henhao
Sign in
Plant BiologybioRxiv
Preprint

Zea Lip: An atlas of glycerolipid profiles across leaf development in maize

Karla A. Juárez-NúñeziD, Guillaume LobetiD, Nirwan TandukariD, Eric Jadidzadeh, Asher PashaiD, Nicholas James ProvartiD, James Brendan HollandiD, Rubén Rellán‐ÁlvareziD, Allison C. BarnesiD

PDFDOI
In-depth full-text digest available · Jump to digest ↓

Abstract

Lipids are the predominant building blocks of plant membranes and are essential for plant growth and development. They are crucial for survival during times of stress as lipids are involved in multiple signaling pathways, and their relative abundances can change in response to environmental factors. To better characterize the lipid composition of the vital food crop maize, we generated a comprehensive glycerolipid atlas using ultra-high-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry. We surveyed the lipid profiles of three different maize genotypes: B73, a temperate inbred; CML312, a subtropical inbred; and Palomero Toluqueno, an open-pollinated variety from the Mexican highlands. We collected leaf samples from 4 developmental stages and 6 leaves. From one growth stage, we also sampled along with three leaf zones: base, center, and tip. The genotype and leaf number were the major drivers of lipid differences. Phosphatidylcholine, lysophosphatidylcholine, and triacylglycerol genotypic differences were particularly high. We generated an eFP browser to be integrated into the maize genome browser, as well as a separate web interface to easily browse and compare lipid levels across tissues and genotypes, available at https://rrellan.shinyapps.io/Zea-Lip/.

02 /Specification & Classification

Click any tag to browse matching papers
Citations
0
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (28 Oct 2026) · CC BY 4.0

UHPLC-MS/MS lipidomics of maize leaves (149 species, three genotypes, four growth stages, three leaf zones) shows genotype and leaf number explain most lipid variation.

Plant lipid profiles shift with tissue, development, genotype and environment, so each published lipidome is effectively a single snapshot, and earlier maize leaf lipid work focused on seedlings, chloroplast differentiation and base-to-tip gradients. This resource extends sampling across plant ages, leaf positions and within-leaf zones in genotypes of contrasting origin, and shares the data through a Shiny App, an eFP browser and MetaboLights.

Key results

  1. ShownBuilt a maize leaf lipid reference of 149 annotated species in 14 LIPID MAPS classes (glycerolipids plus sphingolipids, acylcarnitines and a cholesterol ester) spanning genotype, plant age, leaf position and leaf zone, using a retention-time-corrected library expanded with MS/MS evidence.

    Over 19,000 mass features searched against the m/z-rt library and LipidBlast yielded 149 lipids in 14 classes; 339 MS/MS-supported annotations grew the library from 501 to 840 species; 50 plants gave 151 samples; TG was the most abundant class (42 species). · Plant growth and tissue collection; Data processing and analysis; Separation, annotation, and identification of…

  2. ShownGenotype is the main axis of PCA separation. B73 and highland Palomero Toluqueño (PT) sit at opposite ends, with subtropical CML312 in between, driven largely by phosphatidylcholine (PC), lysophosphatidylcholine (LPC) and triacylglycerol (TG).

    PC1 explains 31.7% and PC2 16.8% of variance across all lipid species, and genotype separates samples on both. PC sums are high in PT and low in B73, LPC is the reverse, and CML312 is intermediate. The authors relate this to the HPC1 phospholipase A1 locus (a highland introgression lowers PC-to-LPC hydrolysis), but HPC1 was not assayed here. · Maize lipid profiles differ by genotype and leaf…; Figure legends

  3. ShownUsing all planting dates, genotype and leaf number were the factors most often kept in per-lipid stepwise linear models. Developmental stage came next, and interaction terms were rarely kept.

    Of 149 lipids, final models kept genotype for 79%, leaf number for 77% and developmental stage for 50%; two- and three-way interactions were kept for 23% or less. By share of sums of squares, genotype, leaf and stage drove most of the variance for most lipids. · Effects of genotype, developmental stage, and…; Maize lipid profiles differ by genotype and leaf…; Figure legends

  4. SuggestedTriacylglycerols were the lipids most tied to interaction effects: only TG species had leaf-by-genotype or leaf-by-stage as their top factor. This suggests that TG turnover is structured by leaf position and plant age.

    Lipids for which the leaf-by-stage interaction occasionally explained a large share of variance were overwhelmingly TG species. The stage main effect was the top factor for a near-even mix of TG and galacto/sulfolipids. The authors cite TG as a transient fatty acid buffer; vegetative TG is typically under 0.1% dry weight. · Maize lipid profiles differ by genotype and leaf…

  5. SuggestedLeaf zone affected many lipids but rarely dominated, and base-to-tip gradients depended on genotype. PT stayed flat, while B73 and CML312 trended toward less monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG) but more phosphatidylglycerol (PG) in tips than in bases.

    In plants from the first planting date (58 days), zone was the top factor for only 13 of 149 lipids but was significant for 84. MGDG, DGDG, PG, TG and PC each had at least one species where zone was the top factor. Unlike earlier reports, tips did not consistently hold the most MGDG, DGDG and PG. · Effect of the leaf zone; Lipid Changes Across Leaf Zones; Figure legends

Methods, data and reproducibility

Three maize genotypes sown in four staggered greenhouse blocks were harvested on one day, sampling leaf blades at four plant ages, six leaf positions and three within-leaf zones. Positive-mode UHPLC-QTOF MS/MS lipid profiles (normalized with SERRF and total lipid signal) were analyzed by PCA and per-lipid linear models chosen by forward selection on Akaike's criterion.

  • Leaf blade lipid extracts from B73, CML312 and Palomero Toluqueño grown in four staggered greenhouse blocks; plants aged 28, 40, 50 and 58 days; leaves 4, 5, 7, 9, 11, 13 · Zea mays (maize) · n = 50 plants; 151 samples · MetaboLights MTBLS15840
  • Leaf-zone subset (base, center, tip) from the first planting date (58-day plants), used for zone modeling · Zea mays (maize)
  • Ionization-mode optimization set: leaf samples of B73, PT and B73 x PT recombinant inbred lines (RILs), MS1 acquired in positive and negative modes · Zea mays (maize) · n = 8 B73, 8 PT, 10 RIL samples
  • Study samples rerun across four m/z windows to check for false-positive lipid annotations · Zea mays (maize) · n = 36 samples
  • In-house MS1 m/z-retention time lipid library, expanded by MSDIAL fragment matching to LipidBlast (Tables S3, S5) · n = 501 expanded to 840 lipid species
Methods & tools
UHPLC-QTOF MS/MS: Agilent 1290 UHPLC, Agilent 6530 QTOF, Waters Acquity CSH C18 column, Matyash MTBE/methanol lipid extraction, Agilent MassHunter Qualitative Analysis B.06.00, MSDIAL 3.40, LipidBlast, MS-FLO, SERRF (Systematic Error Removal using Random Forest), Reifycs Abf converter, stepAIC (AIC-based forward selection) and the emmeans package, R stats prcomp (PCA) and ggplot2, LIPID MAPS classification system, Shiny App and eFP browser for data access
Reagents & environment
Agilent MassHunter Qualitative Analysis B.06.00; MSDIAL version 3.40; MS-FLO; SERRF; stepAIC and emmeans; R stats prcomp and ggplot2 (R and package versions not stated).
Reproducibility gaps
The paper gives no URLs for the Shiny App or eFP browser and shares no analysis scripts; raw files are on MetaboLights (MTBLS15840). Some design combinations come from a single plant.
Limitations (5)
  • Sampling breadth was prioritized over replication, so some genotype, stage, leaf and zone combinations come from a single plant. Values for individual design cells are hypothesis-generating; statistical inference relies on the whole design. · stated by the authors
  • All samples were collected in one late-morning window, so the profiles capture a single moment in a circadian and environmentally regulated cycle. Sampling across the day-night cycle is needed to resolve temporal lipid remodeling. · stated by the authors
  • Leaf-zone patterns differed from earlier maize reports (tips were not consistently richest in MGDG, DGDG and PG). The authors suggest plant age, genotypes, growth conditions or the definition of the leaf base as possible causes. · stated by the authors
  • Only positive ionization mode was analyzed. In the separate optimization set, negative mode detected 16 species (free fatty acid (FA) and PC) with no overlap with positive mode, so free fatty acids and other negative-mode lipids are likely missing from the atlas. · noted in this summary
  • The link from PC/LPC differences to HPC1 or highland adaptation is inferred from three genotypes in one greenhouse. HPC1 alleles and phospholipase activity were not measured, and with one genotype per origin, origin is confounded with genetic background. · noted in this summary

Who should read:Read this if you study maize or plant glycerolipid metabolism and want a leaf lipidome resolved by genotype, leaf and stage, with Shiny and eFP browsers, to generate hypotheses on PC/LPC or galactolipid remodeling.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISBourgault '19Alexander '20Zhou '19Cassim '20Barnes '21Devadasu '25
59% semantic matchPlant Biology
R. Bourgault, S. Matschi, M. Vasquez+7 more·bioRxiv·May 2, 2019