Abstract
Arbuscular mycorrhizal (AM) symbiosis is a mutually beneficial association of plants and fungi of the sub-phylum Glomeromycotina. The endosymbiotic AM fungi colonize the inner cortical cells of the roots, where they form branched hyphae called arbuscules that function in nutrient exchange with the plant. To support arbuscule development and subsequently bidirectional nutrient exchange, the root cortical cells undergo substantial transcriptional re-programming. REDUCED ARBUSCULAR MYCORRHIZA 1 (RAM1), studied in several dicot plant species, is a major regulator of this cortical cell transcriptional program. Here, we generated ram1 mutants and RAM1 overexpressors in a monocot, Brachypodium distachyon. The AM phenotypes of two ram1 lines revealed that RAM1 is only partly required to enable arbuscule development in B. distachyon. Transgenic lines constitutively overexpressing BdRAM1 showed constitutive expression of AM-inducible genes even in the shoots. Following inoculation with AM fungi, BdRAM1-overexpressing roots showed higher arbuscule densities relative to controls, indicating the potential to manipulate the relative proportion of symbiotic interfaces via modulation of RAM1. However, the overexpressors also show altered expression of hormone biosynthesis genes and aberrant growth patterns including stunted bushy shoots and poor seed set. While these phenotypes possibly provide additional clues about BdRAM1’s scope of influence, they also indicate that directed approaches to increase the density of symbiotic interfaces will require a more focused, potentially cell-type specific manipulation of transcription factor gene expression.
Introduction
The GRAS (for GA3 INSENSITIVE [GAI], REPRESSOR OF GAI [RGA], and SCARECROW [SCR]) transcription factor REDUCED ARBUSCULAR MYCORRHIZA (RAM1) has been characterized in three dicot plant species where it is a major regulator of AM symbiosis. In Medicago truncatula, Lotus japonicus, and Petunia hybrida ram1 mutants, AM fungi display limited arbuscule branching and reduced hyphal colonization of the root, which results in a non-functional symbiosis (Gobbato et al., 2013; Park et al., 2015; Rich et al., 2015; Xue et al., 2015; Pimprikar et al., 2016). RAM1 expression is induced in colonized cortical cells and is regulated by CYCLOPS, a transcription factor of the common symbiosis signaling pathway (Pimprikar et al., 2016) and also by DELLA proteins (Park et al., 2015; Pimprikar et al., 2016), negative regulators of GA signaling (Daviere and Achard, 2013). RNA sequencing of ram1 mutants (Luginbuehl et al., 2017) as well as smaller scale gene expression analyses of roots overexpressing RAM1 (Park et al., 2015; Jiang et al., 2017), indicate that RAM1 either directly or indirectly regulates expression of several symbiosis-associated transcription factors, including the GRAS transcription factor RAD1, and three AP2-domain transcription factors of the WRINKLED5 (WRI5) family. RAM1 also either directly or indirectly regulates expression of genes involved in the production and transfer of lipids to the fungal symbiont (e.g. FatM, RAM2, STR) and the phosphate transporter PT4 (Gobbato et al., 2012; Park et al., 2015; Pimprikar et al., 2016; Jiang et al., 2017; Luginbuehl et al., 2017). However, so far, only one lipid biosynthesis gene, RAM2, has been established as a direct target of RAM1 (Gobbato et al., 2012). Regulation of the other lipid biosynthesis and transport genes likely occurs indirectly through the action of the WRI5 family genes (Luginbuehl et al., 2017; Jiang et al., 2018).
RAD1, a GRAS transcription factor very closely related to RAM1 (Supplemental Fig. 1) (Park et al., 2015; Xue et al., 2015) is also required for AM symbiosis. In L. japonicus rad1 mutants, AM fungi display defective arbuscule branching phenotypes reminiscent of those seen in ram1 (Xue et al., 2015); however, in M. truncatula rad1, AM fungi show normal arbuscule branching but reduced colonization levels (Park et al., 2015). In line with this observation, several predicted RAM1 target genes were induced in colonized L. japonicus ram1 mutants but induction was completely abolished in M. truncatula and P. hybrida ram1 (Park et al., 2015; Rich et al., 2015; Pimprikar et al., 2016; Luginbuehl et al., 2017). Thus, there are slight differences in regulation of AM symbiosis genes even between relatively closely related plant species (Pimprikar and Gutjahr, 2018).
Several other GRAS proteins are essential for AM symbiosis including DELLA/SLR1, a negative regulator of GA signaling (Floss et al., 2013; Foo et al., 2013; Yu et al., 2014; Floss et al., 2017). In della mutants, AM fungi show a severely reduced ability to enter cortical cells, and as a result almost no arbuscules are formed (Floss et al., 2013; Foo et al., 2013; Yu et al., 2014). Arbuscules are ephemeral structures, and the few arbuscules that are formed in della mutants display an increased lifespan, indicating that DELLA not only regulates arbuscule formation but also their degradation (Floss et al., 2017). Two other GRAS transcription factors critical for hormone signaling and AM symbiosis are NSP1 and NSP2. These transcription factors regulate phosphate-dependent strigolactone (SL) biosynthesis in M. truncatula and rice (Liu et al., 2011). SLs serve as direct plant communication molecules with AM fungi at the onset of the symbiosis. Mutants impaired in NSP or enzymes required for SL biosynthesis show a reduction in fungal entry into the root and consequently reduced colonization (Gomez-Roldan et al., 2008; Liu et al., 2011; Kobae et al., 2018). Thus, there are several examples of GRAS factors that connect hormone signaling and AM symbiosis.
Many GRAS factors operate in complexes with other GRAS proteins and emerging evidence suggests that this is also true of those involved in AM symbiosis. M. truncatula and L. japonicus RAM1 were reported to interact with RAD1 and NSP2 (which also interact with each other), but not NSP1 (Gobbato et al., 2012; Park et al., 2015; Xue et al., 2015; Heck et al., 2016). In addition, rice RAM1 interacts with the GRAS transcription factor DIP1, which in turn interacts with DELLA (Yu et al., 2014). M. truncatula DELLA proteins were found to interact with numerous other GRAS transcription factors, including RAD1, MIG1, NSP1, and NSP2 (Floss et al., 2016; Fonouni-Farde et al., 2016; Heck et al., 2016; Jin et al., 2016). While their functional significance for symbiosis remains to be determined, the interactions suggest the existence of interconnected transcriptional modules regulated by multiple GRAS transcription factors.
Brachypodium distachyon is a monocot model species capable of forming AM symbiosis (Hong et al., 2012) and amenable to genetic manipulation (Bragg et al., 2015). A recent study identified 48 GRAS transcription factors in the genome of B. distachyon (Niu et al., 2019). Here we report functional analyses of the GRAS transcription factor RAM1 in a monocot and assess the potential to alter the levels of symbiotic interfaces by manipulating RAM1 expression.
Results and Discussion
We identified Bradi4g18390 as the single B. distachyon homolog of the GRAS transcription factor RAM1 (Supplemental Fig. 1), a gene that is conserved in AM host plants and missing from non-hosts (Bravo et al., 2016). Similar to orthologous RAM1 genes of M. truncatula (Gobbato et al., 2013; Park et al., 2015), L. japonicus (Xue et al., 2015; Pimprikar et al., 2016) and P. hybrida (Rich et al., 2015), B. distachyon RAM1 expression is induced in mycorrhizal roots. Following inoculation with the AM fungus Diversispora epigaea (formerly Glomus versiforme), BdRAM1 transcripts increased over time in parallel with increasing colonization of the root system as reported by D. epigaea a-tubulin transcripts and the phosphate transporter gene BdPT7 (Hong et al., 2012), a plant gene marker of AM symbiosis (Figure 1A, B). However, while the transcriptional patterns mirrors the marker genes, it is noticeable that BdRAM1 transcript levels are low, as is often the case for transcriptional regulators.
The role of RAM1 in AM has been established in at least three dicot host plants (Gobbato et al., 2013; Park et al., 2015; Rich et al., 2015; Xue et al., 2015; Pimprikar et al., 2016) where it is essential to support arbuscule development and appears to act in the upper tier of a transcription factor hierarchy (Luginbuehl et al., 2017); when ectopically over-expressed in roots, RAM1 is sufficient to induce expression of several AM-induced genes in the absence of symbiosis (Park et al., 2015; Pimprikar et al., 2016). Given its AM-inducible expression and pivotal regulatory role, we hypothesized that constitutive, high-level expression of RAM1 might increase the occurrence of arbuscules and possibly overall colonization levels, and this might provide an opportunity to evaluate the functional consequences of modifying colonization patterns. To test this hypothesis, we transformed B. distachyon with an overexpression construct, BdRAM1 under the control of two copies of the constitutively active CaMV 35S promoter (35S:BdRAM1). In addition, we generated B. distachyon ram1 loss-of-function mutants via CRISPR/Cas9 editing.
Arbuscule development in B. distachyon ram1 mutants is partly impaired
Five independent transgenic lines carrying a CRISPR/CAS9 construct targeting BdRAM1 were generated and two lines in which BdRAM1 had been edited were chosen for subsequent analysis. In both transgenic lines, the genome had been edited by both guides (Supplemental Fig. 2); editing by the upstream-most guide resulted in premature stop codons and created a truncated protein of 16 amino acids in the first line, designated ram1-1. The second line, designated ram1-2, was bi-allelic with edits resulting in premature stop codons that generated truncated protein products of 16 and 42 amino acids.
ram1-1 and ram1-2 were inoculated with D. epigaea and the fungal colonization patterns were examined. Some ram1 roots showed aberrant infection units, reminiscent of the typical dicot ram1 phenotype, with intraradical hyphae and only small, sparsely branched arbuscules but no fully developed arbuscules (Figure 2A). However, infection units in other roots of the same plant, or even in other parts of the same root, showed an apparent wild-type morphology with some large, well-branched arbuscules (Figure 2B). In comparison with the empty vector control (E.V.) the frequency of aberrant infections in the B. distachyon ram1 mutants was 2.5 to 3-fold higher and their overall root colonization levels were 34 to 52% lower. Similar results were obtained in several experiments across several generations (Supplemental Fig. 2). These results indicate that BdRAM1 is required to enable wild-type levels of arbuscule development similar to its orthologs in dicots; however, the B. distachyon ram1 phenotype is clearly milder than that observed in dicot ram1 mutants. The finding that B. distachyon ram1 can support some full arbuscule development suggests that other proteins or pathways have the potential to compensate for loss of BdRAM1 function. One possible candidate is the GRAS protein RAD1, which is closely related to RAM1 and induced in roots highly colonized by AM fungi (Supplemental Fig. 1, Supplemental Fig. 3). In legumes, there is evidence of a species-specific “micro-diversification” of RAM1 and RAD1, with the relative contributions of the two transcription factors to arbuscule development and symbiotic gene expression varying depending on the host species (Park et al., 2015; Xue et al., 2015; Pimprikar et al., 2016; Pimprikar and Gutjahr, 2018). It is therefore conceivable that some diversification of GRAS factor functions has occurred during the evolution of monocots, which might explain the milder arbuscule development phenotype of B. distachyon ram1 mutants relative to M. truncatula ram1. Interestingly, there are other GRAS factor examples where the converse is true, for example the DELLA proteins, where rice slr (Yu et al., 2014) shows a stronger phenotype than the M. truncatula della double or triple mutants (Floss et al., 2013; Floss et al., 2017). However, in the absence of other monocot ram1 mutants for comparison, it is also possible that the milder ram1 phenotype observed here is a feature specific to B. distachyon.
Overexpression of RAM1 alters plant morphology and results in constitutive expression of AM marker genes
The generation of transgenic B. distachyon plants overexpressing RAM1 was surprisingly challenging; from two, full-scale independent transformation experiments, only three viable independent transgenic 35S:BdRAM1 overexpressing lines (35S:BdRAM1ox) were obtained and the seed production from these lines was exceedingly poor. In addition, we obtained two lines, which carried the 35S:BdRAM1 T-DNA but displayed wild type-like BdRAM1 transcript levels (35S:BdRAM1WT). By contrast, transgenic plants carrying 35S:NLS-GFP-GUS (hereto referred to simply as 35S:NLS-GFP), were generated without difficulty. Seed production from the latter two genotypes was not impaired.
In addition to poor seed production and viability, the shoot and root phenotypes of the three lines with transcriptional up-regulation of BdRAM1 (35S:BdRAM1ox) differed from the vector controls and from the 35S:BdRAM1WT plants. The 35S:BdRAM1ox plants were characterized by a stunted bushy shoot with increased tiller formation and increased leaf angles, as well as a decreased number of node roots (Figure 4A, Supplemental Fig. 4). Thus, constitutive, ectopic, overexpression of BdRAM1, clearly influences plant development. While the cause is unknown, it might be the result of ectopic expression of BdRAM1 target genes and/or perhaps an interference of RAM1 with other GRAS transcriptional networks, many of which regulate development (reviewed in (Cenci and Rouard, 2017). Either way, the aberrant developmental phenotypes likely explain the difficulties in regenerating transgenic lines and their fecundity.
The 35S:BdRAM1ox plants showed constitutive expression of B. distachyon orthologs of RAM2, STR, PT4 and FatM (Figure 3A); elevated expression of these genes would normally occur only in response to colonization by AMF (for example, Harrison et al., 2002; Paszkowski et al., 2002; Gutjahr et al., 2008; Zhang et al., 2010; Gobbato et al., 2012; Gutjahr et al., 2012; Hong et al., 2012; Bravo et al., 2017), and we observed a similar expression pattern of their orthologs in B. distachyon roots (Figure 1B, Supplementary Fig. 3). Given the prior knowledge from dicots, we had anticipated that 35S:BdRAM1ox would increase expression of these genes in roots, but it was surprising to see that expression of these genes was also induced in shoots (Figure 4B). There were some exceptions, expression of BdSTR increased in 35S:BdRAM1ox shoots, but not in roots while BdFatM2 showed the opposite expression pattern (Figure 3A, Figure 4B, Supplemental Fig. 5). Overall, these data indicate that BdRAM1 alone is sufficient to drive increased expression of these genes in the absence of AM fungi and that transcription co-factors - if required for BdRAM1 function - must be present in all tissues.
In dicots, RAM1 regulates expression of a second tier of transcription factors including RAD1 and three members of the WRINKLED family (WRI5a-c); the latter directly regulate expression of lipid genes (Park et al., 2015; Jiang et al., 2017; Luginbuehl et al., 2017). We found that BdRAD1, and the three B. distachyon AP2 family transcription factors most closely related to MtWRI5a-c (further denoted as BdWRI5.1, BdWRI5.2, and BdWRI5.3) were strongly induced in wild-type roots colonized with D. epigaea relative to mock-inoculated controls (Supplemental Fig. 3) but interestingly, only BdWRI5.1 was induced in non-colonized 35S:BdRAM1ox roots (Figure 3B). A similar pattern was observed in 35S:BdRAM1ox shoots (Figure 4B, Supplemental Fig. 5). Thus, in contrast to M. truncatula, a RAM1-independent pathway likely leads to up-regulation of BdRAD1, BdWRI5.2, and BdWRI5.3 in mycorrhizal roots. This points to functional diversification of the regulatory cascade responsible for the transcriptional reprogramming of roots during AM symbiosis in B. distachyon. In addition, it may provide an explanation for the relatively mild ram1 mutant phenotype we observed (Figure 2). Future research in other monocot species is required to determine if such a functional diversification is unique to B. distachyon or a monocot-specific phenomenon.
Arbuscule density is higher in RAM1 overexpressors relative to controls
The initial goal of this study was to test the hypothesis that constitutive overexpression of RAM1 would increase arbuscule density and/or colonization and then to use the plants to address secondary hypotheses about symbiotic performance.
To test the first hypothesis, we grew 35S:BdRAM1ox, 35S:BdRAM1WT and 35S:NLS-GFP control plants in substrate containing D. epigaea spores and evaluated colonization levels and arbuscule morphology. Colonization levels in 35S:BdRAM1ox and control plants did not differ significantly, although the variation was much greater in the 35S:BdRAM1ox plants (Figure 3C). Arbuscules in 35S:BdRAM1ox plants showed a wild-type morphology, but the number of arbuscules, which we assessed within a defined root volume below the hyphopodium, was on average 2-fold greater in the 35S:BdRAMox plants relative to controls (Figure 3D-F, Supplemental Fig. 5). Thus, 35S:BdRAM1ox plants have a higher capacity to establish and/or to maintain arbuscules relative to the control plants. As RAM1 regulates the expression of several other transcription factors, as well as genes involved in lipid biosynthesis and nutrient transport, the increased arbuscule density in the 35S:BdRAM1ox plants may result from a combination of factors including arbuscule initiation and/or regulation of arbuscule lifespan.
Unfortunately, the severe shoot growth and branching phenotype of the BdRAM1 overexpressors prevented a fair evaluation of symbiotic performance (Supplemental Fig. 5). While colonized 35S:BdRAM1ox plants and controls both showed an increase in shoot fresh weight and tiller number relative to their respective mock-inoculated controls, the differences in the developmental architecture of these lines precluded direct physiological comparisons. Consequently, it was not possible to determine whether the increased arbuscule density influenced symbiotic performance.
Hormone biosynthetic and regulatory gene expression is altered in BdRAM1 overexpressors
The shoot architecture phenotype of the 35S:BdRAM1ox plants is reminiscent of the phenotypes of several monocot hormone mutants. For example, rice and B. distachyon mutants defective in GA, SL, and BR biosynthesis or signaling display dwarf phenotypes with increased tillering (e.g. (Spielmeyer et al., 2002; Ishikawa et al., 2005; Asano et al., 2009; Lin et al., 2009; Thole et al., 2012). To obtain further clues about the BdRAM1 overexpression phenotype, we evaluated the expression of several genes associated with SL, GA, and BR signaling. B. distachyon orthologues of genes involved in SL biosynthesis (BdD27, BdD17, BdD10) (Seto and Yamaguchi, 2014) and GA biosynthesis (potential orthologs of A. thaliana GA3ox1 and GA20ox1) (Kakei et al., 2015) were down-regulated, while key BR biosynthesis genes (BdCPD, BdD2/CYP90D) but not BdDWF4 (Kakei et al., 2015) were elevated in non-colonized 35S:BdRAM1ox roots relative to the controls (Figure 5A, B). In contrast, the GA receptor GID1 and the GA-regulator DELLA/SLR1 (Daviere and Achard, 2013) as well as the regulators of SL signaling D3 and D53 (Seto and Yamaguchi, 2014), and the B. distachyon BR receptor BdBRI1 and the BR-responsive transcription factor BdBZR1 (Corvalan and Choe, 2017) were differentially regulated in 35S:BdRAM1ox roots relative to controls (Supplemental Fig. 6). Thus the transcript data indicate a disturbance in hormone biosynthetic and regulatory gene expression likely contributing to the altered shoot architecture. Because of substantial cross-talk between hormone signaling pathways (Itoh et al., 2001; Umehara et al., 2008; Unterholzner et al., 2015; Corvalan and Choe, 2017), it is not possible to predict the initial cause. As GA, SL and BR hormone pathways each involve regulation via GRAS-transcription factors (Tong et al., 2009; Liu et al., 2011; Chen et al., 2013), it is possible that ectopic overexpression of BdRAM1 disturbs GRAS-factor complexes, leading to mis-regulation of these pathways. Alternatively, one of the native functions of BdRAM1 may be to regulate aspects of hormone signaling. For example, in rice, RAM1 interacts with a DELLA interacting protein, DIP, and therefore it is possible that one of RAM1’s native functions is to influence GA signaling and that this is exacerbated in the 35S:BdRAM1ox, leading to further downstream effects on other pathways. If mis-regulation of GA biosynthesis gene expression translates to disturbed GA homeostasis in 35S:BdRAM1ox roots, an imbalance in GA-regulated arbuscule formation and degradation could result (Floss et al., 2013; Floss et al., 2017). Such a scenario might also explain the increased arbuscule numbers in 35S:BdRAM1ox roots as well as a dwarf shoot phenotype.
Conclusion
In conclusion, BdRAM1, similar to its orthologs in dicots, regulates arbuscule development and transcriptional regulation of several AM symbiosis-induced genes, although it is likely that there is some functional redundancy with other GRAS or WRI5 transcription factors. Constitutive overexpression of 35S:BdRAM1 increased the number of arbuscules relative to control plants; although the plants were unsuitable for experiments to assess the functional consequences of increasing the symbiotic interfaces, the data nevertheless indicate that it is possible to manipulate arbuscule density through expression of RAM1. Future research should focus on increasing RAM1 gene expression specifically in the root cortex. We predict such a strategy would increase arbuscule numbers without the accompanying developmental defects and would enable evaluation of the consequences of increasing the density of symbiotic interfaces and the effects on nutrient exchange during AM symbiosis.
Materials and Methods
Plant material and growth conditions
B. distachyon plants were grown in a growth chamber under a 12 h light (24°C)/12 h dark (22°C) regime. For all experiments that were conducted in the absence of an AM fungal symbiont, B. distachyon plants were grown in 20.5 cm long cones filled with sterile Terragreen (Oli-Dri) and play sand (Quikrete) in a ratio of 1:1. For all experiments involving AM symbiosis, B. distachyon plants were grown in cones filled with a sand-gravel mix, and were inoculated with 250 Diversispora epigaea spores (formerly Glomus versiforme) as previously described (Muller et al., 2019). For mock-inoculated controls, we added an appropriate volume of filtered spore wash solution instead of the spores. Unless otherwise stated, B. distachyon plants were fertilized once per week with 1/4-strength Hoagland’s fertilizer containing 20μM Pi and harvested 4-5 weeks after transplanting to cones.
To monitor AM growth responses, seedlings were planted into pots (3 seedling per 11cm diameter pot and 8 pots per genotype) containing a 1:20 mixture of autoclaved N7/N8 soil (Watts-Williams et al., 2019) to sand/gravel mix. The sand/gravel mix is a 2:2:1 mixture of play sand, fine black sand and gravel (as described in (Floss et al., 2017). 250 surface sterilized D. epigaea spores were place below each plant. Beginnning at 3 weeks post planting, the pots were fertilized weekly with 50 ml of 1/4-strength Hoaglands solution lacking phosphate and 9 ml of 0.5mM Ca3(PO4)2. Plants were harvested at 9 weeks post planting. The growth chamber conditions were as described above.
Plasmid generation
To clone the CRISPR/Cas9 construct targeting BdRAM1, we used the vector and cloning system described by Xie et al. (Xie et al., 2015). To design the primers (shown in Supplemental Table 1), gene-specific guide RNA sequences targeting Bradi4g18390 were identified using CRISPR-P (Lei et al., 2014) and CRISPR-PLANT (Xie et al., 2015) and selected based on their location in the coding sequence and low number of off-target sites. We generated a 2-guide CRISPR/Cas9 construct that targeted Bradi4g18390 at positions 32-54 bp (guide RNA1) and 280-302 bp (guide RNA2) downstream of the transcription start site (Supplemental Fig. 2). As a negative control we used the empty vector pRGEB32 (Xie et al., 2015).
To clone 35S:BdRAM1 overexpression constructs, the coding sequence of Bradi4g18390 was amplified using gene-specific primers flanked by attB1 and attB2 recombination sites (Supplemental Table 1), and cloned into pDONR221, resulting in the pENTR1-2 BdRAM1 entry clone. pENTR1-2 clones containing the coding sequence of NLS-GFP-GUS, as well as pENTR4-1 entry clones containing the CaMV35S promoter and pENTR2-3 containing the CaMV35S terminator were cloned previously (Ivanov and Harrison, 2014; Floss et al., 2017). To assemble the binary vectors for B. distachyon transformation, four vectors (pENTR4-1 containing the double CaMV35S promoter, pENTR1-2 containing BdRAM1 or NLS-GFP-GUS, pENTR2-3 containing the CaMV35S terminator and pHb7m34GW (Karimi et al., 2005)), were combined to generate 35S:BdRAM1 or 35S:NLS-GFP using the multi-site gateway cloning system (Invitrogen). All vector sequences were confirmed by Sanger sequencing.
Generation of B. distachyon transformants
The CRISPR/Cas9 constructs targeting BdRAM1 as well as the 35S:BdRAM1 and 35S:NLS-GFP constructs were transformed into B. distachyon (accession Bd21-3) following a previously established protocol(Bragg et al., 2015). Plantlets emerging from transformed calli (selectable marker: Hygromycin) were transplanted into Metro-Mix 350 and genotyped to test for the presence of the construct (see Supplemental table 1 for primer sequences). In addition, in the case of the CRISPR/Cas9 constructs, the CRISPR/Cas9 target loci were amplified using flanking primers and purified PCR products were Sanger-sequenced in order to identify gene edits.
Visualization and quantification of fungal root colonization
Fungal colonization of B. distachyon roots was visualized by staining with Wheat-Germ Agglutinin (WGA) coupled to Alexafluor488 as previously described (Hong et al., 2012). Roots were observed using a Leica M205 stereomicroscope and root colonization was quantified using the gridline-intersect method (McGonigle et al., 1990). To quantify the ram1 phenotype, roots intersecting the gridlines were scored into one of three categories: (1), not colonized; (2), colonized with wildtype-like arbuscules; (3), colonized with aberrant (sparsely branched or collapsed arbuscules) or no arbuscules. The ratio of category 3 over the overall number of intersections of colonized roots (category 2+3) x100 was used to determine the percentage of intersections without arbuscules/total colonization. Total root length colonization was calculated as the percentage of category 2+3 over the total number of intersections counted x 100. To study arbuscule morphology, WGA-Alexafluor488-stained roots were counterstained with propidium iodide to visualize plant cell walls, and observed with a Leica SP5 confocal microscope. To quantify arbuscule numbers in 35S:BdRAM1ox roots, confocal stacks from highly colonized roots were taken so that the fungal hyphopodium was in the center of the image to ensure we capture infections of similar developmental stages. The total number of arbuscules per stack was assessed manually using the Fiji Image Analysis Package (Schindelin et al., 2012). Stack depth (z-plane) was chosen to encompass the whole infection, and arbuscule numbers were normalized against the stack volume (length of x*y*z planes). To avoid potentially confounding effects caused by different B. distachyon root types, we selected only thin lateral roots with a single layer of cortical cells for analysis.
RNA isolation, cDNA synthesis, and quantitative PCR
RNA isolation, cDNA synthesis, and quantitative PCR were performed as previously described (Muller et al., 2019). Primers used to quantify expression of target genes are shown in Supplemental table 1. Ct values of the tested genes were normalized against BdEF1a {resulting in Ct} and relative expression levels were calculated with the formula 2-Ct.
Assessment of plant morphology
Plants were grown in the absence of AM fungi and whole plants were harvested 2, 4, and 6 weeks after planting. Tiller and node root numbers were counted, and maximal root system and shoot length were measured. The angle between individual leaves and the stem was measured on images of the same plants using the Fiji Image Analysis Package (Schindelin et al., 2012).
Phylogenetic analyses
B. distachyon orthologs of M. truncatula RAM1, RAD1, RAM2, PT4, FatM, STR, WRI5a-c as well as O. sativa D27, D17/CCD7, CCD8a, D3, D53, D14, D14L and SLR1 were identified using phylogenetic approaches described before (Supplemental Fig. 1, 7)(Bravo et al., 2016) B. distachyon genes putatively involved in Brassinosteroid and Gibberellic acid biosynthesis and signaling were identified previously (Kakei et al., 2015; Corvalan and Choe, 2017; Niu et al., 2019).
Statistical analyses and data representation
All experiments were performed using biological replicates. All experiments were repeated at least two times. The distribution of residuals was tested for normality using the Shapiro-Wilk test. If normality assumption was met, pairwise comparisons were analyzed using a two-sided Student’s t-test. For multiple comparisons, the raw data was subjected to a one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test. If normality assumption was not met, data were analyzed using the Kruskal-Wallis test followed by Dunn’s post-hoc test (p-values adjusted after Benjamini-Hochberg). All statistical analyses were performed using R software. Quantification data for n>5 biological replicates are represented as box-and-whiskers plots, which show the lower and upper quartiles as well as the minimum and maximum values. The horizontal line in the box plots represents the median. Points represent single measurements. For datasets with less than 5 measurements per genotype, bar plots were chosen. Bars represent the mean, and error bars the standard deviation. Points represent single measurements.
Figure legends
Acknowledgements
We thank Sophia Cotraccia, Cassandra Proctor, and Stephanie Roh for technical assistance and the BTI Biotechnology center for generating some of the B. distachyon transgenic lines. Financial support for the project was provided by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research (grant no. DE-SC0012460) and the TRIAD Foundation. LMM was supported by Postdoctoral Fellowships from the Swiss National Science Foundation (SNF, Early Postdoc.Mobility) and the German Research Foundation (DFG), LCS was supported by a Marie Curie Fellowship (FP7-PEOPLE-2013-IOF-624739).
Footnotes
↵* joint second co-authors