Skip to main content
NCBI home page
Plant Physiology logoLink to Plant Physiology
. 2022 Dec 10;191(2):1122–1137. doi: 10.1093/plphys/kiac571

The double flower variant of yellowhorn is due to a LINE1 transposon-mediated insertion

Hanhui Wang 1,b, Yanan Lu 2,3,b, Tianxu Zhang 4,b, Zhi Liu 5,b, Li Cao 6,7,b, Qiaoying Chang 8,b, Yueying Liu 9,10,b, Xin Lu 11,12,13,b, Song Yu 14,b, Huiyu Li 15,b, Jing Jiang 16,b, Guifeng Liu 17,b, Heike W Sederoff 18,b, Ronald R Sederoff 19,20,b, Qingzhu Zhang 21,b, Zhimin Zheng 22,✉,c,b
PMCID: PMC9922402  PMID: 36494195

Abstract

As essential organs of reproduction in angiosperms, flowers, and the genetic mechanisms of their development have been well characterized in many plant species but not in the woody tree yellowhorn (Xanthoceras sorbifolium). Here, we focused on the double flower phenotype in yellowhorn, which has high ornamental value. We found a candidate C-class gene, AGAMOUS1 (XsAG1), through Bulked Segregant Analysis (BSA) sequencing and genetics analysis with a Long Interpersed Nuclear Elements 1 (LINE1) transposable element fragment (Xsag1-LINE1-1) inserted into its second intron that caused a loss-of-C-function and therefore the double flower phenotype. In situ hybridization of XsAG1 and analysis of the expression levels of other ABC genes were used to identify differences between single- and double-flower development processes. These findings enrich our understanding of double flower formation in yellowhorn and provide evidence that transposon insertions into genes can reshape plant traits in forest trees.

Introduction

Flower development in angiosperms is comprised of flowering induction, flower initiation, and the development of floral organs (Zik and Irish, 2003), which are each under the control of many genes, including homeotic genes that are expressed at specific times and locations (O'Maoileidigh et al., 2014). ABC models of floral development were first inferred from genetic analyses of Arabidopsis (Arabidopsis thaliana) and snapdragon (Antirrhinum majus; Schwarz-Sommer et al., 1990; Bowman et al., 1991). The Class A genes APETALA1 (AP1) and AP2 in Arabidopsis control the development of sepals; Class A and B genes such as AP3 and PISTILLATA (PI) control petals; Class B and C genes like AGAMOUS (AG) control stamens; and only Class C genes control carpels (Bowman et al., 1991, 2012; Causier et al., 2010; O’Maoileidigh et al., 2014; Irish, 2017). Some genes such as SHATTERPROOF1 (SHP1), SHP2, and SEEDSTICK (STK) have been found to share high sequence similarity to AG but have different functions (Liljegren et al., 2000; Pinyopich et al., 2003; Colombo et al., 2008). In addition, Class E SEPALLATA (SEP) proteins can form multimeric complexes with Class B and C proteins to mediate their interactions (Honma and Goto, 2001; Ditta et al., 2004). These genes encode core transcriptional regulators, and all except AP2 belong to MADS-box gene family (Ng and Yanofsky, 2001; Smaczniak et al., 2012).

Many studies have shown that the C-function protein, AGAMOUS, plays an important role in reproductive organ development and the maintenance and termination of floral meristem (Yanofsky et al., 1990; Lenhard et al., 2001; Sun et al., 2009; Liu et al., 2011). It also directly activates the downstream target gene CRABS CLAW (CRC) involved in carpel development (Gomez-Mena et al., 2005; Lee et al., 2005a, 2005b). Loss-of-function of AG leads to the continuous active expression of WUS, resulting in the indeterminacy of flower meristem, thus ag mutants show the “flower-within-flower” phenotype (Bowman et al., 1991; Lenhard et al., 2001). The double flower trait is one of the most important traits in flower culture; its morphology, development, and heredity in many species have been studied (Abdirashid and Lenhard, 2020). In Japanese morning glory (Ipomoea nil or Pharbitis nil), the double flower phenotype is caused by a large deletion of DUPLICATED (DP) gene, due to the insertion of an En/Spm-related transposable element (TE) in the second intron of DP gene (Nitasaka, 2003). Analysis of the “Double White” ThAG1 gene in Thalictrum thalictroides has revealed that an long terminal repeated (LTR) retrotransposon inserted into the fourth exon results in the deletion of K-domain. Another transposon insertion has been also found in Japanese Kerria (Kerria japonica; Rosaceae) double flowers, resulting in the loss of AG gene function (Ma et al., 2018). Although few cases of transposon insertion in MADS-box family genes affecting flower development have been reported, it is undeniable that transposon insertion is of great importance for shaping of plant phenotypes (Lisch, 2013; Dubin et al., 2018).

Yellowhorn (Xanthoceras sorbifolium) is a relict woody tree species in the Sapindaceae family and is endemic to Northern China. It is not only an important woody tree with oil-producing seeds, but also has a high ornamental value because of its flowers with different types (Xiao et al., 2013; Yao et al., 2013; Shen et al., 2018; Li et al., 2020). Yellowhorn double-flowered trees, regarded as sterile mutants spontaneously produced in natural populations, have a high ornamental value because of its large and gorgeous flowers. Interestingly, single-flowered trees can bear both fertile and infertile flowers when one of the reproductive organs aborts (Zhou et al., 2012). In spite of the recent interest in yellowhorn, there has not been much research into its flower, especially the double flower phenotype.

Based on our study of yellowhorn single and double flower populations and forward genetic analysis, we identified the XsAG1 gene. The insertion of a Long Interspersed Nuclear Elements 1 (LINE1) subfamily fragment Xsag1-LINE1-1 into the second intron of XsAG1 likely led to the formation of double flowers in yellowhorn. We suggest that the homozygous Xsag1-LINE1-1 insertion in the trees bearing double flowers affects transcription of the XsAG1 gene and leads to its loss of function. Our results demonstrate the highly conserved function of XsAG1 in determining carpel and stamen identities, shed light on the molecular mechanisms involved in yellowhorn double flower formation, and also provide information that transposon insertion into introns could affect gene function, leading to double flower formation and evolution.

Results

Double flower morphology and development in yellowhorn

We subdivided yellowhorn accessions collected across China (Supplemental Table 1) as either single- or double-flowered trees, according to the numbers of petals and whorls found in their flowers (Figure 1 and Supplemental Figure 1). Double flower trees are uncommon in China, and the number of petals is considerably variable (Figure 1C). A typical double flower generally contains multiple whorls of petals (more than two) and one whorl of sepals. Double flowers do not bear fruit due to the absence of pistils and stamens. The morphology of petals in double flowers is different compared with that of single flowers; they have normal outer-whorl petals and stamen-like inner-whorl petals (Supplemental Figure 1, G–I). A fused-organ (carpel-like organ) appears in the innermost whorl, much like mutated carpels (Supplemental Figure 1, G–I). The appearance of these special organs suggests that yellowhorn double flowers lose the determinacy of the floral organs. In single-flowered trees, Type I flowers contain a normal enlarged ovary surrounded by degenerated stamens with short filaments, which can be fertilized and produce fruit, while Type II flowers contain only a degenerated carpel but do contain normal stamens with long filaments and dehiscent anthers that produce pollen (Figure 1, A–C and Supplemental Figures 1 and 3).

Figure 1.

Figure 1

Open in a new tab

Flowers and floral organ structures of X. sorbifolium. A, Single and double flower phenotypes in X. sorbifolium and Arab. thaliana. Scale bars: 0.5 cm for X. sorbifolium, 2 mm for Arabidopsis. B, Schematic illustration of floral organs in X. sorbifolium and Arabidopsis: bract (dark green), sepal (green), petal (white), horn-like appendage (orange), stamen (yellow and orange), and carpel (blue). C, The number of floral organs of mature flowers in X. sorbifolium. The data represent mean ± SD (n = 60). D, Anatomy and electron microscopy of the development of single and double flowers in X. sorbifolium. The first three are single flowers, and the last three are double flowers. Scale bars: 1 cm for anatomy, 100 nm for electron microscopy figures. Ap, horn-like appendage; Ca, carpel; Pe, petal; Se, sepal; St, stamen.

To better understand differences between single and double flowers, we traced their developmental trajectories. In the early stage of double flower development, the sepal primordia first developed normally, followed by multiple petal primordia, while no stamen primordia or pistil primordia appeared (Figure 1D and Supplemental Figures 2B and 4). The outer-whorl petal development is slower than that of the inner whorl, and stamen-like petals appear near the central fused organ, the inside of which has a new floral meristem (Supplemental Figures 4 and 5). While in single flower development, sepal primordia were first initiated, followed by petal, stamen, and gynoecium at the early stages (Figure 1 and Supplemental Figure 2A). Petal primordia grow but remain relatively small, while stamens develop rapidly and differentiate into anthers and filaments (Figure 1D and Supplemental Figure 3). Then one of the male and female reproductive organs degenerated, forming two types of single flower (Supplemental Figure 3).

In addition, the absence of another important structure in double flowers is also noted: the horn-like appendage (ap; Figure 1 and Supplemental Figure 1). The five horn-like appendages are the most important symbol of X. sorbifolium, and its orange-yellow horn-like shape is the source of “yellowhorn.” It occurs between petals and stamens, and its surfaces secrete mucus (Supplemental Figures 5 and 6). We speculate that it plays an important role in attracting insect pollination as a nectarium (multiple nectary; Alvarez and Smyth, 1999; Bowman and Smyth, 1999; Lee et al., 2005a, 2005b).

Bulked Segregant Analysis (BSA) sequencing locates a candidate gene associated with the double flower phenotype

DNA from natural populations of single- or double-flowered trees was bulked for Bulked Segregant Analysis (BSA) sequencing (Figure 2A). A pool of single flower phenotypes (S-pool) was created by pooling genomic DNA from 98 single-flowered individuals, and a separate pool was created by pooling genomic DNA from 50 double-flowered individuals (D-pool; Figure 2A and Supplemental Tables 2 and 3).These two samples of pooled DNA were then sequenced using an Illumina HiSeq 4000 (Supplemental Table 2), and clean reads from each of the two pools were aligned with the yellowhorn reference genome from the GigaScience Database (GigaDB at http://dx.doi.org/10.5524/100606; Bi et al., 2019; Supplemental Table 3). We used QTLseqr to calculate G values (Mansfeld and Grumet, 2018) and found that one genomic region (Chr5: 1,214,218–3,922,981, ∼2.7 Mb) exhibited a G value greater than the threshold (Figure 2, B and C, Supplemental Table 4). A total of 21,589 single nucleotide polymorphisms (SNPs) were detected at q < 1 × 10−6 (Supplemental Data Set 1). Thus, this ∼2.7-Mb region including 269 genes could contain a locus controlling the double flower phenotype (Figure 2B and Supplemental Data Set 2).

Figure 2.

Figure 2

Open in a new tab

Mapping the genomic interval correlated with the double flower phenotype and identifying candidate genes for floral development by BSA sequencing. A, Flow chart of BSA sequencing of single- and double-flowered trees in X. sorbifolium. B, Major loci for double flower phenotype and the number of genes in the associated region. The candidate gene EVM0013090.1 is shown in red. C, Graph of the G values calculated from BSA analysis of the genetics of single and double flower phenotypes. The x-axis indicates the position of 15 chromosomes in X. sorbifolium, and the y-axis indicates G values.

Then, we performed a resequencing on three double- and eight single-flowered trees (Supplemental Table 5), and genes were filtered by screening SNPs and InDels shared by the ∼2.7-Mb region. SNPs in exon regions, InDels (2–50 bp), and SVs (>50 bp) in genes or promoter regions (2 kb upstream of genes) were identified and a total of 75 SNPs, 20 InDels, and 4 SVs were screened from 16 genes (Supplemental Data Set 3). There were 41 SNPs that could cause amino acid nonsynonymous mutations located in 11 genes, and 8 genes had potentially important InDels in promoter or intron regions. Gene function annotations showed that many enzymes such as GDSL esterase or lipase (Precursor), cyclic phosphodiesterase, alpha, alpha-trehalose-phosphate synthase, serine/threonine-protein kinase, or probable LRR receptor-like serine/threonine-protein kinase were involved (Supplemental Figure 7 and Data Set 2). Only one gene, EVM0013090.1, annotated as a flower homeotic gene, was involved in carpel development (GO:0048440), stamen development (GO:0048443), and maintenance of floral organ identity (GO:0048497) (Supplemental Figure 7).

In order to verify the candidate gene, we conducted RNA-seq on 3 single- and 3 double-flower bud samples, and the expression patterns of 16 genes were analyzed (Supplemental Tables 6 and 7 and Figure 7). Only one of the 16 genes, EVM0013090.1, had a significantly lower expression level in double flower buds than in single flower buds (|Log2FC| = 16.1136, P = 1.16E-28, FDR = 8.78E-25), and 1 InDel on this gene was found (Supplemental Figure 7, Table 7, and Data Set 2). We found that EVM0013090 is a floral homeotic protein belonging to the AG evolutionary clade; it is a 243-aa protein containing N-terminal MADS-box and C-terminal K-box domains (Supplemental Figure 8). Thus, we considered EVM0013090.1 as a candidate gene and named it XsAG1.

The only one InDel in XsAG1 is located in intronic regions, and XsAG1 coding sequences were identical in genomes of both single- and double-flowered trees (Supplemental Figure 9). Reads on both sides of this InDel did not match very well, and through improved read alignments we identified the insertion of a 539-bp additional sequence within the second intron of XsAG1 in double-flowered trees (Figure 3A and Supplemental Figure 10). By searching Repbase databases (https://www.girinst.org/censor/index.php; Kohany et al., 2006), the sequence we identified likely belongs to the LINE1 family of retrotransposons (Supplemental Figure 11), therefore we named this sequence Xsag1-LINE1-1. The 539-bp Xsag1-LINE1-1 sequence was used as a BLASTN query (Supplemental Data Set 4), and a total of 279 LINE1s divided into 5 categories with >90% similarity were found (Supplemental Data Set 5 and Figure 12), and one transposon (Lachesis_group4:1,196,133-1,202,795) belonging to Type I had 98.833% sequence similarity with Xsag1-LINE1-1 but shared only 25 SNPs, indicating that Xsag1-LINE1-1 could be its parental or sibling element (Supplemental Data Set 5).

Figure 3.

Figure 3

Open in a new tab

A 539-bp fragment Xsag1-LINE1-1 insertion into the second intron of XsAG1 in double-flowered X. sorbifolium trees. A, Diagram of X. sorbifolium AGAMOUS1 gene (XsAG1) structure and the 539 bp fragment Xsag1-LINE1-1 insertion into the second intron of XsAG1 in double flowers. Black boxes represent exons, gray boxes represent untranslated regions, green boxes represent Xsag1-LINE1-1 insertion, and lines represent introns. Primers GF, GR, and TF were used for genotyping, while Primers GF′, GR′, GF1, GR1, and UTR-R were used for analysis of gene expression patterns. B, PCR analysis of Xsag1-LINE1-1-based markers (Primers GF, GR, and TF) in single- and double-flowered X. sorbifolium trees sampled from different regions of China. C, Two types of single flower (no Xsag1-LINE1-1 insertion or heterozygous for the Xsag1-LINE1-1 insertion) and one type of double flower (homozygous for the Xsag1-LINE1-1 insertion) were analyzed by RT-PCR using XsACTIN as the internal control. The Primers GF′ and GR′ amplify the full-length XsAG1 coding sequences, GF′ and GR1 amplify the first exon of XsAG1, and GF1 and GR′ amplify sequences from second to the seventh exon. Electropherograms show results of 30 cycles of PCR amplifications. D, XsAG1 RT-PCR results for single and double flower buds with different numbers of amplification cycles (30, 35, or 40 cycles of amplification using the same primers, XsAG1-ATG-F and UTR-R, to amplify XsAG1 with XsACTIN as the internal control. E, Different XsAG1 transcripts resulting from alternative splicing in single and double flowers.

Xsag1-LINE1-1 insertion in XsAG1 results in a recessive mutation

Considering the absence of other genetic differences in XsAG1 sequences, the major insertion Xsag1-LINE1-1 in Intron 2 was suggested to be the most likely cause of the double flower phenotype (Supplemental Figure 9 and Data Set 6). Xsag1-LINE1-1 in XsAG1 Intron 2 was found to be present on both chromosomes in all double flower samples collected nationwide by using GF + GR and TF + GR primers, and its sequences in different individuals were exactly the same, suggesting that Xsag1-LINE1-1 insertion in XsAG1 was homozygous in all double-flowered trees (Figure 3, A and B, Supplemental Figures 11 and 13). Then 238 individuals from natural single-flowered tree populations were detected using the same primers, and the results showed that Xsag1-LINE1-1 insertion was also present in single-flowered trees but only in one chromosome and the frequency was only 10.5% (25/238); most single-flowered trees had no insertion in XsAG1 (Figure 3, A and B, Supplemental Figures 11 and 14). We also found that a tree called CF74 was the only single-flowered ancient tree (about 270 years old) that carries a heterozygous Xsag1-LINE1-1 insertion in XsAG1 among all ancient trees at Fanzong Temple in Chifeng City (Supplemental Table 1 and Figure 15). The occurrence of heterozygous Xsag1-LINE1-1 insertion in XsAG1 in single-flowered trees provides strong genetic evidence that this fragment insertion is responsible for the changes in floral traits.

The expression levels of XsAG1 were substantially different in double flower samples than in single flower samples. With 30 cycles of amplification, no expression of XsAG1 full-length transcripts could be detected in double flower buds, and the abundance of Exon1 transcripts was slightly lower and that of downstream exon transcripts was sharply lower in double flowers than in single flowers (Figure 3C). We also found that in CF74 and other single-flowered trees with the heterozygous Xsag1-LINE1-1 insertion in XsAG1, the expression levels were similar to those in other single-flowered trees (Figure 3C and Supplemental Figure 15). Thus, the inserted transposon fragment Xsag1-LINE1-1 in XsAG1 Intron 2 is the key to the genetic variation between single and double flowers. Therefore, a single-flowered tree without the Xsag1-LINE1-1 insertion corresponds to the wild-type, and a double-flowered tree carrying a homozygous Xsag1-LINE1-1 insertion represents a visible Xsag1 mutant.

Given that an Xsag1-LINE1-1 insertion affects XsAG1 transcription in double-flowered samples, we aimed to determine whether its transcript was changed; however, we were not able to identify the transcriptional termination position of XsAG1 in double flower samples using 3′-RACE, possibly because its transcriptional level in double flowers was too low to detect (Figure 3, C and D). However, the Xsag1-LINE1-1 insertion appears to affect the transcription of XsAG1 subsequent introns and exons, which might also be related to variable splicing (Supplemental Figure 16). Increasing the number of polymerase chain reaction (PCR) cycles also allowed visualization of the XsAG1 transcripts amplified from double-flowered tree samples, including both the original transcript from single-flowered tree samples and other transcripts of different sizes (Figure 3D). These PCR products amplified from samples from double-flowered trees were directly cloned into pEASY-Blunt zero vectors, and single colonies were selected for sequencing. Sequencing these clones revealed that in double-flowered trees, there were original transcripts and aberrant transcripts with alternative splice forms (Figure 3, D and E, Supplemental Figure 17). We detected four aberrant transcripts in samples from double-flowered trees, all of which contained 155- to 202-bp partial Xsag1-LINE1-1 fragments in XsAG1 (Figure 3E and Supplemental Figure 17). The alternative splice sites at the 5′ end of Xsag1-LINE1-1 in all aberrant transcripts were the same, while different splice sites were detected at the 3′ end of the Xsag1-LINE1-1 that resulted in different lengths of aberrant transcripts. It is worth noting that the TGA termination codon appeared in the same position in all aberrant transcripts containing different Xsag1-LINE1-1 lengths detected in double flower samples, suggesting that premature translation termination might have led to a loss of function of the XsAG1 protein (Supplemental Figure 18). The possibility of many alternative spliced transcripts increases the difficulty of isolation of the full-length of these transcripts.

Expression pattern of XsAG1 and other ABC model genes in yellowhorn single- and double-flowered trees

To gain more insights into the differences in XsAG1 expression patterns between single- and double-flowered tree samples, we performed in situ hybridization analyses of XsAG1 using gene-specific probes (designed on Exons 2 to 4; Figure 4A and Supplemental Figure 19). We found that XsAG1 in both single and double flower buds was specifically expressed in the primordia and developing organs, while the probe signal in single flower buds was stronger than that in double flower buds (Figure 4A). In single flower buds, the XsAG1-expressed primordia later become stamens and carpels, while XsAG1 expression was also present in inner whorl organs in double flower buds, but its expression gradually disappeared with development (Figure 4A). These results showed that the expression of XsAG1 was significantly decreased in double flowers (Student's t test, ***P < 0.001; Figure 4B), but weakly expressed, which may be the reason for the appearance of stamen-like petals or carpel-like organs in the inner whorl of double flowers.

Figure 4.

Figure 4

Open in a new tab

Expression pattern of XsAG1 gene and other ABC model genes in single- and double-flowered X. sorbifolium trees. A, In situ hybridization results using sense probe of flower buds from single and double flowers at different developmental stages. Scar, 100 nm. Ap, horn-like appendage; Ca, carpel; Pe, petal; Se, sepal; St, stamen. B, RT-qPCR of ABC model genes in X. sorbifolium single and double flower buds. Ns, no significance; *P < 0.05, **P 0.01, ***P 0.001 (Student's t test). C, RT-qPCR of ABC model genes in X. sorbifolium different flower organs. Bars represent mean ± SD (n = 3).Ap, horn-like appendage; Ca, carpel; dCa, degenerated carpel; D-Pe, normal petal of double flower; D-Se, doubel flower sepal; dSt, degenerated stamen; D-St-like-Pe, Stamen-like petal and carpel-like organ of double flower; Pe, petal; Se, sepal; St, stamen.

We also isolated candidate floral organ identity genes (particularly A-, B-, C-, D-, and E-function genes) and investigated their expression patterns. To elucidate the relationships among the ABC homologs in yellowhorn, 14 genes were obtained (Supplemental Tables 8 and 9, Figure 8C). We analyzed whether ABCDE model homologs in yellowhorn were differentially expressed in single and double flowers by reverse transcription PCR (RT-PCR) and RT-qPCR (Supplemental Figure 20A and Figure 4B). Another gene, EVM0018010.1 (XsCRC), was also examined (Supplemental Table 9). In double flower buds, the expression levels of XsAP1-1, XsAP3-2, XsPI, and XsSEP3-1 are higher, while XsSTK and XsAG1 show considerably decreased expression levels, although the exact levels of expression vary considerably among biological repeats (Figure 4B and Supplemental Figure 20, A and B). To prevent pollen contamination, we selected unopened small flowers for floral organ collection and performed RT-qPCR analysis on these different organs (Figure 4C). Notably, XsCRC is specifically expressed in horn-like appendage; XsAG1 and XsSHP are specifically expressed in horn-like appendages, stamen, and carpels; and XsSTK is only expressed in carpels in yellowhorn single flowers (Figure 4C). While in double flower organs, the expression of XsAG1 and XsSTK was sharply decreased, and A- and B-function genes (such as XsAP1-1, XsAP3-1, XsAP3-2, and XsPI) tended to expand inward. This, together with the observation that double flowers have more petals, suggests that when XsAG1 expression is normal in single flowers, A- and B-function genes are normally expressed in outer whorls, while its expression could expand inward when XsAG1 expression is substantially decreased in yellowhorn double flowers.

XsAG1 exerts conserved C-class function to promote stamen formation

Transformation systems for yellowhorn are not yet well developed, and the time until these trees flower is very long, so we used the Arabidopsis ag-1/ag-1 mutant line. 35S::XsAG1-CDS vector was first transformed into the Arabidopsis AG/ag-1 heterozygous mutant (the ag-1/ag-1 mutant with double flowers is sterile) via Agrobacterium-mediated transformation. Considering the effect of the constitutive expression, we also performed transformation using a complementation construct harboring AtAG promoter. The proAtAG::XsAG1-CDS and 35S::XsAG1-CDS vectors were transformed into the Arabidopsis AG/ag-1 heterozygous mutant. We screened different lines of transgenic plants in the Ler and ag-1/ag-1 background by genotyping T0 and T1 generation positive seedlings and recording images of typical flower phenotypes (Figure 5, A and B). However, both the ag-1/ag-1 proAtAG::XsAG1-CDS and ag-1/ag-1 35S::XsAG1 transgenic plants did not show single flower phenotype (Figure 5, A and B, Supplemental Figures 21 and 22). Some of the wild-type proAtAG::XsAG1-CDS transgenic plants exhibited a phenotype with small and slender petals (Figure 5B), and the wild-type 35S::XsAG1 transgenic plants exhibited a phenotype without petals (Figure 5A), showing that overexpression of XsAG1 could covert petals into stamens. These petal abnormalities were somewhat similar to phenotypes of Class A loss-of-function mutants such as ap1 (Alejandra Mandel et al., 1992; Gustafson-Brown et al., 1994) and ap2 (Huang et al., 2017) as well as AtAG overexpression phenotypes in the wild-type (Castillejo et al., 2005; Airoldi et al., 2010). Although XsAG1 overexpression in ag-1/ag-1 mutants did not rescue the determinacy of flower organs in Arabidopsis, the morphology of ag-1/ag-1 35S::XsAG1-CDS transgenic double flowers changed markedly with smaller flowers, carpel-like sepals and stamens or stamen-like organs (Figure 5B), suggesting that XsAG1 could promote reproductive organ production.

Figure 5.

Figure 5

Open in a new tab

Partial Class C function of XsAG1 and potential ABC model in X. sorbifolium single and double flowers. A, Phenotypes of proAtAG::XsAG1-CDS T0 generation transgenic plants in the Arabidopsis Ler background. The top group is inflorescences and the bottom group is an individual flower. Scale bars: 1 cm. B, Phenotypes of 35S::XsAG1-CDS T1 generation transgenic plants in the Arabidopsis ag-1 mutant and Ler background. Scale bars: 1 cm. C, Genetic variation caused by Xsag1-LINE1-1 insertion may lead to changes in ABC model gene expression patterns, and control the mechanism of single and double flower development in X. sorbifolium. The arrows in A and B indicate abnormal floral organs.

Our developmental, genetic, and molecular biology experiments allow us to propose a model for the potential mechanism of genetic variation in yellowhorn double-flowered trees (Figure 5C). In this model, this LINE1 transposon jumped and integrated into the second intron of XsAG1 gene by an unknown mechanism, retaining a 539-bp fragment on one chromosome, resulting in the first heterozygous Xsag1-LINE1-1 insertion of XsAG1 in a yellowhorn single-flowered tree. This tree has the ability to produce offspring with a homozygous Xsag1-LINE1-1 insertion on the XsAG1 gene, resulting in progeny with mutant phenotypes. Homozygous Xsag1-LINE1-1 insertion leads to a decrease in XsAG1 expression, thereby affecting function and causing abnormal expression levels of other genes in the ABC model. These changes in floral organ identity genes eventually lead to the double flower phenotype of yellowhorn.

Discussion

Forward genetic evidence for screening and identification of the candidate gene related to yellowhorn double flower traits

At present, BSA and improved versions of the method have been applied in rice (Oryza sativa; Wang et al., 2019), maize (Zea mays; Zhang et al., 2019), watermelon (Citrullus lanatus; Liao et al., 2020), groundnut (Arachis hypogaea; Pandey et al., 2017), and other species. BSA has been applied less often in studies of the genetic control of important traits in forest trees but has mainly been applied in studies of orchard tree species such as Malus (Dougherty et al., 2018), pear (Pyrus pyrifolia; Xue et al., 2017), and other fruit trees. Natural populations of yellowhorn are relatively undomesticated compared to many cultivated species, so all of the existing varieties have resulted from natural rather than human-induced variation and have provided a natural mutant library for the study of gene function. Because the single and double flower phenotypes in yellowhorn represent two extreme versions of a floral trait, we were able to create two separate pools of DNA by mixing DNA samples from many single- or double-flowered individuals for BSA sequencing. We located the genomic interval associated with single or double flower phenotypes in a ∼2.7 Mb region according to G′ values (Figure 2, Supplemental Table 4, Data Sets 1 and 2).

A large region (∼2.7 Mb) was mapped using BSA sequencing and contained multiple genes (Figure 2 and Supplemental Data Set 2), only 16 of which were left after filtering SNPs and InDels (Supplemental Data Set 3). Gene expression levels and functional annotations allowed us to focus on one gene, EVM0013090.1, known as XsAG1, which has a C-function in flower development, and an inserted fragment Xsag1-LINE1-1 in its second intron is found to cause the genetic variation between single and double flowers (Figure 3, and Supplemental Figures 7–13). In single-flowered trees, some individuals with no Xsag1-LINE1-1 insertion of XsAG1 are considered as wild-type “AA,” while others with heterozygous Xsag1-LINE1-1 insertion are “Aa.” And all double-flowered individuals are homozygous of the Xsag1-LINE1-1 insertion site, thus are considered as “aa” with mutant phenotypes. Therefore, like petunia (Petunia hybrida; Liu et al., 2016), as a heritable molecular marker, this Xsag1-LINE1-1 insertion could be used in horticultural settings to distinguish young yellowhorn seedlings that have not yet blossomed.

Mutations in C-function genes, which cause loss of reproductive function, can occur naturally in the wild and are often considered an evolutionary “dead end” (Galimba et al., 2012). The mutation of flower structure had no effect on the growth and adaptability of yellowhorn double-flowered trees; however, the disappearance of reproductive organs is unfavorable to its reproduction and evolution. The presence of yellowhorn double-flowered trees in the natural population was attributed to single-flowered trees with heterozygous Xsag1-LINE1-1 insertion of XsAG1, such as CF74, providing strong genetic evidence. And CF74, in particular, may be the oldest tree with the ability to produce double-flowered progeny (Supplemental Figure 15). Seeds from this heterozygous Xsag1-LINE1-1 insertion tree spread to two surrounding forest nurseries and then throughout the country. Because the origin of yellowhorn is inseparable from the spread of Tibetan Buddhism, it is likely that many years ago Buddhist monks dispersed these and other yellowhorn seeds around China. Thus, the serendipitous evolution of yellowhorn has become forever intertwined with the cultural and historical changes in China over many years as these trees have witnessed the passage of time.

Xsag1-LINE1-1 insertion in the second intron of XsAG1 reshapes yellowhorn flower morphology

As a source of genetic variation, TE insertions are different from ordinary structural variation because they contain some regulatory elements that may have an impact on gene expression. A TE fragment inserted in the coding region could change gene structure, while in the promoter or intron region may change gene expression, both of which can reshape traits (Chuong et al., 2017). Numerous studies have shown that TE insertions in plant genomes, such as Arabidopsis, Chinese cabbage (Brassica rapa), oilseed rape (Brassica napus), Capsella rubella, tomato (Solanum lycopersicum), grape (Vitis vinifera), blood orange (Citrus sinensis), rice, maize, and so on, play an important role in phenotypic variation, adaptation, and domestication, which lead to variation in fruit skin color, inflorescence length, seed weight, or other important agronomic traits (Selinger and Chandler, 1999, 2001; Kobayashi et al., 2004; Xiao et al., 2008; Butelli et al., 2012; Stuart et al., 2016; Carpentier et al., 2019; Dominguez et al., 2020; Liu et al., 2020; Song et al., 2020; Cai et al., 2022).

As potential regulators of floral development, TE insertions can exert specific effects on key transcription factors to alter floral developmental processes or cause homeotic transformation of floral organs. In apple, LTR retrotransposons are inserted into the fourth or sixth intron of MdPI, a Class B MADS-box gene, resulting in mutants that form only apetalous flowers (Yao et al., 2001). Petal loss in Nigella apetalous mutants may be due to insertion of miniature inverted repeat transposable elements and disruption of NidaAP3-3 expression (Zhang et al., 2013). Other examples of transposons or similar fragments inserted into C-function genes leading to the conversion of pistils and stamens into petals have also been reported in different species (Nitasaka, 2003; Galimba et al., 2012; Ma et al., 2018). Here, we found that the Xsag1-LINE1-1 fragment inserted in XsAG1 in the yellowhorn genome is a LINE1 transposon (Supplemental Figure 10), and this fragment is responsible for floral homeotic mutation in yellowhorn, providing evidence for the potential role of, LINE1 elements in flower trait reshape and plant evolution. The 25 SNPs between Xsag1-LINE1-1 and its close relative Lachesis_group4:1,196,133-1,202,795 (Type I) suggest that their divergence time may be ∼6.2 × 10−4 mya (r = K/2T = 6.2 × 10−4 mya, K = 0.041, T = 33.07 mya) (Bi et al., 2019; Nishimaki and Sato, 2019), and its transposition time is relatively recent, suggesting the possibility of its active transposition (Supplemental Figure 12). Autonomous transposition of LINEs has been demonstrated in many other systems such as fish and mammals (Ichiyanagi et al., 2007), and we hypothesize that in yellowhorn genome, a LINE1 may have transposable capacity and thereby generate this homeotic mutation; its origin, mechanism of insertion, and whether it is still active are interesting subjects for future research.

Yellowhorn double flower represents a homeotic mutant with XsAG1 loss of function

Double flowers have been well studied in many other species, especially in ornamental plants. The molecular mechanisms controlling the development of double flowers in such diverse species such as rose (Rosa hybrida), Orchidaceae, carnation (Dianthus caryophyllus), petunia, tulip (Tulipa gesneriana), Japanese gentian (Gentiana scabra, Gentiana triflora, and their interspecific hybrids), Prunus lannesiana, and Th. thalictroides have been revealed through identification of ABC model homologuos genes (Kanno et al., 2007; Mondragon-Palomino and Theissen, 2008; Dubois et al., 2010; Galimba et al., 2012; Liu et al., 2013; Nakatsuka et al., 2015; Gattolin et al., 2020). The appearance of double flowers is generally related to the abnormal function of Class B or C genes (Bowman et al., 2012; Irish, 2017). Here, we found a homeotic mutant in yellowhorn had double flowers with stamens and pistils replaced by petals, and its XsAG1 expression was decreased due to insertion of Xsag1-LINE1-1. A simple explanation for this mutant phenotype is a reduction in the amount of functional XsAG1 mRNA caused by the Xsag1-LINE1-1 insertion (Figure 3).

The second intron of AG confers tissue-specific expression, and in Arabidopsis, it is the target of positive and negative regulatory trans-acting factors (Deyholos and Sieburth, 2000). Examples of selective splicing resulting from disruption of the original splicing site by transposon insertion in the intronic region have rarely been reported in plants (Varagona et al., 1992; Supplemental Table 10). Here, we found that Xsag1-LINE1-1 inserted into XsAG1 Intron 2 caused a decrease in the level of the original transcript, and a small number of aberrant transcripts including Xsag1-LINE1-1 appeared in double flowers (Figure 3 and Supplemental Figure 17). Expression of Class C genes is supposed to function in a positive autoregulatory feedback loop (Gomez-Mena et al., 2005), so providing an explanation for the reduced XsAG1 transcript levels of double flowers. Another explanation, not incompatible with the previous one, is that the Xsag1-LINE1-1 insertion changes the original splicing site, causing the aberrant transcripts to be unstable and thus degraded. Therefore, we suggested that yellowhorn double-flowered trees are a homeotic mutant with the loss of XsAG1 function.

Although the double flowers of yellowhorn can also exhibit flower-within-flower phenotypes, they tend to exhibit a Se-Pe-Pe-Pe whorled structure and have lost the special horn-like appendages found in single yellowhorn flowers (Figure 1 and Supplemental Figure 1). The high expression of XsCRC and the mucus on the surface of horn-like appendages suggested that it belonged to a nectarium (Supplemental Figure 6 and Figure 4C). Previous studies have shown that CRC can be activated by AG (Alvarez and Smyth, 1999; Gomez-Mena et al., 2005; Lee et al., 2005a, 2005b), considering the formation of yellowhorn double flowers is caused by XsAG1 transcription level decreasing, and XsCRC expression could not be detected in double flowers (Figure 4, B and C), so loss of horn-like appendages in double flowers may also be affected by XsAG1, also providing some insights into the origin and evolution of nectary organs.

XsAG1 exerts conserved Class C function and its loss of function affects other ABC genes

Flower development is an important process in the growth and development of the forest trees; however, because of their long juvenile phase and lack of mutagenized populations, mutants with floral phenotypes are not as easily available for forest trees. Here, we used a forward genetic approach and cloned the Class C functional gene XsAG1 (Figures 2 and 3). In Arabidopsis, AGAMOUS function is manifested in the identity of the reproductive organs and floral determinacy (Yanofsky et al., 1990; Mizukami and Ma, 1997; Ito et al., 2007; Sun et al., 2009), and its conserved function has been demonstrated in many species (Causier et al., 2010; Irish, 2017; Shan et al., 2019; Pelayo et al., 2021). In yellowhorn, XsAG1-specific expression in stamens, pistils, and horn-like appendages suggests that it may have a conserved function as a Class C gene (Figure 4). Overexpressing XsAG1 in the ag-1/ag-1 transgenic plants did not result in a single flower phenotype either with 35S or the AtAG native promoter; however, according to the developed stamens and carpel-like structures on sepals, XsAG1 has a conserved Class C function (Figure 5C). In the Ler:35S::XsAG1-CDS and Ler:proAtAG::XsAG1-CDS transgenic plants, single flowers did exhibit variation related to small petals and longer stamens and the phenotype is similar to AtAG overexpression in wild-type transgenic plants, in which petals are lost (Castillejo et al., 2005; Airoldi et al., 2010), indicating that XsAG1 plays a similar role in determining reproductive organ identity to AtAG in Arabidopsis and its function is conserved in trees (Figure 5).

In Arabidopsis, only one Class C gene AGAMOUS specifies the development of reproductive organs, while two genes, PLENA (PLE) and FARINELLI (FAR) in Antirrhinum, contribute differently to male and female floral organ development (Yanofsky et al., 1990; Davies et al., 1999; Airoldi et al., 2010; Irish, 2017). XsAG1 homologs XsSTK and XsSHP are identified, and phylogenetic tree shows that XsSHP and XsAG1 belong two evolutionary clades of AGAMOUS and that XsAG1, AtAG, and AmFAR are located in the same clade, while XsSHP and AmPLE are in another clade (Supplemental Figures 8 and 20). Both XsAG1 and XsSHP are expressed in stamens, carpels, and horn-like appendages, their similar expression pattern suggesting similar functions (Figure 4 and Supplemental Figure 8). The expression pattern of floral organ identity genes in different whorls is maintained by mutual regulation, and the interaction of AG with other transcription factors is considered to be essential for its C function (Honma and Goto, 2001; Gomez-Mena et al., 2005; Kaufmann et al., 2010). In yellowhorn, the transcript abundances of other ABC genes, especially XsPI and XsSEP, change significantly when XsAG1 expression decrease (Figure 4). These results also provide support for further understanding and elucidating the ABC model and flower development in yellowhorn.

Here, we offer a possible explanation for the formation of yellowhorn double flowers, a hopping LINE1 transposon provides the condition for its production (Supplemental Figure 20 and Figure 5C). Due to the homozygous Xsag1-LINE1-1 insertion into the second intron of XsAG1, its functional transcript abundance was sharply lower than that of single flowers. The loss of XsAG1 function leads to the loss of floral determination, so that the double flower phenotype is produced in yellowhorn. Meanwhile, the functional analysis of XsAG1 shows that it has strong conservation in the reproductive organ identity. Other ABC homologs in yellowhorn are identified and explored the temporal and spatial specificity of their expression, providing important experimental evidence for the importance of Class C genes in forest tree flower development. Furthermore, the TE Xsag1-LINE1-1 insertion affects the floral phenotype of yellowhorn, which provides evidence for naturally occurring mutations mediated by transposons in forest trees.

Materials and methods

Plant materials

Yellowhorn (X. sorbifolium) trees sampled for the present study are growing in three different regions across China. The Arabidopsis (Arab. thaliana) mutants and transgenic lines are grown under laboratory conditions. See supplementary information for details.

BSA sequencing and analysis

Genomic DNAs were extracted from fresh leaves using the DNAsecure Plant Kit (Tiangen Biotech, Beijing, China). The S-pool and D-pool representing bulks of single- and double-flowered individuals from natural populations, respectively, were constructed by mixing fresh leaves collected from 98 single-flowered individuals (26 from NEFU campus, 36 from Weifang City, and 36 from Chifeng City) and 50 double-flowered individuals (14 from Weifang City, and 36 from Chifeng City). G values (Magwene et al., 2011) were calculated using the QTLseqr package (Mansfeld and Grumet, 2018) in R (https://www.r-project.org/). We filtered out heterozygous sites of the whole genome and any SNPs located on fragments that were not linked to chromosomes. SNPs in each 1-Mb window were used to calculate the G value for a given window. Plots were created using plotQTLStats with the argument q = 1e-6.

Assembly of the EVM0013090 sequence with the Xsag1-LINE1-1 inserted into XsAG1

Based on the BSA results, we used De Bruijn graphs (DBGs; Bresler et al., 2013) to assemble de novo the entire EVM0013090 sequence from yellowhorn that carries the LINE1 insertion. We used the whole genome sequencing (WGS) from double-flowered yellowhorn samples from which adapters and low-quality sequences had been removed. The sequence near the putative TE insertion site in the reference genome was divided into 17-bp oligomers (Kmer = 17) as the basic nodes of the DBG. WGS data were also divided into 17-bp oligomers and any nodes that overlapped by 16 bp (k-1) would be connected.

Analysis and classification of the exogenous inserted fragment in XsAG1 as a LINE1 transposon

We extracted sequence data from the interval containing the Xsag1-LINE1-1 in the double-flowered yellowhorn genome and annotated its sequence context using Repbase (https://www.girinst.org/censor/index.php; Kohany et al., 2006) to identify the inserted sequence. We performed sequence alignment between the Xsag1-LINE1-1 in XsAG1 and other LINE1 sequences annotated by RepeatMasker (http://repeatmasker.org) with high sequence similarity to the Xsag1-LINE1-1in XsAG1. We performed structure analysis of 279 homologous transposons using the Conserved Domain Database at NCBI (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi; Lu et al., 2020).

RNA in situ hybridization

Inflorescence of single- and double-flowered trees at various developmental stages were fixed overnight in Rnase-free FAA (3.7% [v/v] formaldehyde, 5% [v/v] acetic acid, and 50% [v/v] ethanol) at 4°C and then embedded in Paraplast (Sigma, Shanghai, China). Gene-specific fragments were used as templates for synthesis of antisense and sense digoxigenin-labelled RNA probes using the DIG RNA labelling kit (Roche, Shanghai, China; Supplemental Table 10 and 11). Sectioning and subsequent treatments were performed as described. Images were captured with a Leica DM5000 B microscope.

Plasmid construction and generation of transgenic plants

XsAG1 CDS from single-flowered yellowhorn trees were amplified from cDNA using Q5 High-fidelity DNA polymerase (NEB, Beijing, China) and inserted into pEASY®-Blunt Zero Cloning Kit (Transgen Biotech, Beijing, China). Then all the sequences were cloned into SalI-cut-pMCP1 vectors to create 35S::XsAG1 and proAtAG::XsAG1 constructs. All vectors used in the present study were confirmed by Sanger sequencing in Tsingke Biotech. These overexpression vectors were then transformed into Arabidopsis AG/ag-1 mutants using the Agrobacterium tumefaciens-mediated floral-dip method (Zhang et al., 2013). Putative transgenic lines were confirmed using PCR and RT-PCR with transgene-specific primers.

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers PRJNA906480.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Inflorescences, flowers, and floral organ structures in yellowhorn trees.

Supplemental Figure S2. Different development stages of single flower and double flower bud differentiation in X. sorbifolium.

Supplemental Figure S3. Different development stages after primordium differentiation of single flowers in X. sorbifolium.

Supplemental Figure S4. Different development stages after primordium differentiation of double flowers in X. sorbifolium.

Supplemental Figure S5. Petal morphology of double flowers in X. sorbifolium.

Supplemental Figure S6. Development and morphology of horn-like appendages of single flowers in X. sorbifolium.

Supplemental Figure S7. FPKM values of 269 genes in single and double flower bud samples.

Supplemental Figure S8. EVM0013090 in X. sorbifolium is an ortholog of AGAMOUS in Arabidopsis thaliana.

Supplemental Figure S9. Integrative genomics viewer screenshot of the EVM0013090.1 gene in the genomes of single- and double-flowered samples compared with the reference X. sorbifolium genome.

Supplemental Figure S10. Sequences and structure of Xsag1-LINE1-1 insertion in XsAG1 in double-flowered X. sorbifolium trees.

Supplemental Figure S11. Sanger sequencing results for the identification of Xsag1-LINE1-1.

Supplemental Figure S12. Phylogeny of LINE1 transposons homologous to the Xsag1-LINE1-1 variants in XsAG1 in X. sorbifolium.

Supplemental Figure S13. PCR analysis of Xsag1-LINE1-1-based markers (Primers GF, GR, and TF) in all double-flowered X. sorbifolium trees sampled from different regions of China.

Supplemental Figure S14. PCR analysis of Xsag1-LINE1-1-based markers (Primers GF, GR, and TF) in single-flowered X. sorbifolium trees sampled from different regions of China.

Supplemental Figure S15. The ancient X. sorbifolium single-flowered tree CF74 might be the progenitor of modern double-flowered trees.

Supplemental Figure S16. The existence of alternatively spliced XsAG1 transcripts in the double flower samples shown by RT-PCR using primers designed to amplify different regions around the Xsag1-LINE1-1 insertion.

Supplemental Figure S17. Sequences of XsAG1 transcripts and the alternative splice forms expressed in X. sorbifolium double-flowered trees.

Supplemental Figure S18. The alternative splice sites at the 5′ end of the Xsag1-LINE1-1 insertion in XsAG1 that result in different transcripts are in the same position in all samples from double-flowered trees.

Supplemental Figure S19. In situ hybridization of XsAG1 using the antisense probe in inflorescences.

Supplemental Figure S20. The expression pattern of XsAG1 and other ABC model genes in single and double flowers.

Supplemental Figure S21. Phenotypes of proAtAG::XsAG1-CDS T0 generation transgenic plants in the Arabidopsis ag-1 background.

Supplemental Figure S22. Detection of XsAG1 transgene expression in proAtAG::XsAG1-CDS (T0-line23,32,39,47) and 35S::XsAG1-CDS (T1-line14, 83) transgenic Arabidopsis plants. Untransformed Ler wild-type and ag-1/ag-1 plants were used as controls.

Supplemental Table S1. Distribution of single- and double-flowered yellowhorn tree populations in China.

Supplemental Table S2. Statistical table of sequencing data for each BSA mixed pool.

Supplemental Table S3. Chromosome coverage depth of each BSA mixed pool sample.

Supplemental Table S4. Associate interval location information.

Supplemental Table S5. Statistical table of resequencing data for eight single-flowered and three double-flowered trees.

Supplemental Table S6. Statistical table of RNA-seq data for three single and three double bud samples.

Supplemental Table S7. FPKM values of 269 genes within the interval in single and double flower RNA-seq data.

Supplemental Table S8. The amino acid sequences and their GenBank accession number to generate the phylogenetic tree.

Supplemental Table S9. ABCDE model homologous genes in yellowhorn.

Supplemental Table S10. Natural mutants of C-function genes.

Supplemental Table S11. Sequences of oligonucleotides used in this study.

Supplemental Data Set S1. Statistics of SNP mutation loci within the association interval.

Supplemental Data Set S2. Candidate genes and their functional annotations within the correlation interval.

Supplemental Data Set S3. Identical variation types (SNPs, InDels, and SV) and statistics of double-flower sequencing samples in the ∼2.7 Mb associated interval.

Supplemental Data Set S4. Sequence alignment of LINE1 transposons highly similar to Xsag1-LINE1-1 fragments.

Supplemental Data Set S5. The sequence information of LINE1 transposon whose similarities with TE are >90%.

Supplemental Data Set S6. SNP and information statistics of gene EVM0013090.1.

Supplementary Material

kiac571_Supplementary_Data
Click here for additional data file. (18.5MB, zip)

Acknowledgments

The authors thank Yamin Sun for assistance with the processing of BSA sequencing data and Xigang Liu for kindly providing AG/ag-1 lines and critical comments on this manuscript. They also thank Libing Wang, Quanxin Bi, Shouke Li, Xiaoqin Xiang, Yali Wang, and Yindoleng Sai and the Kundu Economic Forest Farm for kindly providing yellowhorn materials. The authors also thank Jiesong Tang and Jingpeng Sun for helping in some experiments.

Contributor Information

Hanhui Wang, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Yanan Lu, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China; College of Life Science, Northeast Forestry University, Harbin 150040, China.

Tianxu Zhang, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Zhi Liu, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Li Cao, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China; College of Life Science, Northeast Forestry University, Harbin 150040, China.

Qiaoying Chang, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Yueying Liu, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China; College of Life Science, Northeast Forestry University, Harbin 150040, China.

Xin Lu, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China; Chifeng Research Institute of Forestry Science, No.77 Yulong Street, Xincheng District, Songshan District, Chifeng City, Inner Mongolia, China; National Forestry and Grassland Shiny-Leaved Yellowhorn Engineering and Technology Research Center, No.77 Yulong Street, Xincheng District, Songshan District, Chifeng City, China.

Song Yu, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Huiyu Li, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Jing Jiang, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Guifeng Liu, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Heike W Sederoff, Department of Plant and Microbial Biology, North Carolina State University, Raleigh, North Carolina 27695, USA.

Ronald R Sederoff, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China; Forest Biotechnology Group, Department of Forestry and Environmental Resources, North Carolina State University, Raleigh, North Carolina 27695, USA.

Qingzhu Zhang, College of Life Science, Northeast Forestry University, Harbin 150040, China.

Zhimin Zheng, State Key Laboratory of Tree Genetics and Breeding, College of Forestry, Northeast Forestry University, Harbin 150040, China.

Funding

This work was supported by grants from Innovation Project of State Key Laboratory of Tree Genetics and Breeding (2019A02) and the Central University Basic Research Fund of China (2572017DA03, 2572018BS01, 2572020DY24).

References

  1. Abdirashid  H, Lenhard  M (2020) Say it with double flowers. J Exp Bot  71(9): 2469–2471 [DOI] [PubMed] [Google Scholar]
  2. Airoldi  CA, Bergonzi  S, Davies  B (2010) Single amino acid change alters the ability to specify Male or female organ identity. Proc Natl Acad Sci U S A  107(44): 18898–18902 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Alejandra Mandel  M, Gustafson-Brown  C, Savidge  B, Yanofsky  MF (1992) Molecular characterization of the Arabidopsis floral homeotic gene APETALA1. Nature  360(6401): 273–277 [DOI] [PubMed] [Google Scholar]
  4. Alvarez  J, Smyth  DR (1999) CRABS CLAW and SPATULA, two Arabidopsis genes that control carpel development in parallel with AGAMOUS. Development  126(11): 2377–2386 [DOI] [PubMed] [Google Scholar]
  5. Bi  Q, Zhao  Y, Du  W, Lu  Y, Gui  L, Zheng  Z, Yu  H, Cui  Y, Liu  Z, Cui  T, et al. (2019) Pseudomolecule-level assembly of the Chinese oil tree yellowhorn (Xanthoceras sorbifolium) genome. Gigascience  8(6): giz070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bowman  JL, Smyth  DR (1999) CRABS CLAW, a gene that regulates carpel and nectary development in Arabidopsis, encodes a novel protein with zinc finger and helix-loop-helix domains. Development  126(11): 2387–2396 [DOI] [PubMed] [Google Scholar]
  7. Bowman  JL, Smyth  DR, Meyerowitz  EM (1991) Genetic interactions among floral homeotic genes of Arabidopsis. Development  112(1): 1–20 [DOI] [PubMed] [Google Scholar]
  8. Bowman  JL, Smyth  DR, Meyerowitz  EM (2012) The ABC model of flower development: then and now. Development  139(22): 4095–4098 [DOI] [PubMed] [Google Scholar]
  9. Butelli  E, Licciardello  C, Zhang  Y, Liu  J, Mackay  S, Bailey  P, Reforgiato-Recupero  G, Martin  C (2012) Retrotransposons control fruit-specific, cold-dependent accumulation of anthocyanins in blood oranges. Plant Cell  24(3): 1242–1255 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cai  X, Lin  R, Liang  J, King  GJ, Wu  J, Wang  X (2022) Transposable element insertion: a hidden major source of domesticated phenotypic variation in Brassica rapa. Plant Biotechnol J  20(7): 1298–1310 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Carpentier  MC, Manfroi  E, Wei  FJ, Wu  HP, Lasserre  E, Llauro  C, Debladis  E, Akakpo  R, Hsing  YI, Panaud  O (2019) Retrotranspositional landscape of Asian rice revealed by 3000 genomes. Nat Commun  10(1): 24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Castillejo  C, Romera-Branchat  M, Pelaz  S (2005) A new role of the Arabidopsis SEPALLATA3 gene revealed by its constitutive expression. Plant J  43(4): 586–596 [DOI] [PubMed] [Google Scholar]
  13. Causier  B, Schwarz-Sommer  Z, Davies  B (2010) Floral organ identity: 20 years of ABCs. Semin Cell Dev Biol  21(1): 73–79 [DOI] [PubMed] [Google Scholar]
  14. Chuong  EB, Elde  NC, Feschotte  C (2017) Regulatory activities of transposable elements: from conflicts to benefits. Nat Rev Genet  18(2): 71–86 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Colombo  L, Battaglia  R, Kater  MM (2008) Arabidopsis ovule development and its evolutionary conservation. Trends Plant Sci  13(8): 444–450 [DOI] [PubMed] [Google Scholar]
  16. Davies B  MP, Keck  E, Saedler  H, Sommer  H, Schwarz-Sommer  Z (1999) PLENA And FARINELLI: redundancy and regulatory interactions between two antirrhinum MADS-box factors controlling flower development. EMBO J  18(14): 4023–4034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Deyholos  MK, Sieburth  LE (2000) Separable whorl-specific expression and negative regulation by enhancer elements within the AGAMOUS second intron. Plant Cell  12(10): 1799–1810 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ditta  G, Pinyopich  A, Robles  P, Pelaz  S, Yanofsky  MF (2004) The SEP4 gene of Arabidopsis thaliana functions in floral organ and meristem identity. Curr Biol  14(21): 1935–1940 [DOI] [PubMed] [Google Scholar]
  19. Dominguez  M, Dugas  E, Benchouaia  M, Leduque  B, Jimenez-Gomez  JM, Colot  V, Quadrana  L (2020) The impact of transposable elements on tomato diversity. Nat Commun  11(1): 4058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Dougherty  L, Singh  R, Brown  S, Dardick  C, Xu  K (2018) Exploring DNA variant segregation types in pooled genome sequencing enables effective mapping of weeping trait in Malus. J Exp Bot  69(7): 1499–1516 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dubin  MJ, Mittelsten Scheid  O, Becker  C (2018) Transposons: a blessing curse. Curr Opin Plant Biol  42: 23–29 [DOI] [PubMed] [Google Scholar]
  22. Dubois  A, Raymond  O, Maene  M, Baudino  S, Langlade  NB, Boltz  V, Vergne  P, Bendahmane  M (2010) Tinkering with the C-function: a molecular frame for the selection of double flowers in cultivated roses. PLoS One  5(2): e9288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Galimba  KD, Tolkin  TR, Sullivan  AM, Melzer  R, Theissen  G, Di Stilio  VS (2012) Loss of deeply conserved C-class floral homeotic gene function and C- and E-class protein interaction in a double-flowered ranunculid mutant. Proc Natl Acad Sci U S A  109(34): E2267–E2275 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gattolin  S, Cirilli  M, Chessa  S, Stella  A, Bassi  D, Rossini  L (2020) Mutations in orthologous PETALOSA TOE-type genes cause a dominant double-flower phenotype in phylogenetically distant eudicots. J Exp Bot  71(9): 2585–2595 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Gomez-Mena  C, de Folter  S, Costa  MM, Angenent  GC, Sablowski  R (2005) Transcriptional program controlled by the floral homeotic gene AGAMOUS during early organogenesis. Development  132(3): 429–438 [DOI] [PubMed] [Google Scholar]
  26. Gustafson-Brown  C, Savidge  B, Yanofsky  MF (1994) Regulation of the arabidopsis floral homeotic gene APETALA1. Cell  76(1): 131–143 [DOI] [PubMed] [Google Scholar]
  27. Honma  T, Goto  K (2001) Complexes of MADS-box proteins are sufficient to convert leaves into floral organs. Nature  409(6819): 525–529 [DOI] [PubMed] [Google Scholar]
  28. Huang  Z, Shi  T, Zheng  B, Yumul  RE, Liu  X, You  C, Gao  Z, Xiao  L, Chen  X (2017) APETALA2 antagonizes the transcriptional activity of AGAMOUS in regulating floral stem cells in Arabidopsis thaliana. New Phytol  215(3): 1197–1209 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Ichiyanagi  K, Nishihara  H, Duvernell  DD, Okada  N (2007) Acquisition of endonuclease specificity during evolution of L1 retrotransposon. Mol Biol Evol  24(9): 2009–2015 [DOI] [PubMed] [Google Scholar]
  30. Irish  V (2017) The ABC model of floral development. Curr Biol  27(17): R887–R890 [DOI] [PubMed] [Google Scholar]
  31. Ito  T, Ng  KH, Lim  TS, Yu  H, Meyerowitz  EM (2007) The homeotic protein AGAMOUS controls late stamen development by regulating a jasmonate biosynthetic gene in Arabidopsis. Plant Cell  19(11): 3516–3529 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kanno  A, Nakada  M, Akita  Y, Hirai  M (2007) Class B gene expression and the modified ABC model in nongrass monocots. Sci World J  7: 268–279 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kaufmann  K, Pajoro  A, Angenent  GC (2010) Regulation of transcription in plants: mechanisms controlling developmental switches. Nat Rev Genet  11(12): 830–842 [DOI] [PubMed] [Google Scholar]
  34. Kimura  M (1980) A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol  16(2): 111–120 [DOI] [PubMed] [Google Scholar]
  35. Kobayashi  S, Goto-Yamamoto  N, Hirochika  H (2004) Retrotransposon-induced mutations in grape skin color. Science  304(5673): 982. [DOI] [PubMed] [Google Scholar]
  36. Kohany  O, Gentles  AJ, Hankus  L, Jurka  J (2006) Annotation, submission and screening of repetitive elements in Repbase: RepbaseSubmitter and Censor. BMC Bioinformatics  7(1): 474. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Lee  JY, Baum  SF, Alvarez  J, Patel  A, Chitwood  DH, Bowman  JL (2005a) Activation of CRABS CLAW in the nectaries and carpels of Arabidopsis. Plant Cell  17(1): 25–36 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Lee  JY, Baum  SF, Oh  SH, Jiang  CZ, Chen  JC, Bowman  JL (2005b) Recruitment of CRABS CLAW to promote nectary development within the eudicot clade. Development  132(22): 5021–5032 [DOI] [PubMed] [Google Scholar]
  39. Lenhard  M, Bohnert  A, Jürgens  G, Laux  T (2001) Termination of stem cell maintenance in Arabidopsis floral meristems by interactions between WUSCHEL and AGAMOUS. Cell  105(6): 805–814 [DOI] [PubMed] [Google Scholar]
  40. Li  W, Lu  Q, Li  X, Liu  H, Sun  L, Lu  X, Zhao  Y, Liu  P (2020) Anti-Alzheimer's disease activity of secondary metabolites from Xanthoceras sorbifolia Bunge. Food Funct  11(3): 2067–2079 [DOI] [PubMed] [Google Scholar]
  41. Liao  N, Hu  Z, Li  Y, Hao  J, Chen  S, Xue  Q, Ma  Y, Zhang  K, Mahmoud  A, Ali  A, et al. (2020) Ethylene-responsive factor 4 is associated with the desirable rind hardness trait conferring cracking resistance in fresh fruits of watermelon. Plant Biotechnol J  18(4): 1066–1077 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Liljegren  SJ, Ditta  GS, Eshed  Y, Savidge  B, Bowman  JL, Yanofsky  MF (2000) SHATTERPROOF MADS-box genes control seed dispersal in Arabidopsis. Nature  404(6779): 766–770 [DOI] [PubMed] [Google Scholar]
  43. Lisch  D (2013) How important are transposons for plant evolution?  Nat Rev Genet  14(1): 49–61 [DOI] [PubMed] [Google Scholar]
  44. Liu  C, He  Y, Gou  T, Li  X, Ning  G, Bao  M (2016) Identification of molecular markers associated with the double flower trait in Petunia hybrida. Sci Hortic  206: 43–50 [Google Scholar]
  45. Liu  X, Kim  YJ, Muller  R, Yumul  RE, Liu  C, Pan  Y, Cao  X, Goodrich  J, Chen  X (2011) AGAMOUS terminates floral stem cell maintenance in Arabidopsis by directly repressing WUSCHEL through recruitment of Polycomb Group proteins. Plant Cell  23(10): 3654–3670 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Liu  Z, Zhang  D, Liu  D, Li  F, Lu  H (2013) Exon skipping of AGAMOUS homolog PrseAG in developing double flowers of Prunus lannesiana (Rosaceae). Plant Cell Rep  32(2): 227–237 [DOI] [PubMed] [Google Scholar]
  47. Liu  J, Zhou  R, Wang  W, Wang  H, Qiu  Y, Raman  R, Mei  D, Raman  H, Hu  Q (2020) A copia-like retrotransposon insertion in the upstream region of the SHATTERPROOF1 gene, BnSHP1.A9, is associated with quantitative variation in pod shattering resistance in oilseed rape. J Exp Bot  71(18): 5402–5413 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Ma  J, Shen  X, Liu  Z, Zhang  D, Liu  W, Liang  H, Wang  Y, He  Z, Chen  F (2018) Isolation and characterization of AGAMOUS-like genes associated with double-flower morphogenesis in Kerria japonica (Rosaceae). Front Plant Sci  9: 959. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Mansfeld  BN, Grumet  R (2018) QTLseqr: an R package for bulk segregant analysis with next-generation sequencing. Plant Genome  11(2): 1–5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Mizukami  Y, Ma  H (1997) Determination of Arabidopsis floral meristem identity by AGAMOUS. Plant Cell  9(3): 393–408 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Mondragon-Palomino  M, Theissen  G (2008) MADS About the evolution of orchid flowers. Trends Plant Sci  13(2): 51–59 [DOI] [PubMed] [Google Scholar]
  52. Nakatsuka  T, Saito  M, Yamada  E, Fujita  K, Yamagishi  N, Yoshikawa  N, Nishihara  M (2015) Isolation and characterization of the C-class MADS-box gene involved in the formation of double flowers in Japanese gentian. BMC Plant Biol  15(1): 182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Ng  M, Yanofsky  MF (2001) Function and evolution of the plant MADS-box gene family. Nat Rev Genet  2(3): 186–195 [DOI] [PubMed] [Google Scholar]
  54. Nishimaki  T, Sato  K (2019) An extension of the kimura two-parameter model to the natural evolutionary process. J Mol Evol  87(1): 60–67 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Nitasaka  E (2003) Insertion of an En/Spm-related transposable element into a floral homeotic gene DUPLICATED causes a double flower phenotype in the Japanese morning glory. Plant J  36(4): 522–531 [DOI] [PubMed] [Google Scholar]
  56. O'Maoileidigh  DS, Graciet  E, Wellmer  F (2014) Gene networks controlling Arabidopsis thaliana flower development. New Phytol  201(1): 16–30 [DOI] [PubMed] [Google Scholar]
  57. Pandey  MK, Khan  AW, Singh  VK, Vishwakarma  MK, Shasidhar  Y, Kumar  V, Garg  V, Bhat  RS, Chitikineni  A, Janila  P, et al. (2017) QTL-seq approach identified genomic regions and diagnostic markers for rust and late leaf spot resistance in groundnut (Arachis hypogaea L). Plant Biotechnol J  15(8): 927–941 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Pelayo  MA, Yamaguchi  N, Ito  T (2021) One factor, many systems: the floral homeotic protein AGAMOUS and its epigenetic regulatory mechanisms. Curr Opin Plant Biol  61: 102009. [DOI] [PubMed] [Google Scholar]
  59. Pinyopich  A, Ditta  GS, Savidge  B, Liljegren  SJ, Baumann  E, Wisman  E, Yanofsky  MF (2003) Assessing the redundancy of MADS-box genes during carpel and ovule development. Nature  424(6944): 85–88 [DOI] [PubMed] [Google Scholar]
  60. Schwarz-Sommer  Z, Huijser  P, Nacken  W, Saedler  H, Sommer  H (1990) Genetic control of flower development by homeotic genes in Antirrhinum majus. Science  250(4983): 931–936 [DOI] [PubMed] [Google Scholar]
  61. Selinger  DA, Chandler  VL (1999) Major recent and independent changes in levels and patterns of expression have occurred at the b gene, a regulatory locus in maize. Proc Natl Acad Sci U S A  96(26): 15007–15012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Selinger  DA, Chandler  VL (2001) B-Bolivia, an allele of the maize b1 gene with variable expression, contains a high copy retrotransposon-related sequence immediately upstream. Plant Physiol  125(3): 1363–1379 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Shan  H, Cheng  J, Zhang  R, Yao  X, Kong  H (2019) Developmental mechanisms involved in the diversification of flowers. Nat Plants  5(9): 917–923 [DOI] [PubMed] [Google Scholar]
  64. Shen  Z, Zhang  K, Ao  Y, Ma  L, Duan  J (2018) Evaluation of biodiesel from Xanthoceras sorbifolia Bunge seed kernel oil from 13 areas in China. J For Res  30(3): 869–877 [Google Scholar]
  65. Smaczniak  C, Immink  RG, Angenent  GC, Kaufmann  K (2012) Developmental and evolutionary diversity of plant MADS-domain factors: insights from recent studies. Development  139(17): 3081–3098 [DOI] [PubMed] [Google Scholar]
  66. Song  JM, Guan  Z, Hu  J, Guo  C, Yang  Z, Wang  S, Liu  D, Wang  B, Lu  S, Zhou  R, et al. (2020) Eight high-quality genomes reveal pan-genome architecture and ecotype differentiation of Brassica napus. Nat Plants  6(1): 34–45 [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Stuart  T, Eichten  SR, Cahn  J, Karpievitch  YV, Borevitz  JO, Lister  R (2016) Population scale mapping of transposable element diversity reveals links to gene regulation and epigenomic variation. Elife  5: e20777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Sun  B, Xu  Y, Ng  KH, Ito  T (2009) A timing mechanism for stem cell maintenance and differentiation in the Arabidopsis floral meristem. Genes Dev  23(15): 1791–1804 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Varagona  MJ, Purugganan  M, Wessler  SR (1992) Alternative splicing induced by insertion of retrotransposons into the maize waxy gene. Plant Cell  4(7): 811–820 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Wang  C, Tang  S, Zhan  Q, Hou  Q, Zhao  Y, Zhao  Q, Feng  Q, Zhou  C, Lyu  D, Cui  L, et al. (2019) Dissecting a heterotic gene through GradedPool-seq mapping informs a rice-improvement strategy. Nat Commun  10(1): 2982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Xiao  H, Jiang  N, Schaffner  E, Stockinger  EJ, van der Knaap  E (2008) A retrotransposon-mediated gene duplication underlies morphological variation of tomato fruit. Science  319(5869): 1527–1530 [DOI] [PubMed] [Google Scholar]
  72. Xiao  W, Wang  Y, Zhang  P, Li  N, Jiang  S, Wang  JH, Huang  J, Li  X (2013) Bioactive barrigenol type triterpenoids from the leaves of Xanthoceras sorbifolia Bunge. Eur J Med Chem  60: 263–270 [DOI] [PubMed] [Google Scholar]
  73. Xue  H, Shi  T, Wang  F, Zhou  H, Yang  J, Wang  L, Wang  S, Su  Y, Zhang  Z, Qiao  Y, et al. (2017) Interval mapping for red/green skin color in Asian pears using a modified QTL-seq method. Hortic Res  4(1): 17053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Yanofsky  MF, Ma  H, Bowman  JL, Drews  GN, Feldmann  KA, Meyerowitz  EM (1990) The protein encoded by the Arabidopsis homeotic gene agamous resembles transcription factors. Nature  346(6279): 35–39 [DOI] [PubMed] [Google Scholar]
  75. Yao  JL, Dong  YH, Morris  BA (2001) Parthenocarpic apple fruit production conferred by transposon insertion mutations in a MADS-box transcription factor. Proc Natl Acad Sci U S A  98(3): 1306–1311 [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Yao  ZY, Qi  JH, Yin  LM (2013) Biodiesel production from Xanthoceras sorbifolia in China: opportunities and challenges. Renew Sustain Energy Rev  24: 57–65 [Google Scholar]
  77. Zhang  R, Guo  C, Zhang  W, Wang  P, Li  L, Duan  X, Du  Q, Zhao  L, Shan  H, Hodges  SA, et al. (2013) Disruption of the petal identity gene APETALA3-3 is highly correlated with loss of petals within the buttercup family (Ranunculaceae). Proc Natl Acad Sci U S A  110(13): 5074–5079 [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Zhang  H, Wang  X, Pan  Q, Li  P, Liu  Y, Lu  X, Zhong  W, Li  M, Han  L, Li  J, et al. (2019) QTG-Seq accelerates QTL fine mapping through QTL partitioning and whole-genome sequencing of bulked segregant samples. Mol Plant  12(3): 426–437 [DOI] [PubMed] [Google Scholar]
  79. Zhou  Y, Gao  S, Zhang  X, Gao  H, Hu  Q, Song  Y, Jiao  Y, Gao  H (2012) Morphology and biochemical characteristics of pistils in the staminate flowers of yellow horn during selective abortion. Aust J Bot  60(2): 143–153 [Google Scholar]
  80. Zik  M, Irish  VF (2003) Flower development: initiation, differentiation, and diversification. Annu Rev Cell Dev Biol  19(1): 119–140 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

kiac571_Supplementary_Data
Click here for additional data file. (18.5MB, zip)

Articles from Plant Physiology are provided here courtesy of Oxford University Press

RESOURCES