A highly selective and sensitive method for the simultaneous analysis of several plant hormones and their metabolites is described. The method combines high-performance liquid chromatography (HPLC) with positive and negative electrospray ionization-tandem mass spectrometry (ESI–MS/MS) to quantify a broad range of chemically and structurally diverse compounds. The addition of deuterium-labeled analogs for these compounds prior to sample extraction permits accurate quantification by multiple reaction monitoring (MRM). Endogenous levels of abscisic acid (ABA), abscisic acid glucose ester (ABA-GE), 7′-hydroxy-abscisic acid (7′-OH-ABA), phaseic acid (PA), dihydrophaseic acid (DPA), indole-3-acetic acid (IAA), indole-3-aspartate (IAAsp), zeatin (Z), zeatin riboside (ZR), isopentenyladenine (2iP), isopentenyladenosine (IPA), and gibberellins (GA)1, GA3, GA4, and GA7 were determined simultaneously in a single run. Detection limits ranged from 0.682 fmol for Z to 1.53 pmol for ABA. The method was applied to the analysis of plant hormones and hormonal metabolites associated with seed dormancy and germination in lettuce (Lactuca sativa L. cv. Grand Rapids), using extracts from only 50 to 100 mg DW of seed. Thermodormancy was induced by incubating seeds at 33°C instead of 23°C. Germinating seeds transiently accumulated high levels of ABA-GE. In contrast, thermodormant seeds transiently accumulated high levels of DPA after 7 days at 33°C. GA1 and GA3 were detected during germination, and levels of GA1 increased during early post-germinative growth. After several days of incubation, thermodormant seeds exhibited a striking transient accumulation of IAA, which did not occur in seeds germinating at 23°C. We conclude that hormone metabolism in thermodormant seeds is surprisingly active and is significantly different from that of germinating seeds.
Plant hormones are low-molecular-weight natural products that act at micromolar (or even lower) concentrations to regulate essentially all physiological and developmental processes during a plant's life cycle. These structurally diverse compounds include auxins, cytokinins (CK), abscisic acid (ABA), gibberellins (GA), ethylene, polyamines, jasmonates, salicylic acid, and brassinosteroids (reviewed by Davies, 1995). Several compounds in the biosynthetic and degradative pathways of plant hormones can exhibit biologic activity, giving rise to a very complex network of signaling molecules at the cellular level. To further complicate the picture, there is mounting evidence of considerable cross-talk among plant-hormone-signaling pathways in regulating developmental and physiological processes. For example, genetic analyses have revealed key interactions between ethylene, ABA, and gibberellins during seed development (Gazzarina and McCourt, 2001), most of which are antagonistic. Additional interactions between hormone response pathways and sugar signaling are just beginning to be elucidated (Finkelstein et al., 2002).
An ideal analytical method would provide quantification of all the plant hormones in a given sample in a single experiment. The development of improved profiling methods is critically important for understanding the role of hormone-induced signaling networks in controlling specific developmental pathways or physiological responses. Such a high-throughput and comprehensive approach is vital for functional genomics research. The problem is challenging. The plant hormones described previously can be neutral, acidic, or basic. Under physiological conditions, these signaling molecules are present at very low concentrations (Davies, 1995), in a background of a wide range of more abundant primary and secondary metabolites. Therefore, methods for the simultaneous and comprehensive analysis of plant hormones and their metabolites entail the use of extraction and analytical methods that can accommodate the wide range of chemical properties represented by the different classes of plant hormones. Specifically, the extraction procedure must be efficient for all the plant hormones and metabolites of interest despite their differing chemistries. Furthermore, the analytical method must be extremely selective to enable quantification of the relatively lower level plant hormones and metabolites in the presence of the hundreds of more abundant compounds known to be present in plant tissue extracts. Fiehn et al. (2000) recently described the metabolomic profiling of over 300 compounds in extracts of rosette leaves from four genotypes of Arabidopsis thaliana by gas chromatography-mass spectrometry (GC–MS). However, none of these compounds was a plant hormone. A number of sensitive and specific methods have been developed for quantification of individual plant hormones. These include GC–MS, high-performance liquid chromatography (HPLC) and liquid chromatography-mass spectrometry (LC–MS) (see Crozier and Moritz, 1999), as well as indirect methods such as enzyme-linked immunosorbent assays (ELISA) (see Beale, 1999).
The utility of mass spectrometry for the profiling and quantification of plant hormones and metabolites is becoming increasingly apparent, because of the high sensitivity and selectivity of this analytical method (Glassbrook et al., 2000). Several recent studies using GC–MS for the analysis of different classes of plant hormones and their metabolites have been published (Duffield and Netting, 2001; Hansen and Dörffling, 1999; Östin et al., 1998; Schmitz et al., 2000; Tam et al., 2000). A method for profiling three classes of acidic plant hormones by GC–MS/MS was recently reported by Müller et al. (2002). The ability to resolve co-eluting compounds using GC–MS with selective ion monitoring (SIM) alleviates the requirement for complete chromatographic separation of compounds. However, GC–MS is limited to the analysis of volatile compounds, and as a result, its application to the investigation of plant hormones means that a majority of these compounds require derivatization prior to analysis (Glassbrook et al., 2000; Wilbert et al., 1998). Contributions of ions from co-eluting major metabolites can overlap with ions from relatively lower level plant hormones, prohibiting quantification. Additionally, the high temperatures used in GC columns can lead to thermal disintegration of labile compounds, such as ABA-GE (Schneider et al., 1997), making precise quantification problematic (Wilbert et al., 1998).
Liquid chromatography-tandem mass spectrometry (LC–MS/MS) is an alternative method that is essentially free of the limitations outlined above (Wilbert et al., 1998; Ross et al., unpublished; Zaharia et al., unpublished; Feurtado et al., unpublished). The selectivity and sensitivity of this method relies on the application of multiple reaction monitoring (MRM), in which each ionized compound gives a distinct precursor-to-product ion transition that is diagnostic for the presence of that particular compound in an extract. Also, the need for complete resolution of compounds prior to analysis is bypassed because peaks containing co-eluting compounds can be resolved by monitoring for specific precursor-to-product ion transitions (Glassbrook et al., 2000; Wilbert et al., 1998; Ross et al., unpublished; Zaharia et al., unpublished; Feurtado et al., unpublished). LC–MS/MS procedures utilizing isotopically labeled internal standards have been published for the quantification of several plant hormones including jasmonic acid, methyl jasmonate and salicylic acid (Wilbert et al., 1998), IAA metabolites (Kowalczyk and Sandberg, 2001), ABA (Gómez-Cadenas et al., 2002) and ABA-GE (Hogge et al., 1993; Schneider et al., 1997). A method for quantification of ABA using LC–MS/MS with non-exchangeable isotope-labeled ABA analogs as internal standards has recently been developed (Ross et al., unpublished). The method has been extended to include the analysis of compounds in the ABA catabolic pathway (Zaharia et al., unpublished; Feurtado et al., unpublished).
The present work reports on the development of a highly sensitive and selective method for the simultaneous profiling and quantification of a wide variety of plant hormone groups and their metabolites using high-performance liquid chromatography (HPLC) coupled with electrospray ionization-tandem mass spectrometry (ESI–MS/MS). One of the most important features of this method is that it can be tailored to take into account the chemical properties of each compound under analysis. First, each compound is analyzed in its native state without any need for derivatization procedures. Second, compounds are separated by HPLC, and therefore, there is no requirement for the use of high temperatures. Third, the different plant hormones and metabolites can be analyzed using either positive- or negative-ion electrospray in a single LC–MS/MS run, depending on their chemical properties. The acidic compounds, e.g. ABA, IAA, and GAs, can be analyzed in the negative-ion mode, while cytokinins are analyzed in the positive-ion mode. Using this method, a wide range of structurally diverse plant hormones and metabolites can be analyzed simultaneously with a high degree of sensitivity and selectivity. Further, we describe the application of this method for the analysis and quantification of 15 compounds in dormant and germinating lettuce seeds.
In the majority of seed plants, primary dormancy is a natural phenomenon that is characterized by a transient inability of mature seeds to germinate under conditions that are conducive to germination (Grappin et al., 2000). Primary dormancy inception occurs during the maturation phase of seed development (reviewed in Bewley, 1997) and is maintained in the mature seed at dispersal. Dormancy can also be induced in mature, already dispersed, non-dormant seeds – induced or secondary dormancy – by environmental conditions that are unfavorable for germination, e.g. anoxia, unsuitable temperatures, or illumination (Bewley, 1997). Dormancy is an adaptive strategy of seed plants that ensures the successful establishment of the next generation; its induction and maintenance are mediated by plant hormones. The termination of seed dormancy can be achieved by moist chilling, after-ripening, or various other signals depending on the species and is generally associated with changes in plant hormone concentrations including ABA, GA, and ethylene (Finkelstein et al., 2002; Grappin et al., 2000; McKeon et al., 1995; Yoshioka et al., 1998). Secondary dormancy of lettuce (Lactuca sativa L. cv. Grand Rapids) is readily induced by manipulating the temperature at which seeds are imbibed. Those imbibed at an optimal temperature (23°C) in darkness readily germinate, while imbibition at a supra-optimal temperature (33°C) results in no germination (Yoshioka et al., 1998). Evidence that ABA synthesis is necessary for the maintenance of secondary dormancy in lettuce seeds incubated at the high temperature was provided by Yoshioka et al. (1998). For example, the carotenoid- and ABA-biosynthesis inhibitor fluridone is effective in permitting germination of seeds at the high temperature (33°C). However, although ABA synthesis occurs in the thermodormant seeds, there is no corresponding increase in the ABA content of the seeds, indicating that ABA is concurrently metabolized. In the present study, changes in hormone metabolism associated with the secondary dormancy of lettuce seeds (induced by supra-optimal temperatures of imbibition) were monitored. While methods for analyzing individual hormones have been developed for studying the role of these signaling molecules in plant processes, the present research aims to provide a comprehensive metabolic profile of plant hormones, affording a snapshot of the overall hormonal status in dormant and germinating seeds. Hormones and hormone metabolites accumulated in lettuce seeds under conditions that inhibit germination were compared to those accumulated under conditions that are favorable for germination. Representative examples of four of the important plant hormone classes were chosen for the present study.
The specific objectives of this study were twofold: (i) to develop an LC–MS/MS method for the simultaneous analysis and quantification of 15 plant hormones and metabolites from four of the nine major classes (auxins, CK, ABA, and GA) using deuterium-labeled analogs as internal standards; and (ii) to apply the newly developed LC–MS/MS method to compare hormone profiles in thermodormant lettuce seeds and during germination.
In this report, we describe the development of a sensitive and selective method for quantifying four classes of plant hormones and several hormone metabolites using LC–MS/MS. The four hormone classes investigated, i.e. auxins, cytokinins, abscisic acid and gibberellins, represent a structurally diverse range of organic compounds with very different chemical properties. The analytical method proved to be quite versatile for the analysis of these four groups of plant hormones and metabolites in their natural state, thus negating any need for derivatization procedures prior to analysis. In addition to the sensitivity and selectivity of this method, another important feature is the ability to switch between the negative and positive ionization modes during a single LC/ESI–MS/MS run. This was illustrated in the case of plant hormones and metabolites whose retention times were between 12 and 16.7 min (Table 1) that were analyzed in functions 2 and 3 in the MRM mode (Figure 5a). In this case, although the retention times specified in the two MRM functions overlapped, the cytokinins (2iP and IPA and their respective internal standards) in function 3 could still be analyzed in the positive ionization mode, while simultaneously, the compounds in function 2 could be analyzed in the negative ionization mode. With the comprehensive approach to analysis of plant hormones and metabolites now afforded by LC/ESI–MS/MS and the number of compounds being analyzed, there is an increased likelihood that the retention times for some of these structurally diverse compounds will overlap. Future MS-based methods must therefore accommodate the different chemical properties of these compounds while maintaining the selectivity and sensitivity required for quantification. For such methods, the ability to switch between ESI and atmospheric pressure chemical ionization as well as positive- and negative-ion modes would be a significant advantage.
In light of the mounting evidence indicating interactions between the different hormone response pathways in regulating developmental and physiological processes (Finkelstein et al., 2002; Gazzarina and McCourt, 2001), there is a requirement for the development of analytical methods that allow the simultaneous and more comprehensive analysis of a wide range of compounds. This will aid in our understanding of hormonal action at the molecular, cellular, and biochemical levels. These methods will be invaluable for complementing high-throughput, functional genomics research. As was recently pointed out by Müller et al. (2002), such broad approaches to the investigation of plant hormones and plant hormone metabolites are likely to yield important information for questions that address the importance of hormone-signaling pathways and underlie a given physiological or developmental process. Herein, we have applied such a method to elucidate hormone metabolism associated with germination and with the induction and maintenance of secondary dormancy in lettuce seeds. Utilizing 12 deuterium-labeled analogs as internal standards, 15 plant hormones and metabolites were quantified simultaneously in a single sample extract derived from seeds at different stages of germination/growth and induction/maintenance of thermodormancy. Thus, 27 different compounds were monitored simultaneously in each LC/ESI–MS/MS run. Despite the wide structural and chemical diversity of the compounds analyzed, only a simple extraction protocol was required. The efficiency of the extraction procedure was demonstrated by the high recovery of the internal standards used for quantification of the endogenous compounds. The method also proved to be efficient for the extraction of the wide variety of plant hormones and metabolites investigated. As demonstrated by the low detection limits for the various compounds, the amount of plant material required for analysis was very small (50–100 mg DW). The compounds could be quantified against a background of hundreds of other organic compounds known to be present in seed extracts, demonstrating the sensitivity and selectivity of LC/ESI–MS/MS, and the application of this method for investigating plant hormone metabolism.
The hormone and hormone metabolite profiles of germinating and thermodormant lettuce seeds were distinct. The major differences can be summarized as follows:
ABA metabolism is different in germinating and thermodormant seeds
Germinating seeds incubated at 23°C accumulated the ABA-glucose ester conjugate over the first 18 h. The decline of ABA-GE between 18 and 24 h was not accompanied by an increase in the level of free ABA. No increase in ABA or ABA-GE occurred during the transition from germination to growth, which occurred between 1 and 2 days in germination conditions at 23°C. These results suggest that there is ABA biosynthesis over the first 18 h of imbibition, with the flux through ABA demonstrated by ABA-GE synthesis.
In contrast to germinating lettuce seeds, there was no significant accumulation of ABA-GE in thermodormant seeds. Instead, these seeds accumulated DPA, possibly indicating a role for 8′-hydroxylation in the maintenance of secondary dormancy. However, the levels of DPA in thermodormant seeds at 7 days were dramatically different from those for ABA. Thus, it is unlikely that the accumulation of DPA can be attributed solely to a sudden and drastic change in the relative rates of ABA synthesis and/or DPA catabolism. An alternative possibility is that DPA arises from a distinct pool of a DPA-related metabolite. This is yet to be investigated. The decline in DPA after day 7 may be caused by a further metabolism of DPA to its 4′-glucoside. Although this compound was not monitored in the present study, it has been reported in several plant tissues (Walton and Li, 1995).
The results described above suggest alternate ABA catabolic pathways in dormant and germinating seeds. These results may be somewhat analogous to differences in ABA metabolism that occur by artificially treating leaves of Xanthium with tetcyclacis, a compound which blocks 8′-hydroxylation of ABA (Krochko et al., 1998; Zeevaart et al., 1988). Under these conditions, ABA-GE is synthesized, in contrast to untreated leaves in which PA is the major ABA metabolite. It is possible that 8′-hydroxylation and associated DPA formation is inhibited and/or is a minor pathway under conditions permitting germination.
Abscisic acid biosynthesis appears to be necessary to maintain thermodormancy of seeds of lettuce (Yoshioka et al., 1998) and primary dormancy of seeds of tobacco (Nicotiana plumbaginifolia; Grappin et al., 2000). In the present study, ABA levels remained relatively constant in thermodormant seeds during the 8-day study period at 33°C, despite notable fluctuations in the detectable catabolites described above. This is consistent with the report of Yoshioka et al. (1998), which showed that ABA levels do not change in thermodormant lettuce seeds even though ABA synthesis is occurring.
The germination-to-growth transition was accompanied by an increase in GA1
Gibberellin A1 and GA3 accumulated in lettuce seeds during germination, and levels of GA1 increased further during early post-germinative growth. The extent of increase in GA1 following germination at 2 days was not evident in dormant seeds after the same incubation period.
Thermodormancy of lettuce seeds was associated with major accumulations of IAA and Z, which were not accompanied by increases in their conjugated metabolites
Hormone conjugation is known to play an important role in regulating the amount of the biologically active auxin, IAA (Normanly and Bartel, 1999; Tam et al., 2000). The levels of the cytokinins Z and 2iP are also modulated through conjugation (McGaw and Burch, 1995). In the present study, IAA was relatively low in dry seeds and in germinating lettuce seeds. However, the amount of the conjugate IAAsp was consistently higher than the amount of IAA. In marked contrast, thermodormant lettuce seeds exhibited a huge, but transient, accumulation of IAA, although there was no accompanying significant increase in the amount of IAAsp. The metabolic fate of this free IAA could not be determined in the present study as only one amide-linked IAA conjugate was analyzed. Other amide-linked IAA conjugates and ester-linked conjugates are known to exist naturally in plants (Normanly and Bartel, 1999). For example, IAA-glutamate and IAA-glucose ester have been identified and quantified in Arabidopsis seedlings (Tam et al., 2000). These authors reported that amide-linked IAA conjugates and ester-linked conjugates represented 90 and 10% of the IAA pool in Arabidopsis seedlings, respectively (Tam et al., 2000). As with DPA accumulating in dormant seed, the source of this IAA peak is puzzling and may be the product of release from a conjugated source in addition to (or instead of) de novo synthesis.
Similarly, an increase in Z occurred during germination of lettuce seeds, but the transition to growth following germination was accompanied by a decline in Z and an increase in ZR. Conditions which maintained thermodormancy (prolonged imbibition of seeds at 33°C for 15 days) also resulted in a major accumulation of Z over the same incubation period (1 day), but no coincident or subsequent comparable increase in the level of ZR. The cytokinin 2iP was undetectable during and following germination. However, thermodormant seeds exhibited a small increase in this hormone.
Thermodormancy is accompanied by substantial changes in hormone metabolism which may reflect hormonal cross-talk
The most striking changes potentially reflective of hormonal cross-talk included the marked accumulation of auxin (IAA) levels in thermodormant seeds at 7 days, which was coincident with the major increase in the level of DPA (Figures 9b and 8b). These results suggest a sudden and significant shift in hormone metabolism occurring after 7 days of imbibition under conditions that maintain dormancy and might also suggest an interaction between auxin biosynthesis and ABA catabolism. Significant fluctuations in the amounts of GAs and cytokinins were particularly evident during very early imbibition at 33°C.
As dormant seed is in developmental stasis, it might be assumed that dormancy is associated with a relatively constant steady-state hormonal metabolism. This assumption was not supported by the present study. Further investigation is required to elucidate whether the changes in hormone levels, and therefore pathway fluxes, are a cause or a consequence of changes in interacting signal transduction pathways and/or of primary metabolism.
In conclusion, we have developed and applied a new reversed-phase LC/ESI–MS/MS method for the simultaneous profiling of 15 plant hormones and metabolites during dormancy and germination of lettuce seeds. We show that this method is applicable for the dissection of basic biological questions related to plant hormone metabolism. Because multiple hormones and metabolites are analyzed in one run, the cost per compound is significantly reduced. Also, experimental errors arising from different extraction methods for individual classes of compounds are eliminated. A growing number of mass spectrometry laboratories at universities are acquiring LC–MS/MS instrumentation for proteomics and genomics research. The mass spectrometry equipment required for hormone profiling is moderately priced. Targeted metabolic profiling of plant hormones using LC/ESI–MS/MS with MRM and mode switching can be expanded to include a greater number of signaling compounds to give a more comprehensive view of hormone levels, hormone metabolism, and changes in intracellular mediators (secondary messengers). Control of seed germination is crucial to the survival of seeds, and there are critical checkpoints at the transitions from dormancy to germination and from germination to growth. Plant hormones and hormone metabolism can mediate the fine-tuned regulation of these transitions. Hormone biosynthesis and metabolism can be controlled by both feed-forward and feedback regulation and by interacting signal transduction pathways that may elicit changes in intracellular mediators and post-translational changes of signal transduction components. In future studies, it will be possible to integrate metabolic profiles with information about changes in gene expression (e.g. microarrays) and protein synthesis to clearly identify events that control changes in hormone biosynthesis and metabolism and the physiological and developmental consequences of these changes.
Gibberellin A1, GA3, GA4, GA7, 17,17-d2 GA1 (d2-GA1), and 17,17-d2 GA4 (d2-GA4) were obtained from Dr Lewis Mander at the Australian National University, Canberra, Australia. Unlabeled Z, ZR, IAAsp, 2iP, IPA, IAA, and ABA were from Sigma–Aldrich, Oakville, ON, Canada. The deuterium-labeled and unlabeled ABA metabolites were available from other studies as described: (–)-DPA (Nelson et al., 1991), (–)-7′,7′,7′-d3 DPA (d3-DPA) (Zaharia et al., unpublished), (+)-ABA-GE (Zaharia et al., unpublished), 4,5,8′,8′,8′-d5 ABA-GE (d5-ABA-GE) (Zaharia et al., unpublished), (–)-PA (Balsevich et al., 1994), (–)-7′,7′,7′-d3 PA (d3-PA) (Zaharia et al., unpublished), (+/–)-7′-OH-ABA (Nelson et al., 1991), (–)-5′,7′,7′,7′-d4 7′-OH-ABA (d4-7′-OH-ABA) (Erosa et al., unpublished), and (–)-5′,8′,8′,8′-d4 ABA (d4-ABA) (Abrams et al., 2003). The cytokinin internal standards 10,14,14-d3 DHZ (d3-DHZ), 10,14,14-d3 DHZR (d3-DHZR), 14,14,14,15,15,15-d6 2iP (d6-2iP), and 14,14,14,15,15,15-d6 IPA (d6-IPA) were obtained from Apex Organics LTD, Honiton, Devon, UK. The auxin internal standard 4,7,8,9,10-d5 IAA (d5-IAA) was obtained from Cambridge Isotope Laboratories, Andover, MA, USA.
Standard solutions (10 µm) for each of the labeled and unlabeled compounds were prepared in 1 : 1 acetonitrile:water acidified with 5% glacial acetic acid (v/v). For the selection of diagnostic precursor-to-product ion transitions, 10 µm standards for each labeled and unlabeled compound were directly infused separately with a Harvard Apparatus Pump II at a flow rate of 20 µl min−1 into a quadrupole tandem mass spectrometer (Quattro Ultima, Micromass, Manchester, UK) outfitted with an electrospray ion source (ESI–MS/MS). ES capillary and cone voltages were optimized for the production of the requisite molecular (precursor) ions in negative or positive ionization mode. IAA, IAAsp, ABA, 7′-OH-ABA, PA, DPA, ABA-GE, GA1, GA3, GA4, GA7, and their respective internal standards (Table 1) were analyzed in the negative-ion mode. The cytokinins (Z, ZR, 2iP, IPA) and their internal standards were analyzed in the positive-ion mode. Collision energy and gas (Ar) pressure were then optimized for dissociation of molecular ions into diagnostic fragment (product) ions for each compound. Once the characteristic precursor-to-product ion transitions had been determined, a mixture containing all the unlabeled compounds under investigation, and their internal standards, was separated by reversed-phase HPLC and analyzed by tandem mass spectrometry (RP-HPLC/ESI–MS/MS) with MRM to determine retention times for the various compounds.
High-performance liquid chromatography conditions
High-performance liquid chromatography was used to separate the plant hormones from bulk tissue extracts. An Alliance 2695 separation module (Waters, Milford, MA, USA) equipped with a 100 mm × 2.1 mm, 4-µm Genesis C18 HPLC column (model FK10960EJ, Jones Chromatography, Hengoed, UK) was used with a ternary solvent system comprising acetonitrile (A), de-ionized water (B), and 5% v/v glacial acetic acid in water (C). A 12.5 mm × 2.1 mm, 5-µm Zorbax XDB-C8 guard column (model Z821125926, Chromatographic Specialties, Agilent, Palo Alto, CA, USA) was also used to maintain the performance of the analytical column. Separations were performed using a gradient of increasing acetonitrile content, a constant glacial acetic acid concentration of 7 mm (pH 3.4), and an initial flow rate of 0.250 ml min−1. The gradient was increased linearly from 1.0% A, 98.2% B, 0.8% C to 45% A, 54.2% B, 0.8% C over 20 min and held for 2 min. The acetonitrile content was then increased linearly to 99.2% A, 0.0% B, and 0.8% C over 6 min. These conditions were held for an additional 2 min with an increased flow rate of 0.350 ml min−1. After 1 min, the initial conditions were restored and allowed to equilibrate for 6 min, giving a gradient program of 37-min duration and a sample-to-sample turn around time of 40 min.
For the creation of calibration curves, 10 µm stock solutions of each of the 15 unlabeled compounds were used to prepare eight standard solutions by adding 0, 5, 10, 20, 50, 100, 150, or 200 µl into 5-ml volumetric flasks. For each of the internal standards, a constant amount was added to the volumetric flasks, and then the standard solutions were made up to 5 ml with the 1 : 1 acetonitrile:water containing 5% v/v glacial acetic acid. The amount of internal standard added to each of the volumetric flasks in the dilution series was such that the final concentration for each of the 12 internal standards was 100 pg µl−1. For the preparation of calibration curves by HPLC/ESI–MS/MS, 200 µl of each of the eight solutions in the dilution series was transferred in triplicate into 2-ml eppendorf tubes and dried under vacuum. The standards were reconstituted in 200 µl of 100% methanol, sonicated and spun as described below, and then transferred into 200-µl HPLC vials. Ten microliters of each standard sample was injected into the HPLC system (2695 Waters HPLC, Waters, Mississauga, ON, Canada) linked to a tandem mass spectrometer (Quattro Ultima, Micromass, Manchester, UK). The area beneath the MRM product ion peak was determined for each analyte and IS in the dilution series. The response was calculated according to the formula:
where IS concentration is the known amount of the internal standard added. Calibration curves were created for the 15 different compounds by plotting the known concentration of each unlabeled compound against the calculated response for each standard solution in the dilution series. MRM calibration curves for each compound were generated from triplicate analyses of these standards using the spectrometer software (MassLynx™ v. 3.5, Micromass, Manchester, UK). This software also calculated the limit of detection (LOD) and the limit of quantification (LOQ) for each compound during each LC/ESI–MS/MS run, from which average LOD and LOQ values (n = 24) were determined for each analyte.
Lettuce seeds (Lactuca sativa L. cv. Grand Rapids) (obtained from Early's Seed Store, Saskatoon, SK, Canada) were sterilized in 90% ethanol for 2 min followed by 25% commercial bleach (5.25% NaOCl) for 5 min, and then rinsed three times with sterilized double-distilled deionized water. The sterilized seeds were placed in Petri dishes (100 per dish) containing two layers of Whatman no. 1 filter paper moistened with 4 ml of sterilized double-distilled deionized water. The Petri dishes were wrapped in foil and incubated in the dark at 23 or 33°C (Yoshioka et al., 1998). Radicle emergence was used as the criterion for germination. Germination percentages were determined on the basis of four replicates of 100 seeds each. Germination percentages are presented as the mean ± SD under the two different temperature regimes. Two independent experiments were undertaken.
Extraction and purification of plant material
At the various time points, seeds were removed from Petri dishes, wrapped in foil, and then immediately frozen in liquid nitrogen and stored at −80°C. When all the required material had been collected, the samples were lyophilized for 24 h. The extraction method used was based on the procedure described by Chiwocha and von Aderkas (2002), with some modifications as detailed below. The samples were ground up, and three replicates of dried material for each time point weighing 50–100 mg (exact weight was recorded) were each placed in 4 ml of 99 : 1 isopropanol:glacial acetic acid (v/v) containing 20 ng of each of the deuterium-labeled internal standards for the various compounds, and left to extract at 4°C on an orbital shaker at 300 r.p.m. for 24 h in the dark. The labeled forms of the compounds were used as internal standards for quantification of each compound except for IAAsp, GA3, and GA7, for which d5-IAA, d2-GA1, and d2-GA4 were used, respectively. Each sample was spun at 290 g for 10 min, and the supernatant was transferred to a clean tube. The pellets were re-suspended in 500 µl of the extraction buffer and spun for 10 min, and the supernatant was combined with the initial extracted volume. The extract was then passed through a Sep-Pak C18 column, which had been equilibrated with 4 ml of 100% methanol followed by 4 ml of extraction buffer with the aid of a vacuum apparatus (Supelco Preppy) (Sigma–Aldrich Canada Ltd, Oakville, ON, Canada). The column was rinsed with 500 µl of 80% methanol acidified with 1% glacial acetic acid. The purified extract was dried in a Speed-Vac (Labconco Centrivap Concentrator, Kansas City, MO, USA) and reconstituted with 200 µl of 100% methanol. Prior to HPLC/ESI–MS/MS analysis, the samples were transferred to 2-ml eppendorf tubes, sonicated for 10 min, and spun at 12 740 g for 10 min to remove any particulate matter that might have carried over through the purification step, and then transferred into 200-µl HPLC vials.
For each sample, 10 µl was injected into the RP-HPLC/ESI–MS/MS, and the eluting ions were monitored by MRM. The levels of plant hormones and metabolites in the samples were quantified in relation to their internal standard using the calibration curves that had been generated for each of the compounds. Each experiment was carried out two times, with three replicates per experiment.
Extraction efficiencies for endogenous compounds and recoveries for internal standards
To determine the efficiency of the extraction method used for the isolation of the various compounds as well as the recoveries for the deuterium-labeled analogs, three replicate samples derived from seeds incubated at 33°C for 4 days containing 20 ng of each of the internal standards were used. Each sample was initially extracted for 24 h, and the supernatant was collected and purified as described above. The pellet was re-suspended in extraction buffer and left to extract for an additional 2 h at 4°C in the dark. The supernatant was purified and collected in a separate vial. This procedure was carried out for a third time. For each sample, the three separate fractions were dried down in a Speed Vac (Labconco Centrivap Concentrator, Kansas City, MO, USA) and 20 ng of d9-ABA was added to each fraction as an external standard. The amount of deuterium-labeled internal standard that was recovered in each of the three successive extractions was quantified against the known amount of d9-ABA added after extraction and purification. The percentage of internal standard recovered in each of the three fractions was calculated for each compound.
To calculate the extraction efficiency, the signal for each endogenous compound in the three successive extractions (response) was determined. For each of the three replicate samples, the signals from the endogenous compounds were detected only in the first two extractions. There was no observable signal in the third extraction. Therefore, the sum of the signals in the first and second extractions was calculated and expressed as a percentage of the total response from the three separate successive extractions.
We would like to thank Tim Squires for his technical help. We are also grateful to Irina Zaharia for the PA and DPA in labeled and unlabeled forms and Marek Galka for the synthesis of ABA-GE and 7′-OH ABA (Plant Biotechnology Institute, Saskatoon, Canada). Their manuscripts describing the synthesis of these compounds are in preparation. This research was supported by a Protein Engineering Network of Centers of Excellence (PENCE) grant awarded to A.R.K., S.R.A., A.R.S.R. and A.J.C., and other investigators.