• In this study a combination of cytoenzymological and immunocytochemical techniques was used in order to demonstrate the presence of cyclic nucleotide metabolism in chloroplasts of higher plants.
• Catalytic cytochemistry was used to localize adenylyl cyclase activity by means of electron microscope investigation on Nicotiana tabacum cv. Petit Havana leaf fragments. Various immunocytochemical techniques were explored to visualize the presence of the second messenger adenosine 3′:5′-cyclic monophosphate.
• Making use of adenylyl imidodiphosphate as a substrate, the enzyme activity was predominantly located at the intermembrane space of the chloroplast envelope. In order to provide further topographical information, intact, isolated chloroplasts were submitted to the same cytoenzymological procedure and revealed stromal adenylyl cyclase activity. Using high-pressure freezing as a physical fixative to obtain an instantaneous metabolic arrest the cellular vitrified water phase was sublimed under ultra-high vacuum by means of molecular distillation drying, avoiding recrystallization and hence redistribution of small highly diffusible molecules. This sequential combination preserved 3′:5′-cAMP epitope retention in chloroplasts as was demonstrated by immunogold labelling.
• These results further substantiate in a unique way the growing evidence of the presence of an organelle-specific cAMP metabolism in higher plants. Furthermore the data presented support the status of chloroplasts as an excellent model to further investigate cAMP metabolism and to correlate it with a variety of physiological functions.
Ever since the first reports on cyclic nucleotide metabolism, its occurrence and function in higher plants have been controversial (Assmann, 1995; Bolwell, 1995; Trewavas, 1997; Newton et al., 1999a; Newton & Smith, 2004). Using mass spectrometry-based identification the natural occurrence of cyclic nucleotides has been repeatedly demonstrated in various plant model systems (Newton et al., 1989; Witters et al., 1996). Biochemical proof for anabolic and catabolic cyclic nucleotide-related enzyme activity in higher plants have been reported in different model species (Newton et al., 1999a; Witters et al., 2004). The recently discovered cDNA, encoding for a soluble enzyme exposing adenylyl cyclase (AC) activity, provided the first evidence at the genetic level for enzymic adenosine cyclic monophosphate production (Moutinho et al., 2001). Similarly, using a molecular biology strategy, a novel protein with guanylyl cyclase activity has been cloned and proven to act as a monomer as well as a homo-oligomer, thereby representing a novel class of guanylyl cyclases (Ludidi & Gehring, 2003). Recent studies describing cyclic nucleotide gated channels (Köhler et al., 1999; Leng et al., 1999; Maathuis & Sanders, 2001; Hua et al., 2003) contributed greatly to the physiological relevance of cyclic nucleotides. With the evidence gathered over the past decade, although it still is fragmentary, renewed interest is emerging. An important consideration one has to make when studying the metabolism of signal molecules is their inherent fast and transient activity profile, making sampling time and spatial resolution crucial. Most data on AC activity in literature have been acquired using biochemical assays, often using tissue homogenates from various parts of the plant that are not necessary homogeneous with respect to their physiological status. As a result the basal metabolic activity can mask or obscure to a great extent the change in activity of the elicited cells. In catalytic histochemistry, both at the tissue level and the cellular level, enzymes are ideally detected in their natural environment at their original site. Cytoenzymological data can therefore provide the means to integrate functional metabolic aspects and morphology. The additional strength of cytoenzymology is its resolution, as it not only can reveal the site of action within a specific cell, but can also show what cells among others in a tissue are affected. The qualitative and to a certain extend also the quantitative information from catalytic cytoenzymological studies can thus be used as a complement to biochemical investigations. Early publications report on the appearance of AC activity at the side of root nodules in Trifolium spp. (Tu, 1974) using ATP as a substrate and lead as capturing agent. Using adenylylimidodiphosphate as a more specific substrate and lead as capturing agent, AC activity in Zea mays root tips was localized at the plasma membrane, endoplasmic reticulum and nuclear membranes (Al-Azzawi & Hall, 1976), at internal membranes of cytoplasmic vacuoles in Pisum sativum (Hilton & Nesius, 1978), at the external side of the host plasma membrane and membranes surrounding the endophyte in root nodules of Alnus glutinosa (Gardner et al., 1979), and at the external side of the plasma membrane of Pisum sativum (Nougarède et al., 1984). Using strontium as a capturing agent and adenylylimidodiphosphate as substrate, AC activity was associated to membrane structures and the cell wall on both pollen and stigma side during pollination (Rougier et al., 1988). They postulated that AC activity was a determinative factor in the compatibility of pollen tube formation in Populus spp. Curvetto & Delmastro (1990) showed that AC activity was localized in Vicia faba guard cells and that it was selectively stimulated by various effectors of the cyclic nucleotide metabolism.
As is the case with histochemistry, immunolocalization can reveal the topographical distribution of proteins at the ultrastructural level but does not give any activity information. Whereas immunolocalization of large biomolecules such as proteins is relatively straightforward, epitope retention and dedicated sample preparation for avoiding redistribution of small diffusible polar compounds such as cyclic nucleotides is far more challenging (Linner et al., 1986). Since those analytes are not encrusted in a matrix such as membranes, microfibres or cell wall, chemical fixation always imposes dilution and redistribution of the solvated molecules as the chemofixative enters the cytosol. Furthermore, chemical fixation is to be considered as a relatively slow process depending on diffusion rates and as a consequence of the covalent binding of the target molecules to the cell matrix, loss of antigenicity by blocking or chemically altering the antigenic recognition site can lead to loss of response. By contrast, cryofixation ensures almost instantaneous biochemical arrest of metabolic processes, does not alter the antigenic recognition sites and dilution and redistribution is kept to a very minimum. Among freeze fixation techniques such as propane-jet freezing, cold metal block slamming, plunge freezing and high-pressure freezing, the last is known to give the best results on ultrastructural preservation of ‘thick’ specimens (Gilkey & Staehelin, 1986). The next crucial step in the preparation procedure is substitution of the water phase with a resin. Resin impregnation via cryosubstitution is excluded a priori, since it involves subzero liquid-phase transfer redistribution, possibly resulting in dilution of unfixed antigens. In order to deal with these particular problems, molecular distillation drying is the method of choice (Linner et al., 1986). This technique enables the removal of amorphous water without recrystallization. Reports on immunolocalization of small diffusible molecules in literature are sparse and most of the reports on topochemical studies of cyclic nucleotides in literature are restricted to 3′:5′-guanosine cyclic monophosphate (3′:5′-cGMP) (Chan-Palay & Palay, 1979; De Vente et al., 1987) and 3′:5′-adenosine cyclic monophosphate (3′:5′-cAMP) (Wedner et al., 1972; Steiner et al., 1976) in paraformaldehyde-fixed animal tissue. Until now, to our knowledge, no reports describe immunolocalization of those compounds in plant tissue. In this study we report on the presence of active AC in chloroplasts as well as the immunolocalization of its product, adenosine 3′:5′-cyclic monophosphate, for the first time in higher plants.
Materials and Methods
For cytoenzymological reactions samples of Nicotiana tabacum cv. Petit Havana SR1 (2 × 2 mm2) were taken from a young fully expanded leaf, excised from a 6-wk-old plant, cultivated in the glasshouse with a 16 h : 8 h light/dark regime. The sample was taken immediately after an expanded (48 h) dark period. For immunolocalization leaf discs (1 mm2) were pinched out of intact leaves obtained from plants treated as described above. Chloroplasts were isolated using a discontinuous Percoll gradient (Nakatani & Barber, 1977).
Salts, buffers, LR White and Spurr's resin, Percoll, glycerol, ATP, AMPP(NH)P, 3′:5′-cAMP, 5′-AMP, alloxan, 3-isobutyl 1-methyl xanthine (IBMX), dithiothreitol, dicyclohexylcarbodiimide, sorbitol, Tween-20, osmium tetroxide, paraformaldehyde and glutaraldehyde were obtained from Sigma (Sigma Chemical Company, Sigma-Aldrich Corp, Bornem, Belgium). Fish gelatin and commercial antibodies were purchased from Aurion (Wageningen, the Netherlands). Phosphate buffered saline (PBS) contained 137 mm NaCl, 2.68 mm KCl, 1.47 mm KH2PO4, 7.67 mm Na2HPO4, pH 7.2.
Cytoenzymological assay conditions
All buffers were freshly prepared at the day of use. For cytoenzymological reactions, leaf fragments were submerged in prefixation buffer (pH 7.2) for 5 min at reduced pressure, containing the basic buffer – 330 mm sorbitol, 80 mm citrate, 2 mm MgCl26H2O, 1 mm dithiothreitol, 4 mm Pb(NO3)2– to which 1.5% formaldehyde and 0.01% Tween-20 was added. Based on the findings of Li et al. (1994), who described a phosphodiesterase inhibition effect using IBMX in palisade parenchyma preparations of Vicia faba, 5 mm IBMX was added to the medium in order to minimize enzymic cAMP hydrolysis. Leaf fragments were rinsed in the same buffer without formaldehyde and transferred to their respective reaction buffer. The reaction buffer contained 4 mm of substrate to which a facultative inhibitor was added. Alloxan was added at a concentration of 10 mm, and DCCD at a concentration of 100 µm. The reaction took place for 30 min at room temperature under reduced pressure in a reaction volume estimated to be at least 100 times the total specimen volume. The specimens were then rinsed in basic buffer and fixed overnight at 4°C in 330 mm sorbitol and 80 mm citrate buffer supplemented with glutaraldehyde (3% final concentration). The tissue was then treated with 2% OsO4 for 2 h at room temperature. The samples were dehydrated in ethanol and embedded in Spurr's resin. The specimens were contrasted with uranyl acetate (4%, w : v) in 50% ethanol.
Immunopurified polyclonal antibodies were raised in chicken against an adenosine 3′:5′-cyclic monophosphate-diphtheria toxoid antigen construct (Roef et al., 1996). Rabbit antichicken 10 nm colloidal gold was used as the secondary antibody in a 1 : 100 dilution of blocking buffer. Washing buffer contained PBS plus 0.025% (w : v) Tween-20 and 20 mm glycine. Blocking buffer contained PBS plus globulin-free bovine serum albumin (BSA) (0.5%, w : v) fish gelatin (0.1%, w : v) and normal rabbit serum for 3′:5′-cAMP localizations. For the sequential procedure see Table 1.
Table 1. Schematic presentation of the protocols used for immunolocalization (the pathway presented in bold was used for the results presented)
For high-pressure freeze fixation the tissue was submersed in 1-hexadecene, placed in an aluminium platelet (2 mm internal diameter, 200 µm deep) and sandwiched with a second platelet, with or without cavity according to the sample thickness, and cryofixed using a Balzers HPM 010 high-pressure freezing device (200 MPa, 1000 Ks−1 cooling rate) (Bal-Tec Products, Guyancourt, France) and stored in liquid nitrogen. For molecular distillation drying, specimens were transferred under liquid nitrogen to a Life Cell molecular distillation dryer unit (MDD-C; Life Cell, Branchburg, NJ, USA) to remove the amorphous water phase, vapour-fixed using paraformaldehyde and vacuum embedded in LRWhite resin. The curve of the microprocessor controlled heating during drying including additional treatment steps (vapour fixation and LRWhite resin embedding) is depicted in Fig. 1. Ultra thin sections (±60 nm) were cut using a LKB Ultratome III and transferred on formvar coated cupper-nickel grids. Sections used for cytoenzymological studies were examined using a Philips EM208 microscope (Philips, Wageningen, the Netherlands), sections used for immunolocalization were examined using a Jeol 1200 STEM (Jeol, Tokyo, Japan).
Results and Discussion
Cytoenzymological localization of adenylyl cyclase activity in leaf tissue
Under the described experimental conditions good morphological structure was preserved (Fig. 2). Adenylyl cyclase activity in palisade parenchyma of excised leaf tissue was located mainly at the chloroplast envelope (Fig. 3; see Fig. 7). Reaction in the chloroplast was visible in the intermembrane compartment. No evidence for reaction at the thylakoid membranes or stroma could be detected. The reaction was relatively homogenous throughout the contour of the envelope, though discrete reaction centres could be located (see Fig. 7). The absence of reaction inside the chloroplast could be explained by the selective permeability of the inner chloroplast membrane. This explanation is further sustained by evidence using isolated chloroplasts, as described below, where the incubation medium contains detergents to permeate the envelopes.
As can be seen from Fig. 3, a different intensity of reaction was observed in the neighbouring cell, indicating either a different infiltration of the prefixative or reaction medium, or a different ‘nonsignalling’ physiological status of the cell. In order to further specify the cytoenzymological reaction enzymic inhibitors were added to the reaction buffer. The use of orthovanadate or molybdate, frequently used as an ATPase inhibitor in biochemical assays, was ruled out since they form insoluble complexes with lead. However, the reaction could be adequately inhibited both by the specific AC inhibitor alloxan (Cohen & Bitensky, 1969) (Fig. 4) and the more general ATPase inhibitor dicyclohexyl carbodiimide (DCCD) (Pick & Racker, 1979) (Fig. 5). Noteworthy, and in agreement with these observations, is the report made by Curvetto & Delmastro (1990) where the occurrence of AC activity was situated at the outer membrane. Another report made by the same group demonstrated the sensitivity of the chloroplast envelope AC towards potassium treatment in guard cells (Morsucci et al., 1991). An earlier report on AC activity in shoot apices of Bryum argenteum demonstrated a uniform distribution of lead precipitate along the envelope of developing plastids (Bhatla & Chopra, 1984). When ATP instead of AMPP(NH)P was used as a substrate (Fig. 6), additional reaction centers in the cytosol could be detected, indicating ATPase activity. Reaction centres were also detected in the chloroplast envelope however, to a lesser extent and more discrete when compared with the reaction pattern obtained using AMPP(NH)P as a substrate (Figs 7 and 8). Interestingly, the longitudinal extended shape of the precipitate in the chloroplast envelope might indicate either a static cluster of complexes of ATPase reactive centres next to each other or a single complex having a restricted mobility in the envelope. This ATPase reaction could adequately be quenched using DCCD as an inhibitor (Fig. 4). In order to acquire additional evidence, in another set of experiments the histocytochemical reaction was carried out on isolated chloroplasts (Figs 9–11). Conditions for incubation time and temperature were based on those used for biochemical experiments (Witters et al., 2004). Based on these experiments, where addition of Tween-20 had a sixfold increase on adenylyl cyclase activity, the reaction buffer was supplemented with 0.01% Tween-20. Compared with the control (Fig. 10), the impact of Tween20 on chloroplast morphology after incubation was quite drastic (Figs 11 and 12). Despite the unfavourable effect of the detergent, the topochemical results can still be interpreted within the context of this study. Whereas AC reaction was also visible on the border of the chloroplast, explicit precipitation occurred in the plastidial stroma (Fig. 11) and alloxan clearly inhibited the AC activity (Fig. 12). These findings are in accordance with and substantiate the results obtained from leaf tissue experiments (Figs 4–8). Experiments using the isolated chloroplasts might prove a useful model in investigating the influence of different effectors on the AC activity since there are no cell walls or various membranes to act as barriers. This experimental setup can also be used as an adequate alternative for biochemical investigations in empirical research for various compounds influencing the AC activity.
Immunocytochemical localization of 3′:5′-cAMP in chloroplasts
To further demonstrate the presence of the cyclic nucleotide metabolic pathway in chloroplasts, a technique for immunocytochemical detection of 3′:5′-cAMP was developed. In a first set of experiments different freezing and labelling techniques were tested to detect-3′:5′-cAMP (Table 1). At light microscopy level no satisfactory protocol was found to immunolocalize 3′:5′-cAMP. In short, all used protocols involving liquid-phase transfer with or without prefixation resulted in poor or no detection of 3′:5′-cAMP and will not be discussed. Initially, a few infiltration experiments concerning the type of resin were conducted. Infiltration seemed best using the hydrophobic epoxy resin (Spurr's resin), however, when related to labelling efficiency it was noted that on a quantitative basis less label was detected in comparison with the acrylic resins such as Lowicryl and LRWhite. In the experiments presented here, LRWhite was used. Using high-pressure freezing, no complete intact cells were observed, a problem inherent to the incompatibility of the highly aerated and vacuolated structure of leaf tissue, avoid from any cryoprotectant, with the freezing technique. Although the outer membrane did not seem to be conserved and therefore with the inherent loss of its epitope retention, the overall ultrastructural morphology of the chloroplast was sufficiently preserved for the purpose of this study (Figs 12–14). Stacked grana and frets were clearly distinguishable in the chloroplast ultrastructure (Fig. 12) in combination with a good epitope retention (Fig. 13). Although the majority of the organelles still suffered from freezing artefacts at a degree that was dependent on their location inside the specimen, the immunolabel could readily be assigned to discrete parts of the organelles. Most of the label locating-3′:5′-cAMP was situated in the stroma (Fig. 13). The immunocytochemical control, where the primary antibody was saturated with the antigen during the incubation, showed almost no label (Fig. 14). Without any form of chemical fixation or cross-linking during the sample preparation until the point where the sample was completely freeze dried, antigenicity for 3′:5′-cAMP was retained. Since no analyte translocation was imposed by liquid transfer, the immunolocalization of 3′:5′-cAMP in the chloroplast is to be considered authentic.
cAMP metabolism in chloroplasts
As mentioned in the Introduction, so far, molecular biology-based strategies that were able to unravel cyclic nucleotide metabolism and signalling in unicellular plants could not be used in higher plants and major contributions have yet to be made. However, as is the case with the pollen tube (Assmann, 1995; Moutinho et al., 2001), the chloroplast as a biological entity can be regarded equally well as a functional model to study cAMP metabolism in higher plants using biochemical and cytochemical techniques. For example, chloroplast physiology is affected by plant growth regulators such as cytokinins (Benkova et al., 1999; Heyl & Schmülling, 2003), phytochrome signalling (Monte et al., 2004) and cryptochrome signalling (Usami et al., 2004). Although these effects were not studied with respect to cAMP metabolism, it is striking that cAMP fluctuations in lower plants have been associated with photosynthetic activity (Segovia et al., 2001; Gordillo et al., 2004), chlorophyll synthesis (Berchtold & Bachofen, 1977), blue light signalling (Iseki et al., 2002) and red light signalling (Okamoto et al., 2004). By contrast, biochemical evidence has been acquired for the presence of phosphodiesterase (Brown et al., 1980) and adenylyl cyclase (Newton et al., 1999b; Witters et al., 2004) in chloroplasts of higher plants. Together with the cytoenzymological data and immunocytochemical data presented here, chloroplasts can be regarded as ultimate candidates for further investigation in order to assign to them a physiological functional cAMP metabolic activity.
The authors thank Prof. C. Hawes and Barry Martin (School of Biological and Molecular Sciences, Oxford Brookes University, Oxford, UK) for the aid with the molecular distillation drying experiments and Dr F. Gaill and J-P. Lechaire (Université Pierre et Marie Curie, Paris France) for assistance with the high-pressure freezing experiments.