Fig 1.
Physiological features of Sierra Mixe maize.
(A) The transition between juvenile and adult phases (black arrow) occurs 5 weeks after planting in Sierra Mixe maize (black bars) and in the tall maize heirloom Hickory King (gray bars). (B) Number of aerial roots observed on Sierra Mixe maize and Hickory King after 14 weeks of growth in the field in Madison, United States of America. Error bars represent standard errors; an asterisk indicates a significant difference between Sierra Mixe maize and Hickory King (Student t test, P < 0.01). (Data at DOI: 10.6084/m9.figshare.6534545).
Fig 2.
The aerial roots of Sierra Mixe maize (left) secrete large quantities of mucilage between 3 and 6 months after planting. The mucilage is carbohydrate rich, with the composition dominated by arabinose, fucose, and galactose (side panel).
Fig 3.
DNA sequencing–based characterization of the microbiome of Sierra Mixe maize.
(A) Samples clustered by PCoA on Bray-Curtis dissimilarity distance matrix. Bacterial communities were assessed using PCR amplification and sequencing of rRNA genes. Each point corresponds to an individual sample. Permanova tests run using adonis in vegan revealed mucilage samples were statistically distinct (e.g., mucilage versus rhizosphere P = 0.002, mucilage versus roots P = 0.03, mucilage versus aerial roots P = 0.03). (B) Metagenomic sequencing–based analysis of homologs of core nif genes (nifH, nifD, nifE, nifK, nifN, and nifB) and alternate nitrogenase (anfG/vnfG). Metagenomic samples were searched for homologs by mapping reads on reference nif trees. The number of hits was normalized to an estimate of the number of bacterial genes in the metagenomic sample (measured using the number of hits to the RecA hidden Markov model). *The lowest abundance of a core nif gene in mucilage and rhizosphere libraries. **Only core nif gene hit in stem library. anfG and vnfG alternate nitrogenase were observed only in mucilage library. PCoA, principal component analysis.
Fig 4.
Nitrogenase and N2 fixation activity in mucilage produced by Sierra Mixe maize.
(A) Mucilage of various Sierra Mixe maize lines collected in Sierra Mixe or field-grown plants in Madison, USA, and of teosinte display strong acetylene reduction activity. (-, no acetylene; +, 10% acetylene). Asterisks indicate significant differences (*P < 0.05; **P < 0.01, Mann-Whitney test). (B) Nitrogen fixation in Sierra Mixe maize mucilage by 15N2 assimilation. Mucilage collected from Sierra Mixe maize grown in Sierra Mixe was incubated in gas-tight vials filled with 15N2 or 14N2 gas for 70 hours at 37 °C. 15N (atom % excess) was determined by IRMS. (C and D) H. seropedicae and A. brasilense display acetylene reduction activity when added to nonfixing mucilage, whereas the same mucilage supplemented with sterile medium (-) or the same bacteria without mucilage (-) do not. (E) Oxygen concentration at 8 mm inside of the mucilage. Means and standard errors are shown. Different letters indicate statistically supported groups (Kruskal-Wallis test). (Data at DOI: 10.6084/m9.figshare.6534545). IRMS, isotope-ratio mass spectrometry.
Fig 5.
Analysis of Sierra Mixe maize samples for 15N2 enrichment in mucilage, aerial roots, and pheophytin from aerial roots.
Mucilage was generated from aerial roots and inoculated with A. brasilense Sp7. 15N2 gas was pumped in and, after incubation mucilage, was separated from the aerial roots. Mucilage alone and aerial roots alone were subjected to analysis by IRMS. Results revealed a significant enrichment of 15N2 in mucilage alone and aerial roots alone (left y-axis). Since the inoculated aerial roots may contain A. brasilense Sp7 attached to the surface, we extracted pheophytin from these aerial roots to test 15N2 incorporation in pheophytin. Results revealed a significant enrichment of 15N2 in these aerial roots, indicating that aerial roots are indeed the sites for transfer of fixed nitrogen to the plants (right y-axis). 14N2 samples were used as negative controls. n = 4 (aerial roots and mucilage), n = 3 (pheophytin). The asterisk (*) indicates a statistically significant difference (p < 0.05). (Data at DOI: 10.6084/m9.figshare.6534545) IRMS, isotope-ratio mass spectrometry.
Table 1.
Natural abundance 15N determinations.
(A) δ15N values from Sierra Mixe maize, a conventional maize variety (Maiz Blanco Conasupo), and reference plants grown in Sierra Mixe, 3 months after planting. Values are given as mean and standard deviation. Statistical comparisons were made between the means of the reference plants, Maiz Blanco Conasupo, and Z. mays S. Mixe, using Student t tests (p < 0.05). Different letters indicate statistically supported groups. (B) δ15N values from Sierra Mixe maize plants grown in Fields 1 and 2 in Sierra Mixe during 2011 and 2012 at 2, 3, 4, 5, and 6 months after planting. Values are given as mean and standard deviation. Statistical comparisons were made between the reference plants mean and Z. mays S. Mixe means at each time point using Student t tests (p < 0.05). Different letters indicate statistically supported groups. Reference plants are listed in S3 Table. (Data at DOI: 10.6084/m9.figshare.6534545).
Table 2.
15N-isotope-enrichment determinations in field trials.
Percent Ndfa and Ndiff were calculated for Sierra Mixe varieties when Atom% 15N excess or Shoot N for the whole plant were significantly different from reference varieties as assessed by ANOVA and single-degree-of-freedom contrasts (p = 0.05), respectively. Values followed by asterisks are significantly different from the reference varieties based on single-degree-of-freedom contrasts (p < 0.05). Percent Ndfa and Ndiff were not calculated for reference (dashes). (Data at DOI: 10.6084/m9.figshare.6534545).