Contributed by Philip N. Benfey, February 1, 2013 (sent for review June 5, 2012)
Analyzing metabolite composition offers a powerful tool for understanding gene function and regulatory processes. Here, we present nontargeted metabolomics assays of five Arabidopsis GFP-tagged lines representing core cell types in the plant root, providing a metabolic map of an organ, composed of its different cell types. Fifty metabolites were putatively identified. The most prominent groups were glucosinolates, phenylpropanoids, and dipeptides. Metabolites were differentially abundant across root cell types and in many cases, this abundance did not correlate with transcript expression, suggesting non–cell-autonomous mechanisms responsible for their targeted localization.
Metabolite composition offers a powerful tool for understanding gene function and regulatory processes. However, metabolomics studies on multicellular organisms have thus far been performed primarily on whole organisms, organs, or cell lines, losing information about individual cell types within a tissue. With the goal of profiling metabolite content in different cell populations within an organ, we used FACS to dissect GFP-marked cells from Arabidopsis roots for metabolomics analysis. Here, we present the metabolic profiles obtained from five GFP-tagged lines representing core cell types in the root. Fifty metabolites were putatively identified, with the most prominent groups being glucosinolates, phenylpropanoids, and dipeptides, the latter of which is not yet explored in roots. The mRNA expression of enzymes or regulators in the corresponding biosynthetic pathways was compared with the relative metabolite abundance. Positive correlations suggest that the rate-limiting steps in biosynthesis of glucosinolates in the root are oxidative modifications of side chains. The current study presents a work flow for metabolomics analyses of cell-type populations.
Recent genome-wide analyses of DNA, RNA, proteins, and metabolites have resulted in a massive amount of novel biological data. Although the data obtained at various levels of biological regulation are of great value, to date, they have mostly been acquired from multicellular organisms using whole-tissue extracts comprising a mixture of cell types or from cultured cell lines. For tissues comprising many cell types, molecules found predominantly in one or a small number of cell types are likely to be diluted when the entire tissue is assayed. For cultured cell lines, the process of establishing and maintaining the cell line frequently leads to changes in molecular composition.
Although metabolism is either directly or indirectly involved with every aspect of cell function, metabolite production and accumulation may be different between various cell types, such that a large portion of our current metabolic knowledge could be misleading. As with other molecules, metabolites that accumulate in specific cell types may not be detected when an entire organ or organism is analyzed. Thus, it is not surprising that single-cell analysis was recently designated as “the new frontier in ‘omics’” (ref. 1, p. 281). Indeed, newly developed methods for genomic analysis at the resolution of a single cell enable new insights into complex biological phenomena (2). In organisms such as plants, the intercell variability in metabolites may be even more pronounced, because it has been estimated that about 200,000 metabolites exist in plants (3). These include a large number of the so-called “secondary” or “specialized” metabolites that are of key importance to the plant’s interaction with the environment. Understanding metabolism and the function of small molecules in specific cell types requires the isolation of individual cell populations for metabolic analysis. For sensitivity reasons, current metabolomics methods require samples that contain a relatively large number of cells (4).
Although plant developmental programs have been extensively studied, the information regarding regulation of central and specialized metabolism is still lacking, and studies that demonstrate nonuniform metabolite accumulation are rare. Laser capture microdissection has been used in several studies to harvest small amounts of cells of specific types. Hölscher and Schneider (5) examined secretory cavities from leaves and flowers of Dilatris species by means of cryogenic 1H NMR spectroscopy and HPLC. This method allowed the identification of several metabolites (e.g., three phenylphenalenones) from specialized cells. Amantonico et al. (6) studied the profile of selected metabolites (ADP, ATP, GTP and UDP-glucose) in single cells of Closterium acerosum using negative-mode MALDI-MS. In an earlier example, the hypothesis that intercellular transport of carbon occurs by diffusion during photosynthesis in C-4 plants was put to a test. Concentration gradients were found between the bundle sheath and mesophyll for 3-phosphoglycerate, triose phosphates, malate, and pyruvate during photosynthesis. These concentration gradients play roles in the regulation of sucrose synthesis and the Calvin cycle (7).
The Arabidopsis root is well suited for the analysis of biological, and especially developmental, processes within individual cell types because of its relatively simple radial organization and its mode of continuous development from a set of stem cells. Previous analyses have examined spatiotemporal gene expression patterns using FACS to sort and profile different GFP-marked cell types in the Arabidopsis root (8, 9). The gene expression map drawn by Birnbaum et al. (8) localized the expression of more than 22,000 genes in five different cell types of the Arabidopsis root and three developmental zones. Their findings strongly suggested that patterns of gene expression traverse traditional anatomical boundaries and correlated groups of genes to specific cell fates. Following this work, Brady et al. (9) studied microarray expression profiles of root developmental time points and provided a comprehensive map of nearly all (14 of 15) cell types within the Arabidopsis root. Their work characterized a multitude of developmental transcriptional programs that are controlled in space and time. Profiling of gene expression from different cell samples of shoot apical meristems by Yadav et al. (10) has further expanded the use of this approach. They isolated three cell-type populations from shoot apical meristems and demonstrated that cell-type-expression profiling is sensitive in identifying transcripts expressed in specific subsets of shoot-meristem cells. Recently, the plant growth regulator auxin was examined for its distribution throughout cell types in the Arabidopsis root following FACS sorting of 14 GFP-marked cell populations (11). This study suggested gradients of auxin concentration within the Arabidopsis root tip, with a maximum in the quiescent center. FACS sorting was also demonstrated not to result in ion leakage or changes in auxin metabolism.
In the current study, we modified the established method for FACS-based, GFP-marked cell isolation from Arabidopsis roots of specific root cell types (12) to carry out metabolomics analysis using high-resolution MS (13, 14). Cell extracts were processed and analyzed by means of ultraperformance liquid chromatography (UPLC) coupled to a quadrupole TOF (qTOF) MS detector. The use of a high-resolution MS detector allows structural elucidation to a certain extent, and thus nontargeted analysis of metabolites. These nontargeted assays were carried out in core root cell populations from five GFP marker lines representing (i) columella (PET111 enhancer), (ii) epidermis and lateral root cap (WEREWOLF gene promoter), (iii) cortex (CORTEX, At1g09750 gene promoter), and (iv) endodermis and quiescent center (SCARECROW gene promoter) and stele (WOODEN LEG gene promoter) (8, 15). Fifty metabolites were putatively identified, with the most prominent classes being glucosinolates (GSLs), phenylpropanoids (PPs), and dipeptides (DPs). The resulting metabolic profiles demonstrated particular metabolite accumulation patterns across cell types. These metabolic profiles were correlated with mRNA expression of the same cell types, revealing three GSL gene expression patterns with strong correlations to metabolite accumulation and a correlation of PP peak of accumulation in the cortex. Our findings should inform the present approaches that predict activity of metabolic pathways through assessment of transcript levels or carry out metabolic analyses of extracts from entire organs or organisms.
The first challenge in carrying out FACS-based metabolomics assays in specific cell populations was optimization of the method used previously for transcriptome analysis (8, 9) (Fig. 1). This procedure typically includes several steps before RNA isolation, namely, seedling growth, protoplasting, and cell isolation using FACS (Fig. 1). Modification of the established procedure was required for three main reasons:
i) The solution used for sorting (i.e., FACS sheath fluid) typically contains PBS, which contaminates the MS detector, resulting in signal deterioration that impedes MS-based analysis. The PBS solution in the FACS sheath fluid also contains large amounts of analytes of nonplant origin.
ii) Metabolites cannot be amplified, as in the case of mRNA; thus, many more cells are required per sample for signal detection.
iii) Metabolites are subject to enzymatic and chemical degradation, and samples were thus quenched on collection using ∼60% (vol/vol) methanol (Methods).
Work flow for cell type-specific metabolite mapping in Arabidopsis roots. Following sample preparation and isolation of cells belonging to a particular cell type, cells are extracted and used for metabolomics assays by means of high-resolution MS. The third phase includes various data analysis steps in which robust mass signals and the corresponding putatively identified metabolites are extracted and used to generate metabolic profiles across different cell types.
To resolve the first issue, we reduced the initial PBS concentration in the FACS sheath fluid by half, without affecting cell isolation. Additionally, we introduced a divert valve after the LC column and before the MS detector. This prevented the insertion of PBS-saturated eluate into the ionization chamber and reduced MS contamination (Fig. S1). Following cell sorting, FACS sheath fluid was collected and served to identify cross-contamination between sample runs and artifactual mass signals originating from the FACS system (herein, samples are termed blanks; Methods).
To resolve the second issue, we used sequential resuspensions in smaller volumes and centrifugations of the extracts after each lyophilization step, resulting in increased sample concentration (see below). We examined the amount of cells required for obtaining quantifiable mass signals in ion chromatograms. The analysis of different amounts of sorted cells per sample (i.e., 5.5 × 105, 11.5 × 105, 14.5 × 105) showed linearity for the masses examined (Dataset S1) and suggested that using 5–7 × 105 cells was sufficient to obtain quantifiable mass signals (Fig. S2). Thus, 7 × 105 cells were used for downstream assays to allow quantification of as many metabolites as possible with the limitations of obtaining these large numbers of cells. To acquire the required amount of cells, each sample was concentrated up to ∼700-fold by means of three sequential lyophilization steps, followed by resuspension of pellets in 75% (vol/vol) methanol with 0.1% formic acid. Whole roots from each marker line were also profiled, and no significant differences were found between the metabolic profiles of the whole roots of different GFP-marked lines. A representative chromatogram of Arabidopsis WT (Col-0) whole roots is provided with the putative identifications of abundant metabolites detected in our system (Fig. S3).
To identify and exclude masses altered due to protoplasting and sorting, the intensities of masses isolated from WT (Col-0) whole roots were compared with masses from WT (Col-0) sorted protoplasts following normalization of data (SI Methods). Masses that were more abundant following the process of protoplasting and sorting (Fig. S4) were excluded from the mass list.
Once the sample preparation had been optimized for UPLC-qTOF-MS assays, we carried out metabolomics analyses of five major cell-type populations in the Arabidopsis root [representative total ion current chromatograms are displayed in Fig. 2]. For the core experiment, we used four biological replicates of WEREWOLF (WER/epidermis) and SCARECROW (SCR/endodermis) and three biological replicates of PET111 (columella), CORTEX, and WOODEN LEG (WOL; stele). The metabolic profiles obtained were analyzed using the XCMS peak picking and alignment software (14, 16, 17) and a metabolomics data quality control (QC) procedure developed by Brodsky et al. (18) and implanted by Rogachev and Aharoni (19). Briefly, the mass signals obtained were clustered (for grouping those masses associated with the same metabolite; SI Methods) and filtered from nonplant origin masses (in comparison to blank samples), and metabolites of interest were putatively identified. A detailed flow chart of the core experiment data analysis procedure is presented in Fig. S4.
Metabolic profiles of specific cell types in the Arabidopsis roots. (A) Representative UPLC-qTOF-MS total ion current (TIC) chromatograms of the five cell-type populations and a blank sample. The mass signal intensity is measured in values of ion current, and the y axes represent the relative peak abundances (%). The y axes of the cell population chromatograms are linked, and 100% abundance of each chromatogram corresponds to the TIC of 5⋅104, to enable comparison of the different chromatograms. Selected GSLs are marked: (1) 4-methylthiobutyl GSL (4MTB; metabolite 25 in Table S1); (2) 8-methylsulfinyloctyl GSL (8MSOO; metabolite 27 in Table S1); (3) 4-methoxyindole I3M GSL (4MO-I3M: metabolite 31 in Table S1); (4) 7-methylthioheptyl GSL (7MTH; metabolite 33 in Table S1); and (5) 8-methylthio-octyl GSL (8MTO; metabolite 34 in Table S1). These nontargeted assays were carried out in core root cell populations from five GFP marker lines representing the following: endodermis, SCARECROW gene promoter; epidermis, WEREWOLF gene promoter; columella, PET111 line; cortex, CORTEX line; and stele, WOODEN LEG gene promoter. (B) Typical TIC chromatogram of an Arabidopsis whole-root extract (250 mg). This chromatogram has another y-scale, because it contains high-intensity signals, derived from a relatively much bigger sample.
Clustering of 19,410 mass signals (in the negative ionization mode) and 22,887 mass signals (in the positive ionization mode) resulted in 12,275 and 11,465 clusters from the negative and positive modes, respectively (including singletons; SI Methods and Fig. S4). This output was filtered by the following: (i) comparing mass intensities from samples with the corresponding intensities from blanks to remove artifact masses, (ii) setting a threshold of the natural logarithm of the mass intensity to 5 to reduce noise. The application of the mass filtering procedure described above (see also SI Methods) resulted in 714 and 962 mass signals in the negative and positive modes, respectively. The selected masses were then examined for robustness, by crossing the XCMS output from the main dataset with two additional XCMS outputs (chosen using a QC procedure; SI Methods). Finally (iii) masses with higher intensity in the protoplasts that likely originated during protoplasting and sorting were excluded. By manually picking masses, examining the main ion mass in each cluster, and verifying a reliable accumulation profile in the different replicates throughout the different cell types, we were able to detect 112 of the above robust masses as the molecular ion masses of metabolites (Dataset S2). Fifty of these masses were putatively identified by using their accurate mass and tandem MS (MS/MS) fragmentation patterns (Table S1).
The reproducibility of the biological replicates was checked using a Pearson correlation matrix (Fig. 3A). While correlations among different cell-type samples mostly ranged between 0.32 and 0.85, all correlations among samples of the same cell type were above 0.85 (Fig. 3A and Fig. S5). Reproducibility was also exemplified by overlaying the chromatograms obtained from the biological replicates of each particular cell type (chromatograms of a representative ion mass are illustrated in Fig. S6). The similarity in metabolic profiles of biological replicates and differences between the profiles of cell type-specific extracts can be clearly observed with principal component analysis (Fig. 3B). Although the epidermis and stele cell types displayed high similarity in metabolic profiles, the endodermis and cortex were the most diverse cell types. A scatterplot superimposing mass abundances from intra- (endodermis vs. endodermis) and inter- (endodermis vs. epidermis) cell-type samples (Fig. 3C) also demonstrated the reproducibility of the data and differences between cell-type metabolome. The diversity of mass intensities from the endodermis/epidermis samples is shown in Fig. 3C by the wide distribution of the blue dots, as opposed to the tight distribution of the endodermis/endodermis mass intensities around the diagonal line, suggesting high similarity.
Reproducibility of the cell-type metabolic profiling procedure. (A) Pearson correlation coefficient of robust masses of plant origin among replicates of all five cell-type populations is higher than 0.85 (pink, correlation >0.95; dark orange, correlation of 0.90–0.95; light orange, correlation of 0.85–0.90) (all correlations are shown in Fig. S5). (B) Principal component analysis (PCA) of robust masses (detected in the negative mode) derived from biological replicates of the different cell-type populations. The columella separates from other cell types on the PC2 component; however, this component represents 11.6% of the total variance, whereas the PC1 component represents 62.5%. (C) Scatterplot between different cell-type mass abundances (endodermis-epidermis comparison, dark blue), superimposed by a scatterplot of intrapopulation mass abundances (endodermis-endodermis comparison, pink). The high variability between the two groups is shown by the wider distribution of the blue dots, whereas a tight distribution is seen in the scatterplot of endodermis-endodermis mass intensities around the diagonal line.
The putatively identified metabolites represent diverse chemical classes, with GSLs being the most abundant ones detected (Fig. 2). The analysis further showed that in our experimental setup, cortex cells contain the highest amount of abundant metabolites, whereas stele cells contain the lowest, as seen in the chromatograms (Fig. 2) and the heat map of masses (Fig. 4). The full chromatograms are presented in Fig. S7. Fig. 4 also shows the relative abundance of masses in whole roots following quantile normalization (Methods). Although the current study focuses on specific cell-type metabolic profiles, whole roots were also focused to confirm, whenever possible, that the masses detected are of plant origin. They were also used for carrying out MS-MS experiments in cases when the mass signal appeared highest in the whole-root sample compared with either one of the cell-type samples. As mentioned previously, whole-root analysis was also performed to allow subtraction of the protoplasting and sorting effects. Normalized intensities of most masses in the whole-root samples were lower than the corresponding intensities in at least one cell type, further underlining the importance of the cell type-specific analysis for high-resolution metabolite profiling. Thus, most metabolites, including GSLs and DPs, were enriched in one or more cell-type populations. An exception was a group of masses that included flavonols, which were more abundant in the normalized whole-root samples.
Differential accumulation of all robust masses in the different cell types. (A) Global heat map of all robust mass signals obtained in the ESI (−) mode. Following quantile normalization (Methods), average ion mass intensities of sample replicates (logE-transformed) are presented as the ratio to the maximal intensity detected for each mass. The yellow line separates the cell layers and the whole-root samples. (B) Five representative clusters of the heat map provided in a greater detail. K-hex-deox, kaempferol hexose deoxyhexose (metabolite 46 in Table S1); Q-dideox, quercetin dideoxyhexose (metabolite 47 in Table S1); Q-O-dihex-O-deox, quercetin-O-dihexose-O-deoxyhexose (metabolite 43 in Table S1); Q-deox-hex-deox, quercetin deoxyhexose-hexose-deoxyhexose (metabolite 44 in Table S1); 7MTH, 7-methylthioheptyl GSL (metabolite 33 in Table S1); I3M (metabolite 26 in Table S1); 3BZO (metabolite 30 in Table S1); 8MTO, 8-methylthio-octyl GSL (metabolite 34 in Table S1). (C) Relative accumulation of representative metabolites of the clusters shown in B. Bars represent SEs. Statistically significant differences are represented for all pairwise comparisons using a two-way ANOVA. Pairwise comparisons are as follows: a = cortex_columella, b = cortex_epidermis, c = cortex_stele, d = cortex_endodermis, e = columella_epidermis, f = columella_stele, g = columella_endodermis, h = epidermis_stele, I = epidermis_endodermis, and j = stele_endodermis. ***P < 0.001; **P < 0.01; *P < 0.05.
Thirteen GSLs were identified in this study. Those of the aliphatic class predominantly accumulated in the cortex, whereas indole GSLs largely accumulated in the columella cells (Fig. 5A). The abundance of GSLs could be further divided into three covarying groups with distinct differential accumulation as determined by k-means clustering (Pearson correlation coefficient is greater than 0.91 for all groups). Across all three groups, accumulation was highest in the cortex and columella and lowest in the endodermis and stele. In the first group (Fig. 5A), high correlation was observed for methylsulfinyl GSLs, including methylsulfonyloctyl GSL (metabolite 28 in Table S1), 6-methylsulfinylhexyl GSL (metabolite 22 in Table S1), 7-methylsulfinylheptyl GSL (metabolite 23 in Table S1), and 8-methylsulfinyloctyl GSL (metabolite 27 in Table S1), which showed similar levels in the cortex and columella and lower levels in the epidermis, endodermis, and stele. In the second group (Fig. 5B), primarily indole GSLs were represented: indol-3-yl-methyl GSL (I3M; metabolite 26 in Table S1), 1-methoxyindole I3M GSL (1MO-I3M; metabolite 29 in Table S1), and 4-methoxyindole I3M GSL (4MO-I3M; metabolite 31 in Table S1), as well as the modified secondary side chain 3-benzoyloxypropyl GSL (3BZO; metabolite 30 in Table S1), showed a peak of accumulation in the columella. Interestingly, in Arabidopsis, benzoyl-modified GSLs (e.g., 3BZO in the second group) have previously only been identified in the seed (20, 21). In the third group (Fig. 5C), aliphatic methylthio-GSLs, represented by 4-methylthiobutyl GSL (4MTB; metabolite 25 in Table S1), 7-methylthioheptyl GSL (7MTH; metabolite 33 in Table S1), and 8-methylthio-octyl GSL (8MTO; metabolite 34 in Table S1), predominantly accumulated in the cortex layer and to a lower extent in all other cell types. A second benzoyl-modified GSL, 4-benzoyloxybutyl GSL (4BZO; metabolite 32 in Table S1), showed a similar accumulation pattern.
Abundance of GSLs in specific cell types is tightly coregulated with transcripts encoding side chain redox enzymes. (A–C) GSLs display three distinct patterns of accumulation as determined by k-means clustering, with a Pearson correlation coefficient greater than 0.91 in all three groups. Metabolite abundance is visualized as log10, and error bars represent SD. The abundance of 4-methylsulfinylbutyl GSL (4MSOB; metabolite 24 in Table S1) could not be distinguished from that of the m/z = 420.05 fragment of 4-methylthiobutyl GSL (4MTB; metabolite 25 in Table S1). MSO, methylsulfonyloctyl GSL (metabolite 28 in Table S1). (D) Expression of GSL biosynthesis genes across the Arabidopsis root (9). Red, expressed in the five cell types examined; pink, not expressed in a root cell type examined for metabolite content; gray, not expressed in roots. Blue quadrangles indicate genes whose relative expression is highly correlated with the relative abundance of GSLs. 1MO-I3M, 1-methoxyindole I3M GSL; 3MSOP, 3-methylsulfinylpropyl GSL; 4MO-I3M, 4-methoxyindole I3M GSL; 4MSOB, 4-methylsulfinylbutyl GSL; 4MTB, 4-methylthiobutyl GSL; 5MSOP, 5-methylsulfinylpentyl GSL; 6MSOH, 6-methylsulfinylhexyl GSL; 7MSOH, 7-methylsulfinyl heptyl GSL; 7MTH, 7-methylthioheptyl GSL; 8MSOO, 8-methylsulfinyloctyl GSL; 8MTO, 8-methylthio-octyl GSL.
In recent years, multiple studies have demonstrated an important role of regulation of structural gene expression and corresponding metabolite accumulation (22). Of the 36 GSL-associated genes examined, 29 were expressed in the root (9). Seventeen of these genes were expressed in the five root cell types studied, such that they could be assayed for their correlation with metabolite accumulation (Fig. 5D). Similar correlation analyses, based on patterns of gene expression and chromatin state, were recently used in human cell types to identify putative target genes and to predict the cell type-specific activators and repressors that modulate them (23). We determined the Pearson correlation between the identified GSLs of the three accumulation patterns and the mRNA expression of the corresponding GSL pathway enzymes (SI Methods and Datasets S3 and S4). The methylsulfinyl GSLs represented in group “a” were best correlated with the gene expression of GLUCOSINOLATE HYDROXYLASE (GS-OH), encoding a side chain-modifying enzyme involved in hydroxylating alkenyl GSLs (24), as well as with FLAVIN-MONOOXYGENASE GLUCOSINOLATE OXYGENASE 5 (FMO-GSOX5), encoding an enzyme catalyzing the thiol oxidation (i.e., the conversion of methylthioalkyl GSLs into methylsulfinylalkyl GSLs) (22). The accumulation of group “b” GSLs, including indole GSLs, was correlated extremely well with the gene encoding FMO-GSOX5 and with the gene encoding FMO-GSOX4, which also catalyzes thiol oxidation. Group “c” GSLs, comprising methylthio-GSLs and 4BZO, showed a striking correlation with GS-OH gene expression, with low correlation to the other GSL genes examined (Datasets S3 and S4). Interestingly, all genes exhibiting mRNA expression that was well correlated to metabolite accumulation corresponded to oxidative enzymes taking part in secondary modifications of the core structure of aliphatic GSLs.
Ten glycosylated flavonols that were putatively identified in the various cell types, predominantly accumulated in the cortex (Fig. 6 and Table S1). Eleven genes in the PP biosynthesis pathway showed preferential mRNA accumulation in the cortex. Of these, 4-COUMARATE:COA LIGASE 3 (4CL3) and FLAVONOID 3′-HYDROXYLASE (F3′H) displayed the best-correlated expression with flavonol accumulation (Dataset S5). Whereas quercetin glycosides correlated best with F3′H, kaempferol glycosides were better correlated with expression of 4CL3 (Dataset S5).
Phenylpropanoids (PP) are enriched in the cortex cell type. A scheme of the PP pathway with relative mRNA expression and metabolite accumulation in the core five cell types examined is shown. Gene expression (9) and metabolite accumulation are colored in representations of a root transverse section and a cut-away of a root tip. (A) Expression of flavonoid biosynthesis genes (log2) is enriched in the cortex relative to other cell types. (B) The majority of PPs show maximal abundance (log10) in the cortex relative to other cell types.
Although small peptides (e.g., DPs and tripeptides) are considered as a nitrogen source for plant growth, data regarding these molecules, their abundance, and their roles in plants are limited (25). Nontargeted analysis of extracts derived from the five sorted cell-type populations suggested the accumulation of DPs in specific cell types. Most of these DPs were composed of branched-chain amino acids, predominantly leucine and isoleucine (Fig. 7). The abundance of most DPs was significantly higher in the epidermis and endodermis compared with the other cell types (Fig. 7 and Dataset S6). Overall, we detected 14 DPs in cells of the Arabidopsis root (Table S1); of these, 5 were identified based on synthesized standards, whereas an additional 5 were putatively identified as isomers of the standards by comparison with mass fragments. The DPs’ accumulation profile in cell populations showed low correlation to the gene expression of the Arabidopsis DP transporters, PEPTIDE TRANSPORTER 2 (AtPTR2) and AtPTR5 (Dataset S7), which were suggested earlier to be involved in the local accumulation of these metabolites (25).
Dipeptides (DPs) are highly abundant in the root endodermis and epidermis cell types. Leu-Val (metabolite 10 in Table S1) and Val-Leu (metabolite 11 in Table S1) as well as Leu-Val (metabolite 13 in Table S1) and Val-Leu (metabolite 14 in Table S1) isomers are presented together, because their mass intensities could not be separated. Error bars represent SE (see correlation matrices for DPs in all cell types in Dataset S6).
In this study, we demonstrate the value of performing metabolic analyses on sorted cell populations as a tool for high-resolution investigation of metabolism in specific cell types. By combining two separation technologies (i.e., FACS, LC), we were able to dissect metabolites of an organ into five different cell-type populations with typical metabolic profiles. Due to sensitivity reasons, metabolomics studies are limited by the need to analyze large numbers of cells, which limited the number of cell types we could examine. To accomplish this, we had to overcome several technical issues during both sample preparation and analysis. Among these were (i) contamination during sample preparation, especially during FACS sorting (e.g., contaminants originating in the FACS instrument and the sheath fluid); (ii) low amount of starting material, requiring intense sample concentration in some cases; and (iii) potential alteration of metabolic profiles during protoplasting and FACS sorting. By addressing these issues, we were able to detect metabolites that are highly enriched in certain cell types, some of which have not been previously described in Arabidopsis roots. Our results shed light on the location and mechanism of root GSL and PP biosynthesis, as well as introducing a yet unexplored group of Arabidopsis root metabolites, DPs, with distinct patterns of accumulation.
The effects of the protoplasting and sorting procedures on Arabidopsis root metabolites were examined by mining those that were more abundant in the protoplasted and sorted sample set (compared with the unprotoplasted/sorted set), because it is practically impossible to determine the reason for a decrease in intensity of protoplasted cells samples: Not all cells are going through protoplasting, and more cell types are present in the Arabidopsis root than studied in the current work. The analysis of the sorting and protoplasting effects on the intracellular metabolic profile suggested that, as in the case of auxin (11), many metabolites are not influenced by these procedures during the high-throughput protocol used in the current study. Roughly 10% of the robust masses were significantly higher in protoplasts compared with whole-root masses following quantile normalization (64 of 692 and 107 of 948 in the negatively and positively ionized mass signals, respectively; Fig. S4), and these were removed from the dataset.
Of the five cell-type populations examined, the epidermis and stele were most similar in their accumulation of metabolites, implying that these cell types share metabolic features. The enrichment of GSLs in cortex cells suggested a role for protective specialized metabolites in this cell type. Previous reports provide evidence for a role of the root cortex in the defense of Brassicaceae plants against Plasmodiophora brassicae infection (26, 27), which may be related to these findings, especially because clubroot disease is one of the most damaging diseases within this plant family (26). In other plant species, the prime location of pathogen or symbiotic bacterial infection of the root also occurs through the cortex as with the potato soil-borne pathogen Dickeya sp. IPO2254, which primarily infects root cortex cells (28). PPs, especially flavonols, which are also relatively highly accumulated in the cortex, provide protection from different biotic and abiotic stresses (29). Taken together, metabolites from these two distinct pathways are highly accumulated in the cortex and offer higher levels of chemical protectants than in other cell types within the root. Not all GSL biosynthesis genes are expressed in the root (Fig. 5D); moreover, none of the previously identified GSL transcriptional regulators was shown to be expressed in the root according to previously defined and validated thresholds (15). This corroborates previous evidence indicating that GSLs are transported from the leaves to the roots via the phloem (30). However, the accumulation of GSLs across root cell types in this study showed high correlation with the expression of oxidation enzymes, suggesting that side chain oxidation could be a rate-limiting factor in the biosynthesis of GSLs in the root. Because many of the enzymes in the first steps of the pathway are not present in the root, we propose that GSLs are transferred from a tissue in which they are produced (probably shoot) into the root and are further metabolized by FMO-GSOX4 and FMO-GSOX5, which provide a controlling step to the synthesis of the downstream root GSL metabolites. Modifications in GSL side chains are of particular importance because the biological activity of the GSL hydrolysis products is determined to a large extent by the structure of the side chain. FMO-GSOX enzymes play important roles in such modifications (31), and the correlation between FMO-GSOX and methylsulfinyl GSL accumulation was to be expected. The striking correlation of the patterns of FMO-GSOX5 and indole GSL accumulation may suggest a possible involvement of this enzyme in the biosynthesis of indole GSLs as well. Another surprising correlation was found between the GS-OH gene expression and methylthio-GSL accumulation, implying that it may have a yet unknown role in the methylthio-GSL pathway. It should be noted, however, that the literature describing these genes strongly suggests a role for GS-OH as a short-chain modifier and the modification of indole GSL by cytochrome P450. The different accumulation of 3BZO and 4BZO in the columella and epidermis (Fig. 5 B and C) is surprising; it is probably due to the high variability in the accumulation of 4BZO in the columella (as seen by the large error bar in Fig. 5C; this is indeed the only identified metabolite with such high variability) and does not necessarily imply relatively selective expression of the two benzyl GSLs in these cell types.
Although large differences in abundance of the GSLs (5- to 25-fold in most cases) were observed between the different cell-type populations, the differences in the accumulation of flavonoids across cell types were of a much lower magnitude (1.5- to 6-fold). Thus, it seems that the accumulation of flavonoids is less cell type-dependent, although enriched in the cortex, in agreement with gene expression data (9). Because some PPs are lost during protoplasting due to accumulation in the cell wall domain, the comparison of metabolome and transcriptome data is not straightforward. Assuming that this loss is not different between the different cell types, the transcripts that correlate best with the pattern of flavonol accumulation correspond to 4CL3 and F3′H. The Pearson correlation between the accumulation of metabolites and mRNAs is considerably lower in the flavonol pathway than in the GSL pathway. This suggests an additional role for transport, posttranscriptional or posttranslational regulatory mechanisms in the PP pathway in the Arabidopsis root, which are yet to be discovered.
The relatively high concentration of DPs in root epidermis is in good agreement with earlier work, such as the study by Komarova et al. (25), which suggested that DPs should be considered as a nitrogen source and transport form in plants. The high DP concentration in the endodermis was unexpected and suggests that there may be additional roles for these metabolites. The origin of the DPs in roots may be as products of protein degradation taking place differentially among cell types. The particular accumulation of DPs may also be a result of the specific activity of a transporter resembling members of the nitrate/peptide transporter (NRT/PTR) family (25), although the comparison of metabolite abundance profiles to the gene expression profiles of AtPTR5 and AtPTR2 failed to show a positive correlation. The identification of these DPs in Arabidopsis roots, including their enrichment in particular cell layers, might provide clues to understanding their role in plants.
To summarize, this study represents an initial attempt to carry out nontargeted metabolomics analysis of an organ at a cell-type level. Although the methodology presented has some limitations, such as in the case of metabolites that show rapid turnover in vivo or after tissue disruption and the lack of cell wall metabolites in protoplasts, it may open the way for additional efforts to follow more precisely the metabolism of multicellular organisms. Such future endeavors may include (i) cell-type metabolic profiling of plant tissues other than roots and in additional plant species; (ii) cell-type metabolomics in organisms outside the plant kingdom (e.g., mammalian systems); (iii) profiling of additional classes of metabolites, such as lipids or ions (through lipidomics and ionomics approaches); and (iv) combining other analytical systems, such as NMR and GC-MS. When applying the methodology presented in the current study to examine rapidly turning over metabolites, one should examine recovery in specific cell types to correct for differential loss of metabolites. The findings here clearly indicate that many yet undiscovered mechanisms of metabolic regulation are active in a cell type-specific manner, introducing a unique perspective to our understanding of cell heterogeneity. They provide a framework for metabolomics analyses and chemical phenotyping at a cell-type level.
Plants were germinated, grown, and harvested according to an established protocol (12). Briefly, five marker lines were used to sort different cell populations in the Arabidopsis Col-0 root, expressing GFP in the epidermis and lateral root cap (WEREWOLF gene promoter), the cortex (CORTEX gene promoter line), the endodermis and quiescent center (SCARECROW gene promoter), the columella (PET111 line), and the stele (WOODEN LEG gene promoter) (8). For the final experiment, we used four biological replicates of WEREWOLF and SCARECROW and three of PET111, CORTEX, and WOODEN LEG. Nylon mesh was placed on top of the solidified media [1.0% agar (10 g), 0.5 g of Mes (M-2933; Sigma), 1% sucrose (10 g), 4.33 g of MS salts (catalog no. 11117-066; Invitrogen), pH to 5.7–5.8 with KOH]. Sterilized seeds were evenly planted in two rows with about 500–1,000 seeds per row. Following vernalization (1 d at 4 °C), roots were grown in 16 h of light and 8 h of dark conditions at 22 °C. Roots were disconnected from the plant and cut into pieces with a no. 10 surgical blade 6 d after introducing the seeds to the growth chamber. Roots were then digested with enzymes for 1 h to release protoplasts, which were collected by centrifugation and then sorted by FACS directly into a cold methanol solution. Seedlings harvested from five to six plates were placed on a cell strainer in one Petri dish and incubated in an enzyme buffer [600 mM mannitol, 2 mM MgCl2, 0.1% BSA, 2 mM CaCl, 2 mM Mes, 10 mM KCl, 1.5% cellulase, 0.1% pectolyase (vol/vol; pH 5.5)] at room temperature for 60 min, swirling on a benchtop orbital shaker. Roots were agitated by spreading them over the cell strainer and applying gentle pressure on them over the filter. Protoplasts were then moved to 50-mL tubes and centrifuged at 170 × g for 5 min. Supernatant was removed, and the cells were resuspended in 500 μL to 1 mL of the above enzyme buffer, this time with no enzymes.
GFP-expressing cells were isolated using a FACS (FACSVantage; Becton Dickinson) fit with a 70-μm nozzle at a rate of 5,000 events per second and fluid pressure of 20 psi. Cells were sorted as described by Brady et al. (9). Blanks were taken in the following manner: After each sample, FACS was switched to the test mode and sheath fluid was collected in FACS tubes with the same volume as for a sample.
Cells were sorted directly into a methanol solution, making a final solution of ∼60% (vol/vol) methanol for immediate enzyme quenching, and were kept at −80 °C until sample preparation and on ice during sample preparation. Blanks taken from the FACS were treated in the same manner. Protoplasted and sorted cells were unified to attain 7 × 105 cells per sample and were ultrasonicated for better extraction of metabolites. The extracts were concentrated by three sequential lyophilizations and resuspensions in 75% (vol/vol) methanol + 0.1% (vol/vol) formic acid in decreasing volumes. The final sample preparation contained 70 μL of cell extract, and 5 μL was injected into LC-MS instrument.
Metabolite analyses were performed using a UPLC-qTOF system (HDMS Synapt; Waters), with the UPLC column connected online to a photodiode array (PDA) detector and then to the MS detector. A 100 × 2.1-mm i.d., 1.7-μm UPLC BEH C18 column (Waters) was used for the separation of metabolites. The mobile phase consisted of 0.1% formic acid in acetonitrile: water [5:95, (vol/vol); phase A] and 0.1% formic acid (vol/vol) in acetonitrile (phase B). The linear gradient program was as follows: 100 to 72% phase A over 22 min, 72 to 63% phase A over 7 min, 63 to 35% phase A over 10 min, and 35 to 0% phase A over 7 min; held at 100% phase B for further 2 min; and then returned to the initial conditions (100% phase A) in 0.5 min and conditioning at 100% phase A for 1.5 min. Following preliminary experiments, the retention period of 1.8–35 min was used for analysis. The flow rate was 0.3 mL/min, and the column temperature was kept at 35 °C. Masses of the eluted compounds were detected by a qTOF HDMS Synapt mass spectrometer, equipped with an electrospray ionization (ESI) source. Acquisition was performed separately in positive and negative ESI modes. The following settings were applied during the LC-MS runs: divert valve (Rheodine), excluding 0–1.8 min and 43–50 min following injection; capillary spray at 3.0 kV; cone voltage at 28 eV; and collision energy at 4 eV. For the MS/MS experiments, collision energies were 10–25 eV for positive mode and 15–40 eV for negative mode. Full-scan mass spectra were acquired from 50–1,500 Da. Argon was used as the collision gas for collision-induced dissociation MS/MS experiments. The mass spectrometer was calibrated using sodium formate, and leucine enkephalin was used as the lock mass. The UV spectra were acquired from 200 to 500 nm. A mixture of 15 standard compounds, injected after each 10 samples, was used for QC. MassLynx software version 4.1 (Waters) was used to control the instrument and calculate accurate masses. When available, metabolites were identified using standard compounds by comparison of their retention times, UV spectra, and MS/MS fragments. In case the corresponding standards were not available, compounds were putatively identified as previously described (14).
For relative quantification of metabolites, the LC-MS chromatograms were analyzed by the XCMS peak-picking/peak-alignment software (14, 16, 17) for positive and negative ionization modes separately. The analyses were performed in two separate batches. In preliminary experiments, no statistically significant association was detected between different cell-type samples and blanks (Figs. S8 and S9); however, we further examined the mass signals with increased intensities in protoplasts and excluded them from the analysis. The first set, the main dataset, comprised samples of five cell types (n = 3–4 for every cell type), with the corresponding technical controls (blanks; Methods, FACS) and one group of whole roots (n = 4). We used this dataset for the selection of peaks with significant cell layer vs. blank differentiation. In the second set, the “protoplasting dataset,” the main dataset was analyzed, together with Arabidopsis WT whole roots and protoplasts of Arabidopsis WT from the same batch to evaluate the effect of the protoplasting and sorting procedures on metabolic profiles. For each dataset separately, the XCMS parameters were optimized as previously described (18), based on the general QC measure: an average Fisher z-correlation–based similarity of samples in replicate groups. For the best XCMS outputs in each set (for the negative and positive modes separately), a quantile normalization of peak log intensities was performed in two metagroups of samples: (i) biological samples and (ii) technical controls. The quantile normalization method was adapted from microarray data preprocessing (32). In an earlier study, such statistical quantile normalization was performed and demonstrated its application to the analysis of biological replicates in LC-MS data (such as the ones used here). It showed comparable performance to that of the currently applied standard-based methods (18). The regions of concentrated discrepancies between replicates were detected, and deviating samples were mutually equalized by weighted averaging across these regions.
We thank Dr. Sara Rubinraut for her help in the synthesis of DP standards. The work in the A.A. laboratory was supported by the European Research Council Project SAMIT (Seventh Framework Programme program). The work in the P.N.B. laboratory was funded by grants from the National Science Foundation Arabidopsis 2010 program and by a National Institutes of Health P50 grant. A.A. is the incumbent of the Peter J. Cohn Professorial Chair.
Author contributions: A.M., S.M.B., P.N.B., and A.A. designed research; A.M., H.B., M.Y., and S.M.B. performed research; A.M., I.R., L.B., S.M., T.W.T., S.M.B., and A.A. analyzed data; and A.M., S.M.B., P.N.B., and A.A. wrote the paper.
The authors declare no conflict of interest.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1302019110/-/DCSupplemental.
