Chromosome-scale pearl millet genomes reveal CLAMT1b as key determinant of strigolactone pattern and Striga susceptibility

Abstract

The yield of pearl millet, a resilient cereal crop crucial for African food security, is severely impacted by the root parasitic weed Striga hermonthica, which requires host-released hormones, called strigolactones (SLs), for seed germination. Herein, we identify four SLs present in the Striga-susceptible line SOSAT-C88-P10 (P10) but absent in the resistant 29Aw (Aw). We generate chromosome-scale genome assemblies, including four gapless chromosomes for each line. The Striga-resistant Aw lacks a 0.7 Mb genome segment containing two putative CARLACTONOIC ACID METHYLTRANSFERASE1 (CLAMT1) genes, which may contribute to SL biosynthesis. Functional assays show that P10CLAMT1b produces the SL-biosynthesis intermediate methyl carlactonoate (MeCLA) and that MeCLA is the precursor of P10-specific SLs. Screening a diverse pearl millet panel confirms the pivotal role of the CLAMT1 section for SL diversity and Striga susceptibility. Our results reveal a reason for Striga susceptibility in pearl millet and pave the way for generating resistant lines through marker-assisted breeding or direct genetic modification.

Fig. 1. Aw and P10 are contrasting lines for Striga susceptibility and strigolactone production.

a The wild accession, Aw, was shorter and had several tillers, whereas the domesticated P10 was a tiller monoculm. b Aw and P10 grown in Striga-infested soil. Aw exhibited no emerging Striga and was minimally affected in growth, while P10 was heavily infested by emerging Striga, severely impacting its growth and development. c Less than 10% of the Striga seeds exposed to Aw’s root exudate germinated, whereas exposure to P10’s root exudate stimulated germination in over 70% of the Striga seeds (n = 6 plates; P < 0.0001). d When grown in Striga-infested soil, P10 induced an average of 20 emerging Striga plants, while Aw induced none after 70 days (n = 6 plants; P = 0.0053). e The root exudate of both Aw and P10 contained orobanchol and orobanchyl acetate, with significantly lower levels in P10 (n = 3 seedlings; Oro, P = 0.0066; Oro ace, P = 0.0459). f Four new strigolactones were observed in P10’s root exudate but not detectable in Aw’s (n = 3 seedlings; except n = 5 for PL2 and PL3 Aw; P values for PL1 = 0.0397; PL2 = 0.0047; PL3 = 0.0038; PL4 = 0.0065). g Fractionation of SLs and subsequent assessment of individual fractions revealed that fractions G through J, containing PL and PL3, significantly induced Striga seed germination (n = 3 seedlings). Size bars indicate 20 cm. Error bars represent the mean ± s.d. Significant differences were tested using a two-tailed t test (*P  <  0.05, **P  <  0.01, ***P  <  0.001, ****P < 0.0001). Source data are provided as a Source Data file.

Fig. 2. Genome assembly for Aw and P10.

a Circos plot of the genomes of Aw and P10. Layers from the edge to the center are as follows: I. Location of predicted strigolactone biosynthesis genes, II. GC content (line) and gene density (bars), III. GC skew, IV. Chromosome names, and V. Synteny plot. The insets provide the total assembly size and number of high-confidence (HC) and low-confidence (LC) genes. The synteny plot revealed no large-scale rearrangements between P10 and Aw. b The 0.7 Mb section of chromosome 2, present in P10 but absent in Aw, extended from CLAMT1c to CLAMT1b and included CYP51 and an acyl transferase. The flanking regions exhibited strong synteny, as evidenced by three HC genes on each side of the region.

Fig. 3. P10CLAMT1b produced methyl carlactonoate, which is a precursor of pennilactone1.

a The expression of pearl millet homologs of known SL biosynthetic genes was generally upregulated in both Aw and P10 under low Pi conditions and even more so with the addition of the SL analog MP3. P10CLAMT1b was strongly co-expressed with PgD27, PgCCD8, and PgMAX1-1400, while P10CLAMT1c was induced more weakly. PgCLAMT1a was not significantly induced in Aw or P10 (n = 4 biological replicates of three plants each; differential expression analized using Fisher’s exact test and the p-values were adjusted for multiple testing using the Benjamini–Hochberg method; P values < 0.0003). b Recreating the SL biosynthetic pathway up to carlactonoic acid (CLA) through transient expression in N. benthamiana leaves produced no peak at the same retention time as the MeCLA standard. However, when P10CLAMTb was added to the experiment, a peak was observed, indicating that P10CLAMTb could convert CLA to MeCLA, as previously shown for OsCLAMT and AtCLAMT. c Total ion chromatograms of Aw roots fed with MeCLA produced PL1 (Extracted ion chromatogram: 359.14893; Retention time: 11.7), confirming that MeCLA is a precursor to PL1. The feeding led to two isomers with overlapping peaks, the second of which (blue arrow) corresponds to PL1. d The proposed SL biosynthesis pathway in pearl millet, where the boxed section is only present in P10 because of the presence of CLAMT1b. Consequently, only P10 produced downstream SLs, such as PL1, which can induce Striga seed germination when exuded from the roots, making P10 more susceptible to Striga infestation.

Fig. 4. Pearl millet cultivars with the CLAMT1b gene exude PLs and induced more Striga seed germination.

The root exudates from P10, Aw, and eight other sequenced pearl millet accessions were analyzed for the presence of SLs and their ability to induce Striga seed germination. Only P10, PI537069, and PI583800 contained CLAMT1b. The lines are arranged in descending order of Striga germination rates. Error bars represent the mean ± s.d.; n = 6 biological replicates (except for PI583800, n = 3; PI521612, n = 5). Significant differences regarding P10 were assessed using a two-tailed t-test: PI537069, P = 0.74 (ns); PI583800, P = 0.70 (ns); PI343841, P = 0.0010; PI521612, P = 0.0004; PI526529, P = 0.0039; PI587025, P = 0.022; PI592791, P = 0.0005; Aw, P < 0.0001; PI250656, P < 0.0001 (* P  <  0.05, ** P  <  0.01, *** P  <  0.001, **** P < 0.0001). Source data are provided as a Source Data file.