Hongkun Yng,Yun Xio,Peng He,Dilong Ai,Qiosheng Zou,Jin Hu,Qiong Liu,Xiuln Hung,Ting Zheng,Goqiong Fn,b,c,*

a Crop Ecophysiology and Cultivation Key Laboratory of Sichuan Province,Sichuan Agricultural University,Chengdu 611130,Sichuan,China

b Key Laboratory of Crop Eco-Physiology &Farming System in Southwest China,Ministry of Agriculture and Rural Affairs,Chengdu 611130,Sichuan,China

c State Key Laboratory of Crop Gene Exploration and Utilization in Southwest China,Ministry of Science and Technology,Chengdu 611130,Sichuan,China

Keywords:Straw mulch-based no-tillage Wheat tillering Stable isotope tracing Transcriptome Metabolomics

ABSTRACT The moisture-conserving effect of straw mulch-based no-tillage (SMNT) is expected to increase fertile spikes and grain yield in environments with rainfall less than 200 mm.However,the mechanisms underlying the positive effect of SMNT on wheat tillering are not fully elucidated.A split-plot experiment was designed to investigate the combined effects of SMNT and cultivars on tillering of dryland wheat grown under both dry and favorable climates.Application of SMNT to a cultivar with 1–2 tillers exploited both tillering and kernel-number plasticity,increasing the mean grain yield by 20.5%.This increase was attributed primarily to an increased first-tiller emergence rate resulting from increased N uptake,leaf N content,and N remobilization from tillers to their grain.The second and third tillers,as transient sinks,contributed to the tiller survival rate,which depends on tiller leaf number.The increased total N uptake by SMNT also increased the dry mass yield of tillers and the C:N ratio,reducing the asymmetric competition between main stem and tillers.Owing to these beneficial effects,reduced mitogen-activated protein kinase (MAPK) and abscisic acid signals were observed under SMNT,whereas indole-3-acetic acid(IAA) signals and genes involved in DNA replication and mismatch repair were increased.These signals activated three critical transcription factors (the calmodulin-binding transcription activator,GRAS domain,and cysteine-2/histidine-2 family)and further increased rapid drought response and tiller maintenance after stem extension.Phenylpropanoid biosynthesis,sphingolipid biosynthesis,and galactose metabolism were most relevant to increased tillering under SMNT because of their critical role in drought response and lignin biosynthesis.Our results suggest that straw mulch-based no-tillage activates rapid drought response and improved wheat tillering by coordinating root N uptake,N remobilization,and asymmetric competition between main stem and tillers.

1.Introduction

Tillering is a crucial agronomic trait of cereal crops,because the number of green tillers is directly related to the number of fertile spikes,which is a critical component of grain yield [1,2].Among the three stages of tiller development(axillary meristem initiation,axillary bud development,and axillary bud outgrowth),the first stage is determined mainly by genetic factors [3],whereas the third is regulated mainly by environmental factors and management practices [4].Thus,efforts to increase fertile spikes via crop management should focus on the third stage.We previously reported [5] that straw mulch-based no-tillage (SMNT) conserved soil moisture and alleviated drought stress in the tillering stage,thereby increasing grain yield in a dryland farming system.However,the responses to SMNT of cultivars with contrasting tillering capability remain unclear.Excessive tiller production can increase infertile-tiller numbers,inducing resource competition between the main stem and tillers and increasing lodging probability[6,7].Cultivars with 1-2 tillers show higher yield potential under severe drought stress conditions(rainfall ＜200 mm)[8,9].Experiments employing the use of reduced tillering (e.g.,tin) genes [10]have shown that inhibition of tillering promotes the development of deeper roots,increases the tiller population,and increases the formation of large spikes under drought environments[8–10].This positive effect declined after grain yields exceeded 1.6 t ha1.Breaking the current yield bottleneck and maximize the tillerpromoting effect of SMNT in dryland farming systems will require knowledge about the combined effect of SMNT and cultivars on wheat (Triticum aestivuml L.) tillering and tiller survival.

Tiller outgrowth and survival are inhibited by drought stress[11].An inappropriate tillage method and residue management[12] that resulted in the eventual death of 50%–80% of tillers caused irreversible loss of assimilation and grain yield [10,13,14].In general,the main stem and first primary tillers contribute more to grain yield than late-emerging tillers,owing to asymmetric competitive advantages under drought,and such advantages have been associated with increased leaf number [15].Infertile tillers can also function as transient sinks of plant assimilates that can later be remobilized and transferred to the main stem via tiller nodes in wheat[16]and rice[17].The transport of water and nutrients between the main stem and tillers and through the vascular bundles of tiller nodes is critical to tiller development and survival[18].Thus,the supply and partitioning of C and N between main stems and tillers play a critical role in producing ear-bearing tillers[19,20].Grain yield was reduced by asymmetric competition between main stems and tillers for C and N[21,22],in consequence of later panicle differentiation of tillers than of the main stem.This advantage increases with time [15,23,24].We previously showed[5] that SMNT increased root N uptake,but whether increased N of tillers under SMNT can alleviate asymmetric competition for C and N nutrition between the main stem and tillers remains unclear.

Previous studies have investigated the roles of transcription factors,hormone signals,miRNA,and quantitative trait loci on tillering in rice,wheat,barley,and maize.In rice,MOC1,which encodes a GRAS family nuclear protein,promoted axillary bud outgrowth[1],and CsI transcript regulation affected carbon partitioning and tillering [3].Phytohormone signaling regulates tiller outgrowth and increases drought tolerance by mediating growth,development,and nutrient allocation between the main stem and tillers[25–28].Auxin can modify plant architecture and coordinate plant growth under drought stress [28].In rice seedlings,the tld1-D/OsGH3.13 encoding IAA-amino synthetase has been reported [28]to increase the expression of a late embryogenesis-abundant gene,increasing tiller number and drought tolerance.The reduction of free IAA promotes tiller production by expressing the IAAglucose synthase gene OsIAGLU,and exogenous IAA application reduces tiller number [25,26].TaD27-B,regulating enzymes involved in strigolactone biosynthesis,induces tiller growth by regulating axillary bud outgrowth [29].Abscisic acid (ABA)down-regulated strigolactone biosynthesis genes,thereby promoting tiller production[30,31].LAX2 is a gene that acts together with LAX1 to regulate axillary meristem initiation [32].These studies focused on tiller buds,and little is known about the critical role of tillering nodes in tiller survival and mortality.The tillering node is of great interest in investigating signal transduction and asymmetric competition between main stem and tillers,because tillers are grown on non-elongated tiller nodes that function as a nutrition and signal transporter mediator,regulating tillering and mortality.

We previously reported[5]that SMNT conserved soil moisture,alleviated drought stress in the tillering stage,and increased fertile spikes and grain yield in dryland farming systems.Cultivars adapt to soil drought via both ABA-dependent and ABA-independent signal-transduction pathways [33].The objectives of the present study were to test the hypothesis that SMNT increases wheat tillering by rapid drought response,root N uptake,N remobilization,and asymmetric competition between main stems and tillers for C and N nutrition.The experimental approach employed14C and15N isotope-based tracing to characterize the asymmetric competition between main stems and tillers,and RNA-seq and metabolomics to identify tillering-associated genes and metabolites under SMNT.

2.Materials and methods

2.1.Experimental site and experimental design

The experiment was conducted at the Renshou Experimental Station (33°19′N,120°45′E) of Sichuan Agriculture University,which is located in a dryland region with cumulative rainfall ＜200 mm during the wheat season (Fig.1A).The parent soil in this area is purple lithomorphic soil[34]and is characterized by low levels of organic matter,moderate levels of root-available soil K,and a clay texture (Fig.1B).Rainfall records from 1997 to 2017 reveal that severe droughts have become more frequent [5].

During typical dry (2017–2018) and favorable (2018–2019)growing seasons,a split-plot design,which included SMNT and no-mulch no-tillage (NM),was used to assess the effect of SMNT on tillering and tiller survival of wheat cultivars (CN16,CN30,and CM104)grown in a wheat-maize rotation system.The selected cultivars had similar plant height (78–85 cm) and growth durations(181–186 d)but varied in tiller score[35]:the mean numbers of tillers present at the stem extension stage were 1.8,1.9,and 2.6,respectively.Six treatments with three replications were randomly distributed in the field,with each plot measuring 18 m long and 12 m wide.During both growing seasons,straw mulch composed of crushed maize stalks was applied after maize harvest (August 25),and wheat seeds were sown on October 30 to yield a density of 225 seedlings m2[5].N,P,and K were deployed at the rates of 120 kg ha1urea (46% N),75 kg ha1superphosphate (12%P2O5,12%S),and 75 kg ha1potassium chloride(50%K2O),respectively;100% of the P and K and 60% of the N were applied at sowing,with the remaining 40%of the N applied at the stem extension stage.No irrigation was employed in either growing season.Pest management followed local practices.

2.2.Sampling and measurements

2.2.1.Dynamic tiller population,critical agronomic traits,and grain yield

Fifteen adjacent wheat plants were collected from the middle row of each plot at the tillering,jointing,booting,heading,grainfilling,and maturation stages.After removal of all roots,each plant was divided with scissors into main stem and first,second,and third tillers.The leaves,grain,and main stem were separated,held at 105 °C for 30 min,and dried at 70 °C to constant weight.The ratio of dry mass yield of fertile tillers to total dry mass yield was calculated as previously described [36].Plant number (Np),maximum number of active tillers (Nmax),and number of ears per m2(Ne) were measured in a randomly selected 4 m2square of each plot 37,49,and 157 days after sowing(DAS),as previously described [37].Tiller survival rate (TSR) and tillering capability(TC) were calculated as follows:

Fig.1.Environmental conditions(A)and soil chemical characteristics(B)of the study site during dry and favorable growing seasons.The red dotted line represents the mean air temperature.The pink area represents diurnal temperature.SOM,soil organic matter;Av-N,plant-available soil nitrogen;Av-P,plant-available soil phosphorus;Av-K,plant-available soil potassium.

Tillering is driven by thermal time (τ,°C d),assuming a linear response to temperature and a base temperature of 0 °C,which is widely accepted for the wheat [37].Depending on the tillering curves of each plot,the thermal times required to achieve a plant number (τMS,Np),plant with 1 tiller (τT1,NT1),plant with 2 tiller(τT2,NT2),maximum number of active tillers(τBS,Nmax),and number of ears per m2(τES,Ne) were estimated by fitting a previously described model [37] to experimental values of green tillers per m2and thermal time(τ,°C d).Tiller emergence rate(TER)and tiller mortality rate (TMR) were calculated as follows:

The N uptake of each tiller was calculated as the product of dry matter yield and corresponding N concentrations[36].The amount of N remobilization was calculated as the difference between the total N uptake of each tiller at anthesis and that of the straw,leaf,and chaff at maturity.Grain yield was measured by harvesting a representative plot of 4 m2for each treatment at crop maturity.The fertile spikes in 4 m2representative plots were also counted,and the grain from the spikes was threshed and air-dried.Wheat kernels in each spike (kernel number spike1) was calculated as the number of kernels from 15 wheat plants divided by the number of spikes collected from those plants.The air-dried kernels were used to measure 1000-kernel weight at a grain moisture content of 13.5%.Harvest index(HI)was calculated as the ratio of grain yield to total plant dry mass yield at maturity.N harvest index(NHI)was calculated as the ratio of grain N yield to total N uptake at maturity.

2.2.2.Stable isotope(δ14CO2and δ15N)tracing between main stem and tillers

The experimental approach employed14C and15N isotopebased tracing to characterize the asymmetric competition between main stems and tillers.A micro-plot design and δ14CO2tracing were used for each plot to investigate the asymmetric competition between the main stem and tillers for carbohydrates.An air mixture that contained14CO2was generated from NaOH (0.1 mol L1) and 74 MBq mL1NaH14CO3in a small polyethylene bottle.The resulting δ14CO2was added to an airtight rubber bag that contained 1 L air,and 2 mL H2SO4(1 mol L1)was injected into the bag to prevent the loss of14CO2.The experiment with14CO2-labeled air was conducted at 10:00 am when leaf photosynthesis achieved its maximum value.Finally,NaOH (1 mol L1) was used to remove12CO2,and the labeled air mixture,which contained 3.7 104Bq L1δ14CO2,was then injected into a confined plastic chamber(0.6 0.6 1 m).An infrared CO2analyzer (GXH-3010-E;Beijing Huayun Co.,Beijing,China) was used to monitor the CO2concentration in the chamber.One short-term (1 d) and two long-term(anthesis and maturation)assimilation period pulse labelings were performed on the fixed plants in each micro-plot,and they were controlled using a12CO2pulse-labeling method [38].

Five plants for each plot were collected from the δ14C-labeled micro-plot chamber at the tillering,stem extension,heading,and maturation stages.The samples were collected on bright,cloudless days(762±31 μmol m2s1of adaxial photosynthetic photon ux density),separated into their constituent parts,and oven-dried as described above.Water-soluble carbohydrates (WSC%) were extracted three times from each sample using 80% ethanol and an 80 °C water bath (30 min),and were quantified using a 96-well polystyrene plate,a Benchmark microplate reader (Bio-Rad,Hercules,CA,USA),and standards.The powdered samples were subsequently subjected to δ14C and C content (%) measurements using a stable isotope ratio mass spectrometer (Delta XL Plus EAIRMS,Thermo Finnigan,Waltham,MA,USA)and element analyzer(Flash SMART 2000 HT,Thermo Finnegan),respectively.The δ14CO2fixation rates of main stems and tillers and the δ14C yield gap between the main stems and tillers were calculated as follows:

A micro-plot design and δ15N-tracing were used to investigate the asymmetric competition between main stems and tillers for N.In each micro-plot,separation plates were installed (60 cm below ground,20 cm above ground)to separate 1 m2micro-plots,as previously described [39].N fertilizer was applied to the soil in each micro-plot using15N-labeled urea (10% abundance;Shanghai Research Institute,Shanghai,China) at the recommended rate(120 kg ha1),with 60% applied at sowing and the remaining 40%applied at the stem extension stage.P and K were independently applied at the recommended rate of 75 kg P ha1using superphosphate(12%P2O5,12%S)and potassium chloride(50%K2O).

Five plants were collected from the15N-labeled micro-plots at the tillering,jointing,heading,and maturation stages,separated into their constituent parts,and oven-dried,as described above.Their14N contents were measured using an Automatic Kjeldahl apparatus (FOSS-8400;Nils Foss,Copenhagen,Denmark),and δ15N was measured using a stable isotope ratio mass spectrometer(Delta XL Plus EA-IRMS,Thermo Finnigan).The amount of15N derived from fertilizer(%)and the δ15N yield gap and N partitioning ratio between main stems and tillers [40] were calculated as follows:

2.2.3.cDNA library construction and RNA sequencing of tiller node

Integrating RNA-seq and metabolomics was used to identify tillering-associated genes and metabolites under SMNT.Total RNA was extracted from non-elongated tiller nodes with TRIzol(Invitrogen,Carlsbad,CA,USA) and purified with the RNAeasy Total RNA Kit (Qiagen,Valencia,CA,USA).RNA concentration and integrity were measured with a NanoDrop spectrophotometer(ND-2000,Thermo Finnigan) and Bioanalyzer 2100 (Alignment;Santa Clara,CA,USA),respectively.A cDNA library was prepared with the TruSeq RNA Sample Preparation Kit v2 (Illumina,San Diego,San Diego,CA,USA),and then sequenced using Illumina HiSeq 2000 system (Illumina).Raw reads were filtered (Q ＞20)to obtain high-quality reads and then mapped onto the International Wheat Genome Sequencing Consortium (IWGSC) genome using Bowtie2 2.2.9[2].The expression of each gene was evaluated using RNA-Seq by Expectation-Maximization (RSEM 1.1.11) [41].Differential expression levels were estimated using the ‘‘edgeR”package (version 2.3.52) in R 4.0 [2].A non-parametric algorithm was developed to identify differentially expressed genes (DEGs)between samples,and the Benjamini-Yekutieli[2]hypergeometric test was used to calculate the false discovery rate(FDR).DEGs that had expression levels with an absolute log2Fold change ≥1 and an absolute–log10P ≥1.41 were considered significant.Principal component analysis(PCA)was performed with the RNA-seq count datasets generated from SMNT and NM plants at both the stem and heading stages.

2.2.4.Gene ontology (GO) and kyoto encyclopedia of genes and genomes (KEGG) enrichment analyses

GO analysis was used to annotate SMNT-induced DEGs by biological process(BP),cellular component(CC),or molecular function(MF),using the IWGSC genome as a reference [2].GO enrichment was performed using AgriGO 2.0 with a hypergeometric statistical test and the FDR method.A cutoff of FDR ＜0.05 was employed to remove unwanted GO terms from the data pool.Pathway enrichment analysis was performed with KEGG (http://www.genome.jp/kegg/) to identify enriched signal transduction and metabolic pathways by comparison with the whole genome background [2].

2.2.5.Ultra performance liquid Chromatography-Tandem mass spectrometry (UHPLC-MS) based non-target metabolomics

UHPLC-MS-based non-target metabolomics was used to identify patterns of metabolic alterations associated with tillering and tiller mortality.At the stem extension and heading stages,nonelongated tiller nodes were collected from 30 plants in the middle row of each plot,and metabolites were extracted with 1.5 mL of a 3:1 mixture of pre-cooled methanol and water [2].The resulting extracts were vortexed,ultrasonicated for 15 min in an ice bath,and incubated in a shaker at 4 °C.On the following day,samples were centrifuged at 12,000 g for 15 min at 4 °C,and the supernatants were transferred to 2 mL glass vials for chromatographic separation,and metabolite identification using a UHPLC-MS instrument (Ultimate 3000LC,Orbitrap Elite,Thermo Finnegan) coupled to a T3 column (100 2.1 mm,1.8 μm;Waters,London,UK).The metabolites were eluted with 0.1% aqueous formic acid and 100%acetonitrile (50:50,v:v) at the ow rate of 0.4 mL min1.The auto-sampler and column temperatures were 4 and 40 °C,respectively,and the injection volume was 2 μL.Mass data were collected using an AB Sciex QTOF mass spectrometer(Waters Acquity UPLC,London,UK).Quality control samples were injected at 10 intervals during the analytical run.

The UHPLC-MS data of each sample were preprocessed with SIEVE software(Thermo Finnegan,Waltham,MA,USA).The retention time,peak area,molecular weight,and peak density of each sample were calculated to form the normalized data by sum.Unsupervised PCA and supervised orthogonal partial least-squaresdiscriminant analysis (OPLS-DA) were performed using SIMCA-P 13.0 (Umetrics,Umea,Sweden).Metabolites with variable importance in the projection (VIP) values of ＞1.0 were considered associated with SMNT-enhanced tillering or tiller mortality.SMNT-induced differentially expressed metabolites (DEMs) at the tillering and tiller mortality stages were evaluated by log2Fold change,and each metabolite was log10-transformed for normalization across samples before pathway enrichment analysis.DEMs were identified using mass-to-charge ratios and the Metline database(https://metlin.scripps.edu).Metabolic pathways were generated by using MetaboAnalyst 5.0(https://metaboanalyst.ca)[42].A correlation-based network was generated and visualized using the‘‘FELLA”package in R(method=‘‘diffusion”,approx=‘‘normality”,and threshold=0.05) to identify DEMs,enzymes,reactions,and pathways associated with SMNT-increased tillering at the stem extension stage.The network was constructed by determining Pvalue thresholds (＜0.01),and then network density,average node degree,diameter,and clustering coefficient were tested across a range of P scores to determine correlation coefficients.The number of nodes was limited to ‘‘nlimit=150”.

2.3.Statistical analysis

Experimental values were subjected to analysis of variance(ANOVA) for the split-plot design using SAS version 9.1.3 (SAS Institute Inc,Cary,NC,USA).Means were compared using Tukey’s studentized range test,with the SMNT set as the main plot and the cultivars with contrasting tillering scores set as subplots.Graphs were plotted using Origin 9.0 software (OriginLab,Northampton,MA,USA).The contribution of yield components to total yield variation was quantified using dominant analysis in Stata 16.0 software (StataCorp,College Station,TX,USA).

3.Results

3.1.Grain yield and yield components

SMNT increased grain yield by 17.7% and 22.5% under dry(2017–2018) and favorable climates (2018–2019),respectively,relative to NM treatment (Table 1).The grain yields of the medium-and high-tillering cultivars were 16.0%and 19.7%greater(P ＜0.01) than that of the low-tillering cultivar across cropping seasons.The low-tillering cultivar maintained a high kernel number and HI,and the cultivars with high-tillering scores did not always produce high grain yield.SMNT and cultivar affected fertile spikes(FS)and kernel number per spike(KN),and minor variation observed in 1000-kernel weight(TKW).The application of SMNT to the medium-tillering cultivar resulted in an increase in grain yield by a mean of 20.5%,and FS,KN,and TKW explained total yield variations of 72.5%,19.1%,and 8.4%,respectively.

3.2.Tillering dynamics

The Nmaxwas observed at 56 DAS(800°C d)and was positively correlated with the tillering score (P ＜0.05) (Fig.2A).SMNT increased the Nmaxof the low-,medium-,and high-tillering cultivars.Tiller mortality was initiated at 800 °C d and terminated at 1800 °C d.At maturity,SMNT increased the fertile spikes of the low-,medium-,and high-tillering cultivars by 3.5%,17.4%,and 19.8%,respectively.

Wheat plants produced means of 1.3 tillers and 1.9 tillers at the stem extension stage under dry and favorable climates (Table 2),with third tillers observed only under SMNT and favorable conditions (Fig.2B).SMNT increased the tillering capacities of the lowand high-tillering cultivars by 19.3% and 32.3%,and first tillers were produced in favorable climates at markedly higher rates than in dry climates.Straw mulching increased the first TER of the lowand high-tillering cultivars by 35.2%and 40.0%,respectively.However,in dry and favorable climates,the TER of second tillers was highly variable and relatively low.The occurrence rate of first and second primary tillers observed at the stem extension stage was positively correlated with the leaf number of tillers(Table S1).The percentages of first primary tiller occurrence for low-and high-tillering cultivars grown under NM were respectively 14.3% and 17.4% greater than that of tillers under SMNT.The number of first primary tillers was significantly higher than that of second primary tiller,and SMNT reduced the difference between first and second primary tillers by 16.3% and 80.4% in dry and favorable climates,respectively.All second and third tillers eventually died,with only 14%–61%of tillers developing into fertile spikes (Fig.2B).The SMNT reduced TSR of the medium-and hightillering cultivars by 7.6%and 12.3%,respectively(Table 2;Fig.2C).SMNT increased the TER of first tillers,which compensated for the slightly increased tiller mortality,thereby increasing the fertile spike number under both dry and favorable climates.

3.3.N uptake,remobilization,and asymmetric competition among tillers

SMNT increased total N uptake by 72.7%,24.8%,and 26.2%at the stem extension,anthesis,and maturation stages,respectively,compared with the NM treatment(Fig.3A).Averaged across cropping seasons,the total N uptake of the high-tillering cultivars was 16.0% greater (P ＜0.01) than that of the low-tillering cultivar.SMNT increased the aboveground dry mass yield of low-and high-tillering cultivars by 13.1% and 18.2%,respectively(Table S2).The proportions of fertile tillers relative to the total dry mass yield under SMNT were 32.3%and 20.5%higher than that under NM.However,irreversible dry matter loss (infertile tillers)accounted for a mean of only 1.7% of total dry mass yield.Fertile spikes increased with increasing ag leaf N% at both the stem extension and anthesis stages.If ag leaf N%was ＞3.4%at the stem extension stage,there was a 90.5% probability of achieving spike densities and grain yields of 450 m2and 8 t ha1,respectively(Fig.3B).

Table 1 Effects of straw mulch-based no-tillage and cultivar with contrasting tillering capability on yield components of wheat and HI in both dry and favorable climates.

SMNT increased N remobilization from each tiller to its grain in low-and high-tillering cultivars by respectively 29.9%and 35.1%in dry climates and 67.5% and 27.3% in favorable climates (Fig.4A).The N% of tillers grown under dry climate was 2.35%,a value 37.4%lower than that of tillers(3.23%of N)grown under favorable climate,whereas the WSC% and C:N were 128% and 178% greater(Table 3),respectively.Irrespective of growing season or cultivar,SMNT increased the N%,WSC%,and C:N of tillers by 7.2%,17.1%,and 9.3%,respectively.The WSC% and N% of the main stems were 12.2%and 4.0%greater than those of tillers,and the C:N of the main stem under NM and SMNT was 5.6% and 10.9% greater than the main stem that of tillers.

Table 2 Tillering emergence and survival processes of wheat cultivars grown under straw mulch-based no-tillage and no-mulch treatments in both dry and favorable climates.

Table 3 Effects of straw mulch-based no-tillage on the water-soluble carbohydrates (WSC%) and nitrogen contents (N%) of main stems and tillers of wheat cultivars in stem extension stage.

The δ14C(mg plant1)yield gap between main stems and tillers decreased with wheat growth,whereas the δ15N(mg plant1)yield gap showed the opposite trend (Fig.4B).The δ14C and δ15N yield gaps were greater in the low-tillering than in the high-tillering cultivars.This finding indicates that the SMNT optimized WSC and N ratio of the tillers and that the asymmetric competition between the main stem and tillers was reduced.The proportion of total plant N in grain (NHI) was affected by environment,SMNT,and cultivar (P ＜0.05,Fig.4C).The NHI in the dry climate was 15.8%higher than that in the favorable climate,and the NHI decreased with increasing tiller position.The SMNT increased the NHI of the main stem and first tillers by 12.1% and 4.8%,respectively.Thus,SMNT increased tiller emergence rate and fertile tiller number by increasing total N uptake and N remobilization from each tiller to the grain.Consequently,asymmetric competition between main stem and tillers decreased,increasing the probability of tillers developing into spikes after stem extension.

3.4.Tillering-associated DEGs

A total of 16,133 DEGs (q ＜0.05,log2FC ＞1) were associated with increased tillering under SMNT,with 5425 genes upregulated and 10,497 down-regulated (Fig.S1).Compared with NM,the CAMTA transcription factor (TF) family showed the highest counts under SMNT,followed by the GRAS and C2H2TF families(Fig.S1D).The three most enriched TFs and associated genes are involved in the receptor-like protein kinase and calciumdependent protein kinase.Among the biological process DEGs,the most enriched GO terms were cell division and water deprivation responses (Fig.5A).Among the cellular component DEGs,the most enriched GO terms were nucleosomes,DNA packaging complex,and chromatin (Fig.S2).Finally,the most enriched GO terms among the molecular function DEGs included protein heterodimerization activity,catalytic activity,and protein dimerization activity.KEGG analysis assigned a total of 16,133 DEGs to 123 pathways(Fig.5B),and SMNT induced the most significant enrichment in environmental information processing (mitogen-activated protein kinase [MAPK] signaling,plant hormone signal transduction),genetic information processing (DNA replication and mismatch repair),and metabolism (phenylpropanoid biosynthesis,and porphyrin,chlorophyll,and phenylalanine metabolism).

Among plant hormone signaling transduction pathways,the 112 DEGs that encoded GH2,PYR/PYL,ERF1/2,TCH4,and JAZ,which regulate senescence and drought responses,were rapidly down-regulated in response to SMNT,whereas auxin response factor (ARF),A-ARR,and CYCD3,which regulate cell division and cell enlargement were up-regulated (Fig.S3).Among MAPK signaling pathways,the 94 DEGs encoded p12,p74,p36,p46,and p4 MAPK,PYL,and PYR were rapidly down-regulated in response to SMNT,whereas EPF 1/2,ERL,and SPCH,which regulate stomatal development,were up-regulated(Fig.S4).Among genetic information processing pathways,DEGs encoding DNA polymerases α,δ,and ε,the MCM complex,DNA ligase,and helicase were all up-regulated under SMNT.There were no significant differences in DNA polymerase B or D expression levels,primase,or the DNA polymerase III holoenzyme.Among DEGs associated with eukaryotic DNA mismatch repair,the components of MSH2 and MSH6 were significantly up-regulated under SMNT.

3.5.Tillering-associated DEMs

The OPLS-DA model was relatively well supported (Fig.6A,R2=0.9).A total of 661 credible metabolites (VIP ＞ 1,P ＜ 0.05) were identified for both positive-and negative-ion models.A total of 13 and 50 metabolites were down-and upregulated (fold change ＞1.4,P ＜0.05),respectively,by SMNT(Fig.6B).The metabolites whose levels increased the most were melezitose (2.3-fold),Arg (1.9-fold),and raffinose (1.8-fold).The metabolites whose levels decreased the most were 6-o-B-Dgalactosyl-D-glucose (4.0-fold),lupeoic acid (2.5-fold),6-hydroxyluteolin 6,7,3′-trimethyl ether 4′-glucoside (2.4-fold),and phenylbutyrylglutamine (2.4-fold,Fig.6D).Among the upregulated DEMs,eight metabolites (among them tyrosine,phenyl ethyl alcohol,trans-2-hydroxycinnamate,and trans-3-hydroxycinnamate) were associated with phenylalanine metabolism,and four (tyrosine,4-hydroxystyrene,trans-2-hydroxycinnamate,and sinapyl alcohol) were associated with phenylpropanoid biosynthesis (Fig.6C).Overall,the SMNTinduced tillering-associated DEMs were associated with sphingolipid,phenylpropanoid,and galactose.

Fig.2.Effects of straw mulch-based no-tillage(SMNT),cultivar identity,and climate on wheat tillering.(A)Tiller-promoting effect of SMNT in dry and favorable climates.τMS,τT1,τT2,τBS,and τES,are the thermal times for the emergence of the main stem,first tiller,second tiller,the beginning,and the end of the senescence phase,respectively.The NP,NT1,NT2,Nmax,and Ne are the green tillers number per m2 at plant emergence,plants with 1 tiller,plants with 2 tillers,maximum active tillers,and number of ears per m2,respectively;(B) Effects of SMNT and cultivars on green tiller number at each position observed at the stem extension stage when the plants achieve the maximum tiller population.LT and HT represent the cultivars with low and high tillering capacity,respectively.(C) Relationship between maximum tiller number at stem extension and fertile spike number at maturity.The gray triangular area represents the tiller mortality domain.The dotted line represents the 1:1 ratio.All values were measured with three replications and are expressed as mean ± standard error (n=3).

3.6.Tillering-associated metabolites,enzymes,reactions,and pathways in tiller nodes

Integrating pathways,reactions,enzymes,and metabolites provided the five-key pathways control wheat tillering in SMNT(Fig.S5).The correlation-based network included 27 compounds,86 reactions,24 enzymes,7 modules,and 5 pathways.Galactose metabolism was most closely associated with the expression of inositol 3-alpha-galactosyltransferase,galactinol-sucrose galactosyltransferase,and β-frutofuranosidase.Similarly,amidase and arogenate dehydrogenase levels were most closely associated with phenylalanine,tyrosine,and tryptophan biosynthesis,and changes in phenylalanine ammonia lyase and 4-monooxygenase levels were most closely associated with changes in phenylalanine metabolism.Trans-cinnamate and 4-coumarate CoA ligase levels were most closely associated with phenylpropanoid biosynthesis.The network analysis indicated that galactose metabolism and phenylpropanoid biosynthesis were most strongly associated with increased tillering under SMNT.

Fig.3.Effects of straw mulch-based no-tillage(SMNT)and cultivars on the root N uptake of each tiller(A),and the relationship between fertile tiller density and leaf N%of dryland wheat cultivars(B).HT,high-tillering cultivar;LT,low-tillering cultivar;NM,no-mulch control;SMNT,straw mulch-based no-tillage;G,genotype;E,environment;M,management.Values are expressed as mean ± standard error (n=3).**, P ＜0.01;*, P ＜0.05;ns,not significant (P ≥0.05).

4.Discussion

4.1.SMNT promotes grain yield by coordinating tiller-and kernelnumber plasticity

The finding that the grain yield of the medium-tillering cultivar exceeded that of the high-tillering cultivar in agricultural areas that receive ＜200 mm rainfall,which agreed with previous results[8,11,43].The increased kernel and fertile tiller numbers under SMNT can be attributed to increased total N uptake,leaf N%,and the amount of remobilized N from each tiller to its grain (Fig.7).This is because cultivars with fewer tillers can mobilize more nutrients to support root development,stem growth,and oret development,thereby increasing biomass and HI for maximum oret primordia number,and consequently,increased kernel numbers and yield [22,44,45].Typically for the high-yielding population at this ecological site,the fertile spike number was 350 spikes m2,and the present study suggests that cultivars with a main stem and 1.9 tillers can exploit both tiller and kernel-number plasticity to maintain high fertile tillers and kernel numbers in both dry and favorable environments.The moistureconserving effects of straw mulching alleviate drought stress in the tillering stage,increase total N uptake and fruiting efficiency,and promote a balanced tillering and fruiting efficiency [5].Thus,a cultivar with reduced tillering traits can respond more actively to increased soil water availability under SMNT and achieve high grain yield in both dry and favorable climates.

4.2.SMNT maintained the advantage of reduced-tillering traits by optimizing tillering curves

Fig.4.Effect of straw mulch-based no-tillage(SMNT)on N remobilization(A),asymmetric C and N competition between main stem and tillers(B),and harvest index(C)of low-and high-tillering wheat cultivars.Asymmetric N competition between main stems and tillers was estimated using the δ15N yield gap between main stems and tillers(see material and methods).HT,high-tillering cultivar;LT,low-tillering cultivar;NM,no-mulch control;SMNT,straw mulch-based no-tillage.Values are expressed as mean ± standard error (n=3).Different letters indicate significant differences at the 0.05 probability level.*, P ＜0.05;**, P ＜0.01;ns,not significant (P ≥0.05).

The combined effects of SMNT and cultivars on tillering of dryland wheat grown under both dry and favorable climates demonstrate the importance of optimizing tillering curve rather than the maximum tiller population alone to achieve the maximum advantages of SMNT.SMNT did not affect the phyllochron(the final leaf number),as previously reported for both wheat and barley[37],but did increase TER and effective tillering duration,thereby increasing TC.Increased TER of first tillers was mediated by tiller node cell division and elongation,owing to up-regulation of a series of DEGs associated with DNA replication and mismatch repair under SMNT.Cell division is closely associated with tillering capability,given that wheat tillering is dependent on the cell division of exuberant meristems [2].SMNT increased N uptake,WSC,and N levels of tillers,thereby increasing cell division,tiller emergence rate,and tillering capability.This is because,as previously reported[46,47],N status did not affect tiller bud initiation but did affect cell division-mediated stem elongation.Similarly,overexpression of a low-affinity nitrate transporter gene (OsNPF7.2) increased nitrate absorption of roots,thereby increasing tiller number [48],likely through increased cell division in tiller buds [49].Collectively,these results indicate that the increased emergence rate of first tillers and fertile spike under SMNT is mediated by tiller node cell division,which is increased by increased total N uptake and N%of tillers (Fig.7).

Fig.5.The most enriched GO (A) and KEGG (B) pathway analysis of validated differentially expressed genes (DEGs) in straw mulch-based no-tillage (SMNT) vs.no-mulch(NM).(A)Bubble colors represent the enrichment Q values of the corresponding pathways.(B)The vertical axis represents the corresponding DEG count illustrating the DEG number participating in environmental information processing,genetic information procession,and metabolism.

The effects of SMNT on TER and tiller occurrence (%) varied according to tiller position.Much less variation was observed for first than for second tiller production,possibly because of reduced asymmetric competition between the main stem and tillers under SMNT.The emergence rate of second tillers was lower and more variable than that of first tillers,and second tillers more often failed to develop into fertile spikes after stem extension.These findings can be attributed to the weak sink strength of second tillers with fewer than four leaves.Such tillers could have supplied nutrition to first tillers after stem extension [16,17],thereby increasing first tiller survival.However,in a study[18]of the relationship between early-and late-emerging tillers,the presence of nonproductive tillers failed to improve the performance of primary tillers of wheat plants under drought stress.These debatable findings might be attributed to the mobilization of nutrients among tillers.

Although the TC was the main contributor to the 20.5% yield increase observed under SMNT,reduced tillering also played a role because of its positive effects on tiller survival after stem extension.Cultivars with fewer tillers at the stem extension stage exhibited lower tiller mortality and consequently more fertile spikes in dry and favorable climates.Both the emergence rate of second tillers and TMR were unaffected by cultivar but varied under SMNT and by cropping season.This unexpected result suggests an excellent opportunity to increase grain yield through SMNT.Because the appearance of second and subsequent tillers generally coincides with the frequency of cold-stress events,and tiller mortality is associated with the mobilization of nutrients from main stems and productive organs,it was not surprising that environment and SMNT played dominant roles in determining the formation rate of second tillers.The positive association between tiller survival and mortality,which determined compensation between these two variables,is relatively common in gramineous crops[50,51].The positive effect of SMNT on fertile spikes is the basis for increasing fertile spikes in agricultural areas that receive ＜200 mm rainfall.Although another study [11] has suggested that reduced-tillering cultivars may underperform in high-yield situations,the effects of SMNT can be harnessed to maintain the advantage of reduced-tillering traits.

4.3.SMNT improved TER by reducing asymmetric competition between main stem and tillers

Most of the yield-promoting effects of SMNT can be attributed to the reduction of asymmetric competition between main stems and tillers for WSC and N% nutrition.This conclusion is supported by tillering dynamics and isotope (δ14C and δ15N)-based tracing.Grain yield was reduced by asymmetric competition between main stems and tillers for C and N nutrition[15,21,22].The increased N%of tillers under SMNT resulted from increased total N uptake,which alleviated asymmetric competition between main stems and tillers for N nutrition,thereby promoting tiller emergence and tiller capability [23,24].The increase in the δ15N yield gap between main stems and tillers with plant maturity was expected because main stem dominance increases with plant growth.However,the δ14C yield gap between main stems and tillers unexpectedly decreased with plant maturity,and the high-tillering cultivar exhibited lower δ14C and δ15N yield gaps than did the low-tillering cultivar.These results indicate that tiller plasticity was strongly affected by optimized WSC and N levels.The use of reducedtillering cultivars or de-tillering treatments can increase the WSC contents of main stems and the availability of nutrients for spike development,thereby increasing kernel number and yield[22,52].The negative relationship between organ number and carbohydrate accumulation is common in cereal crops [53] and explains the incongruence of increases in grain yield and tillering score.Along with the results of a similar study in rice[54],the present study’s findings suggest that wheat plants with more tillers and late-emerging tillers achieve a lower WSC content at anthesis(Table 3).SMNT increased the N contents of both main stems and tillers,which resulted from increased carbohydrate partitioning to roots.Increased carbohydrate level improved root branching and root surface area [11],thereby promoting N and water absorption[5].Allometric increases in WSC and N% under SMNT resulted in increased C:N,which could provide enough energy and nutrients for increasing the number of ear-bearing spikes,thereby increasing kernel number per m2as observed under SMNT (Fig.7).

Fig.6.Comparative analyses of tillering-associated differently expressed metabolites (DEMs) in straw-mulch based no-tillage (SMNT) vs.no-mulch (NM) treatments.(A)Supervised orthogonal partial least-squares-discriminant analysis (OPLS-DA) model of class membership resulting in the largest predicted indicator variable.(B) Effect of straw mulching on significant DEMs in tiller node cells.(C)Pathway enrichment analysis of tillering(left)and tiller survival(right)-associated pathways(SMNT vs.NM).(D)Up-or down-regulated metabolites associated with tillering of wheat in comparison of SMNT with NM.Blue and red color represent differently expressed metabolites (–log10P ≥1.41,log2FC ＞1,VIP score ＞1).(For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)

4.4.SMNT promotes tillering by altering signaling transduction,rapid drought response,and tiller maintenance after stem extension

The present study results suggest that MAPK and phytohormone signal transduction,as preferred signal sensing pathways,respond to contrasting environments under SMNT,thereby regulating both tillering and tiller mortality.Similarly,OsMPK3 phosphorylates OsWRKY30,which is known to increase drought tolerance [55,56],and PtMAPK6 has been reported [57] to phosphorylate PtMYB4,which regulates tiller outgrowth in the early stages of xylem development in tillers of rice.The MAPK signal can also activate transcription factors(GRAS,WRKY,NAC),thereby eliciting rapid responses to drought stressors [58,59].Three major differentially expressed TFs(CAMTA,GRAS,and C2H2)were identified in the present study,a finding similar to that of a previous study [2,60].CAMTA represses salicylic acid (SA) biosynthesis and N-hydroxypipecolic acid levels by modulating SARD1 and CBP60 expression,thereby eliciting a rapid drought response[61–64].In maize plants,the reduced-tillering gene tin1,which harbors the C2H2-zinc-finger domain,functions as a C2H2TF and regulates tiller bud outgrowth and drought tolerance [65].In rice plants,the MOC1/LAS/Ls gene encodes GRAS family proteins [66],and overexpression of MOC1 was reported to increase TC because of their critical role in gibberellic acid transduction,root development,and axillary shoot development [1,67,68].GRAS proteins can also maintain meristem formation competence and subsequent initiation of axillary meristems [69].Thus,CAMTA,GRAS,and C2H2function in SMNT-increased tillering by increasing rapid drought response and tiller bud maintenance (Fig.7).

Fig.7.A cause-and-effect diagram illustrating the regulation of tillering capability and grain yield of dryland winter wheat by straw mulch-based no-tillage.KN,kernel number per spike;PYR/PYL,pyrabactin resistance/pyrabactin resistance-like;PP2C,protein phosphatases 2C;SnRK2,sucrose nonfermenting 1-related subfamily2;IAA,auxin;CTKs,cytokinin;CAMTA,calmodulin-binding transcription activator;GRAS,GAI,RGA,SCR family;C2H2,cysteine-2/histidine-2.

In the present study,both ARF and A-ARR were up-regulated by SMNT,further promoting cell enlargement and tiller growth and confirming previous findings [38] that have suggested that auxin tightly controls tiller formation and drought stress responses.Indeed,both ABA and jasmonic acid (JA) have been reported[70,71] to play essential roles in stress-exposed plant responses.In one study[72],drought elevated the ABA concentration of inferior tillers and inhibited their growth.ABA signals have also been reported to affect tillering by modulating strigolactone-mediated axillary bud dormancy in barley[30] and rice[73].The expression profiles of ABA,SA,and JA pathway-associated DEGs in the present study suggest that SMNT down-regulates genes associated with drought stress responses and further regulates the dormancy of tiller buds and tiller emergence rate,a finding similar to a previous one[31].Thus,based on these findings,it is likely that SMNT alleviates the effect of drought on wheat tillering by promoting ABAdependent and ABA-independent signaling transduction (Fig.7).Further work is needed to determine whether the regulation is direct or mediated by a signal transmitted from roots to tiller buds.

Phenylpropanoid biosynthesis serves as a starting point for producing the lignin required for wheat tillering [74].In wheat,tillering-associated long noncoding RNAs have been reported[75] to regulate tiller development by in uencing phenylpropanoid biosynthesis.In rice,strigolactones have been reported [76] to inhibit bud growth by regulating the expression of genes involved in phenylpropanoid and avonoid biosynthesis.In foxtail millet,genes and metabolites of the phenylpropanoid pathway function in regulating resistance to drought stress[77].In the present study,the DEGs encoding phenylalanine ammonia lyase,trans-cinnamate 4-monooxygenase,shikimate O-hydroxy-cinnamoyl-transferase,F5H,and coniferyl-aldehyde dehydrogenase were down-regulated under SMNT,whereas 4-coumarate-CoA ligase,cinnamoyl-CoA reductase,cinnamylalcohol dehydrogenase,and peroxidase were up-regulated.Phenylalanine ammonia lyase is the first enzyme to catalyze Lphenylalanine conversion to ammonia and trans-cinnamic acid,which functions in defense against drought stressors [75,78].Cinnamic,p-coumaric,caffeic,ferulic,and sinapic acids are reportedly [79] regulated by both 4-coumarate-CoA ligase and cinnamoyl-CoA reductase.Overexpression or suppression of phenylpropanoid metabolism alters plant anatomy and lignification and affects plant fitness.Thus,SMNT-up-regulated genes involved in phenylpropanoid biosynthesis increased lignin biosynthesis required for tiller growth.Sphingolipid metabolism was the most strongly affected by tillering under SMNT.The functions of sphingolipids in plants were long overlooked,until genetic and post-genomic models were used to demonstrate their essential roles in cell-to-cell interactions and cell wall formation [80,81].In the present study,DEGs involved in sphingolipid biosynthesis may have contributed essential defense mechanisms,given that sphingolipid inositol has been reported to protect young leaves from drought stress [82] and regulate plant growth [83].The DEGs encoding aspartate transaminase,tyrosine transaminase,polyphenol oxidase,and aromatic-Lamino-acid decarboxylase were down-regulated under SMNT,whereas those encoding primary-amine oxidases were upregulated.Galactose metabolism is responsible for converting D-fructose-6-P to ascorbate following a dominant pathway of ascorbate biosynthesis in Arabidopsis thaliana [84] and rice[85].The potential higher activities of the corresponding enzymes may reduce osmotic pressure,thereby increasing drought tolerance and tillering capability.Thus,phenylpropanoid,galactose,and sphingolipids act in SMNT-increased tillering by increasing drought tolerance and lignin biosynthesis required for tiller growth (Fig.7).

5.Conclusions

Application of SMNT to a medium-tillering cultivar tended to coordinate both tiller and kernel-number plasticity,thereby increasing mean grain yield by 20.5%.An increase in fertile spikes under SMNT was attributed mainly to the augmentation of the first TER by increasing root N uptake,tiller N%,and the amount of remobilized N from tillers to their grain.The second and third tillers functioned as transient sinks and contributed to the improved TER,depending on the leaf number of tillers.Because of allometric increases in N uptake and assimilation production of tillers,asymmetric competition between main stem and tillers was reduced by SMNT.The multi-omics integration approach confirmed that SMNT alleviated drought stress and regulated wheat tillering by both ABA-dependent and ABA-independent signals.These signals activate three critical transcription factors (CAMTA,GRAS,and C2H2)that enhance rapid drought response and tiller bud maintenance.Phenylpropanoid biosynthesis,sphingolipid biosynthesis,and galactose metabolism positively regulated tillering under SMNT because of their critical role in drought response and lignin biosynthesis.Further investigation is needed to elucidate the crosstalk between signals,TFs,and metabolites in tiller node cells under SMNT.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to in uence the work reported in this paper.

CRediT authorship contribution statement

Hongkun Yang:Conceptualization,Data curation,Methodology,Software,Supervision,Writing– review &editing.Yun Xiao:Investigation,Methodology.Peng He:Investigation,Validation.Dailong Ai:Investigation,Validation.Qiaosheng Zou:Investigation,Validation.Jian Hu:Validation.Qiong Liu:Validation.Xiulan Huang:Validation.Ting Zheng:Validation.Gaoqiong Fan:Funding acquisition,Project administration.

Acknowledgments

We are grateful for financial support from the Sichuan Province Science and Technology Support Program (2021YJ0504,2021YFYZ0002),National Key Research and Development Program of China (2016YFD0300406),Special Fund for Agro-scientific Research in the Public Interest (20150312705),and the Crops Breeding Project in Sichuan Province (2016NYZ0051,22ZDZX0018).We also thank the editor and two reviewers for their help in revising the manuscript.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.09.011.