Fine mapping of the tiller inhibition gene TIN5 in Triticum urartu

A tiller inhibition gene TIN5 was delimited to an approximate 2.1 Mb region on chromosome Tu7 that contains 24 annotated genes. Grain yield in wheat (Triticum aestivum L.) is a polygenic trait representing many developmental processes and their interactions with the environments. Among them, tillering capacity is an important agronomic trait for plant architecture and grain yield, but the genetic basis of tiller formation in wheat remains largely unknown. In this study, we identified a tiller inhibition 5 (tin5) mutant from ethyl methane sulfonate treated G1812 (Triticum urartu Thumanjan ex Gandilyan). A mapping population was constructed with tin5/G3146. Based on the sequence differences between G1812 and G3146, large insertions and deletions (≥ 5 bp) were selected and verified, and a skeleton physical map was constructed with genome-wide 168 polymorphic InDel markers. Genetic analysis revealed that the low-tiller phenotype was controlled by a single recessive locus, which we named TIN5. This locus was mapped to a 2.1-Mb region that contained 24 annotated genes on chromosome Tu7. Among these annotated genes, only TuG1812G0700004539 showed a non-synonymous polymorphism between tin5 and the wild type. Our finding will facilitate its map-based cloning and pave the way for an in-depth analysis of the underlying genetic basis of tiller formation and regulation patterns.


Introduction
Wheat (Triticum aestivum L.) is an important crop for global food security, providing about 20% of the calories and dietary proteins consumed by humans (FAOSTAT 2017). Wheat yield is usually dissected into three component traits, including productive tiller number per unit area (PTN, also referred to as spike number per unit area), number of seeds per spike (SPS), and seed weight (SW) (Ma et al. 2007), among which PTN is mainly affected by tiller number per unit area (TN) (Elhani et al. 2007). As each tiller has the potential to form a fertile inflorescence, the number of tillers is a critical determinant of grain yield (Sakuma et al. 2019). Additionally, tiller number per plant (TNP) is also a crucial trait for plant architecture (Naruoka et al. 2011). Despite its importance for grain yield, the molecular basis of tiller initiation is largely unknown in wheat since tillering is a complex trait controlled by multiple genes and their interactions with environmental factors (Kebrom et al. 2012). Therefore, it is important to map and isolate genes responsible for tiller formation in wheat.
Yaoqi Si, Qiao Lu and Shuiquan Tian have equally contributed to this work. 1 3 most well-known tillering-related gene in maize inhibits the growth of axillary organs under the high expression condition (Doebley et al. 1997). However, TaTB1 mainly increases the spikelet number per spike by promoting inflorescence branching (Dixon et al. 2018) and regulates plant height and stem internode length in wheat (Dixon et al. 2020). These findings suggested that the genetic control of tiller formation in wheat is somewhat different from that of rice and maize.
In wheat, at least six tiller inhibition genes have been reported, such as tin1 (Richards 1988;Spielmeyer and Richards 2004), tin2 (Peng et al. 1998), tin3 (Kuraparthy et al. 2007), ftin (Zhang et al. 2013), dmc , and TIN4 . One high-tillering dwarf mutant NAUH167 was characterized. Its TNP and plant height were controlled by a partial recessive gene mapped to the short arm of chromosome 2D flanked by markers Xcfd11 and Xgpw361 (Xu et al. 2017). Moreover, a number of quality trait loci (QTLs) affecting tiller number have been reported in wheat, including QTn.mst-6B (Naruoka et al. 2011), QMTN.sicau-4D (Hu et al. 2017, QPtn.sau-4B , cqTN-2D.2 (Ren et al. 2018), Qltn.sicau-2D , and Qetn-sau-1B.1 (Liu et al. 2020). Additionally, several genes were reported to associate with tiller formation in wheat. The overexpression of tae-miR156 in common wheat (cv. Kenong199) led to an increased TNP (Liu et al. 2017). Conversely, PHYTOCHROME-INTER-ACTING FACTOR-LIKE (PIL) family transcription factors has been reported as depressors of tillering in wheat, overexpression of TaPIL1 reduces wheat TNP . Similarly, down-regulation of TaPIN1s, the auxin efflux carrier, increased the tiller number per plant of wheat . TaD27, the ortholog of rice Dwarf27 (D27), encoding an enzyme involved in the strigolactone biosynthesis in wheat, played a critical role in regulating wheat tiller number. TaD27-RNAi wheat plants had more tillers, and TaD27-B-OE wheat plants had fewer tillers. Unlike rice D27, plant height was not affected in the transgenic wheat plants suggesting the divergent functions of D27 and its wheat ortholog (Zhao et al. 2019).
Although a number of QTLs responsible for the tiller number have been reported, the genes underlying these loci have not been isolated and characterized. The compensatory effects of homoeologous genes and the complexity of wheat genome are the big restriction for gene cloning studies of common wheat. The diploid wheat species Triticum urartu (AA, 2n = 14) is the wheat A-genome progenitor. Compared with common wheat, T. urartu has a smaller genome, making it an ideal model plant for studying complex agronomic traits. Moreover, the genome sequence of T. urartu acc. G1812 with 41,507 protein-coding gene models was released (Ling et al. 2018), which facilitates the map-based cloning of genes from T. urartu.
In this study, we reported a reduced tiller mutant tin5 obtained from ethyl methane sulfonate (EMS)-treated accession G1812, from which genome was assembled. Compared with the wild type, tin5 was defective in the outgrowth of axillary buds, and had less tillers and aboveground biomass. We finely mapped the low-tiller phenotype locus (herein named TIN5) with large F 2 segregating populations and predicted its candidate genes. This study provides a foundation for further cloning of TIN5 and its uses in wheat breeding.

Plant materials
The mutant tin5 was obtained from 0.3% (v/v) EMS (Sigma-Aldrich) mutagenesis population (approximately 7000 M 2 individuals) of T. urartu accession G1812 (PI 428,198), from which the genome was assembled (Ling et al., 2018). To map TIN5, an F 2 population derived from the cross of tin5 and T. urartu accession G3146 (PI 538,737) was developed. The F 2:3 families subsequently produced by self-pollination of recombinant individuals selected from the F 2 population were applied for fine mapping.

Growing conditions and phenotypic evaluations
The F 2 population and F 2:3 families were evaluated at Zhaoxian County in Hebei Province, China (37° 50′ N, 114° 49′ E) in wheat growing season 2016-2017 and 2017-2018, respectively. All lines were sown as one-meter row plots with 11 plants, 10 cm between plants, and 25 cm between rows. The seedlings were covered with plastic films in winter to avoid winterkill. The field management was followed by the local wheat production practices.
The TNPs of 240 individual plants randomly selected from 2545 F 2 plants were counted manually at the grain filling stage in the field and their whole aboveground biomass (AGB) including stem/spike were measured at the maturity stage after air-dried. For the rest F 2 plants and their selected F 2:3 families, we divided the individuals into the mutant type and the wild type by visual inspection. Briefly, a plant with 20 tillers or less and reduced plant height was characterized as the mutant type, whereas the others were considered as the wild type (high tiller number). The phenotype of the key recombinants was confirmed by the phenotype segregation analyses of their corresponding F 2:3 families.

Molecular marker development
Genomic DNA was extracted from fresh leaves of young seedlings according to the CTAB method (Chatterjee et al. 2002). The accession G3146 was re-sequenced for a 1 3 sequencing depth of 10 × genome coverage, and the highquality reads were aligned to the G1812 reference genome (Ling et al. 2018). Genome-wide single nucleotide polymorphism (SNP) and InDel between G3146 and G1812 were called using the HaplotypeCaller module (McKenna et al. 2010). Based on the flanking sequence of InDels (sequence difference ≥ 5 bp), 440 InDel primers were developed. Additional simple sequence repeat (SSR) markers and SNP primers linked to TIN5 were developed according to the G1812 genome sequence (Ling et al. 2018). The 10 μL PCR system contained 5 μL 2 × Taq PCR Starmix (GenStar, China), 3 μL ddH 2 O, 1 μL of 80 ng/μL genomic DNA, and 0.5 μL of 10 μM each primer. The PCR reactions were carried out as described by Si et al. (2021). All the primers used in the study are listed in Supplementary Table S1.

Genetic map construction and QTL detection
The genetic map was constructed with software JoinMap 4.0 using the Kosambi mapping function. QTL analysis was performed by inclusive composite interval mapping (ICIM) using the software IciMapping 4.1 (Zhao et al. 2020) with 1000 permutations at P = 0.05 in the F 2 population. A logarithm of the odds (LOD) threshold of 2.5 was set to declare a significant QTL.

Sequence analysis and RT-PCR
To identify the genetic lesion in tin5, we explored the expression patterns of candidate genes within the TIN5 mapping interval with the existing G1812 expression data (leaf, spike, and root) (Ling et al. 2018). Moreover, we analyzed the coding sequences, 3'-UTR region and approximately 1.6-kb promoter sequence upstream of candidate genes within the mapping interval from both G1812 and tin5 with the primers listed in Table S1.
The expression analysis was performed using tiller nodes from the greenhouse seedlings. The seeds of G1812 and tin5 were sown into pots (dimensions 25 × 25 × 30 cm) containing the mixture of vermiculite and peat (volume:volume, 1:1) supplied with the slow release fertilizer (Osmocote, 14-14−14) according to the manufacturer's instructions. The plants were grown under the glasshouse conditions (16-18 °C; 16 h light: 8 h dark) for 4-5 weeks. The tiller nodes were sampled from tin5 and G1812 seedlings with three replicates when the fourth tiller nodes were fully developed. Total RNA was extracted from tiller nodes using TRIzol Reagent (Invitrogen, USA). After removing genomic DNA contamination by DNase I (NEB, USA), SuperScript II (Invitrogen, USA) was utilized for first-strand cDNA synthesis according to the manufacturer's instructions. qRT-PCR experiments were performed using a total volume of 10 μL with 4 μL cDNA template, 1 μL gene-specific primer (0.5 μL of each 10 μM sense and 10 μM antisense primers) and 5 μL SYBR Green Master mix (TaKaRa, Japan) on LightCycler 480 Real-Time PCR System (Roche Diagnostics, Switzerland) according to the manufacturer's recommendations. The relative gene expression level was calculated using the 2 −ΔΔCt method with three biological replicates. The ACTIN gene was used as the endogenous control (Zou et al. 2018). Primers for RT-PCR are listed in Table S1.

Phenotype and genetic analysis of tin5
To investigate the molecular basis underlying wheat grain yield, we generated a population with approximately 7000 M 2 individuals in a T. urartu accession, G1812 (Ling et al. 2018), by the EMS mutagenesis, and screened for mutants exhibiting altered important agricultural traits, including tiller number and plant height. One such mutant, a tiller inhibition 5 (tin5) mutant was selected for detailed studies. Comparing to G1812, young tin5 plants were compromised in its tillering ability in the field (Fig. 1a). The decreased tiller number per plant (TNP) of tin5 was associated with the inhibition of tiller bud outgrowth rather than with fewer tiller buds (Fig. 1b-d). At the heading stage, tin5 exhibited reduced TNP and plant height compared with G1812 (Fig. 1e). The average TNPs of G1812 and tin5 were 44.87 and 17.53, respectively.
A normal phenotype was observed for all F 1 plants derived from the cross between tin5 and G3146, indicating that the tiller inhibition phenotype is a recessive trait. The F 1 plants were self-pollinated to generate F 2 mapping populations (total 2545 plants). Then, a genetic linkage analysis was performed by using 240 plants of the F 2 population, and the tiller habits of the corresponding individuals were recorded and verified in their F 2:3 families in the field at Zhaoxian, Hebei Province. The TNP distribution of the 240 F 2 plants was multimodal (Fig. 1f). There were 188 normal plants and 52 reduced tiller plants in this F 2 population, which fitted a 3:1 Mendelian ratio (188:52; χ 2 = 1.25 < χ 2 0.05 , 1 = 3.84). This result indicated that the reduced tiller phenotype is controlled by a single recessive gene.

InDel variation analysis between G1812 and G3146
In order to develop polymorphic markers between tin5 and G3146, we re-sequenced G3146 and aligned high-quality reads to G1812 reference genome (Ling et al. 2018). Totally, 29,565 InDels between G3146 and G1812 were obtained by the HaplotypeCaller module. Among them, 1,747 InDels were larger than 5 bp (Table 1). Based on the flanking sequences of partial InDels (≥ 5 bp), 440 InDel 1 3 markers were designed for covering all seven chromosomes, and 168 (37.8%) markers were found to be polymorphic in 5% agarose gel visualization between tin5 and G3146 (Table 1, Fig S1).

Single-marker analysis and molecular mapping of TIN5
Considering tillering is greatly affected by the environment, aboveground biomass (AGB) of the F 2 population was also selected as a proxy for tiller number in order to more accurately map the TIN5 locus in the study. Consistent with the tiller number trait, the AGB trait of the 240 F 2 plants also presented continuous distribution (Fig. 1g). To map TIN5, the same 240 F 2 plants of tin5/G3146 cross for genetic analysis were genotyped with 70 polymorphic InDel markers (ten markers on each chromosome). Among them, markers (7 T-33, 7 T-74 and 7 T-81) located on chromosome Tu7 were significantly (P < 0.05) associated with TNP and AGB (aboveground biomass) by single-marker analysis performed on WinQTLCart version 2.5 ( Table 2). The remaining 18 polymorphic InDel markers on chromosome Tu7 were then used to genotype the 240 F 2 plants. Subsequently, we chose 16 markers to construct the chromosome Tu7 genetic map of this F 2 population with a total genetic length of 33.3 cM (Fig. S2). With the genetic linkage map, QTL mapping for TNP and AGB was conducted with the additive model of inclusive composite interval mapping (ICIM). Finally, we found that QTNP.ucas-7A for TNP was coincident with QAGB.ucas-7A for AGB, which explained 18.28-26.59% and 26.65-33.80% of the phenotypic variations, respectively, indicating TIN5 was mapped to the interval of 12.5-17.5 cM flanked by markers 7 T-S117 and 7 T-S187 (Fig. S2, Fig. 2a).

Fine mapping of TIN5
The flanking markers 7 T-S117 and 7 T-S187 were used to screen the remaining 2305 F 2 plants, and 46 recombinants were identified (Fig. 2b). To further narrow down the genetic region of the TIN5 locus, we systemically genotyped 1,408 self-pollinating progenies (F 3 ) of the recombinants with the flanking markers (7 T-S117 and 7 T-S187), and with 7 newly developed polymorphic markers and phenotyped F 3 individuals by visual inspection. Taken together, TIN5 was identified to co-segregate with the marker 7 T-4188 and was placed within a 2.1 Mb physical interval delimited by the markers 7 T-S167 and 7 T-S174 (Fig. 2c).

Mining of candidate genes
In the approximately 2.1 Mb genomic region defined by markers 7 T-S167 and 7 T-S174, a total of 24 predicted genes (TuG1812G0700004532 to TuG1812G0700004555) were found on the Triticum urartu v2.0 reference genome (Table S2). To identify the candidate genes for TIN5, we analyzed the orthologues of these 24 genes in rice. Among these genes, TuG1812G0700004540 (TuD27) encoded betacarotene isomerase D27 was the ortholog of rice Dwarf 27 (D27) (Fig. S3). Rice D27, an iron-containing protein participating in the biosynthesis of strigolactones, plays an important role in the regulation of tiller number in rice (Lin et al. 2009). Meanwhile, a recent report has shown that TaD27-RNAi wheat plants had more tillers, and TaD27-B-OE wheat plants exhibited fewer tillers, suggesting that TaD27 (the ortholog of D27 in bread wheat) also plays a critical role in wheat tiller development (Zhao et al. 2019). Thus, TuD27 may be the candidate gene of TIN5. To confirm that whether TuD27 is responsible for the few tiller phenotype of tin5, we analyzed the promoter (1.6-kb), exon, intron and 3'-UTR region sequences of TuD27, and checked the expression levels of TuD27 in the tiller nodes at the four-leaf stage of tin5 and G1812. However, neither sequence nor expression differences were detected for TuD27 between tin5 and G1812 (Fig. S3, S4a).  On the other hand, eight out of other 23 genes beside TuD27 were not expressed in all three tissues investigated (leaf, spike, and root), whereas the remained 15 genes showed expression at least in one tissue (Table S3). Notably, we did not detect any changes both in these 15 expressed genes and eight unexpressed genes between G1812 and tin5, except TuG1812G0700004539. The tin5 mutant had a G → T base change causing single amino acid substitution from Aps to Tyr in TuG1812G0700004539 (Fig. 3a). Furthermore, we analyzed the expression abundance of TuG1812G0700004539 and did not observed significant difference between tin5 and G1812 (Fig. 3b). TuG1812G0700004539 encodes a putative 674 amino acid protein containing a typical pentatricopeptide repeat structure. OGR1, the homologue of TuG1812G0700004539 in rice, encodes a pentatricopeptide repeat protein containing the DYW motif, is essential for RNA editing on five mitochondrial transcripts in rice. The ogr1 mutant exhibited retarded growth, dwarfism and reduced tiller phenotype (Kim et al. 2009), which is similar to the morphological phenotypes of tin5. From these results, we speculate that TuG1812G0700004539 may be the candidate gene for TIN5.

T. urartu with smaller genome facilitates the mining of genes controlling complex agronomic traits
Although promising progress has been made in the assembly and annotation of the allohexaploid wheat genome, compensatory effects of homoeologous genes, and its large and complex genome are still the big restriction for gene cloning and functional studies in common wheat (IWGSC 2018). In addition, most yield-related traits of wheat, especially tiller number, are quantitative traits affected by multiple gene loci and environmental conditions. So, it is difficult to isolate genes for the tillering traits (Kebrom et al. 2012). Till now, none of the underlying genes for tillering trait have been cloned in wheat through the map-based cloning methodology . Therefore, researchers studied the homologues of the tillering genes cloned in model species, such as rice and maize, taking advantages of gene editing methodology. However, recent studies suggested that the homologous genes TaMoc1, a homolog of MOC1 in rice, and TaTB1, a homolog of TB1 in maize, mainly affect the 7T-S174 7T-S117  (Dixon et al. 2018;Zhang et al. 2015). These results indicated that the function of genes controlling tillering was differentiated between rice, maize and wheat. Therefore, it is of great significance to excavate the genes controlling tillering from diploid wheat relatives, such as T. urartu.
As the donor of wheat A subgenome, T. urartu is closely related to common wheat and has high genetic homology (Ling et al., 2018). Meanwhile, the smaller and simpler T. urartu genome compared to hexaploid wheat provides an opportunity to identify tiller mutants and clone corresponding genes in T. urartu. Here, we identified the tiller mutant tin5 from the EMS mutagenesis library of T. urartu accession G1812, finely mapped TIN5, and predicted the candidate genes. This research can be an example of studying genes for complex traits in T. urartu.

An InDel map facilitates mapping of mutated genes
In order to improve the efficiency of marker development in this study, we called the InDel variations at the whole genome level between G1812 (the wild type of tin5) and G3146, and developed a physical map with 168 InDel markers covering seven chromosomes of T. urartu (Table 1). Yu et al. (2016) constructed a T. urartu genetic map containing 926 molecular markers. Among them, 584 are DArT markers, which need special equipment to reproduce. Since our InDel markers were developed from InDels larger than five base pairs, they can be easily distinguished by agarose gels (Fig. S1). Moreover, this set of markers has been used to preliminary map several T. urartu mutant genes under G1812 genetic background (data not shown) in our laboratory. So, these InDel markers are good supplement for published markers, especially for those using G1812 (PI 428,198) as the material.

The candidate genes for TIN5
To date, a few tiller inhibition genes have been identified in wheat, such as tin1, tin2, tin3, ftin, dmc, and TIN4 . Meanwhile, several QTLs for tiller number were subsequently reported, such as QTn. ipk-1B, Qetn-sau-1B.1, cqTN-2D.2, QPtn.sau-4B, QMTN. sicau-4D, QTn.ocs.5A.1, and QTn.mst-6B (Liu et al. 2020;Ren et al. 2018;Wang et al. 2021). However, no genes or QTLs for tiller number were isolated and cloned. Here, we mapped a tiller inhibition gene TIN5 to a 2.1 Mb physical region on the chromosome Tu7 of T. urartu genome. Among the 24 candidate genes (Table S2), TuG1812G0700004540 (TuD27) is an ortholog of Dwarf 27 (D27) regulating rice tiller bud outgrowth (Fig. S3). Unexpectedly, no sequence difference was detected for TuD27 between tin5 and G1812 (Fig. S3). In rice, plastidlocalized enzymes encoded by D27, D10, and D17 play crucial roles in catalyzing the formation of carlactone, which is further catalyzed by cytochrome P450 to generate strigolactones involved in tiller regulation (Zhang et al. 2014). Therefore, we further analyzed the relative expression levels of TuD27, TuD10 (TuG1812G030000321), and TuD17 (TuG1812G0200004646) (the orthologs of D10 and D17 in rice, respectively) in the tiller nodes at the four-leaf stage between tin5 and G1812. No significant expression differences were observed for TuD17, TuD10 and TuD17 between tin5 and G1812 (Fig. S4). From these results, it is still hard to rule out TuD27 as a candidate gene. For instance, a candidate gene Csl for tin1 in hexaploid wheat was associated only with the variation in the length of a microsatellite repeat in the 5'-UTR. In the study, any SNPs at the locus would be difficult to detect via sequencing due to alignment errors, and the expression difference was only evident at some particular time-points during early stem elongation (Hyles et al. 2017 of all candidate genes might be helpful in the future work for determining the causal gene of low tillers of tin5. To anchor the candidate genes, we then sequenced other 15 genes, which were expressed in seedling leaves, roots or spikes (Table S3). Compared to the wild type G1812, only TuG1812G0700004539, a putative protein containing pentatricopeptide repeat, had a SNP change ( G 541 T ) leading to an amino acid substitution ( D 181 Y ) in the mutant tin5 (Fig. 3a). In rice, pentatricopeptide repeat proteins mainly affect chloroplast development, and defects of these genes led to growth retardation, reduced plant height and tillering, sterile phenotype, and leaf color changes (Kim et al. 2009;Toda et al. 2012). In contrast, the leaf color of tin5 did not change significantly compared with G1812 (Fig. 1a, e). To explore other possible candidate genes, we sequenced these eight unexpressed genes and found that there was no SNP for these genes between tin5 and G1812. Based on these observations, TuG1812G0700004539 is a strong candidate gene of TIN5.
Since the transgenic method is yet to be developed in T. urartu, future studies will include editing the orthologous genes of TuG1812G0700004539 in common wheat and elucidating the molecular mechanisms of tillering in wheat.
In summary, this study provided a set of low-density physical map of T. urartu constructed with agarose-resolvable InDel markers and finely mapped a new tiller inhibition gene TIN5 on chromosome Tu7, and predicted its causal gene. TuG1812G0700004539 has an SNP variation causing an amino acid change between tin5 and G1812. It is a strong candidate gene for reduced tiller phenotype. The identification and mapping of TIN5 will facilitate its uses in wheat breeding and the genetic dissection of molecular mechanism of wheat tillering.
Author contribution statement SZ and YS developed the mapping populations; YS, QL, ST, and JN carried out the experiments and analyzed the data; ST analyzed the candidate genes; MC and XL assisted in phenotyping, genotyping and field work; QG and XS analyzed the data of re-sequencing; YS wrote the manuscript; H-Q L and SZ designed the project and revised the manuscript.
Funding This research was jointly supported by grants from the National Natural Science Foundation of China (31971877) and the Strategic Priority Research Program of Chinese Academy of Sciences (grant no. XDA24010104-1).
Availability of data and material All data generated or analyzed during this study are included in the main text article and its supplementary files.
Code availability Not applicable.

Conflicts of interest
The authors declare that they have no conflicts of interest.
Ethical approval Not applicable.

Consent to participate Not applicable.
Consent for publication Not applicable.