ReportsBOTANY

Plants transfer lipids to sustain colonization by mutualistic mycorrhizal and parasitic fungi

See allHide authors and affiliations

Science  16 Jun 2017:
Vol. 356, Issue 6343, pp. 1172-1175
DOI: 10.1126/science.aam9970

Food for fungi

A wide variety of plants form symbiotic relationships in their roots with arbuscular mycorrhizal fungi. The fungi channel inorganic and micronutrients from soil to the plant, and the plant supplies the fungi with organic nutrients. Jiang et al. and Luginbuehl et al. found that as part of this exchange, the plant supplies lipids to its symbiotic fungi, thus providing the fungi with a robust source of carbon for their metabolic needs.

Science, this issue p. 1172; p. 1175

Abstract

Arbuscular mycorrhizal (AM) fungi facilitate plant uptake of mineral nutrients and draw organic nutrients from the plant. Organic nutrients are thought to be supplied primarily in the form of sugars. Here we show that the AM fungus Rhizophagus irregularis is a fatty acid auxotroph and that fatty acids synthesized in the host plants are transferred to the fungus to sustain mycorrhizal colonization. The transfer is dependent on RAM2 (REQUIRED FOR ARBUSCULAR MYCORRHIZATION 2) and the ATP binding cassette transporter–mediated plant lipid export pathway. We further show that plant fatty acids can be transferred to the pathogenic fungus Golovinomyces cichoracerum and are required for colonization by pathogens. We suggest that the mutualistic mycorrhizal and pathogenic fungi similarly recruit the fatty acid biosynthesis program to facilitate host invasion.

About 80 to 90% of plant species are colonized by arbuscular mycorrhizal (AM) fungi, which facilitate the uptake of mineral nutrients such as phosphate and nitrogen from the soil (13). Evidence suggests that the fungus receives carbon from the plant in the form of sugars (hexoses) in return (47); however, plants could transfer alternative carbon sources. The AM fungal genome of Rhizophagus irregularis lacks genes encoding type I multidomain fatty acid synthases (FASs), which synthesize palmitic acid (C16:0) in fungi (810). Although the mitochondrial type II FAS genes are present in R. irregularis, the genes support production of octanoic acid (C8:0) rather than palmitic acid (1012).

To investigate whether AM fungi can synthesize fatty acids de novo, we performed isotope labeling experiments. Isotope ratios can be used as tracers to understand complex substrate metabolism (fig. S1A) (13). We used the R. irregularis–carrot (Daucus carota) root monoxenic culture system in divided petri plates, which prevents diffusion of nonvolatile solutes between compartments (fig. S1, A to C) (4, 5, 7). When isotopically labeled [1,3-13C]glycerol was added to the fungal extraradical mycelium (ERM) (fig. S1B), we found that 10.52 ± 2.02% and 3.47 ± 0.53% of glycerol moieties in the ERM and the intraradical mycelium/root compartment (IRM/R), respectively, were labeled at the fragment containing C1 or C3 atoms (Fig. 1A), indicating that glycerol absorbed by the ERM can be transferred between it and the IRM/R. Also, 6.54 ± 0.73% and 1.01 ± 0.21% of glucose moieties in the ERM and IRM/R, respectively, were labeled at the fragment containing C2 and C3 atoms (Fig. 1A), indicating that fungi can synthesize glucose from glycerol. We did not detect an enrichment of labeling in fatty acid moieties (Fig. 1A). Similar results were observed when [2-13C]glycerol was supplied in the ERM (fig. S1D). Thus, the AM fungus R. irregularis cannot synthesize fatty acids de novo from glucose and glycerol.

Fig. 1 Plant hosts synthesize and transfer fatty acids to sustain mycorrhizal colonization.

(A and B) Isotope tracers showed percentages of 13C labeled in glucose (C2-3-Glu), glycerol (C1/3-Gly), and the C2 unit of C16:0 fatty acid (16-FA) after [1,3-13C]glycerol was supplied to the ERM (A) and the IRM/R (B) in the R. irregularis–carrot root culture system. (C) Quantification of AM colonization in plant roots overexpressing MtPK, MtKAS II, MtKAR, and MtFatM at 42 days post-inoculation with R. irregularis (dpi). RLC, root length colonization. (D) Quantification of AM colonization in MtPK, MtKAS II, and MtFatM RNA interference (RNAi) plant roots at 42 dpi. (MtKAR RNAi plants failed to generate hairy roots.) (E) Lauric acid content of UcFatB-overexpressing (UcFatB) and empty vector (EV) roots without mycorrhizal fungus infection, expressed as mole percent of total fatty acids. (F) Lauric acid content in the ERM in the UcFatB-overexpressing and EV plant inoculation system (fig. S4A). The ERM was collected from R. irregularis colonization sand at 42 dpi. These experiments were repeated three times with similar results. Values are the mean ± SE of measurements from three independent plates, 8 to 12 plants of each genotype, or the ERM from six pots. *P < 0.05; **P < 0.01; ns, not significant (Student’s t test).

To follow carbon transfer during AM symbiosis, we added [1,3-13C]glycerol to the IRM/R in the R. irregularis–carrot root culture system (fig. S1C). If sugars are the sole form of carbon transferred from the plant to the AM fungus, we would expect higher or equal labeling ratios of sugars in the IRM/R relative to the ERM when labeled glycerol is supplied in the IRM/R. We observed 7.62 ± 0.35% of glucose moieties in the IRM/R labeled at the fragment containing C2 and C3 atoms, one-fourth that in the ERM, where 28.31 ± 0.71% of glucose moieties were labeled at the fragment containing C2 and C3 atoms (Fig. 1B). [2-13C]glycerol supplied in the IRM/R (fig. S1E) had similar effects. The greater enrichment of 13C that we detected in glucose in the ERM compared with that in the IRM/R (Fig. 1B) suggests that glycerol is not directly metabolized to glucose within plant roots for transfer to the fungus. Instead, these data suggest that glycerol might be incorporated into the backbone of glycerolipids that are transferred to the ERM, then converted to glucose by gluconeogenesis in the fungus.

Several fatty acid biosynthesis genes—including those encoding pyruvate kinase (MtPK), ketoacyl-ACP synthase II (MtKAS II), ketoacyl–acyl carrier protein (ACP) reductase (MtKAR), enoyl-ACP reductase I (MtENR I), and acyl-ACP thioesterase B (MtFatM, also known as FatB in Arabidopsis)—were induced by mycorrhizal fungal infection in Medicago truncatula (fig. S2, A and B, and table S1) (14, 15). Overexpression of MtPK, MtKAS II, MtKAR, and MtFatM increased AM colonization, whereas knockdown of these genes reduced AM colonization (Fig. 1, C and D, and fig. S2, C and D). Furthermore, a Mtfatm mutant obtained from a Tnt1-insertion line was impaired in AM fungal colonization (fig. S3, A and B), consistent with previous observations (16). MtFatM is expressed in arbuscule-containing cells, revealed by a β-glucuronidase (GUS) reporter driven by the MtFatM promoter (fig. S3, C and D). Thus, M. truncatula fatty acid biosynthesis genes support AM symbiosis.

To determine whether fatty acids synthesized in the host plant can be transferred to the AM fungus, we genetically engineered lauric acid (C12:0 fatty acid) synthesis in M. truncatula roots. Umbellularia californica lauroyl-ACP thioesterase (UcFatB) terminates acyl-chain elongation in fatty acid biosynthesis early by releasing the medium-chain fatty acid, lauric acid, from ACP (17). We found that overexpression of UcFatB increased the amount of lauric acid to 2.5 mol % in root fatty acids, representing a >25-fold increase relative to the empty vector–transformed control (Fig. 1E and fig. S4). Cis-palmitvaccenic acid (C16:1ω5) is normally absent from plant roots and has been used as an AM fungal marker (6). Plants overexpressing UcFatB displayed a >12-fold increase compared with the empty vector–transformed control in fungal lauric acids or triglyceride (TAG) (C44:2, C16:1–C16:1–C12:0) containing C16:1ω5 fatty acid in addition to C12:0 in the ERM or IRM/R (Fig. 1F and fig. S4). This result supports that fatty acids are transferred from the host root to the AM fungus.

RAM2 synthesizes 2-monoacylglycerols for accumulation of extracellular lipid polyesters such as cutin on the surface of shoot organs (1820). Mutant ram2 plants are deficient in AM fungal colonization when grown as a monoculture (individual plant) (20) or cultivated together with another ram2 plant (Fig. 2A). However, this mutant can be colonized with high expression of PHOSPHATE TRANSPORTER 4 by R. irregularis when cultivated together with a wild-type nurse plant to increase inoculums strength, although ram2 roots contain fewer fully developed arbuscules than wild-type roots in nurse plants (Fig. 2, A and B, and fig. S5). The fact that R. irregularis colonization of ram2 roots occurs in the presence of a nurse plant (Fig. 2, A and B) allowed us to test whether RAM2 is required for the accumulation of plant-derived fatty acids in AM fungi. When we expressed UcFatB in ram2 roots (ram2-UcFatB), the fungal TAG (C44:2) and lauric acid did not accumulate in the IRM/R and ERM associated with these plants in the presence of a wild-type nurse plant (Fig. 2, C and D, and fig. S6). These data suggest that 2-monoacylglycerols synthesized by RAM2 are required for the transport of fatty acid species from the root to the AM fungus.

Fig. 2 Fatty acid transfer from plant to AM fungus requires RAM2.

(A) Quantification of R. irregularis colonization in ram2 and wild-type (WT) plants grown with a ram2 or WT nurse plant at 42 dpi. The tester plant is labeled in red, and the nurse plant is labeled in black. (B) Images of WGA-AF488–stained arbuscules (left); bright-field (BF) images (middle); and WGA-AF488–BF merge images (right). (C) Lauric acid content of WT-EV, WT-UcFatB, ram2-EV, and ram2-UcFatB roots without R. irregularis infection. (D) Lauric acid content in ERM associated with WT-EV, WT-UcFatB, ram2-EV, and ram2-UcFatB plants. ram2 plants grew in the presence of a WT nurse plant (fig. S6A). WT plants grew as a monoculture. ram2 plants are deficient in AM fungal colonization when cultivated together with another ram2 plant (A). The ERM was collected from R. irregularis–colonized sand at 42 dpi. These experiments were repeated three times with similar results. Scale bar, 30 μm. Values are the mean ± SE of measurements performed on 8 to 12 plants. *P < 0.01 (Student’s t test).

The heterodimeric adenosine triphosphate (ATP)–binding cassette (ABC) transporters STR (stunted arbuscule) and STR2 are specifically localized in the peri-arbuscular membrane and are required for AM symbiosis (21, 22). STR and STR2 are half-transporter proteins of the ABCG subfamily, of which ABCG11 and -12 are required for the accumulation of wax and cutin (23). Co-overexpressing STR and STR2 in Medicago roots and Arabidopsis leaves led to higher accumulation of extracellular lipid polyesters, such as cutin monomers (Fig. 3A and figs. S7 and S8). The effect was limited to specific monomers—notably, 16:0-α,ω-dicarboxylic acid (DCA) and 18:2-DCA (Fig. 3B). Consistent with the low level of STR and STR2 expression under nonmycorrhizal conditions (21), we found no differences in the composition of cutin monomers in leaves or roots of wild-type and str plants without mycorrhizal infection (fig. S9).

Fig. 3 STR and STR2 mediate lipid export from the plant to the AM fungus.

(A and B) Cutin analysis of total load (A) and monomer load (B) in EV and STR-STR2 co-overexpression (STR-STR2-OE) roots. Each value is the mean ± SE of measurements performed on 15 to 20 plants. FW, fresh weight. (C) Lauric acid content of WT-EV, WT-UcFatB, str-EV, and str-UcFatB hairy roots without R. irregularis infection. (D) Lauric acid content in ERM associated with WT-EV, WT-UcFatB, str-EV, and str-UcFatB plants. DCA, α,ω-dicarboxylic acid; OHFA, hydroxylated fatty acid. str plants grew in the presence of a WT nurse plant (fig. S11A). WT plants grew as a monoculture. The ERM was collected from R. irregularis–colonized sand at 42 dpi. These experiments were repeated three times with similar results. Values are the mean ± SE of measurements performed on 8 to 12 plants. *P < 0.05; **P < 0.01 (Student’s t test).

The R. irregularis colonization of str roots occurs with more fully developed arbuscules in the presence of a nurse plant than when str is grown as a monoculture (fig. S10), reminiscent of the ram2 phenotype. When we expressed UcFatB in str roots (str-UcFatB), the fungal TAG (C44:2) and lauric acid did not accumulate in the IRM/R and ERM associated with these plants in the presence of a wild-type nurse plant (Fig. 3, C and D, and fig. S11). Altogether, we conclude that the heterodimeric ABC transporters STR and STR2 are responsible for delivery of lipids to the AM fungus.

Powdery mildew (PM) and Ustilago maydis fungi are parasitic pathogens that infect a large number of plant species. We noticed that these pathogenic fungi induce expression of fatty acid biosynthesis genes in Arabidopsis and maize, respectively (fig. S12A and table S2) (24, 25). The induction of KAR and KAS I by PM was further verified by a GUS reporter driven by the AtKAR and AtKAS I promoters in Arabidopsis (fig. S12B). Columbia-0 (Col-0) plants are susceptible to the PM pathogen Golovinomyces cichoracerum UCSC1, supporting the production of a large number of spores (26, 27) that can be used for lipid analysis. We expressed UcFatB in Col-0 and observed a 3- to 4.5-fold increase relative to the empty vector–transformed control in lauric acid of G. cichoracerum fatty acids associated with UcFatB overexpression in Arabidopsis leaves (Fig. 4, A and B, and fig. S13). Furthermore, we found that kas1, kar1, and fatb-1 mutant plants were smaller and showed enhanced disease resistance compared with wild-type plants: Fungal colonies on mutant leaves had fewer spores and conidiophores (Fig. 4, C and D, and fig. S14). Clones containing KAS I, KAR, and FatB genes complemented the kas1, kar1, and fatb-1 mutations, respectively (fig. S15). Altogether, our data indicate that G. cichoracearum can take up fatty acids from the plant and that reduced plant fatty acid biosynthesis impairs pathogenic fungal infection.

Fig. 4 Plant hosts supply fatty acids to the pathogenic fungus.

(A) Lauric acid content in 21-day-old WT, EV, and UcFatB-overexpression plant rosette leaves (three independent transgenic lines: 16, 20, and 22). (B) Lauric acid content of G. cichoracearum collected from WT, EV, and UcFatB-overexpression rosette leaves at 8 dpi. (C) Four-week-old wild-type (Col-0), kas1, kar1, fatb-1, and fatb-2 plants were inoculated with G. cichoracearum. Trypan blue staining of the leaves was used to visualize fungal structures and plant cell death at 8 dpi. Extensive plant cell death was observed in the kas1 mutant. Scale bar, 100 μm. (D) Quantitative analysis of conidiophore formation on 4-week-old wild-type and mutant plants at 5 dpi. Values are the mean ± SD of measurements performed on 30 independent colonies in one experiment. The experiment was repeated twice with similar results. *P < 0.05; **P < 0.01 (Student’s t test).

We suggest that sources of carbon for mutualistic AM fungi include fatty acids exported from the host plant, as well as sugars (fig. S16). Consistent with the nutrient role of fatty acids in fungi, genes encoding enzymes involved in fatty acid degradation (fig. S17) and fatty acid elongation (fig. S18) are found in the R. irregularis genome (28). 2-monoacylglycerols, likely exported by ABCG transporters, accumulate as extracellular lipid polyesters at the plant surface in Arabidopsis leaves (18, 19, 23). In mycorrhizal symbiosis, the 2-monoacylglycerols synthesized by RAM2 are likely exported by the peri-arbuscular membrane-localized heterodimeric ABC transporters STR and STR2 into the interface space and then taken up by AM fungus (fig. S16) (16, 21, 22). Induction of RAM2 expression by the Myc factor signal recognition pathway requires the GRAS-domain transcription factor RAM1 (29). The expression of fatty acid biosynthesis genes appears to be regulated by transcription factors that regulate arbuscule development (fig. S19) (30). Fatty acid biosynthesis in plants is usurped by parasitic pathogenic fungi to secure fatty acids and promote infection.

Supplementary Materials

www.sciencemag.org/content/356/6343/1172/suppl/DC1

Materials and Methods

Figs. S1 to S20

Tables S1 to S3

References (3155)

References and Notes

  1. Acknowledgments: We thank J. Murray, C. Stonoha, and D. Wang for helpful suggestions on the manuscript; M. Harrison for providing str mutant seeds; W. Hu for gas chromatography–mass spectrometry and gas chromatography–quadrupole time-of-flight mass spectrometry analysis; Y. Shan for liquid chromatography–high-resolution mass spectrometry analysis; and the company Hanyu Bio for reanalysis of R. irregularis. The research was supported by the 973 National Key Basic Research Program in China (grant 2015CB158300), the Ministry of Agriculture of China for Transgenic Research (grant 2016ZX08009003005-003), the National Science Foundation of China (grant 31522007), the China Postdoctoral Science Foundation (2014M560358), and the Strategic Priority Research Program “Molecular Mechanism of Plant Growth and Development” of the Chinese Academy of Sciences (grant XDPB0404). Y.J. and E.W. directed the research. Y.J., W.W., and Q.X. performed most of the experiments. Other authors contributed to the analytical, molecular cloning, and transformation work. Y.J. and E.W. wrote the manuscript. The authors declare that they have no competing interests. The supplementary materials contain additional data.
View Abstract

Navigate This Article