Capparis spinosa (caper), a winter‐deciduous perennial shrub, is a consistent floristic element of Mediterranean ecosystems, growing from May to October, i.e. entirely during the prolonged summer drought. The internal architecture of young and fully expanded leaves was studied, along with certain physiological characteristics. Capparis spinosa possesses thick, amphistomatic and homobaric leaves with a multilayered mesophyll. The latter possesses an increased number of photosynthesizing cells per unit leaf surface, a large surface area of mesophyll cells facing intercellular spaces (Smes) and a low percentage of intercellular space per tissue volume. Smes and chlorophyll content attain their maximum values synchronously, slightly before full leaf expansion. Nitrogen investment is also completed before full leaf expansion. The structural features, in combination with the water status, could contribute to enhanced rates of transpiration and photosynthesis under field water shortage conditions.
Received: 10 February 2003; Returned for revision: 23 April 2003; Accepted: 26 May 2003 Published electronically: 9 July 2003
In the Mediterranean Basin, plants either avoid or tolerate the prolonged, dry and hot summer (Di Castri, 1973). Evergreen sclerophylls and drought semi‐deciduous shrubs that shed their leaves during dry periods are the dominant plants in this region (Margaris, 1981; Werner et al., 1999). Capparis spinosa L. (Capparaceae), a winter‐deciduous species, is one of the few perennial shrubs that grow and flower entirely during summer. In Dioscorides’s herbal, C. spinosa is referred to as a species distinct enough not to be confused with anything else (Raven, 1990). Rhizopoulou (1990) argued that the physiological response of C. spinosa to drought is based upon osmotic adjustment, regulation of stomatal opening, modification of cell wall properties and an extensive root system. It is likely that C.spinosa is a stenohydric plant, largely free of competition for water with other species in the Mediterranean region (Rhizopoulou et al., 1997; Sozzi, 2001).
Anatomical features of leaves are linked to CO2 assimilation rates (Miyazawa and Terashima, 2001) and to components of water status (Rascio et al., 1990). Thus, free mesophyll surface (Smes) is the real CO2 absorbing surface (Nobel, 1974; Caemmerer and Evans, 1991; Terashima, 1992; Miyazawa and Terashima, 2001). The ratio of Smes per leaf surface is correlated with the habitat (Fahn and Cutler, 1992), the tissue photosynthetic pathway (Dengler et al., 1994) and light conditions during the species’ growing period (Longstreth et al., 1985). Photo synthesis is also influenced by the volume of intercellular space, via the partial pressure of internal CO2 (Parkhurst, 1994).
There is another anatomical feature that seems to affect leaf photosynthetic performance. In many dicotyledonous leaves, parenchyma or sclerenchyma cells of the vascular bundle sheath extend to the epidermis on both leaf sides, forming the so‐called bundle sheath extensions. Bundle sheath extensions constitute partitions in the mesophyll, resulting in many mesophyll compartments (Terashima, 1992). These leaves have been characterized as heterobaric. Leaves possessing uniform mesophyll and evenly distributed stomata are characterized as homobaric. Most heterobaric leaves show stomatal patchiness (Beyschlag and Pfanz, 1990). The existence of mesophyll compartments seems to be an advantage in protecting mesophyll against water stress (Terashima, 1992).
The aim of this study was to investigate the structural and functional features of expanding leaves of C. spinosa, in particular, focusing on the mesophyll development and structure, as well as leaf water relations. An attempt was made to find structural and functional features that might influence leaf development.
MATERIALS AND METHODS
Five branches of the same developmental stage, i.e. of similar length and bearing up to seven leaves, were selected for leaf anatomical study. Each branch was from a different individual. Care was taken to collect branches with similar positions within the shrubs of Capparis spinosa that grew in an open field on Athens University Campus (38°57·5′N, 23°48·0′E, altitude 250 m). During the experimental period (May–October) night air temperature averaged 16–18 °C and day air temperature averaged 24–29 °C, with air humidity fluctuating between 25 and 48 %. In the following text L1 represents the youngest leaf, i.e. the first leaf emerging from the shoot tip, and L6 the oldest leaf. Leaf length was measured with a ruler.
Two samplings were made at the end of May in the morning (0800 h). Leaf blade samples from each leaf developmental stage (L1–L6) were carefully cut in the field and fixed in 5 % glutaraldehyde, in phosphate buffer at pH 7, at room temperature for 2 h. Care was taken to sample parts from the central leaf blade beneath mid‐vein, between one‐third to one‐half of the distance from mid‐vein to leaf margin. Tissue was then post‐fixed in 1 % OsO4 at 4 °C, dehydrated in ethanol, and embedded in Durcupan ACM (Fluka, Buchs, Switzerland). Semi‐thin sections (1–2 µm) of plastic‐embedded tissue made on a LKB Ultrotome III microtome were stained with Toluidine Blue O. At least three transverse and a series of paradermal sections from each leaf sample were examined and photographed using a Zeiss Axioplan microscope. Data concerning the adaxial leaf mesophyll were obtained from paradermal sections through mesophyll cell layers between upper epidermis and vascular bundles. Abaxial mesophyll was considered to be the mesophyll tissue restricted between lower epidermis and vascular bundles.
At least ten randomly selected micrographs of the same magnification of cross or paradermal leaf sections were used for measuring leaf thickness, cell number and cell dimensions for each developmental stage; for cell dimensions, at least 100 cells were measured per developmental stage. Intercellular air space (ICS) and the mesophyll cell surface exposed to internal leaf atmosphere (Smes) were measured using the methods of Parkhurst (1982). Some of the measurements were confirmed using the method of Turrell (1936) (see also Psaras and Rhizopoulou, 1995). Stomatal density was determined on both adaxial and abaxial surface, from paradermal sections and from leaf imprints made with a dental material (Sandro‐sil preci flow, Cologne, Germany); data presented are derived from five leaves, ten replicates for each leaf surface at each developmental stage. Leaves were infiltrated with water according to Beyschlag and Pfanz (1990) to verify ICS. Leaf area was measured with a Delta area meter (Delta‐T Devices, Cambridge, UK). Leaves were oven dried at 70 °C for 48 h to determine dry mass.
Leaf discs (6 mm diameter) were used for the estimation of water potential (ψ) values. Leaf discs were taken from leaves similar to those described above, and from the same individuals, early in the morning (0700–0800 h). They were placed in C‐52 psychrometric chambers (Wescor Inc., Logan, UT, USA) attached to a microvoltmeter (HR‐33T; Wescor Inc.). Solute potential (ψs) values were obtained from the same discs that had been used for measurements of water potential after freezing and thawing; the equilibration time for each of ψ and ψs was 2 h. Turgor potential (ψp) was calculated as the difference between ψ and ψs. Each reported value of ψ and ψs is the mean of five determinations per leaf. Chlorophyll content (Chl) was determined after Linder (1974), free proline accumulation using the method of Bates et al. (1973), soluble sugars according to Dubois et al. (1956) and starch content using the anthrone method (McCready et al., 1950). The total amounts of carbon and nitrogen in dried and ground samples were determined in a CHNOS analyser (vario EL; ELEMENTER Analyzen Supreme, Hanau, Germany). Each reported value of all the above‐mentioned parameters is the mean of three replicates per leaf.
Morphometry and quantitative anatomy
Morphometric characteristics of successive leaves along a shoot are given in Fig. 1. Leaf area increased from L1 (youngest leaf) concurrently with an increase in dry weight. Specific leaf area (SLA) decreased from L1 to L6 (Fig. 1A). Mesophyll and whole leaf thickness increased from L1 to L3 (Fig. 1B); the thickness of a single epidermis remained constant (Fig. 1B). In adaxial and abaxial mesophyll, the number of cells per unit of leaf surface decreased from L1 to L6 (Fig. 1C), whereas the diameter of photosynthetic cells increased from L1 to L6 (Fig. 1D). ICS increased during leaf expansion, being approx. 15 % of the volume in a mature leaf (Fig. 1E). Smes, as well as the ratio Smes /leaf surface, was very low in young expanding leaves, but increased substantially from L1 to L3 (Fig. 1F). Stomata (10 µm wide × 28 µm long) were uniformly distributed on both epidermises, though stomatal density on the adaxial epidermis was slightly higher than that on the abaxial epidermis (Fig. 1G). Stomatal density decreased from L1 to L5 (Fig. 1G). Leaf length plotted against leaf position on axis, indicates that the maximum length was found in L6 (Fig. 1H).
Young leaves of Capparis spinosa (L1, Fig. 2A–C) consisted of five or six layers of mesophyll cells with uniseriate upper and lower epidermises (Fig. 2A). Both epidermises consisted of rectangular cells, with vacuoles larger than those of the mesophyll cells (Fig. 2A). Undifferentiated parenchyma cells of adaxial and abaxial mesophyll were slightly elongated (Fig. 2A); they possessed dense cytoplasm and small vacuoles and were tightly packed, forming small, triangular intercellular spaces (Fig. 2B and C). L2 (Fig. 2D–G) was thicker than in L1; there was an increased number of cell layers and the height of mesophyll cells was greater (Fig. 2D). The external wall of epidermal cells was thicker than in L1 (Fig. 2D and G). As seen in paradermal sections, young mesophyll cells possessed larger vacuoles than in L1 (Fig. 2E and F), though smaller than those observed in mature mesophyll cells, and a slightly larger, irregular ICS. Some stomata seemed to be mature (Fig. 2G). In L3 (Fig. 2H–J) and L4 (Fig. 2K–N) both epidermises consisted of differentiated cells with thick external walls (Fig. 2H and K). Mesophyll cells possessed a central vacuole and several chloroplasts (Fig. 2I, J and L–N). Both abaxial and adaxial mesophyll possessed small ICS (Fig. 2I, J, L and M) when compared with L5 and L6 (Fig. 3A–D and F).
Older leaves (L5 and L6) possessed seven or eight layers of mesophyll cells (Fig. 3A and D). Adaxial mesophyll cells (Fig. 3B and E) appeared more cylindrical than abaxial cells (Fig. 3C and F). Mesophyll cells formed small, abundant ICS (Fig. 3E and F). The cylindrical photosynthetic cells exposed most of their surface to the internal leaf atmosphere, and the chloroplasts were distributed along the cell wall sites exposed to air spaces (Fig. 3B and E).
As seen in cross‐sectioned leaves of all developmental stages, vascular bundles lack bundle sheath extensions (Figs 1A, D, H and K and 2A and D), indicating homobaric leaves. The fresh weight of mature leaves infiltrated with water increased by 14 %, which indicates an analogous percentage of ICS. This value is not far from the 15 % shown in Fig. 1E.
Chlorophyll, proline, soluble sugars, starch, C, N and water relations
Total chlorophyll increased substantially from L1 to L3 (Fig. 4A), while the highest value of free proline accumulation was recorded in L1 (Fig. 4A). The highest value for soluble sugars was found in young leaves (L1, Fig. 4B), whereas starch content showed its highest value in mature leaves (L6, Fig. 4C). The C content in developing leaves of caper was 42·0 % (Fig. 4B) and the N content 1·5 % (Fig. 4C). In mature leaves, more negative values of ψ and ψs were detected, in comparison to the young leaves (Fig. 4D). A substantial decline of ψ and ψs observed from L3 to L4, as well as from L5 to L6, is related to the high values of ψp in expanding leaves (Fig. 4D).
The relatively small and thick leaves, the multilayered mesophyll, the thick outermost epidermal cell walls and the small leaf ICS percentage (15 %) of C. spinosa constitute features of xerophytes (Shields, 1950; Fahn and Cutler, 1992). Thick leaves, tending to be amphistomatic (Parkhurst, 1994; Pearson et al., 1995), are more common in xeric habitats (Parkhurst, 1978) and exhibit higher photosynthetic rates than hypostomatic leaves (Mott et al., 1982; Peat and Fitter, 1994). The presence of stomata on both leaf surfaces shortens the distance of CO2 diffusion to mesophyll cells (Parkhurst et al., 1988).
Leaves of C. spinosa seem to be homobaric. They lack bundle sheath extensions, they possess uniform mesophyll, and their stomata are evenly distributed across both leaf surfaces. Thus, CO2 diffusion across the mesophyll is unobstructed. Usually, stomatal density of homobaric leaves is smaller than that of heterobaric leaves (Terashima, 1992). However, stomatal density of C. spinosa is higher than that of the seasonally deciduous species, and is the highest recorded among plant life forms grown in arid environments (Sundberg, 1986). Stomata of C. spinosa are longer (average length 28 µm; data not shown) than the average stomata length (20 µm) of 134 desert species studied by Sundberg (1986). Among these 134 species, only two species were found to have slightly longer stomata [29·97 µm, Andromischus cristatus (Haw.) Lem.; 30·97 µm, Gasteria verrucosa (Mill.) Haw.], with the rest of the 132 species possessing shorter stomata than C. spinosa, and one species (Grewia flava DC.) possessing stomata 7·33 µm long. Although it is difficult to correlate directly the size and density of stomata with conductance, high stomatal density and large pores may support elevated stomatal conductance (Sundberg, 1986; Fahn and Cutler, 1992). In fact, stomatal conductance (g) found in C. spinosa (data not shown) is much higher than that of other species grown in the same environment (Tenhunen et al., 1990; Rhizopoulou et al., 1991; Tognetti et al., 1998).
Stomatal density in the adaxial epidermis of C. spinosa is 25 % higher than that in the abaxial; this is not unexpected, since perfect amphistomaty is rare (Parkhurst, 1994). This may result in differences in stomatal conductance between adaxial and abaxial surface of a mature leaf (Rhizopoulou et al., 1997). The environments of the adaxial and abaxial leaf surface differ in many ways. The adaxial surface is exposed to very high photon flux densities of solar radiation, while the abaxial surface, which is shaded by the mesophyll, usually receives a lower percentage of the photon flux densities incident to the adaxial surface (Lu et al., 1993). Also, it is possible that different vapour pressure gradients between the upper and lower surface of leaves that grow close to the ground, where mid‐day temperature exceeds 35 °C, influence stomatal aperture and, thus, stomatal conductance (Rhizopoulou et al., 1997).
The number of mesophyll cells per unit of leaf surface in C. spinosa is high. For example, it is six‐fold higher than that of the xerophyte Ballota acetabulosa L. (Psaras and Rhizopoulou, 1995). This can be attributed to (a) the small cell diameter (15–20 µm), (b) the leaf thickness and (c) the tight arrangement of cylindrical photosynthetic cells in both the adaxial and the abaxial mesophyll. Mesophyll cells of C. spinosa are closely packed, forming narrow air spaces. The high number of mesophyll cells in leaves of C. spinosa results in high Smes, which is 10–30 times that of the external leaf area. This particular feature seems to be important for the photosynthetic capacity, since total chloroplast surface adjacent to Smes has been positively correlated with CO2 assimilation rate during photosynthesis (Caemmerer and Evans, 1991). In C. spinosa, a high Smes accompanied by a low proportion of air space (15 % of the mature leaf) provides the structural basis for enhanced gas exchange rates under the intensive irradiance and the elevated temperature in the Mediterranean region. This happens in coincidence with high stomatal conductance, via an unlimited water supply to the expanding leaf, due to the extensive root system (Rhizopoulou, 1990).
A differential development of chloroplasts and Smes has been found in several species. According to Miyazawa and Terashima (2001) and Kursar and Coley (1992), species where the peak of photosynthesis develops synchronously with Smes can be characterized as having a ‘normal’ greening process, in contrast to other species which exhibit a ‘delayed’ greening process and reach their peak in photosynthesis and chloroplast development after full leaf expansion. Smes in C. spinosa has been developed considerably before full leaf expansion (L3, Fig. 1F). Chlorophyll content attains its maximum at the same leaf age, concomitantly with photosynthetic rate (data not shown). These results show that mesophyll cell expansion and chloroplast development proceeds synchronously before full leaf expansion.
Starch, soluble sugars and proline content in developing leaves of C. spinosa are low when compared with those of the dominant species grown in the same environment (Rhizopoulou et al., 1989, 1991). In expanding leaves, soluble sugar and proline content decreased as both leaf area and dry weight increased, indicating a declining contribution of osmolyte to the maintenance of turgor. N in leaves of this species, which is widespread on arid soil, is higher than N in other winter deciduous shrubs (Donovan et al., 1996). This may result from N2‐fixing micro‐organisms in its rhizosphere (Andrade et al., 1997).
Differences in ψ and ψs between L4–L6 and L1–L3 indicate an increase in solute accumulation and an increased tendency of water uptake as leaves expand; this probably reflects the xylem water status. In fact, the fine structure of the xylem of C. spinosa indicates a high hydraulic conductivity and adequate water supply to the leaves (Psaras and Sofroniou, 1999). In accordance with an earlier report (Rhizopoulou, 1990), C. spinosa exhibits more negative ψ values during midday elevated transpiration rates, when compared with the morning and afternoon partial stomatal closure (Rhizopoulou et al., 1997).
Caper exhibits a deep and extensive root system (Rhizopoulou, 1990) and a highly specialized conducting tissue (Psaras and Sofroniou, 1999). Apart from these features, this work revealed that C. spinosa possesses thick, amphistomatous and homobaric leaves that bear a large number of mesophyll cells and have a high free mesophyll surface and a low percentage of intercellular space. The synchronous development of Smes and chloroplasts under adequate water supply before full leaf expansion might be another feature enabling the plant to complete its biological cycle during the drought period.
Measurements with CHNOS vario EL were conducted in Bielefeld, via a joint research project (70/3/3867) between Greece and Germany. We thank Mr K. Philippou and Mrs A. Papavassiliou for help with fieldwork and anonymous referees for critical comments on an earlier draft of the manuscript.
1Section of Botany, Department of Biology, University of Athens, Athens 157 81, Greece and 2Section of Plant Biology, Department of Biology, University of Patras, Patras 265 00, Greece