The modified vector pCAMBIA1300-cFlag (Yao et al., 2016) carrying the full coding sequence of Arabidopsis thaliana D14 (AtD14), S. spontaneum D14b (SsD14b), N-terminal (amino acids 1–49) truncated S. spontaneum D14a (SsD14aΔN), N-terminal (amino acids 1–44) truncated S. spontaneum D14b (SsD14bΔN) and S. spontaneum MAX2 (SsMAX2) under the control of the CaMV 35S promoter was introduced into the Atd14-1 (Salk_057876) (Waters et al., 2012) or Atmax2-3 (Salk_092836) (Jia et al., 2014) mutant by using the Agrobacterium-mediated floral dip method.

Similarly, we used GoldenBraid 2.0 system (Addgene¹) (Sarrion-Perdigones et al., 2013) to construct binary plant expression vectors: 35S:SsD14b–35S:SsMAX2 (P35s:SsD14b:Tnos–P35s:SsMAX2:Tnos–Pnos:NptII:Tnos), 35S:SsD14bP304R–SsMAX2 (P35s:SsD14bP304R:Tnos–P35s:SsMAX2:Tnos–Pnos:NptII:Tnos), and 35S:AtD14–35S:AtMAX2 (P35s:AtD14:Tnos–P35s:AtMAX2:Tnos–Pnos:NptII:Tnos), which were introduced into the Atd14-1 Atmax2-1 double mutant, respectively, to generate transgenic plants. The primary rosette branching numbers were counted for 5-week-old plants, which were germinated on plates and grown in soil under a light/dark photoperiod of 16 h/8 h at 22°C.

To construct plasmids for yeast two-hybrid (Y2H) assays, the CDS of SsD14a/b and SsD14a/b truncations (SsD14aΔN49 and SsD14bΔN44) were cloned into yeast expression vector pGBKT7 to generate BD-SsD14a/b and BD-SsD14a/b-ΔN, and we also constructed the mutations BD-SsD14aR310P and BD-SsD14bP304R. Similarly, we obtained BD-OsD14 and BD-OsD14ΔN53. The CDS of SsMAX2, AtMAX2, and AtSMXL7 were cloned into pGADT7 to make Gal4 DNA activation domain (AD) constructs, respectively. Y2H assays were performed using the Yeastmaker Yeast Transformation System 2 (Clontech, United States). In brief, yeast strain AH109 cells were co-transformed with specific bait and prey constructs and coating on selective growth medium SD/-Leu/-Trp for 3 days at 30°C, pick the positive constructs into liquid-selective growth medium SD/-Leu/-Trp for 36 h at 30°C, 200 rpm. Washed yeast cells three times and diluted, make sure OD₆₀₀ reached 2.5, then serial 10-fold dilutions of yeast cultures were spotted onto selective growth medium that was supplemented with 5 μM rac-GR24 or dimethyl sulfoxide (DMSO). All yeast transformants were grown on selective growth medium at 30°C, 4 days.

The positive clones of SsD14aΔN (residues 1–49) proved by DNA sequencing were transformed into Escherichia coli strain BL21 (DE3) for protein expression. Kanamycin-resistant colonies were picked to grow in the Luria–Bertani (LB) medium (10 g/L tryptone, 10 g/L NaCl, and 5 g/L yeast extract) at 37°C until OD₆₀₀ reached 0.6–1.0. Then 0.5 mM isopropyl-beta-D-thiogalactopyranoside (IPTG) was added to induce protein expression at 16°C for 18 h. The cell pellet was resuspended in phosphate-buffered saline (PBS) buffer containing 30 mM imidazole, and homogenized by using an ultrahigh pressure cell disrupter (JNBIO, Guangzhou, China). The lysate was centrifuged at 15,000 rpm for 1 h, and soluble proteins were loaded onto the Ni-NTA column. Target proteins were eluted by the PBS buffer containing 300 mM imidazole. The eluted SsD14aΔN (residues 1–49) was further purified by Superdex^TM75 (GE Healthcare, United States) at 16°C with the buffer containing 150 mM NaCl, 2 mM MgCl₂, 20 mM Tris pH 8.0, and 10% glycerol.

Purified SsD14aΔN (residues 1–49) (roughly 10 mg/ml) were dissolved in the buffer containing 150 mM NaCl, 2 mM MgCl₂, 20 mM Tris pH 8.0, and 10% glycerol. The crystals of SsD14aΔN (residues 1–49) were obtained using the hanging-drop method by mixing 1 μl protein with equal volume of reservoir solution containing 0.01 M magnesium chloride hexahydrate, 0.05 M Tris hydrochloride pH 7.5, 5% v/v 2-Propanol at 16°C for 1 week. The data of the SsD14aΔN (residues 1–49) crystal were collected on beamline BL17U1 at Shanghai Synchrotron Radiation Facility (SSRF) and processed by XDS (20124692). The structure of SsD14aΔN (residues 1–49) was determined by molecular replacement, using the structure of OsD14ΔN (residues 1–51) (PDB ID: 3VXK) as the initial searching template. Model building and structural refinement were performed by using COOT (20383002) and PHENIX (22505256), respectively. In the final model, more than 97% residues fall in the favored region in the Ramachandran plot, and the final R_work/R_free is 0.1914/0.2275. Data collection and refinement statistics are summarized in Table 1. The atomic coordinates and structure factors have been deposited in the Protein Data Bank.

The SL biosynthesis and core signaling pathways have been thoroughly studied in many plant species including Arabidopsis thaliana and Oryza sativa (Figure 1A), but remain to be investigated in sugarcane. To identify and investigate the SL receptor(s) D14 in S. spontaneum (SsD14), we searched Saccharum Genome Database (SGD)² (Zhang et al., 2018) using BLAST with Arabidopsis thaliana D14 (AtD14) and Oryza sativa D14 (OsD14) as queries to obtain the predicted sequences of D14 orthologs from S. spontaneum. Accordingly, we found two putative D14 orthologous genes SsD14a (Sspon.001B0005800) and SsD14b (Sspon.001B0005830) in S. spontaneum.

The phylogenetic analysis showed that SsD14s exhibit closer relationships with OsD14 from rice, which belongs to Gramineae too (Figure 1B). The similarity between SsD14a and OsD14 is 84.91% at the amino sequence, and the similarity between SsD14b and OsD14 is 85.94%. Sequence alignment and structural annotation showed that SsD14a/b, AtD14, and OsD14 exhibit both considerable identities at the primary amino acid sequence level and have the same catalytic triad Ser-His-Asp (Figure 1C). These information implies conserved physiological functions of SsD14 proteins.

Similarly, we searched SGD to obtain the predicted sequences of MAX2 and SMXL7 orthologs from S. spontaneum. Then, we found the putative orthologous genes SsMAX2 (Sspon.008D0018870) and SsSMXL7 (Sspon.007A0023280). To determine the biochemical function of SsD14a and SsD14b, we used Y2H assays to examine the interaction of SsD14s proteins with SsMAX2, AtMAX2, SsSMXL7, and AtSMXL7. Surprisingly, there were significant binding ability differences between SsD14a and SsD14b. The results showed that SsD14a interacted with SsSMXL7 and AtSMXL7 and interacted with AtMAX2 slightly. However, SsD14b interacted with neither AtMAX2 nor AtSMXL7. Meanwhile, Y2H results showed a strong interaction of SsMAX2 with both SsD14a and SsD14b (Figure 2). In other words, although SsD14a and SsD14b share 97.47% similarity in amino acid sequence, they have different preferences in binding downstream signaling partners, which leading us to speculate that the differences in interactions are attributed to some of these different residues.

Previous reports showed that many D14 of Gramineae species contain an extra N-terminal peptides when compared to AtD14 (Yao et al., 2018). Related studies have proved that both the full-length OsD14 and the N-terminally truncated OsD14 were able to complement the multi-branching mutant Arabidopsis d14-5, even the N-truncated D14 have more stronger interaction with AtMAX2 and complement d14 mutant better than the full-length version (Yao et al., 2018). According to our results of Y2H assays, SsD14aΔN can interact with AtMAX2 and AtSMXL7 as the full-length SsD14a did (Supplementary Figure 1). SsD14aΔN was introduced to complement Arabidopsis d14-1 mutants. We also generated the 35S:AtD14 d14-1 plants as positive control. The results showed no significant difference between the number of primary branches between 35S:SsD14aΔN d14-1 and 35S:AtD14 d14-1 (Figures 3A,B), which means that SsD14aΔN was able to rescue the multi-branching phenotypes. In addition, the leaf morphology (length/width ratio) was also recovered by SsD14aΔN (Supplementary Figure 2A). Therefore, SsD14a is functionally conserved when compared with AtD14.

However, the complementation results were quite different for SsD14b. According to the Y2H results, neither N-terminal truncated SsD14b nor full-length SsD14b could interact with AtMAX2 and AtSMXL7 (Figure 2 and Supplementary Figure 1). We transferred the full-length SsD14b to the Arabidopsis d14-1 mutant and obtained 35S:SsD14b d14-1 plants. We found that SsD14b cannot rescue the d14-1 multi-branching phenotype (Figure 4A). But interestingly, we found that the height of transgenic 35S:SsD14b d14-1 seemed to have a partial restoration (Figure 4B), which will be further investigated in the future project. We found that the multi-branched phenotype of one complemented line was only partially restored in 35S:SsD14bΔN d14-1 transgenic lines and still differed from WT, indicating that 35S:SsD14bΔN cannot fully complement Atd14-1. The difference between 35S:SsD14b d14-1 and 35S:SsD14bΔN d14-1 was that multi-branching and the leaf morphology of 35S:SsD14bΔN d14-1 were rescued in different degrees but both not thoroughly (Figure 4 and Supplementary Figure 2B). No obvious interactions of SsD14b/SsD14bΔN with AtSMXLs were detected in our work, which is probably because that the interactions were too weak to be detected in our current Y2H system. Consistent with this, the complementation effect of SsD14bΔN is significantly lower compared to SsD14a. Taken together, SsD14a and SsD14b may function as conserved and divergent SL receptors in sugarcane, respectively.

The crystal structure of SsD14aΔN was determined at a resolution of 1.65 Å (Table 1). SsD14a belongs to the α/β hydrolase superfamily, of which the structure consists of an α/β hydrolase core domain and a four-helix lid domain (αT1, αT2, αT3, and αT4) (Figure 5). The catalytic triad residues of S145, D266, and H295, distributed on the loops following the β4, β6, and β7 strands, are located at the bottom of the hydrophobic substrate-binding pocket. The rest of the core domain is made up of seven β strands (β1–β7) and six α helices (α1, α2, α3, α8, α9, and α10). R310 is located at the α10 helix of SsD14a.

To gain insights into the conformational state of SsD14a, we performed structural comparisons between SsD14a and other D14 orthologs from other plants. Structure comparisons revealed that the overall structure of SsD14a was identical to those from other plants in the open state (Figure 5B), with root-mean-square deviations (RMSD) ranging from 0.250 to 0.301 Å (Figure 5B). Notably, the overall structure of SsD14a in the open state was apparently larger than the closed state of AtD14-CLIM (covalently linked intermediate molecule, a hydrolysis intermediate of SL molecule), thus these two structures cannot be well aligned, with an RMSD of 0.662 Å. Furthermore, results of docking approaches demonstrated extensive binding of GR24 by residues in the catalytic pocket of SsD14a (Figures 5C,D). In general, the structural characteristics of SsD14a are highly conserved and guarantee its branching inhibition function.

To clarify the differences on SL transduction between the two SsD14 proteins, SsMAX2, another key SL signaling transduction component, was obtained and verified its function. SsMAX2 interacted with AtD14 in an SL-dependent manner with the intensity similar to AtMAX2 (Figure 2). We further investigated the physiological function of SsMAX2 proteins in Arabidopsis. We introduced full-length S. spontaneum MAX2 into the Arabidopsis max2-3 mutant under the control of a 35S promoter. As shown in Figure 6 and Supplementary Figure 2C, 35S:SsMAX2 max2-3 rescued the branching and leaf phenotype of max2-3 to a level comparable with the wild-type Col-0. These genetic data indicated that SsMAX2 could inhibit axillary branching of Arabidopsis. Our results demonstrated that SsMAX2 can resemble AtMAX2 to play a physiological role in Arabidopsis.

Further sequence comparison with AtD14 and OsD14 found that only SsD14b had a proline (P304) at the α10 helix, whereas, other D14 proteins contained an arginine (R) (Figure 1B). To further explore the mechanism underlying the differences in protein interactions, we made point mutations to SsD14a and SsD14b to obtain BD-SsD14aR310P and BD-SsD14bP304R, respectively. We were surprised to find that the point mutation SsD14aR310P no longer interacted with SsSMXL7 and AtSMXL7 (Figure 7A), but still interacted with SsMAX2. The point mutation SsD14bP304R turn out to obviously interact with AtSMXL7. Inferring from these results, for D14, residue R (like R310 of SsD14a) at the α10 helix might be the key residue contributing to the association with repressor factors SMXLs.

To further investigate the importance and the widespread of the amino acid site of R310 (in SsD14a), we performed point mutation validation in AtD14 and OsD14. We obtained BD-AtD14R262P and BD-OsD14ΔNR261P by site-directed mutagenesis PCR. Y2H results showed that AtD14R262P substitution largely affected the interaction with AtSMXL7 and also greatly weakened the interaction with AtMAX2 (Figure 7B). Similar observation was also found in BD-OsD14ΔNR261P (Figure 7C). Unlike AtD14, OsD14 showed hormone-dependent interaction with SsSMXL7. We speculate that this is a result of the higher sequence similarity between rice and sugarcane, which both belong to Gramineae. Interestingly, AtD14R262P and OsD14ΔNR261P, like wild-type proteins, still have strong hormone-dependent interactions with SsMAX2. The SsMAX2 protein can bind strongly with mutant proteins, which may have application in resolving the crystal structures of certain important D14 mutant proteins in complex with SsMAX2. In general, for AtD14 and OsD14, we further verified the importance of this site for binding downstream signal components.

In the SL signaling pathway, the D14 receptor senses SL before binding the F-box protein MAX2 to form the D14–MAX2 complex. Later, the complex would recruit and degrade the downstream repressor protein AtSMXLs through ubiquitination–proteasome pathway to regulate plant branching (Jiang et al., 2013; Zhou et al., 2013). To investigate whether SsD14bP304R has gained the capability in plant branching control, we generated and compared the transgenic Arabidopsis 35S:SsD14b 35S:SsMAX2 d14-1 max2-1 and 35S:SsD14b-P304R 35S:SsMAX2 d14-1 max2-1 by introducing full-length SsMAX2 together with SsD14b or SsD14b-P304R into the d14-1 max2-1 double mutant. We also generated the 35:AtD14 35S:AtMAX2 d14-1 max2-1 plants as positive control. We found that 35S:SsD14b-P304R 35S:SsMAX2 d14-1 max2-1 showed similar primary branching and leaf morphology as 35S:AtD14 35S:AtMAX2 d14-1 max2-1 (Figures 7D,E and Supplementary Figure 2D). However, the complex of SsD14b–SsMAX2 was unable to inhibit the branching of d14-1 max2-1 double mutant, consistent with the capability of SsD14b or SsD14b-P304R to bind AtSMXL7 (Figures 2, 7A). These results demonstrated that P304R single-residue substitution endows SsD14b with the branching inhibition function, indicating the close correlation between SL responses and receptor–repressor interaction.