The genetic architecture of host response reveals the importance of arbuscular mycorrhizae to maize cultivation

Abstract

Arbuscular mycorrhizal fungi (AMF) are ubiquitous in cultivated soils, forming symbiotic relationships with the roots of major crop species. Studies in controlled conditions have demonstrated the potential of AMF to enhance the growth of host plants. However, it is difficult to estimate the actual benefit in the field, not least because of the lack of suitable AMF-free controls. Here we implement a novel strategy using the selective incorporation of AMF-resistance into a genetic mapping population to evaluate maize response to AMF. We found AMF to account for about one-third of the grain production in a medium input field, as well as to affect the relative performance of different plant genotypes. Characterization of the genetic architecture of the host response indicated a trade-off between mycorrhizal dependence and benefit. We identified several QTL linked to host benefit, supporting the feasibility of breeding crops to maximize profit from symbiosis with AMF.

Keywords: Rhizophagus irregularis; arbuscular mycorrhiza; benefit; dependence; maize; plant biology; trade-off.

Figure 1.. Mutation of maize Castor results in resistance to AMF.

(A) Schematic representation of the maize Castor gene (GRMZM2G099160_T01 gene model). Boxes indicate coding regions. Red triangles indicate the insertion of Mutator transposable elements in the alleles castor-1 and castor-2. (B) Root segments stained with trypan blue 7 weeks after inoculation with Rhizophagus irregularis. Characteristic root-internal hyphal structures, such as intraradical hyphae (IH) and arbuscules (A) , seen in wild-type (WT) and castor-1 plants are absent from the castor-2 mutant. (C) RT-PCR analysis of Castor transcript accumulation in the roots of WT and castor-2 plants that were non-inoculated (NM) or inoculated with plate (M_p. In vitro produced) or crude (M_c. Sand-pot produced) inoculum. Castor cDNA was amplified using a primer set spanning the castor-2 mutator insertion site (5’) and a second set spanning the 3’-most intron (3’). Primers to the maize actin gene Act1 were used as a control. (D) Colonization by AMF (% root length) of the plants analyzed in C. TC, total colonization; EH, external hyphae; IH, internal hyphae; A, arbuscules; V, vesicles; S, spores; H, hyphapodia. C, crude inoculum. P, plate inoculum. Lowercase letters indicate significant differences among genotypes and inoculum, determined by the Kruskal-Wallis test and LSD.

Figure 1—figure supplement 1.. Generation of the F_2:3 population.

(A) An F₁hybrid was generated from the cross between the CML312 inbred line and a plant homozygous for the castor-2 allele in the W22 background. The F₁ was self-pollinated to generate an F₂ stock segregating castor-2 along with CML312 and W22 genome content. Homozygous wild-type and mutant individuals were identified by PCR genotyping and self-pollinated to generate F_2:3 families. Additional tissue from the F₂ parents was used for genotypic analysis on an Illumina 3047 SNP microarray chip as described in the main text. (B) An F₂ ear from the cross of the color-converted bz-mum9 Uniform Mu stock carrying castor-2 and the white-kernelled subtropical inbred line CML312.

Figure 1—figure supplement 2.. Maize castor-2 mutants are resistant to colonization by AMF.

Field evaluation of castor-2 mutants, Irapuato, Mexico. Two complete blocks with four treatments were used: (1) AMF-S (wt) families without inoculation with exogenous AMF (NC), (2) AMF-R families (castor-2) without inoculation with AMF (NC), (3) AMF-S families (wt) with inoculation with AMF (M), and (4) AMF-R families (castor-2) with inoculation of AMF (M). The inoculum used was the commercial consortium BioMic which includes: Glomus constrictum, Glomus geosporum, Glomus tortuosum, Acaulospora scrobiculata, Gigaspora margarita, and Glomus sp. Field colonization of wt families without application of exogenous inoculum indicates the action of the native community of AMF.

Figure 1—figure supplement 3.. Evaluation of AMF-S (wt) and AMF-R (castor) families in heavily managed (PV) and a rain-fed, medium input (Ameca) fields.

A subset of 20 AMF-S and 20 AMF-R families was evaluated in a winter season in a managed (fertigation, herbicide, pesticide; sandy soil allowing free availability of nutrients to plants) field site near Puerto Vallarta on the Mexican Pacific Coast. No difference was seen between AMF-S and AMF-R families in plant height or days-to-silking (female flowering). By contrast, in the medium-input Ameca site used for the full evaluation, AMF-S and AMF-R families were distinct. The box in violin plots represents the interquartile range with the horizontal line representing the median and whiskers extend 1.5 times the interquartile ranges. The shape of the violin plot represents the probability density of data at different values along the y-axis. Results were based on two-group Wilcoxon tests with Bonferonni adjusted p-values. Note: *: p<0.05; **: p<0.01; ***: p<0.001; NS: not significant.

Figure 2.. Mycorrhizal symbiosis promotes maize growth under cultivation.

(A) Susceptible (AMF-S) and resistant (AMF-R) families segregate genomic content from two founder parents, shown as red and blue bars, but are homozygous for the wild-type (green) or mutant (yellow) allele at Castor (black arrow), respectively, blocking AM symbiosis in AMF-R. (B), Border between representative AMF-S and AMF-R plots, Ameca, Mexico, 2019. (C), Representative AMF-S ears. (D), Representative AMF-R ears. (E), Plant height (PH) of 73 AMF-S (green) and 64 AMF-R (yellow) families. (F), Anthesis-silking interval (ASI). (G), Kernel shape (KS) based on the analysis of scanned kernel images . (H), Total kernel number (TKN). The violin plots in (E, F and H) are a hybrid of boxplot and density plot. The box represents the interquartile range with the horizontal line representing the median and whiskers extending 1.5 times the interquartile range. The shape of the violin plot represents the probability density at different values along the y-axis. ***, difference between AMF-S and AMF-R significant at p<0.001 (Wilcoxon test; Bonferonni adjustment based on the total trait number).

Figure 2—figure supplement 1.. Location of experimental fields.

(A) Satellite view of the Ameca site used for the principal field evaluation (20.57 lat, −104.04 lon, 1268 masl). Yellow rectangle shows the field area. (B) Satellite view of the Puerto Vallarta (PV) site used for population generation and preliminary high-input evaluation (20.78 lat, −105.243 lon, 41 masl). Red rectangle shows the field area. (C) Mexico showing the two sites.

Figure 2—figure supplement 2.. Representative AMF-S and AMF-R families in high- and medium- input sites.

(A) Winter PV field in 2020, located in Nayarit, Mexico. (B) Summer Ameca field in 2019, located in Jalisco, Mexico.

Figure 2—figure supplement 3.. Plan of the Ameca field design.

Summer experimental field 2019, located in Ameca, Jalisco. Three complete blocks were evaluated. Forty-five seeds were sown per three-row plot (15 seed per row). AMF-S and AMF-R families were alternated with the order of the families within each subpopulation randomized within each block. The rows marked with red, blue, and green correspond to blocks 1, 2, and 3, respectively. A commercial UNISEM hybrid was planted on the border of the experiment and used as a check throughout the field. The field was fertilized at planting with 250 kg / Ha of diammonium phosphate (DAP) as a source of nitrogen and phosphorus (18-46-00, NPK). A further application of 250 kg / Ha of urea was given (46-00-00, NPK) at 40 days after planting.

Figure 2—figure supplement 4.. Comparison of plant phenotypic traits between AMF-S (wt) and AMF-R (castor) families.

The box in violin plots represents the interquartile range with the horizontal line representing the median and whiskers representing 1.5 times the interquartile ranges. The shape of the violin plot represents the probability density of data at different values along the y-axis. Results were based on two-group Wilcoxon tests with Bonferonni adjusted p-values. Note: *: p<0.05; **: p<0.01; ***: p<0.001; NS: not significant. Trait codes as in Table 1.

Figure 2—figure supplement 5.. Ear and kernel image analysis.

The shape of maize ear (A), cob (B), and kernel (C) of AMF-S (wt) and AMF-R (castor) families. Median shapes of ear, cob, and kernel from two families were shown in thick green and yellow lines, whereas individual genotypes of two groups were shown in thin, semi-transparent green, and yellow lines. Shapes were extracted from scanned images from 3 ears per plot, 50 kernels per ear.

Figure 2—figure supplement 6.. Principal component analysis.

A, Phenotypic variation captured by the first five principal components (PCs). The first two principal components (PCs) explained more than 50% of the phenotypic variance. PC1 (40.1%) was primarily associated with ASI and maize yield-related traits, such as ear weight (EW), total kernel number (TKN), and total kernel weight (TKW). B, Loading of AMF-S and AMF-R families on PC1 and PC2. The position of trait names (as Table 1) with respect to the origin indicates contribution to each PC.

Figure 3.. QTL x AMF effects contribute significantly to variation in host response.

(A) Genomic position of QTL identified in this study. Boxes indicate the position of the peak marker. Bars represent the drop-1 LOD interval. Colors denote patterns of AMF × QTL interaction as described in the text. (B) The contribution of different components to phenotypic variance in different plant traits among AMF-S and AMF-R families. Total genetic variance was estimated based on differences between experimental blocks and partitioned into variation explained by the additive effect of AMF, the additional variation explained by interactive QTL and QTL × AMF interaction (QTL × AMF), the additional variation explained by additive QTL (QTL) and residual genetic variation (Res H²). The balance of the phenotypic variance was assigned to the environment (E). Traits codes as in Table 1.

Figure 3—figure supplement 1.. Genotypic analysis.

Plot of genetic (y axis) and physical (x axis) position for 1050 genetic markers that compose the genetic map of the CML312 x W22 castor-2 F₂ population.

Figure 3—figure supplement 2.. Genetic map.

Plot representing the estimated genetic map and the marker distribution across the ten chromosomes of maize for the CML312 × W22 castor-2 F₂ population.

Figure 4.. Mycorrhiza response confounds benefit and dependence.

(A) Host response (R) is the difference in trait value between mycorrhizal (M; green) and non-mycorrhizal (NM, yellow) plants. Increased response can result from either greater dependence (D; poorer performance of NM plants) or greater benefit (B; greater performance of M plants). (B), QTL × AMF effects underlying variation in response reveal the balance of benefit and dependence. In this theoretical example, the effect is conditional on mycorrhizal colonization, reflecting a difference in benefit more than dependence. (C, D) Effect plots for major QTL associated with PC1 and PC2, respectively. Effect of the homozygous CML312 (red), homozygous W22 (blue), or heterozygous (brown) genotypes in AMF-S and AMF-R families. The PC1 QTL is conditional on AMF-R, indicating a difference in dependence. The PC2 QTL is conditional on AMF-S, indicating a difference in benefit.

Figure 4—figure supplement 1.. Scenarios in AMF x genotype interaction.

Theoretical performance (arbitrary units; y axis) of mycorrhizal (M) and non-mycorrhizal (NM) plants of two genotypes (red and blue). The difference between M and NM defines response for a given genotype. (Top left) No AMF × genotype interaction. Both genotypes respond equally to AMF; the difference between varieties is constant between M and NM. (Top right) NM conditional. Response is higher in the blue genotype; difference between genotypes is conditional on NM treatment; response variation is driven by greater dependence of the blue genotype. (Bottom left) M conditional. Response is higher in the red genotype; difference between genotypes is conditional on M treatment; response variation is driven by greater benefit of the blue genotype. (Bottom right) Antagonistic pleiotropy. Response is higher in the red genotype; difference between the genotypes is expressed in both M and NM, but the sign of the effect changes; the red genotype is superior in M growth but inferior in NM growth.

Figure 5.. The genetic architecture of mycorrhiza response implies a trade-off between dependence and benefit.

(A) Extreme QTL × AMF effects result in antagonistic pleiotropy. In this theoretical example, the blue allele is superior in mycorrhizal (M) plants, while the red allele is superior in non-mycorrhizal (NM) plants. Such a QTL is linked to differences in both dependence and benefit. (B) Effect plot for a total kernel weight (TKW) QTL located on Chromosome 1. TKW of families homozygous for the CML312 allele at this locus (red) is stable across susceptible (AMF-S) and resistant (AMF-R) families. In contrast, families homozygous for the W22 allele (blue) or heterozygous at this locus (brown), show superior TKW if AMF-S but lower TKW if AMF-R. (C) Theoretical trade-off between the performance of non-mycorrhizal (NM) and mycorrhizal (M) plants. For any position on this plot, the vertical distance above the diagonal indicates the magnitude of host response. Trade-off prevents occupancy of the upper-right of the plot, defining a so-called ‘Pareto front’ (solid arc). Here an increase in the NM performance of variety X is necessarily accompanied by a reduction in M performance, as described by movement along the arc to condition Y. Biological constraint prevents the path towards condition Z. (D) Fitted ear weight values for all genotypes from a multiple QTL model under both levels of AMF (AMF-S and AMF-R). Line segments connect the same plant genotype under the two AMF levels. (E) Observed ear weight against whole-genome prediction for 73 susceptible (AMF-S) and 64 resistant (AMF-R) families. Best fit regression line in dashed blue. (F) Predicted ear weights for all 137 genotypes based on AMF-S (M) and AMF-R (NM) whole-genome models. Dashed red line shows AMF-S = AMF R, that is, no host response.