Tissue specific diurnal rhythms of metabolites and their regulation during herbivore attack in a native tobacco, Nicotiana attenuata

Abstract

Ecological performance is all about timing and the endogenous clock that allows the entrainment of rhythms and anticipation of fitness-determining events is being rapidly characterized. How plants anticipate daily abiotic stresses, such as cold in early mornings and drought at noon, as well as biotic stresses, such as the timing of pathogen infections, is being explored, but little is known about the clock's role in regulating responses to insect herbivores and mutualists, whose behaviors are known to be strongly diurnally regulated and whose attack is known to reconfigure plant metabolomes. We developed a liquid chromatography-mass spectrometry procedure and analyzed its output with model-based peak picking algorithms to identify metabolites with diurnal accumulation patterns in sink/source leaves and roots in an unbiased manner. The response of metabolites with strong diurnal patterns to simulated attack from the specialist herbivore, Manduca sexta larvae was analyzed and annotated with in-house and public databases. Roots and leaves had largely different rhythms and only 10 ions of 182 oscillating ions in leaves and 179 oscillating ions in roots were rhythmic in both tissues: root metabolites mainly peaked at dusk or night, while leaf metabolites peaked during the day. Many oscillating metabolites showed tissue-specific regulation by simulated herbivory of which systemic responses in unattacked tissues were particularly pronounced. Diurnal and herbivory-elicited accumulation patterns of disaccharide, phenylalanine, tyrosine, lyciumoside I, coumaroyl tyramine, 12-oxophytodienoic acid and jasmonic acid and those of their related biosynthetic transcripts were examined in detail. We conclude that oscillating metabolites of N. attenuata accumulate in a highly tissue-specific manner and the patterns reveal pronounced diurnal rhythms in the generalized and specialized metabolism that mediates the plant's responses to herbivores and mutualists. We propose that diurnal regulation will prove to an important element in orchestrating a plant's responses to herbivore attack.

Figure 1. Experimental procedures used to identify oscillating and herbivore-induced metabolites and their associated genes in different tissues of Nicotiana attenuata.

(A) Wild type (WT) N. attenuata plants were harvested every 4 h for two days during the initiation of stem elongation. To mimic herbivory, oral secretions (OS) of the larvae of the specialist herbivore, M. sexta, were immediately applied to puncture wounds made in leaves at 1 pm. Water treatment of puncture wounds on separate plants was used to distinguish OS-specific from wound-induced changes in metabolites and transcripts. (B) Metabolites from three different tissues, source leaves, sink leaves, and roots of N. attenuata were isolated. The leaf at node 0 had completed the sink to source transition and the leaf at node +1 was older by one leaf position than the leaf at node 0 and so forth. Source leaves (at nodes +2, +1, 0) were wounded with a fabric pattern wheel and treated with 20 µl of M. sexta OS, which was diluted 1∶5 with water. Untreated leaves (at nodes −1, −2) and roots were harvested to monitor systemic responses. (C) After sample preparation from six biological replicates, a 40% methanol extraction method optimized for the defense metabolites of N. attenuata was used and the metabolites separated with a rapid separation liquid chromatography (RSLC) on a C₁₈ column and detected by ESI-TOF-MS (electrospray ionization time-of-flight mass spectrometer) for parents and their daughter ions. Peak picking and alignments were performed with the XCMS package . Diurnal oscillating metabolites were extracted by the pattern matching algorithms of HAYSTACK tool . In-house and public databases were used to identify oscillating metabolites and a 44K Agilent microarray designed for N. attenuata was used to examine the expression of metabolite-related genes.

Figure 2. Accumulation of oscillating metabolites in N. attenuata show tissue-specific rhythms.

Oscillating metabolites in source leaves (A) and roots (C) were roughly divided into two groups, one peaking during the day, and the other peaking at night. The heat map displays all of the Z-transformed oscillating metabolites levels in a false-color scale where green indicates low and red indicates high values. Each metabolite is represented by a single row with the average linkage hierarchical clustering tree obtained using Euclidean distances as metric. We counted the number of ions (y-axis) that peak at a particular harvest times (x-axis) in source leaves (B) and roots (D). Gray boxes depict the dark period. (E) Venn diagram of the oscillating metabolites selected across source leaves (black solid line) and roots (gray dashed line). (F) Venn diagram of the oscillating metabolites after removing adduct and daughter ions. CAMERA package and Pearson correlation were used to select adduct and daughter ions detected from extracts of source leaves (black solid line) and roots (gray dashed line).

Figure 3. Accumulation of disaccharides and sugar related genes in three different tissues.

(A) Mean (±SE) levels of normalized intensity of disaccharides (m/z 341.11 at 90 s, C₁₂H₂₁O₁₁ ⁻) in source leaves, sink leaves and roots at each harvest time for two days (gray dashed lines) in control (Con) plants. After wounding and treating puncture wounds with either water (W+W, dashed lines with colors) or M. sexta OS (W+OS, solid lines with colors), disaccharides levels were examined in treated leaves (red), untreated systemic leaves (blue) and roots (green). Gray boxes depict the dark period. Asterisks indicate significant differences among the treatments at the indicated harvest time (* = P<0.05, one-way ANOVA with Bonferroni post hoc test). (B) Schematic overview of sucrose (suc) metabolism. (C) Two diurnal patterns of sugar metabolism-related genes (Table S2) accumulation in source leaves. Gray boxes depict the dark period. Sweet: sugar exporter . Ratio to average: Ratio of transcript abundance at the time point shown, to the mean abundance of the same transcript across all time points. (D) Volcano plot analysis of the transcript levels of sugar-related genes in treated leaves (red dot), systemic untreated leaves (blue dot) and roots (green dot) 1 and 5 h after M. sexta OS-elicitation. The log₂ ratio of mean intensities (OS-elicited/Con, with microarray expression data) plotted against the negative log₁₀-transformed P value derived from Student's t tests. The horizontal dashed line indicates the threshold for statistically significant expression at P = 0.05 and the vertical dashed line, the threshold for two-fold changes in gene expression.

Figure 4. Diurnal rhythms and OS elicitation of glucose-containing secondary metabolites in different tissues.

Mean (±SE) levels of normalized intensity of lyciumoside I (m/z 629.35 at 339 s, C₃₂H₅₃O₁₂ ⁻) in treated (A), and untreated systemic leaves (C) at each harvest time for two days (gray dotted lines) in control plants. Lyciumoside I was not detected in roots. After W+W (dashed lines with colors) or W+OS (solid lines with colors) treatments, lyciumoside I levels were examined in treated leaves (A) and untreated systemic leaves (C). Gray boxes depict the dark period. Asterisks indicate significant differences among the treatments at the indicated time points (* = P<0.05, one-way ANOVA with Bonferroni post hoc test). (B), (D) Effects of W+W and W+OS on relative transcript abundance (±SE) of NaGGPPS (N. attenuata geranylgeranyl diphosphate synthase), a gene involved in producing the diterpenoid precursor, geranylgeranyl diphosphate . Different letters (a, b and c) reflect significant differences among the treatments at the indicated time points (P<0.05, one-way ANOVA with Bonferroni post hoc test).

Figure 5. Diurnal rhythms and OS elicitation of Phe and Tyr and their related genes in different tissues.

(A) Schematic overview of Phenylalanine (Phe) and Tyrosine (Tyr) metabolism. CM, chorismate mutase; ADT, arogenate dehydratase; PAL, phenylalanine ammonia lyase; C4H, cinnamate 4-hydroxylase; 4CL, 4-coumarate-coa ligase; TyrA, arogenate dehydrogenase; THT, tyramine N-hydroxycinnamoyltransferase; SAMDC, S-adenosylmethionine decarboxylase; SPDS, spemidine synthase. (B) Diurnal expression of genes encoding Phe or Tyr metabolism enzymes in source leaves. Gray box depicts the dark period. Ratio to average: Ratio of transcript abundance at the time point shown, to the mean abundance of the same transcript across all time points. Mean (±SE) levels of normalized intensity of Phe (C; m/z 164.07 at 192 s, C₉H₁₀NO₂ ⁻) and Tyr (G; m/z 180.07 at 144 s, C₉H₁₀NO₃ ⁻) in treated leaves, untreated leaves and roots at each harvest time for two days (gray dashed lines) in control plants. After W+W (dashed lines with colors) or W+OS (solid lines with colors) treatments, Phe (C) and Tyr (G) levels were quantified in treated leaves (red), untreated systemic leaves (blue) and roots (green). Effects of W+W and W+OS treatments on relative transcript abundance (±SE) of genes related Phe (D–F) and Tyr (H–J) metabolism. Gray box depicts the dark period. Different symbols (* and #) indicate significant differences among the treatments at the indicated time point (P<0.05, one-way ANOVA with Bonferroni post hoc test).

Figure 6. OS-elicitation affects secondary metabolites in the phenylpropanoid pathway.

Mean (±SE) levels of normalized intensity of coumaroyl tyramine (A; m/z 284.10 at 165 s, C₁₇H₁₈NO₃ ⁺), feruloyl putrescine (B; m/z 265.15 at 212 s, C₁₄H₂₁N₂O₃ ⁺) and N-feruloyl tyramine (C; m/z 314.14 at 319 s, C₁₈H₂₀NO₄ ⁺) in treated and systemic leaves at each harvest time for two days (gray dotted lines) in control plants. Feruloyl putrescine and N-feruloyl tyramine were not detected in roots. After W+W (dashed lines) or W+OS (solid lines) treatments their levels were examined in treated leaves (red) and untreated systemic leaves (blue). Gray boxes depict the dark period. Asterisks indicate significant differences among the treatments at the indicated time point (* = P<0.05, one-way ANOVA with Bonferroni post hoc test).

Figure 7. Diurnal rhythms and OS elicitation of OPDA, jasmonic acid and JA-related genes in roots.

(A), (B) Mean (±SE) levels of normalized intensity of 12-oxophytodienoic acid (OPDA, m/z 291.20 at 479 s, C₁₈H₂₇O₃ ⁻) and jasmonic acid (JA, m/z 209.17 at 368 s, C₁₂H₁₇O₃ ⁻) in treated leaves and roots at each harvest time for two days (gray dotted lines) in control plants. After W+W (dashed lines with colors) or W+OS (solid lines with colors) treatments, their levels were examined in treated leaves (red) and roots (green). Gray boxes depict the dark period. Different symbols (* and #) indicate significant differences among the treatments at the indicated time point (P<0.05, one-way ANOVA with Bonferroni post hoc test). (C) Diurnal rhythms of gene accumulation involved in JA biosynthesis or signaling. AOS, allene oxide synthase; JAZ, jasmonate-ZIM-domain protein; COI1, coronatine insensitive 1.