There are a variety of methods for characterising gene expression at the level of individual cells and for demonstrating that the cells also contain the encoded proteins. However, measuring the activity of enzymes at the resolution of single cells in complex tissues, such as leaves, is problematic. We have addressed this by using single-cell sampling to extract 10–100 pl droplets of sap from individual plant cells and then measuring enzyme activities in these droplets with nanolitre-scale fluorescence-based assays. We have optimised these assays and used them to measure and characterise the activities of acid phosphatase, cysteine protease and nitrate reductase in sap samples from epidermal and mesophyll cells of barley (Hordeum vulgare L.) and Arabidopsis thaliana leaves exposed to different developmental and environmental conditions. During leaf senescence in barley, we found that the dynamics with which acid phosphatase and protease activities changed were different in each cell type and did not mirror the changes occurring at the whole-leaf level. Increases in nitrate reductase activities after exposure of barley plants to nitrate were large in mesophyll cells but small in epidermal cells. The technique was applied successfully to Arabidopsis and, as in barley, revealed cell-specific differences in the activities of both acid phosphatase and nitrate reductase. The assays add to the spectrum of techniques available for characterising cells within complex plant tissues, thus extending the opportunity to relate gene expression to biochemical activities at the single-cell level.
Functional genomics will eventually provide an understanding of how the temporal and spatial expression of all genes in plants controls the levels of proteins and the rate of key developmental, metabolic and physiological processes. In due course, this information will be needed at the level of individual cells so that the contribution each cell type makes to the development, function and performance of tissues, organs and plants can be elucidated. This will include determining how changes in gene expression are related to alterations in the rates of biochemical processes, for which knowledge of enzyme activities will be required. Therefore, the development of robust methods to measure and characterise enzyme activities in identifiable, individual cells in intact tissues is needed.
A range of techniques is available for demonstrating the distribution of enzymes and proteins among different cell types. These include in situ enzyme histochemistry (de Block, 1995; Sergeeva and Vreugdenhil, 2002), the use of reporters for gene expression (Guivarch et al., 1996; Haseloff and Amos, 1995), and immunogold labelling and immunocytochemistry (Harris, 1994). While these techniques yield information about the presence of particular proteins in specific cells, they provide little quantitative data about the activities of enzymes. For instance, fixation of tissues prior to histochemistry may result in the loss of up to 70% of the activity of some enzymes (Sergeeva and Vreugdenhil, 2002), preventing full quantification and leading to artefacts if activity in one cell is more susceptible than that in another. Imaging of changes in the concentration of fluorescent substrates or products in cells can provide information about differences in enzyme activities between different cell types, as well as some kinetic information (Meyer et al., 2001 and references therein). However, it is limited to cells that can be easily imaged, making it difficult to apply to cells that are deep within tissues. Physical dissection of cells coupled with sensitive enzyme assays can be used to measure enzymes and metabolites quantitatively in single cells, but stabilisation of the tissues by freeze substitution or freeze drying is needed and some labile enzymes can be lost (Outlaw and Zhang, 2001).
Here, we describe the optimisation of nanolitre-scale assays for acid phosphatase, cysteine protease and nitrate reductase in sap samples from individual epidermal and mesophyll cells of barley and Arabidopsis leaves. We used these assays to characterise the activities and to measure how they altered during senescence or in response to a change in environmental conditions. The results show that the activities change at different rates in different cell types and in ways that are not predictable from measurements made at the whole-leaf level.
Optimisation of single-cell enzyme assays
In view of the mixed success of previous attempts to measure enzyme activities in samples from single cells (Fricke et al., 1994a; Koroleva et al., 1997), initial experiments concentrated on optimising the conditions for assaying the three chosen enzymes (acid phosphatase, protease and nitrate reductase) in single-cell extracts from barley leaves. Enzyme activities were measured fluorometrically in 4 or 5 nl droplets on a microscope slide by following the accumulation of fluorescent hydrolysis products from MUP (acid phosphatase) and Z-Phe-Arg-AMC (protease), or the decrease in fluorescence accompanying the nitrate-dependent oxidation of NADH (nitrate reductase).
For acid phosphatase, care had to be taken to limit the assay incubation period because fluorescence quenching occurred if MU accumulated to concentrations greater than about 100 µm (see John, 1998). Reliable measurement of acid phosphatase activity required the addition of BSA to the assay medium. When this was omitted, substrate kinetics were inconsistent within and between experiments (Figure 1a). When acid phosphatase was assayed under optimised conditions, it had a pH optimum of 6.0 (Figure 1b), and the release of MU was linear with time (Figure 1c). The activity in sap samples from both mesophyll and epidermal cells was completely inhibited by 100 µm ammonium molybdate, a potent inhibitor of acid phosphatases (results not shown; Gallagher and Leonard, 1982; Leigh and Walker, 1980). To characterise the activity further, the substrate dependence was measured and found to conform to Michaelis–Menten kinetics (Figure 2a). The Km for MUP in mesophyll cells was significantly different (P ≤ 0.05) from that in the lower epidermis but not the upper epidermis (Figure 2a).
Addition of DTT was necessary to detect protease activity (Figure 3a), which had a pH optimum of 5.0 (Figure 3b), and showed a linear time course after an initial lag of 5 min (Figure 3c). The activities from all cells were decreased by 74–88% in the presence of 10 µmN-(trans-epoxysuccinyl)-l-leucine-4-guanidinobutylamide (E-64) but were unaffected by 0.4 mm phenylmethanesulfonyl fluoride (PMSF). Z-Phe-Arg-AMC is mainly hydrolysed by cysteine proteases (Harris et al., 2000), and the sensitivity to E-64, but not PMSF, indicates that the activity in the cell extracts is that of a cysteine protease (Beers et al., 2000; Ménard and Storer, 1998). The requirement for DTT is also consistent with this conclusion (Minami and Fukuda, 1995). The activity showed Michaelis–Menten kinetics when Z-Phe-Arg-AMC concentration was varied but, in contrast to acid phosphatase, the Km values were not significantly different between various cell types (Figure 2b).
Acid phosphatase and cysteine protease have been shown to be localised in the vacuole (Boller and Kende, 1979; Heck et al., 1981). To demonstrate that the method is not limited to vacuolar enzymes, we also measured nitrate reductase activity in samples from barley cells. Initial experiments used commercially prepared nitrate reductase from Aspergillus nidulans to test the assay. This showed that the rate of fluorescence change was proportional to the amount of enzyme added (Figure 4a). When used with a cell extract, the reaction was linear for approximately 30 min (Figure 4b), and displayed Michaelis–Menten kinetics with respect to nitrate concentration, yielding a Km for nitrate of 0.4 mm (Figure 4c), in line with reported values for plant nitrate reductases (Hageman and Hucklesby, 1971).
Developmentally and environmentally induced changes in single-cell enzyme activities
To demonstrate the utility of the optimised assays, changes in the three enzyme activities were measured in barley under conditions in which they were expected to increase. Acid phosphatase and cysteine protease activities were measured in sap samples taken from cells in detached leaves that were induced to senesce by incubating them in the dark (Miller and Huffaker, 1985). Changes in nitrate reductase were measured in response to provision of nitrate to plants previously grown in the absence of nitrate (Beevers and Hageman, 1969).
Activity of acid phosphatase in mesophyll cells increased substantially over the first 2 days after leaf detachment but then declined for the remainder of the experiment (Figure 5a). The activity in the lower epidermal cells declined initially and then rose to a transient peak on day 3. In contrast, there was no substantial change in acid phosphatase activity in upper epidermal cells. None of these patterns were mirrored in the changes measured in whole leaf extracts which, after a small increase over the first 2 days, declined for the remainder of the experimental period (Figure 5c). Treatment of the leaves with 50 µm cycloheximide inhibited the increase in acid phosphatase activity in mesophyll cells (not shown).
The changes in cysteine protease activity were quite different from those for acid phosphatase. Activity increased in all three cell types during the first day after detachment, then declined and increased again on day 3 (Figure 5b). The peak on day 1 was largest in mesophyll cells, while activity on day 3 was similar in all three cell types. Like acid phosphatase, the changes in cysteine protease activity in the cells did not match those in the whole-leaf extracts, which rose to a transient peak on days 2 and 3 (Figure 5c). As with acid phosphatase, cycloheximide prevented the increase in cysteine protease activity in mesophyll cells (not shown).
An advantage of single-cell sampling is that it can be used to map changes in activities with both position and time. As a test of this, increases in cysteine protease activity over the first day of senescence were measured at different points along the leaf blade. This showed that there were gradients of activity along the leaf, but the characteristics of these were different in each cell type. In mesophyll cells and upper epidermal cells, activity increased the most near the tip and the least near the base (Table 1). In contrast, increases in activity in lower epidermal cells were similar at all points sampled.
Table 1. Increase in cysteine protease activity in cells located at different distances from the tip of barley leaves
Protease (amol AMC min−1 pl−1)
Leaves were 14 cm long. Activities (mean of two independent experiments) are the difference between those measured immediately after leaf detachment and those measured 1 day later.
Provision of nitrate to barley plants, previously grown without nitrate, increased nitrate reductase activity in a cell-specific manner. Mesophyll cells showed the largest increase (40-fold), while there was no change in lower epidermal cells. Upper epidermal cells (eightfold increase) were intermediate in their behaviour (Figure 6).
To demonstrate that these methods are transferable to other species, measurements of acid phosphatase and nitrate reductase activities were made on sap samples from leaf epidermal and mesophyll cells of Arabidopsis. Acid phosphatase activity was similar in upper epidermal and mesophyll cells but was significantly less in lower epidermal cells (Figure 7a). Activity in all cell types was inhibited by the addition of 100 µm molybdate. When plants were grown in the presence of nitrate, significant nitrate reductase activity was measured only in mesophyll cells (Figure 7b).
The techniques described here allow detailed quantitative biochemical investigations of enzyme activities in single cells at defined locations in intact plant organs. Although fluorescent flow cytometry and similar techniques can be used to measure enzyme activities within individual cells (Deutsch et al., 2000; Haugland and Johnson, 1993), they mainly have been used with cell suspensions, such as spermatozoa (Schaller and Glander, 2000) or cultured cells (e.g. mouse thymocytes; Telford et al., 2001), but not intact tissues where spatial location of cells is important. In situ enzyme histochemistry (de Block, 1995; Sergeeva and Vreugdenhil, 2002) and immunocytochemistry (Harris, 1994) retain spatial information but do not provide quantitative measurements of enzyme activity (see Introduction). Imaging can provide some limited information about cell-specific enzyme activities in living cells, but, at present, is restricted in the range of parameters that can be measured, and the cells that can be accessed (Meyer et al., 2001). Although single-cell dissection allows both spatial and quantitative information to be obtained simultaneously, the technique requires cryofixation and stabilisation of the sample before enzymes are assayed, which may result in loss of some labile enzymes (Outlaw and Zhang, 2001). The technique described in this paper has similar advantages to those of single-cell dissection but requires no fixation step, so samples can be taken almost instantaneously from virtually any cell in any tissue and assayed immediately. It should have wide applicability because in addition to barley and Arabidopsis (this paper; Fricke et al., 1994a,b; Koroleva et al., 1997, 2000a,b; Laval et al., 2002), single-cell sampling has been successfully applied to a range of other species and tissues including leaves from tomato, wheat, potato, cucumber and Thlaspi caerulescens (Brandt et al., 1999; Karrer et al., 1995; Kehr et al., 1999; Küpper et al., 1999; Malone et al., 1991), carrot tap roots (Korolev et al., 2000), and maize primary roots (Pritchard et al., 1996).
The ability to measure and characterise enzyme activities in single cells using single-cell sampling and nanolitre-scale assays extends previous studies, which have been limited to simple demonstrations that a particular activity is present or absent (Fricke et al., 1994a; Koroleva et al., 1997). The results presented here show that it is possible to measure substrate kinetics (Figures 2 and 4), inhibitor sensitivities and changes in activity induced by imposed treatments (Figures 5–7). The main requirement is to optimise the composition of the assay media (Figures 1a and 3a). This need to optimise assay conditions for single-cell samples may explain why Koroleva et al. (1997) failed to measure acid invertase activity in samples from single cells despite its presence in whole-leaf extracts. They omitted BSA during the 3 h incubation of cell sap with sucrose but included it in the subsequent assay of the released hexoses. It is possible that this omission allowed the acid invertase to be substantially degraded by proteases. Although they could measure activity in cell homogenates, the dilution of enzymes during the homogenisation in buffer may have slowed the rate of invertase degradation compared with undiluted sap taken directly from cells.
The advantages of being able to measure enzyme activities at the resolution of individual cells is illustrated by the differences in behaviour of acid phosphatase and cysteine protease in different cell types and in whole leaves of barley (Figure 5). The lack of correspondence between changes at the leaf and cell levels confirms that measurements on whole leaves reveal little about the activities in individual cells. Further, the changes in each cell type are more dynamic than the whole leaf measurements would suggest. Thus, cysteine protease activity at the whole-leaf level increases after a lag of 2 days and then declines slowly, whereas activities at the cell level show two distinct peaks during the same period. The lack of correspondence between behaviour measured at the different levels arises because the whole-leaf measurements include all cell types, some of which were not measured here, and because cells at different parts of the leaf begin to senescence at different times, leading to a gradient of activities along the leaf (Table 1). Thus, a more detailed sampling to map changes in more cell types over a wider area of the leaf is needed to explain fully events at the whole-leaf level in terms of activities in individual cells. This is feasible with the technique described.
The differences in acid phosphatase and cysteine protease activities between epidermal and mesophyll cells of senescing barley leaves (Figure 5) confirm that these cell types senesce at different rates (Matile, 1997). There are sound physiological reasons for these differences. The mesophyll is the major store of nitrogen and phosphorus in the leaf, and re-mobilisation of these limiting resources to growing organs, such as new leaves or developing seeds, is an important aspect of leaf senescence. In contrast, maintenance of epidermal cells until late in senescence is essential to retain leaf integrity, to prevent pathogen invasion and to ensure continuation of stomatal function and transpiration so that water is available for the export of catabolites. Perhaps more surprising are the differences in behaviour between the upper and lower epidermal cells, exemplified by the contrasting changes in acid phosphatase activity (Figure 5a). These differences are consistent with other observations, showing that these tissues differ in other respects, e.g. nitrate and chloride accumulation (Fricke et al., 1994b; Cuin and Leigh, unpublished). There is a need to study in more detail the changes occurring in epidermal cells as senescence progresses and to relate this to the maintenance of structural integrity and key functions of the leaf.
Finally, the increase in nitrate reductase in mesophyll cells in response to the provision of nitrate (Figures 6 and 7) confirms that the method can be used to measure enzyme induction in response to short-term environmental changes. Coupled with techniques to measure gene expression in single cells (Brandt et al., 1999, 2002; Gallagher et al., 2001; Laval et al., 2002), it will now be possible to relate changes in enzyme activities to alterations in gene expression. When combined with measurements of ion and solute concentrations and water relations parameters in the same cells (Tomos and Leigh, 1999; Tomos et al., 1994), these approaches offer the opportunity to integrate molecular, biochemical and biophysical parameters at the level of individual cell types.
Barley (Hordeum vulgare L. cv. Halcyon) seeds were germinated on moist filter paper in the dark at 25°C. Plants used for measurements of acid phosphatase and protease activities were transferred to soil when the primary root was approximately 1 cm in length. To induce senescence, fully expanded primary leaves were detached from 12-day-old seedlings, placed with their cut end in 0.2 mm CaSO4 and left in the dark at 25°C (Miller and Huffaker, 1985). Senescence was complete 7 days later. Leaf samples used for measurement of whole-leaf enzyme activities were taken from the region where cells were sampled (see below), frozen in liquid N2 and stored at −80°C. Plants used for measurements of nitrate reductase activities were germinated as above, and 5-day-old seedlings were transferred to a modified Hoagland's solution (Walker et al., 1996), containing 1 mm NH4Cl and 4.5 mm CaCl2 instead of KNO3 and Ca(NO3)2, respectively. They were grown for 21 days at 24°C, 60% relative humidity, 800 µmol photons m−2 sec−1, and a 16 h light period. Nitrate reductase activity was induced by transferring the plants to the same nutrient solution but with Ca(NO3)2 and KNO3 replacing the CaCl2 and NH4Cl. Cells were sampled 6 h later.
Arabidopsis thaliana L. ecotype Columbia was germinated and grown for 2 weeks on 4.4 g l−1 Murashige and Skoog medium (Duchefa, Haarlem, Holland), 1% sucrose and 1% Bactoagar, before being transferred to a nutrient solution containing 1 mm KNO3, 1 mm KH2PO4, 1 mm CaSO4, 0.5 mm MgSO4, 50 µm NaFeEDTA and trace elements. After 1 week, half of the plants were transferred to a nitrate-free solution in which KCl replaced KNO3. They were grown for a further 7 days before being used in experiments.
Sap samples were removed from cells using a silicone oil-filled, salinised microcapillary (Tomos et al., 1994) made from unfilamented borosilicate glass tubing with an internal diameter of 0.58 mm (Harvard Apparatus, Edenbridge, Kent, UK). The microcapillaries were fabricated using a vertical microelectrode puller (Narishige, Tokyo, Japan) and were broken against a fine metal wire to a tip diameter of approximately 5 µm. Cells chosen for sampling in barley leaves were located 1.5 cm from the tip of a detached primary leaf (acid phosphatase and protease experiments) or midway along the fully expanded attached third leaf (nitrate reductase experiments). Those in Arabidopsis were from the first true leaves. All epidermal samples from barley were taken from ‘trough’ cells (Fricke et al., 1995) and in Arabidopsis from the centre of the leaf half way between the mid-vein and the edge. Samples from mesophyll cells were obtained by either inserting the microcapillary through open stomatal pores (attached leaves of barley; Fricke et al., 1994a) or puncturing an upper epidermal cell, expelling any cell contents that had entered the microcapillary, and then proceeding to the mesophyll cell below (detached barley leaves and Arabidopsis). As a check on the puncturing method, some mesophyll cells in intact non-senescing attached leaves in barley were accessed through open stomatal pores. Acid phosphatase activities in these samples were not significantly different from those obtained by cell puncturing. The sampling time for all cells was <1 sec. Samples (30–100 pl in barley and 10–20 pl in Arabidopsis) were expelled into paraffin oil on a microscope slide and assayed immediately. In some experiments, sap from several cells of the same type was pooled. Subsampling of sap samples and dispensing of solutions in enzyme assays were carried out with constriction pipettes (10 pl−5 nl) prepared from borosilicate glass tubing (see above) and backfilled with paraffin oil (Tomos et al., 1994). Pipette volumes were approximate. To maintain consistency, the same set of pipettes was used throughout an individual experiment.
Measurement of enzyme activities in extracts from single cells
Acid phosphatase was measured in an assay medium (final volume 5 nl) containing 50 mm acetic acid-tris(hydroxymethyl)methylamine (Tris), pH 6.0, 0.1% (w/v) bovine serum albumin (BSA), 3.3 mm 4-methylumbelliferyl-7-phosphate (MUP; Sigma Chemical Company, Poole, Dorset, UK) dissolved in dimethyl sulfoxide (DMSO; final concentration 3% v/v), and, where added, 100 µm ammonium molybdate. When needed, pH was varied by altering the proportions of acetic acid and Tris. Assay droplets were prepared and kept under water-saturated paraffin oil on a microscope slide. The reaction was started by the addition of 10 pl of cell sap and the droplet mixed by agitation with the tip of the constriction pipette. Samples were incubated for 10 min at 25°C. The reaction was stopped by the addition of 500 pl of 1 m NaHCO3/2 m NaOH to raise the pH to 10.5, the optimal pH for 4-methylumbelliferone (MU; Sigma) fluorescence. The amount of MU was determined fluorometrically using a Leica IRB fluorescence microscope fitted with filter system A (excitation filter 340–380 nm, dichroic mirror 400 nm, suppression filter long pass 425 nm) and an MPV-C2 photomultiplier (Leica Microsystems, Milton Keynes, UK). Output from the photomultiplier was collected on a PC running the Leica mpv-meas software. A standard curve (0–250 fmol MU) was prepared and measured simultaneously.
Protease was measured in a medium (final volume 5 nl) containing 50 mm 2-N-(morpholino)ethanesulfonic acid (Mes)-NaOH, pH 5.0, 0.5 mm dithiothreitol (DTT), 1 mm ethylenediaminetetraacetic acid, 150 mm NaCl, 200 µm benzyloxycarbonyl-Phe-Arg-7-amino-4-methylcoumarin (Z-Phe-Arg-AMC; Bachem, St Helens, Merseyside, UK) dissolved in DMSO (final concentration 5% v/v), and, where added, inhibitors at the concentration indicated. For the experiment in Figure 3(b), pH was varied using 50 mm Mes-NaOH (pH 4–6) and 50 mm 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (Hepes)-NaOH (pH 6–8). The change in buffer at pH 6 did not significantly affect the activity. Reactions were started and mixed as above. The samples were incubated for 15 min at 25°C, and 7-amino-4-methylcoumarin (AMC) was measured by microfluorometry using the same settings as above. The amount of AMC released was determined from standards (0–120 fmol AMC) measured simultaneously. AMC fluoresces maximally at pH 3.5–8.5 permitting continuous measurement of its release.
The nitrate reductase assay medium contained 50 mm KH2PO4-KOH, pH 7.25, 0.1% (w/v) BSA, 10 mm KNO3, 10 mm NADH and 4 µm FAD in a total volume of 4 nl. Reaction was started by the addition of 12 pl of cell sap or nitrate reductase from A. nidulans (Roche, Lewes, UK). The decrease in NADH fluorescence was measured over 15–30 min at room temperature using the same equipment and settings as above. The amount of NADH consumed was determined from a standard curve (0–40 pmol NADH) measured simultaneously.
In the absence of an assay for total protein content of the sap samples, all activities were related to the volume of sap used in the assays. To check that changes in activities of acid phosphatase and cysteine protease in senescing detached barley leaves were not the result of changes in the water content of cells, the K+ concentrations in representative cell sap samples were measured at different times after leaf detachment (see Tomos et al., 1994 for method). Any changes in water content of the cells should alter this parameter. Potassium concentrations did vary as senescence proceeded (e.g. between 200 and 260 mm in mesophyll cells), but there was no clear pattern in the changes, and they were small in relation to the changes in enzyme activities. Thus, relating activities to sap volume did not create artefacts.
Extraction and measurement of enzymes from leaf tissue
To extract acid phosphatase, approximately 0.1 g of frozen barley leaf was homogenised in 1 ml of 50 mm acetic acid-Tris (pH 5.0) using a cooled mortar and pestle. The extract was centrifuged at 10 000 g for 10 min at 4°C in a microfuge and the supernatant diluted 40-fold with extraction buffer. Assays were conducted under conditions similar to those used for the single-cell extracts except that the assay volume was 2.5 ml, reaction was started by the addition of 10 µl of leaf extract, the samples were incubated at 25°C for 30 min, and the amount of MU released was measured using a Perkin-Elmer LS50 fluorimeter (excitation at 360 nm, emission at 450 nm, and a slit width of 5 nm) after the reaction had been stopped by adding 100 µl of 1 m NaHCO3/2 m NaOH. A standard curve (0–600 pmol MU) was prepared and measured simultaneously.
For protease activity, frozen leaf tissue was homogenised and centrifuged as above except that the extraction medium was 50 mm Mes-NaOH, pH 5.0. The activity was assayed in a total volume of 200 µl in 96-well microplates using the same assay medium as for single-cell extracts. The reaction was started by the addition of 20 µl of leaf extract, the samples were incubated at 25°C for 20 min, and the amount of AMC released was determined fluorometrically using a Perkin-Elmer LS50 fluorimeter fitted with a microplate reader. The settings included excitation at 380 nm, emission at 440 nm, and a slit width of 5 nm. A standard curve (0–6 nmol AMC) was prepared and measured simultaneously.
The work was supported by a grant from the Brooks Fund (University of Cambridge) and by a Biotechnology and Biological Sciences Research Council postgraduate studentship to SJR. We are grateful to Dr Wieland Fricke (University of Paisley, Scotland) for his help in establishing the single-cell sampling technique.