A rapid and sensitive method for the simultaneous analysis of aliphatic and polar molecules containing free carboxyl groups in plant extracts by LC-MS/MS

Background Aliphatic molecules containing free carboxyl groups are important intermediates in many metabolic and signalling reactions, however, they accumulate to low levels in tissues and are not efficiently ionized by electrospray ionization (ESI) compared to more polar substances. Quantification of aliphatic molecules becomes therefore difficult when small amounts of tissue are available for analysis. Traditional methods for analysis of these molecules require purification or enrichment steps, which are onerous when multiple samples need to be analyzed. In contrast to aliphatic molecules, more polar substances containing free carboxyl groups such as some phytohormones are efficiently ionized by ESI and suitable for analysis by LC-MS/MS. Thus, the development of a method with which aliphatic and polar molecules -which their unmodified forms differ dramatically in their efficiencies of ionization by ESI- can be simultaneously detected with similar sensitivities would substantially simplify the analysis of complex biological matrices. Results A simple, rapid, specific and sensitive method for the simultaneous detection and quantification of free aliphatic molecules (e.g., free fatty acids (FFA)) and small polar molecules (e.g., jasmonic acid (JA), salicylic acid (SA)) containing free carboxyl groups by direct derivatization of leaf extracts with Picolinyl reagent followed by LC-MS/MS analysis is presented. The presence of the N atom in the esterified pyridine moiety allowed the efficient ionization of 25 compounds tested irrespective of their chemical structure. The method was validated by comparing the results obtained after analysis of Nicotiana attenuata leaf material with previously described analytical methods. Conclusion The method presented was used to detect 16 compounds in leaf extracts of N. attenuata plants. Importantly, the method can be adapted based on the specific analytes of interest with the only consideration that the molecules must contain at least one free carboxyl group.


Background
The analysis of low abundant signalling molecules such as phytohormones (e.g., jasmonic acid (JA), salicylic acid (SA) and abscisic acid (ABA)) and intermediates of meta-bolic pathways (e.g., free fatty acids (FFA), oxygenated forms of fatty acids) in plants is an important tool to understand how plants grow, develop and respond to stress conditions. For high-throughput biochemical phe-notyping of, for example, wild-type or genetically modified plants grown under diverse conditions, it is essential to develop methods for the rapid, simultaneous and reliable quantitative analysis of a broad range of molecules.
The use of tandem mass spectrometry (MS/MS) coupled to liquid chromatography (LC) is ideal for the analysis of complex mixtures of compounds which are commonly found in biological matrices such as plant tissues. One of the advantages of LC-MS/MS is that separation and structural elucidation of compounds can be achieved in a continuous manner without the need for purification or derivatization steps. Another advantage of LC-MS/MS is the use of tandem MS, in which a precursor ion is massselected by mass analyzer 1, focused into a collision region preceding a second mass analyzer (collision chamber), and their mass fragments analyzed in a third mass analyzer [1]. The capacity to perform multi reaction monitoring (MRM) has the advantage of rapid and sensitive detection of several compounds even if they show similar retention times during LC.
Ionization by electrospray (ESI) is one of the most widely used tools for LC-MS/MS analysis, and the ions can be selectively monitored in negative and positive mode [2]. In the ESI negative mode, analysis of small molecules containing free carboxyl groups, yields mainly the ion [M-H] -, corresponding to their carboxylate anion. However, the efficiency of formation of carboxylate anions differs widely among compounds depending on their chemical structure [3]. For example, while small molecules containing free carboxyl groups and high numbers of heteroatoms such as the phytohormones JA, SA and ABA or aromatic substances are ionized efficiently by ESI [4][5][6], aliphatic molecules such as FFAs are relatively poorly ionized by this technique, in particular, as the aliphatic chain becomes longer and the degree of saturation higher [7]. This lower efficiency of carboxylate anion formation restricts the analysis of aliphatic molecules, specially, when small amounts of tissue are available for analysis (e.g., embryos, pollen, tissue sections obtained by laser micro-dissection).
Analytical methods to quantify aliphatic molecules are laborious, involving enrichment steps using chromatographic techniques such as thin layer chromatography (TLC), solid phase extraction (SPE) or LC previous to derivatization and gas chromatography (GC) for separation and analysis [8]. One strategy to improve the sensitivity of analysis of aliphatic molecules containing carboxyl groups by LC-MS/MS is the use of chemical derivatization that generates a strong ion in the ESI source [8]. Among these, the generation of pyridinium compounds has many advantages [9]. The presence of the N atom in the pyridine moiety allows for the efficient ionization of the com-pounds [9] and due to the mild conditions used to generate the Picolinyl ester intermediates, analysis of sensitive molecules containing conjugated double bonds is possible [10]. Moreover, since ester bonds are resistant to the conditions used to generate Picolinyl ester derivatives, esterified fatty acids do not interfere with the analysis [10].
In this study, a method for the simultaneous determination of aliphatic molecules (e.g., FFAs and their oxygenated derivatives) and more polar (phytohormones, phenolics) compounds by using Picolinyl ester derivatives of leaf extracts coupled with LC-MS/MS is presented. The method was applied for the analysis of leaves of N. attenuata plants before and after wounding and elicitation by the insect elicitors fatty acid-amino acid conjugates (FACs) [11]; which are treatments known to stimulate the production of a large number of phytohormones and secondary metabolites [12].

Method development
Commercial chemical standards (Table 1) were first derivatized to their respective Picolinyl esters using the mild method proposed in [10]. Some examples of the molecules used are shown in Figure 1. This method allows the quantitative derivatization of free carboxyl group-containing molecules within 10 min using conditions that preserve sensitive molecules containing conjugated double bonds. The first step in the reaction involves the activation of the carboxyl group with 1,1'-carbonyldiimidazole to form an active carbimidazol amid (1, Figure 2). The second step involves the reaction of 1 with 3-(hydroxymethyl)-pyridine to form the corresponding β-Picolinyl ester (2, Figure 2).
As mentioned in the background section, the formation of Picolinyl esters by this method is restricted to free carboxyl groups [10]. To confirm the absence of hydrolysis of ester bonds from esterified fatty acids, 10 μg of commercial glycerolipids (monogalactosyldiglycerol: MGDG, phosphatidylcholine: PC and phosphatidylglycerol: PG) were first purified by TLC and then subjected to the reaction. No free fatty acids could be detected (data not shown and see below for the methodology used for detection), confirming that the reaction does not induce the hydrolysis of esterified fatty acids.
Analysis of the Picolinyl ester derivatives of the commercial standards was first accomplished by their direct injection into the MS-interface to determine their [M+H] + parent ion and their MS/MS fragmentation pattern by collision induced dissociation (CID) using increasing voltage energies. The third mass analyzer was set in the scan mode for ions with m/z between 50 and 500. Two major frag-ments were generated, m/z = 92 and m/z = 108, resulting from the loss of the methyl-pyridine fragment and the hydroxymethyl-pyridine fragment, respectively ( Figure 2). The ion m/z = 92 gave the strongest intensity at a collision energy of -25.5 V and therefore the [M+H] + > 92 m/z transition was used for specific detection of Picolinyl ester derivatives.
The chromatographic separation of Picolinyl ester derivatives was performed on a reverse-phase column using acidic water and methanol as solvents in a gradient mode.
In this case, the LC method was optimized for the analysis in plant extracts of small polar molecules and aliphatic molecules containing no more than 18 carbons. After a pre-run of 1.5 min, all substances of interest eluted from the column in 18.5 min. An additional post run of 6.5 min was added for column conditioning for a final run time of 25 min. An example of chromatograms (total ion current, TIC) for derivatized commercial standards and a derivatized leaf extract from N. attenuata is shown in Fig  Mixtures of derivatized commercial standards ranging from 100 to 1000 pg/μL were first used to determine their linear range of detection and their limit of detection (LOD). These concentrations were in the same range as the endogenous compounds quantified in derivatized leaf extracts (see below). Within this range, most analytes presented a linear response (concentration vs. area) with r values higher than 0.97 with the exception of ICA, GA 3 and 2 H 2 -OPC 8:0 which presented r values between 0.93 and 0.94 (Table 1). The LODs, calculated based on the calibration plot method, were between 5 and 42 pg/μL (Table 1). To determine the injection precision, each derivatized commercial standard was injected 10 times at different concentrations and the coefficient of variation (CV) was calculated. For all compounds, the CV values were less than 0.1 for all concentrations tested (Additional file 1, Table S1).
To determine matrix suppression effects in a leaf extract, mixtures of derivatized standards were spiked at different concentrations in underivatized leaf extracts of N. attenuata plants and the linear range of detection and LOD were calculated (Table 1). In this case, the linearity of the response (r value, concentration vs. area) was similar to that presented by the derivatized standards in pure solvent, however, the LOD values were increased for all standards to amounts between 70 and 350 pg/μL.

Extraction and analysis of N. attenuata leaves
The extraction of polar and lipophilic compounds from N. attenuata leaves was performed as indicated by [13] with the modifications adopted by [14]. Additional modifications were included to increase the number of samples that can be processed simultaneously and to reduce the amount of sample material necessary for analysis (see Methods section).
The extraction method was first validated by performing 10 biological replicates of N. attenuata non-elicited and FAC-elicited leaves after 60 min of the treatment. For each replicate, 300 mg of leaf tissue were extracted, derivatized and analyzed by LC-MS/MS (Additional file 1, Table S2). The standard deviations were below 10% of the average values for all detectable compounds. To determine the extraction recovery rate, the residual leaf material obtained after the first extraction was re-extracted, derivatized and analyzed. The recovery rates were higher than 98% for all molecules tested (Additional file 1, Table S2). Sample stability was tested by re-analyzing the derivatized leaf extracts after two days of the first analysis (samples were kept at 10°C). The amounts of all compounds did not differ significantly between the first and second analyses (paired t-test, data not shown). The CV was analyzed for the detectable compounds in derivatized leaf extracts (analyzed the same day of extraction) from FAC-elicited leaves (60 min after the treatment). The CV values of twoday old extracts (kept at 10°C) were also analyzed. The CV values were below 0.1 for all detectable compounds (Additional file 1, Table S1). Finally, the efficiency of derivatization was evaluated by the analysis of the respective free compounds in derivatized leaf extracts (60 min after FAC elicitation). No free molecules were detected, corroborating that the derivatization is quantitative [10].
To validate the results obtained with the method presented in this study, 300 mg of non-elicited leaves of N. attenuata plants and leaves wounded and elicited with FAC (30 and 60 min) were analyzed with the present method and additionally with well established analytical methods for aliphatic compounds and phytohormes. Aliphatic compounds (free fatty acids in this case) were analyzed by performing separation by TLC and GC-MS analysis of their methyl-ester derivatives (see Methods section) and phytohormones were analyzed by LC-MS/MS using underivatized leaf extracts [6].
A total of 25 compounds were analyzed, including FFAs, derivatives and intermediates of the lipoxygenase pathway (e.g., OPDA, dnOPDA, OPCs) and some phytohormones related to stress responses or growth (e.g., JA, SA, ABA, JA-Ile, GAs, IAA). The analysis showed that 16 compounds could be detected in derivatized N. attenuata leaf extracts after elicitation. ICA, CA, traumatol, dnOPDA, IAA, 16:2, GA 3 , OPC-8:0 and GA 4 could not be detected by any of the methods used ( Table 2). The phytohormones IAA, GA 3 , and GA 4 accumulate to low levels in leaf tissue and their detection usually requires purification steps. Thus, these molecules were most likely below their LOD Reaction mechanisms for the formation of Picolinyl ester derivatives of carboxylic acids    (Table 1). Likewise, levels of free 16:2, CA, OPC-8:0, dnOPDA and traumatol could be either below their LOD or in some cases (CA and dnOPDA) absent in N. attenuata leaves. JA-Ile and traumatic acid could not be detected by the present method in non-elicited tissue however they were detected in low amounts by analysis of underivatized extracts (Table 2). Nevertheless, as the amount of these molecules increased after FAC elicitation, they became detectable in their Picolinyl ester forms (Table 2). In contrast, OPC-4:0 and OPC-6:0 could be detected only by the present method after FAC elicitation but not by the analysis of underivatized extracts. For the remaining compounds, a good correlation was observed between the methods used (Table 2).
Finally, increasing amounts of leaf material (5 to 200 mg; fresh weight) were also extracted from N. attenuata plants after 60 min of FAC elicitation to determine the range of tissue amounts in which a linear correlation with the amounts of detectable compounds was observed. The results showed a linear correlation between the amount of 11 compounds and the amount of leaf material extracted (Figure 4), indicating that as little as 5 mg of leaf tissue was sufficient for the reliable quantification of these molecules. In the case of JA-Ile, at least 50 mg of leaf tissue were required for detection while for OPC-6:0, OPC-4:0, traumatin and traumatic acid more than 100 mg of tissue were required (Additional file 1, Table S3).

Conclusion
A method was developed that enables the rapid, specific and simultaneous analysis of aliphatic compounds such as free fatty acids and small polar compounds such as phytohormones in N. attenuata leaves. Although in this study only a selected number of molecules were tested, the method can be adapted to the needs of the investigator. This method may prove useful in cases in which tissue amounts are limiting (e.g., embryos, ovules, pollen or specific tissue sections obtained by laser micro-dissection) and multiple analytical techniques cannot be used. Additionally, the rapid extraction procedure without additional purification steps may also facilitate the analysis of multiple samples quickly, making it compatible with high-throughput biochemical screenings.
were dried under a stream of nitrogen and 0.5 mL of water were added. Picolinyl ester derivatives were extracted twice with 0.5 mL hexane, the solvent evaporated under a stream of nitrogen and the samples reconstituted in 70/30 (v/v) methanol/water for LC-MS/MS analysis. Example of the linear correlation between the amounts of leaf tissue extracted from FAC elicited N. attenuata leaves and the amounts of compounds detected Figure 4 Example of the linear correlation between the amounts of leaf tissue extracted from FAC elicited N. attenuata leaves and the amounts of compounds detected. Different amounts of leaf tissue (5 to 250 mg; fresh weight) from N. attenuata plants were collected after 60 min of FAC elicitation. Leaf material was extracted, derivatized and analyzed by LC-MS/ MS. Two biological replicates were performed for each amount of tissue.

LC-MS/MS analysis
The MS was used in ion positive mode and ions detected using multiple reaction monitoring (MRM) and their respective m/z transitions [M+H] + > 92 (Table 1) after collision induced fragmentation with argon gas under -25.5 V collision energy. For ionization, the needle was set at 5000 V and the drying gas (nitrogen) at 300°C and 20 psi (housing 50°C). The detector was set at 1800 V. Samples were kept at room temperature (RT) and separation was also performed at RT. Analysis of underivatized phytohormones from crude leaf extracts was performed as previously described [6].

Analysis of FFA by GC-MS
FFA extraction was performed according to [14]. Heptadecanoic acid (0.5 μg per sample) was added as internal standard for quantification. After extraction, samples were reconstituted in 0.2 mL of chloroform and lipids were separated on Partisil ® K6 silica plates (Whatman, Dassel, Germany) which were developed with 70/30/1 (v/v/v) hexane/diethyl ether/acetic acid. Commercial FFAs were used as standards and plates were stained with 2% (v/v) 2'-7'-dichlorofluorescin/methanol and lipids were visualized under UV light. FFA were eluted from the silica with 3 mL of 2/1 (v/v) chloroform/methanol and methylated in 1% (v/v) H 2 SO 4 /methanol for 1 h at 75°C. Fatty acid methyl esters were extracted with hexane and analyzed in a Varian CP-3800 GC coupled with a Varian Saturn 3800 ion trap MS in electron ionisation (EI; 70 eV) mode (Varian, Palo Alto, CA). One μl of the sample was injected in splitless mode on a DB-WAX column (30 m × 0.25 mm I.D., 0.25 μm film thickness, Agilent, Boeblingen, Germany) with helium at a constant flow of 1 mL min -1 as the carrier gas. The injector was at 230°C. The oven temperature program was: 130°C for 5 min, 220°C at 3.0°C/min, 5°C/min ramp to 240°C and hold for 1 min. EI spectra were recorded on Scan mode from 40 to 400 amu. Quantification was performed in the linear range of detection and based on calibration curves generated with increasing concentrations of commercial FAMES mixes (Matreya, Pleasant Gap, PA) and the IS (17:0).