The effects of gibberellin on the expression of symbiosis-related genes in Paris-type arbuscular mycorrhizal symbiosis in Eustoma grandiflorum

ABSTRACT

Arbuscular mycorrhiza (AM) is a symbiotic interaction in terrestrial plants that is colonized by fungi in the Glomeromycotina. The morphological types of AM, including the Arum-type and Paris-type, are distinct, depending on the host plant species. A part of the regulatory pathways in Arum-type AM symbiosis has been revealed because most model plants form the Arum-type AM with a model AM fungus, Rhizophagus irregularis. Moreover, gibberellin (GA) is known to severely inhibit AM fungal colonization in Arum-type AM symbiosis. Recently, we showed that exogenous GA treatment significantly promoted AM fungal colonization in Paris-type AM symbiosis in Eustoma grandiflorum. In this study, we focused on the transcriptional changes in AM symbiosis-related genes in GA-treated E. grandiflorum. The expression levels of all examined E. grandiflorum genes were maintained or increased by GA treatment compared with those of the control treatment. Our new results suggest that signaling pathway(s) required for establishing AM symbiosis in E. grandiflorum may be distinct from the well-characterized pathway for that in model plants.

Most terrestrial plants associate with Glomeromycotina fungi to establish arbuscular mycorrhizal (AM) symbioses, and these plants supply carbohydrates to the symbiotic fungi in return for inorganic nutrients, especially inorganic phosphate.¹ The morphology of AM roots is divided into two types depending on the host plant species, namely Arum-type and Paris-type.^2,3 The Arum-type AM shows the intercellular elongation of AM fungal hyphae and a highly branched structure, arbuscule, in the cortical cells, whereas the Paris-type AM exhibits fungal hyphae that grow intracellularly and show a hyphal coil with emerging arbuscules in the cortical cells.^2,3 Many model plants, such as Lotus japonicus, Medicago truncatula, and Oryza sativa, form Arum-type AM.^2,4-7 In addition, the Arum-type AM is generally found in herbaceous plants in light areas.⁸ On another note, the Paris-type AM was often observed in trees and herbaceous plants such as Brachypodium distachyon and Acer species.^6,9 Moreover, mycoheterotrophic plants that obtained their carbohydrates from AM fungi are known to form the Paris-type AM as well.^3,9,10

Recently, the molecular mechanisms that regulate AM symbiosis have been revealed by using model plants that form Arum-type AM. Moreover, several studies showed that gibberellin (GA) negatively regulated AM fungal colonization in Arum-type AM symbioses in L. japonicus, M. truncatula, Pisum sativum, Solanum lycopersicum, and Oryza sativa roots, while certain level of GA was required for hyphal branching in the root cortex of L. japonicus.^5,7,11-13 Based on these findings, it has been generally thought that AM symbiosis is suppressed by GA. However, we found that the exogenous GA₃ treatment promoted AM fungal colonization in E. grandiflorum that forms Paris-type AM with the model AM fungus, Rhizophagus irregularis.¹⁴ It was also revealed that GA-treated E. grandiflorum roots exudated an unidentified branching factor, which resulted in increased fungal colonization in the root cortical cells.¹⁴ In this study, we focused on how exogenous GA affects the expression of AM symbiosis-related genes in E. grandiflorum.

RNA-sequencing (RNA-seq) data have been so far available only in E. grandiflorum cv. Bolero White, which is different from the Pink Thumb cultivar that was used in our previous study.^14,15 In this study, RNA-seq was conducted using E. grandiflorum cv. Pink Thumb to obtain its cDNA sequences. First, RNA-seq libraries were prepared from the total RNA extracted from 4-week-old E. grandiflorum seedlings and roots grown under axenic condition using an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Next, the libraries were sequenced using Illumina NovaSeq6000 by Macrogen Japan (Kyoto, Japan), and then de novo short read sequence assembly was conducted using Trinity v2.8.4 with default settings.¹⁶ In this study, E. gradiflorum roots inoculated with AM fungi were not used for the de novo assembly to avoid the contamination of R. irregularis reads in E. grandiflorum data. In addition, similar contigs were organized using CD-HIT (sequence identity threshold = 1) to reduce sequence redundancy.¹⁷ We conducted tblastx search with an E-value cutoff of 1 − e⁻⁵ using Medicago truncatula or Arabidopsis thaliana sequences as queries, and retrieved 6–23 candidate genes for each of target genes from the assembled data. To identify the E. grandiflorum orthologs of well-known genes relating to AM symbiosis in model plants, all sequences were aligned using ClustalW, and the phylogenetic trees were constructed through the maximum-likelihood method using MEGA7 with 1,000 bootstrap replications.¹⁸ This phylogenetic analysis allowed us to determine the E. grandiflorum orthologs among the candidate sequences (Supplementary Figure S1; Supplementary Data 1). Quantitative RT-PCR was conducted as previously described by Tominaga et al. (2020) to elucidate the effects of exogenous GA₃ and a GA biosynthesis inhibitor, Uniconazole-P (Uni), on Paris-type AM symbiosis in E. grandiflorum (Supplementary Table S1). For the qRT-PCR analysis, the growth condition of E. grandiflorum and extraction of total RNA were followed to our previous study.¹⁴ The colonization rates in E. grandiflorum used for the qRT-PCR analysis showed the positive effect of GA as we previously reported¹⁴ (Supplementary Figure S2A).

RAM1 (Required for Arbuscular Mycorrhization 1), RAD1 (Required for Arbuscule Development 1), PT4 (Phosphate Transporter 4), STR/STR2 (Stunted Arbuscule), Vpy (Vapyrin), and DLK2 (DWARF LIKE-2) are known to be transcriptionally upregulated in model plants colonized by AM fungi.^19,20 To decide the orthologous genes of DLK2, OsDLK2b was added as a query since expression of this gene has been reported to be relatively higher than that of its paralogous gene, OsDLK2a, in rice inoculated with R. irregularis.²⁰ Among these, RAM1 and RAD1 belong to GRAS (GIBBERELLIC-ACID INSENSITIVE, REPRESSOR OF GAI and SCARECROW) transcription factor family.^21–23 LjRAM1 expression is directly upregulated by a GA-degradable repressor, DELLA protein, in L. japonicus.^11,24-26 On another note, the MtRAD1 transcript is increased by MtRAM1 in M. truncatula.²³ Therefore, these two GRAS transcription factors may be directly/indirectly upregulated by DELLA protein. PT4 is a transporter of inorganic phosphate derived from symbionts, and STR/STR2 are predicted to transport lipids synthesized by host plants, respectively.^27,28 Vpy is thought to function in exocytosis for periarbusclar membrane formation and arbuscular development.^29,30 In addition, the expression of these genes in model plants may be regulated by the DELLA protein, RAM1, and RAD1, which means that the expression of these symbiosis-related genes is inhibited by GA.^11,22,23 On the contrary, EgRAD1, EgPT4, EgSTR, EgSTR2, and EgVpy expression levels were further increased by GA treatment compared with control E. grandiflorum colonized by R. irregularis, although the differences were statistically insignificant except for that of EgRAD1 (Figure 1a). Regarding to the transcriptional response of phosphate transporters, Hong et al. (2012) have already revealed that the expression levels of several genes encoding the transporters were significantly increased in B. distachyon forming Paris-type AM as well. The expression of EgPT4 was similarly responded to the AM fungal colonization, even though exogenous GA was added to the soil (Figure 1a). Furthermore, Uni application significantly reduced EgRAD1, EgPT4, EgSTR, and EgVpy expression levels compared with that of the GA-treated E. grandiflorum mycorrhizas (Figure 1a). On another note, EgRAM1 expression levels were not significantly altered by R. irregularis colonization, GA treatment, and Uni treatment (Figure 1a). Incidentally, while the expression levels of EgVpy and EgDLK2 were significantly reduced by Uni and GA treatment, respectively, the expression levels of rest examined genes were not affected in 4-week-old E. grandiflorum roots that were axenically cultured (Supplementary Figure S2B).

Figure 1.

Expression levels of symbiosis-related genes in gibberellin-treated Eustoma grandiflorum roots. Quantitative RT-PCR was conducted to reveal the regulatory mechanism of Paris-type AM symbiosis in E. grandiflorum colonized by Rhizophagus irregularis (at 4 weeks post-inoculation). In this analysis, E. grandiflorum seedlings were treated with 0.01% ethanol as control treatment (NC: non-colonized roots, AM: colonized roots), 1 μM GA₃ (GA–AM), and 1 μM Uniconazole-P (Uni–AM). R. irregularis spores were also inoculated in the soil at 1000 spores per soil container.¹⁴ (a) The expression levels of E. grandiflorum orthologues of RAM1, RAD1, PT4, STR/STR2, Vpy. The inset graphs are added due to the relatively low expression levels of EgRAM1, EgRAD1, EgSTR, EgSTR2, and EgVpy. (b) The same analysis as (A) was performed to investigate the expression level of DLK2 (DWARF LIKE-2) in the conditions. The expression levels of these genes were normalized by that of EgACT.¹⁴ Mean values ± SDs were calculated for each samples (n = 3). This experiment was conducted at 3 times, and the representative data is shown here. Different alphabets indicate significant differences using Tukey–Kramer test (P < .05) (Red alphabets: P < .1).

The expression level of EgDLK2 encoding a protein included in D14 family was also significantly increased by R. irregularis inoculation and GA treatment (Figure 1b). The D14 family consists of three subclades. The proteins in the D14 and D14-like (KAI2) clade, respectively, function as strigolactone (SL) and karrikin receptors, and DLK2 has been proven not to bind to natural SL, which implies DLK2 does not functions in SL signaling^31,32 (Supplementary Figure S1E). Of these proteins, it was reported that D14-like is required for AM fungal colonization in rice.^33,34 In addition, the proteins that belong to DLK2 clade are transcriptionally upregulated by AM fungal inoculation in rice, although their functions in symbiotic associations are still unknown.²⁰ Intriguingly, EgDLK2 expression level was significantly increased upon R. irregularis inoculation and further enhanced by GA treatment (Figure 1b). In contrast, Uni treatment significantly reduced its expression level compared with control and GA-treated mycorrhizal roots (Figure 1b). This result suggests that DLK2 also functions in the GA-promoted AM symbiosis in E. grandiflorum, although further studies are necessary to investigate the detailed functions of DLK2 in AM symbioses in E. grandiflorum and model plants. However, the treatments of GA and Uni would associate with other phytohormonal pathways, which may also affect the AM symbiosis in E. grandiflorum.^35–38

In this study, we revealed that the effects of GA on the signaling pathways regulating AM symbiosis in E. grandiflorum and model plants would be different from each other. While EgRAM1 expression level was not affected by GA treatment, that of the other GRAS transcription factor, EgRAD1 was significantly promoted in the presence of GA (Figure 1a). In addition, the exogenous GA biosynthesis inhibitor Uni treatment did not affect EgRAM1 expression (Figure 1a), whereas GA biosynthesis inhibition and DELLA stabilization enhanced RAM1 expression in L. japonicus.¹¹ These results indicate that the EgRAM1 and EgRAD1 expression would be regulated by protein(s) other than DELLA, although it appears that E. grandiflorum possesses two genes encoding DELLA proteins (Figure 2; Supplementary Figure S1A). This unique system regulating the GRAS transcription factors in E. grandiflorum may enable the host plant to establish AM symbiosis in the presence of GA (Figure 2). In Arum-type AM symbiosis, several additional transcription factors may also function in the activation of downstream genes, because the symbiosis-related genes, such as PT4, are significantly expressed even in the colonized roots of the L. japonicus ram1 mutant.¹¹ Moreover, EgRAM1 expression was not affected even by AM fungal colonization, which indicates that EgRAM1 may not be required for the establishment of AM symbiosis in E. grandiflorum (Figure 1a). Therefore, the involvement of DELLA and GRAS transcription factors in downstream gene expression in E. grandiflorum should be investigated to further elucidate the regulatory mechanisms of the GA-promoted AM symbiosis. Furthermore, DELLA also negatively regulates AM symbiosis; it functions in the arbuscule degradation in concert with a transcription factor, MYB1.³⁹ This raises a question whether the negative role of DELLA is also conserved in the GA-resistant AM symbiosis of E. grandiflorum, and investigating it would be interesting. Incidentally, E. grandiflorum orthologues of AM symbiosis-related genes in model plants were expressed only at low levels even upon R. irregularis inoculation, although they appeared to be slightly upregulated (Figure 1a). This result may be attributed to the relatively low colonization rate of R. irregularis in E. grandiflorum roots compared with other plants, as shown in our previous study.¹⁴

Figure 2.

Proposed signaling pathway(s) underlying Paris-type AM symbiosis in Eustoma grandiflorum. In Arum-type AM symbiosis in model plants, the presence of bioactive gibberellins (GAs) triggers the degradation of DELLA protein, which suppresses the expression of AM symbiosis-related genes. In contrast, the exogenous GA treatment may somehow promote EgRAD1 expression (Fig. 1A), which supposes the existence of other protein X. If so, the protein would directly/indirectly regulate the expression of EgRAD1, which may result in the promotion of AM symbiosis-related genes’ expression (see the text for details). In addition, crosstalk between GA and other phytohormones might be included in the establishment of AM symbiosis in E. grandiflorum. The dotted lines indicate possible pathways regulating AM symbiosis in E. grandiflorum.

The opposite effects of GA on AM symbioses in model plants and E. grandiflorum raise the question of why GA-mediated regulatory mechanisms of them are different from each other. Most of mycoheterotrophic plants associated with AM fungi form Paris-type AM; however, this type of AM can be observed in autotrophic plants as well.^2,6,10,14 Thus, the physiological function of Paris-type AM may be related to the occurrence of mycoheterotrophic plants. Intriguingly, it has been recently reported that autotrophic host plants that showed Paris-type AM, including Anemone nemorosa and Paris quadrifolia, tended to accumulate ¹³ C that originated from associating AM fungi.⁴⁰ Furthermore, a Gentian species, Ptrerygocalyx volubilis, has been also indicated to accumulate ¹³ C and ¹⁵ N isotopes.⁴¹ This implies that even autotrophic plants forming Paris-type AM behave more like parasites toward the symbionts. However, more convincing evidence showing the dependencies of hosts including trees such as Acer species on AM fungi should be necessary since these studies mainly focus on the isotopic fluxes in herbaceous species. In addition, this study revealed that Paris-type AM symbiosis in E. grandiflorum would be differentially regulated compared with Arum-type AM symbiosis in model plants. Moreover, the Gentianaceae family, to which E. grandiflorum belongs, includes several mycoheterotrophic species.^10,42 Therefore, the signaling pathway to establish Paris-type AM symbiosis in the Gentianaceae family may have uniquely evolved to obtain AM fungal carbohydrates, which resulted in the occurrence of GA-resistant signaling pathway(s) regulating the symbiosis. In this study, we could not clarify why the regulatory mechanism of AM symbiosis is different depending on the morphological types of AM. The development of Paris-type AM is relatively slow due to the repeated hyphal penetration into the cortical cells; hence, forming Paris-type AM could be reasonable for host plants to serve their photosynthates in dark areas.^4,43 In a shaded area like the forest floor, plants activate a shade avoidance response by elongating their stem and suppressing leaf growth, which involves homeostatic/transcriptional changes in the biosynthesis/metabolism of several phytohormones, such as auxin, cytokinin, brassinosteroid, and GA.^44–46 In addition, the biosynthesis and accumulation of bioactive GAs have been reported to be promoted by shade treatment.^47,48 Therefore, it could be hypothesized that host plants have evolved to establish Paris-type AM symbiosis so that they can obtain more nutrients and struggle for existence under conditions where GA biosynthesis is promoted. To support this hypothesis, it is necessary to reveal how shade avoidance affects the regulation of AM symbiosis by testing more Paris-type AM-forming species including mycoheterotrophic ones.

In conclusion, we revealed that the expression levels of E. grandiflorum orthologues of AM symbiosis-related genes in model plants were stable or rather promoted (significant upregulation in EgRAD1 and EgDLK2, but slight promotion in EgPT4 and EgSTR) in GA-treated E. grandiflorum roots colonized by R. irregularis (Figure 1).¹⁴ The elucidation of the signaling pathway regulating the expression of these genes would shed light on understanding AM symbiosis. In addition, the opposite effect GA on Arum- and Paris-type AM symbioses may be related to the adaptation of the Paris-type AM to specific environments, such as shaded areas, which may explain the reason why mycoheterotrophic plants generally form Paris-type AM. However, the effects of GA on Paris-type AM symbiosis in more host plants should be investigated to support this hypothesis since this study used only one species, E. grandiflorum. Furthermore, investigating the involvement of the crosstalk between GA and other phytohormones in the regulation of AM symbiosis in E. grandiflorum would contribute to the comprehensive understanding of the GA-resistant signaling pathway(s) underlying the symbiosis.

Funding Statement

This work was supported by the NIBB Cooperative Research Programs (Next-generation DNA Sequencing Initiative: 19-433).

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Data availability

The nucleotide sequence data obtained by RNA-seq in this study have been deposited into the DDBJ Sequence Read Archive under the accession number DRA010085.

Supplemental materials

Supplemental data for this article can be accessed on the publisher’s website.