Gene-edited Mtsoc1 triple mutant Medicago plants do not flower

Abstract

Optimized flowering time is an important trait that ensures successful plant adaptation and crop productivity. SOC1-like genes encode MADS transcription factors, which are known to play important roles in flowering control in many plants. This includes the best-characterized eudicot model Arabidopsis thaliana (Arabidopsis), where SOC1 promotes flowering and functions as a floral integrator gene integrating signals from different flowering-time regulatory pathways. Medicago truncatula (Medicago) is a temperate reference legume with strong genomic and genetic resources used to study flowering pathways in legumes. Interestingly, despite responding to similar floral-inductive cues of extended cold (vernalization) followed by warm long days (VLD), such as in winter annual Arabidopsis, Medicago lacks FLC and CO which are key regulators of flowering in Arabidopsis. Unlike Arabidopsis with one SOC1 gene, multiple gene duplication events have given rise to three MtSOC1 paralogs within the Medicago genus in legumes: one Fabaceae group A SOC1 gene, MtSOC1a, and two tandemly repeated Fabaceae group B SOC1 genes, MtSOC1b and MtSOC1c. Previously, we showed that MtSOC1a has unique functions in floral promotion in Medicago. The Mtsoc1a Tnt1 retroelement insertion single mutant showed moderately delayed flowering in long- and short-day photoperiods, with and without prior vernalization, compared to the wild-type. In contrast, Mtsoc1b Tnt1 single mutants did not have altered flowering time or flower development, indicating that it was redundant in an otherwise wild-type background. Here, we describe the generation of Mtsoc1a Mtsoc1b Mtsoc1c triple mutant lines using CRISPR-Cas9 gene editing. We studied two independent triple mutant lines that segregated plants that did not flower and were bushy under floral inductive VLD. Genotyping indicated that these non-flowering plants were homozygous for the predicted strong mutant alleles of the three MtSOC1 genes. Gene expression analyses using RNA-seq and RT-qPCR indicated that these plants remained vegetative. Overall, the non-flowering triple mutants were dramatically different from the single Mtsoc1a mutant and the Arabidopsis soc1 mutant; implicating multiple MtSOC1 genes in critical overlapping roles in the transition to flowering in Medicago.

Keywords: CRISPR-Cas9; Medicago; MtSOC1a; MtSOC1b; MtSOC1c; flowering time; gene editing; legume.

Figure 1

The Mtsoc1a Mtsoc1b Mtsoc1c triple mutant line Mtsoc1-1 segregates plants that do not flower. (A, B) Distribution graphs in days (A) and node numbers on the primary axis (B) of the first flower of WT R108 (WT) and Mtsoc1-1 T1 segregating line (R545) under VLD. The sample sizes are shown above each bar (C) Box plots showing the number of days to first flower of the WT and Mtsoc1-1 T2 segregating lines (R760 and R761) under VLD. The homozygous (ho) and heterozygous (het) genotypes are shown. Statistical significance was determined using the Wilcoxon test (P-value with Bonferroni correction: ****P ≤0.0001). (D) Photographs of WT and non-flowering Mtsoc1-1 triple mutants (R656-6, right top and left bottom, and R545-2, right bottom) under VLD. The white stealth arrows indicate the apex of the primary axis. Scale is 2 cm.

Figure 2

The triple mutant line Mtsoc1-2 segregates non-flowering plants. (A, B) Box plots showing the number of days (A) and node number on the primary axis (B) to the first flower of the WT and Mtsoc1-2 T2 segregating line (R647) under VLD. Homozygous (ho) and heterozygous (het) genotypes are shown. Statistical significance was determined using the Wilcoxon test (P-value with Bonferroni correction: ****P ≤0.0001). (C) Photographs of the WT and non-flowering Mtsoc1-2 triple mutant (R647-26) taken on days 50, 87, and 147 under VLD. White stealth arrows indicate the apices of the primary axis. Scale is 2 cm.

Figure 3

Aerial architecture of plants segregating in triple mutant lines Mtsoc1-1 and Mtsoc1-2. (A–D) Aerial architecture of the WT and Mtsoc1-1 T2 lines (R760 and R761) under VLD, shown as boxplots of primary axis length for each plant (A), Boxplots of number on the primary axis for each plant (B), Boxplots of the longest secondary axis length for each plant (C), Boxplots of the node number on the longest secondary axis for each plant (D). Homozygous (ho) and heterozygous (het) genotypes are shown. (E) Left: Photographs of primary axes on day 84 under VLD of the WT and non-flowering Mtsoc1-1 triple mutant (R760-11), with leaves and branches removed. Right: Close-up view of the Mtsoc1-1 triple mutant primary axis on day 84 under VLD. (F) Photographs of the longest secondary axes on day 84 under VLD of the WT (left) and Mtsoc1-1 triple mutant (R760-11, right). (G–J) Aerial architecture of the WT and Mtsoc1-2 T2 segregating line (R647) under VLD, shown as a boxplot of the primary axis length for each plant (G), boxplot of the longest secondary axis length for each plant (H), boxplot of the node number on the primary axis for each plant (I), and boxplot of the node number on the longest secondary axis for each plant (J). (K) Left: Photographs of primary axes on day 87 under VLD of the WT and non-flowering Mtsoc1-2 triple mutant (R647-18), with leaves and branches removed. Right: Close-up view of the Mtsoc1-2 triple mutant primary axis on day 87 under VLD. (L) Photographs of the longest secondary axes on day 87 under VLD of the WT (left) and Mtsoc1-2 triple mutant (R647-18, right). Photographs on the top right are close-ups of the regions in dashed rectangles. Yellow stealth arrows indicate the lateral structures. Scale is 2 cm. Statistical significance was determined using a Wilcoxon test (P-value with Bonferroni correction: *P ≤0.05, **P ≤0.01, ***P ≤0.001, ****P ≤0.0001). ns, not significant.

Figure 4

Gene expression in the non-flowering Mtsoc1-2 triple mutants was consistent with their vegetative phenotypes. (A) Euler diagrams showing the number of upregulated and downregulated genes (two-fold change) in the shoot apex of the Mtsoc1-2 triple mutant (83 days old) and the Mtsoc1a Tnt1 single mutant (15 days old) compared to WT (15 days old) (P _adj <0.05). (B) Expression of Medicago homologs of Arabidopsis genes bound and regulated by AtSOC1 in ChIP-seq from Immink et al. (2012), shown as z-scores extracted from the TPM values. (C) Heat map of Mtsoc1-2 triple mutant (83 days old) and the Mtsoc1a Tnt1 single mutant (15 days old) and vegetative WT (15 days old) (left, heatmap log2(TPM) is shown), and mutants Mtfda (30 days old), Mtfta1 (30 days old, ft1, 63 days old, ft2), MtfdaMttfta1 (63 days old), and flowering WT (30 days old) from Cheng et al. (2021) (right part of the heatmap (log2(FKPM) is shown). The 93 genes shown are a list extracted from Cheng et al. (2021) of the selected candidate flowering genes. The DEG in Mtsoc1-2 genes were upregulated or downregulated with a fold change of 2 and padj ≤0.05. (D) Upper panel: RT-qPCR analysis of candidate flowering genes at the apex of the Mtsoc1-2 triple mutant and WT. Graph showing the relative expression of candidate flowering genes in the apex samples of the Mtsoc1-2 triple mutant (day 83) and the WT (day 14) under VLD. Relative gene expression was calculated using the formula 2^−ΔCT, where ΔCT was obtained by normalizing the gene of interest to the reference gene, MtPP2A, and presenting the log2 fold change in the mutant relative to WT. the box plot shows three biological replicates. Asterisks indicate significant differences between the mutant and WT strains using the t-test, assuming unequal variance (p ≤0.05). Lower panel: Expression shown as z-scores extracted from the TPM values from RNA-seq.