INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Mitochondrial function is intimately associated with energy production to meet the requirements of cellular metabolism. Because the metabolic needs of plant cells are likely to vary widely depending on tissue-type, developmental stage, and environmental conditions, it can be argued that mitochondrial functions must be subject to some sort of regulation, either of the activity or the synthesis of the components involved in energy production.

Accumulating evidence suggests that the expression of genes encoding mitochondrial components, which are encoded in two separate genomes, is a regulated process. Regarding mitochondrial genes, transcript abundance has been found to increase during microsporogenesis in maize (Monéger et al., 1992) and, more specifically, in microsporocytes and developing tapetal tissue in sunflower (Helianthus annuus) anthers (Smart et al., 1994). Specific patterns of localization were also found in maize anthers for different proteins encoded by mitochondrial genes (Conley and Hanson, 1994). In addition, Topping and Leaver (1990) reported that mitochondrial transcript abundance decreases as wheat leaf cells develop and acquire photosynthetic competence. Li et al. (1996) have also shown the existence of cell-specific expression of different mitochondrial transcripts in maize seedlings with increased levels in vascular bundles and root meristematic cells with high division activity.

In the case of nuclear genes, Huang et al. (1994) observed that transcript levels for the Rieske iron-sulfur protein of complex III increase severalfold in tobacco flowers with respect to leaves. Based on measurements of specific protein levels in isolated mitochondria and whole cell extracts, they concluded that the increase was related with a higher number of mitochondria per cell in flowers than in leaves. Several nuclear genes encoding mitochondrial components show elevated transcript levels in flowers of different plant species (Landschütze et al., 1995; Felitti et al., 1997; Heiser et al., 1997; Zabaleta et al., 1998). For three components of the NADH dehydrogenase (complex I) from Arabidopsis, pollen-specific expression regulated at the transcriptional level has been demonstrated (Zabaleta et al., 1998). The emerging picture on the regulation of mitochondrial biogenesis in plants then shows a general increase of transcript levels in flowers, superimposed with specific regulation of the abundance of some mitochondrial transcripts in different tissues. To show cell-specific expression of mitochondrial transcripts, in situ hybridization or immunolocalization techniques have been used (Conley and Hanson, 1994; Smart et al., 1994; Li et al., 1996). However, except for studies performed by Leaver and coworkers (Smart et al., 1994; Balk and Leaver, 1998), similar experiments have not been conducted with nuclear-encoded genes, specially those for respiratory chain components.

We recently have demonstrated that the sunflower nuclear gene encoding mitochondrial cytochrome c is regulated by both tissue type and environmental factors, such as light, nitrate, and carbon source (Felitti et al., 1997; Felitti and Gonzalez, 1998). In the present study, we have analyzed the cell-specific expression of transcripts encoding cytochrome c in sunflower with emphasis in the process of flower development. Sunflower belongs to the Compositae family, with a terminal inflorescence (head or capitulum) composed of hundreds of flowers of two different types: ray (sterile) flowers in the periphery, and rings of disc (fertile) flowers in the center (actually formed by radiating arcs from the center of the head) (Seiler, 1997). Fertile flowers develop sequentially from the periphery to the center of the head (Hernández and Green, 1993). This constitutes an interesting system to observe changes in expression patterns dependent on flower development since an inflorescence in a given stage contains flowers at different developmental stages. Our results indicate that transcript levels for cytochrome c are elevated as early as flower meristems develop. At different developmental stages, expression is progressively localized to developing floral organ primordia, sex organs, anthers, and tapetal cells. Cell-specific expression has also been observed in roots. Our results demonstrate the existence of cell-specific expression of a nuclear gene encoding a component of the mitochondrial respiratory chain and argue in favor of the existence of coordination in the expression of nuclear and mitochondrial genes encoding such components.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Previous studies have shown that transcript abundance for the cytochrome c gene are much higher in developed flowers than in mature leaves (Felitti et al., 1997). To analyze the expression during flower development, we have isolated total RNA from the apical portion of the sunflower plant at different developmental stages, either before (shoot apical meristem) or after inflorescence meristem formation. We have classified the several stages in flower development from R-1 to R-5, according to Schneiter and Miller (1981). The R-1 stage refers to the time when the inflorescence begins to enlarge and is visible as a bud with the bracts closed forming a star-like structure. In the R-2 stage, the inflorescence has further enlarged and separated 0.5 to 2 cm from the youngest leaves. R-3 inflorescences have opened their bracts upright and are separated more than 2 cm from the youngest leaves. In R-4 inflorescences, bracts are already open and ray flowers (yellow) are visible. Flowering begins with the R-5 stage, and can be divided in several substages according to the percentage of open disc flowers.

As shown in Figure 1, cytochrome c transcript levels show only a slight variation during inflorescence and flower development when considering the entire upper part of the plant. It is interesting that the apical portion of the plant, including meristem and leaf primordia, contains relatively high levels of cytochrome c transcripts related to what has been observed in mature leaves (Felitti et al., 1997).

View larger version (67K): [in this window] [in a new window]   Figure 1.   Northern-blot analysis of sunflower inflorescence RNA at different developmental stages. Total RNA was isolated from the apical portion of plants either before (M) or after inflorescence formation (R1-R3), and analyzed as described in "Materials and Methods" using probes for cytochrome c (A) or the 25S rRNA (B). Twenty micrograms of total RNA was loaded per lane. R1 to R3 indicate different stages of inflorescence development (see text for details).

Since the parts of the plant used for the northern analysis are rather complex in structure and show a mixture of different tissues, we have decided to use RNA in situ hybridization to determine the relative levels of expression in different cell types. In shoot apex sections at the vegetative stage, comprising the meristem and young leaf primordia, a general increase in labeling over the whole tissue section, compared with sections treated with the sense probe, was observed (not shown). However, no cell-specific labeling was evident except in vascular bundles, which also stained with the control probe. In contrast, intense labeling was observed in floral meristems at different stages of development in R-1 stage inflorescences (Fig. 2, A, D, G, and J). In comparison the central portion of the capitulum, bearing the inflorescence meristem, showed no specific labeling. This result points to the existence of enhanced expression as soon as flowers begin to develop. The external part of the inflorescence contains flowers in which petal primordia have begun to differentiate. In those flowers, enhanced expression over the whole flower was also observed.

View larger version (192K): [in this window] [in a new window]   Figure 2.   Cytochrome c transcript localization in longitudinal sections of sunflower inflorescences at different developmental stages. Sections from inflorescences at stages R-1 (A, D, G, and J), R-2 (B, E, H, and K), and R-3 (C, F, I, and L) were hybridized with antisense (A-I) or sense (J-L) cytochrome c probes. A shows the central portion of an R-1 inflorescence including the inflorescence meristem and developing floral meristems. D shows the peripheral part of the same inflorescence as in A; note the development of floral organ primordia. G and J are enlargements of the peripheral part of the inflorescence. B and K show developing flowers of an R-2 inflorescence with stamen and carpel primordia at different developmental stages. E and H are enlargements of flowers from different portions of the same inflorescence. C, F, I, and L show flowers at different developmental stages from an R-3 inflorescence. Anthers in C are at the premeiosis stage; those in F and L are at the leptotene stage; anthers in I are at the pachytene stage. Scale bars = 200 µm.

Figure 2, B, E, H, and K, shows flowers at different developmental stages within an R-2 inflorescence. Less developed flowers contain petals that surround sex organ primordia. Developing stamen and carpel primordia progressively begin to be discernible. In less developed flowers, enhanced expression was observed in sex organ primordia and the tips of growing petals (not shown in the figure). After that, expression in petals disappeared (or was less pronounced), and the signal was observed only in stamen and carpel primordia.

At the R-3 stage, anther development is clearly discernible, and the ovary has acquired its basal position. Anthers from flowers at different positions show several stages of pollen development from archesporial tissue at premeiotic stage to young microspores. In R-3 flowers, hybridization signals were obtained mainly in anthers (Fig. 2, C, F, I, and L). Less developed flowers showed expression through the archesporial tissue (Fig. 2, C and F). Later in development, the label was observed mainly in tapetal cells (Fig. 2I).

In flowers of the R-4 stage, pollen grains have completed their development and an embryo sac surrounded by the nucella was discernible in ovules. At this stage, intense labeling was observed in fully developed pollen grains, but a similar result was obtained using the sense probe (not shown). We presume this is due to the precipitation of dye on the walls of pollen grains. The technique used does not allow the analysis of cytochrome c gene expression in mature pollen. Northern blots of total pollen RNA hybridized with a cytochrome c probe did not reveal any considerable increase in transcript levels (not shown).

To analyze if the expression patterns observed are characteristic of sunflower, we also performed in situ hybridization studies using Arabidopsis flowers at different developmental stages. In this case, a homologous antisense probe synthesized from expressed sequence tag (EST) clone 172G24T7 (accession no. H35987; Newman et al., 1994) was used. Figure 3, A and B, shows that in Arabidopsis increased expression was observed in floral meristems at very early stages of development. Upon progression of organ differentiation, the signal in sepals and petals disappeared and mainly inner organ primordia were labeled. This label became more intense in anthers especially in archesporial tissue when flowers further developed (Fig. 3, C and D). At more advanced stages, a discernible signal was observed in anther tapetal cells (data not shown). Sequence similarity searches indicate that the Arabidopsis genome contains two genes that encode mitochondrial apocytochrome c polypeptides located on chromosome I and IV. To obtain a more complete picture of the expression of cytochrome c encoding transcripts in Arabidopsis flowers, we have also performed in situ hybridization studies using probes for the chromosome IV gene (EST clone APZ69d09; accession no. AV521939; Asamizu et al., 2000). With these probes, however, we did not observe significant label. Since the probes were able to hybridize to homologous DNA fixed on a nylon filter, and experiments conducted in parallel using the chromosome I gene probe showed intense signals, we hypothesize that transcript levels for the chromosome IV gene must be significantly lower in flowers. The expression patterns shown in Figure 3 would then correspond mainly to RNA transcribed from the chromosome I gene but would fairly represent the expression of total cytochrome c encoding transcripts. The results obtained indicate that the expression patterns of cytochrome c encoding transcripts are similar in sunflower and Arabidopsis, suggesting the existence of conserved mechanisms of regulation.

View larger version (96K): [in this window] [in a new window]   Figure 3.   Cytochrome c transcript localization in longitudinal sections of Arabidopsis inflorescences at different developmental stages. Sections were hybridized with antisense (A and C) or sense (B and D) cytochrome c probes. A and B show a group of developing flowers at different stages from a floral meristem (at the center) to a stage where organ primordia are readily discernible. C and D show a flower at a later stage in development; note staining in anthers. Scale bars = 200 µm.

We have also used in situ hybridization to analyze the occurrence of cell-specific expression of cytochrome c transcripts in other parts of the sunflower plant. As mentioned above, defined patterns of expression were not observed in shoot apex preparations, including apical meristem and leaf primordia. A similar result was obtained using sections prepared from mature leaves, stems, hypocotyls, and cotyledons. In young root tips, however, specific expression patterns were observed with increased levels of expression in dividing cells of the meristematic region, especially in the developing endodermis and pericycle, and in protoxylem initials (Fig. 4). It should be mentioned that, together with flowers, roots show an increase in cytochrome c transcript levels when analyzed by northern blots (Felitti et al., 1997). As suggested by our results, this increase may be related to the induction of cytochrome c gene expression in specific cell types.

View larger version (67K): [in this window] [in a new window]   Figure 4.   Cytochrome c transcript localization in longitudinal sections of sunflower developing roots. Sections were hybridized with antisense (A-D) or sense (E-H) cytochrome c probes. B through D and F through H show different portions of the same root. A is an enlargement of the labeled region (left part) shown in B. E is an enlargement of F in the same region. a, b, and c, Endodermis, pericycle, and protoxylem initials, respectively. Scale bars = 200 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Previous studies have suggested the existence of tissue-specific expression of cytochrome c transcript levels in sunflower (Felitti et al., 1997). In the present study, we have extended these studies to show the presence of defined patterns of expression in specific cell types both in roots and flowers of sunflower and Arabidopsis. Expression in flowers has been followed through different developmental stages in sunflower (from the appearance of floral meristems to the process of pollen and embryo sac formation). The general conclusion of these studies is that an increase in expression with respect to other parts of the sunflower capitulum, including the inflorescence meristem, can be observed in flowers as soon as floral meristems form. The expression pattern changes becoming more localized to specific cell types when flowers further develop.

Our results indicate that cytochrome c transcript levels vary widely among different cell types. A general view of the results obtained may suggest that higher transcript levels are present in tissues with high cell division activity, perhaps reflecting the requirement for more active mitochondrial biogenesis in these tissues. Active mitochondrial biogenesis, however, is not always correlated with mitotic activity, at least in roots (Kuroiwa et al., 1992). In our case, no labeling was observed in the central region of the inflorescence meristem in R-1 inflorescences, which is known to have mitotic activity comparable with that of peripheral zones (Steeves et al., 1969; Marc and Palmer, 1982). It was similar that no specific labeling has been observed in different parts of the vegetative shoot apex, which comprises regions with considerably different mitotic activity (Steeves et al., 1969). The increase in transcript abundance must be related to a cell- and/or tissue-specific developmental process, perhaps more related to mitochondrial biogenesis than to mitotic activity.

The importance of mitochondrial gene expression during flower development is well documented. Aberrant expression of mitochondrial genes is the cause of cytoplasmic male sterility, reflected in defects in pollen production (Hanson, 1991). The idea is that higher demands on mitochondrial ATP production exist during microsporogenesis so that partially defective mitochondria are not able to meet these demands. An approximately 40-fold increase in the number of mitochondria accordingly occurs during the formation of tapetal cells (Lee and Warmke, 1979).

Cell-specific expression of several mitochondrial genes in developing anthers is well documented (Conley and Hanson, 1994; Smart et al., 1994). On the other hand, a general increase in the expression of nuclear genes encoding mitochondrial components is thought to occur in flowers with respect to other organs (Huang et al., 1994; Landschütze et al., 1995; Felitti et al., 1997; Heiser et al., 1997; Zabaleta et al., 1998). However, since these analyses have dealt with total RNA preparations, differences in expression in specific cell types have not been reported, except for the -subunit of ATP synthase in sunflower (Balk and Leaver, 1998). These authors, using a homologous probe, have observed accumulation of transcripts encoding this subunit in meiocytes and tapetal cells at premeiosis and only in tapetal cells at later stages (Balk and Leaver, 1998). Using a different approach, Zabaleta et al. (1998) have observed that the promoters of three nuclear genes encoding complex I components direct enhanced -glucuronidase expression in anthers. Our results on cytochrome c gene expression seem to correlate with those obtained by Balk and Leaver (1998) and Zabaleta et al. (1998), although we were not able to detect specific expression in mature pollen. They also correlate quite well with observations on cell specific expression of mitochondrial genes during anther development in sunflower (Smart et al., 1994). We then postulate that common mechanisms should operate in the regulation of the expression of the nuclear-encoded cytochrome c gene and the expression of mitochondrial genes probably through the regulation of nuclear-encoded components involved in mitochondrial biogenesis. This model, similar to those proposed for the regulation of mitochondrial biogenesis in yeast (de Winde and Grivell, 1993) and mammals (Grossman and Lomax, 1997), deserves further investigation to determine if other nuclear genes show similar expression patterns. The fact that other genes increase their expression in flowers suggests that this may be the case. Results on components involved in mitochondrial biogenesis, however, are almost entirely lacking.

Not only in flowers the expression patterns of cytochrome c encoding transcripts show similarities with the expression of mitochondrial genes. Li et al. (1996) have detected higher mitochondrial transcript levels in meristematic cells and in the vascular cylinder of maize roots. Our results indicate that cytochrome c transcript levels are also high in these parts of the sunflower root. It is noteworthy that active mitochondrial biogenesis has been shown to occur in the root apical meristem of Arabidopsis and Pelargonium zonale (Kuroiwa et al., 1992; Fujie et al., 1993). A correlation between active mitochondrial biogenesis, higher mitochondrial transcript levels, and increased cytochrome c gene expression then is also evident in roots.

So far, we have emphasized the correlations between cytochrome c transcript abundance and mitochondrial biogenesis, assuming that both features should be somehow related. Although our studies suggest that this is mostly the case, it should be kept in mind that heterogeneity in mitochondrial subpopulations (Dai et al., 1998) as well as tissue-specific differences in protein content and/or respiratory activity of mitochondria (Day et al., 1985; Rios et al., 1991) have been observed. This opens the possibility that specific regulatory processes may operate for individual genes. In addition, cytochrome c function is not only related to respiratory activity. This protein recently has been implicated in early events that conduct to programmed cell death (Liu et al., 1996). Regulation of the amount of cytochrome c per cell or mitochondria may also be related to this process, which plays important roles during the life cycle of plants.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

Sunflower (Helianthus annuus L. cv contiflor 15, from Zeneca seeds) plants were grown in pots in a greenhouse under natural light. Roots were collected 72 h after germination. Inflorescences at different developmental stages were harvested and classified as defined by Schneiter and Miller (1981). Arabidopsis Heyhn. ecotype Columbia (Col-0) was purchased from Lehle Seeds (Tucson, AZ). Plants were grown in pots in a growth chamber at 22°C to 24°C under long-day photoperiods (16 h of illumination by a mixture of cool-white and GroLux fluorescent lamps) at an intensity of approximately 200 µE m² s¹ until flowering. Floral buds were collected and sorted by size. The size was measured along the longitudinal axis.

In Situ Hybridization

Tissue preparation and in situ hybridization were carried out essentially as described by Burgess (1995). Plant material was fixed overnight in 3.7% (w/v) formaldehyde, 5% (v/v) acetic acid, 47.5% (v/v) ethanol at room temperature, dehydrated through an ethanol series, and embedded in Histoplast (Biopack, Buenos Aires). Sections (5-7 µm thick depending on the material) were mounted on slides coated with 50 µg/mL poly-D-Lys (Sigma, St. Louis) in 10 mM Tris-HCl, pH 8.0, and dried overnight at 42°C. After removing the paraffin with xylene, sections were rehydrated by an ethanol series and treated with 1 µg/mL proteinase K for 30 min at 37°C and then with acetic anhydride in 100 mM triethanolamine, pH 8.0, for 10 min at room temperature. After a brief wash with water, sections were used for hybridization.

Digoxigenin-labeled RNA sense and antisense probes were synthesized using the DIG-RNA labeling mix (Boehringer Mannheim, Basel) and T3 or T7 RNA polymerase, according to the manufacturer's instructions. Labeled RNA was precipitated with LiCl, and ethanol, and its concentration and integrity were checked in agarose gels. Cytochrome c cDNA clones from sunflower (Felitti et al., 1997) or Arabidopsis (EST clones 172G24T7 or APZ69d09) constructed in pBluescript SK and linearized with appropriate restriction enzymes were used as templates.

Tissue sections were prehybridized in a moist chamber for 60 min at 44°C in 300 µL per slide of a solution containing 50% (v/v) formamide, 4× SSC (1× SSC is 0.15 M NaCl, 0.015 M Na₃-citrate, pH 7.0), 5% (w/v) dextran sulfate, 0.02% (w/v) polyvinylpirrolidone, 0.02% (w/v) bovine serum albumin, 0.02% (w/v) Ficoll, and 0.25 mg/mL of yeast tRNA. Hybridization was carried out overnight under similar conditions with a 1-µg/mL probe. After they were hybridized, sections were washed twice in 2× SSC at 42°C for 15 min and once in 500 mM NaCl, 10 mM Tris-HCl, pH 8.0, 0.1 mM EDTA at 37°C for the same time. Single-stranded RNAs were removed from the sections with a 30-min incubation in 10 µg/mL RNase A in the same buffer, followed by washes (15 min each) in 2× SSC, 1× SSC, 0.5× SSC (at 37°C), and 0.2× SSC (at room temperature). For probe detection, slides were washed twice with 20 mM Tris-HCl, pH 7.5, 0.48 M NaCl, blocked for 30 min in 1% (w/v) Blocking Reagent (Boehringer Mannheim), incubated for 2 h at room temperature with alkaline phosphatase labeled anti-digoxigenin (Boehringer Mannheim) diluted 1:500 in the same buffer, and developed after two new washes using the 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium colorimetric system. Development was carried out until satisfactory signals were obtained (usually overnight). After stopping the reaction with 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, sections were mounted with 15% (w/v) gelatin, 50% (w/v) glycerol in water with traces of phenol. Digital images were generated using Image ProLite software.

RNA Isolation and Analysis

Total RNA was isolated by phenol extraction and LiCl precipitation according to Ausubel et al. (1987). For northern-blot analysis, specific amounts of RNA were electrophoresed through 1.5% (w/v) agarose/6% (w/v) formaldehyde gels. The integrity of the RNA and equality of RNA loading were verified by ethidium bromide staining. RNA was transferred to nylon membranes (Hybond N, Amersham) and hybridized overnight at 65°C to a ³²P-labeled sunflower cytochrome c probe (full length 590-bp insert, Felitti et al., 1997) in buffer containing 6× SSC, 0.1% (w/v) polyvinylpirrolidone, 0.1% (w/v) bovine serum albumin, 0.1% (w/v) Ficoll, 0.2% (w/v) SDS, and 10% (w/v) polyethylene glycol 8,000. Filters were washed with 2× SSC plus 0.1% (w/v) SDS at 65°C (4 times, 15 min each), 0.1× SSC plus 0.1% (w/v) SDS at 37°C during 15 min, dried and exposed to Kodak X-AR films. To check the amount of total RNA loaded in each lane, filters were then reprobed with a 25S rDNA from Vicia faba under similar conditions as those described above, except that hybridization was performed at 62°C and the wash with 0.1× SSC was omitted.