Mechanism of Early Flowering in a Landrace Naked Barley eam8.l Mutant

Abstract

The EARLY MATURITY 8 (EAM8) gene of barley is homologous to the EARLY FLOWERING 3 (ELF3) gene in Arabidopsis, as loss-of-function mutations in this circadian clock gene promote rapid flowering. A previous study demonstrated that the early flowering phenotype of a hulless barley, Lalu, was due to allele eam8.l carrying an alternative splicing mutation in intron 3 that led to intron retention. In the present study, we verified that eam8.l encoded a truncated protein. Although EAM8 was expressed at a higher level in Lalu than in other barley lines with a longer growth period, it did not negatively regulate flowering time. This result further proved that the eam8.l protein was nonfunctional in regulating flowering in barley. The early flowering phenotype of Lalu plants was strongly dependent on the biosynthesis of gibberellin (GA). The eam8.l mutation stopped the suppression of GA biosynthesis, and Lalu accumulated excessive GA, especially in leaves. This was achieved through the upregulated expression of genes in the GA pathway, including GA20ox2, LFY1, SOC1, PAP2, and FPF3. The mutation of the EAM8 gene also abolished the inhibition of FLOWERING LOCUS T-like (FT1) gene expression at night. During the night, expression levels of the FT1 gene were higher than those during the day in Lalu. However, the GA-dependent pathway and FT1 gene mechanism are two independent pathways that promote flowering in Lalu. Alleles of EAM8, therefore, demonstrated an important breeding value in barley, which is probably effective in many other day-length-sensitive crop plants as well; thus, they could be used to tune adaptation in different geographic regions and climatic conditions, a critical issue in times of global warming.

1. Introduction

Plant flowering is a complex regulatory process of development that is influenced by a number of internal and external factors. Under the influence of internal and external factors, plants initiate related signal transduction processes and promote the control of plant gene expression, thereby determining plant flowering. Based on an in-depth study of the model plant Arabidopsis thaliana, at least four types of pathways that control plant flowering have been discovered, including photoperiod, vernalization, autonomic promotion, and various hormonal pathways in response to specific exogenous signals, which synergistically control the flowering process of plants [1,2]. These pathways construct a very complicated and large regulatory network involving a large number of genes. For example, FLOWERING LOCUS T (FT) encodes a protein with a phosphatidylethanolamine-binding protein domain, and the FT protein interacts with the leucine zipper transcription factor FLOWERING LOCUS D, activating the floral meristem gene APETALA 1 and enhancing the expression of LEAFY (LFY), thereby promoting flowering [3]. During crop domestication, FLOWERING LOCUS T-like genes (FT1 in barley) were a key determinant of the flowering response. The reduced expression of FT1 always leads to delayed flowering. EARLY FLOWERING 3 (ELF3) is usually regarded as a zeitnehmer (time taker) and an essential component of the plant circadian clock. It plays a crucial role as a negative regulator of flowering time in A. thaliana [4]. Mutants that lose the function of this gene have attenuated photoperiod sensitivity and flower much earlier under both short-day (SD) and long-day (LD) conditions [5]. The survival of such mutations, occupying broad geographical ranges, indicates that early flowering is advantageous for crops in any environment where the growing season is short [5,6,7]. Breeders have identified the pattern in crops by reducing photoperiod sensitivity to assist with migration to high-latitude regions where SDs impede floral induction [5,7,8].

In barley, the EARLY MATURITY 8 (EAM8) gene, also known as Mat-a (Praematurum-a), is responsible for the generation of circadian rhythms and the regulation of photoperiodic flowering. It is located on the long arm of barley chromosome 1H and is homologous to the ELF3 gene in Arabidopsis [5]. The loss of EAM8 gene function can induce photoperiod insensitivity in barley, increasing the expression of the FT1 gene and promoting rapid heading and flowering of barley (Hordeum vulgare) under LD and SD conditions. Zakhrabekova et al. found that among 87 induced eam8 mutants, 6 carried a splice-site mutation [5]. Among them, there were only two spontaneous mutations in EAM8: eam8.k (found in cvs. Kinai 5 and Kagoshima Gold) and eam8.w (cv. Early Russian) [5]. The heading of eam8_ea8.k and eam8_ert.016 with a Bowman or an Igri genetic background occurred earlier than that of the parental cultivars under both SD (25 days earlier) and LD (~2 weeks earlier) conditions [9]. Studies revealed that the early flowering phenotype in mat.a-8 mutants was explained by increased production of the hormone gibberellin (GA), and that GA played an important role in the transition from vegetative to reproductive growth in barley, cooperating with the FT1 gene [9,10]. A third spontaneous EAM8 mutation (eam8.l) was identified in a six-rowed hulless barley landrace Lalu adapted to the Qinghai–Tibetan Plateau [11]. The eam8.l line flowered approximately 21 days and 15 days ahead of Diqing1, a normal six-rowed hulless cultivated barley line, under LD and SD conditions, respectively. eam8.l carries an A/G alternative splicing mutation at position 3257 in intron 3, which results in intron retention and a putative truncated protein [11].

In the present study, the early flowering mechanism of Lalu plants (eam8.l) was further investigated. The objectives of this study were as follows: (1) to verify the existence of the truncated protein; (2) to show the expression pattern of eam8.l; (3) to confirm whether the early flowering phenotype of Lalu was caused by GA synthesis regulation; and (4) to show how GA synthesis regulates early flowering. Here, the hypothesis was verified that the precocious phenotype was dependent on an increase in GA synthesis. In eam8.l, the results showed a significant increase in the expression of genes related to GA synthesis, especially GA20oxidase 2 (GA20ox2). Moreover, FT1 also played an important role in promoting Lalu flowering, but its expression was not affected by GA. Thus, GA and FT1 acted via two independent pathways to promote the early flowering of Lalu plants. This study provides a theoretical basis for future barley precocity breeding under SD conditions.

2. Materials and Methods

2.1. Plant Lines

The plant lines comprised landrace barley Lalu (eam8.l mutant, with a G to A mutation at the 3257 bp site of EAM8 leading to early flowering), the cultivar Diqing1 (wild type, flowering later than Lalu), the cultivar Bonus (wild type, a general reference line of barley, kindly provided by the Qinghai Academy of Agricultural Sciences), and Tibetan wild barley 86587 (wild type, different to the 3257 bp site of EAM8; the remaining sequences of the gene were the same as Lalu; flowering later than Lalu). All plants were grown in gallon pots (3L) filled with general nutritional soil in standard growth rooms at 25 °C with an air humidity of 53% and all plants were under the same SD photoperiod (8 h light/16 h dark).

Germination assays, chlorophyll extraction and measurements, GA and PAC (paclobutrazol, used to inhibit GA synthesis) treatments, apex dissection, developmental flowering time measurements, and GA extraction and quantification were performed as described by Boden et al. [10].

2.2. Western Blotting

The antibodies were prepared by QiWei YiCheng Tech Co., Ltd. (Beijing, China). The polypeptide was synthesized according to the partial exon 1 amino acid sequence of the EAM8 gene: MRRAGGGGGGGSGEDKVMGPLFPRLHVNDTTLKGGGPRAPPRNKMALYEQFSVPSQRFAANTAPAAHRPAASFAAVSSASAGQIGGIDRPLF. After compounding of the polypeptide, a rabbit anti-EAM8 immune serum was prepared as the primary antibody, and anti-rabbit IgG H+L (HRP) was used as the second antibody.

The protein of plant lines was extracted using a total protein extraction kit (Solarbio, Beijing, China). After polyacrylamide gel electrophoresis, the gel area with the targeted band was selected for transmembrane use. Then, the following procedures were performed: blocking, primary antibody incubation, and secondary antibody incubation. Finally, the membrane containing the protein compound was exposed in a dark room.

2.3. RNA Extraction and Expression Analysis

RNA was extracted from the barley lines as follows: (1) The sampling time for the lines used for spatial expression was 22:00. (2) The sampling time for the lines (leaves and apices) treated with GA, PAC, PAC+GA, and the control was also 22:00 at the four-leaf stage. (3) Leaves from Lalu under PAC-treatment (1 μM) conditions and leaves from Lalu, Diqing1, 86587, and Bonus without any treatment were harvested at defined intervals until the four-leaf stage. Each leaf was the youngest tissue that emerged from the fourth leaf at the four-leaf stage. Each apex RNA sample contained eight pooled apices and was harvested from the main stem. All barley lines used in the experiment contained three biological replicates. Total RNA extraction kits (TaKaRa MiniBEST Plant RNA Extraction Kit), reverse transcription kits (PrimeScript™ RT Reagent Kit with gDNA Eraser), and kits for real-time PCR (RT-qPCR) (TB Green™ Premix Ex Taq™ II) were from TaKaRa Bio Inc. Details of the nucleotide primer sequences used in this study are shown in Supplementary Table S1. The internal reference gene used in the quantitative experiment was α-tubulin.

2.4. Statistical Analysis

Student’s t-test of the significance of the differences between treatment groups was performed with GraphPad Prism software using a 95% confidence level (p < 0.05). The results are shown in the figures as mean ± SE.

3. Results

3.1. eam8.l Encodes a Truncated Protein

Previous studies have demonstrated that the G/A mutation at 3257 bp in the intron of eam8.l in Lalu results in a translation frame shift and leads to truncated proteins [11]. A bioinformatics analysis predicted that the theoretical molecular weight of the eam8.l protein in Lalu would be 46.40 kDa, whereas that of the EAM8 protein in Dingqing1 would be 84.32 kDa (Figure 1A). To verify this conjecture, Western blotting was performed to determine the total proteins of new leaves of Lalu and Diqing1 at the four-leaf stage (Figure 1B). The results showed that the molecular weight of the EAM8 protein in Lalu was apparently lower than that of the corresponding protein in Diqing1, indicating that the protein encoded by eam8.l in Lalu was indeed a truncated incomplete sequence (Figure 1). This deficiency causes a change in protein function.

3.2. Spatial Expression of the EAM8 Gene at Different Developmental Stages

Owing to the mutation of gene EAM8, the flowering time of Lalu is significantly earlier than that of other lines. To further investigate the mechanism of the earlier flowering phenotype, the expression levels of the EAM8 gene were tested in different organs of Lalu, Diqing1, 86587, and Bonus at different stages. Under SD conditions, none of the lines other than Lalu could be headed. Therefore, the expression test was carried out at three periods: the four-leaf stage, tillering stage, and jointing stage.

At the four-leaf stage, the expression of the EAM8 gene in the Lalu apex was significantly higher than that in Diqing1 and 86587 (Figure 2A), but in leaves and roots, there was no significant difference among Lalu, Diqing1, and 86587 in the expression level of EAM8 at this stage (Figure 2B,D). At the jointing stage, the expression levels of the EAM8 gene in the apex and leaves of Lalu were significantly higher than those in Diqing1 and 86587 (Figure 2E,F), whereas the difference in expression in stems and roots was not significant (Figure 2G,H). In the tillering stage, the transcript levels for EAM8 were significantly higher in Lalu than in Diqing1 and 86587 except for in the root; especially in the apex, the difference was extremely significant (Figure 2I–L).

From an organ perspective, in the apex the expression of EAM8 in Lalu was significantly higher compared with that in Diqing1 and 86587 at all three stages (Figure 2A,E,I). In contrast, no significant difference was detected among the three lines in the three stages in the roots (Figure 2D,H,L). Given that EAM8 is regarded as a negative regulator of flowering time [11], the higher its expression, the later the flowering should be. In the present study, although the transcript levels in various organs during the growth process were higher in Lalu compared with those in other barley lines, Lalu flowered much earlier. This result further indicated that the EAM8 protein in Lalu was nonfunctional and promoted flowering.

3.3. The Early Flowering Phenotype in Lalu Depends on Increasing GA Production

To further analyze the mechanism of the early flowering of Lalu caused by eam8.l, four experimental lines (Lalu, Diqing1, 86857, and Bonus) were planted under SD conditions to compare the flowering phenotype. At early vegetative growth, it was observed that Lalu was superior to Diqing1, 86587, and Bonus; for instance, it had longer coleoptiles and leaf length, but a lighter leaf color than the other three lines. Qualitatively measuring these phenotypes showed that Lalu had advantages over Diqing1, 86587, and Bonus in terms of germination rate, leaf growth, and coleoptile development (Figure 3A–C), whereas the chlorophyll content of Lalu was significantly lower than that of the other three lines (Figure 3D).

The trend of these phenotypic changes was similar to the response to increasing GA [12,13,14]. GA can promote plant development in many ways; thus, it was suspected that the GA-dependent pathway might also be involved in the early flowering pathway of Lalu plants. To this end, we analyzed the flowering phenotypes of Lalu, Diqing1, 86587, and Bonus under the GA3 and GA biosynthesis inhibitor PAC. Anatomic photographs of the developmental inflorescence of Lalu, Diqing1, 86587, and Bonus were taken under treatment and control conditions. PAC strongly inhibited the development of the inflorescence of the four accessions, and Diqing1, 86587, and Bonus did not flower during the course of the experiment (Figure 4A). In addition, the higher the PAC concentration, the stronger the inhibition (Figure S1A). This inhibition was significantly relieved by the application of GA3 (Figure 4A; Figure S1A). Additionally, increasing GA content markedly promoted flowering in Lalu but did not lead to the complete flowering of Diqing1, 86587, and Bonus (Figure S1). Therefore, GA-dependent pathways have a major impact on the early flowering of Lalu.

Four lines were treated with the GA biosynthesis inhibitor PAC at different concentrations to verify whether Lalu synthesizes higher levels of GA or is more sensitive to GA. The leaf growth rates of all four lines slowed down and reached an equal level as the concentration of PAC increased (Figure 4B). Significantly, the inhibition of Lalu was more remarkable. As Lalu was more sensitive to the increase in PAC concentration, the assumption that Lalu had a constitutive GA response was disproved. Four lines were treated with different concentration gradients of GA3 to verify whether Lalu displayed a similar trend in GA sensitivity. Diqing1, 86587, and Bonus plants showed similar responses to exogenous GA, and their leaves grew as fast as those of Lalu plants under all GA3 concentrations. Lalu was not as sensitive to exogenous GA as the other lines (Figure 4C). Taken together, these results indicate that the growth of Lalu plants is highly dependent on endogenous GA biosynthesis.

3.4. The GA-Dependent Pathway and FT1 Gene Represent Two Independent Pathways That Promote Flowering in Lalu

The FT1 gene is located at the junction of various flowering pathways in plants and has a key role in the regulation of plant flowering. Under appropriate light conditions, the FT1 gene is exposed to vascular-bundle-specific zinc finger proteins in the photoperiod pathway. The FT1 protein synthesized in plant leaves is transferred to the shoot apex meristem to regulate downstream gene expression and promote flowering [10]. Therefore, we presumed that the PAC treatment would affect the expression of the FT1 gene in Lalu plants. The FT1 transcript levels from leaves of PAC-treated and control Lalu plants at the four-leaf stage were compared. No transcripts of FT1 were detected in Diqing1 or 86587 leaves under PAC-treatment and control conditions. FT1 gene expression was not affected significantly by the PAC treatment in Lalu (Figure 5). Combined with the previous results, this demonstrated that flowering could be delayed by reducing GA biosynthesis in Lalu plants, but this reduction did not affect the expression of FT1.

To further explore the mechanism of early flowering in Lalu plants, we measured the expression of flowering-related genes downstream of the GA pathway, including LEAFY (LFY1), SUPPRESSOR OF CONSTANS 1 (SOC1), FLORAL PROMOTING FACTOR 3 (FPF3), and PANICLE PHYTOMER 2 (PAP2). Likewise, the expression of MADS8 and MADS14, which are downstream genes of FT1, was also measured. The transcription of these genes was compared in the developing spikes of Lalu, Diqing1, and 86587. The expression levels of the LFY1, SOC1, FPF3 and PAP2 genes in Lalu were significantly higher than those in Diqing1 and 86587 (Figure S2A). Furthermore, we compared the expression of these genes in developing spikes from control Lalu plants and plants that had been treated with PAC and PAC/GA to determine whether their expression was GA dependent. When PAC was applied to Lalu plants, the expression levels of these four genes in the GA pathway decreased significantly, and the suppressive effect was eliminated by the application of GA (Figure S2A). The expression levels of MADS8 and MADS14 genes in Lalu plants were also significantly higher than those in Diqing1 and 86587, but there was no significant change resulting from the treatment with PAC or PAC/GA (Figure S2B). These results indicate that GA promotes the early flowering phenotype in Lalu plants by increasing the expression of genes including LFY1, SOC1, FPF3, and PAP2.

3.5. Mutation of the EAM8 Gene (eam8.l) Suppresses GA Synthesis in Lalu

The abovementioned experimental results demonstrated that Lalu plants exhibited a strong response to GA biosynthesis inhibitors (Figure 4C). It was presumed that the mutation of EAM8 made Lalu synthesize more GA than Diqing1 and 86587. First, the amount of GA molecules in leaves was measured, including bioactive GA1, precursor molecules GA19 and GA20, and the inactivate product GA8, as direct evidence to support this assumption. It was found that the four small GA molecules in Lalu plants were all higher than those in Bonus, 86587 and Diqing1 plants (Figure S3, Table S1). Moreover, the GA1, GA19, and GA8 contents in Lalu plants were significantly higher than those in Diqing1 plants (Figure S3, Table S1). The results indicated that the leaves of Lalu contain excessive GA production. Previous studies have shown that the production of GA19 and GA20 needs to be catalyzed by a GA20ox gene, GA20 is used to form GA1, and GA8 is produced after catabolism of GA1 [15,16]. Therefore, the expression of GA20ox genes is very important in the GA synthesis process and has been measured. The expression pattern over the day–night cycle of the GA20ox2 gene in leaves showed that the expression of the GA20ox2 gene in Lalu plants was significantly higher than that of Diqing1 and 86587 during the whole day, and this relatively high expression trend was much more evident at night (Figure 6A). We also found that the expression levels of genes GA20ox1, GA20ox3, and GA20ox4 in the leaves of Lalu plants were higher than those in Diqing1 and 86587 during the night at ZT 10 h (Figure 6B). In contrast to the GA20ox2 gene, the EAM8 gene showed significantly higher expression during the daytime than at night (Figure 6A,C). Therefore, although the expression of the EAM8 gene decreased, the production of GA showed an increasing trend. Moreover, when GA3 (10⁻⁴ M) was applied to either Lalu, Diqing1, or 86587, the expression levels of the GA20ox2 and GA20ox3 genes were significantly lower than that in the control (Figure S4A). Taken together, these results demonstrate that the trend of GA synthesis gene expression was positively correlated with GA abundance.

We also analyzed the expression levels of GA20ox in the apex of Lalu, Diqing1, and 86587. GA20ox1 was not expressed in the apex, and the expression levels of GA20ox3 and GA20ox4 did not differ significantly between the treatment and control groups (Figure S4B). However, under PAC-treated conditions, the expression level of the GA20ox2 gene in the apex of Lalu plants was significantly higher than that in the Lalu control group, and the expression decreased when GA3 was applied (Figure S4C). Taken together, these results indicate that GA does not accumulate excessively in the apex of Lalu plants with the eam8.l mutation.

The transcripts of other GA synthesis-related genes, including GA3ox1, GA3ox2, GA2ox3, and GA2ox4, could be detected in leaves. The quantitative RT-qPCR showed that during the daytime, GA3ox1 transcripts were significantly higher in Lalu than in Diqing1 and 86587, although the transcript level was very weak compared with GA3ox2 and GA2ox3. During the nighttime, the GA2ox3 transcript level of Lalu was much higher than that of the other two lines, and there was no significant difference in GA3ox1, GA3ox2, or GA2ox4 transcript levels among the three lines (Figure 6D). The GA2ox3 transcription pattern was consistent with the feedback mechanism of the GA biosynthesis pathway. In the apex of Lalu, Diqing1, and 86587, we only detected the GA3ox2 and GA2ox3 genes. The expression level of the GA3ox2 gene in Diqing1 and 86587 plants was significantly higher than that in Lalu plants. The differences in the expression levels of the GA2ox3 gene between Lalu and the other two lines (Diqing1 and 86587) were not significant (Figure S4D). Furthermore, the expression levels of these genes in Lalu plants responded to PAC and PAC/GA treatment to different degrees. These results further indicated that the increase in GA production occurred in the leaves of Lalu plants, but not in apices.

4. Discussion

Plants that flower rapidly can be grown in marginal environments where the growing season is restricted by the climate. Photoperiod insensitivity is also beneficial in regions with a long growing season, as it extends the length of the vegetative phase, with positive effects on reproductive potential [5,17].

The EAM8 gene in barley is a homolog of ELF3 that is involved in regulating the sensitivity of the barley photoperiod. Loss of function of the EAM8 gene causes photoperiod insensitivity and precocity in barley under both LD and SD conditions [5,9,10]. Mutations in EAM8 have aroused the interest of barley breeders. Lalu is a six-rowed hulless landrace barley with an early flowering phenotype that is natively grown on the Tibetan Plateau. Lalu was demonstrated to be a spontaneous recessive EAM8 mutant (eam8.l) by members of our laboratory, and the mutation was shown to lead to intron retention and a putative truncated protein [11]. In the present study, both a bioinformatics analysis and Western blotting verified that the EAM8 protein in Lalu was indeed a truncated protein. EAM8, a negative regulator of flowering, was expressed at higher levels in Lalu than in other barley lines that flowered later than Lalu. These results provided the evidence that the eam8.l protein was nonfunctional. The loss of function was responsible for the early flowering of Lalu.

The present study aimed to determine how eam8.l regulates the early flowering of Lalu. Previous studies have shown that GA, a phytohormone, is not the only factor regulating flower formation, but works in concert with other factors under the conditions of an induced photoperiod [18,19,20]. Our results showed that the eam8.l gene can regulate the expression of GA biosynthesis-related genes in Lalu plants, especially GA20ox2, which is also confirmed as a factor in the early flowering phenotype of Lalu plants. Moreover, the role of the EAM8 gene in Lalu plants was consistent with the increased expression of GA20ox2 in the circadian clock mutant of A. thaliana [21]. In Lalu, high levels of GA20ox2 expression and GA accumulation were found mainly in leaves rather than apices; however, genes associated with flowering in the GA pathway, including LFY1, SOC1, FPF3, and PAP2, were highly expressed in apices. Therefore, the GA high-content signal may be transmitted from leaves to apices, followed by the expression of downstream genes in the GA pathway.

In barley, FT1 is the main FT-like gene involved in the transition from vegetative to reproductive development and is highly expressed under LD photoperiods relative to SDs [9,22]. The circadian cycle expression patterns of the FT1 and GA20ox2 genes showed that the mutation of the EAM8 gene makes early flowering inseparable from these two genes in Lalu plants. In Lalu plants, the expression of GA20ox2 was significantly higher than that of Diqing1 and 86587 in the diurnal expression pattern, which was consistent with ELF3 being a key inhibitor in the evening loop of the circadian cycle [23,24]. Interestingly, the development of Lalu was obviously inhibited by PAC, but was not sensitive to exogenous GA3. Moreover, the FT1 expression level in Lalu was higher than that in Diqing1 and 86587; however, it was not affected by the PAC treatment. Therefore, increasing GA production and FT1 expression are two independent means of accelerating the flowering process in Lalu (Figure S5). This pattern is consistent with the early flowering mechanism of the elf3 mutant mat.a-8 [10]. The mechanism was further confirmed by the relationship between GA and FT1 found in Lolium, but not in Arabidopsis [25,26,27].

5. Conclusions

In conclusion, eam8.l indeed encoded a truncated protein. The loss of function relieved the suppression of GA synthesis by EAM8, and more GA was produced in Lalu than in wild types. The early flowering phenotype in Lalu was responsible for the excessive production of GA and the high expression of the FT1 gene. These two pathways were independent of each other in promoting the flowering process in Lalu. This study clearly elucidates the early flowering mechanism in Lalu and could provide help for barley breeding programs. Lalu could be used as a parental line by breeders to modify flowering time and create new cultivars adapted to broader geographical ranges.