To date, variation in nectar chemistry of flowering plants has not been studied in detail. Such variation exerts considerable influence on pollinator–plant interactions, as well as on flower traits that play important roles in the selection of a plant for visitation by specific pollinators. Over the past 60 years the Aquilegia genus has been used as a key model for speciation studies. In this study, we defined the metabolomic profiles of flower samples of two Aquilegia species, A. Canadensis and A. pubescens. We identified a total of 75 metabolites that were classified into six main categories: organic acids, fatty acids, amino acids, esters, sugars, and unknowns. The mean abundances of 25 of these metabolites were significantly different between the two species, providing insights into interspecies variation in floral chemistry. Using the PlantSEED biochemistry database, we found that the majority of these metabolites are involved in biosynthetic pathways. Finally, we explored the annotated genome of A. coerulea, using the PlantSEED pipeline and reconstructed the metabolic network of Aquilegia. This network, which contains the metabolic pathways involved in generating the observed chemical variation, is now publicly available from the DOE Systems Biology Knowledge Base (KBase; http://kbase.us).
Citation: Noutsos C, Perera AM, Nikolau BJ, Seaver SMD, Ware DH (2015) Metabolomic Profiling of the Nectars of Aquilegia pubescens and A. Canadensis. PLoS ONE 10(5): e0124501. https://doi.org/10.1371/journal.pone.0124501
Academic Editor: Andrea Motta, National Research Council of Italy, ITALY
Received: October 21, 2014; Accepted: March 2, 2015; Published: May 1, 2015
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication
Data Availability: All relevant data are within the paper and its Supporting Information files and from the http://narrative.kbase.us/functional-site/#/ws/objects/Acoerulea site.
Funding: This work was funded by grants from the National Science Foundation of United States of America under the Division of Biological Infrastructure to Dr. Doreen Ware who is Co-Principal Investigator to the DBI-1265382. Financial support was provided by the Department of Energy (DE-AC02-98CH10886) for Seaver SMD and to Christos Noutsos from the NSF by the BDI-1265383, the iPlant Collaborative Project. Basil J. Nikolau and and Ann M. Perera received no specific funding for this work. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Nectar plays an essential role in the interaction between animal- and insect-pollinated plants and their pollinators . Plant–pollinator interactions are beneficial to plants, which require pollination in order to successfully reproduce, as well as to pollinators, which derive food from plant visitation. Key factors affecting the frequency and duration of visitation of pollinators to plants include nectar production rate  and the chemical composition of nectar, including the types and relative amounts of sugars, amino acids, organic acids, and lipids [3,4]. In particular, nectar sugar composition has been thoroughly studied in relation to pollinator assemblages in angiosperms [5,6]. Although several studies have compared sugar composition between species, only a few reports have compared individuals, populations, or subspecies of the same species [7,8].
The Aquilegia genus consists of 70 species[9,10]; two of these species, Aquilegia canadensis and Aquilegia pubescens, exhibit extensive variation in flower morphology . A. canadensis, which is native to eastern and central North America, has petals that are prolonged backwards into a tubular spur; the petals are primarily yellow, but redden toward the tip of the spur, which is red. Sepals are petal-like. Plants grow from early spring, bloom from March to July, and set fruit in mid-to-late summer. A. pubescens, which is native to the South Sierra mountains in California, grows only in alpine and subalpine climates; both the sepals and spurs of the flowers are yellow. Flower positioning also differs between A. canadensis and A. pubescens: the A. canadensis flower faces downwards, whereas A. pubescens flowers are erect rather than drooping. Flower morphology and positioning may be essential determinants of pollinator visitation (Fig 1). Further details regarding these species can be found in .
Black braces denote the spurs, which contain the nectar. a) A. canadensis; b) A. pubescens. The darker-colored region (bottom to middle part) of the spur indicates the nectar level
Aquilegia species are pollinated by three major categories of pollinators: bumblebees, hummingbirds, and hawk moths. Speciation in genus Aquilegia has been proposed to be driven by the pollinator shift model . In this model, nectar spurs evolve increasing length via a directional shift to pollinators with longer tongues. Therefore, the long-spurred A. pubescens is primarily pollinated by hawk moths, whereas A. canadensis is pollinated by hummingbirds. Despite the important role played by flower morphology in determining the identity of primary pollinators, nectar composition may also have a larger impact on pollinator attraction, thereby driving diversification. Recently, an elegant study provided evidence that alternations in taste receptor function contributed to the acquisition of nectar-feeding behavior and enabled the extensive radiation of hummingbird species .
To identify the key nectar elements that attract different types of pollinators, we sought to characterize the metabolic profiles of A. pubescens and A. canadensis. Using computational analysis, we mapped the experimentally identified metabolites to known metabolic pathways in the PlantSEED database. Finally, we extended the reconstruction of metabolic networks of a model species of genus Aquilegia, A. coerulea, to allow researchers to gain insight into the metabolism of a newly sequenced basal eudicot flowering plant.
Material and Methods
Plant growth and collection of flowers
A. pubescens and A. canadensis plants were grown from seeds in pots in a greenhouse under the following conditions: day temperature, 22°C; night temperature, 19°C; and daylight period, 16 hours. Flowers were monitored as they emerged. To ensure consistency in the comparison of their nectar , three male-phased (1 day old),flowers from each species were collected and immediately frozen in liquid nitrogen Flower samples were stored at -80°C prior to metabolic profiling.
To isolate the metabolites and for the GC-MS analysis, we followed the methods described by the Metabolomics Standards Initiative. More specifically, the identification of the metabolites was done by comparing the mass spectral to NIST08 Library and using Retention Indices. Because metabolites (known and unknown) respond variably to the MS detector, we quantitated our metabolites relative to internal standards, as described in . Prior to metabolite extraction, ~20 mg of tissue was precisely weighed, spiked with internal standards (20 μg of ribitol and 20 μg of nonadecanoic acid), and homogenized with 0.35 mL of hot methanol (60°C). The mixture was immediately incubated for 10 min at 60°C and sonicated for 10 min. Chloroform (0.3 mL) and 0.3 mL of water were added, and the mixture was vortexed for 3 min. After centrifugation to separate phases, 0.2 mL of the upper polar phase and 0.2 mL of the lower non-polar phase were removed and transferred into 2-mL glass vials. Both fractions were dried in a SpeedVac concentrator. The extracts were subjected to methoximation with methoxyamine hydrochloride at 30°C for 90 min. Samples were silylated with BSTFA/TCMS at 60°C for 30 min, and then subjected to gas chromatography–mass spectrometry (GC-MS) on an 6890N gas chromatograph in tandem with a 5973MSD detector equipped with a gas ionization detector (gas flow: UHP Helium, 1.0 ml/min.; ionization mode: El; polarity: positive; skimmer/focusing lens voltages: 70 eV). Samples were loaded onto the GC with an 7683B automatic liquid sampler. The mass range was set from 40–1000 m/z. The separation column was an HP5MSI (30 m long, 0.250 mm ID, 0.25 μm film thickness). All instruments and the column were obtained from Agilent Technologies. The GC-MS protocol is described in in details in . The GC was controlled by the ChemStation software (Agilent)
Upon analysis and transformation of analyte peaks to numerical values, we implemented filtering of the data. Specifically, values of metabolites detected in three flower samples from a species were selected for further analysis. All statistical analysis was performed on iPlant’s cloud-computing resource, Atmosphere , using an image in which R v3.0.1 was installed. For Student’s t-test, we used the built-in t.test function with the confidence interval set at 0.95. For principal component analysis (PCA), we used the built-in prcomp function with the default settings. For visualization of the principal components, we used the R package ggplot2. We performed Pearson correlation analysis using the cor function.
Reconstruction of the metabolic network
The Aquilegia metabolic network was constructed by annotating the reference genome of A. coerulea v1.1; sequence data were produced by the US Department of Energy Joint Genome Institute. Specifically, the annotation originated from the PlantSEED database , which contains a number of plant protein families curated according to metabolic functions pertaining to plant primary metabolism. Annotation was performed using the RAST server within SEED . All protein sequences in the genome were searched for k-mers of 8 amino acids that were associated with any of the PlantSEED protein families. Once a ‘hit’ was identified, the annotation was propagated to the Aquilegia sequence. Once the annotation was complete, a metabolic model was generated from the annotation, using the plant template created by the PlantSEED project. The plant template consists of curated reactions involved in plant primary metabolism and an extensive plant biomass. When the template is used, the annotation in the genome is linked to the reactions in the template, which are then used to create a metabolic reconstruction. Finally, the plant biomass is added to the reconstruction to allow it to function as a model of plant primary metabolism in Aquilegia. A gap-filling algorithm was used to find and fill gaps in the reconstruction that would ensure that the metabolic network could generate the same plant biomass using metabolic flux-analysis methods such as flux balance analysis (FBA; see PlantSEED paper for more details). Each step in the generation of the metabolic network was performed within the DOE Systems Biology Knowledgebase environment and framework (KBase; http://kbase.us). The annotated genome and metabolic reconstruction are available for public use in the KBase workspace named “Acoerulea”, which can be accessed by users via a free KBase account (http://narrative.kbase.us/functional-site/#/ws/objects/Acoerulea). The same workspace contains results obtained by simulating heterotrophic growth using flux balance analysis, allowing users to browse reaction activity.
Identification of metabolites to A. canadensis and A. pubescens species
We performed metabolomic profiling of flowers from three independent plants for each of the two species. We were able to identify a total of 106 analyte peaks, which were integrated with the ChemStation software from Agilent and converted to numerical values based on internal standards, and then filtered. Metabolites were quantitated relative to internal standards, for reasons described in Materials and Methods. Hence, we calculated the standard deviation for the internal standards, highlighted with yellow in S1 Table, for each of the species; all values were very low (nonadecanoic acid: 5.17•10–5 for A. canadensis and 5.14•10–5 for A. pubescens; ribitol: 9.35•10–5 for A. canadensis and 9.29•10–5 for A. pubescens). Using t-test, we ended up with a list of 75 metabolites, including 21 organic acids, 1 fatty acid, 5 amino acids, 4 esters, 25 sugars, and 19 chemically undefined compounds (S1 Table). Of these, 63 were present in both species, 9 were specific to A. pubescens, and 3 were specific to A. canadensis (S1 Table). The mean levels of 27% of the metabolites (17 out of 63) metabolites differed significantly (p-value<0.05) between species (Table 1), and 12 additional metabolites were species-specific; thus we identified at total of 29 metabolites that distinguished the two species.
After identifying metabolites, we examined the two species’ compositions of compounds in all metabolic categories. In the A. canadensis samples, total metabolites consisted of 63.197% sugars, 21.502% organic acids, 0.402% esters, 10.042% amino acids, and 3.734% unknown compounds and 1.123% other, whereas in A. pubescens, total metabolites consisted of 81.356% sugars, 12.019% organic acids, 0.395% esters, 2.115% amino acids, and 3.865% unknown compounds and 0.251% other (Fig 2). Thus nectar composition varied significantly between species.
A) Percentage of each type of compound in A. canadensis and A. pubescens, relative to the total number metabolites. B-C) Pie chart representation of the percentages shown in A.
Sugars constituted over 90% of all metabolites that differed significantly between the two species. Of the 15 different sugar metabolites identified, the mean levels of 10 differed significantly between the two species, including the three most abundant: fructose, glucose and sucrose. Notably, fructose and glucose were more abundant than sucrose in A. pubescens, whereas the sucrose concentration was as high as those of fructose and glucose in A. canadensis. At the species level, A. canadensis exhibited slightly higher net total sugars and total metabolites than A. pubescens. Additionally, we found five sugars (D-xylofuranose, arabino-hexos-2-ulose, α-D-xylopyranose, xylulose, and 2-deoxy-D-erythro-pentofuranose) that were present only in A. canadensis, suggesting that the two species exhibit qualitative as well as quantitative differences in sugar metabolites.
The second major category of metabolites we detected was organic acids. Of the 21 organic acids we identified, the levels of only three were significantly different between the two species (Table 1). Species-specific organic acids were also detected; citric acid, tetradecanoic acid (fatty acid), and 2-isopropylmalic acid were only detectable in A. canadensis, whereas quinolinic acid was only detectable in A. pubescens. The qualitative and quantitative variation in organic acids, in addition to that of sugars, suggests that nectar composition is a complex trait.
Amino acids were also detected as metabolites in nectar. The mean levels of two amino acids (L-threonine, 5-oxo-proline) differed significantly between the two species, while a third amino acid, serine, was present only in A. pubescens. Glycine and cytosine levels did not differ significantly. In addition to amino acids, we also found esters; two out of four esters were present only in A. pubescens. Moreover, a sizable fraction of 17 unknown metabolites was detected, only two of which (Unknown #28 and Unknown #29) differed significantly between species; these compounds remain to be identified. The variation in metabolite identities and concentrations prompted us to look for correlations between flower spur length and the levels of metabolites that differed significantly between the two species. With some exceptions, metabolites tended to be positively correlated with spur length in one species and negatively correlated in the other, (Table 2). The most striking exception was the strong positive correlations in A. pubescens and the strong negative correlations in A. canadensis, of the amino acids 5-oxo-proline, and L-threonine. Of all sugars, sucrose, scyllo-inositol and β-D- galactofuranose exhibited the highest correlations with spur length, whereas fructose and glucose exhibited very low positive correlations to spur length in both species. Unknown metabolite #29 was among the compounds most strongly positively correlated with spur length in both species. Moreover, at the species level, A. pubescens consistently exhibited very strong positive correlations of spur length with metabolites that differed significantly between the two species,
In addition, we performed principal component analysis(PCA) to define the metabolites that contributed the most to differentiation between the two species (Fig 3; Table 3 and S2 Table). The PCA revealed that PC1 explained 90% of the total variance, whereas PC2 explained another 8% (Table 3).
The black numbers inside the plot represent each of the metabolites (for a key, see S2 Table) and the red arrow shows the relative loadings of the species to the first and second principal components. B) Is a zoom version of A for a better visualization of the metabolites.
Mapping of metabolites to pathways and reconstruction of the metabolic network of Aquilegia
We mapped all of the identified metabolites to known pathways in the PlantSEED database. Of the 75 metabolites, we were able to map 21. Each metabolite mapped to between 4 and 281 pathways (S3 Table). Because the metabolites are involved in many different pathways, those pathways might be elaborated.
The process of annotating the Aquilegia genome resulted in assignment of metabolic annotations to 1,202 different features, and the resulting working metabolic model consisted of 1,573 reactions (S4 Table). The majority of these reactions (72.3%) were assigned to the cytosol, 16.89% to plastids, 4.41% to mitochondria, 3.58% to the endoplasmic reticulum, 1.25% to the cell wall, 0.75% to the vacuole, and 0.25% to peroxisomes and 0.58% to other compartments (Fig 4; S4 Table). According to KEGG pathway annotations (S5 Table), many gene models are involved in metabolism, with the greatest number associated with amino-acid metabolism, fatty acid biosynthesis, and phenylpropanoid biosynthesis, followed by carbon fixation. The whole model is available through KBASE at http://narrative.kbase.us/functional-site/#/ws/objects/Acoerulea.
A) Number of metabolic reactions in each subcellular compartment. B) Pie chart representation of the distribution of metabolic reactions in each compartment, expressed as percentage. Families with fewer than three members were assigned to the ‘Other’ category.
In this study, by performing metabolomic profiling of whole flowers from two Aquilegia species, we sought to obtain insights into the metabolites that influence attraction of pollinators. Some metabolites were present in both species, whereas others were species-specific. Of the shared metabolites, the levels of around 25% differed significantly between the two species. Moreover, we mapped the identified metabolites on to metabolic pathways in the PlantSEED database, and constructed a model metabolic network of A. coerulea.
Sugars are the most prominent component of total metabolites in both species, as described elsewhere [21–23]. Our results revealed that the levels of ten different sugar metabolites differed significantly between the two species we examined. Although the total sugar concentration did not differ significantly between the two species, subtle differences were detected in glucose and fructose levels, as previously reported for species with similar pollinators [2,3,5,24]. Of the three major sugar metabolites, the most pronounced difference found in this study was in the level of sucrose: sucrose levels in A. canadensis, a hummingbird-pollinated species, was 3.75 times higher than A. pubescens, a moth-pollinated species. These results are consistent with the sucrose levels detected in naturally grown plants that are visited primarily by hummingbirds [25,26]. Whether pollinator-mediated selection has shaped and/or maintained the differences in nectar between A. pubescens and A. canadensis remains to be studied.
Several other sugars also differed significantly between A. pubescens and A. canadensis. The disaccharide turanose, which has a high energy density, is also found in the nectar of sunflowers . Glucitol, also known as sorbitol, is found in traces in the Rosaceae species and seems to be rapidly converted into sucrose and fructose . The levels of scyllo-inositol and glycerol also differed significantly between species. Xylopyranose, a sugar found in the cell walls of plant cells , has been previously reported to exhibit species-specific differences in its levels , a finding supported by this study.
The results of this study provide insights not only into sugars, which are certainly major metabolites, but also into other types of compounds such as amino acids, organic acids, and esters, which have been neglected in previous studies, despite the fact that they may play significant roles in pollination-driven differentiation. To date, the majority of studies on nectar composition have focused on sugar content , even though other metabolites have been demonstrated to be crucial for the attraction of pollinators [31–33].
Amino acids play an important role in the taste of nectar [34,35]. Although the mean total amino-acid content of A. pubescens was 2-fold greater than that of A. canadensis, this difference was not statistically significant statistically, as previously reported . However, of the five amino acids we detected, the individual levels of two amino acids (5-oxo-proline and threonine) differed significantly between the two species, and serine was detected only in A. pubescens. A study that tested the preference of forager honeybees for proline-, serine-, and alanine-enriched nectars reported that proline-enriched nectar was clearly preferred over nectars containing only sugars, whereas serine-enriched nectar was disfavored . Whether similar trends exist for hummingbirds or hawk moths remains to be elucidated.
Organic acids in nectars have not been studied in detail, despite the fact that they may play an important role in nectar quality. For example, studies of honey have revealed that organic acids serve as antioxidants and play a role in preventing infestation . In addition, organic acids provide flavor to nectars, as well as aromas that can attract pollinators. In our study, of the 22 organic acids we identified, three were present in both species at significantly different levels: propanoic acid, 2,3-bis[oxy],4-aminobutyric acid, glucaric acid. Citric acid, 2-isopropylmalic acid, and tetradecanoic acid were unique to A. canadensis, and quinolinic acid was unique to A. pubescens. Notably, the proportion of organic acids in nectar was 2-fold greater in A. canadensis than in A. pubescens.
Esters tend to provide odor, which along with visual stimuli such as flower color is a vital means of attracting pollinators . Two esters, myristic acid-2,3-bis(oxy)propylester and catechollactate, also differed significantly between the two species, as did the levels of two unknown metabolites (S1 Table). The identities and roles of these compounds should be further explored in future studies.
Early studies  elegantly demonstrated the co-evolution of spur length with pollinators in the Aquilegia genus. Based on these previous findings, we also examined the correlations between the lengths of the spurs, which contain the nectar, with the levels of various metabolites. Some metabolites were correlated positively with spur length in one species and negatively in the other. A. pubescens exhibited stronger correlations of spur length with amino acids, acids, and sugars than A. canadensis, suggesting that various metabolite pathways may be involved in the development of longer spurs. These metabolites should be further tested to provide insights in Aquilegia diversification.
Furthermore, we used the PlantSEED database  to map experimentally identified metabolites onto metabolic pathways. Only 21 out of our 75 metabolites were mapped to pathways related to biosynthesis. Also, using the PlantSEED pipeline, we annotated the genome of the model species A. coerulea, whose sequence data are available from JGI, and constructed a metabolic network of the Aquilegia in KBase. This metabolic model will enable more in silico research on the differential regulation of the enzymes involved in the biosynthesis of the metabolites described in this study.
Variation in nectar production variation among populations, plants, and flowers of a single plant has been observed in multiple species . Moreover, in some cases, nectar volumes and concentrations are strongly affected by environmental factors . Notably, a previous study  in Aquilegia species pollinated by bees revealed major differences in sucrose and fructose content, not only between the species studied, but also between samples of the same species collected in the field. Because Aquilegia species intercross to produce vital hybrids, it would be possible to identify the inheritance of nectar qualitative and quantitative traits by generating crosses between A. canadensis and A. pubescens and testing them in a controlled environment.
S1 Table. Quantification of all identified metabolites.
All metabolites and their relative quantities to the internal standards. In the last column are denoted the metabolites different significantly between the two species by performing the t-test
S2 Table. Loadings of Principal Components.
The loading are reported for each of the metabolites. In column two there is an ID number for each of the metabolites shown in Fig 4 of the paper.
S3 Table. Metabolites mapped to BioCyc and KEGG Pathways.
Metabolites vary to the number of pathways they belong to. We mapped 22 out of our 79 metabolites
S4 Table. Mapping of metabolites to cell compartments and annotations of metabolites.
In column F we mention an Arabidopsis exemplar that belongs to each of the pathways. Most of the metabolites belong to cytosol some to the rest of the compartments.
We would like to thank the members of the Ware lab for critical reading of the paper. We would like to thank Ivan Iossifov for critical discussion on the statistics performed. We would like to thank the iPlant collaborative team for providing access to their cloud services for data analysis, as well as the KBase project for hosting our metabolic data and making them publicly available to the scientific community. The work of C. Noutsos was supported by NSF iPlant Collaborative Project DBI-1265383 and S. M. D. Seaver was supported by the Office of Science, Office of Biological and Environmental Research, of the US Department of Energy under contract number DE-ACO2-06CH11357, as part of the DOE Systems Biology Knowledgebase.
Conceived and designed the experiments: CN BJN SMDS. Performed the experiments: CN MAP SMDS. Analyzed the data: CN SMDS. Contributed reagents/materials/analysis tools: CN BJN MAP SMDS DW. Wrote the paper: CN SMDS BJN.
- 1. Jones CE, John R. Little Handbook of experimental pollination biology. New York. 1983
- 2. Fenster CB, Cheely G, Dudash MR, Reynolds RJ. Nectar Reward and advertisement in hummingbird pollinated Silene virginia (Caryophyllaceae). Am J Bot 2006;93: 1800–1807. pmid:21642126
- 3. Baker HG BI. Floral nectar sugar constituents in relation to pollinator type; Jones CE LR, editor. New York: Van Norstrand Reinhond. 1983
- 4. Bernardello G, Galetto L, Forcone A. Floral nectar chemical composition of some species from Patagonia II. Biochemical Systematics and Ecology 1999;27: 779–790.
- 5. Jurgens A. Nectar sugar composition and floral scent compounds of diurnal and nocturnal Conophytum species (Aizoaceae). South African Journal of Botany 2004;70: 191–205.
- 6. Dupont YL, Hansen DM, Rasmussen JT, Olesen JM. Evolutionary changes in nectar sugar composition associated with switches between bird and insect pollination: the Canarian bird-flower element revisited. Functional Ecology 2004;18: 670–676.
- 7. Roldán-Serrano AS, Guerra-Sanz J. Dynamics and sugar composition of sweet pepper (Capsicum annuum L.) nectar. Journal of Horticultural Science and Biotechnology 2004;79: 717–722.
- 8. Wit T, Jürgens A, Geyer R, Gottsberger G. Nectar dynamics and sugar composition in flowers of Silene and Saponaria species (Caryophyllaceae). Plant Biology 1999;1: 334–345.
- 9. Grant V. Isolation and hybridization between Aquilegia formosa and Aquilegia pubescens. Aliso 1952;2: 341–360.
- 10. Whittall JB, Hodges SA. Pollinator shifts drive increasingly long nectar spurs in columbine flowers. Nature 2007;447: 706–709. pmid:17554306
- 11. Hodges SA, Arnold ML. Columbines—a Geographically Widespread Species Flock. Proceedings of the National Academy of Sciences of the United States of America 1994;91: 5129–5132. pmid:8197196
- 12. Payson EB The North American Species of Aquilegia; museum SiUsn, editor. Washington: Goverment Printing Office. 1918
- 13. Baldwin MW, Toda Y, Nakagita T, O'Connell MJ, Klasing KC, Misaka T. et al. Evolution of sweet taste perception in hummingbirds by transformation of the ancestral umami receptor. Science 2014;345: 929–933. pmid:25146290
- 14. Canto A, Perez R, Medrano M, Castellanos MC, Herrera CM. Intra-plant variation in nectar sugar composition in two Aquilegia species (Ranunculaceae): Contrasting patterns under field and glasshouse conditions. Annals of Botany 2007:99: 653–660. pmid:17259227
- 15. Sumner LW, Amberg A, Barrett D, Beale MH, Beger R, Daykin CA et al. Proposed minimum reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics: Official journal of the Metabolomic Society 2007;3: 211–221. pmid:24039616
- 16. Schmidt MA, Barbazuk WB, Sandford M, May G, Song ZH, Zhou WX et al. Silencing of Soybean Seed Storage Proteins Results in a Rebalanced Protein Composition Preserving Seed Protein Content without Major Collateral Changes in the Metabolome and Transcriptome. Plant Physiology 2011;156: 330–345. pmid:21398260
- 17. Quanbeck SM, Brachova L, Campbell AA, Guan X, Perera A, He K. et al. Metabolomics as a hypothesis-generating functional genomics tool for the annotation of Arabidopsis thaliana genes of "unknown function". Frontiers in plant science 2012;3.
- 18. Goff SA, Vaughn M, McKay S, Lyons E, Stapleton AE, Gessler D. et al. The iPlant Collaborative: Cyberinfrastructure for Plant Biology. Frontiers in plant science 2011;2: 34. pmid:22645531
- 19. Seaver SMD, Gerdes S, Frelin O, Lerma-Ortiz C, Bradbury LMT, Zallot R. et al. High-throughput comparison, functional annotation, and metabolic modeling of plant genomes using the PlantSEED resource. Proceedings of the National Academy of Sciences of the United States of America 2014;111: 9645–9650. pmid:24927599
- 20. Overbeek R, Olson R, Pusch GD, Olsen GJ, Davis JJ, Disz T. et al. The SEED and the Rapid Annotation of microbial genomes using Subsystems Technology (RAST). Nucleic Acids Research 2014;42: D206–D214. pmid:24293654
- 21. Kaczorowski RL, Gardener MC, Holtsford TP. Nectar traits in Nicotiana section Alatae (Solanaceae) in relation to floral traits, pollinators, and mating system. American Journal of Botany 2005;92: 1270–1283. pmid:21646148
- 22. Wolff D. Nectar sugar composition and volumes of 47 species of gentianales from a southern Ecuadorian montane forest. Annals of Botany 2006;97: 767–777. pmid:16495315
- 23. Gottsberger G, Schrauwen J, Linskens HF. Amino-Acids and Sugars in Nectar, and Their Putative Evolutionary Significance. Plant Systematics and Evolution 1984;145: 55–77.
- 24. Fenster CB, Armbruster WS, Wilson P, Dudash MR, Thomson JD. Pollination syndromes and floral specialization. Annual Review of Ecology Evolution and Systematics 2004;35: 375–403.
- 25. Burke JM, Wyatt R, dePamphilis CW, Arnold ML. Nectar characteristics of interspecific hybrids and their parents in Aesculus (Hippocastanaceae) and Iris (Iridaceae). Journal of the Torrey Botanical Society 2000;127: 200–206.
- 26. Pyke GH, Waser NM. The Production of Dilute Nectars by Hummingbird and Honeyeater Flowers. Biotropica 1981;13: 260–270.
- 27. Zajacz E. Nectar production and some site-related issues of sunflower hybrids under different agro-ecological conditions. Godollo: Szent Istvan University. 2011
- 28. Bieleski RL, Redgwell RJ. Sorbitol Metabolism in Nectaries from Flowers of Rosaceae. Australian Journal of Plant Physiology 1980;7: 15–25.
- 29. Payan F, Leone P, Porciero S, Furniss C, Tahir T, Williamson G. et al. The dual nature of the wheat xylanase protein inhibitor XIP-I—Structural basis for the inhibition of family 10 and family 11 xylanases. Journal of Biological Chemistry 2004;279: 36029–36037. pmid:15181003
- 30. Bender R, Klinkenberg P, Jiang ZH, Bauer B, Karypis G, Nguyen N. et al. Functional genomics of nectar production in the Brassicaceae. Flora 2012;207: 491–496.
- 31. Gijbels P, Van den Ende W, Honnay O. Landscape variation in nectar amino acid and sugar composition in lepidoptera pollinated orchid species and its relation with fruit set. Journal of Ecology 2014;102: 136–144.
- 32. Nepi M. Beyond nectar sweetness: the hidden ecological role of non-protein amino acids in nectar. Journal of Ecology 2014;102: 108–115.
- 33. Kirk H, Cheng DD, Choi YH, Vrieling K, Klinkhamer PGL. Transgressive segregation of primary and secondary metabolites in F-2 hybrids between Jacobaea aquatica and J. vulgaris. Metabolomics 2012;8: 211–219. pmid:22448153
- 34. Baker HG, Baker I. Intraspecific Constancy of Floral Nectar Amino-Acid Complements. Botanical Gazette 1977;138: 183–191.
- 35. Baker I, Baker H.G. Analysis of amino acids in flower nectars of hybrids and their parents, with phylogenetic implications. New Phytology 1976;76: 87–98.
- 36. Brice AT, Grau CR. Protein-Requirements of Costas Hummingbirds Calypte-Costae. Physiological Zoology 1991;64: 611–626.
- 37. Bertazzini M, Medrzycki P, Bortolotti L, Maistrello L, Forlani G. Amino acid content and nectar choice by forager honeybees (Apis mellifera L.). Amino Acids 2010;39: 315–318. pmid:20091414
- 38. Mato I, Huidobro J., Simal-Lozano J., & Sancho M. T. Significance of nonaromatic organic acids in honey. Journal of Food Protection 2003;66: 2371–2376. pmid:14672241
- 39. Shalit M, Guterman I, Volpin H, Bar E, Tamari T, Menda N. et al. Volatile ester formation in roses. Identification of an acetyl-coenzyme A. Geraniol/citronellol acetyltransferase in developing rose petals. Plant Physiology 2003;131: 1868–1876. pmid:12692346
- 40. Hodges SA. Consistent Interplant Variation in Nectar Characteristics of Mirabilis-Multiflora. Ecology 1993;74: 542–548.
- 41. Wyatt R, Broyles SB, Derda GS. Environmental-Influences on Nectar Production in Milkweeds (Asclepias-Syriaca and a-Exaltata). American Journal of Botany 1992;79: 636–642.