200 (PUKKILA
and CASSELTON 1991) was mixed with 2 pg of either a single
clone or a pool of clones. Alternatively, 1 p g of cosmid was
mixed with 5 p g of cloned DNA.
PCR The 9785- and 8909-bp PCR clones shown in Figure
7 were constructed by amplification from cosmid 1B11,which
is also shown in Figure 7. Primers used were 23 or 27 nt in
length, and the protocol and reagents of the Expand Long
 rad9 Gene of Cojm'nzts 1107

Template PCRSystem (Boehringer Mannheim) were used.

For the amplification, 100 ng of template DNA was amplified
in a Perkin-Elmer GeneAmp PCRSystem 2400, using 1.8 U
enzyme in a .50 p1 reaction, with the following cycling condi-
tions: 94". 2 min; followed by 94", 15 sec; 62", 30 sec; 68", 8
min for 10 cycles; followed by 94", 15 sec; 62", 30 sec; 68". 8
min plus 20 sec additional each cycle, for 10 cycles; followed
by 68",7 min. Amplified fragments were cloned directly (with-
out gel purification) into the pCRII vector (Invitrogen).
For RT-PCR, total RNA from caps at 1 or 6 hr postkaryo-
gamy was used for first strand cDNA synthesis. For each region
to be amplified, -20 pg of RNA was mixed with 144 or 198
pmoles of the appropriate downstream primer, and heated
at 70" for 15 min. After chilling on ice for 2 min, samples
were spun briefly in a microcentrifuge, and reverse transcrip
tion was carried out at 37" for 60 min, using 100 U M-MLV
reverse transcriptase (BRL) in buffer supplied by the manu-
facturer, supplemented to 10 mM DTT and 2.5 mM each
dATP, dCTP, dGTP and dTTP (Promega), andcontaining 40
U RNasin (Promega). Samples were held at 100" for 5 min,
and then treated with DNAse-free RNkse (Boehringer Mann-
heim) at 400 pg/ml for 30 min at 37". PCR was then carried
out in a 25-pI reaction containing 5pl of the reverse transcrip-
tion reaction and using 5 U Taq polymerase (Promega) in
buffer supplied by the manufacturer, supplemented to 1 mM
MgC12, 200 nM for each dNTP, and 1 mM for the appropriate
upstream primer. Amplification conditions were: 94", 2 min;
followed by 94", 15 sec; 55", 30 sec; 72", 1 min, for 30 cycles;
followed by72". 7 min. Reactions were then separated on
agarose gels, and sample lanes were blotted and hybridized
with a labeled insert of a clone containing the corresponding
genomic regions of the rad9 gene. The smallest hybridizing
fragments were isolated (using the EluQuick kit, Schleicher
and Schuell) and cloned into the pCRII vector.

RESULTS
Meioticchromosomeprogression in wild-type C.
c
ine
reu: The time course of meiosis in C. cinereus has
s
been examined in intact basidia, by both light (RAJU
and LU 1970; PUKKILA et al. 1984) and electron micros-
copy (HOLMet al. 1981). Recently, cell disruption and
surface spreading techniques have been refined (PUK-
KlLA and Lu 1985; ZoLAN et d . 1988; PumLA et al.
1992), enabling many features of meiotic progression
to be monitored at high resolution for numerous cells.
To provide a framework for understanding thebehavior
of meiotic chromosomes in rad mutants, we examined
the time course of meiosis ina congenicwild-type strain,
using both light and electron microscopy of surface-
spread preparations.
Meiosis in C. cinereus begins with karyogamy, which is
followed rapidly by the fusion of nucleoli (Figure 1, A
and B). We define karyogamy as the time at which
roughly half of all basidia have nuclei.
fused In our obser-
vations, this occurs 1 hr before the lights are turned on,
when cultures are maintained at 25" with a 16hr light,
8-hr dark cycle. We find that 290% of basidia undergo
karyogamy in a wave that lasts -20 min. Thus, the time
point defined as karyogamy in our figures will naturally
show a range of early meiotic stages (Figures 1, A and
B, and 2; see also w j u and Lu 1970). After karyogamy,
condensation of chromatin progresses until distinct,
FIGURE1.-Time course of chromatincondcnsation in
wild-type C. n'nmms. Samples were fixed, spread, and stained
with acridine orange. Photographs are from samples taken at
the following times: (A and B) karyogamy (as defined in the
text); (C) Karyogamy plus 1 hr; (D) Karyogamy plus 4 hr; (E)
Karyogamy plus 6 hr; ( F ) Karyogamy plus 8 hr; (G and H)
karyogamy plus 9 hr. The arrow in A marks a nucleolus, and
the arrow in D marks synapsed telomeres. All size bars indicate
a length of 2 pm.
condensed chromosomes are apparent (Figure 1C; the
stagesof condensation intermediate between those
shown in B and C are not shown). In previous studies
of intact basidia, LU and R,qu (1970) found distinct
chromosomes in fusing nuclei. The reason for the dis-
crepancy between their data and ours is not known, al-
though it is possible that our spreading method disrupts
early, partial chromatin condensation. We normally see
condensation of the level shown in Figure 1Cbefore we
observe synapsis(Figure 1D), and synapsis often appears
to initiate at telomeres (Figure lD, arrow). Full pachy-
tene synapsis (Figure 1E) is observed for -90% of basi-
dia at 6 hr postkaryogamy and is the major class of chro-
matin observed at 4, 6, and 7 hr postkaryogamy (Figure
1108 L. C. Seitz et al.
100 C I A
.- I (1985).
u
3
c 75i
0 1 2 4 6 7 8 9 m H
Hours after karyogamy
FIGURE2.-Percent of nuclei at stages shown in Figure 1
for time points sampled duringmeiotic development. Because
Figure 1 does not show all stages observed, percentages do
not sum to 100. The letters A-H refer to panels of Figure 1.
Karyogamy time (0)is defined in the text. Numbers of nuclei
scored for each time point were: 0 hr, 17; 1 hr, 27; 2 hr, 16;
4 hr, 38; 6 hr, 16; 7 hr, 4; 8 hr, 10; 9 hr, 23.
2). Pachytene is followed by a “diffuse stage” (Figure
1 F see also LU and RAJU 1970), which is similar to that
described for Saccharomyces cerevisiae (DRESSER and GI-
ROUX 1988), for Swdmia mucrospwa (ZICKLER1977), and
for many organisms with large chromosomes (WILSON
1925). This stage of decondensation is followed by re-
condensation (Figure lG) until metaphase I figures (Fig-
ure 1H; note also the loss of the nucleolus) are observed,
at 9 hr postkaryogamy. By 12 hr after karyogamy,all
basidiahave undergone both meiotic divisions ( R A ~
and Lu 1970; VALENTINE et al. 1995).
Time courseof SC formation in wild-type C. ci-:
Samples prepared as for chromosomecondensation
studies were spread on plasticcoated slides, stained with
silver nitrate, and transferred to grids for viewing by
electron microscopy (PUKKILA and LU 1985;PUWLAet
al. 1992; Figure 3). Using this procedure, the lateral
elements of the SC are clearly visible, but the central
elements are not seen. After karyogamy and nucleolar
fusion (Figure 3, A and B),axial core formation occurs
rapidly. At 1 hr postkaryogamy, 92% of nuclei examined
were at orbeyond the stage shown in Figure 3C (Figure
4). For both the karyogamy and 1 hr postkaryogamy
samples, spreads such as that shown in Figure 3D were
20% of those observed (Figure 4). These spreads fea-
turelong axial cores and one or more regions a p
proaching synapsis, and it is common for the synapsed
regions to look like chromosome ends, as shownin
Figure 3D. Complete homolog alignment (Figure 3E)
is followed by progressive tightening of the SC structure
(Figure 3F) until full pachytene synapsisis reached
(Figure 3G). The beginnings ofSC degradation are
apparent at about 8 hr postkaryogamy (Figure 3H). Ten
pachytene spreads (the stage shown in Figure 3G) were
measured, and the average total SC length was 76 t
9.5 pm (95% confidence limits), in good agreement
with the 70 pm average reported byPUKKILAand Lu
Meiotic defects in rad9-1 basidia: Basidia of rad9-I
mushrooms undergo karyogamy at the same time, rela-
tive to the light cycle, as wild-typecultures. In addition,
nucleolar fusion appears to occurwith normal kinetics
(Figures 5C and 6A are spreads from samples taken at
a karyogamy time point). At early time points (Figure
5, B and C) a large proportion of the spreads show a
ring-like region of condensation against a noncon-
densed background. At karyogamy,51%of spreads
have a ring structure, and at 1 hr postkaryogamy, 33%
contain a ring. This stage is alsoobserved (in -5%
of the spreads) in wild-type basidia at karyogamy time
points, but is not observed thereafter; its persistence in
rad9-l indicates that it may be a normal stage that the
wild-type passes through quickly, but that lasts longer
in the mutant.
Other features of prophase I in rad9-1 basidia are
consistent with both arrest and abnormalities of the
processesof chromatincondensation, chromosome
pairing, axial core development, and synapsis. Examina-
tion of acridine orangestained spreads (Figure 5 ) re-
veals that fullpachytene-like chromosome condensa-
tion is never observed. Instead, short stretches of
condensed and apparently paired regions are visible,
against a diffuse background, at 6 and 8 hr postkaryo-
gamy (Figure 5, arrows in D and F).By 10 hr postkaryo-
gamy, these structures have disappeared (Figure 5G).
At this time point, 92% of spreads observed had the
diffuse appearance shown in Figure 5G, and 8% had
the condensed appearance shown in Figure 5H. At 12
hr postkaryogamy, when wild-type basidia have com-
pleted both meiotic divisions (VALENTINEet al. 1995),
we scored 213 rad91 nuclear spreads, and found that
117, or 55%, were condensed asin Figure 5H, and
96, or 45%, were very diffuse, and retained the large
nucleolus characteristic of pre-metaphase I spreads.
Thus, no rad91 basidia were observed to undergo a
normal levelof pachytene condensation. About half
arrest at the diffuse stage which follows pachytene, and
about half progress into metaphase I. However, the
metaphase I stage of rad9-l mutants is not normal; the
compaction observed is not as great as that seen in wild-
type spreads (compare Figures 1H and 5H), and it is
most likely that homologs are not paired.
Higher resolution of the rad9-I phenotype was ob-
tained by silver staining and electron microscopy (Fig-
ure 6). By this technique, it is clear that axial core
formation is incomplete; in fact, the short stretches of
condensed and paired chromosomes seen in the light
microscope (Figure 5) are echoedby short stretches of
synapsed axial core elements, superimposed on a dif-
fuse background (Figures 6, B-E). The total amount
of axialcore development and synapsis inrad91 basidia
seems to be established early after karyogamy (Figure
 rad9 Gene of Copn'nu.5
- .. .
6B) and does not progress (Figure 6E). Measurement
of six spreads of samples taken at 6 hr postkaryogamy
yielded a range of SC length from 0.7 to 10 pm, with
the average length 3 ? 0.6 pm (95% confidence limits).
At 12 hr postkaryogamy (Figure 6F), most SC elements
have degraded, leaving cellsthat have not enteredmeta-
phase I with a verydiffuse appearance (Figure 6F).
Three other features of rad91 axial core and SC devel-
opment are notable. First, most or all of the axial core
segments that form are also synapsed (Figure 6, B-E).
Second, many of the synapsed regions have the appear-
ance characteristic of chromosome ends (arrows in Fig-
ure 6C). Third, numerous tripartite stretches are ob-
1109
FIGURE3.-Time course of SC forma-
tion inwild-type C. cinereus. Photographs
are from samples takenat the following
times: (A-D) Karyogamy (as definedin the
text); (E and F ) Karyogamy plus 4 hr; ( G )
Karyogamy plus 6 hr; ( H ) Karyogamy plus
8 hr. The arrowinAmarksa nucleolus,
the arrow in D marks synapsing telomeres,
and the arrows in E and F mark synapsed
telomeres. AI! size bars indicate a length
of 1 pm.
served (arrowheads in Figure 6, B-D). While it is
possible that these stretches represent canonical tripar-
tite SC regions, the fact that the central elementof the
SC is not observed in wild-typespreads (Figure 3) makes
it more likely that the tripartite stretches observed in
md9-I cells are instead regions of triple synapsis. This
interpretation is supported by the observation of pair-
ing partner switches. An example of such a switchis
indicated by the left-most arrowhead of Figure 6B. In
this case, the central element switchesfromsynapsis
with an element on its left to synapsis with an element
on its right. In addition, in most cases of
apparent triple
synapsis the middle element of the tripartite stretch is
1110 L. C. Seitz et nl.

100 I ,OA
n

0 1 4 6 a
Hours after karyogamy
FIGURE4.-Percent of nuclei at stages shown in Figure 3,
for time points sampledduringmeioticdevelopment. Be-
cause Figure 3 does not show all stages observed, percentages
do not sum to 100. The letters A-H refer to panels of Figure
3, and stages intermediatein SC structure between the nuclei
shown in D and E are grouped with E. Karyogamy time (0)
is defined in the text. Numbers of nuclei scored foreach time
point were: 0 hr, 26; 1 hr, 24; 4 hr, 56; 6 hr, 51; 8 hr, 42.

the same thickness as the outer elements. Were these

true tripartite, individual SC stretches, the central ele-
ments wouldlikely be much thinner, relative to the
outer, lateral elements (MOSES 1968; DRESSER and GI-
ROUX 1988).
Molecular analysis of the md9 gene: We previously
constructed a cosmid library of the chromosome 8,s
doublet from strain Okayama-7 and screened it for
clones that could complement boththe radiation sensi-
tivity and the meiotic defect in rad91 (ZOLAN et al.
1992). We readily obtained complementation and used
the process of sib-selection, in which we transformed
with progressively smaller pools of clones, to isolate an
individual clone, termed 1B11, whichcontains the rad9
gene (Figure 7). Further evidence that this clone con-
tains the complete rad9 gene, and not an extragenic
suppressor, are the observations that the clonemaps to
the position of the rad9-1 mutation and that it comple-
ments this mutation when present in transformants in
a single copy (ZOLAN et al. 1992).
To define the region of cosmid l B l l that contains
the rad9 gene, cotransformations were performed. In
these experiments, an intact cosmid vector containing
the C. cinereus t r p l gene was mixed with gel-purified
fragments of 1B11, and the DNA was transformed into
rad9-l;trpl-l, 1-6 mutants using standardprocedures
(BINNINGER et al. 1987; ZOLAN et al. 1992). Trans-
formants were selected for tryptophan prototrophyand
screened for radiation resistance and the formation of
spores. The 10.8-kb SpeI-SmaI fragment shown in Figure
7 complemented bothof the rad9-l mutant phenotypes.
The location of the rad9 gene was further confirmed
by a series of transformations using overlapping cos-
mids. One cosmid, 1D7, was found that overlaps l B l l
significantly, but does not complement the defects in
FI(;uKE .5.--Chromatin conrlensarion i n I-udY-I. I ' h o r o -
graphs are from samples taken at the following times: (A-C)
Karyogamy; (D and E) Karyogamy plus 6 hr: (F) Karyogamy
plus 8 hr; (G and H) Karyogamy plus 10 hr. The arrow in
A indicates a nucleolus, the arrow in C indicates a ring of
condensation, and the arrows i n D and F indicate short con-
densed regions that also appear to he synapsed chromosome
ends, as discussed in the text. Numbers of nuclei scored were:
0 hr, 46; 1 hr, 18, 6 hr, 32; 8 hr, 16: 10 hr, 24; 12 hr, 213. All
size bars indicate a length of 2 pm.
rad9-1 strains (Figure 7). In separate transformation
experiments, the levelof complementation of rad9-I
with cosmid l B l l ranged from 52 to 89% (ZOIAN et
al. 1992), whereas no complementation, of either the
radiation sensitivity or meiotic defects of 1-ad9-1,was
obtained for 50 transformants with cosmid 1D7. We
therefore concluded that the rad9gene is probably lo-
cated near the endof cosmid 1B11 that is truncated in
cosmid 1D7 (Figure 7).
Using a probemade from a subcloneof cosmid 1B11,
which overlaps the end of cosmid 1D7, we examined
RNA from vegetative monokaryons and dikaryons, and
from fruiting body caps. With this probe, we identified
 rad9 Gene of Coprinus 1111
.
i
.

t. '.

.a
..*
.i
I . FIGURE 6.-SC in formation
A. from
rad9-I. arePhotographs
_...
samplestakenat the following
times: (A) Karyogamy; (B) Karyo-
gamyplus 1 hr; (C) Karyogamy
plus 4 hr; (D) Karyogamy plus 8
hr; (E) Karyogamy plus 10 hr; (F)
Karyogamy plus 12 hr. The arrow
in A indicates a nucleolus, the
arrows in C indicate regions
whichappear to be synapsed te-
lomeres, and arrowheads indicate
regions of apparent triple synap
sis, as discussed in the text. Num-
bers of nuclei scored were: 0 hr,
19; 1 hr, 35; 4 hr, 40; 6 hr, 32; 8
hr, 8; 10 hr, 15; 12 hr, 31.All size
E bars indicate a length of 1 pm.

.
+ :

..
. .

a transcript of -6.6 kb, which is induced after gamma
irradiation (data not shown) and during meiosis (Fig-
ure 8; the probeused for this northern was a rad9cDNA
clone insert, but identical results were obtained using
genomic DNA clones). By doing northern hybridiza-
tions with small fragments from across the region, we
determined that the 6.6-kb transcript was roughly cen-
tered in the 10.8-kb SpeI-SmaI fragment shown in Figure
7 and that anothertranscript, of -2.7 kb, was detected
by sequences at the SpeI end. Both transcripts were also
present in the rad9-l mutant(datanotshown), al-
though transcript levels were not compared between
mutant and wild-type strains.
We sequenced the entire10.8-kb genomic region and
found, using the BLAST procedure (ALTSCHUL et al.
1990), that the 2.7-kb transcript encodes an ATPdepen-
dent RNA helicase of the DEAH box family (M. E. Zo-
LAN, unpublished data; LINDERet al. 1989). The two
closest homologs of the C. cinereus helicase are the PRP2
gene of S. cereoisiae (CHEN and LIN 1990) and the PRHl
gene of Schizosaccharomycespombe (INOUE et al. 1992),
both of which function in RNA splicing. We estimate
that the 10.8-kb SpeI-SmaI fragment contains -65% of
the C. cinereus helicase homolog. Because this fragment
contains the complete coding sequence of the 6.6-kb
transcript and completely complements the defects in
rad91 strains in transformation experiments, we rea-
soned that the6.6-kb transcript is most likely the actual
rad9 gene.
To definitively rule out the helicase homolog as the
source of complementation of the rd9-I mutation, we
used PCR to amplify, from cosmid 1B11, the regions
shown as A and B in Figure 7. In cotransformation
experiments with the trpl vector, clones of each frag-
ment complemented the defects of rad9-l strains. Be-
cause fragment B includes only 325bp upstream of the
rad9 translational start site (Figure 7), and none of the
helicase coding sequence, we conclude that the 6.6-kb
1112 L. C. Seitz et al.
Sml
I I
I I
-1201
-325
transcript encoded at the end of cosmid l B l l is indeed
derived from the rad9 gene.
To determine the structure and coding sequenceof
the rad9gene, thecDNA sequence was determined and
compared with the genomic sequence. The rad9 open
reading frame is 6471 bp in length and is interrupted
by 26 introns (Figures 7 and 9), which range in size
from 47 to 60 bp. The introns are found in all three
rad9- I I QI, a et al. 1987) or the YNYRAYeukaryotic consensus (SHARP
6 13 23 1 5 9 3 bases upstream of the 3' splicejunction. This positional
relative expression
FIGURE 8.-Expression of the rad9 gene during meiosis.
For each lane, 25 pg of total RNA from either caps (cap), a
vegetative dikaryon (vd), or stipes (stipe), was treated with
RNAse-free DNAse,and gels wererun, blotted and hybridized
as described (YEAGERSTASEN et al. 1996). The probe used
was the insert of a cDNA clone corresponding to the 3' end
of the rad9 gene. After the hybridized blot was exposed to
X-ray film, hybridization information was recorded using a
phosphorimager (Molecular Dynamics). Without stripping,
the blot was then hybridized with a probe for the C. cinereus
ribosomal RNA gene repeat (Wu et al. 1983), and the amount
of hybridization of the rad9 probe was normalized to that of
the RNA gene probe. The numbers above each lane except
the vd lane indicate the time, relative to karyogamy, of collec-
tion of each sample. The numbers below each lane indicate
the level of normalized radPspecific hybridization, relative to
the vd sample.
FIGURE 7.-Organization of the rad9
gene. Cosmids 1D7 and lBll are from a
library of chromosomes 8 and 9 (ZOLAN et
1D7(-) al. 1992). The unlabeled lines on the cos-
mid mapsindicate Hind111 sites. In the mid-
dle panel, the region labeled helicase has
lBll(+)
sequence similarity with the Saccharomyces
cerevisiae gene PRP2 and the Schizosaccham
myces pombegene PRHI. The stippled boxes
indicate regions ofcDNA outside of the
presumptive translated region, and the
ATG and TAG represent the presumptive
start and stop sites for translation. Solid
boxes indicate exons, and open boxes indi-
cate introns. In the bottom, A and B repre-
sent PCR products and clones, which are
described in the text. A (+) indicates that
1 a givenDNA fragment or clone comple-
ments the defects in rad91 strains; a (-)
I
indicates that a givenDNA fragment or
4584 clone does not complement rad91 for ei-
ther the survival of gamma irradiation or
for meiosis.
possible codon phases (Figure 9), with an even distribu-
tion of interruptions between codons (46%)and within
codons (54%).Exons usually begin with a purine resi-
due and end with a guanine, and all intron ends are
defined by the 5' GT and 3' AG observed for virtually
all spliceosomal introns. The 5' splice-site consensus
sequence, GTRNGT (Figure 9), agrees with that gener-
ally found forfilamentous fungi (GURRet al. 1987; EDEL
MANN and STABEN 1994), as does the3' splice site, YAG.
In two cases, the Y at the3' splice site was replaced by an
A, as has been observed for other fungal genes (GURR et
al. 1987).
The 26 rad9 introns do notconsistently display inter-
nal sequences similar to either the YGCTAAC branch
point consensus described for filamentous fungi (GURR
1987). However, 22 of the 26 introns contain the se-
quence CTNA (underlined in Figure 9), which is similar
to the NYRA core of YNYRAY. In 21 of these introns,
the A residue of the CTNA sequence is located 12 to 20
conservation, coupled with the preservation of the
CTNA sequence itself, ispossibly indicative of func-
tional properties conferred by this consensus. If CTNA
does represent a partially conserved branch point in C.
cinereus introns, then the associated adenosine residue
may serve as the attacking nucleophile at the 5' splice
site, as seen in other cases ofeukaryotic splicing (SHARP
1987).
The predicted Rad9 protein has 2157 amino acid
residues, a molecular weight of 239 kD, a PI of 7.3, and
a proline-rich amino terminus (Figure 10). The first
430 amino acids are 17.7%proline, whereas the remain-
der of the protein is 4.3% proline, which is consistent
with the average composition of eukaryotic proteins
(4.6%proline; DOOLITTLE 1986). Analysis of hydropa-
thy (KWE and DOOLITTLE 1982) for Rad9 indicated that
 1113

4 l574 1 52
5 2267 0 48
6 2792 0 49
7 2941 2 55
8 3602 1 48
9 3754 0 58
10 4015 2 51
11 4259 0 5 4
I2 4460 0 49
l3 4953 0 50
14 5ll4 0 52
ls 5310 0 51
16 5455 1 48
17 5644 1 5¶
18 5822 1 57
19 6043 1 51
ao 6179 a 51
a1 6366 0 57
22 6601 1 5 4
23 6795 0 51
24 7108 1 47
as 7255 2 56
26 7462 0 49

5 ' COllSePSUS: G6a g l o o t l o o lh-9 g88 e 7 7 3 consensus: ys a l o o g ~ o o R 7 9

RCURE 9.-Intron sequences of the rad9 gene. Uppercase letters represent flanking exon sequences and lowercase letters
indicate intron sequences.No, intron number; Pos, intron position relative to the ATG shown in Figure 7; Ph, intron phase (0,
between two codons; 1, between the first and second nucleotides of a codon; 2, between the second and third nucleotides of a
codon). Underlined sequences are the conservedCTNA sites discussed in the text.

there are no membrane-spanning domains. There are
four sites in the Rad9 protein that are consensus se-
quences for phosphorylation by mitogen-activated pro-
tein kinase (MANSOUR et al. 1994; these are underlined
in Figure 10). Additionally, the size and basic amino
acid composition of the sequencePVKKGSRKKK (posi-
tions 607-616, Figure 10) lead us to speculate that this
region may function in nuclear localization (GARCIA-
BUSTOSet al. 1991).
Database searches (using the BLAST algorithm; ALT-
SCHUL et al. 1990) identified sequences with similarity
to two different regions of the Rad9 protein. Significant
matches (with random hit probabilities of 7.8 e-52 and
5.9 e-33, respectively), were to predicted proteins of
unknown function in S. pombe (chromosome I cosmid,
Genbank No. 250113) and S. cereoisiae (chromosome N
cosmid, Genbank No. SC9395). Sequence alignments
revealed that these proteins are probably homologous;
for each pair-wise comparison, there is 23% identity
(excluding the 21-29small gaps), over a region of
-1000 amino acids. However, the S. pombe protein, at
1583 amino acids, and the S. meuisiae protein, at 1494
amino acids, are significantly smaller than the Rad9
protein, and neither contains a dramatically proline-
rich amino terminus.
The next highest match in BLAST searches was be-
tween the proline-rich amino terminus ofRad9 and
various proline-rich proteins, such as the plant cell wall
protein extensin (KIELISZEWSKI and LAMPORT 1994).
However, these BLAST probability scores (2.4 e-16-1.5
e-") are less supportive of homology than those ob-
tained for the yeast genes, and, unlike Rad9, extensins
are proline-rich throughout their protein sequences.
Attempts at alignments of Rad9with extensins (not
shown) further indicated that these proteins areproba-
bly not evolutionarily homologous.
DISCUSSION
Correlation of information from light and electron
micrographs of wild-typespreads: Our data indicate
that the condensedand paired figures shown in Figure
1E are representative of fullyaligned chromosomes with
some variation in SC width (Figure 3, F and G, would
both correspond to the spread shownin Figure 1E).
The diffuse stage of chromatin decondensation corre-
sponds to the beginning of dissolution of the SC and
lossofaxial core structure (Figures 1F and 3H; also
compare Figures 2 and 4). A logical interpretation of
these data is that SC elements contributeto the compac-
tion observed at pachytene and that the dissolution of
axial cores leads to the diffusion of chromatin observed
at early diplotene (Figure 1F). While it is clear that
chromosomes donot completely decondense after
1114 L. C . Seitz et al.

1
l(193)'
SWFiPWPR IPSQQSPUTU m
QFT
M M
WQTVQHAQQM LMFPXWAT PALQPVREIA
-AAPPS
+ Z(259)
FVQPlwYsSQ AQPDYTHVQQ

101 ENMEIPPPP YTfyDPBwFl LSAPQVPQNR mHWETNUQ SVRtrarrpnS RSYPYQEPQQ

+ 3(847)
PSMWPASSYQ
STPFAQSVFQ RTTPUSAYPT PPPWISTSS

'201 88m-P PQPPPRQQTV TPPLWPTQA PLPQPAPW PAPVPQATW

PPPSUPPAAQ
U9-D SAS?FMSFLE R
KTT
-H QQTTSAKSIE

301 PPAPAKLPLP
VPVASKPPAR
I(I(TPPPREQK
AVPDSSPDPL SLS8lslULPS STP- EIHSPPPKRI QBYIW(TL1SQ-SRQAp LTpTSRASN&

401
+
SSVPTTASSS TSITSMSSSS r m v a g P T P QRI-
4(115)
PDLQQYGSSE DTPWUUWS SVKRTQERDD d P
5(641)
w VDEIFESEDS LPADLQHEDL

901

1001

1101

1101

1301

1401

1501

1601

1701

1801

1901

2001

1101 LPBVCSWOA
YLIHTLDILF C
- R T F A F V m R TLSRMlLSSV VVIIXPW

FIGURE 10.-Amino acid sequence of the predicted Rad9 protein. Arrows indicate the positions of introns. Numbers above
the amino acid sequence indicate exon number, and numbers in parentheses indicate exon sizes. Asterisks indicate that the
lengths of the first and last exons are for their coding segments only; 5' and 3' UTR regions are not included because their
lengths were not established. Potential target sites for phosphorylation by mitogen-activated protein kinase (amino acid positions
382-385, 390-393, 425-428, and 1502-1505) and a potential nuclear localization signal (positions 607-616) are underlined.

pachytene (Figure IF), anincrease in compaction does
occur between the diffuse stage and metaphase I (Fig-
ure 1H; see also LU and RAJU 1970).
Defects in rad9-1 basidia: The defects in meiosis of
rad9-1 are seen consistently in light and electron micro-
graphs (Figures 5 and 6). With both types of analysis,
the short stretchesof condensation and synapsis against
the diffuse background are the hallmark of the rad9-1
phenotype. Manyof the synapsed stretches of rad9-1
chromatin resemble telomeres, and it is possible that
the ring-like regions of condensation observed at early
time points (Figure 5, B and C) represent telomeric
chromatin. In addition, apparentlyinterstitial stretches
of SC are also observed (Figure 6). However, in neither
case can it be assumed that synapsed regions are homol-
ogous; because the SC can form between nonhomolo-
gous chromosomes (LOIDL et al. 1991) and nonhomolo-
gous regions of chromosomes (MCCLINTOCK 1933;
MACUIRE and RIESS 1994), it is possible that the small
amount of condensed chromatin of rad9-1 cells is syn-
apsed with other, available condensed chromatin re-
gions. Evidence for aberrant synapsis in rad9-1 comes
from our observation (Figure 6, B-D) of regions of
triple pairing. The useof telomeric and interstitial
probes in fluorescence in situ hybridization (LOIDLet
al. 1994;WEINER and KLECKNER 1994) will help to deter-
mine whether the small amount of synapsis observed
for rad9-1 is in fact homology-based or fortuitous.
A striking feature of rad91 nuclei is their failure to
develop during prophase I. We hypothesize that the
defect in rad9-l is manifest very early in meiosis. The
processes ofkaryogamy and nucleolar fusion appear
normal, and the condensation of telomeres may initiate
as for wild-type cells, but further condensationand axial
core development are not possible. This defect may be
a direct consequenceof the absence of functional Rad9
protein in these cells, or the Rad9 protein may be indi-
rectly required for chromatin condensation; for exam-
ple, it may have a role in the resolution of interlocks
that is required for the progression of condensation
and axial core development (HOLMet al. 1981).
It is apparent, however, that meiosis continues, at
least to some extent, in rad9-l mutants. The small
amount of condensation observed givesway,by 8 hr
postkaryogamy, to a more diffuse state (Figures 5G and
6 F ) . Interestingly, about half of all rad91 basidia pro-
gress into the metaphase I stage of the cell cycle, with
shrinkage of the nucleolus and at least some compac-
tion of the chromatin (Figure 5H), although it is not
possible to determine from our data whether homologs
are insome way paired inrad9-l metaphase I. The chro-
matin of the remaining half of rad9-1 basidia remains
 rad9 Gene of Copnnus
difise and graduallybecomes more and more dis-
persed in appearance. These data are consistent with
our previous observations, madeon intact basidia (VAL
ENTINE et al. 1995), which showed that about half of
rad91 basidia arrest in prophase I and about half con-
tinue to metaphase I or anaphase I. Because no rad91
basidia exhibit full pachytene-stage chromosome con-
densation, but fully half showat least some compaction
at metaphase I, we conclude that metaphase I condensa-
tion is not dependent onsuccessful pachyteneconden-
sation. Therefore, these two processes may be distinct,
differently regulated events. In a large survey of reces-
sive sexual-phase mutants in wild strains of N. crussa,
RAJU and LESLIE(1992)describedseveralmutants
which are similar to rad91 in that the initial meiotic
defect is observed in zygotene, but the cell cycle pro-
gesses such that either meiotic arrest or cell death oc-
curs at a later stage.
The specific meiotic phenotype of rad91 basidia is,
to our knowledge, unique. Numerous mutants, charac-
terized infungi, plants, and animals, fail atthe comple-
tion of axial core formation or synapsis (for recent re-
views, see RAJU 1992; HAWLEY et al. 1993; STAIGER and
CANDE 1993; PUKKILA 1994). Phenotypes reported in
fungi range from the almost complete absenceof axial
core development, in severe rad50 strains (ALANI et al.
1990),to the full axialcore development and alignment
withoutsynapsisseenin zip1 disruptions (SYMet al.
1993). However, the rad91 phenotype, of short axial
core elements, essentially all of which are synapsed, a
preponderance of telomere-like regions,and a correla-
tion between condensation and synapsis, is previously
unreported.
The rad9 gene: The rad9 gene contains 26 introns,
which is the largest number reported, to our knowl-
edge, for a fungal gene. The size range of theseintrons,
47-60 bp, is consistent with the small size (usually <lo0
bp) of introns found in other filamentous fungi(GURR
et al. 1987), but is unusually tight, even for C. cinereus
(YEAGER STASSEN et al. 1996). Intron frequency for rad9
(4.0 introns/kb of coding sequence) isless than the
average value (6.3 introns/kb) for other genes of C.
n’nereus (L. BAUNSGAARD, J. VINDand H. DALBOGE,un-
published data, Genbank No. X70789; 0.ISHIBASHI and
K SHISHIDO, unpublished data, Genbank Nos. Dl3295
and D13226; SKRZYNIA et al. 1989; TYMON et al. 1992;
LOGSDON et al. 1995). The intron splice junctions of
rad9 fit those previously reported for filamentous fungi
(GURRet al. 1987). Furthermore, potential internal
branch points have been identified (Figure 9); however,
experimental support is necessary to demonstrate any
functional significance of these regions.
The features of Rad9 necessary for its dual roles in
meiosis and the survival of gamma irradiation are not
yet known. The identification of a putative nuclear lo-
calization signal (Figure 10) suggests that Rad9 could
be a nuclear protein, which would be consistent with a
1115
role in DNA metabolism. The cosmid 1D7, which does
not complement rad91 mutants for either gamma irra-
diation survival or meiosis, is missing 650 bp from the
3’ end of the rad9 transcript (Figure7). Therefore, this
portion of the gene is required for Rad9 function in
both cellular events. We have not yet determined the
mutational site in the rad91 mutant. The behavior of
this mutant, which is strongly defective in meiosis but
only slightly sensitive to gamma irradiation (ZOLANet
al. 1988; VALENTINEet al. 1995), is reminiscent of the
rad50S mutants described by ALANI et al. (1990), in that
they exhibit a “separation of function” phenotype. In
screens of >20,000 mutagenized cultures (ZOLAN et al.
1988;VALENTINE et al. 1995),only one rad9 mutant was
recovered. We are currently screening directlyfor lack
of complementation of the meiotic defect in radP1, as
an alternative strategy for the isolation of additional
rad9 alleles;if the primary type of viablerad9 mutant is
similar in phenotype to rad91, then, because the gene
is so large, this new screen should readily yield new
mutants.
The peak of accumulation of rad9transcript is at -6 hr
after karyogamy (Figure 8 and data not shown), although
defects inthe rad91 mutant are apparent by 1 hr postkar-
yogamy (Figures 5 and 6). The rad9 gene product is
therefore likelytobenecessaryfromearlyinmeiosis
through at least pachytene, the stage of all wild-type basi-
dia at 6 hr postkaryogamy. However, as previously de-
scribed forRADs0, regulated levels oftranscript accumu-
lation do not necessarily correspond tovarying levels of
protein product (RAYMOND and KLECKNER 1993), and
therefore the precise temporal relationship between rad9
expression and Rad9 function can only be determined
by direct analysis of the Rad9 protein.
The Rad9 protein is unrelated in sequence to any
other protein found in previous screens for genes nec-
essary for meiosis or DNA repair. Although homologs
of Rad9 exist inS.mevisiae and S. pombe, the functional
and evolutionary relationships among these gene prod-
ucts and the C. cinereus protein are not yet known. In
our screens for radiation-sensitive,meioticmutants
( Z O W et al. 1988; VALENTINE et al. 1995), we have
deliberately followedthe path set by GAME(1983,1993)
and others (BAKERet al. 1976) in the investigation of
genes with roles in both DNA repair and meiosis. This
laboratory hasso far isolated and sequenced two genes,
rad9 and rad51, the C. cinereus ortholog of the S. cerevis-
iae RAD51 gene (YEAGER STASSEN et al. 1996). The C.
cinereus rad9 gene is not a homolog of any of the genes
denoted by the S. cereuisiae rad mutants, nor was a rad51
mutant found in our genetic screens,and yet homologs
of these genesare present in both organisms. InNeuros-
pura crussa, alleles of the mi-3 gene, an ortholog of
RAD51 (CHENG et al. 1993; H A T A K E Y M et al. 1995),
have been found in screensfor strains exhibiting muta-
gen sensitivity (DELANGE and MISHRA1981).Therefore,
the different biology of ascomycetes and basidiomycetes
1116 L. C . Seitz et al.
could make the recovery of certain classes of mutants
more likely. The analysis of diverse
organisms will hence
provide a broad perspective on the meiotic process, and
the characterization of genes with meiotic function in
multiple organisms will facilitate an understanding of
the genus-specific and fundamental aspects of meiotic
pathways.
We thank Dr. F. RUDOLFTURNER for taking all of the electron
micrographs shown in this paper. We also thank Dr. KEMING SONG
for advice about PCR, Dr. BENJAMIN Lu for help with chromosome
spreads, Ms. KAREN JEPSON-INNES for assistance with figure prepara-
tion, Mr. JON JNmfor helpwith DNA sequencing, Dr. LISAVAILLAN-
COURT and Dr. EUNICEFROELICER for information aboutbasidiomy-
ceteintrons, and Dr. JEFF PALMERandmembers of the ZOLAN
laboratory for helpful discussion and critical reading of the manu-
script. Thiswork was supported by grant GM-43930 from theNational
Institute of General Medical Sciences. W.J.C.was supported by a
fellowship from the Indiana Institute for Molecular and Cellular Biol-
ogy and by NIGMS Training Grant GM-07227.
LITERATURE CITED
A L A NE.,
I , R. PADMORE and N. KLECKNER,1990 Analysis of wild-type
and rad50 mutants of yeast suggests an intimate relationship
between meiotic chromosome synapsis and recombination. Cell
61: 419-436.
ALTSCHUL, S. F., W. GISH,W. MILLER,E. W. MYERSand D. J. LIPMAN,
1990 Basic local alignmentsearchtool. J. Mol. Biol. 215:
403-410.
BAKER, B.S., A. T. C. CARPENTER,M. S. ESPOSITO, R. E. ESPOSITOand
L. SANDLER, 1976 The genetic control of meiosis. Annu. Rev.
Genet. 1 0 53-134.
BERNSTEIN, C., and H. BERNSTEIN, 1991 Aging, Sex, and DNA Repair.
Academic Press, Inc., San Diego.
BINNINGER, D. M., C. SKRZWIA, P. J. PUKIULA and L.A. CASSELTON,
1987 DNA-mediated transformation of the basidiomycete Copn-
nus cinereus. EMBO J. 6 835-840.
CARPENTER, A. T. C., 1987 Gene conversion,recombinationnod-
ules, and theinitiation of meiotic synapsis. Bioessays 6 232-236.
CHEN,J.-H., and R.-J. LIN, 1990 The yeast PRPP protein, a putative
RNAdependent ATPase, shares extensive sequence homology
with two other pre-mRNA splicing factors. Nucleic Acids Res. 1 8
6447-6447.
CHENG,R., T. I. BAKER,C. E.CORDS and R. J. RADLOFF, 1993 mei-
3, a recombination and repair geneof Neurospura crassa, encodes
a RecA-like protein. Mutat. Res. 294: 223-234.
DELANGE, A.M., and N. C. MISHRA,1981 The isolation of MMS
and histidine-sensitive mutants in Neurospora crassa. Genetics 97:
247-259.
DIETRICH,A. J., E. KOK, H. H. OFFENBERG, C. HEWING,P. DE BOER
et al., 1992 The sequential appearance of components of the
synaptonemal complex during meiosis of the female rat. Ge-
nome 3 5 492-497.
DOBSON,M. J., R. E. PEARLMAN, A. KARAISKAKIS,B. SPYROPO~LOS and
P. B. MOENS,1994 Synaptonemalcomplexproteins:occur-
rence, epitope mapping and chromosome disjunction. J. Cell
Sci. 107: 2749-2760.
DOOLITTLE,R. F., 1986 Of lJr/.s an,d W s . University Science Books,
Mill Valley, CA.
DRESSER, M. E., and G. N. GIROUX,1988 Meiotic chromosome behav-
ior in spread preparations of yeast. J. Cell Biol. 106 567-573.
EDELMANN, S. E., and C. STABEN,1994 A statistical analysis of se-
quence features within genes fromNeurospora crassa. Exp. Mycol.
1 8 70-81.
ENGEBRECHT, J., J. HIRSCHand G. S. ROEDER,1990 Meiotic gene
conversion and crossing over: their relationship to each other
and to chromosomesynapsis and segregation. Cell 6 2 927-937.
GAME,J. C., 1983 Radiation-sensitive mutants and repair in yeast,
pp. 109-137 in YeastGenetics: Fundamental and Applied Aspects,
edited by J. F. T. SPENCER,D. M. SPENCER and A.R. W. SMITH.
Springer-Verlag, New York.
GAME, J. C., 1993 DNA double-strand breaks and the RAD5GRAD57
genes in Saccharomyces. Semin. Cancer Biol. 4: 73-83.
GARCIA-BUSTOS, J., J. HEITMANand M.N. HALL,1991 Nuclear pro-
tein localization. Biochim. Biophys. Acta 1071: 83-101.
GURR,S. J., S. E. UNKLESand J. R. KINGHORN, 1987 The structure
and organization of nuclear genesof filamentous fungi, pp.93-
139 in Gene Structure in Eukaryotic Mimobes, edited by J. R. KINGH-
ORN.I R L Press, London.
HAWLEY, R. S., K. S. MCKIMand T. ARBEL, 1993 Meiotic segregation
in Drosophila mlanog&er females: molecules, mechanisms, and
myths. Annu. Rev. Genet. 27: 281-317.
HATAKEYM, S., C. ISHII andH. INOUE,1995 Identification and
expression of Neurospora mama m&3 gene which encodes a pro-
tein homologous to the Rad51 of Saccharomyces mmisiae. Mol.
Gen. Genet. 249 439-446.
HOLM,P. B., S. W. RASMUSSEN, D. ZICKLER, B. C. Lu and J. SAGE,
1981 Chromosome pairing, recombination nodules and chi-
asma formation in the basidiomycete Copnnus cinereus. Carlsberg
Res. Commun. 46: 305-346.
INOUE,S. B., H. SAKAMOTO, H. SAWA and Y. SHIMURA, 1992 Nucleo-
tide sequence of a fission yeast gene encoding the DEAH-box
RNA helicase. Nucleic Acids Res. 20: 5841-5841.
KIELISZEWSKI, M. J., and D. T. LAMPORT, 1994 Extensin: repetitive
motifs, functional sites, post-translational codes, and phylogeny.
Plant J. 5 157-172.
KLECKNER, N., R. PADMORE and D. K. BISHOP,1991 Meiotic chromo-
some metabolism: one view. Cold Spring Harbor Symp. Quant.
Biol. 5 6 729-743.
KWE, J., and R. F. DOOLITTLE,1982 A simplemethodfor dis-
playing the hydropathic character of a protein. J. Mol. Biol.
157: 105-132.
LINDER,P., P. F. LASKO, M. ASHBURNER, P. LEROY,P.J. NIELSENet al.,
1989 Birth of the D-E-A-D box. Nature 337: 121-122.
LOGSDON, J. M. J., M. G. TISHENKO, C. DIXON,J. D.JAFARI, V. K.
WALKERet al., 1995 Seven newly discovered intron positions in
the triose-phosphate isomerase gene: evidence for the introns-
late theory. Proc. Natl. Acad. Sci. USA 9 2 8507-8511
LOIDL,J., K. NAIRZand F. KLEIN,1991 Meiotic chromosome synapsis
in a haploid yeast. Chromosoma 100: 221-228.
LOIDL,J., F. KLEIN and H. SCHERTHAN, 1994 Homologous pairing
is reduced but not abolished in asynaptic mutants of yeast. J.
Cell Biol. 125: 1191-1200.
Lu, B. C., and N. B. RAJU,1970 Meiosis in CopnnzrJII.Chromosome
pairing and the lampbrush diplotenestage of meiotic prophase.
Chromosoma 29: 305-316.
MACUIRE, M. P., and R. W. RIESS, 1994 The relationship of homolo-
gous synapsis and crossing over in a maize inversion. Genetics
137: 281-288.
ANSO SOUR, S. J., K. A. RESING,J. M. CANDI, A. S. HERMANN, J. W.
GLOOR et al., 1994 Mitogen-activated protein ( M A P ) kinase
phosphorylation of MAP kinase kinase: determination of phos-
phorylation sites by mass spectrometry and sitedirected muta-
genesis. J. Biochem. 116: 304-314.
MCCLINTOCK, B., 1933 The association of non-homologous parts of
chromosomes in the mid-prophase of meiosis in zea mays, pp.
191 -237
in zeitschrifi fur ZeUfmchungund Mikroskopische Anatomie,
edited by R.GOLDSCHMiDT and w. VON MOI~IENDROFF. Verbg
von Julius Springer, Berlin.
MOSES,M. J., 1968 Synaptinemalcomplex. Annu. Rev. Genet. 2:
363-412.
PUKKILA,P. J., 1994 Analysis of meiosis in filamentous fungi, pp.
267-281 in The Mycota I Growth, Dtfferentiation and Sexuality, ed-
ited by J. G. H. WESSELSand F. MEINHARDT. Springer-Verlag,
New York.
PUKKILA,P. J., and L. A. CASSELTON,1991 Molecular genetics of the
agaric Copinus cinereus, pp. 126-150 in More Gene Manipulations
in Fungi, edited by J. W. BENNETT and L. L. LASURE. Academic
Press, San Diego.
PUKKILA,P. J., and B. C. Lu, 1985 Silver staining of meiotic chro-
mosomes in thefungus, Coprinus cinereus. Chromosoma 91:
108-112.
PUKKILA,P. J., B.M. YASHAR and D. M. BINNINGER, 1984 Analysis of
meiotic development inCOprnus cinereus,pp. 177-194 in Control-
1ingEuents in Meiosis, edited by C. W. EVANSand H. G. DICKINSON.
Society for Experimental Biology, Cambridge.
PUKKILA,P. J., C. SKRZYNIA and B. C. Lu, 1992 The rad3-1 mutant
 rad9 Gene of Copnnus 1117
is defective in axial core assembly and homologous chromosome W O N , A. M., U. KUES,W. V. RICHARDSON and L.A. CASSELTON,
pairing during meiosis in the basidiomycete Coptinus cinereus. 1992 Afungalmating type protein that regulates sexual and
Dev. Genet. IS: 403-410. asexual development contains a POU- related domain. EMBO J.
RAJu, N. B., 1992 Genetic control of the sexual cycle in Neurospwa. 11: 1805-1813.
Mycol. Res. 96: 241-262. VALENTINE, G., Y. J. WALLACE, F. R. TURNER and M. E. ZOLAN, 1995
RAJu, N. B., and B. C. Lu, 1970 Meiosis in Copnnus. 111. Timing of Pathway analysis of radiationsensitive meiotic mutants of Copn-
meiotic eventsin C. lagopus (sensu Buller). Can. J. Bot. 4 8 2183- nus cinereus. Mol. Gen. Genet. 247: 169-179.
2 186. VON WEITSTEIN,D., S. W. RASMUSSEN and P. B. HOLM,1984 The
RAMESH, M. A., and M. E. ZOLW, 1995 Chromosome dynamics in synaptonemalcomplex in geneticsegregation.Annu. Rev.
rad12 mutants of Coprinus c i w e u s . Chromosoma 104: 189-202. Genet. 18: 331-413.
-BAND, W. S., and D. S. BRIGHT,1995 NIH image: a public domain WEINER,B. M., and N. KLECKNER, 1994 Chromosome pairing via
image processing program for the Macintosh. Microbeam Anal. multiple interstitial interactions before and during meiosis in
SOC.J. 4 137-149. yeast. Cell 77: 977-991.
RAYMOND, W. E., and N. KLECKNER, 1993 Expression of the Saccham WILSON,E. B., 1925 The (211 in Deuelopmat and Heredity. The MacMil-
myces cmeuisiaelUD5Ogeneduring meiosis: steady-state transcript Ian Company, New York.
levels rise and fall while steady-state protein levels remain con- Wu, M.M. J., J. R. CASSIDY and P. J. PUKKILA, 1983 Polymorphisms
stant. Mol. Gen. Genet. 238: 390-400. in DNA of Copnnus cinereus. Curr. Genet. 7: 385-392.
ROEDER,G. S., 1990 Chromosome synapsis and genetic recombina- ~ G E STASSEN,
R N., J. LOCSDON,H. OFFENBERG, G. J. Vow, J. D.
tion: their roles in meiotic chromosome segregation. Trends PALMER et al., 1996 Isolation and analysis of RAD51 orthologs
Genet. 6: 385-389. from Copinus cinereus a n d L y c e i c o n esculentum, and phyloge-
SANCER, F., S. NICKLEN and A.R. COULSON,1977 DNA sequencing netic analysis of eukaryotic r e d homologs. Curr.Genet.(in
with chain- terminating inhibitors. Proc. Natl. Acad. Sci. USA 74: press).
5463-5467. ZICKLER, D.,1977 Development of the synaptonemal complex and
the “recombination nodules” during meiotic prophase in the
SHARP, P. A,, 1987 Splicing of messenger RNA precursors. Science
seven bivalents of the fungus Swdana macrospwa Auersw. Chro-
235: 766-771.
mosoma 61: 289-316.
SKRZWIA, C., D. M. BINNINGER, J. A. I. ALSPAUCH and P. J. PUKKILA, ZOLAN,M. E., and P. J. PUKKILA, 1986 Inheritance of DNA methyla-
1989 Molecular characterizationof T W l , a genecodingfor tion in Copnnus cinereus. Mol. Cell. Biol. 6: 195-200.
tryptophan synthetase inthe basidiomycete Coprinus cinereus. ZOLAN,M. E., C. J. TREMELand P. J. PUKKILA, 1988 Production
Gene 81: 73-82. and characterization of radiation-sensitive meiotic mutants of
STAIGER,C. J., and W. Z. CANDE, 1993 Cytoskeletal analysis of maize Copinus cinereus. Genetics 120: 379-387.
meiotic mutants, pp. 157-171 in Molecular and Cell Biology of the ZOLW, M. E., J. R. CRI~TENDEN, N. K HEXERand L. C. SErrz, 1992
Plant CellCycle, edited by J. ORMRODand D. FRANCIS. Kluwer Efficient isolation and mappingof rad genes of the fungus Copn-
Academic Publishers, The Netherlands. nus cinereus using chromosome-specific libraries. Nucleic Acids
SYM,M.,J. A. ENCEBRECHT and G. S. ROEDER,1993 ZIP1 isa synapto- Res. 20: 3993-3999.
nemal complex protein required formeiotic chromosome synap
sis. Cell 7 2 365-78. Communicating editor: C. BRONSON