Candidate Gene Identification and Transcriptome Analysis of Tomato male sterile-30 and Functional Marker Development for ms-30 and Its Alleles, ms-33, 7B-1, and stamenless-2

Male sterility is a valuable trait for hybrid seed production in tomato (Solanum lycopersicum). The mutants male sterile-30 (ms-30) and ms-33 of tomato exhibit twisted stamens, exposed stigmas, and complete male sterility, thus holding potential for application in hybrid seed production. In this study, the ms-30 and ms-33 loci were fine-mapped to 53.3 kb and 111.2 kb intervals, respectively. Tomato PISTILLATA (TPI, syn. SlGLO2), a B-class MADS-box transcription factor gene, was identified as the most likely candidate gene for both loci. TPI is also the candidate gene of tomato male sterile mutant 7B-1 and sl-2. Allelism tests revealed that ms-30, ms-33, 7B-1, and sl-2 were allelic. Sequencing analysis showed sequence alterations in the TPI gene in all these mutants, with ms-30 exhibiting a transversion (G to T) that resulted in a missense mutation (S to I); ms-33 showing a transition (A to T) that led to alternative splicing, resulting in a loss of 46 amino acids in protein; and 7B-1 and sl-2 mutants showing the insertion of an approximately 4.8 kb retrotransposon. On the basis of these sequence alterations, a Kompetitive Allele Specific PCR marker, a sequencing marker, and an Insertion/Deletion marker were developed. Phenotypic analysis of the TPI gene-edited mutants and allelism tests indicated that the gene TPI is responsible for ms-30 and its alleles. Transcriptome analysis of ms-30 and quantitative RT-PCR revealed some differentially expressed genes associated with stamen and carpel development. These findings will aid in the marker-assisted selection for ms-30 and its alleles in tomato breeding and support the functional analysis of the TPI gene.


Introduction
Tomato (Solanum lycopersicum) is among the most crucial vegetables globally.According to the Food and Agriculture Organization of the United Nations statistical database 2021, approximately 189 million tons of fresh tomatoes are produced worldwide (https: //www.fao.org/faostat/zh/#data/QCL).The widespread adoption of tomato F 1 hybrids can be attributed to their consistently higher yields, improved quality, and enhanced resistance to diseases compared with open-pollinated varieties [1,2].However, the traditional method of producing tomato hybrid seeds involves labor-intensive processes, such as manual emasculation and hand pollination, leading to increased costs and the risk of self-pollination impurities [3,4].To address these challenges, researchers have explored the suitability of male sterile plants as female progenitors in the production of tomato hybrid seeds [4,5].Different genes controlling male sterility have been examined, and their applications in plant breeding have been evaluated.
Since the first report of a male sterile mutant in tomatoes in 1915, over 55 male sterile alleles have been identified and categorized into three types: sporogenous, functional, and structural sterility [5,6].Most tomato male sterile mutants are of the sporogenous type [7], with only a few having been cloned, such as SlMS10 and SlMS32.Both of these encode a basic helix-loop-helix (bHLH) transcription factor [8][9][10] and are homologous to DYT1 and bHLH10/89/90 in Arabidopsis, respectively [8,10].DYT1 and bHLH10/89/90 are part of the DYT1-TDF1-AMS-MS188-MS1 pathway that regulates tapetum development and pollen formation in Arabidopsis [11].Positional sterile (ps) is the first reported mutant of functional male sterility in tomato [12].It produces normal pollen; however, its pollen grains remain enclosed within the anther locule due to non-splitting anthers [13].The second reported mutant of this type is ps-2.PS-2 encodes a polygalacturonase gene (PG) [14,15], and it has been used for tomato hybrid seed production in Eastern Europe [16].Moreover, functional male sterility can result from exserted stigma.In the exserted stigma (ex) mutant, the anthers are normal and capable of producing and releasing viable pollens, but the excessively long style prevents effective pollination [17].Other mutants of this type include Style2.1,Stigma Exertion 3.1 (SE3.1), and SlLst (Long styles) [18][19][20].Stamenless (sl), the first structural male sterile mutant reported in tomato, exhibited the complete transformation of stamens into carpels and the partial transformation of petals into sepals [21].Tomato APETALA3 (TAP3) may be responsible for the sl trait [22,23].In the sl-2 mutant, petals appear almost typical, but the stamens are twisted and distorted, bearing naked ovules [24].The male sterile mutant 7B-1 displays a high degree of phenotypic similarity with sl-2 and an allelism test confirmed that 7B-1 and sl-2 are allelic [25].Genetic mapping and molecular analysis indicated that the B-class MADS-box gene Tomato PISTILLATA (TPI, syn.SlGLO2) is a candidate gene of 7B-1 and sl-2; however, the specific sequence variations of TPI in 7B-1 and sl-2 mutants remain elusive [25].Tomato ms-15 and its alleles, ms-15 26 and ms-15 47 , producing flowers with homeotic conversion of stamens into carpels, have been fine-mapped to a 44.6 kb interval that contains the TM6 gene [26].A recent study revealed that TM6 maintains the balance of alternative splicing of MADS-box genes involved in flower development [27].
Each type of male sterility in tomato presents both advantages and disadvantages for use in hybrid seed production.The lines ms10 and ms32 are considered beneficial for hybrid seed production [28], because they exhibit exserted stigma (accessible for pollination without emasculation), complete sterility, and stable expression of sterility regardless of environmental conditions.However, the application of these two sterile lines has some major drawbacks, such as the variation in the percentage of flowers with accessible stigmas depending on the environment and genotype [4].The exserted stigma (ex) sterility type was initially considered advantageous for hybrid seed production due to the easy maintenance of sterile lines, absence of a need for stamen emasculation, and high hybrid seed yield.However, it presents the challenge of managing two polygenic traits (anther and style length) and the variability of phenotypic expression under different environmental conditions.ps-2 has disadvantages due to the risk of undesirable self-pollination and the necessity for stamen emasculation [4,16].Although 7B-1 can be used for hybrid seed production [29], the absence of a functional codominant marker for 7B-1 reduces the efficiency of backcross transfer.These deficiencies in male sterile lines for tomato hybrid seed production necessitate further exploration of male sterility genes and the development of codominant molecular markers for marker-assisted selection (MAS).
Tomato male sterile mutants ms-30 and ms-33 feature shorter, twisted stamens and exerted stigmas, facilitating direct pollination without the need for emasculation [30,31].Among these mutants, ms-33 appears highly promising as a female parent in the production of tomato hybrid seeds [32].Linkage analysis revealed that the ms-33 locus is located on chromosome 6 [33].The fertility of ms-33 can be partially or completely restored with the application of exogenous GA3, resulting in seeds that produce 100% male sterile plants [7,34].Given these characteristics, ms-30 and ms-33 can be valuable for hybrid seed production.However, the specific genes responsible for ms-30 and ms-33 sterility remain unknown, and there are currently no codominant functional markers available for selecting these mutants.
The objective of this study was to characterize the genes MS-30 and MS-33, along with their sequence variations in wild-type plants and the male sterile mutants ms-30 and ms-33.Subsequently, the study aimed to develop codominant functional markers for MAS utilizing these sequence variations, and to perform an initial investigation into the role of the MS-30 gene in stamen development through RNA-seq analysis.

Phenotypes of the ms-30 and ms-33 Mutants
The flowers of WT-30 (the wild-type corresponding to ms-30) and WT-33 exhibited normal architecture, in contrast to the stamens of ms-30 and ms-33, which displayed distinctive features, such as twisted stamens, lateral separation, reduced size, and pale color, resulting in an exerted stigma (Figures 1 and S1A).Additionally, these distorted stamens were observed to transform into carpel-like structures, including carpelloid structures (CS), naked external ovules on the adaxial surface (EO), and complete transformation into carpels (TC).Notably, the occurrence of EO was more frequent than that of CS and TC in both ms-30 and ms-33 (Table S1).Due to these alterations, neither mutant was capable of producing pollens (Figure S1B), and both ms-30 and ms-33 were unable to set fruit when subjected to self-pollination using an electric vibrator.However, manual pollination with pollens from wild-type plants resulted in a successful fruit set (Figure S1C).
Int. J. Mol.Sci.2024, 25, x 3 of 18 Among these mutants, ms-33 appears highly promising as a female parent in the production of tomato hybrid seeds [32].Linkage analysis revealed that the ms-33 locus is located on chromosome 6 [33].The fertility of ms-33 can be partially or completely restored with the application of exogenous GA3, resulting in seeds that produce 100% male sterile plants [7,34].Given these characteristics, ms-30 and ms-33 can be valuable for hybrid seed production.However, the specific genes responsible for ms-30 and ms-33 sterility remain unknown, and there are currently no codominant functional markers available for selecting these mutants.
The objective of this study was to characterize the genes MS-30 and MS-33, along with their sequence variations in wild-type plants and the male sterile mutants ms-30 and ms-33.Subsequently, the study aimed to develop codominant functional markers for MAS utilizing these sequence variations, and to perform an initial investigation into the role of the MS-30 gene in stamen development through RNA-seq analysis.

Phenotypes of the ms-30 and ms-33 Mutants
The flowers of WT-30 (the wild-type corresponding to ms-30) and WT-33 exhibited normal architecture, in contrast to the stamens of ms-30 and ms-33, which displayed distinctive features, such as twisted stamens, lateral separation, reduced size, and pale color, resulting in an exerted stigma (Figures 1 and S1A).Additionally, these distorted stamens were observed to transform into carpel-like structures, including carpelloid structures (CS), naked external ovules on the adaxial surface (EO), and complete transformation into carpels (TC).Notably, the occurrence of EO was more frequent than that of CS and TC in both ms-30 and ms-33 (Table S1).Due to these alterations, neither mutant was capable of producing pollens (Figure S1B), and both ms-30 and ms-33 were unable to set fruit when subjected to self-pollination using an electric vibrator.However, manual pollination with pollens from wild-type plants resulted in a successful fruit set (Figure S1C).

Genetic Analysis of ms-30 and ms-33 Loci
The F 1 plants resulting from crosses between ms-30 and LA1589, as well as between ms-33 and LA1589, displayed complete fertility.Additionally, a preliminary analysis involved 200 plants in each of the F 2 populations derived from these crosses.The ratio of fertile to sterile plants for both ms-30 and ms-33 conformed to the expected 3:1 segregation ratio, as confirmed by the chi-square test (Table 1).The observed segregation pattern and the statistical analysis suggested that the male sterility in both ms-30 and ms-33 is controlled by a single recessive gene.

Fine-Mapping of ms-30 and ms-33 Loci
For preliminary mapping of the ms-30 and ms-33 loci, markers located on chromosome 6 were utilized.This approach was based on prior reports indicating that ms-33 is located on chromosome 6 of tomato.Furthermore, the ms-30 and ms-33 mutants exhibited similar phenotypes, suggesting a potential linkage (Figures 1 and S1; Table S1) [30,31,33,35,36].In this preliminary mapping phase, 94 male sterile plants from each F 2 population were genotyped, and their recombination rates were calculated.As anticipated, the findings confirmed that both ms-30 and ms-33 are located within the region bounded by markers HP149 and HP3661 on chromosome 6 (Figures 2A and S2A).To further narrow down the locations of these loci, 76 and 39 recombinants were selected from the entire F 2 populations of ms-30 and ms-33, respectively, and genotyped with additional markers.This detailed analysis allowed the ms-30 locus to be narrowed down to a 53.3 kb region, delimited by the markers CP207 and CP203 (Figure 2B).Similarly, the ms-33 locus was confined to an interval of 111.2 kb, defined by markers CP207 and HP3921 (Figure S2B).Notably, both loci shared an identical 53.3 kb region.

Candidate Gene Analysis and Allelism Testing among ms-30, ms-33, and sl-2
According to the ITAG4.0genome annotation [37], Solyc06g059970, which encodes Tomato PISTILLATA (TPI, synonymously referred to as SlGLO2) [38,39], was identified within the fine-mapped regions of both the ms-30 and ms-33 loci (Figures 2C and S2C; Tables 2 and S2).TPI plays a pivotal role in stamen development and has been identified as the candidate gene for 7B-1 and sl-2 [25], which exhibit phenotypes similar to those of ms-30 and ms-33 (Figures 1 and S1).To further ascertain the genetic relationships among ms-30, ms-33, and sl-2, allelism tests were performed.The outcomes of these tests confirmed that ms-30, ms-33, and sl-2 are allelic, indicating that they arise from mutations in the same gene (Table S3).Thus, TPI was identified as the most probable candidate gene responsible for the phenotypic traits observed in ms-30 and ms-33.

Candidate Gene Analysis and Allelism Testing among ms-30, ms-33, and sl-2
According to the ITAG4.0genome annotation [37], Solyc06g059970, which encode Tomato PISTILLATA (TPI, synonymously referred to as SlGLO2) [38,39], was identified within the fine-mapped regions of both the ms-30 and ms-33 loci (Figures 2C and S2C Tables 2 and S2).TPI plays a pivotal role in stamen development and has been identified as the candidate gene for 7B-1 and sl-2 [25], which exhibit phenotypes similar to those o ms-30 and ms-33 (Figures 1 and S1).To further ascertain the genetic relationships amon ms-30, ms-33, and sl-2, allelism tests were performed.The outcomes of these tests con firmed that ms-30, ms-33, and sl-2 are allelic, indicating that they arise from mutations in the same gene (Table S3).Thus, TPI was identified as the most probable candidate gen responsible for the phenotypic traits observed in ms-30 and ms-33.To identify sequence variations within the TPI gene between wild-type plants and ms mutants, the genomic sequence of the TPI gene was sequenced in WT-30, ms-30, WT-33, and ms-33.Sequencing data analysis revealed specific mutations in each mutant.In ms-30, a single-nucleotide polymorphism (SNP) was discovered in the first exon of the TPI gene (Figure 3A), leading to a missense mutation.This mutation altered a serine residue to an isoleucine within the MADS domain of the protein, impacting its function (Figure S3).In ms-33, a different SNP was identified in the first intron of the TPI gene (Figure 3B).Amplification of cDNAs from ms-33 produced several bands, with one being smaller than the corresponding band from the wild-type (Figure S4A).Further analysis indicated that this SNP might introduce an alternative 5' splice site in ms-33's gene transcript.This alternative splicing resulted in a 138 bp deletion in cDNA, leading to the protein lacking a 46 amino-acid fragment (Figure S3).
Pucci et al. (2017) [25] reported that the primer pair LeGLO2-6F/R did not amplify the last exon and part of the 3' UTR region of the TPI gene in 7B-1 and sl-2.To investigate sequence variation in this region, PCR amplification was conducted using the LeGLO2-6F/R primer pair with an extended elongation time.This method enabled the production of approximately 5000 bp long amplicons from both 7B-1 and sl-2 (Figures S4B and S5).Sequencing of these amplicons showed no sequence variations.However, a fragment of 4868 bp in the amplicon was identified as a Copia-like retrotransposon, featuring a 398 bp long terminal repeat (LTR) sequence and a 4 bp target site duplication (TSD) sequence (Figures 3C and S5).Based on the SNPs in the TPI gene of ms-30 and ms-33 and the LTR retrotransposon insertion in the TPI gene of 7B-1 and sl-2, three codominant markers were developed, namely a Kompetitive Allele Specific PCR (KASP) marker, MS30KASP; a sequence marker, MS33SEQ; and an Insertion/Deletion (InDel) marker, LeGLO2-6/SL-2TE (Figure 3D-F).
To identify sequence variations within the TPI gene between wild-type plants and ms mutants, the genomic sequence of the TPI gene was sequenced in WT-30, ms-30, WT-33, and ms-33.Sequencing data analysis revealed specific mutations in each mutant.In ms-30, a single-nucleotide polymorphism (SNP) was discovered in the first exon of the TPI gene (Figure 3A), leading to a missense mutation.This mutation altered a serine residue to an isoleucine within the MADS domain of the protein, impacting its function (Figure S3).In ms-33, a different SNP was identified in the first intron of the TPI gene (Figure 3B).Amplification of cDNAs from ms-33 produced several bands, with one being smaller than the corresponding band from the wild-type (Figure S4A).Further analysis indicated that this SNP might introduce an alternative 5' splice site in ms-33′s gene transcript.This alternative splicing resulted in a 138 bp deletion in cDNA, leading to the protein lacking a 46 amino-acid fragment (Figure S3).[25] reported that the primer pair LeGLO2-6F/R did not amplify the last exon and part of the 3' UTR region of the TPI gene in 7B-1 and sl-2.To investigate sequence variation in this region, PCR amplification was conducted using the LeGLO2-6F/R primer pair with an extended elongation time.This method enabled the production of approximately 5000 bp long amplicons from both 7B-1 and sl-2 (Figures S4B and S5).Sequencing of these amplicons showed no sequence variations.However, a fragment of 4868 bp in the amplicon was identified as a Copia-like retrotransposon, featuring a 398 bp long terminal repeat (LTR) sequence and a 4 bp target site duplication (TSD) sequence (Figures 3C and S5).Based on the SNPs in the TPI gene of ms-30 and ms-33 and the LTR

Targeted Knockout of the TPI Gene by CRISPR/Cas9
To confirm that the TPI gene confers the phenotype observed in ms-30 and its alleles, the TPI gene was knocked out using the CRISPR/Cas9 system, and allelism tests were conducted between the ms-30 and TPI gene-edited mutants.Two vectors were designed to disrupt the TPI gene, with target sites on the first and last exons for vector-1 and vector-2, respectively.This resulted in two null mutants of the TPI gene.The tpicr-e1 mutant exhibited a 4 bp deletion, whereas tpicr-e7 had a 66 bp deletion (Figure 4A).The stamens of both tpicr-e1 and tpicr-e7 mutants showed curling, wrinkling, greenish coloration, and were incapable of producing viable pollen (Figure 4B-J), similar to phenotypic traits observed in the ms-30 and ms-33 mutants.Allelism tests between tpicr-e1, tpicr-e7, and ms-30 indicated that they were allelic (Table S3).These findings support the conclusion that TPI is responsible for the observed phenotypes in ms-30 and its alleles, including ms-33, 7B-1, and sl-2.

Comparative RNA-Seq Analysis of WT-30 and ms-30 Floral Organs from the Meiotic to Tetrad Stage
To further explore the role of TPI in floral organ development, RNA-seq analysis was conducted on the four floral organ whorls of young flower buds from WT-30 and ms-30 at the meiotic and tetrad stages (3-5 mm in length) [40].RNA-Seq generated approximately 568 million paired-end reads, with each sample yielding between 20 million and 30 million reads.On average, 98% of these reads were successfully mapped to the ITAG4.0tomato reference genome (Table S4).Principal component analysis (PCA) results showed that biological replicates clustered together, indicating good consistency within the data (Figure 5A).Notably, the stamens of ms-30 (ST_ms-30) were closely related to the carpels of WT-30 (CA_WT-30) and the carpels of ms-30 (CA_ms-30) in the PCA plot.This observation aligns with the homeotic transformation of stamens into carpels observed in the ms-30 mutant, providing molecular evidence for this phenotypic change.
mato reference genome (Table S4).Principal component analysis (PCA) results showed that biological replicates clustered together, indicating good consistency within the data (Figure 5A).Notably, the stamens of ms-30 (ST_ms-30) were closely related to the carpels of WT-30 (CA_WT-30) and the carpels of ms-30 (CA_ms-30) in the PCA plot.This observation aligns with the homeotic transformation of stamens into carpels observed in the ms-30 mutant, providing molecular evidence for this phenotypic change.Among the differentially expressed genes (DEGs) identified between WT-30 and ms-30, 96 were in sepals, 738 in petals, 5391 in stamens, and 131 in carpels.Specifically, within the stamens, 2238 genes were upregulated, whereas 3153 genes were downregulated (Figure 5B, Tables S5-S9).Regarding the ABCDE model of floral organ development, the expression of the D-class gene TAGL11 was not detected in any of the four floral organ whorls.The expression levels of other ABCDE model genes, with the exceptions of SlGLO1 and TAG1, differed significantly in the stamens between WT-30 and ms-30 (Figure S6).
Gene ontology (GO) analysis revealed that upregulated DEGs in stamens were enriched in specific biological process categories, such as cell division, regulation of transcription, DNA-templated synthesis, response to auxin, ethylene-activated signaling pathways, and response to light stimulus (Figure 5C).Conversely, DEGs downregulated in stamens were associated with categories such as response to water deprivation, response to abscisic acid, pollen sperm cell differentiation, and pollen tube growth (Figure 5D).These findings suggest that the expression of the identified DEGs is associated with key processes such as pollen development, hormonal responses, and responses to environmental stimuli.

Discussion
Using male sterile lines as the female parent in hybrid tomato seed production is considered a highly efficient and cost-effective method [28].Several male sterile lines, including ms-10, ms-32, sl, 7B-1, ex, ps, and ps-2, [4,29] have been successfully applied in tomato hybrid seed production [4,29].However, these lines have drawbacks, such as the risk of self-pollination, the need for anther emasculation, environmental dependency of phenotypic expression, reduced hybrid seed yield, or the lack of codominant markers for MAS [4].Therefore, identifying additional genes responsible for male sterility in tomatoes, particularly those resulting in exposed stigmas and stable, complete male sterility, is crucial.Developing codominant markers for MAS of these male sterile loci is equally important.The ms-30 and ms-33 mutants are promising for hybrid tomato seed production due to their complete male sterility, exerted stigmas (Figure 1), and the ability to be manually pollinated directly without the need for hand emasculation.Both ms-30 and ms-33 exhibit single-gene recessive inheritance, with only homozygous plants displaying the male sterile phenotype, making MAS more efficient than conventional phenotypic selection for these mutants in breeding programs [47].The ms-30 and ms-33 loci have been fine-mapped, identifying the TPI gene as the candidate gene for both loci.These mutants were found to be allelic to 7B-1 and sl-2 [25], which also involve TPI.Phenotypic analysis and allelism tests of TPI gene-edited mutants confirmed that TPI underlies the male sterile phenotype in ms-30 and its alleles, including ms-33, 7B-1, and sl-2.Sequence analysis identified variations in the TPI gene across these mutants.Based on these variations, three codominant markers-KASP marker MS30KASP, sequencing marker MS33SEQ, and InDel marker LeG-LO2-6/SL-2TE-were developed, facilitating rapid incorporation of these loci into tomato breeding.Additionally, CRISPR/Cas9-mediated targeted mutagenesis represents another approach to rapidly introduce male sterility into elite breeding lines [9,[48][49][50][51][52].The phenotypes of two TPI gene-edited mutants closely matched those of ms-30 and its alleles, suggesting that gene editing of TPI could efficiently induce male sterility in elite lines, thereby reducing the costs associated with hybrid seed production.
Carpelloid stamens are occasionally observed in angiosperms, with mutations in Bclass genes identified as the primary genetic cause of this phenomenon [53].The number of B-class genes varies across different plant species; for instance, Arabidopsis has two B-class genes (AP3 and PI), whereas tomato contains four genes (TAP3, TM6, TPI, and Sl-GLO1) [38,54].This variation suggests potential functional conservation and diversification of B-class genes among plant species [54].The role of B-class genes in stamen specification and development has been extensively studied since the 1990s [55][56][57]; however, further research is warranted.In this study, the ms-30 and ms-33 mutants displayed carpelloid stamens (Figure 1) and were found to be allelic to 7B-1 and (Table S3).The B-class gene TPI was identified as the underlying gene for ms-30 and its alleles (Figures 2 and S2).Transcriptome analysis of ms-30 was conducted to examine the function of TPI in stamen development.PCA of the transcriptome data showed that the clusters for stamens and carpels were closely related (Figure 5A), aligning with the observed phenotype of carpelloid stamens in ms-30 (Figure 1).The number of DEGs was higher in stamens than in the other three floral organs (Figure 5B).Some DEGs in stamens, specifically expressed in wild-type stamens, were downregulated in the ms-30 mutant (Figure 6A).Examples of such genes are SlMS10 (Solyc02g079810), Solyc03g113530 (AtTDF1-like1), SlAMS (Solyc08g062780), SlMS32 (Solyc01g081100), and SlPHD_MS1 (Solyc04g008420) (Figure 6C), which are key components of the DYT1-TDF1-AMS-MS188-MS1 regulatory network [8][9][10]41,42].Conversely, DEGs that were specifically expressed in wild-type carpels and upregulated in ms-30 stamens included SlCRCa (Solyc01g010240), SlCRCb (Solyc05g012050), SlINO (Solyc05g005240), LYRATE (Solyc05g009380), and SlMBP3 (Solyc06g064840) (Figure 6D), suggesting their involvement in carpel or ovule development in tomato [43][44][45][46].These findings suggest that TPI positively regulates the genes associated with stamen and pollen development while negatively influencing the genes involved in carpel and ovule development within stamens.Consequently, loss of function of the TPI gene leads to the low or absent expression of genes associated with stamen development and the ectopic expression of genes related to carpel and ovule development in stamens, potentially resulting in the formation of carpelloid stamens in mutants.However, further investigation is needed to elucidate the specific molecular regulatory mechanisms.

Plant Materials
Seeds for the wild-type (WT) plants and ms mutants (ms-30, accession number 2-455; ms-33, 2-511; sl-2, LA1801; Ailsa Craig [AC]; LA1589 and Heinz 1706) were obtained from the Tomato Genetics Resource Center (Davis, CA, USA).The male sterile plants derived from the accessions 2-455, 2-511, and LA1801 were designated as ms-30, ms-33, and sl-2, respectively.Similarly, the homozygous male fertile plants from 2-455 and 2-511 were identified as WT-30 and WT-33, respectively.AC was used for gene editing and as WT controls for the tpicr-e1 and tpicr-e7 mutants.Two F 2 populations were created by crossing both ms-30 and ms-33 with the wild tomato species Solanum pimpinellifolium accession LA1589.These populations were cultivated in an open field in Shunyi District, Beijing, China, during the spring and summer of 2016, comprising 949 and 861 plants, respectively.
To confirm the allelism among ms-30, ms-33, and sl-2, the flowers of sl-2 were pollinated using pollens from the heterozygous ms-30 or ms-33 plants.Additionally, to determine whether the TPI gene is responsible for conferring the ms-30 phenotype, flowers of ms-30 were pollinated with pollens from the heterozygous TPI gene-edited mutants.Conversely, male sterile flowers of TPI gene-edited mutants were pollinated with pollens from the heterozygous ms-30 plants.The progeny from these crosses were grown in a greenhouse located in Haidian District, Beijing, China, during the springs and summers of 2019, 2020, and 2023.

Phenotypic Analysis
For phenotypic and expression analyses, WT-30, ms-30, WT-33, ms-33, and AC plants, as well as gene-edited mutants, were grown in a greenhouse in Haidian District, Beijing, China, across the springs and summers of 2019, 2020, 2021, and 2023.The flower morphology of both WT and ms mutants was examined at the anthesis stage using the previously described methods [25].The entire flower was photographed by a camera (Canon EOS 70D, Canon Inc., Tokyo, Japan).The anther cones and dissected stamens were observed by a stereomicroscope (Carl Zeiss MicroImaging GmbH, Gottingen, Germany).Furthermore, referring to the procedures outlined by Cao et al., whole plants were photographed and pollen viability was evaluated [26].Pollen was collected from the WT and ms plants, stained with 1% aceto-carmine, and examined using a microscope (BX51, Olympus Corporation, Tokyo, Japan).Based on crossing of the male sterile mutants ms-30 and ms-33 with Heinz 1706, we collected the fruit set data.

Molecular Marker Development and Genotyping
InDels and SNPs were identified by comparing the sequence of chromosome 6 between the tomato lines Heinz 1706 and LA1589.Whole-genome sequences for these lines were obtained from the Sol Genomics Network (SGN, https://solgenomics.net/) [58].To evaluate polymorphisms between ms-30/ms-33 and LA1589, PCR primers were designed based on the flanking regions of these InDels and SNPs.
KASP genotyping reactions were performed using a LightCycler480I (Roche Diagnostics GmbH, Mannheim, Germany) with KASP-TF V4.0 2× Master Mix with Low Rox (Cat.No. KBS-1050-122, LGC Genomic Limited, Middlesex, United Kingdom) following the provided manual protocol (https://www.biosearchtech.com/support/education/kaspgenotyping-reagents/running-kasp-genotyping-reactions).The TPI gene fragment for ms-33 was amplified using the marker MS33SEQ and sequenced at the Beijing Genomics Institute (Beijing, China).To identify the insertion of the LTR retrotransposon in the 7B-1 and sl-2 mutants, PCR was conducted separately with primers LeGLO2-6F/R and SL-2TE-F/R.The resulting PCR products were then combined in a single tube and analyzed through electrophoresis.

Genetic Analysis and Fine-Mapping
Genetic analysis and preliminary mapping were performed using 200 plants and 94 male sterile plants selected from each F 2 population.For fine-mapping of the ms-30 and ms-33 loci, all plants in the F 2 populations were used.The goodness-of-fit to a 3:1 segregation ratio at the ms-30 and ms-33 loci was tested using the chi-square test with 200 individuals from each F 2 population.For preliminary mapping of the ms-30 and ms-33 loci, the first 94 sterile plants in each F 2 population were genotyped with 2 and 6 InDel markers, respectively.For fine-mapping of the ms-30 and ms-33 loci, 76 and 39 recombinants were selected from each F 2 population using markers HP149 and HP3661, respectively.These recombinants were identified as homozygous for either the ms-30 or ms-33 mutant allele and for only one of these two markers.The recombinants were further genotyped with additional markers (Table S12), and their fertility/sterility was assessed by examining the flowers on at least three inflorescences per plant.

Gene Prediction and Sequence Polymorphism Analysis
The putative genes in the fine-mapped region containing the ms-30 and ms-33 loci were identified using the tomato gene model (ITAG release 4.0) [37] in the SGN (https: //solgenomics.net).
The genomic sequence of TPI from the tomato lines WT-30, ms-30, WT-33, and ms-33 was obtained using overlapping PCR.The last exon and part of the 3 ′ UTR region of the TPI gene in 7B-1 and sl-2 were amplified by PCR with an elongation time of 7 min, using primers LeGLO2-6F/R [25], and 2× Phanta Max Master Mix (Dye Plus) (Cat.No. P525-03, Vazyme, Nanjing, China).The coding sequences (CDS) of the TPI gene were obtained through reverse-transcription PCR (RT-PCR).Fragments were amplified using primers LeGLO2-6F/R, and the CDS of TPI were cloned into the pEASY-Blunt Zero Cloning Vector (pEASY-Blunt Zero Cloning Kit, Cat.No. CB501-02; TransGen Biotech, Beijing, China).Both the amplified fragments and plasmids were sequenced at the Beijing Genomics Institute (Beijing, China).All primers utilized in this study are detailed in Table S12.Genomic DNA of 7B-1 was provided by Professor Huolin Shen from the College of Horticulture, China Agricultural University.

RNA Extraction, cDNA Synthesis, and qRT-PCR Analysis
The sepals, petals, stamens, and carpels of young flower buds at the meiotic and tetrad stages (3-5 mm in length) [40] were collected from WT-30, ms-30, WT-33, ms-33, AC, tpicr-e1, and tpicr-e7 plants and then immediately frozen in liquid nitrogen.Each tissue type included three biological replicates, with each replicate comprising samples from at least three plants.Total RNA extraction, cDNA synthesis, and qRT-PCR analysis were conducted following the methods previously described by Cao et al. (2019) [26] and Liu et al. (2019) [10].All primers used for these analyses are detailed in Table S12.

Construction of Gene Editing Vectors, Plant Transformation, and Selection of Mutant Alleles
For the CRISPR/Cas9 constructs, two sgRNA binding sites per vector were identified using CRISPR-GE [62] (http://skl.scau.edu.cn/).Primers incorporating sgRNAs and Bsa I recognition sites were used to amplify the sgRNA(1)_SlU6-2t_SlU3-5p_sgRNA(2) fragments, with the pCBC-S1 vector as a template.The PCR fragments were then assembled into the binary vector pMGET through the Golden Gate cloning method, following the protocols described by Xing et al. (2014) [63] and Yang et al. (2023) [64].These constructs were introduced into AC tomato plants via Agrobacterium tumefaciens (GV3101)-mediated cotyledon explant transformation, adhering to the method outlined by Yang et al. (2023) [64].To confirm whether transgenic plants were successfully generated, PCR was performed using primers Cas9F and Cas9R.To identify specific types of edits, two primer pairs, TPI1 and TPI2, were used to amplify sequences including the targets of vector1 and vector2, respectively.The resulting PCR fragments were cloned into the pEASY-Blunt Zero Cloning Vector.At least 12 clones per PCR product were sequenced at the Beijing Genomics Institute.Two heterozygous T 0 plants, each with a 4 bp or 66 bp deletion, were self-pollinated to produce T 1 generation seeds.Homozygous mutants from the T 1 generation, lacking CAS9, were identified using primers TPI1, TPI2, and Cas9.All primers used are detailed in Table S12.

Data Statistical Analysis
The mean values of floral organs length and relative expression were analyzed using IBM SPSS Statistics 20.0 software, with a t-test employed to assess significant differences among the means.The chi-square goodness-of-fit test was conducted by online data analysis platform SPSSPRO (https://www.spsspro.com).

Conclusions
Overall, this study demonstrated that the TPI gene (Solyc06g059970) is the causal factor for the phenotypic abnormalities observed in the mutants ms-30, ms-33, 7B-1, and sl-2.Sequence alterations in TPI across these mutants included an SNP each in ms-30 and ms-33 and an LTR retrotransposon insertion in both 7B-1 and sl-2.Leveraging these sequence variations, three codominant markers-MS30KASP, MS33SEQ, and LeGLO2-6/SL-2TEwere developed.RNA-seq and qRT-PCR analyses indicated that loss-of-function mutations in TPI altered the expression of genes involved in stamen and carpel development.These findings may be useful for the rapid development of male sterile lines through molecular marker-assisted backcrossing or CRISPR/Cas9-mediated mutagenesis targeting TPI, contributing to hybrid seed production and laying the groundwork for further functional analysis of the TPI gene.

Figure 2 .
Figure 2. Fine-mapping of ms-30.(A) Preliminary mapping of ms-30; PS indicates population size (B) Fine-mapping of ms-30; N indicates the number of recombinants, and 76 recombinants wer selected from 949 plants of F2 population.(C) ITAG4.0 annotated genes of ms-30.Arrows indicat the direction of transcription, and the solid arrow represents the most likely candidate gene for ms 30.

Figure 2 .
Figure 2. Fine-mapping of ms-30.(A) Preliminary mapping of ms-30; PS indicates population size.(B) Fine-mapping of ms-30; N indicates the number of recombinants, and 76 recombinants were selected from 949 plants of F 2 population.(C) ITAG4.0 annotated genes of ms-30.Arrows indicate the direction of transcription, and the solid arrow represents the most likely candidate gene for ms-30.2.5.Sequence Analysis of the TPI Gene in ms-30, ms-33, 7B-1, and sl-2

Figure 3 .
Figure 3. Sequence alterations of the TPI gene in ms-30, ms-33, 7B-1, and sl-2 and the development of molecular markers (A,B) Single-nucleotide polymorphisms (SNPs) in the TPI gene of ms-30 and ms-33; black boxes indicate the exon of TPI; (C) structure of the TPI gene according to ITAG4.0 (upper) and the structure of the long terminal repeat (LTR) retrotransposon inserted in the last exon of TPI in 7B-1 and sl-2 (lower); gray boxes indicate the LTR sequence of retrotransposon, hollow blocks are two ORFs predicted on the retrotransposon, and arrows indicate the direction of transcription."TATC" is the TSD sequences.LeGLO2-6F, LeGLO2-6R, SL-2TE-F, and SL-2TE-R are the primers used in generating the LeGLO2-6/SL-2TE marker.(D) Endpoint fluorescence scatter plot of the Kompetitive Allele Specific PCR marker MS30KASP.(E) The chromatogram of MS33SEQ; red box denotes the SNP in TPI of ms-33.(F) Agarose gel electrophoresis of PCR fragments amplified from Ailsa Craig, 7B-1, and plants from LA1801 (sl-2) using the marker LeGLO2-6/SL-2TE; F indicates the fertile plant, whereas S indicates the sterile plant.

Figure 3 .
Figure 3. Sequence alterations of the TPI gene in ms-30, ms-33, 7B-1, and sl-2 and the development of molecular markers (A,B) Single-nucleotide polymorphisms (SNPs) in the TPI gene of ms-30 and ms-33; black boxes indicate the exon of TPI; (C) structure of the TPI gene according to ITAG4.0 (upper) and the structure of the long terminal repeat (LTR) retrotransposon inserted in the last exon of TPI in 7B-1 and sl-2 (lower); gray boxes indicate the LTR sequence of retrotransposon, hollow blocks are two ORFs predicted on the retrotransposon, and arrows indicate the direction of transcription."TATC" is the TSD sequences.LeGLO2-6F, LeGLO2-6R, SL-2TE-F, and SL-2TE-R are the primers used in generating the LeGLO2-6/SL-2TE marker.(D) Endpoint fluorescence scatter plot of the Kompetitive Allele Specific PCR marker MS30KASP.(E) The chromatogram of MS33SEQ; red box denotes the SNP in TPI of ms-33.(F) Agarose gel electrophoresis of PCR fragments amplified from Ailsa Craig, 7B-1, and plants from LA1801 (sl-2) using the marker LeGLO2-6/SL-2TE; F indicates the fertile plant, whereas S indicates the sterile plant.

Figure 4 .
Figure 4. Sequence alterations and phenotypes of TPI gene-edited mutants.(A) TPI mutations generated through CRISPR/Cas9 gene editing.Red letters mark the sequence of targets.Blue letters indicate the protospacer-adjacent motif.Green hyphens indicate deletions in tpicr mutants.(B-J) Representative flower, anther cone, and pollen viability images of AC, tpicr-e1, and tpicr-e7.The white bar indicates 0.5 cm (B-G), and the black bar indicates 1 mm (H-J).

Figure 5 .
Figure 5. Loss of function of TPI in ms-30 affects multiple biological processes during stamen development.(A) Principal component analysis of RNA-seq data.(B) The number of differentially

Figure 5 .
Figure 5. Loss of function of TPI in ms-30 affects multiple biological processes during stamen development.(A) Principal component analysis of RNA-seq data.(B) The number of differentially expressed genes that were upregulated and downregulated in SE, PE, ST, and CA.(C) Gene ontology (GO) enrichment analysis of upregulated differentially expressed genes in ST. (D) GO enrichment analysis of downregulated differentially expressed genes in ST.SE, Sepals; PE, Petals; ST, Stamens; and CA, Carpels.

Figure 6 .
Figure 6.Expression pattern of the DEGs related to stamen or carpel development between WT-30 and ms-30.(A) Heatmap of the DEGs whose homologs in Arabidopsis are related to stamen development.The genes in the red box represent those expressed mainly in the stamen of WT-30 and significantly downregulated in the stamen of ms-30.(B) Heatmap of DEGs whose homologs in Arabidopsis are related to carpel development.Genes in the red box represent those mainly expressed in the carpel and significantly upregulated in the stamen of ms-30.(C,D) FPKM and qRT-PCR analysis of the DEGs related to stamen and carpel development in four floral organs of ms-30.Asterisks indicate a significant difference (**, p < 0.01) between WT-30 and ms-30; WT-30 and ms-30 are homozygous male fertile plants and male sterile plants derived from the tomato line 2-455, respectively.SE, sepals; PE, petals; ST, stamens; and CA, carpels.

Figure 6 .
Figure 6.Expression pattern of the DEGs related to stamen or carpel development between WT-30 and ms-30.(A) Heatmap of the DEGs whose homologs in Arabidopsis are related to stamen development.The genes in the red box represent those expressed mainly in the stamen of WT-30 and significantly downregulated in the stamen of ms-30.(B) Heatmap of DEGs whose homologs in Arabidopsis are related to carpel development.Genes in the red box represent those mainly expressed in the carpel and significantly upregulated in the stamen of ms-30.(C,D) FPKM and qRT-PCR analysis of the DEGs related to stamen and carpel development in four floral organs of ms-30.Asterisks indicate a significant difference (**, p < 0.01) between WT-30 and ms-30; WT-30 and ms-30 are homozygous male fertile plants and male sterile plants derived from the tomato line 2-455, respectively.SE, sepals; PE, petals; ST, stamens; and CA, carpels.

Table 1 .
Chi-square goodness-of-fit test for segregation ratios of F 2 populations at ms-30 and ms-33 loci.

Table 2 .
Predicted genes in the ms-30 region.

Table 2 .
Predicted genes in the ms-30 region.