Using Simple Sequence Repeats in 9 Brassica Complex Species to Assess Hypertrophic Curd Induction

: Five Simple Sequence Repeats (SSRs) were used to assess the relationship between inﬂorescence characteristics and their allelic variation in 53 Brassica oleracea and Brassica wild relatives ( n = 9). Curd morphometric traits, such as weight (CW), height (CH), diameter (CD1), shape (CS) inﬂorescence curvature angle (CA), and its curd stem diameter (CD2), were measured. The aim of the work was to analyze the relationships among the allelic patterns of the SSRs primers utilized, and their status of homo or heterozygosity registered at each locus, as well as the inﬂorescence morphometric traits in order to individuate genomic regions stimulating the hypertrophy of this reproductive organ. The relationships found explain the diversity among B . oleracea complex species ( n = 9) for the inﬂorescence size and structure, allowing important time reduction during the breeding process by crossing wild species, transferring useful resistance, and organoleptic and nutraceutical traits. The ﬁve SSRs loci were BoABI1, BoAP1, BoPLD1, BoTHL1, and PBCGSSRBo39. According to the allelic variation ascertained, we evaluated the heterozygosity index (H) for each SSR above cited. The results showed a signiﬁcant interaction between the H index of the BoPLD1 gene and the inﬂorescence characteristics, summarized by the First Principal Component (PC1) ( p -value = 0.0244); we ascertained a negative correlation between the H index and inﬂorescence characteristics, namely CW, CH, CD1, CD2, CA. The homozygosity BoPLD1 alelles, indicated by the H index, affect the inﬂorescence characteristics and broccoli and cauliﬂower yields.


Introduction
Brassica crops include several interesting species which are strictly related to crop wild relatives (CWRs) during their domestication process [1]. The Mediterranean region represents one of the main domestication and diversification centers of Brassica genus, in particular in Sicily where the cytodeme is represented by several wild relatives such as Brassica macrocarpa Guss., B. villosa Biv., B. rupestris, and B. incana [2].
The brassica genus includes three diploids (2n) (AA, BB, CC) and three tetraploids (4n) (AABB, AACC, BBCC) main species as described in the U's triangle model [3]. The B. oleracea complex species (n = 9) belongs to genome C (n = 9) and it represents the primary gene pool of the Brassica genus. This genus shows high genetic variability due to the genetic self-incompatibility characterizing the landraces and their CWRs and to several domestication processes [4]. Genetic diversity of B. oleracea is shown by the several varieties obtained by different domestication processes in a number of geographic areas which include broccoli, cauliflower, cabbage, kale, kohlrabi, savoy cabbage, and Brussel sprouts.
Brassica wild relatives could be a source of cytoplasmic male sterility (androsterility) for the development of hybrid seed of Brassica crops and they can provide genes for resistance to different diseases and pests and for these traits they can be used in breeding programs [5].
Cauliflower and broccoli are characterized by the floral induction of hypertrophic inflorescence [6]. Broccoli crop, as reported by Viani, originated from wild cabbage while cauliflower head derived from the improving process of broccoli addressed to the reduction of branches length, flower bud size, and the absence of their pigmentation.
Flower development genes were studied by Bowman et al. who found several genes involved such as apetala 1 and cauliflower in Arabidopsis [7]. These genes are closely related to members of the MADH-box genes family and a mutant copy of them is present in B. oleracea genome. Irish and Sussex characterized several floral phenotypes produced by the recessive homeotic apetala 1 (ap1) mutation in Arabidopsis and the homozygote for this mutation showed weak inflorescence affecting floral primordia formation [8].
Smith and King proposed a simple genetic model based on segregation of recessive alleles for BoAP1 and BoCAL candidate genes which showed differences in stage of arrest between cauliflower and Calabrese broccoli [9]. According to Smith and King's allelic distribution genetic model, the domestication process reduced the allelic diversity by promoting loci affecting the arrest of floral development which determined the inflorescence hypertrophy and then the domestication for cauliflower's curd phenotype; the Sicilian Purple was indicated as an important intermediate of this domestication process [10].
BoABI1, BoAP1, BoPLD1, and BoTHL1were designed to amplify genomic DNA region by Tonguc and Griffiths [11]. They were used to assess genetic similarity between several B. oleracea cultivars, belonging to three varietal groups (cabbage, cauliflower, and broccoli) while PBCGSSRBo39 was designed by Burgess et al. [12] to provide a useful molecular marker for crop improvement which was derived from shotgun sequencing programs.
Simple sequence repeats markers can be a useful tool to find genetic relationships among genotypes and related species provided from different countries. They can be used also as chloroplastic SSRs (cpSSRs) to avoid multiple gene copy number problems in polyploidy species [13].
In this study, the inflorescence morphometric traits of several accessions of broccoli and cauliflower landraces and commercial varieties, and Brassica relatives were measured [14], and additionally the five SSRs above cited were utilized to analyze the allelic variation among the accessions used and to associate them with inflorescence characteristics.

Materials and Methods
Plant material was represented by fifty-three accessions belonging to the Department of Agriculture, Food and Environment (Di3A) of the University of Catania-UNICT (Table 1). Seeds were sown in the first week of July in cellular trays placed under greenhouse conditions. The seedlings were transplanted after 5 weeks on the experimental farm of the University of Catania, (37 • 27 N, 15 • 40 E, 10 m a.s.l.) in single rows, with 1.0 m between the rows and 0.5 m between the plants along the rows, at crop density of 2 plants/m 2 . The experimental design was composed of four replicates (10 plants each) placed in randomized blocks as described by Branca et al. [15]; plants were grown in open fields.
For the accessions, inflorescence morphological data were registered following the International Board for Plant Genetic Resources (IBPGR) [16] descriptors related to the curd. Inflorescence morpho-biometric traits such as weight (CW), height (CH), diameter (CD1), shape (CS), angle of curvature (CA), and inflorescence stem thickness (CD2) were measured and calculated at the laboratory of Biotechnology of Vegetable and Flower Crops of the Di3A UNICT department. The inflorescence before anthesis was cut five centimeters before the first branch of inflorescence and for it the CW was registered by analytical balance, the CH and CD1 were calculated using a meter rule while CD2 was calculated using a caliber. The inflorescence shape (CS) parameter can be used to distinguish broccoli and cauliflowers from the CWRs and is derived from the ratio between CH and CD1. Curvature angle CA was registered with a goniometer by calculating the angle limited to between the central vertical inflorescence axes and the tangent to the extreme part of it. For morphological data, the mean values of the analyzed parameters of every accession were used to prepare a numerical matrix.
Genomic DNA was extracted from seedlings upon reaching the 6-8 leaved stage in young leaves tissues as reported by Tonguç and Griffiths utilizing the kit GenEluteTM Plant Genomic DNA Miniprep (Sigma Aldrich Inc.).
Extracted DNA was measured using a spectrophotometer Shimadzu at wavelengths of 260 and 280 nm, quantified by visual comparison on ethidium bromide-stained agarose gels. The final DNA concentration followed the protocol established by Branca et al. (2018) which includes 200 ng of template DNA.
The primers flanking SSR sequences (Table 2) were obtained in accordance with Tonguc and Griffiths (2004) for BoTHL1, BoAP1, BoPLD1, and BoABI1; concerning the PBCGSSRBO39 primers sequence, this was retrieved by Burgess et al. The position of the primers was checked using Assembly: GCA_0006955251.1 within Ensembl. Table 2. List of primers utilized with their sequences and chromosome position.

GenBank
Primers Name SSR Motif Primer Sequence (Forward, Reverse) Chromosome Five SSRs primers used were chosen following Branca et al., selecting them from ten primers, performed by Branca et al. for phylogenetic analysis and to assess the genetic similarity between several B. oleracea cultivars and wild Brassica species, belonging to two varietal groups (cauliflower and broccoli) as well as to estimate genetic divergence using F ST statistic; broccoli cultivars clustered with cauliflower cultivars as predicted and wild species showed major genetic differences [13].
The basic local alignment search tool (BLAST) was performed to check amplicon size and to compare results with amplificated sequences registered in an online database which was represented by BLAST (version 1.17) and Ensembl. The Uniprot database (release 2021, version 3) was used to study encoding regions close to the gene of interest.
The SSRs studied are located in different regions of the plant genome: BoABI1 is located in chromosome 1 region: 1,229,915,511-12,992,170 within the gene Bo1g041870 coding the ABI1 protein. The second SSR BoTHL1 is located on chromosome: 17,254,558: 17,255,176 within the Bo9g058820 gene, a homolog of thioredoxin 3 in Arabidopsis thaliana. The microsatellite PBCGSSRBo39 is located inside the Bo7g105720 gene on chromosome 7, BoAP1 is located inside chromosome 6: 33,883,667-33,887,357 inside the Bo6g108600 gene, one of MADS-box gene family members (Ap1Like).
BoPLD1 marker is located in the fifth chromosome in B. oleracea from 46,037,340 bp to 46,037,606 bp.
After DNA purification, PCR-based amplification was performed in 20 µL of final volume. The reaction mixture was composed of 200 ng of DNA template, 200 µM of each dNTP 3.75 mM MgCl2, 1X Taq DNA polymerase buffer, and 2 mM Primer according to Branca et al. (2018). DNA amplification was conducted in a Perkin Elmer 9700 thermocycler (ABI, Foster City, CA, USA) with the following parameters: initial denaturation at 94 • C for 5 min, followed by 35 cycles of denaturation at 94 • C for 30 s, primer annealing at 50 • C for 1 min, and extension at 72 • C for 1 min, with a final extension at 72 • C for 7 min. At the end of reaction, amplicons were stored at 4 • C. PCR products were loaded into 4% agarose gels (UNILAB Life Science, Taipei, Taiwan) and the electrophoresis run at a voltage of 100 V for 5-6 h in 1 X TBE buffer [15]. Capillary electrophoresis was performed using ABI PRISM 3130 Genetic 191 Analyser (Applied Biosystems, Waltham, MA, USA) as described by Branca et al. (2013) and Branca et al. Fragment sizes were determined by the GeneMapper 3.7 software (Applied Biosystems, Waltham, MA, USA). The allele peaks were checked by performing capillary electrophoresis and also checked using GeneMapper software. Each allele peak was manually rechecked by the operator.

Data Analysis
Allelic detection occurred in coding allelic status on the basis of their molecular weights using numeric scores: 2 (homozygosity), 1 (heterozygosity), and 0 (absence of any allele).
Allelic frequency data were elaborated to calculate heterozygosity index (H) which indicates the frequencies of heterozygosity in a population; an H value close to 1 suggests a large degree of heterozygosity within the populations while an H value close to 0 suggests homozygosity. Statistical analysis was performed to evaluate the correlation between the heterozygosity index for each locus and the inflorescence morpho-biometric traits as CW, CD1, CD2, and PC1.
For statistical analysis, the main inflorescence morpho-biometric characteristics, with exception of CS were used to calculate the Principal Component Analysis (PCA) to obtain a single parameter to summarize the inflorescence characteristics. CS was discarded due to the origin of this parameter which is derived from CH and CD1. PCA was performed using RStudio software 3.6.3 and a linear regression model was used to obtain information about the relationships between the heterozygosity index of each locus and the inflorescence bio-morphometric traits. PCA data were scaled to have unit values.

Results
Inflorescence morphometric traits CW, CH, CD1, CD2, and angle of curvature CA were registered for the Di3A accessions establishing a morphological database ( Table 3).
The inflorescence morphometric variance was detected using data on plant biomorphometric parameters recorded on Sicilian broccoli and cauliflower landraces and their F1 hybrids which show a large diversity among the genotypes.
The curd weight showed higher values for cauliflower landraces and F1 hybrids, than the CWRs analyzed, which registered lower values. Among cauliflower accessions, CV 171 Menhir F1 recorded the highest value (10,958 g); landrace curds weighed less than those collected from F1 hybrids and that explain the worldwide diffusion of these genotypes due to their yield. Curd diameter (CD1) was related to CW and, as registered, showed lower values in CWRs accessions while cultivated accessions with higher CD1 values, such as CV 99 S2 B recorded the largest curd diameter, respectively 21.1 cm (Table 3).
Broccoli types showed an elongated inflorescence as shown by CS values while cauliflowers showed more compact and flattened curd such as compared to the CWRs. Broccoli landraces showed the highest CS values compared to commercial cultivars.
The curvature angle (CA) also showed the large phenotypical variability among broccoli and cauliflowers landraces and hybrids F1, and their CWRs. CA distinguish well cauliflowers from broccoli and CWRs; cauliflowers were characterized by the highest CA value. Broccoli accessions showed lower CA values than cauliflowers. CWRs were characterized by the absence of the hypertrophic inflorescence developed from the apical meristem and they showed the lowest CA values; B. macrocarpa accession Favignana 1 showed the lowest CA value (7 • ).
Phenotypical variability was explained using a correlation model for each bio-morphometric descriptor; PC1 showed 46.75% of the total variance among the accessions (Table 4) and it is significantly correlated to CW, CD1, CD2 and CA. PC2 overlaps with CH which is one of the major traits affecting inflorescence morphology and therefore it was not used for this analysis.  Morphometric traits were subsequently elaborated and correlated to the genetic data by statistical analysis.
Each SSR locus exhibited a different number of alleles among the accessions studied: BoTHL1 showed eight alleles, PBCGSSRBo39 eleven alleles, BoPLD1 six alleles, BoAP1 showed twelve alleles, and BoABI1 nine alleles. Allelic data were processed to measure genetic diversity for each locus within the different Brassica accessions examined and to calculate the H index (Table 5). The correlation between CW and the five locus H index did not show significant p-value, although the BoPLD1 one was weakly significant (smaller than 0.10). On the other hand, significant correlations were observed among BoPLD1 H index and CD1, CD2, and PC1 (Table 5). The negative sign of the estimate coefficient confirms the association between the heterozygosity index and BoPLD1; when the H index increases the size of the inflorescence and the thickness of the stem decrease. The analysis also showed no significant correlation between the H index of the other loci (BoAP1, BoTHL1, BoAB1, PBCGSSRBo39) and inflorescence characteristics.

Discussion
The Di3A core collection describes the evolution of the domestication process from the Brassica wild species (n = 9) to the broccoli and cauliflower crops by comparing the main morphometric traits of the inflorescence and the allele diversity of the molecular primers utilized during human selection.
The domestication process is explained by the use of five SSR primers which show a wide range of alleles among the growing and wild species belonging to the B. oleracea complex species (n = 9). Some alleles, useful for increasing the inflorescence size, were unconsciously subject of selection by the growers in order to fix the hypertrophic inflorescence of broccoli and cauliflower. In addition, the broccoli and cauliflower domestication processes have been affected by the genetic flux among the Brassica wild relatives (n = 9), and the first domesticated sprouting broccoli was permitted to enlarge the inflorescence size gradually and define its shape of the hybrids F1 of broccoli and cauliflower [17]. In Branca et al. Fst was calculated in order to measure the genetic distance among accessions; the genetic diversity shown by the five SSRs primers utilized in relation to the B. oleracea complex species (n = 9) accessions, permitted us to classify them in relation to their domestication process (CWRs, landraces of cauliflowers and broccoli and their hybrids F1. MADS-box genes family includes several transcriptional factors involved in the growth and development of the inflorescence after its reproductive induction, flowering time, fruit development, and ripening [18]. During the last decades, several genomic studies were reported to explain the role played by some homeotic genes involved in the development of the hypertropic inflorescence, called head for broccoli and curd for cauliflower. Several genes such as apetala 1 (AP1) were reported to be involved in the inflorescence structure controlling its meristematic development. The transcript of AP1 gene (RefSeq ID: XP_013590290.1) was described by Sheng et al., 2019 as showing high levels of expression in different tissues and in particular in the curd and the flower [19]. These genes are related to the development of the reproductive organs and they belong to the MADH-box genes family [20]. BoCal is one of the related genes involved in curd formation; the mutant alleles seem to stop flower development and a simple genetic model has been proposed [9]. BoPLD1 marker is located in the fifth chromosome of B. oleracea from 46,037,340 bp to 46,037,606 bp in an untranscribed region (accession: LR031877), near the region encoding Phospolipase D (UniProtKB-A0A3P6FGA7). This catalytic enzyme is encoded from BOLC5T33808H gene and is involved in glycerolphospholipids hydrolysis at the terminal phosphodiesteric bond.
Taking into consideration the stem diameter (CD2) and the diameter of the inflorescence (CD1), the correlation between their values and the H index is significant only for the BoPLD1 locus while it is weakly significant with respect to the weight of the inflorescence ( Table 5). The analysis of the negative correlation coefficient between these two inflorescence morphometric traits (CD1 and CD2) permits us to deduce that the more the H index associated with BoPLD1 increases, and the related locus tends to show higher heterozygosity, the more such a parameter affects the inflorescence size. This could be correlated with the observation that in the CWRs there is greater heterozygosity than in the hybrids F1 but the size of the inflorescence, its stem, and its weight decrease for the former. The sequence hosting the microsatellite placed in the initial portion of the gene Bo5g126670, just before the first exon, could not exclude the presence of a repeat affecting the transcription of the gene itself.
The data acquired consent to delineate the next steps of this study sequencing the polymorphisms present in the upstream region of the Bo5g126670 gene that could be involved in the inflorescence hypertrophy of the B. oleracea complex species (n = 9). These variations could be used for marker-assisted selection (MAS) and for individuating in advance, during the breeding program utilized CWRs-the individuals who express hypertrophic inflorescence are an object of interest for further field evaluation.

Conclusions
BoPLD1 marker heterozygosity (H index) shows significant interaction with several inflorescence morpho-biometric characteristics and when BoPLD1 alleles tend to homozygosity an increase of inflorescence and curd size are observed. These results permit us to continue to investigate by sequencing these primers to individuate the SNPs useful for distinguishing the broccoli types with hypertrophic inflorescence during the organic breeding programs. The crossing plans among the broccoli breeding lines and the Brassica wild relatives, aiming to transfer forgotten alleles during the domestication process, will be useful for increasing the resistance against biotic and abiotic stresses, and for nutritional, organoleptic and nutraceutical traits. The molecular marker will reduce the cost of evaluation field transplanting with only the selected individuals expressing the broccoli inflorescence phenotype, reducing the number of individuals to grow and to analyze.

Conflicts of Interest:
The authors declare no conflict of interest.