Download Duplicative Transfer of a MADS Box Gene to a Plant Y Chromosome

Duplicative Transfer of a MADS Box Gene to a Plant Y Chromosome
Sachihiro Matsunaga,*1 Erika Isono, Eduard Kejnovsky,à Boris Vyskot,à
Jaroslav Dolezel,§ Shigeyuki Kawano,* and Deborah Charlesworthk
*Department of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Chiba, Japan; Department of
Biological Sciences, Graduate School of Science, University of Tokyo, Tokyo, Japan; àInstitute of Biophysics, Academy of Sciences
of the Czech Republic, Brno, Czech Republic; §Institute of Experimental Botany, Academy of Sciences of the Czech Republic,
Olomouc, Czech Republic; and kInstitute of Cell, Animal and Population Biology, University of Edinburgh, Edinburgh, United
Kingdom
Y chromosomes carry genes with functions in male reproduction and often have few other loci. Their evolution and the
causes of genetic degeneration are of great interest. In addition to genetic degeneration, the acquisition of autosomal
genes may be important in Y chromosome evolution. We here report that the dioecious plant Silene latifolia harbors
a complete MADS box gene, SlAP3Y, duplicated onto the Y chromosome. This gene has no X-linked homologs but only
an autosomal paralog, SlAP3A, and sequence divergence suggests that the duplication is a quite old event that occurred
soon after the evolution of the sex chromosomes. Evolutionary sequence analyses using homologs of closely related
species, including hermaphroditic Silene conica and dioecious Silene dioica and Silene diclinis, suggest that both SlAP3A
and SlAP3Y genes encode functional proteins. Indeed, quantitative RT-PCR and in situ hybridization analyses showed
that SlAP3A is expressed specifically in developing petals, but SlAP3Y is much more strongly expressed in developing
stamens. The S. conica homolog, ScAP3A, is expressed in developing petals, suggesting subfunctionalization with
evolution of male-specific functions, possibly due to evolutionary change in regulatory elements. Our results suggest that
the acquisition of autosomal genes is an important event in the evolution of plant Y chromosomes.
Introduction
Studies of sex chromosomes have concentrated
mainly on animal species, since most animals are
dioecious. Although most flowering plant species are
hermaphroditic, with bisexual flowers that contain both
male and female reproductive organs (stamens and pistils),
about 6% of angiosperms are dioecious (separate individuals produce staminate and pistillate flowers; see Renner
and Ricklefs 1995), and some have heteromorphic sex
chromosomes. The Y chromosome of Silene latifolia, in
the family Caryophyllaceae, has been extensively analyzed (Matsunaga and Kawano 2001; Negrutiu et al.
2001). There are striking similarities with animal sex
chromosomes, as well as some differences. Classical
cytogenetic analyses of deletion mutants suggested that the
Y chromosome has three functions: (1) suppression of
pistil development, (2) initiation of stamen development,
and (3) completion of stamen development (Westergaard
1958). Recent studies of Y chromosome deletions induced
by c-rays or X-rays support this general conclusion
(Negrutiu et al. 2001; Lebel-Hardenack et al. 2002).
Compared with human or fruitfly Y chromosomes, the S.
latifolia Y contains little heterochromatin (Matsunaga et
al. 1999; Charlesworth and Charlesworth 2000).
It is currently unclear how many X-linked genes have
Y chromosome homologs in plants, but two apparently
functional Y-linked genes, SlY1 and SlY4, have so far been
discovered in S. latifolia (Delichere et al. 1999; Atanassov
et al. 2001). SlY1 encodes a WD-repeat protein and is
preferentially expressed in the stamens of male plants,
1
Present address: Department of Biotechnology, Graduate School of
Engineering, Osaka University, Osaka, Japan.
Key words: dioecious plant, Y chromosome, sex differentiation,
MADS box gene, subfunctionalization, duplication.
E-mail: [email protected].
Mol. Biol. Evol. 20(7):1062–1069. 2003
DOI: 10.1093/molbev/msg114
Molecular Biology and Evolution, Vol. 20, No. 7,
Ó Society for Molecular Biology and Evolution 2003; all rights reserved.
1062
whereas SlY4 encodes a fructose-2, 6-bisphosphatase. Both
of them have X-linked homologs. Except for the maledetermining gene SRY, genes on the human Y chromosome fall into two groups (Lahn and Page 1997). The first
group consists of housekeeping genes, which have Xlinked homologs. Genes in the second group are expressed
exclusively in testes and form gene families on the Y
chromosome. The previously reported S. latifolia Y-linked
genes are comparable to the human Y-linked housekeeping genes. We here report the discovery of a Y-linked
MADS-box gene with no X-linked counterpart.
Extant dioecious species of Silene include a group of
close relatives, S. latifolia, Silene diclinis, and Silene
dioica. A phylogenetic tree based on internal transcribed
spacer data for nuclear rRNA genes of 22 Silene suggested
that dioecy in the genus evolved from gynodioecious
ancestors (Desfeux et al. 1996). This is consistent with the
fact that, whereas the majority of species in the 80 genera
in the family Caryophyllaceae are hermaphroditic, many
Silene species are gynodioecious and must carry male
sterility factors (Defeux et al. 1996). The hermaphroditic
and gynodioecious species, S. conica, and S. vulgaris,
which are related to the dioecious species, do not have
heteromorphic chromosomes. Chromosome heteromorphism therefore reflects de novo evolution of sex
chromosomes during the evolution of dioecy in this plant
lineage, a relatively recent event within this genus. This
gives an opportunity to study processes that occur in
young sex chromosomes that are still in earlier stages of
their evolution than those of humans or Drosophila.
The evolution of sex chromosomes is believed to
involve at least two types of processes. The complete are
almost complete genetic degeneration is a dramatic effect
that is well known (Lahn and Page 1997; Charlesworth
and Charlesworth 2000). Another hypothesized process is
the addition to the Y chromosome of genes that are
advantageous in males but disadvantageous in females
(Charlesworth and Charlesworth 1980; Rice 1997).
A MADS Box Gene on a Plant Y Chromosome 1063
Because Y chromosomes function in the development of
male reproductive organs, new genes may be added to the
Y chromosome and may be able to persist despite the
forces tending to lead to genetic degeneration. An example
of duplicative transfer of autosomal genes to the human Y
chromosome is the DAZ gene, which functions in
spermatogenesis (Saxena et al. 1996). DAZ reached the
Y chromosome by interchromosomal transposition of an
autosomal progenitor (Saxena et al. 1996).
We have discovered that a MADS box gene has been
duplicated onto the Y chromosome of S. latifolia. The
MADS box genes encode transcription factors that share
a common DNA-binding domain and represent crucial
regulatory genes in plant development, perhaps comparable in importance to the HOX homeobox transcription
factor genes in animal development (Ng and Yanofsky
2001; Meyerowitz 2002). The Y-linked MADS box gene
evolved from an ancestral autosomal gene through
duplication accompanied by increased expression in male
reproductive organs. Our finding suggests that duplication
of genes onto the Y chromosome may be important in
plant, as well as mammalian, Y chromosome evolution,
consistent with the theory mentioned above.
Materials and Methods
Plant Material
We used an inbred Silene latifolia line, K1, for the
molecular experiments. Seeds of Silene conica, S. diclinis,
S. dioica, and other strains of S. latifolia were obtained
from the seed banks of the Royal Botanical Gardens (Kew,
England), from the botanical gardens at Ulm University
(Germany), Regensburg University (Germany), Graz
University (Austria), and Portugal University (Portugal),
and from the Lake Kawaguchi herb garden (Yamanashi,
Japan). Plants were grown in a temperature-controlled
chamber at 228C. Stages of flower buds, including young,
early mature, intermediate, and late mature stages,
correspond to previously classified stages B1, B2, B3,
and B4 plus B5 (Matsunaga et al. 1996).
Cloning of SlAP3 and Orthologs
We designed two sets of degenerate PCR primers,
including 59-CGGAATTCATGAARMGIATIGAIAA-39
and 59-CGGGATCCITCIARYTGICBYTCIA-39 or 59CGGGATCCYTCIGCRTCRCAIAGIAC-39 based on the
highly conserved MADS box domain and K domain. The
symbols I, R, Y, and B denote inosine, purines (A and G),
pyrimidines (C and T), and mixtures without A (C, G, and
T), respectively. Hemi-nested PCR amplification was
performed using young flower-bud cDNA (five cycles:
948C for 25 s, 378C for 2 min, and 728C for 1 min and 30
cycles: 948C for 20 s, 558C for 1 min, and 728C for 1 min).
Amplified fragments were subcloned using a TOPO TA
Cloning kit (Invitrogen). The nucleotide sequences were
determined using an ABI 3100 genetic analyzer and
a BigDye terminator cycle sequencing kit (Applied
Biosystems). We obtained full-length cDNAs from
screening of our constructed Tripl Ex2 cDNA library
(Clontech) or RACE-PCR using SMART RACE cDNA
amplification kit (Clontech).
Genomic Distribution
Genomic DNA for southern hybridization was
isolated from young leaves using an automatic DNA
isolation system PI50 (Kurabo). Preparation of membranes
was performed as described previously. Probe preparation,
hybridization, and detection were performed using the
AlkPhos direct labeling and detection system with CDPstar (Amersham). PCR analyses using genomic DNA and
flow-sorted chromosomes were done as described previously (Kejnovsky et al. 2001). The oligonucleotide
primers used for the K domains of SlAP3 were as follows:
59-GTACGATGAGTACCAGAAGA-39 and 59-GATCCATGAGGAGATCTCCA-39. Genomic clones were isolated from a genomic phage library derived from male
leaves as described previously (Uchida et al. 2002).
Expression Analyses
Total RNA was isolated from nitrogen-frozen organs
using Trizol (Lifetech). Northern hybridization was
performed using the Gene Images random-prime labeling
and detection system (Amersham). Total RNA was reverse
transcribed into cDNA using a first-strand cDNA synthesis
kit (Amersham). RT-PCR was performed using Ready-toGo RT-PCR beads (Amersham) with cDNA. Quantitative
RT-PCR was performed using two different systems, the
LightCycler (Roche) with LightCycler-DNA master
SYBR green I kit (Roche) and the Smart Cycler (Takara)
with QuantiTect SYBR green kit (Qiagen). The gene for
the GTPase beta subunit (SlGb), which is expressed
constitutively in all organs (Matsunaga, unpublished data),
was used as an internal standard to estimate the relative
expression of mRNA. The relative expression of mRNA
for a given tested gene was defined as the mean value that
was divided by the mean for SlGb, with the same cDNA
used as template. Relative expression values and corresponding standard deviations for the transcripts were calculated from four to six experimental replicates with each
of the two real-time PCR systems. The oligonucleotide
primer sets used for quantitative RT-PCR were as follows:
59-GACATGGTGACAGCCATAGCAACA-39 and 59TCACGAGAAGCAGAGACTATCTGT-39 for SlGb;
59-GGCATGGAGATCTCCTCATGGATC-39 and 59ATACTGGAGATAACACAGCCTAGT-39 for SlAP3A;
59-GGCATGGAGATCTCCTCATGGATC-39 and 59TATATTCGAGACAACATGGCCTGG-39 for SlAP3Y;
and 59-GGCATGGAGATCTCCTCATGGATC-39 and 59ATATTCGAGACAACATGGCCTAGT-39 for ScAP3A.
Using these primers, only a single fragment was observed
in agarose gel electrophoresis after the RT-PCR. The band
lengths for SlGb, SlAP3A, SlAP3Y, and ScAP3A, were 330
bp, 313 bp, 310 bp, and 312 bp, respectively. In situ
hybridization was performed as described previously
(Matsunaga et al. 1996), with an automatic ISH robot
AIH-101B (Aloka) using tyramide amplification in the
GenPoint system (Dako).
1064 Matsunaga et al.
Sequence Diversity and Divergence Analyses
The coding sequences were aligned by eye, with
adjustments using BioEdit software (http://www.mbio.
ncsu.edu/BioEdit/bioedit.html), and analyses were performed on the aligned sequences using DNAsp version
3.95 (Rozas and Rozas 1999). The average numbers of
pairwise differences, ps and pa, for synonymous and nonsynonymous sites, and their standard deviations were
estimated for the samples of 12 S. latifolia SlAP3A and
SlAP3Y sequences, each from a different population of
origin. The mean pairwise divergence between these sets, in
the core and noncore regions (Ng and Yanofsky 2001), and
between each of them and the S. conica ScAP3A sequence at
silent and replacement sites (Ks and Ka), were also
estimated, as were the same quantities for the smaller
samples of the two other dioecious species (S. dioica, n ¼ 3;
and S. diclinis, n ¼ 1). The same software was used for the
McDonald-Kreitman tests (McDonald and Kreitman 1991)
and Tajima’s D statistic (Tajima 1989). To estimate changes
in the separate Y-linked and autosomal lineages since the
duplication, the ‘‘preferred and unpreferred synonymous
substitutions’’ analysis of DNAsp version 3.95 was used
with S. conica as the outgroup, assuming parsimony.
Tajima’s relative rate test (1993) was done by using
parsimony to infer sites fixed within the Y chromosome
lineage before the divergence of the three dioecious
species, sites fixed in the autosomal lineage, and sites that
are unique to the single S. conica sequence. A NeighborJoining tree based on the sequence data from all the
species was made using MEGA2 (Kumar et al. 2000).
Results and Discussion
Identification of a Y-linked MADS Box Gene with an
Autosomal Paralog
Degenerate primers whose sequences were designed
from the conserved MADS and K domains were used in
RT-PCR utilizing RNA from young S. latifolia male
flower buds. Sequencing revealed seven different products.
Three sequences were identified as known autosomal
MADS box SLM genes (Hardenack et al. 1994; Guttman
and Charlesworth 1998). Another sequence showed
significant similarity to APETALA3, which in Arabidopsis
thaliana functions in defining the floral identity of petals
and stamens (Jack et al. 1992). This sequence differs from
a previously reported S. latifolia AP3 homolog, SLM3,
which functions in the identity of petals and stamens in
floral meristems (Hardenack et al. 1994). We therefore
named it SlAP3 (S. latifolia AP3). SlAP3 has 78% amino
acid identity with an AP3 homolog, CMB2, of Dianthus
caryophyllus in the tribe Caryophyllidae (Baudinette et al.
2000). Higher eudicots have two AP3 gene lineages
(Kramer, Dorit, and Irish 1998). SLM3 is in the eu-AP3
lineage, whereas SlAP3 and CMB2 fall among genes in the
TM6 lineage, most of which are expressed in developing
floral organs, rather than in floral meristems. For example,
TM6 itself is highly expressed in developing petals,
stamens, and carpels but does not play a role in the
identity of petals and stamens (Pnueli et al. 1991).
Genomic southern hybridization of SlAP3 with
FIG. 1.—Evidence for the chromosomal locations of the two SlAP3
paralogs. (A) The upper panel shows the genomic southern hybridization
analysis of SlAP3. Genomic DNA was digested with HindIII. The SlAP3A
genomic clone includes an internal HindIII restriction site, whereas
SlAP3Y has no HindIII site. The figure shows the results from two parents
(#1 and $1: male and female parents) and some male and female
offspring (# and $: 2 to 5), hybridized with the K domain of SlAP3.
Based on the presence of a HindIII restriction site within the sequences,
the expected autosomal pattern has two hybridizing fragments, whereas
a single fragment represents the sequence present in the male parent and
male progeny. The middle and lower panels show the electrophoretic
analysis of PCR products obtained with the SlAP3A-specific and SlAP3Yspecific primers, respectively. (B) PCR analyses on flow-sorted
chromosomes. The PCR products were amplified with the primers for
the K domain of SlAP3-specific, SlAP3A-specific, or MROS3-specific
primers using female ($) and male (#) genomic DNA, no chromosomes
(0), flow-sorted autosomes (A), or X chromosomes (X) as templates. The
primers for the K domain and SlAP3A amplify 1,094-bp and 411-bp
fragments, respectively. The MROS3-specific primers amplify 433-bp
fragments from both autosomal and X-linked paralogs (Kejnovsky et al.
2001).
genomic DNA of a pair of male and female parents and
four progeny of each sex showed that there are at least two
paralogous sequences, one or more apparently autosomal
or X-linked, and one present only in males and therefore
probably Y-linked (fig. 1A). To investigate these loci more
fully, we obtained and sequenced full-length homologous
cDNAs from a male flower bud library. Twenty-two
cDNAs isolated from 2 3 105 recombinants fell into only
two different sequence classes. The putative coding and 39
noncoding sequences of these cDNAs are 92.9% and
85.0% identical, respectively, so we could design PCR
primers specific for the two loci. PCR with these primers
detected one type of sequence (denoted here by SlAP3A) in
both males and females, and a paralog (SlAP3Y) present
only in males (fig. 1A, lower part), confirming the Ylinkage of SlAP3Y. To determine the chromosomal
location(s) of SlAP3A sequences, we performed PCR with
flow-sorted X chromosomes and autosomes using primers
for the conserved K domains of SlAP3A and SlAP3Y or
specific primers for SlAP3A (Kejnovsky et al. 2001),
which showed clearly that SlAP3A has no X-linked
homolog (fig. 1B). Moreover, genomic clones of both
SlAP3A and SlAP3Y have five exons and four introns. The
SlAP3A genomic clone includes an internal HindIII
A MADS Box Gene on a Plant Y Chromosome 1065
Table 1
Comparison of Sequence Divergence Between Paralogs
and Orthologs of SlAP3
Ka
Ks
Ka/Ks
Autosomal versus Y-linked
S. latifolia
S. dioica
S. diclinis
0.053
0.045
0.046
0.130
0.107
0.119
0.411
0.423
0.390
Autosomal versus S. conica
S. latifolia
S. dioica
S. diclinis
0.052
0.048
0.048
0.164
0.146
0.148
0.314
0.327
0.325
Y-linked versus S. conica
S. latifolia
S. dioica
S. diclinis
0.043
0.038
0.040
0.176
0.171
0.187
0.246
0.224
0.217
restriction site, whereas SlAP3Y has no HindIII site
(Matsunaga, unpublished data). Our results thus indicate
that SlAP3 has only two paralogs, SlAP3A and SlAP3Y, in
the male genome.
SlAP3A and SlAP3Y Are Functional Duplicates,
Evolving Under Selective Constraint
It is very interesting to find a Y-linked locus with no
X-linked homolog. One possibility is that this is simply
a nonfunctional duplication. This seems unlikely since
both paralogs, including SlAP3Y, are present in cDNA, but
we have several further pieces of evidence. First, we
investigated SlAP3 evolution in the genus and estimated
selective constraint on the protein encoded by the
duplicates by measuring Ka and Ks (the estimated mean
numbers of nonsynonymous substitutions per nonsynonymous site and synonymous substitutions per synonymous
site) between different sequences. Loss of function of
a gene is indicated by ratios of Ka/Ks greater than 1 (neutral
evolution), and values greater than 0.5 might thus be
observed if duplicate genes are compared and one of them
has lost its function. These analyses used SlAP3 orthologs
from the related hermaphrodite species S. conica and
included two dioecious species that are close relatives of S.
latifolia. 39-RACE and 59-RACE yielded two sets of
orthologs from male flower buds, SdAP3A and SdAP3Y
from S. dioica and SiAP3A and SiAP3Y from S. diclinis.
PCR with specific primers for SdAP3Y and SiAP3Y
showed male-specific amplification; thus both these
species also have Y-linked SlAP3Y homologs. From the
bisexual flower buds of S. conica, only one ortholog,
ScAP3A, could be isolated. PCR with primers for ScAP3A,
and genomic southern hybridization analysis (data not
shown) indicated that the SlAP3 ortholog is a single-copy
gene in the S. conica genome. In all comparisons involving
sufficient divergence, the Ka/Ks ratios of SlAP3A and also
SlAP3Y are considerably less than 1 (table 1), suggesting
that both SlAP3Y and SlAP3A evolved under selective
constraint, that is, that SlAP3Y has retained functionality
for most of its evolutionary history (although this test
cannot exclude recent loss of function). In support of this
conclusion, the 59 half of the sequence, which encodes the
FIG. 2.—Expression analyses of SlAP3A and SlAP3Y. (A) The first
panel shows northern hybridization analysis of SlAP3. Each lane
contained 20 lg of total RNA from male and female organs hybridized
with SlAP3. Panels 2 to 4 show RT-PCR analyses using specific primers
for SlAP3A, SlAP3Y, and SlGb, respectively. PCR products were
amplified using 1 lg of the RNA that was used in the northern blot.
(B) In situ hybridization of SlAP3A (left) and SlAP3Y (right) in
longitudinal sections of young female and male flower buds, respectively.
The hybridization signals appear reddish brown. G ¼ suppressed
gynoecium; O ¼ ovary; P ¼ petal; Se ¼ sepal; St ¼ suppressed stamen;
T ¼ stamen; Y ¼ style.
‘‘core region,’’ including the MADS box (Ng and
Yanofsky 2001), has particularly low Ka/Ks ratios in both
paralogs (using divergence from the S. conica sequence,
the values for SlAP3A and SlAP3Y are, respectively, 0.125
and 0.086 for the core versus 0.415 and 0.435 for the
remainder of the coding sequence; using the divergence
between SlAP3A and SlAP3Y, the two regions’ values are
0.276 and 0.487). This differs from the findings in
Hawaiian silverswords, where an AP3 gene is present
as duplicate copies because the species are tetraploid; in
these (unlike their diploid North American relatives) mean
Ka/Ks ratios were close to 0.5 (Barrier, Robichaux, and
Purugganan 2001).
SlAP3A and SlAP3Y Are Expressed Differently During
Flower Development
Northern hybridization and RT-PCR analyses using
reproductive and vegetative organs of males and females
1066 Matsunaga et al.
FIG. 3.—Quantitative analyses of the SlAP3A, SlAP3Y, and ScAP3A transcripts. Bars represent normalized relative expression values. The blue,
pink, and green bars represent the relative expression levels in male, female, and bisexual flower buds, respectively. The data are shown as mean 6
SEM, where n ¼ 4 to 6. (A) Temporal gene-expression patterns during the development of flower buds. (B) Spatial gene-expression patterns in
developing floral organs. (C) Quantitative comparisons of gene expression between the petals of male, female, and bisexual flower buds. ‘‘Male’’
designates expression in petals of SlAP3A plus SlAP3Y, and ‘‘female’’ and ‘‘bisexual’’ represent the expression levels in petals of SlAP3A and ScAP3A,
respectively.
showed that both SlAP3A and SlAP3Y are expressed
specifically in reproductive organs, including flower buds
and flowers (fig. 2A). To determine which regions of the
floral organs contained these transcripts, we performed in
situ hybridization analysis using as probes the locusspecific 39 noncoding sequences mentioned above. No
transcripts of SlAP3A and SlAP3Y were detected in the
second or third whorls of floral meristems of either sex
until the stamen primordia emerged. Moreover, levels of
both transcripts were low in very young petals and
stamen primordia. In contrast, large amounts of both gene
transcripts were detected in developing floral organs (fig.
2B). These findings are consistent with data for most
other TM6 lineage genes, which are reported to be
expressed in developing floral organs, rather than floral
meristems (Kramer, Dorit, and Irish 1998). In female
flower buds, low SlAP3A transcription levels were
detected in the ovary walls but were significantly
elevated in developing petals, style primordia, and ovules
(fig. 2B). The growth of stamen primordia is arrested in
female flowers and the tissues degenerate before the
flowers open (Grant, Hunkirchen, and Saedler 1994). No
SlAP3A transcription was detected in arrested stamens
(fig. 2B). In male flower buds, the carpel primordium is
suppressed and later becomes a rudimentary gynoecium
(Grant, Hunkirchen, and Saedler 1994), in which SlAP3Y
transcription was detected at low levels (fig. 2B). The
SlAP3Y transcript accumulated at high levels in stamens
and petals.
Although SlAP3Y was absent from the female flower
buds, a hybridization signal above background was
detected in female buds, probably due to cross-hybridization between the SlAP3A and SlAP3Y probes. To more
clearly distinguish the SlAP3A and SlAP3Y expression
patterns, we therefore performed quantitative PCR with
locus-specific primers based on the 39 noncoding region
sequences (figs. 3A and B). Relative expression values
and corresponding standard deviations for the transcripts
were calculated from at least four experimental replicates
with two different real-time PCR systems. SlAP3A and
SlAP3Y showed different developmental expression
profiles during flower bud maturation. In both male and
female flower buds, expression of SlAP3A increased
dramatically in early mature buds and then gradually
decreased (fig. 3A). SlAP3A transcript accumulated
abundantly in petals of both male and female buds and
was also present at high levels in the females’ styles (fig.
3B). In contrast, SlAP3Y transcription increased steadily
throughout male flower bud maturation (fig. 3A), with
high transcript levels in petals and higher transcript levels
in stamens (fig. 3B).
Gene duplications are a source of evolutionary
novelties, including new gene functions and expression
patterns (Ohno 1970). Our results suggest that both
duplicates in the dioecious Silene species have escaped
silencing by degenerative mutations. The expression data
suggest that the duplicate copy retains only some of the
original functions and tissue expression locations. These
changes in expression may have involved mutations in the
genes’ regulatory regions or in the coding regions (Force et
al. 1999). Estimated sequence changes in the Y-linked
versus autosomal gene lineages since the duplication do
not differ significantly (the numbers were, respectively,
nine versus seven synonymous changes and seven versus
11 nonsynonymous changes). Again, unlike Hawaiian
silverswords (Barrier, Robichaux, and Purugganan 2001),
there is no evidence for accelerated subsitution in either
Silene duplicate and thus no suggestion that either is
evolving a new function. This is supported by nonsignificant McDonald-Kreitman tests (McDonald and
Kreitman 1991).
The evolution of the SlAP3 duplicates may thus have
involved mainly regulatory mutations changing the
temporal expression of SlAP3A, so that expression of
SlAP3Y has changed from petal-specific to largely anther
expression. Even without changes in the coding regions,
quantitative subfunctionalization is possible, resulting
from mutations reducing the expression of both copies
(Force et al. 1999). To test for this possibility, we did
quantitative RT-PCR with primers from identical regions
of ScAP3A, SlAP3A, and SlAP3Y (fig. 3C). If SlAP3A and
SlAP3Y are both expressed in petals in the same manner as
the ancestral gene, the total transcript level in male flowers
should be twice that in the bisexual flower of S. conica.
However, the level of SlAP3A plus SlAP3Y in petals of
male flowers was similar to the level of the single-copy
ScAP3A, whereas SlAP3A in female flowers’ petals was
lower (fig. 3C).
A MADS Box Gene on a Plant Y Chromosome 1067
duplication occurred soon after the evolution of the sex
chromosomes.
SlAP3Y Acquired a Male Function Despite Degenerative
Forces on the Y Chromosome
FIG. 4.—Schematic representation of the evolutionary history of
SlAP3. Duplication of the ancestral APETALA3 (white arrowhead)
produced the eu-AP3 and TM6 lineages. In the TM6 lineage, duplication
and transposition to the Y chromosome (black arrowhead) occurred after
the evolution of the sex chromosomes.
The Duplication Occurred After the Evolution of the
Sex Chromosomes
Given the evidence for altered gene expression, it is
of interest to ask when the duplication that created the
autosomal and Y-linked paralogs occurred, relative to the
origin of the sex chromosomes. This can be estimated from
the synonymous nucleotide divergence in the coding
regions of the SlAP3 homologs (table 1), assuming that
synonymous substitutions are close to neutral. Divergence
between the autosomal and Y-linked paralogs is lower than
divergence of either of them from ScAP3A (table 1),
suggesting that the duplication event occurred after the
divergence of S. conica and the three dioecious species
(fig. 4). Tajima’s relative rate test (1993) was nonsignificant for the entire sequence (800 nucleotides) or
for the coding sequence, for which the alignment is more
certain. In the Neighbor-Joining tree, there was 99%
bootstrap support for the branch of the tree that includes all
the Y sequences and for the autosomal sequences (data not
shown). Under the alternative hypothesis that S. conica
had two loci and lost one of them, whereas the dioecious
species retained both (one remaining autosomal and the
other being on the chromosome that evolved to become the
Y), the divergence values would be expected to be very
different from this. Sequence divergence should then be
lowest between the S. conica ScAP3A and its ortholog in
the dioecious species, whereas divergence between
ScAP3A and the nonorthologous sequence should be much
larger, since the duplication that created the two genes
must predate the split between the ancestors of S. conica
and the dioecious species. However, both sequences in the
dioecious species are roughly equally diverged from
ScAP3A, which rules out this possibility. Moreover, the
Ks values for SlAP3A versus SlAP3Y are consistent with
those for the two known S. latifolia X-linked and Y-linked
gene pairs (Atanassov et al. 2001), suggesting that the
SlAP3Y is expressed almost exclusively in developing
stamens, which strongly suggests that it functions in the
maturation of anthers and pollen grains (figs. 2B and 3B).
SlAP3Y has a different function in male fertility from that
of the autosomal SlAP3A, which appears to have preserved
ancestral functions expressed during petal development,
that is, SlAP3Y probably acquired a new function in anther
maturation after its duplication to the Y chromosome. This
presumably accounts for its having maintained function,
despite the degenerative forces acting on nonrecombining
Y chromosomes (Charlesworth and Charlesworth 2000).
Processes involved in the genetic degeneration have been
inferred from evidence for a reduction in effective
population size for Y-linked genes. Assuming equilibrium
under neutrality, Y-linked genes should have one quarter
of the diversity of autosomal loci. Significantly lower
diversity than this neutral expectation has been found for
certain genes on the neo-Y chromosome of a fruitfly,
Drosophila miranda, based on lower levels of synonymous nucleotide variability (Bachtrog and Charlesworth
2002). In S. latifolia, the Y-linked genes SlY1 and SlY4
also have lower nucleotide variability than predicted
(Filatov et al. 2000; Laporte, Filatov, and Charlesworth,
unpublished data). However, the only autosomal locus so
far surveyed also had low variability, and analysis of its
variants suggests that diversity may have been reduced by
a recent selective sweep (Filatov et al. 2001), so that it is
uncertain whether Y chromosomal variability is low or the
X-linked genes have unusually high diversity.
We have compared the species-wide nucleotide
diversity of SlAP3Y with that of the autosomal SlAP3A
using sequences of these genes from 12 different populations. First, the autosomal locus diversity is consistent with that predicted from the previously studied
X-linked genes, making an elevated X-linked diversity
unlikely. Comparisons with autosomal genes are not
possible at present, since, as just mentioned, the only
other such gene cannot be assumed to be at equilibrium,
but may have recently lost diversity. Diversity of X-linked
loci could, however, be used to predict the expected
neutral diversity value for autosomal genes. Two loci have
been surveyed. The species-wide synonymous site p
values were about 0.027 for SlX1 (Filatov et al. 2001),
and the very high value of 0.059 for SlX4, possibly due to
introgression from S. dioica (Laporte, Filatov, and
Charlesworth, unpublished data). These results suggest
Table 2
Nucleotide Diversity for Nonsynonymous and Synonymous Sites of the Autosomal and
Y-Linked Paralogs
Locus
SlAP3A
SlAP3Y
Sample Size
No. of Segregating Sites
pa
ps
h 6 Standard Deviation
12
12
11
2
0.0020
0.00083
0.0108
0.0012
0.0058 6 0.0027
0.0011 6 0.0008
1068 Matsunaga et al.
an expected autosomal value of about 3% to 7%, and 1%
to 2% for Y chromosomal genes. Both these are
considerably higher than our observed values (note,
however, that the samples are comparable, although not
from the same set of populations). If these high values are
typical, which is not certain, given the high variance of
such estimates, our autosomal locus has lower than the
predicted diversity. Even based on the low SlAP3A value,
SlAP3Y diversity, measured as either the mean pairwise
proportion of sites differing between alleles (p) or from the
number of segregating sites (h) was six to nine times lower
than the SlAP3A value (table 2); based on the standard
errors of the h values, this is just significantly below the
predicted value of four times lower. Thus, as for the other
Y-linked genes in this species, we conclude the diversity
of SlAP3Y is smaller than that of SLAP3A to an extent
greater than expected. The effect is less than for SlY1, but
it again suggests that selective events at other Y-linked loci
may be reducing diversity, as expected if this Y
chromosome is undergoing a process of degeneration
(Charlesworth and Charlesworth 2000).
Duplications to the Y chromosome may be a regular
event in the evolution of sex chromosome systems,
allowing the accumulation of genes that are advantageous
in males but disadvantageous in females (Charlesworth
and Charlesworth 1980; Rice 1997). Such a hypothesis is
plausible for SlAP3Y, with its increased stamen expression.
Fixation of an interchromosomal transposition would be
a unique event, which would initially completely eliminate
Y-linked diversity and might potentially be detectable
from a pattern of excess low-frequency synonymous
variants, which is the signature of a selective sweep
associated with the recent spread of an advantageous
mutation (Tajima 1989). Such a result would be very
interesting because it would support the hypothesis of
selection promoting the spread of the Y-linked duplication.
However, the SlAP3Y diversity is too low to use its
variants in this test. Moreover, we estimate that the event
occurred too long ago for any effects of a selective sweep
to be detectable, even if one occurred. It seems more likely
that the low SlAP3Y diversity is due to currently operating
causes, most likely processes involved in Y chromosome
degeneration, rather than to such a selective sweep.
Interestingly, in contrast to the testis-specific expression of DAZ, SlAP3Y retains some autosomal ancestral
expression in petals (as well as its higher anther
expression). The extent of the transposed genomic region
that includes SlAP3Y is not yet known, but it probably
encompasses not only introns and exons but also the
promoter region. The changed SlAP3Y male function may
have been caused, in part, by degenerative mutations in the
promoter, leading to reduced petal expression. Our results
suggest that the acquisition of autosomal genes is an
important component of plant Y chromosomes. If so, it is
probably important in Y chromosome evolution in general,
which has until now been a purely speculative idea.
Supplementary Material
The sequences of the SlAP3A, SlAP3Y, SdAP3A,
SdAP3Y, SiAP3A, SiAP3Y, and ScAP3A cDNA have been
deposited in GenBank. Accession numbers are AB090863,
AB090864,
AB090865,
AB090866,
AB090867,
AB090868, and AB090869, respectively.
Acknowledgments
Financial support was provided by the Ministry of
Education, Sports and Culture of Japan (S.M. and S.K.),
Grant Agency of the Czech Republic (E.K., B.V., and J.D.:
No. 521/02/0427 and 204/02/0417), and the Natural
Environment Research Council of Great Britain (D.C.).
Literature Cited
Atanassov, I., C. Delichere, D. A. Filatov, D. Charlesworth, I.
Negrutiu, and F. Moneger. 2001. Analysis and evolution of
two functional Y-linked loci in a plant sex chromosome
system. Mol. Biol. Evol. 18:2162–2168.
Bachtrog, D., and B. Charlesworth. 2002. Reduced adaptation
of a non-recombining neo-Y chromosome. Nature 416:
323–326.
Barrier, M., R. H. Robichaux, and M. D. Purugganan. 2001.
Accelerated regulatory gene evolution in an adaptive
radiation. Proc. Natl. Acad. Sci. USA 98:10208–10213.
Baudinette, S. C., T. W. Stevenson, and K. W. Savin. 2000.
Isolation and characterisation of the carnation floral-specific
MADS box gene, CMB2. Plant Sci. 155:123–131.
Charlesworth, B., and D. Charlesworth. 2000. The degeneration
of Y chromosomes. Proc. R. Soc. Lond. B Biol. Sci. 355:
1563–1572.
Charlesworth, D., and B. Charlesworth. 1980. Sex differences in
fitness and selection for centric fusions between sexchromosomes and autosomes. Genet. Res. 35:205–214.
Delichere, C., J. Veuskens, M. Hernould, N. Barbacar, A.
Mouras, I. Negrutiu, and F. Moneger. 1999. SlY1, the first
active gene cloned from a plant Y chromosome, encodes
a WD-repeat protein. EMBO J. 18:4169–4179.
Desfeux, C., S. Maurice, J.-P. Henry, B. Lejeune, and P.-H.
Gouyon. 1996. Evolution of reproductive systems in the
genus Silene. Proc. R. Soc. Lond. B Biol. Sci. 263:409–414.
Filatov, D. A., V. Laporte, C. Vitte, and D. Charlesworth. 2001.
DNA diversity in sex linked and autosomal genes of the plant
species Silene latifolia and S. dioica. Mol. Biol. Evol. 18:
1442–1454.
Filatov, D. A., F. Moneger, I. Negrutiu, and D. Charlesworth.
2000. Low variability in a Y-linked plant gene and its implications for Y-chromosome evolution. Nature 404:388–390.
Force, A., M. Lynch, F. B. Pickett, A. Amores, Y. Yan, and J.
Postlethwait. 1999. Preservation of duplicate genes by
complementary, degenerative mutations. Genetics 151:1531–
1545.
Grant, S., B. Hunkirchen, and H. Saedler. 1994. Developmental
differences between male and female flowers in the dioecious
plant Silene latifolia. Plant J. 6:471–480.
Guttman, D. S., and D. Charlesworth. 1998. An X-linked gene
has a degenerate Y-linked homologue in the dioecious plant
Silene latifolia. Nature 393:263–266.
Hardenack, S., D. Ye, H. Saedler, and S. Grant. 1994.
Comparison of MADS box gene expression in developing
male and female flowers of the dioecious plant white campion.
Plant Cell 6:1775–1787.
Jack, T., L. L. Brockman, and E. M. Meyerowitz. 1992. The
homeotic gene APETALA2 of Arabidopsis thaliana encodes
a MADS box and is expressed in petals and stamens. Cell
68:683–697.
Kejnovsky, E., J. Vrana, S. Matsunaga, P. Soucek, J. Siroky, J.
A MADS Box Gene on a Plant Y Chromosome 1069
Dolezel, and B. Vyskot. 2001. Localization of malespecifically expressed MROS genes of Silene latifolia by
PCR on flow-sorted X chromosomes and autosomes. Genetics
158:1269–1277.
Kramer, E. M., R. L. Dorit, and V. F. Irish. 1998. Molecular
evolution of genes controlling petal and stamen development:
duplication and divergence within the APETALA3 and
PISTILLATA MADS-box gene lineages. Genetics 149:
765–783.
Kumar, S., K. Tamur, I. Jacobsen, and M. Nei. 2000. MEGA2:
molecular evolutionary genetics analysis. Version 2.0.
Pennsylvania and Arizona State Universities, University Park,
Pennsylvania and Tempe, Arizona.
Lahn, B.T. and D. C. Page. 1997. Functional coherence of the
human Y chromosme. Science 278:675–679.
Lebel-Hardenack, S., E. Hauser, T. F. Law, J. Schmid, and S. R.
Grant. 2002. Mapping of sex determination loci on the white
campion (Silene latifolia) Y chromosome using amplified
fragment length polymorphism. Genetics 160:717–725.
Matsunaga, S., and S. Kawano. 2001. Sex determination by sex
chromosomes in dioecious plants. Plant Biology 3:481–488.
Matsunaga, S., S. Kawano, T. Michimoto, T. Higashiyama, S.
Nakao, A. Sakai, and T. Kuroiwa. 1999. Semi-automatic laser
beam microdissection of the Y chromosome and analysis of Y
chromosome DNA in a dioecious plant, Silene latifolia. Plant
Cell Physiol. 40:60–68.
Matsunaga, S., S. Kawano, H. Takano, H. Uchida, A. Sakai, and
T. Kuroiwa. 1996. Isolation and developmental expression of
male reproductive organ-specific genes in a dioecious campion, Melandrium album (Silene latifolia). Plant J. 10:
679–689.
McDonald, J. H., and M. Kreitman. 1991. Accelerated protein
evolution at the Adh locus in Drosophila. Nature 351:652–
654.
Meyerowitz, E. M. 2002. Plants compared to animals: the
broadest comparative study of development. Science 295:
1482–1485.
Negrutiu, I., B. Vyskot, N. Barbacar, S. Georgiev, and F.
Moneger. 2001. Dioecious plants: a key to the early events of
sex chromosome evolution. Plant Physiol. 127:1418–1424.
Ng, M., and M. F. Yanofsky, 2001. Function and evolution of the
plant MADS-box gene family. Nature Rev. Genet. 2:186–195.
Ohno, S. 1970. Evolution by gene duplication. Springer-Verlag,
Heidelberg, Germany.
Pnueli, L., M. Abu-Abeid, D. Zamir, W. Nacken, Z. SchwarzSommer, and E. Lifschitz. 1991. The MADS box gene family
in tomato: temporal expression during floral development,
conserved secondary structures and homology with homeotic
genes from Antirrhinum and Arabidopsis. Plant J. 1:255–266.
Renner, S. S., and R. E. Ricklefs. 1995. Dioecy and its correlates
in the flowering plants. Am. J. Bot. 82:596–606.
Rice, W. R. 1997. The accumulation of sexually antagonistic
genes as a selective agent promoting the evolution of reduced
recombination between primitive sex-chromosomes. Evolution 41:911–914.
Rozas, J., and R. Rozas. 1999. DnaSP version 3.0: an integrated
program for molecular population genetics and molecular
evolution analysis. Bioinformatics 15:174–175.
Saxena, R., L. G. Brown, T. Hawkins et al. (11 co-authors). 1996.
The DAZ gene cluster on the human Y chromosome arose
from an autosomal gene that was transposed, repeatedly
amplified and pruned. Nat. Genet. 14:292–299.
Tajima, F. 1989. Statistical method for testing the neutral
mutation hypothesis. Genetics 123:585–595.
———.1993. Simple methods for testing the molecular evolutionary clock hypothesis. Genetics 135:599–607.
Uchida W., S. Matsunaga, R. Sugiyama, F. Shibata, Y. Kazama,
Y. Miyazawa, M. Hizume, and S. Kawano. 2002. Distribution of interstitial telomere-like repeats and their adjacent sequences in a dioecious plant, Silene latifolia. Chromosoma
111:313–320.
Westergaard, M. 1958. The mechanism of sex determination in
dioecious flowering plants. Adv. Genet. 9:217–281.
Brandon Gaut, Associate Editor
Accepted February 26, 2003