LATERAL BRANCHING OXIDOREDUCTASE specificity for strigolactone branching inhibition in barley

Abstract

Strigolactone (SL) mutants display a range of phenotypes, such as increased branching, reduced stature, and a loss of SLs exuded from roots into the soil. SL biosynthesis is complex and divergent between plant species. Recently, mutants defective in specific SL biosynthesis genes have shown a loss of exuded SLs, but no obvious change in branching (tillering). This means that functional specification may exist between certain SL subtypes. It has been suggested that the LATERAL BRANCHING OXIDOREDUCTASE (LBO) enzyme acts in a subpathway of SLs that is specific for branching. Here we report that barley plants mutant in hvlbo have increased tillering, but normal production of SLs detected in roots and root exudates. This finding supports the idea that SLs have functional or tissue-specific differences and that the LBO pathway has specificity for bud outgrowth rather than exudates.

Barley lateral branching oxidoreductase mutants reveal functional divergence between strigolactones involved in tillering and those involved in root exudates.

Barley lateral branching oxidoreductase mutants reveal functional divergence between strigolactones involved in tillering and those involved in root exudates.

Introduction

Strigolactone (SL) biosynthesis consists of a core conserved pathway from β-carotene to carlactone (CL), which then diverges through a range of enzymatic pathways to produce various SL types depending on each plant species (Marzec et al., 2020b; Yoneyama and Brewer, 2021; Dun et al., 2023; Nomura et al., 2024). SLs are plant hormones perceived by the DWARF14 (D14) receptor, which is encoded by a single gene copy in most species (Waters et al., 2012). Mutants that fail to produce the core biosynthesis pathway, such as rice dwarf17 (d17) or maize zmccd8, display the full range of SL-related phenotypes, such as excessive branching/tillering, dwarfism, and the absence of SLs exuded from roots (Brewer et al., 2013; Ito et al., 2022; Chen et al., 2023; Li et al., 2023). Failure to exude SLs can result in reduced fungal symbiosis or parasitic weed infestation (Gomez-Roldan et al., 2008; Umehara et al., 2008). SLs, including those that are exuded, can inhibit branching when artificially added onto buds (Arite et al., 2009). Hence, it has been assumed that any SL exuded from roots also contributes to branching inhibition as an internal hormone. However, recently, SL mutants have been reported to be defective in exudates (and some plant phenotypes), but normal in terms of branching/tillering (Wakabayashi et al., 2019; Ito et al., 2022; Chen et al., 2023; Li et al., 2023, 2024; Wang et al., 2024; Zhou et al., 2025).

The phenotypic separation of exudate content and shoot branching suggests that there are SL biosynthesis pathways for distinctive compounds with different physiological functions. For instance, Os900 is a CYP711A2 enzyme in rice involved in the production of deoxyorobanchol (4DO), which is further converted to orobanchol (oro) by Os1400 (CYP711A3). However, although rice os1400 and/or os900 mutants showed loss of SL exudate, they did not show the increased tillering phenotype of the core biosynthesis mutant, d17 (Ito et al., 2022; Chen et al., 2023). Moreover, os1400 single mutants were specifically shorter due to a build-up of 4DO (Chen et al., 2023). In maize, ZmCYP706C37 and ZmMAX1b (a CYP711A enzyme) are required for the production of prominent exuded SLs, including zealactone. However, plants defective in those enzymes did not show the branching increase that was observed in the core biosynthesis mutant, zmccd8 (Li et al., 2023). Also, OsCYP706C2 and two methyl transferases are involved in converting CL to 4-oxo-methyl carlactonoate (4-oxo-MeCLA). However, although oscyp706c2 mutants were defective in 4-oxo-MeCLA exudate, root mycorrhizal colonization, and root architecture, they did not show increased tillering (Li et al., 2024). CYP722C is involved in the production of oro. In tomato, loss of slcyp722c caused deficiencies of oro and solanacol in exudates, but did not affect branching or height (Wakabayashi et al., 2019). Moreover, recent findings showed that CYP722A converts carlactonoic acid (CLA) into 16-hydroxy-CLA. However, Arabidopsis atcyp722a mutants showed no increase in branching (Zhou et al., 2025).

The production and identity of hormone-specific SLs remain unclear, together with how they may function specifically in branching. The most likely pathway involves the enzymes CLA METHYLTRANSFERASE (CLAMT) and LATERAL BRANCHING OXIDOREDUCTASE (LBO) (Brewer et al., 2016; Yoneyama et al., 2020a; Wakabayashi et al., 2021; Mashiguchi et al., 2022; Haider et al., 2023; Li et al., 2023; Zhou et al., 2025). CLAMT is a methyltransferase that catalyses the conversion of CLA to MeCLA. Arabidopsis atclamt mutants still produced CLA, but were deficient in MeCLA, and yet showed increased branching (Wakabayashi et al., 2021; Mashiguchi et al., 2022).

MeCLA can be further converted to 1′-OH-MeCLA by LBO, which is a 2-oxoglutarate and Fe(II)-dependent dioxygenase (2-OGD) (Brewer et al., 2016; Yoneyama et al., 2020a; Zhou et al., 2025). However, the role of 1′-OH-MeCLA is uncertain (Brewer et al., 2016). Arabidopsis atlbo mutants still produced upstream compounds CL and MeCLA, and yet branching was not fully repressed (Brewer et al., 2016). This means that LBO produces from MeCLA one or more SLs that inhibit branching, which contrasts with the pathway steps described above that relate to SL exudates, but not branching. Moreover, clamt and lbo double mutant plants were additive in branching, attaining the same level as the core biosynthesis mutant, atccd7 (Zhou et al., 2025). This confirms that the CLAMT–LBO pathway is a key regulator of branching, and MeCLA bioactivity is enhanced by LBO for full branching inhibition.

Hydroxylated versions of CLA can also be processed into hydroxylated derivatives by CLAMT and LBO (Yoneyama et al., 2020a; Zhou et al., 2025). Accordingly, a group of related compounds could be produced internally in plants as branching inhibitors (Zhou et al., 2025). Moreover, it was recently reported that branching could be repressed in Arabidopsis clamt mutants by engineering plants to produce 5-deoxystrigol (5DS) using GaCYP722C from cotton (Mizutori et al., 2025). However, mutants engineered to produce 18-oxo-CLA (using SlCYP722C from tomato) or 4DO (using Os900 from rice) did not show branching repression (Mizutori et al., 2025). This suggests that 5DS can be an endogenous branching inhibitor, in addition to the CLAMT–LBO pathway.

Further research into the CLAMT–LBO pathway may help to identify branching-specific SLs and discover how they relate to exuded SLs and other plant traits. Arabidopsis lacks the complexity in SL production and exudates required to fully understand LBO function (Mizutori et al., 2025). Although barley is a critical part of international grain production, relatively little is known about barley SL biosynthesis. We recently detected 6-epi-heliolactone, barleylactone 1, and barleylactone 2 in root exudates (Inoue et al., 2024), which allowed us to test whether CLAMT–LBO is involved in the production of these SLs. LBO is encoded by only a single gene in most species with one known catalytic action on MeCLA (Brewer et al., 2016; Yoneyama et al., 2020a; Zhou et al., 2025). By contrast, CLAMT enzymes are more diverse in phylogeny and can contribute to the production of both tillering and exudate SLs (Wakabayashi et al., 2021; Li et al., 2023; Kuijer et al., 2024). Therefore, we decided to firstly target HvLBO and generated a range of barley hvlbo mutants by TILLING (Targeting Induced Local Lesions in Genomes) and gene editing. Intriguingly, the resultant mutants displayed hallmark phenotypes that suggest a lack of SL production specific for branching. However, the mutants still produced exuded root SLs and showed no obvious defect in fungal symbiosis. These data support a model whereby (i) the LBO pathway produces branching SLs; (ii) exuded SLs are produced via a split in the SL pathway upstream of LBO; and (iii) exuded SLs do not normally inhibit branching.

We attempted to discover a branching SL in barley. However, we could not yet identify an SL missing in the shoot tissue of hvlbo mutants. We previously showed in Arabidopsis that the product of LBO was poorly mobile over a graft (Brewer et al., 2016). Hence, the branching influence of LBO is likely to break down quickly in tissue extracts. Further research is required to discover this branching SL and to determine why the root-exuded SLs in hvlbo are not involved in internal branching inhibition.

Materials and methods

Plant materials

For work at the University of Milan, seeds of the Arabidopsis thaliana lbo-1 mutant line [Ws-4 ecotype; T-DNA insertional mutant; code 119G09 (DYK9)] and the corresponding wild-type (WT) background (Ws-4, code 530AV) were acquired from the French National Institute for Agricultural Research (INRA). Barley lbo.f homozygous mutants in the Sebastian background and Sebastian WT seeds originated from the University of Silesia in Katowice stocks (see below). For the production of gene-edited hvlbo alleles at The University of Adelaide, Golden Promise was acquired from University seed stocks.

Screening for mutants in the hvlbo gene using the HorTILLUS TILLING population

To isolate mutants carrying point mutations in the HvLBO gene, a TILLING strategy was used. The procedure was carried out on the population developed at the University of Silesia in Katowice, Poland called HorTILLUS (Hordeum vulgare  TILLING University of Silesia), which was created by double treatment of spring barley variety Sebastian with two chemical mutagens: sodium azide (NaN₃) and N-methyl-N-nitrosourea (MNU) (Szurman-Zubrzycka et al., 2018). DNA extracted from nearly 7000 M₂ individuals from the HorTILLUS population was combined into 8-fold DNA pools and screened for the presence of mutations in a gene fragment encompassing the LBO catalytic domain. The HorTILLUS primers (Supplementary Table S1) were anchored in the second exon of the HvLBO gene. The protocol for TILLING analysis followed previous recommendations (Szurman-Zubrzycka et al., 2017). Briefly, the target amplicons were generated using 8-fold pools of DNAs from M₂ individuals, with a 20 µl PCR mixture contacting 5 pmol of each primer, 0.2 mM of each of dNTP, 1 U of Color Taq DNA polymerase (EurX, Gdansk, Poland), 150 ng of template DNA, and 1×PCR buffer containing 1.5 mM MgCl₂. The PCR profile was as follows: 3 min at 95 °C, eight cycles of touchdown with denaturation at 95 °C for 30 s, annealing at 64 °C for 50 s (decreasing 0.5 °C per cycle), and amplification at 72 °C for 1 min 30 s, followed by 30 cycles with denaturation at 95 °C for 30 s, annealing at 60 °C for 50 s, and amplification at 72 °C for 1 min 30 s. The reaction was completed by a final amplification step of 5 min at 72 °C. Next, the heteroduplex formation step was performed at 95 °C for 3 min for initial denaturation, followed by slow renaturation in a touchdown type of reaction (70 cycles of 20 s starting from 70 °C with 0.1 °C decrease per cycle). Then, the samples were treated for 15 min at 45 °C with 20 μl of 0.1×Celery Juice Extract (CJE) containing the CelI enzyme that specifically cuts DNA mismatches. The products of cleavage were analysed using PAGE on a LI-COR 4300 DNA Analyzer. Gel images were manually screened to detect the signals indicating putative mutations in the 8-fold DNA pools. The identification of single plants carrying the mutations was done by mixing equal amounts of DNA from each of the M₂ plants that comprised the specific pool with DNA from the parent variety Sebastian. Such mixtures were then analysed using the same protocol as described for the 8-fold DNA pools. All putative mutations were confirmed by Sanger sequencing.

Isolation of hvlbo gene-edited mutant alleles

Construction of pY-LCRISPR-HvLBO was conducted following a prior protocol (Ma et al., 2015), including the thermocycler setups and Golden Gate assembly (GGA) reactions, unless otherwise stated. The construct targeted two positions, one in exon 1, HvLBO-Target-1: GAAGCTCAGGAGAGCCTGCGAGG (PAM sequences in bold and italic), and one in exon2, HvLBO-Target-2: CGGGTAGAAGTTCATCCGCACGG, and was driven by the OsU6a promoter and OsU6b promoter, respectively. Phusion® High-Fidelity DNA Polymerase (NEB) was used for all PCRs during plasmid construction. Overlapping PCR was used to introduce the target sequence into the single guide RNA (sgRNA) cassette. All PCR fragments were confirmed by Sanger sequencing prior to use. Following sequence verification to confirm successful introduction of target sequences 1 and 2, both verified fragments were combined with the binary vector pYLCRISPR/Cas9Pubi-H in a GGA reaction. For a 10 µl reaction, ∼40 ng of pYLCRISPR/Cas9Pubi-H plasmid was combined with ∼10 ng of each sgRNA cassette fragment (7:1 vector:insert ratio) and 1 U of BsaI-HFv2 (NEB) in 1× T4 ligase buffer (NEB). The reaction was incubated at 37 °C for 3 min and 16 °C for 4 min (30 cycles) followed by 5 min at 60 °C for enzyme inactivation. A 2 µl aliquot of the reaction was transformed into 20 µl of Escherichia coli TOP10 competent cells. Transformants were selected on LB media plates containing 50 µg ml⁻¹. Prior to transformation into Agrobacterium AGL1, the resulting plasmid, pYLCRISPR-HvLBO, was verified using restriction enzymes and Sanger sequencing with the M13F primer. The construct was transformed into immature embryos (Harwood, 2014). Following transgenic plant regeneration, transformants were screened for successful edits using PCR and Sanger sequencing of PCR products overlapping the regions using primers for Target 1: LBO-T1-F and LBO-T1-R and Target 2: LBO-T2-F and LBO-T2-R. Only HvLBO-Target-1 was successful in generating hvlbo mutants. Primer sequences are given in Supplementary Table S1.

HvLBO and hvlbo.f complementation test in the Arabidopsis atlbo-1 mutant

To obtain the HvLBO coding sequence (CDS), gene expression in different tissues and developmental stages was analysed by mining publicly available transcriptomic data through the IPK website (https://apex.ipk-gatersleben.de/apex/f?p=116:1). The seedling stage was chosen for RNA extraction because of the early HvLBO expression. To clone HvLBO and hvlbo.f CDSs, RNA was extracted by the Trizol protocol (Invitrogen®) from root samples of 8-day-old WT Sebastian and hvlbo.f homozygous mutant seedlings. After DNase I treatment and LiCl purification, reverse transcription was performed by the SUPERSCRIPT III Reverse-Transcriptase (Invitrogen®) according to the manufacturer’s protocol. The CDS was amplified with high-fidelity Platinum SuperFi II DNA Polymerase (Invitrogen®) by specific tailed primers, AK365108-FW and AK365108-RV, and cloned by Gateway reaction in the p221 pDONR and pH2GW7 pDEST containing the 35S cauliflower mosaic virus (CaMV) promoter and terminator.

Agrobacterium tumefaciens GV3101 was used to transform atlbo-1 homozygous mutants by floral dip (Clough and Bent, 1998). T₁ seeds were screened on hygromycin selective medium, and resistant seedlings were checked for transgene presence by PCR using primers HvLBO/hvlbo.f-FW and HvLBO/hvlbo.f-RV. Transgene expression was assessed by quantitative real-time PCR (qRT-PCR). To test expression in transgenic Arabidopsis plants, RNA was extracted from rosette leaves using the LiCl method (Verwoerd et al., 1989). After DNase treatment, for each sample, RNA was reverse transcribed using the iScript kit (BioRad). qRT-PCR assay was performed using iTaq Universal SYBR Green supermix (BioRad) in a Bio-Rad iCycler iQ Optical System (software version 3.0a). Relative transcript enrichment of the targets of interest was calculated by normalizing the amount of mRNA against the ubiquitin-like RCE1 transcript (At4g36800) (Pribil et al., 2010) and the EIF4 transcript (At3g13920) (Yamaguchi et al., 2009). Expression of the genes was calculated using the 2−ΔΔCt method, employing the WT as the normalizer. The primers HvLBO-SEED1 and HvLBO-SEED2 for qPCR were used to evaluate the expression level of both the HvLBO and hvlbo.f transcript, RCE1-147 and RCE1-148 for the RCE1 transcript, and RT2414 and RT2415 for the EIF4 transcript. T₂ and T₃ seeds were selected on hygromycin medium in order to track plants bearing the transgene in homozygosity or with a copy number that does not give WT segregation in up to 40 T₃ plants. Those T₃ seeds were used for the phenotypic analysis, and three plants of each T₃ line were checked for transgene expression and silencing. Primer sequences are shown in Supplementary Table S1.

Plant treatments and phenotypic analyses

Arabidopsis seeds were sown in multipot plateaus (2.5 cm diameter each pot) and stratified in soil for 2 d. Plants were grown in short-day conditions (8 h day/16 h night, 20–22 °C) for 14 d and then moved to long days (16 h day/8 h night, 20–22 °C). The short-day condition was used to delay flowering and increase rosette leaf production and branching. As near as possible to the first senescent silique, the number of rosette branches >5 mm was counted (Brewer et al., 2016). A total of 25–27 plants per genotype were used and grown in a partially block-randomized setting.

A greenhouse experiment (2019–2020) was set up to phenotype barley HorTILLUS lines carrying point mutations in hvlbo (Table 1) and the WT cultivar Sebastian, following a random block design, with 19 incomplete blocks. In the final design, we included 27 Sebastian plants and 19 hvlbo.f plants. Seeds were sterilized for 1 min in 1% sodium hypochlorite, rinsed in water once, then 2 min in 70% ethanol, and rinsed three times in water. Seeds were then sown on a rooting medium containing 2.15 g l⁻¹ Murashige and Skoog salts with vitamins (Duchefa Biochemie), 30 g l⁻¹ sucrose, 8 g l⁻¹ plant agar (Duchefa Biochemie), and 5 ml l⁻¹ indole-3-butyric acid (IBA; 0.1 mg ml⁻¹ stock) (Sparkes et al., 2006), and incubated in a growth room in the dark at 24 °C. When they reached the one-leaf stage, seedlings were transplanted in soil in 14×14×16 cm pots (one plant per pot) and grown in a greenhouse at Orto Botanico Città Studi, located in Milan, under a long-day photoperiod (16 h light at 20 °C and 8 h dark at 17 °C), 60–65% humidity, 200 μmol m⁻² s⁻¹ light. The number of tillers on different days after sowing (DAS) was recorded for each plant, along with plant height (distance between the point of emergence from the soil and the base of the spike), and tiller number at harvest stage.

Table 1.

HvLBO alleles isolated from screening of the HorTILLUS mutagenized population

Allele	Nucleotide change (from ATG)	Effect	Mutation type
lbo.d	G815A	D241N	Missense
lbo.e	G834A	S247N	Missense
lbo.f	G845A	D251N	Missense
lbo.ia	C915T	A274V	Missense

Gene-edited alleles were developed and assessed in a glasshouse under long days (16/8 h) at 21/16 °C with daylight supplemented with 400 W Metal Halide lamps (Plusrite) at the South Australia Research & Development Institute (SARDI) Plant Research Centre, University of Adelaide Waite Campus. Plants were grown in SARDI’s proprietary Cocopeat soil mix, containing 56% coir fibre, 44% quarried drainage sand, and controlled-release fertilizer (N added separately) at 2 cm sowing depth in 17.5 cm diameter, 11.6 cm height, 2.8 litre pots, with calcium nitrate watered in at the start at the specified rates. Hormone treatments contained either 2 μM GR24 (rac-GR24, PhytoTech Labs, G3324) or mock without hormone addition. Treatment solutions contained one drop of Triton-X surfactant and 10 μl of acetone (GR24 was dissolved in acetone) per 50 ml of MilliQ water. Hormone solutions were applied to tiller buds under the leaf sheaths using a pipette starting from mid-tillering phase (Zadok stage 24). Treatments were applied three times per week (Monday, Wednesday, and Friday) over a 6 week period. A 200 μl aliquot of solution was applied for early development, increasing to 400, 800, and 1200 μl as plant size progressively increased. Tiller number was manually scored at the end of development.

For hydroponic experiments in Japan, barley plants were assessed in half-strength Tadano and Tanaka (TT) (Tadano and Tanaka, 1980) liquid culture using our reported plastic cup hydroponic method (Yoneyama et al., 2020b) and including the specified nutrient restrictions.

Strigolactone collection from root tissues and exudates, and LC-MS/MS

SLs were collected and analysed following our reported methods for root tissues and exudates (Yoneyama et al., 2022; Inoue et al., 2024). Barley SLs were analysed using ultra performance LC coupled to tandem MS (UPLC-MS/MS). For this, the Acquity UPLC System (Waters) coupled to a Xevo TQD triple-quadrupole mass spectrometer (Waters MS Technologies) with an electrospray ionization (ESI) interface was used. Chromatographic separation was achieved using an ODS column (ACQUITY UPLC, BEH C18, 2.1×100 mm, 1.7 µm; Waters) with a water–MeOH gradient containing 4% 50 mM ammonium acetate to promote ionization. Separation started at 35% MeOH, followed by a 2 min gradient to 55% MeOH, followed by a 13 min gradient to 95%, kept at 96% MeOH for 2 min to wash the column, and then back to 35% MeOH for 3 min. The column was equilibrated at this solvent composition for 5 min before the next run. Total run time was 25 min. The column oven temperature was maintained at 40 °C with a flow rate of 0.2 ml min⁻¹ (sample injection volume of 1 µl). Multiple reaction monitoring (MRM) transitions for 6-epi-heliolactone eluting at 5.5 min were monitored for m/z 361/97 at a collision energy (CE) of 18 V and m/z 361/233 at a CE of 18 V with a cone voltage of 25 V. The MRM transitions of m/z 377/97 at a CE of 20 V and m/z 377/231 at a CE of 15 V with a cone voltage of 25 V were used for the detection of barleylactone 2 (a novel epoxy variant of 6-epi-heliolactone) eluting at 4.4 min. SLs are inherently low-abundance molecules that are difficult to precisely measure, hence we could only determine relative levels (‘peak areas’) between samples rather than absolute concentrations.

Arbuscular mycorrhizal fungi colonization assay

Arbuscular mycorrhizal (AM) fungi colonization assays were carried out at the Adelaide SARDI facilities using a clay loam that was dried and then sieved to <2 mm. A fine sand and the sieved soil were then sterilized by autoclaving at 121 °C for 20 min twice, with >24 h between autoclave runs. Free-draining, 0.8 litre plastic pots were used, and each pot held 1.1 kg of the soil mix (loam soil:dry sand, 1:1). Soil and sand were homogenously mixed prior to potting. The AM treatment was inoculated with four isolates of the AM fungus Rhizophagus irregularis as a commercial product (Start Up Ultra, Microbe Smart Pty Ltd, Australia) at a rate of 1 g kg⁻¹ soil (corresponding to ∼800 spores per pot) and then mixed thoroughly with the soil. The non-AM control pots had no inoculum mixed with the soil. For AM fungus uninoculated (–AMF) treatment, there were six pots prepared for each genotype (Golden Promise WT, lbo-2, and lbo-3). For AM fungus inoculated (+AMF), there were 10 pots prepared for each genotype, where six pots were for the main experiment and four extra pots were prepared for mid-experiment destructive sampling; 3–4 seeds were directly sown onto the soils. Plants were watered approximately every 2 d with 50 ml of water. Pots in which >3–4 seedlings emerged were thinned to two seedlings at 14 DAS. At 48 DAS, plants were gently removed from the soil and any roots with attached soil were washed clean. A representative subsample of fresh roots (∼ 200 mg) was taken and placed in a tissue cassette, submerged in 50% ethanol for at least 24 h to fix. The fresh root subsamples were subsequently rinsed and placed into 10% potassium hydroxide (KOH) (w/v) solution for clearing at room temperature for 7 d. Then, the samples were rinsed well and stained in a 5% ink in vinegar solution (Vierheilig et al., 1998), and then de-stained in acidified reverse osmosis (RO) water for at least 24 h, before being stored in a 50% glycerol solution. The roots were then quantified for percentage root length colonized by AM fungi under a stereomicroscope using the gridline intersect method (Giovannetti and Mosse, 1980).

Parasitic weed seed germination bioassay

A 2.5 ml aliquot of each collected sample of exudate was concentrated and applied to seeds of branched broomrape (Phelipanche ramosa) and Egyptian broomrape (Phelipanche aegyptiaca), and assessed for percentage germination, according to our published method (Inoue et al., 2024).

Data analysis

Data obtained from the above experiments were obtained via manual collection and were arranged in Microsoft Excel. All data were transferred to GraphPad Prism for statistical analysis and visualization. Comparisons between two groups used parametric t-tests to determine statistical relationships. For multiple comparison tests between >2 groups, one- and two-way ANOVA with Tukey post-hoc tests were used to determine statistical relationships.

Results

Barley HvLBO shows functional conservation in Arabidopsis thaliana

Sequence searches using Arabidopsis AtLBO At3g21420 identified a single barley gene (HORVU.MOREX.r3.3HG0314210) in the Morex V3 reference genome (Mascher et al., 2021), encoding a protein of 375 amino acids with high similarity to proteins from related species, rice, Brachypodium distachyon, and sorghum, and to Arabidopsis AtLBO (59% identity, 77% similarity) (Supplementary Fig. S1). The sequence contains within exon 2 a double-stranded β-helix conserved catalytic domain (DSBH), which is a hallmark of 2-OGDs (Fig. 1A). We previously showed that the enzymatic activity of this protein was the same whether it came from tomato, maize, sorghum, or Arabidopsis (Yoneyama et al., 2020a). To validate functional orthology of this candidate HvLBO from barley, an interspecific complementation of the Arabidopsis atlbo-1 mutant was employed. Whereas atlbo-1 showed increased branching (Brewer et al., 2016), 35S-driven constitutive expression of the HvLBO CDS in the Arabidopsis mutant fully restored branching to WT (Ws-4) levels in four independent lines (Fig. 1B). This confirms functional conservation of LBO across monocotyledonous and dicotyledonous plants. Expression of HvLBO in Arabidopsis did not inhibit branching below WT levels (Fig. 1C). However, high SL production would be expected to repress the endogenous SL pathway in Arabidopsis as a form of feedback (Hellens et al., 2023).

Fig. 1.

Analysis of the hvlbo.f mutant. (A) HvLBO encodes a 2-oxoglutarate and Fe(II)-dependent dioxygenase (2-OGD) across two exons, including a conserved double-stranded β-helix conserved catalytic domain (DSBH). A start codon was predicted (pale arrow). However, a second site is likely to be the real start codon (arrow), because this matches the protein sequence of BAJ96311 derived from a full-length cDNA library, which we confirmed by PCR of a cDNA sample. The locations of three independent mutant alleles are shown (arrowheads) (an in-frame mutation, hvlbo-1, was not included in this study). (B) A 35S:HvLBO construct was transformed into the atlbo-1 mutant allele of Arabidopsis. The rosette branch number was reduced to WT (Ws-4) levels in four independent transgenic T₃ lines. Data are means ±SEM, n=23–26. (C) Primers specific for HvLBO were used to detect relative expression of the transgene in independent plants from each T₃ line. Data are means ±SEM, n=3. (D) Mutant hvlbo.f plants isolated from the HorTILLUS TILLING population showed increased tiller number compared with the WT Sebastian cultivar (Seb) during the later phase of growth, but not earlier. Data are means ±SEM, n=19–27. (E) Plant height was reduced in hvlbo.f. Data are means ±SEM, n=19–27. (F) A 35S construct expressing the barley mutant hvlbo.f CDS was also transformed into the atlbo-1 mutant of Arabidopsis. The rosette branch number was only partially reduced to WT (Ws-4) levels in two out of four independent transgenic T₃ lines. Data are means ±SEM, n=23–27. (G) Primers specific for HvLBO/hvlbo.f were used to detect relative expression of the transgene in independent plants from each T₃ line. Data are means ±SEM, n=3. Different letters indicate statistically significant differences by one-way ANOVA Tukey’s test, P<0.05. Asterisks show significant differences by Student’s t-test, ****P≤0.0001.

Barley hvlbo mutants display a subset of strigolactone phenotypes

We employed TILLING using the HorTILLUS mutagenized population of two-row spring barley (Hordeum vulgare L. cv. ‘Sebastian’) (Szurman-Zubrzycka et al., 2018) to find potential hvlbo mutant alleles. Screening of 6912 M₂ plants resulted in the identification of 12 lines harbouring distinctive point mutations in the target gene. Out of these, four lines carrying independent missense mutations (Table 1) showed viable seed germination and were further examined under glasshouse conditions. All mutations in these hvlbo alleles fell within the putative LBO catalytic domain near the C-terminus. One allele in particular, hvlbo.f, contained a point mutation in the coding sequence at G845A, which would result in residue change D251N (Supplementary Fig. S1; Fig. 1A). Residue 251 is a conserved aspartic acid (D) residue that is predicted to play a critical role in iron binding within the LBO catalytic domain (Brewer et al., 2016). Thus, this mutation is predicted to significantly interrupt LBO enzymatic function, and we analysed this line further. Consequently, the hvlbo.f line exhibited increased tiller number at harvest compared with WT Sebastian (Fig. 1D) and reduced plant height (Fig. 1E), similar to SL-defective phenotypes reported in rice and other species (Gomez-Roldan et al., 2008; Umehara et al., 2008). Tiller number reached a plateau at 64–70 DAS in the WT, but continued to increase for a further 2 weeks in the hvlbo.f line (Fig. 1D).

To validate these results, a construct expressing the hvlbo.f CDS under the control of the CaMV 35S constitutive promoter was introduced in the Arabidopsis atlbo-1 mutant, and branch number (Fig. 1F) and the relative expression of the transgene was evaluated in T₃ lines (Fig. 1G). According to the phenotype of four T₃ independent lines, hvlbo.f was not able to fully complement the atlbo-1 branching phenotype in any of the analysed lines. Indeed, we detected a partial complementation only in two out of the four analysed lines (Fig. 1F). These results suggest only a partial functionality of the lbo.f allele.

Further barley mutant alleles were generated by targeting exon 1 and exon 2 of HvLBO with a CRISPR/Cas9 [clustered regularly interspaced palindromic repeats (CRISPR)/CRISPR-associated protein 9] site-directed nuclease construct transformed into barley (H. vulgare L. cv. ‘Golden Promise’). Two new alleles, hvlbo-2 and hvlbo-3, were identified that carry single base changes resulting in frameshifts early in the CDS, which would result in severe protein truncation and loss of the catalytic domain (Fig. 1A). We grew hvlbo-2 and hvlbo-3 mutants individually in hydroponic cups and compared plants with hvd14.d, a previously isolated mutant in the SL receptor gene (Marzec et al., 2016, 2020a). The hvlbo mutants showed increased tillers (Fig. 2A–C). Mutant plants of hvd14 showed on average 2.2 (±0.4) extra tillers at 18 d old compared with the WT variety (H. vulgare L. cv. ‘Sebastian’), while hvlbo-2 and hvlbo-3 showed 1.3 (±0.3) and 1.6 (±0.5) extra tillers, respectively, compared with Golden Promise (Fig. 2B). Plants in hydroponic cups have a limited lifespan and can experience some stress, which may affect tillering. Hence, we examined other plants grown individually in pots with normal soil and nutrients. Analysis of tiller number over time revealed that hvlbo-2 mutant tiller numbers were not significantly different during early growth compared with the WT (Golden Promise), but hvlbo-2 ended up with more tillers (Fig. 2C), consistent with results obtained with hvlbo.f (Fig. 1D). This confirms that SLs are more important for tiller repression in older plants. Further experiments are required to understand whether this represents a direct effect on older tiller bud outgrowth.

Fig. 2.

Analysis of hvlbo gene-edited mutants. (A) Mutant hvlbo-2 and hvlbo-3 plants grown in hydroponic cups. Arrowheads indicate tillers. (B) Mutant hvlbo-2 and hvlbo-3 plants showed increased tillering at 18 d old compared with the WT variety [Golden Promise (GP)], but not as much as hvd14.d [compared with WT Sebastian (Seb)]. (C) In pots with soil and normal nutrients, hvlbo-2 tiller number was not different at 33 d post-emergence, but became greater in older plants. Root tissue was collected from the plants in hydroponic cups at 28 d old, after 10 d of P starvation. Extracts from the root tissues were analysed for SLs by LC-MS/MS. (D) 6-Epi-heliolactone was detected in the WTs and mutants, at similar relative levels. (E) Barleylactone 2 was detected in the WTs and mutants, at similar levels. Other SLs were not detected. (F) Root biomass was not significantly different. Data are means ±SEM, n=4–8. Asterisks show significant differences by Student’s t-test, *P≤0.05, **P≤0.01.

Other individual plants were grown in hydroponic cups and transferred at 18 d into phosphorus (P) starvation conditions to boost SL production. At 28 d, the root tissues were extracted and assessed for any SLs. Recently 6-epi-heliolactone and barleylactone 2 (a novel epoxy variant of 6-epi-heliolactone) have been detected as predominant SLs in barley (Inoue et al., 2024). We also detected these SLs in root tissues in all genotypes (Fig. 2D, E). Surprisingly, the peak areas of the mass spectrum were all very similar, including from hvlbo-2 and hvlbo-3, indicating similar levels of SL production, regardless of the genetics (Supplementary Fig. S2; Fig. 2D, E). Root biomass was not greatly different, suggesting that root growth is not a compounding factor that may affect SL production (Fig. 2F). Similar SL levels in WT and lbo mutants suggest that LBO is not required for 6-epi-heliolactone and barleylactone 2 production within roots. Lack of SL signalling in d14 might be expected to increase SL production due to feedback (Arite et al., 2009). However, we did not see increased SL in hvd14.d, which may indicate a lack of D14-dependent feedback in barley (at least for 6-epi-heliolactone and barleylactone 2 content within roots).

Barley hvlbo mutant tiller number is greater and less reactive to nutrient conditions

As noted above, nutrients can repress SL production, and P or nitrogen (N) starvation is a common way to boost SL production for research (Brewer et al., 2013). Nutrients promote the number of grain-bearing shoots (branches/tillers), which enables plants to increase progeny in proportion to greater resource availability (Brewer et al., 2013). Hence, fertilizer can increase yield in crops by boosting the number of tillers. Conversely, this relationship can have adverse effects on final tiller number and yield if growers seek to reduce fertilizer use (Francis et al., 2023; Kelly et al., 2023). SL biosynthesis gene expression, including of LBO, can also be highly suppressed by N (Lailheugue et al., 2023). We supposed that LBO, as a key SL enzyme, would also be part of nutrient responses in barley. Hence, we examined hvlbo-2 in more detail. Plants were grown in pots in soil with low, intermediate, and normal N levels, and phenotypic changes were examined. Mutant hvlbo-2 plants consistently displayed greater tiller numbers than the WT control variety (Golden Promise) (Fig. 3A). Tiller number aligned well with grain number (Fig. 3B) and yield per plant (Fig. 3C), and inversely with average individual grain (kernel) mass (Fig. 3D). In each case, WT response to N was greater than that of hvlbo-2 and tended to converge with the mutant at higher N. This aligns with the known role of SLs in adapting crop plants to nutrient levels (Brewer et al., 2013; Sigalas et al., 2024; Huang et al., 2025). Mutant and WT phenotypes tend to converge in high N, because the SL pathway becomes suppressed by N in the WT (Wu et al., 2020; Sun et al., 2023). Hence, under optimal conditions, SL levels in the WT and mutant plants are similarly low and the plants appear more similar in phenotype.

Fig. 3.

hvlbo mutant phenotype on low (250 mg kg^–1 soil), intermediate (500 mg kg^–1), and normal (750 mg kg^–1) nitrate (N) added at sowing, compared with the WT (Golden Promise). (A) hvlbo-2 mutant tiller number, (B) grains per plant, (C) yield per plant, (D) average grain (kernel) mass per plant, (E) harvest index (yield as a proportion of total shoot biomass), (F) percentage of seed set (a grain present) at each seed position, (G) total shoot biomass including grains, and (H) plant height. Data are means ±SEM, n=10. Different letters indicate statistically significant differences by two-way ANOVA Tukey’s test, P<0.05.

Barley hvlbo mutants retain yield and fertility at low N, but produced grains of reduced mass

Interestingly, hvlbo-2 showed 29% greater yield than the WT at 250 mg kg^–1 N (Fig. 3C). Harvest index (Fig. 3E) and seed set (Fig. 3F) were also higher at 250 mg kg^–1 N. This means that, in the WT, SLs produced by LBO contribute to a repressive effect on yield, harvest index, and fertility in low N, because these inhibitions were not as strong in the mutant. Hence, mutant variants of lbo may be useful in breeding to prevent yield loss from poorly fertile soil or low fertilizer use. On the other hand, yield was 13% lower in hvlbo-2 at 750 mg kg^–1 N. This was mostly due to lower individual grain mass that offset increased tiller and grain number. Smaller grain size has been observed in SL mutants in rice and maize (Yamada et al., 2019; Guan et al., 2023). This implies that SLs produced by LBO in the WT help plants to maintain the high mass of individual grains, even if grain number increases. Hence, LBO-related SLs may have positive or negative effects on yield and grain mass, depending on the conditions. Shoot biomass (Fig. 3G) and plant height (Fig. 3H) were only slightly different, indicating that the mutants did not show a strong growth or stature defect, which might otherwise repress grain mass. Likewise, root biomass was not greatly different in various experiments (Fig. 2F and see below).

Barley hvlbo responded to strigolactone treatment, but showed normal root fungal symbiosis

To further test hvlbo mutants, we added a synthetic SL (rac-GR24) to hvlbo-2 mutants and observed that tiller number was repressed to near the level of the WT (Fig. 4A). This suggests that barley tiller buds are responsive to SLs and that hvlbo mutants are deficient in the production of a branching-inhibiting SL. SL branching repression below that of the control WT was not observed. SL feedback homeostasis, whereby SLs repress SL biosynthesis, is proposed to explain resistance to exogenous SLs in the WT (Dun et al., 2023). The lack of increased SL production in hvd14.d roots (Fig. 2D, E) may contradict that, but autoregulation of the signalling complex may also provide SL resistance (Dun et al., 2023).

Fig. 4.

The effects of SL treatment and AM fungal inoculation (Rhizophagus irregularis) on barley hvlbo mutants. (A) Doses of 2 µM rac-GR24 repressed tiller number in hvlbo-2 but not in WT Golden Promise (GP). Data are means ±SEM, n=10. (B) The percentage presence of AM hyphae, vesicles, and arbuscules was not obviously different between GP, hvlbo-2, and hvlbo-3 at 48 dpi. Data are means ±SEM, n=6. (C) Shoot FW at 34 dpi trended upwards in the mutants compared with the WT (GP), and downwards with AM fungal inoculation. (D) Root FW trended downwards both in the mutants and with AM fungal inoculation. Data are means ±SEM, n=5–6. Different letters indicate statistically significant differences by two-way ANOVA Tukey’s test, P<0.05.

We also examined the capacity for AM fungal colonization in hvlbo-2 and hvlbo-3 mutant roots at 48 days post-inoculation (dpi). Hyphae, vesicles, and arbuscules were observed in roots, indicating that colonization had occurred. However, no significant differences were observed between mutants and the WT in terms of overall colonization and fungal morphology (Fig. 4B). In addition, we confirmed no change in AM fungal colonization in hvlbo-2 and hvlbo-3 in an independent laboratory using alternative inoculation and scoring methods (Supplementary Fig. S3; Supplementary Protocol S1). AM fungal colonization coincided with slightly repressed root and shoot biomass at 34 dpi (Fig. 4C, D). This result may indicate a carbon drain, negatively affecting plant growth, as has been reported in barley previously (Grace et al., 2009; Kaur et al., 2022). Importantly, similar biomass reduction trends were seen across WT, hvlbo-2, and hvlbo-3 (Fig. 4C, D), suggesting that the fungal interaction is functioning in a similar way in the mutant and WT. Root biomass was slightly reduced compared with the WT in hvlbo-2 and hvlbo-3, and the shoot biomass trended slightly higher, which suggests that the LBO pathway may play a role in balancing root and shoot growth, in a manner unrelated to AM fungal colonization. A thorough examination of hvlbo root architecture responses to nutrients would be useful for future studies.

Barley hvlbo mutants have normal strigolactone root exudates

Having established that lbo mutant plants show phenotypes that align with SL deficiency, and yet show normal SL production in root tissues, and no obvious signs of deficient AM fungal symbiosis activity, we next sought to test whether exuded SL was defective in mutants. We grew hvlbo-2 mutant plants in P- and N-deficient conditions to boost SL production, and the SLs exuded from roots were collected and analysed by LC-MS/MS. We detected and identified the same two SLs (Fig. 2D, E): 6-epi-heliolactone and barleylactone 2 (Supplementary Fig. S4; Fig. 5A, B). As expected, secretion of both SLs, 6-epi-heliolactone (Fig. 5A) and barleylactone 2 (Fig. 5B), greatly increased in the WT. This was especially so for N deficiency. Remarkably, hvlbo-2 mutants showed very similar increases. Parasitic weed seed germination is a reliable bioassay used to verify that bioactive SLs are present in a sample, particularly when SL levels are near the limit of detection (Ablazov et al., 2025). The WT and mutant-collected exudate stimulated parasitic weed seed germination, with trends matching the SL detection data (Fig. 5C, D), confirming similarity of SL levels. This implies that the product of LBO is not an important contributor to SL exudate production.

Fig. 5.

hvlbo mutant plants grown in hydroponics with phosphorus (–P) or nitrogen (–N) starvation conditions. SL exudates analysed by LC-MS/MS. The WT (Golden Promise; GP) and mutant show similar large increases in (A) 6-epi-heliolactone and (B) barleylactone 2. Data are means ±SEM, n=2–5. SL exudate collected samples were also applied to parasitic weed seeds, (C) branched broomrape (Phelipanche ramosa), and (D) Egyptian broomrape (Phelipanche aegyptiaca), and assessed for percentage germination. Data are means ±SEM, n=9. Not significant (ns) is as per two-way ANOVA Tukey’s test, P>0.05.

Discussion

Until recently, all reported SL mutants were highly branched. Also, SLs can inhibit buds when applied exogenously, and there usually exists only one copy of the SL receptor in a species. Hence, it had been assumed that all SLs produced within plants will inhibit buds. However, research into the subpathway of SL biosynthesis that involves CLAMT and LBO, along with new SL biosynthesis mutants that are not highly branched, suggests that there may be (at least) exudate- and branching-specific SL pathways (Wakabayashi et al., 2019; Ito et al., 2022; Chen et al., 2023; Li et al., 2023; Wang et al., 2024). This may mean that, despite years of research into branching and SLs, the branching-specific SL is as yet unknown (Yoneyama and Brewer, 2021; Dun et al., 2023; Nomura et al., 2024). Therefore, it is useful to discover whether lbo mutants are defective in exuded SLs. In this study, we show that lbo mutants in barley fail to produce the SL-related branching inhibitor, but are still able to make exuded SLs.

Exudate SLs may include various structures, including four-ringed types (canonical) and non-four-ringed types (non-canonical). The production of known canonical SLs mostly involves consecutive reactions from P450 enzymes deriving from CLA, a non-bioactive, two-ringed SL intermediate (Dun et al., 2023). Plant species and varieties can exude their own distinctive mix of SLs. This is presumably to avoid parasitic weed infection but still retain AM fungal interactions (Yoneyama and Brewer, 2021). It is not yet well known how the diversity in exuded SLs relates to internal hormone function and evolution. Branching-specific SLs are likely to be CLA-derived and non-canonical. MeCLA is produced from CLA by the CLAMT enzyme (Wakabayashi et al., 2021; Mashiguchi et al., 2022; Haider et al., 2023; Li et al., 2023), and MeCLA is a precursor for non-canonical SLs (so far). MeCLA was shown to be the precursor to lotuslactone, zealactone, or avenaol, all found in root exudates (Yoneyama and Brewer, 2021; Dun et al., 2023). LBO is a 2-OGD enzyme that acts to convert MeCLA into 1'-HO-MeCLA, a function that appears to be conserved in monocotyledonous and dicotyledonous plants (Brewer et al., 2016; Yoneyama et al., 2020a). Mutants of clamt or lbo in Arabidopsis show increased branching, suggesting that this pathway produces a branching-inhibiting SL (Brewer et al., 2016; Zhou et al., 2025). This also means that LBO is currently the enzyme in the most distal position in the pathway that produces the branching SL. However, the identity of that SL or what further conversions or transport may occur are unknown. The LBO enzyme, when heterologously expressed in E. coli, produces 1'-HO-MeCLA, but it also produces CLA, and at greater levels (Yoneyama et al., 2020a). Mutants of lbo have high levels of MeCLA and undetectable levels of 1'-HO-MeCLA (Yoneyama et al., 2020a). 1'-HO-MeCLA is predicted to be unstable, and CLA may be a non-enzymatic reversion from 1'-HO-MeCLA that perhaps occurs in the absence of further enzymatic processing or perception (such as would be absent in E. coli). However, the instability of 1'-HO-MeCLA may mean it cannot be chemically synthesized for further testing. Likewise, further research into the in planta function of LBO remains challenging.

Meanwhile, lbo mutants so far have only been revealed in Arabidopsis (Brewer et al., 2016; Zhou et al., 2025). However, Arabidopsis seems to have a simplified SL biosynthesis pathway, perhaps because hormone function has been retained but AM fungal symbiosis has been lost during evolution. Hence, subpathways of SL biosynthesis relating to AM fungi may also be lost. Therefore, it may be useful to study lbo mutants in species that have the capacity to associate with AM fungi. Hence, we isolated hvlbo mutants in barley by TILLING and gene editing. The mutants showed the expected increased tillering phenotype of SL mutants (Figs 1D, 2A–C, 3A, 4A), although not as strong as hvd14 (Fig. 2B). Previously, Arabidopsis lbo mutants showed intermediate branching, perhaps due to a buildup of MeCLA, which presumably is only weakly bioactive (Brewer et al., 2016). It is also important to consider that SLs are not the only signal that regulates branching. Cytokinins, sugar signalling, and bud auxin transport also promote tillering and still function in SL mutants (although they may be altered by complex feedback mechanisms) (Beveridge et al., 1997; Zhang et al., 2020; Salam et al., 2021).

Mutants of hvlbo also showed a decrease in average individual grain mass (Fig. 3D). This is similar to rice and maize SL mutants (Yamada et al., 2019; Guan et al., 2023). Increased grain number in hvlbo was largely offset by reduced grain mass, which tended to make yield per plant similar to that of the WT (Fig. 3C). Further research is needed to determine whether the lower grain mass in hvlbo-2 is due to resource limitation from too many tillers or any defect directly in grain development. The latter is currently favoured based on tiller removal in rice SL mutants that did not restore the smaller grain size (Yamada et al., 2019). The hvlbo mutants had levels of two SLs in root tissues that were similar to those of WT and the hvd14.d mutant (Fig. 2D, E), yet tiller number was repressed by adding SL (Fig. 4A). This may indicate that LBO is defective in making some type of SL, but not these two. It is unclear if other SLs are made by barley. There are five MAX1 homologue genes in barley that are partly conserved and partly divergent from the MAX1 homologues in rice (Marzec et al., 2020b), so other SLs may exist. 5DS was previously reported for barley (Wang et al., 2018), however, we could not detect it or any other SLs. We are currently isolating a range of SL biosynthesis mutants, which will help to elucidate the entire SL pathway in barley, and clarify the role of LBO.

The hvlbo mutants showed a general lack of response to nitrate, which would be expected for an SL mutant. The lack of response ultimately meant that yield was higher at low nitrate, but lower at higher nitrate (Fig. 3C). This opens up the potential to develop lbo-related genetic variants to employ in poorly fertile soils or when less fertilizer is used (Kelly et al., 2023). Field trials will be useful to further explore this concept. In the case of barley, consistent grain size, germination, and protein content is important for malting. So, grain quality analysis, such as germination and protein levels, will be useful in future. A trade-off of low grain mass may not be desirable for malting barley, but may be acceptable for livestock feed. Also, because barley hvd14 mutants were shown to be defective in drought response (Marzec et al., 2020a), hvlbo mutants should also be assessed for this trait in the future. We do not yet know exactly what traits may be regulated by the LBO pathway, in addition to branching and grain size. For instance, AM fungi symbiosis was not obviously affected in the mutants (Fig. 4B), but the stress response is as yet unknown. Plants were shorter in hvlbo.f (Fig. 1E) compared with lbo-2 (Fig. 3H). Further research is needed to test whether the growth conditions, cultivar background, or a second-site mutation may have specifically enhanced dwarfing in hvlbo.f. To date, we would not yet rule out shade avoidance as a trait regulated by LBO.

We observed normal SL exudation and effects in hvlbo mutants. AM fungi symbiosis (Fig. 4B), the production of two SLs in root tissue (Fig. 2E, F) and exudates (Fig. 5A, B), and exudate germination stimulation of parasitic weed seeds (that are extremely sensitive to SLs) (Fig. 5C, D) suggest that SL production in roots is normal in hvlbo mutants. SLs affect root architecture (Brewer et al., 2013). The root biomass of hvlbo mutants was not greatly different, but it will be important to discover whether there are subtle differences. Remaining important questions are: does LBO have any role in roots, and can root SL exudate production be separated from root architecture?

We uncovered two SLs at similar levels in WT and hvlbo roots and exudates. Importantly, this means that LBO is not required to produce those SLs and that those SLs are not involved in branching. 6-Epi-heliolactone has been identified as a biosynthetic intermediate between MeCLA and avenaol in black oat (Moriyama et al., 2022). Treatment with prohexadione, a 2-OGD inhibitor, repressed the conversion of 6-epi-heliolactone to avenaol in black oat (Moriyama et al., 2022). However, our results suggest that LBO is not involved in 6-epi-heliolactone production, at least in barley. Another 2-OGD may produce avenaol, such as LOTUSLACTONE-DEFECTIVE (LLD), a 2-OGD involved in lotuslactone production in the legume Lotus japonicus (Mori et al., 2020), or an as yet unknown 2-OGD.

The detection of normal levels of SLs in roots and exudates in the lbo mutant suggests that the LBO enzyme is not significantly involved in the biosynthesis or exudation, or even the feedback, of those SLs. This supports the hypothesis that LBO has specificity for branching, but not exudation. hvlbo mutant plants can make SLs, and those SLs are not able to significantly inhibit branching as an internal hormone. It is possible that they simply cannot travel to tissues where branching inhibition occurs. Transport can be influenced by SL structure (Xie et al., 2016). It has been suggested that the hormonal function of LBO is to convert MeCLA into branching SLs close to buds (Brewer et al., 2016; Yoneyama and Brewer, 2021; Dun et al., 2023). Differences in chemistry may mean that CLA is a soluble transported form that is converted to MeCLA in target tissues for further processing by LBO (Dun et al., 2023). Thus, specialized transport and delivery near buds may be a key specification for the branching SL and may explain why other SLs cannot function in that way. Novel SLs are regularly being discovered, even in well-studied species. Work will continue to discover the gamut of SLs in barley, especially those particularly involved in bud outgrowth.

Abbreviations

AM

arbuscular mycorrhizal

CaMV

cauliflower mosaic virus

CDS

coding sequence

CL

carlactone

CLA

carlactonoic acid

CLAMT

CLA METHYLTRANSFERASE

DSBH

double-stranded β-helix conserved catalytic domain

1′-HO-MeCLA

hydroxymethyl carlactonoate

HorTILLUS

Hordeum vulgare TILLING University of Silesia

LBO

LATERAL BRANCHING OXIDOREDUCTASE

MeCLA

methyl carlactonoate

2-OGD

2-oxoglutarate and Fe(II)-dependent dioxygenase

SL

strigolactone

TILLING

Targeting Induced Local Lesions in Genomes

WT

wild type

Supplementary data

The following supplementary data are available at JXB  online.

Table S1. List of primers used in this study.

Fig. S1. Amino acid sequence alignment of barley HvLBO.

Fig. S2. Mass spectrum data for Fig. 2.

Fig. S3. Arbuscular mycorrhizal fungal colonization in 6-week-old plants.

Fig. S4. Mass-spectrum data for Fig. 5.

Protocol S1. AM fungi colonization in barley hvlbo mutants for Supplementary Fig. S3.

Author contributions

LR, KY, and PBB: conceptualization; AS, JHK, WC, AJ, LR, KY, and PBB: methodology; MI, AS, JHK, WC, HZ, CCF, SS, AB, SJWW, AJ, RA, BC, LR, KY, and PBB: formal analysis; MI, AS, JHK, WC, HZ, CCF, VG, AV, SS, AB, SJWW, AJ, RA, BC, LR, KY, and PBB: investigation; DZ, AJ, MM, LR, KY, and PBB: resources; LR and PBB: writing—original draft; AS, JHK, CCF, VG, MTW, MRT, SJWW, AJ, LR, and PBB; writing—review and editing; KY and PBB: visualization; VG, MRT, DZ, LR, KY, and PBB: supervision; JHK, VG, DZ, AJ, MM, LR, KY, and PBB: funding acquisition.

Funding

This research was supported by Australian Research Council (ARC) Fellowship (FT180100081) and South Australian Grain Industry Trust grant (UAD 02023) to PBB; the ARC Discovery Project (DP240102441) to MTW and PBB; ARC Centre of Excellence in Plants for Space (CE230100015) to MRT and PBB; an Australian Government Research Training Program (RTP) Scholarship and a Thyne Reid Foundation/Playford Trust Scholarship to JHK, the ERA-NET Cofund FACCE SURPLUS (BarPLUS grant id. 93) funded by the Italian Ministry for Universities and Research to LR and by the National Centre for Research and Development in Poland to AJ. AR was supported by a scholarship from the Saudi Arabia Ministry of Higher Education-King Saud University. Work at the University of Milan was in part supported by the CLIMBER project funded under the University’s 2019 SEED call (grant no. 1253 to LR and VG). LR also acknowledges financial support under the National Recovery and Resilience Plan (NRRP), Mission 4, Component 2, Investment 1.1, Call for tender No. 104 published on 2.2.2022 by the Italian Ministry of University and Research (MUR), funded by the European Union—NextGenerationEU– Project Title ‘RADICALS—root and shoot developmental insights for crop agricultural traits affecting resilience, competitiveness and sustainability’—CUP G53D23004030006—Grant Assignment Decree No. 1048 adopted on 14/07/2023 by the Italian Ministry of Ministry of University and Research (MUR). KY was supported by the Japan Science and Technology Agency (FOREST, JPMJFR220F) and the Japan Society for the Promotion of Science (KAKENHI 21H02125, 22H02270).

Data availability

All data are presented in the article and in the supplementary figures.