In this study, we used liquid chromatography high-resolution tandem mass spectrometry to analyze the lipidome of turtlegrass (Thalassia testudinum) leaves with either extremely high phosphorus content or extremely low phosphorus content. Most species of phospholipids were significantly down-regulated in phosphorus-deplete leaves, whereas diacylglyceryltrimethylhomoserine (DGTS), triglycerides (TG), galactolipid digalactosyldiacylglycerol (DGDG), certain species of glucuronosyldiacylglycerols (GlcADG), and certain species of sulfoquinovosyl diacylglycerol (SQDG) were significantly upregulated, accounting for the change in phosphorus content, as well as structural differences in the leaves of plants growing across regions of varying elemental availability. These data suggest that seagrasses are able to modify the phosphorus content in leaf membranes dependent upon environmental availability.
Citation: Koelmel JP, Campbell JE, Guingab-Cagmat J, Meke L, Garrett TJ, Stingl U (2019) Re-modeling of foliar membrane lipids in a seagrass allows for growth in phosphorus-deplete conditions. PLoS ONE 14(11): e0218690. https://doi.org/10.1371/journal.pone.0218690
Editor: Silvia Mazzuca, Università della Calabria, ITALY
Received: June 17, 2019; Accepted: October 10, 2019; Published: November 27, 2019
Copyright: © 2019 Koelmel et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Seagrasses are a widely distributed group of marine plants that provide a range of ecological services to coastal habitats around the world. However, due to a variety of natural and anthropogenic stressors, many seagrass beds are declining globally. For over 30 years, it has been known that many seagrass species can display shifts in foliar phosphorus (P) content in response to environmental availability[2,3], an adaptation that allows for growth in a wide range of habitats of varying nutrient regimes. In Thalassia testudinum (turtlegrass), a dominant species in South Florida, elemental C:P ratios can differ by nearly 8-fold, from around 400 to 3000, often dependent upon environmental P availability[3,5,6]. Thalassia testudinum is distributed along the western Atlantic from Florida, USA to Venezuela, throughout the Gulf of Mexico and the Caribbean Sea. Thalassia hemprichii, the other species in this genus, is also widely distributed in the coastal waters of the Indian Ocean and the western Pacific. While the morphology of turtlegrass leaves and canopy structure changes with decreased P content, areal production rates can remain relatively high, indicating metabolically active plants. Changes in C:P ratios and P content of turtlegrass occur along natural P gradients[4,6,9], but can also be induced by fertilization experiments in P-depleted habitats[5,10]. The exact cellular mechanisms on how turtlegrass lowers its P content are mostly unknown. In this study, we used liquid chromatography high-resolution tandem mass spectrometry (LC-HRMS/MS) to analyze the lipidome of turtlegrass leaves that contained either a high percentage of P or a low percentage of P as a result of a fertilization experiment.
Materials and methods
In 2014, samples of Thalassia testudinum leaf tissue were collected in Key Largo, Florida, as part long-term fertilization experiment (see Campbell et al 2018 for full details). In brief, 60 replicate 0.25 m2 plots were established in a shallow (1 m deep) T. testudinum meadow in Largo Sound (25° 7.58’ N, 80° 24.29’ W). Thirty of these plots were haphazardly selected to receive amendments of slow release fertilizer. Nutrient enriched plots received 350 g of slow release Osmocote fertilizer (NPK 14-14-14) enclosed in fiberglass mesh bags. Each bag was attached to a PVC post positioned in the center of the plot and was replaced every 4 weeks to ensure consistent nutrient delivery. After 14 weeks, 3–4 separate shoots were harvested from each plot and transported to the laboratory on ice. Shoots from each plot were pooled, rinsed in deionized water, scraped free of epiphytic growth, and dried to a constant weight at 60°C. All samples were then pulverized, homogenized and stored in boroscillicate glass vials. Nitrogen content (% dry mass) was measured via CHN analysis (Fisons NA1500 elemental analyzer). Phosphorus (P) content (% dry mass) was measured colorimetrically after a dry-oxidation, acid hydrolysis extraction procedure. Plots displaying the highest and lowest P content were then selected for further lipidome analysis. Elemental C, N, P contents of leaves of selected plots are shown in Table 1.
Samples were extracted using the folch extraction procedure. Briefly, 25 mg of seagrass was weighed and 10 μL of 10x diluted internal standard mixture (stock solution of 50 ppm, w:v) was added. Internal standards were purchased from Avanti Polar Lipids, inc. (Alabaster, AL, USA) and consisted of: LPC (17:0), PC (17:0/17:0), PG (14:0/14:0), PE (15:0/15:0), PS (14:0/14:0), PI (8:0/8:0), SM (d18:1/17:0), Cer (d18:1/17:0), DG (14:0/14:0), CL (15:0_15:0_15:0_16:1), Sphingosine (d17:1), PAzePC, Glucosyl (β) Cer (d18:1/12:0), BMP (14:0/14:0) (S,R), and LSM (d17:1), except for TG (15:0/15:0/15:0), which was obtain from Nu-Chek (Elysian, MN, USA). Samples were extracted using 1:2:4 water:methanol:chloroform (v:v:v), and the organic phase was collected, dried down, and reconstituted in 75 μL of isopropanol plus 1 μL of injection standard mixture (100 ppm, w:v). Injection standards were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL, USA) and consisted of: LPC (19:0), PC (19:0/19:0), PG (17:0/17:0), PE (17:0/17:0), PS (17:0/17:0), and TG (17:0/17:0/17:0). Extraction blanks (without internal standard), neat quality controls (QCs, blanks with internal standards), solvent blanks, and Red Cross plasma for QC purposes were also prepared.
Data was acquired using high-performance liquid chromatography high-resolution tandem mass spectrometry (LC-HRMS/MS). Chromatographic separation was achieved using reverse phase chromatography (Dionex Ultimate 3000 RS UHLPC system, Thermo Scientific, San Jose, CA, USA) with a Waters Acquity C18 BEH column maintained at 30°C (2.1 × 100 mm, 1.7 μm particle size, Waters, Milford, MA, US). The gradient (S1 Table) consisted of solvent A (60:40 acetonitrile:water) and solvent B (90:8:2 isopropanol:acetonitrile:water), both with 10 mM ammonium formate and 0.1% formic acid. The flow rate was 500 μL/min. Ammonium formate is not only needed for separation, but also for ionization of neutral lipids as [M+NH4]+ in electrospray ionization.
For acquiring mass spectra and MS/MS, a Q-Exactive orbitrap (Thermo Scientific, San Jose, CA) was used. Mass spectral parameters are shown in S2 and S3 Tables. The sequence consisted of three blanks followed by a neat QC, and one blank and QCs inserted between every 10 samples. Data was acquired for six low P containing seagrass samples and six high P containing seagrass samples injected at 2 μL in positive ion mode, and 4 μL in negative ion mode. Both data-dependent (ddMS2-top10) and all-ion fragmentation (AIF) data were obtained on two samples per group for identification purposes. In addition, full-scan data was acquire for all 12 samples without MS/MS for comparing lipid intensities across groups.
LipidMatch Flow was used for file conversion, peak picking (implementing MZMine 2), blank filtration[14,15], lipid annotation, and combining positive and negative datasets. LipidMatch Flow software and tutorials (including video tutorials) can be found at <http://secim.ufl.edu/secim-tools/>. In addition to LipidMatch annotation, MS-DIAL annotations were appended to the feature table obtained from LipidMatch Flow using an in-house R script. MS-DIAL was only used to identify lipids using data-dependent analysis while LipidMatch was used to annotate ions using both all-ion fragmentation (AIF) data and data-dependent analysis.
Multivariate statistical analysis was performed using Metaboanalyst 3. Raw intensity values were normalized by sum, log transformed, and mean centered. Principal component analysis (PCA) was performed on the resulting normalized lipid values, with samples color coded by low P and control. For univariate statistics a two-tailed heteroscedastic t-tests was performed on low P versus control samples (peak areas). To account for multiple comparison errors, the Benjamini–Hochberg method was used to obtain false discovery rate (FDR) corrected p-values. Features with lipid annotations and an FDR corrected p-value less than 0.05 were considered significant. To determine trends across lipid classes, a Fisher's exact test was performed using an in-house R script. For the Fisher's exact test, features were considered upregulated with a lipid peak area fold change greater than 1.5 and downregulated with a fold change less than 0.67. In either case, features were only included if the FDR corrected p-value was less than 0.2. Lipid features and their respective classes according to this inclusion criteria were highlighted in a volcano plot color coded by lipid class. Fisher's exact test was used to calculate p-values based on whether lipids of a certain class tended to be more significantly upregulated or downregulated compared to lipids across all classes. Therefore, the lipid classes with the most significant change between low P and controls was determined.
Results and discussion
In total, 600 unique molecular lipid species across 36 lipid classes (S5 Table) were tentatively annotated using exact mass and MS/MS information by LipidMatch Flow[21,22]. Acronyms of the lipid classes covered in this manuscript are defined in S5 Table. The total lipidome of the samples grouped based on foliar P content without any exceptions (Fig 1).
Data was normalized by sum, log transformed and mean centered. PC1 explained 43.5% of the variation, and PC2 explained 13.9% of the variation.
Most classes of phospholipids were significantly down-regulated in P-depleted leaves including PC and PE, which have been reported as the most abundant phospholipids in three species of seagrasses, whereas diacylglyceryltrimethylhomoserine (DGTS), triglycerides (TG), galactolipid digalactosyldiacylglycerol (DGDG), certain species of glucuronosyldiacylglycerols (GlcADG), and certain species of sulfoquinovosyl diacylglycerol (SQDG) were significantly upregulated (Table 2, S1 Fig, S4 and S5 Tables) and presumably replace phospholipids in the membranes.
Structures of certain upregulated and downregulated lipids are shown in Fig 2. It is interesting to note that total DGTS had the greatest fold change increase in low P, as compared to other non-phosphorus containing membrane lipids, suggesting partial replacement of the dominant PC membrane lipid. Substitution of phospholipids by non-phosphate containing lipids was first reported in Proteobacteria, where glycolipids replaced a large part of phospholipids in Pseudomonas diminuta so dramatically that in P-limited cultures, phosphate lipids were barely detectable (< 0.3% of total polar lipids). Since this landmark discovery, several lipid classes have been identified in a variety of diverse organisms to be involved in membrane lipid reconstructions during P starvation: SQDG was detected to substitute for phospholipids and thus to reduce P needs in Arabidopsis and certain species of picocyanobacteria[25–27]. Similar modifications of membrane lipids, but with DGDG replacing phospholipids, have been reported in oat as well as in seven other species of monocots and dicots. DGTS (a P-free betaine-lipid analog of PC) has been reported to replace PC in fungi. So far, the only study revealing that membrane re-modeling is an important adaptation to low P concentrations in environmental mixed communities was reported for phytoplankton communities in the Sargasso Sea.
Lipid acronyms are defined as follows: phosphatidylcholine (PC), phosphatidic acid (PA), phosphatidylethanolamine (PE), diacylglyceryltrimethylhomoserine (DGTS), glucuronosyldiacylglycerols (GlcADG), and galactolipid digalactosyldiacylglycerol (DGDG). *for GlcADG only a few molecular lipid species were significantly upregulated.
Using LC-tandem MS and LipidMatch Flow software[21,22], we were able to identify that not all molecular species in a given lipid class showed the same trend and thus the data in Table 2 only shows a simplistic overview of the changes in lipid composition. Under P-depleted conditions, the most significantly upregulated lipid species in Thalassia in terms of fold-change were actually GlcADG (S1 Fig, S5 Table), which were only recently discovered in the context of P starvation in Arabidopsis. Specifically GlcADG(16:0_18:2), fold change of 21, GlcADG(16:0_16:0), fold change of 7, and GlcADG (18:0_18:2), fold change of 7, were significantly higher under P-deplete conditions compared to high P (S2 Fig). Interestingly, of the twelve GlcADG molecular species that were identified, only four were significantly upregulated (Hochberg corrected p-value < 0.05) and only the three listed above had fold changes above 2 (S5 Table). S2 Fig shows examples of three GlcADG species identified by both MS-DIAL and LipidMatch, which had greatly differing fold changes. This impressively illustrates the use of MS and the urgent need for the identification of single molecular lipid species over other techniques that only analyze lipid classes (e.g. 2-D TLC), and explains why GlcADGs are not included in Table 2, which only shows average changes in lipid classes.
Similar to GlcADGs, TGs were highly upregulated in P-deplete Thalassia leaves. TGs were also upregulated in nitrogen studies in the alga Chlamydomonas reinhardtii. In general, in starvation conditions, membrane phospholipids are expected to decrease due to a shift towards TG synthesis as well as due to replacement by DGTS and DGDG. We found that a significant number of DGDG species increased in P-deficient seagrass leaves (S4 and S5 Tables). Other lipids, which were downregulated under P-deplete conditions, were diglycerides (DG) and ceramides (Cer-NS), (Table 2, S4 and S5 Tables). DGs are involved in DGTS, DGDG, and TG synthesis, all of which were upregulated in P-deficient Thalassia leaves. Still, more research is needed to understand the downregulation of DG and Cer-NS in P-deficient seagrass plants.
While the majority of the 32 SQDG species that were identified had fold changes greater than one (indicating upregulation; 27/32), only two were found to be significantly upregulated (Hochberg corrected p-value < 0.05), namely SQDG (16:0_18:4) and SQDG (40:11) (S5 Table). Therefore, according to our study, SQDG had only minor to no upregulation in concentration compared to TG, DGDG, DGTS, and certain GlcADG species and does not seem to play a major role in remodeling of foliar membrane lipids under different P concentrations. While we cannot completely exclude that some of the 600 detected lipids originate from epiphytes or microbial (endo)symbionts that were not completely removed by our washing steps, we are certain that the decrease in P-containing lipids reflect actual changes in the seagrass lipidome as phosphatidylcholine (PC), phosphatidic acid (PA), and phosphatidylethanolamine (PE) have previously been reported to be the main P-lipids in seagrasses.
In conclusion, we present evidence of a key cellular mechanism employed by a widely distributed marine plant to thrive in nutrient-poor, oligotrophic conditions. These results not only explain the cellular mechanisms driving variability in turtlegrass P content, but also may potentially explain broader shifts in leaf structure or morphology under P-limitation, as membrane fluidity may be heavily influenced by lipid re-modelling. Understanding the biology of seagrasses and their adaptation to changing nutrient concentrations can help in conservation efforts. The lipid composition of seagrasses could be used as a biomarker to identify long-term nutrient limitation, which might not be detectable from periodic monitoring of nutrient concentrations in the surrounding waters.
S1 Fig. Volcano plot (low phosphorus versus high phosphorus) colored by lipid class.
S2 Fig. Box pots of GlcADG identified in negative polarity in both MS-DIAL and LipidMatch (raw data prior to normalization).
Boxplots show that the degree/occurrence of upregulation was quite different depending on the GlcADG fatty acyl species. In bold is the fold change (low / high P). A) GlcADG(16:0_16:0) [M-H]− (7), B) GlcADG(16:0_18:1) [M-H]− (2), C) GlcADG(16:0_18:2) [M-H]− (20).
S1 Table. Reverse phase liquid chromatography mobile phase gradient.
S2 Table. Heated electrospray (HESI) source parameters.
S3 Table. Q-Exactive mass spectrometers scan parameters.
S4 Table. Upregulated, downregulated, and unchanged lipid classes based on a fisher's exact test (see methods for details) and t-test based on the sum of species total intensities for each lipid class.
Based on a fisher's exact test: upregulated means a significant number of lipids in a given class were higher in P deficient substrates, while downregulated means that a significant number of lipids in a given class were lower in P deficient substrates as compared to controls. For total intensity and t-tests, upregulated or downregulated mean the total intensity of that lipid class was higher or lower, respectively, in low P versus high P leaves. Bold p-values are considered significant, while highlighted red values are highly significant. Acronyms are defined in S5 Table.
Publication of this article was funded in part by the University of Florida Open Access Publishing Fund. This is contribution #170 from the Center for Coastal Oceans in the Institute of Water and Environment at Florida International University.
- 1. Waycott M, Duarte CM, Carruthers TJB, Orth RJ, Dennison WC, Olyarnik S, et al. Accelerating loss of seagrasses across the globe threatens coastal ecosystems. PNAS. 2009;106: 12377–12381. pmid:19587236
- 2. Atkinson MJ, Smith SV. C:N:P ratios of benthic marine plants1. Limnology and Oceanography. 1983;28: 568–574.
- 3. Duarte CM. Seagrass nutrient content. Marine Ecology Progress Series. 1990;67: 201–207.
- 4. Fourqurean JW, Zieman JC, Powell GVN. Phosphorus limitation of primary production in Florida Bay: Evidence from C:N:P ratios of the dominant seagrass Thalassia testudinum. Limnology and Oceanography. 1992;37: 162–171.
- 5. Campbell JE, Altieri AH, Johnston LN, Kuempel CD, Paperno R, Paul VJ, et al. Herbivore community determines the magnitude and mechanism of nutrient effects on subtropical and tropical seagrasses. Van Alstyne K, editor. Journal of Ecology. 2018;106: 401–412.
- 6. Barry SC, Jacoby CA, Frazer TK. Environmental influences on growth and morphology of Thalassia testudinum. Marine Ecology Progress Series. 2017;570: 57–70.
- 7. The IUCN Red List of Threatened Species. In: IUCN Red List of Threatened Species [Internet]. [cited 28 Jan 2019]. Available: https://www.iucnredlist.org/en
- 8. Larkum AWD, Orth RJ, Duarte C, editors. Seagrasses: Biology, Ecology and Conservation [Internet]. Springer Netherlands; 2006. Available: //www.springer.com/us/book/9781402029424
- 9. Fourqurean JW, Jones RD, Zieman JC. Process Influencing Water Column Nutrient Characteristics and Phosphorus Limitation of Phytoplankton Biomass in Florida Bay, FL, USA:Inferences from Spatial Distributions. Estuarine, Coastal and Shelf Science. 1993;36: 295–314.
- 10. Armitage AR, Frankovich TA, Fourqurean JW. Long-Term Effects of Adding Nutrients to an Oligotrophic Coastal Environment. Ecosystems. 2011;14: 430–444.
- 11. Fourqurean JW, Zieman JC, Powell GVN. Relationships between porewater nutrients and seagrasses in a subtropical carbonate environment. Marine Biology. 1992;114: 57–65.
- 12. Folch J, Lees M, Sloane Stanley GH. A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem. 1957;226: 497–509. pmid:13428781
- 13. Pluskal T, Castillo S, Villar-Briones A, Orešič M. MZmine 2: Modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile data. BMC Bioinformatics. 2010;11: 395. pmid:20650010
- 14. Kirpich AS, Ibarra M, Moskalenko O, Fear JM, Gerken J, Mi X, et al. SECIMTools: a suite of metabolomics data analysis tools. BMC Bioinformatics. 2018;19: 151. pmid:29678131
- 15. Patterson RE, Kirpich AS, Koelmel JP, Kalavalapalli S, Morse AM, Cusi K, et al. Improved experimental data processing for UHPLC–HRMS/MS lipidomics applied to nonalcoholic fatty liver disease. Metabolomics. 2017;13: 142.
- 16. Koelmel JP, Kroeger NM, Ulmer CZ, Bowden JA, Patterson RE, Cochran JA, et al. LipidMatch: an automated workflow for rule-based lipid identification using untargeted high-resolution tandem mass spectrometry data. BMC Bioinformatics. 2017;18: 331. pmid:28693421
- 17. Tsugawa H, Cajka T, Kind T, Ma Y, Higgins B, Ikeda K, et al. MS-DIAL: data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nat Meth. 2015;12: 523–526. pmid:25938372
- 18. R Development Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2016.
- 19. Xia J, Wishart DS. Metabolomic data processing, analysis, and interpretation using MetaboAnalyst. Curr Protoc Bioinformatics. 2011;Chapter 14: Unit 14.10. pmid:21633943
- 20. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society Series B (Methodological). 1995;57: 289–300.
- 21. LipidMatch Flow (Beta) [Internet]. [cited 28 Jan 2019]. Available: http://secim.ufl.edu/secim-tools/lipidmatchflow/
- 22. Koelmel JP, Kroeger NM, Ulmer CZ, Bowden JA, Patterson RE, Cochran JA, et al. LipidMatch: an automated workflow for rule-based lipid identification using untargeted high-resolution tandem mass spectrometry data. BMC Bioinformatics. 2017;18: 331. pmid:28693421
- 23. Khotimchenko SV. Fatty acids and polar lipids of seagrasses from the sea of Japan. Phytochemistry. 1993;33: 369–372.
- 24. Minnikin DE, Abdolrahimzadeh H, Baddiley J. Replacement of acidic phospholipids by acidic glycolipids in Pseudomonas diminuta. Nature. 1974;249: 268–269. pmid:4833243
- 25. Benning C. Biosynthesis and Function of the Sulfolipid Sulfoquinovosyl Diacylglycerol. Annual Review of Plant Physiology and Plant Molecular Biology. 1998;49: 53–75. pmid:15012227
- 26. Essigmann B, Güler S, Narang RA, Linke D, Benning C. Phosphate availability affects the thylakoid lipid composition and the expression of SQD1, a gene required for sulfolipid biosynthesis in Arabidopsis thaliana. Proc Natl Acad Sci USA. 1998;95: 1950–1955. pmid:9465123
- 27. Van Mooy BAS, Fredricks HF, Pedler BE, Dyhrman ST, Karl DM, Koblížek M, et al. Phytoplankton in the ocean use non-phosphorus lipids in response to phosphorus scarcity. Nature. 2009;458: 69–72. pmid:19182781
- 28. Andersson MX, Stridh MH, Larsson KE, Liljenberg C, Sandelius AS. Phosphate-deficient oat replaces a major portion of the plasma membrane phospholipids with the galactolipid digalactosyldiacylglycerol. FEBS Lett. 2003;537: 128–132. pmid:12606044
- 29. Tjellström H, Andersson MX, Larsson KE, Sandelius AS. Membrane phospholipids as a phosphate reserve: the dynamic nature of phospholipid-to-digalactosyl diacylglycerol exchange in higher plants. Plant Cell Environ. 2008;31: 1388–1398. pmid:18643953
- 30. Riekhof WR, Naik S, Bertrand H, Benning C, Voelker DR. Phosphate starvation in fungi induces the replacement of phosphatidylcholine with the phosphorus-free betaine lipid diacylglyceryl-N,N,N-trimethylhomoserine. Eukaryotic Cell. 2014;13: 749–757. pmid:24728191
- 31. Okazaki Y, Otsuki H, Narisawa T, Kobayashi M, Sawai S, Kamide Y, et al. A new class of plant lipid is essential for protection against phosphorus depletion. Nat Commun. 2013;4: 1510. pmid:23443538
- 32. Dean AP, Sigee DC, Estrada B, Pittman JK. Using FTIR spectroscopy for rapid determination of lipid accumulation in response to nitrogen limitation in freshwater microalgae. Bioresour Technol. 2010;101: 4499–4507. pmid:20153176
- 33. Hu Q, Sommerfeld M, Jarvis E, Ghirardi M, Posewitz M, Seibert M, et al. Microalgal triacylglycerols as feedstocks for biofuel production: Perspectives and advances. The Plant Journal. 2008;54: 621–639. pmid:18476868