Loss-of-function of triacylglycerol lipases are associated with low flour rancidity in pearl millet [ Pennisetum glaucum (L.) R. Br.]

Abstract

Pearl millet is an important cereal crop of semi-arid regions since it is highly nutritious and climate resilient. However, pearl millet is underutilized commercially due to the rapid onset of hydrolytic rancidity of seed lipids post-milling. We investigated the underlying biochemical and molecular mechanisms of rancidity development in the flour from contrasting inbred lines under accelerated aging conditions. The breakdown of storage lipids (triacylglycerols; TAG) was accompanied by free fatty acid accumulation over the time course for all lines. The high rancidity lines had the highest amount of FFA by day 21, suggesting that TAG lipases may be the cause of rancidity. Additionally, the high rancidity lines manifested substantial amounts of volatile aldehyde compounds, which are characteristic products of lipid oxidation. Lipases with expression in seed post-milling were sequenced from low and high rancidity lines. Polymorphisms were identified in two TAG lipase genes (PgTAGLip1 and PgTAGLip2) from the low rancidity line. Expression in a yeast model system confirmed these mutants were non-functional. We provide a direct mechanism to alleviate rancidity in pearl millet flour by identifying mutations in key TAG lipase genes that are associated with low rancidity. These genetic variations can be exploited through molecular breeding or precision genome technologies to develop elite pearl millet cultivars with improved flour shelf life.

Keywords: Pennisetum glaucum (L.) R. Br.; lipase; milled flour; millet; pearl millet; rancidity; shelf life; triacylglycerol.

Figure 1

(A) Comparison of acid value (AV) of flour of 12 pearl millet inbred and germplasm association lines (PMiGAP) during 14-day accelerated storage conditions. The genotypes include ICMB-843 (I1); ICMB-88004 (I2); ICMB-95222 (I3); ICMB-81 (I4); ICMB-89111 (I5); ICMP-842 (I6); ICMB-863 (I7); ICMB-98222 (I8); IP5931 (P13); IP13840 (P14); IP 22419 (P19); IP6099 (P20). (B) Change in AV in the flour of selected contrasting lines on day 10 and 21. The data are represented as the mean ± standard error of three biological replicates. Variables showing F-value less than 0.05 were significant. Also, the means displaying non-matching lowercase letters were significantly different at p ≤ 0.05 by the Duncan test.

Figure 2

Lipid catabolism in milled flour under accelerated-aging conditions. (A) Fatty acid profiles obtained by GC-FID. (B) Total fatty acid quantitation by GC-FID, average over three replicates. Each brace corresponds to a Mann–Whitney test with the resulting value of p indicated above the brace. (C) TAG and FFA levels measured by HPLC-ELSD in two datasets (2019 and 2021). Log-linear hierarchical mixed models were utilized to assess the differences in TAG and FFA levels of these three inbred lines. Points represent observed data, lines represent posterior mean of the predictive distribution, and the ribbon corresponds to the 95% credible interval of the posterior mean predictive distribution.

Figure 3

Volatile analysis of pearl millet I3, I5 and I7 lines using Solid Phase Micro-Extraction Electron Impact Ionization Gas Chromatography Mass Spectrometry (SPME-EI-GC-MS). (A) First and second principal components for processed auto-scaled datasets for the I3, I5 and I7 lines (blue triangles, aquamarine inverted triangles and pink stars, respectively). A 95% confidence range is indicated by a dashed line. (B) Most abundant fragment m/z total peak intensity bar charts for lowest value of p aldehydes (C₅–C₁₀) for the I3, I5 and I7 lines. (C,D) Volcano plots generated from 2 groups sample comparison tests for I3 vs. I5 (C) and I3 vs. I7 (D). Most statistically significant aldehydes as measured by value of p are shown labeled. See Supplemental Table S5 for the detailed statistical analysis.

Figure 4

Phylogenetic relationships and expression analysis of the pearl millet lipases. (A) The phylogenetic tree constructed using MacVector software by the NJ method for PgTAGLips. Asterisks (^*) denotes the genes used for further analysis: Pgl_GLEAN_10023115 was renamed PgTAGLip1, Pgl_GLEAN_10024115 renamed PgTAGLip2, and Pgl_GLEAN_10011213 renamed PgTAGLip3. (B) Expression analysis of three selected PgTAGLips in the various developmental stages of pearl millet variety ICMB 9333. Data points represent the expression values, obtained after normalization against the reference genes EUKARYOTIC INITIATION FACTOR4Α (PgEIF4α) and MALATE DEHYDROGENASE (PgMDH). Each data point represents the mean of three biological replications with standard error (±SE; representing the mean of three technical replicates). (C) Expression of the selected PgTAGLips in contrasting genotypes (I3, I5, and I7) under accelerated storage of milled flour. The x-axis represents the identities of flour samples stored at 0 h, 1 h, 5 h, 24 h. The y-axis indicates the normalized gene expression values against the reference genes PgEIF4α and PgMDH. Each data point represents the mean of three biological replications with standard error (±SE; representing the mean of three technical replicates).

Figure 5

Structural variations of selected TAG lipases. (A) Comparative amino acid sequence alignment of Pgl_GLEAN_10023115 (PgTAGLip1) in high (I7, 86 M88, I5) and low rancidity genotypes (I3, Super Boss, RHB177). The LID domain is in bold, and the Lip serine active motif (GxSxG) is in bold and underlined. (B) Comparative amino acid sequence alignment of PgTAGLip2 in high rancid (I7 and I5) and low rancid (I3) lines. The LIP domain is bold and Lip serine active motif (GxSxG) is bold and underlined. (C) Predicted protein models and docking affinities of three PgTAGLips (PgTAGLip1, PglTAGLip2, and PgTAGLip3) with triolein and trilinolein. The predicted substrate binding is shown in a ball and stick model, with dashed pink lines denoting the hydrogen bonds. The putative LID domain GxSxG motif is in green, the catalytic triad is in yellow with ball and stick, the putative glycosylation sites are in red, and the ligand ball and stick is in pink.

Figure 6

Functional validation in yeast expression system and seedling growth of the contrasting pearl millet seeds lines during germination (A) GC-MS TAG analysis of yeast cells expressing TAG lipase variants. Each data point represents the mean of three biological replications with standard error (±SE; B) in vivo mobilization of TAG reserve in yeast triple lipase mutant expressing TAG lipase variants from 2 h to 6 h. Each data point represents the mean of two biological replications with standard error (±SE). (C) Seedling establishment of I3 and I7 lines. (D) TAG mobilization during germination and post-germinative growth. Each data point represents the mean of three biological replications with standard error (±SE).