Assessing photosystem I and II distribution in leaves from C4 plants using confocal laser scanning microscopy


Correspondence: E.Pfündel. Fax: + 49 931 888 6235; E-mail:


Images of chlorophyll fluorescence emitted at wavelengths above and below 700 nm were recorded from leaf sections of C4 species using confocal laser scanning microscopy (LSM). We investigated species exhibiting both NAD-malic enzyme (NAD-ME) C4 photosynthesis and NADP-malic enzyme (NADP-ME) C4 photosynthesis. Comparing LSM fluorescence of leaf sections with flow-cytometrically determined fluorescence from individual chloroplasts revealed that LSM fluorescence was distorted by the optical properties of leaf sections. Leaf section fluorescence, when corrected by transmission data derived from light transmission images, agreed with flow cytometry data. The corrected LSM fluorescence yielded information on the distribution of the individual photosystems in the C4 leaf sections: PSII concentrations in bundle sheath cells were elevated in NAD-ME species but diminished in most of the NADP-ME species investigated. The NADP-ME species, Arundinella hirta, however, showed normal PSII and increased PSI concentration in bundle sheath chloroplasts. Finally, a gradient of PSI was observed within the bundle sheath cells from Euphorbia maculata.


The photosynthetic performance of an entire leaf reflects the action of the anatomically and physiologically different photosynthetic cells inside the leaf, and, consequently, the understanding of whole-leaf photosynthesis requires information on the intra-leaf gradients of the different components involved in the photosynthetic processes. A conspicuous example of the heterogeneity of photosynthetic components is the gradient of sun- and shade-type chloroplasts within C3 leaves grown under high light conditions ( Outlaw 1987; Terashima 1989; Nishio, Sun & Vogelmann 1994; Sun, Nishio & Vogelmann 1996). Important factors for this chloroplast differentiation are the gradients in light intensity and light quality inside the leaf ( Nishio, Sun & Vogelmann 1993; Vogelmann 1993).

Studies, using whole leaves that were either sun-adapted or shade-adapted, indicated that the different light environments resulted in varying ratios of photosystem I/photosystem II (PSI/PSII; Boardman 1977; Lichtenthaler et al. 1981 ; Anderson 1986). Therefore, gradients of the PSI/PSII ratio may also exist inside C3 leaves, and, hence, such gradients could reveal the spatial pattern of the intra-leaf light gradients. Principally, the concentrations of PSI and PSII can be assessed by measuring the chlorophyll fluorescence emitted above and below 700 nm because the former spectral window is dominated by PSI emission and the latter by PSII emission ( Govindjee 1995; Croce et al. 1996 ). However, fluorescence at wavelengths shorter than 700 nm is re-absorbed by chlorophyll to varying degrees, whereas fluorescence re-absorption at wavelengths longer than 700 nm is negligible ( Weis 1985; Richter & Fukshansky 1994). Consequently, the assessment of photosystem ratios from fluorescence measurements requires correction for re-absorption artefacts.

In this work, we analyse the effects of re-absorption on the fluorescence signal from leaf cross-sections using confocal laser scanning microscopy (LSM). To quantify such effects, one requires detailed information on the true fluorescence emission of chloroplasts inside the specimen. Therefore, we confine our investigations to C4 leaves exhibiting either NADP-malic enzyme (NADP-ME) or NAD-malic enzyme (NAD-ME) photosynthesis for two reasons: firstly, each of the two biochemical types of C4 photosynthesis exhibits two distinct photosynthetic compartments which differ in their PSI/PSII ratio ( Edwards & Walker 1983; Hatch 1987). Hence, a coarse estimate on the correctness of the fluorescence from leaf sections can be made by comparing the measured ratios of long-wavelength to short-wavelength chlorophyll fluorescence with the known PSI/PSII ratios. Secondly, for the plants chosen, the fluorescence characteristics of the individual chloroplasts from the two photosynthetic domains are known from earlier investigations using flow cytometry ( Pfündel & Meister 1996; Pfündel, Nagel & Meister 1996). Therefore, we are able to carry out a detailed analysis of fluorescence re-absorption by comparing the fluorescence from leaf sections with that of single chloroplasts. On the basis of our analysis, we demonstrate new aspects of the distribution of PSI and PSII within C4 leaves.



All plants were cultivated in the greenhouse in pots of garden mulch. From October to April, supplementary light of 150–200 μE m−2 s−1 intensity, measured at leaf levels, was given from 0600 to 2200 h using 400 W high-pressure sodium lamps (SON-T AGRO 400, Phillips, Belgium). Only fully developed leaves were investigated. The NADP-ME species studied were Arundinella hirta (Thunb.) Tanaka, Cyperus papyrus L., Euphorbia maculata L., Flaveria australasica Hook, Flaveria trinervia (Spreng.) C. Mohr, Portulaca grandiflora Hooker, Saccharum officinarum L., Setaria viridis (L.) Beauv. and Zea mays L. Based on the fluorescence characteristics determined earlier ( Pfündel & Pfeffer 1997), we group the ‘C4-like’ ( Ku et al. 1991 ) Flaveria palmeri J.R. Johnston among the NADP-ME C4 species. The NAD-ME species examined were Atriplex rosea L., Panicum miliaceum L. and Portulaca oleracea L. A summary of the distribution of the C4 types of photosynthesis in our species is reported in Pfündel et al. (1996) and Pfündel & Pfeffer (1997); the distribution in A. hirta and P. miliaceum is described in Dengler, Dengler & Grenville (1990) and Mateu-Andrés (1993), respectively.

Confocal laser scanning microscopy

Fragments of freshly harvested leaves were embedded in tissue freezing medium (Leica Instruments GmbH, Nussloch, Germany) at – 25°C. Cross-sections of 20–40 μm thickness were prepared using a cryo-microtome (Leica CM3050-Cryostat, Leica Instruments GmbH) at –15 °C, and collected on poly- L-lysine-coated slides (Sigma, St Louis, Missouri, USA). Probably, the freeze–thaw cycle could affect the chloroplast ultra-structure; however, flow cytometric investigations showed that even membrane particles released from chloroplasts were unaltered in their fluorescence properties ( Pfündel & Meister 1996). Hence, structural changes during sample preparation do not affect the fluorescence signal that we consider here. To obtain maximum PSII fluorescence, leaf sections were overlaid with 10 μM of the PSII inhibitor 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) dissolved in 100 m M potassium phosphate buffer pH 7·0. Prior to microscopy, DCMU was allowed to diffuse into the specimen for at least 30 min.

Microscopy of sections was carried out using a Laser-Scan-Microscope LSM 410 invert (Carl Zeiss, Oberkochen, Germany) that includes an Axiovert 135M microscope. Chlorophyll fluorescence was excited by 488 nm light from a 25 mW argon ion laser (Carl Zeiss). Excitation and emission light was separated by a FT 488/510 dichroic mirror (Carl Zeiss). Chlorophyll fluorescence was separated into two wavebands by a F43-700 dichroic mirror (AHF Analysentechnik, Tübingen, Germany) which, at 45° to the beam, preferably reflects light below 700 nm but transmits light above 700 nm (FR and FFR fluorescence, respectively). The FR fluorescence was further confined by a combination of short-pass and long-pass filters (700 FL 07-25, LOT-Oriel, Darmstadt, Germany, and RG 645, Schott, Mainz, Germany, respectively). The FFR fluorescence was confined to the spectral region above 700 nm by a long-pass filter (RG 715, Schott). In both cases, red-sensitive photo- multiplyer tubes (R 3896, Hamamatsu, Toyooka, Japan) were used for signal detection. Confocal blue-green fluorescence from cell walls was excited by 365 nm light from a 50 mW argon ion laser (INNOVA ENT 651, Coherent, Palo Alto, California, USA). An FT 395 dichroic mirror (Carl Zeiss) was used to separate the UV excitation light from blue-green fluorescence which was passed through a BP 450–490 band-pass filter (Carl Zeiss) and measured by an R 4632 photo-multiplyer tube (Hamamatsu). Non-confocal transmission images were obtained using 543 nm light from a 0·5 mW HeNe laser (Carl Zeiss) with the laser intensity attenuated to 1%. The LSM 4/V 3·96 software (Carl Zeiss) was used for instrument control and image analyses. Images were obtained by averaging 8 or 16 measurements in line mode. In this mode, each line is scanned repeatedly and the signal is averaged before the laser beam proceeds to the next line in the frame. To quantify the relative intensities of chlorophyll fluorescence, mean digital values of a fluorescing area in the mesophyll compartment and from an adjacent fluorescing area in the bundle sheath compartment were determined. The background signal was taken from the nearest region outside the leaf section. For the calculation of bundle sheath/mesophyll fluorescence quotients, only background-corrected fluorescence intensities were used. Five fluorescence ratios were determined per image and averaged. We analysed between 5 and 32 images per species.

For the analyses of transmission images, only areas exhibiting chlorophyll fluorescence were considered. Similar to the chlorophyll fluorescence images (see above), relative transmittance values were taken from adjacent mesophyll and bundle sheath areas, and from a neighbouring blank region. The ratio of the ‘background minus bundle sheath’ value to the ‘background minus mesophyll’ value provided a relative absorptance ratio (where absorptance is defined as 1 – transmittance). Five such ratios were calculated per image, and between 3 and 16 images per species were analysed. To obtain intensity profiles from fluorescence and transmission images, we used the Osiris Medical Imaging Software Version 3·12 (University Hospital of Geneva, Switzerland).

Flow cytometry

Preparations enriched in either mesophyll or bundle sheath chloroplasts were obtained according to Höfer, Santore & Westhoff (1992) with the modifications specified in Pfündel & Meister (1996). Flow cytometrical analysis of chloroplasts was done with a FACStarPlus sorting flow cytometer (Becton Dickinson, San Jose, California, USA) as described previously ( Pfündel & Meister 1996). Chlorophyll fluorescence of single particles was excited by 488 nm light from an argon ion laser operated at 500 mW output power. The chlorophyll fluorescence was separated into short and long wavebands using the filter combination specified for the microscope (see above). We noticed that identically specified filters exhibited different transmission properties; consequently, the spectral windows used in the microscope are not entirely identical to those of the flow cytometer. Mesophyll and bundle sheath chloroplasts were identified by simultaneously monitoring the FR and FFR fluorescence. The number of flow cytometric determinations ranged from 3 to 19 per species. A flow cytometric analysis of most of the species investigated here is presented in Pfündel et al. (1996) .


Figure 1 shows microscopic images of the NADP-ME C4 species C. papyrus, A. hirta and E. maculata. For the former two species, the blue-green fluorescence from cell walls is shown in blue to outline the anatomy of the C4 leaf [panels (a)–(h)]. Because E. maculata did not show prominent blue-green fluorescence, a transmission image is drawn in grey to be used as the background [panels (i)–(l)]. Chlorophyll fluorescence is placed in different colours over the background images. All species in Fig. 1 display the typical Kranz-type anatomy described by Haberlandt (1881) which is characterized by a ring of chlorophyll-containing cells around the vascular bundle, i.e. the ‘bundle sheath cells’. The bundle sheath cells, in turn, are surrounded by the smaller chlorophyll-containing ‘mesophyll cells’.

Figure 1.

Fluorescence images from leaf cross-sections from C. papyrus (a–d), A. hirta (e–h), and E. maculata (i–l). For C. papyrus and A. hirta, blue-green fluorescence from cell walls is shown in blue and used as the background for chlorophyll fluorescence images. For E. maculata, the background was a transmission image. The images for FFR (a,e,i) and for FR (b,f,j) use red and green colour scales, respectively. Within a particular colour scale, dark and light tones indicate low and high fluorescence intensities, respectively. The colours of the fluorescence image in (c), (g) and (k) arise from blending red and green fluorescence images. In (d), (h) and (l), we show images for FFR/FR which were calculated from the digital values of FFR and FR images. The colours of the ratio images are translated into digital values by the colour bars in the figure. MC and BC indicate mesophyll and bundle sheath cells, respectively. DC indicates particular cells in A. hirta which are similar to bundle sheath cells but not adjacent to a vascular bundle ( Dengler et al. 1990 ). The inserts in (i)–(l) show images from E. maculata at low magnification.

To visualize different fluorescence intensities, 8-bit single-colour scales were used, where bright tones correspond to high intensities and dark tones correspond to low intensities: in Fig. 1(a, e & i), a red colour scale was used to display the FFR fluorescence, and a green colour scale was selected for the FR fluorescence in Fig. 1(b, f & j). Combinations of the images for FFR and for FR are presented in Fig. 1(c, g & k). The coloration and the brightness of the latter images result from the pixelwise blending of the original single-colour images and hence depend on the ratios and on the intensities of FFR and FR. Largely independent from the fluorescence intensities, information on the ratio of FFR/FR can be gained from the division of the FFR image by the FR image ( Fig. 1d, h & l), where the colours in ratio images are translated into digital values by the colour bars in Fig. 1.

In C. papyrus, the FR intensity from bundle sheath cells is considerably lower than that from the mesophyll cells ( Fig. 1b), whereas the signal for FFR was similar in both cell types ( Fig. 1a). Blending the FFR and FR images stresses that the bundle sheath cells in C. papyrus emit mostly FFR fluorescence: the bundle sheath cells appear reddish ( Fig. 1c). In comparison, the yellow coloration in mesophyll cells indicates comparable digital values in the original images. Correspondingly, the FFR/FR fluorescence ratio image ( Fig. 1d) exhibits much higher values in bundle sheath cells compared to mesophyll cells. The fluorescence ratio data from C. papyrus agree with the known PSI/PSII ratios in leaves from NADP-ME C4 species which exhibit higher ratios of PSI/PSII in the bundle sheath cells than in the mesophyll cells ( Edwards & Walker 1983; Hatch 1987; Pfündel et al. 1996 ). The latter references also specify that the group of NAD-ME C4 species show higher PSI/PSII ratios in the mesophyll cells than in the bundle sheath cells. However, fluorescence microscopy with our NAD-ME species did not reveal increased FFR/FR fluorescence ratios in mesophyll cells (images not shown). This observation indicated that fluorescence from leaf sections is potentially artefactual and caused us to examine in detail the relation between the fluorescence from leaf sections and that from single chloroplasts as determined by flow cytometry.

In flow cytometry, we require the ratio of FFR/FR to differentiate between mesophyll and bundle sheath chloroplasts ( Pfündel & Meister 1996). However, FFR/FR depends on the flow cytometer settings which may vary between experiments. Because these variations affect the FFR/FR ratios of mesophyll and bundle sheath chloroplasts in a proportional way, we obtained a fluorescence parameter which is independent of flow cytometer settings by expressing the FFR/FR ratio from bundle sheath chloroplasts relative to the FFR/FR ratio from mesophyll chloroplasts. To compare flow cytometry with fluorescence microscopy, the analogous fluorescence quotient was derived from chlorophyll fluorescence images (see Material and methods).

For the three NAD-ME and nine NADP-ME species, we plotted the bundle sheath/mesophyll fluorescence quotients, determined by microscopy, against the corresponding data from flow cytometry ( Fig. 2). Figure 2a shows that our relative FFR/FR ratio from microscopy increased parallel to the data from flow cytometry. Linear regression analysis yielded a slope of 0·56 for the relation between microscopy and flow cytometry. That the slope differed from unity can be readily explained by different spectral characteristics of the two instruments used. Independent of the spectral differences between instruments, however, the regression line should meet the point (1,1). This prediction is evident if one considers a C4 plant with mesophyll and bundle sheath chloroplasts that exhibit identical fluorescence emission spectra. In this hypothetical case, the flow cytometer as well as the microscope should yield a value of 1 for the bundle sheath to mesophyll ratio for FFR/FR. Using the 95% confidence interval as the criterion, however, the regression line in Fig. 2a misses the point (1,1).

Figure 2.

Relation between chlorophyll fluorescence from leaf sections and from individual chloroplasts. The figure includes data from 12 C4 species which are identified in the lower right panel. Closed and open symbols indicate NAD-ME C4 species and NADP-ME C4 species, respectively. (a) Data from microscopy (LSM) plotted against data from single chloroplasts measured by flow cytometry (FlCy). In both cases, the FFR/FR values from the bundle sheath are expressed relative to the FFR/FR values from the mesophyll: (FFR/FR)B/(FFR/FR)M. (b) The quotients of the ordinates of (a) and the abscissas of (a) plotted against the corresponding bundle sheath to mesophyll ratios for the absorptance of leaf sections (Abs: absorptance = 1 – transmittance). (c) uses the same abscissas as (a), the ordinates are the (FFR/FR)B/(FFR/FR)M ratios obtained from microscopy (the ordinates of (a)) normalized to the bundle sheath to mesophyll absorptance ratio (the abscissas of (b)). In (a) and (c), the cross (+) marks the point (1,1). Standard deviations are shown as bars. Straight lines result from first-order regressions, with 95% confidence intervals included in (a) and (c). Values for r2 are displayed in all panels, and analysis of variance yielded P values < 0·0001 for all regressions. The slope and the y-axis intercept of the regression lines were 0·557 (SE = 0·078) and 0·696 (SE = 0·141), 1·136 (SE = 0·176) and – 0·348 (SE = 0·220), and 0·680 (SE = 0·057) and 0·240 (SE = 0·104), for (a), (b) and (c), respectively.

Because instrumental variances fail to explain the disagreement between microscopy and flow cytometry, we pursued the idea that re-absorption affects the FR fluorescence from our optically dense leaf sections, but does not noticeably influence the fluorescence from single chloroplasts. Our notion agrees with recent data which demonstrate fluorescence re-absorption for single protoplasts ( Pfündel & Meister 1998); that is, for an optical light path which is smaller than that of our cross-sections. Artefacts caused by varying light absorption in cross-sections do not explain the shifted relationship between fluorescence quotients from microscopy and flow cytometry ( Fig. 2a) because such effects would cancel out in our bundle sheath to mesophyll ratio for FFR/FR.

For fluorescence re-absorption, the absorptance in the wavelength interval at which chlorophyll absorption overlaps the chlorophyll fluorescence is important; that is, in the range between 680 and 700 nm. This interval corresponds to the long-wavelength flank of the red chlorophyll a absorptance ( Govindjee 1995) at which light absorption is significantly smaller than that at peak wavelengths. Therefore, absorptance data for wavelengths which are efficiently absorbed by chlorophyll a are not appropriate to correct for re-absorption artefacts. In the absence of laser light in the 680–700 nm range, we chose the weakly absorbed 543 nm emission from a HeNe laser to characterize our cross-sections.

Typical transmission images taken with 543 nm light are shown for A. rosea and Z. mays in Fig. 3. The optical properties of our leaf sections were quantified by calculating the relative absorptance values for bundle sheath and mesophyll cells (see Material and methods). Ratios of bundle sheath/mesophyll absorptance indicate that most of our species absorb more light in bundle sheath cells than in mesophyll cells: the group of NAD-ME species exhibited the greatest difference between the two cell types ( Fig. 2b). We next quantified the assumed relative error in microscopy by expressing our bundle sheath to mesophyll fluorescence parameter from microscopy (the ordinates of Fig. 2a) relative to the analogous parameter from flow cytometry (the abscissas of Fig. 2a). Figure 2b shows that this error parameter appears to be linearly related to the bundle sheath/mesophyll absorptance ratio. This relationship can be rationalized by considering that re-absorption decreases the true FR fluorescence to a greater degree in cells with high absorptance than in cells with low absorptance. Consequently, the more the absorptances differ between bundle sheath and mesophyll cells, the greater becomes the distortion of the true mesophyll/bundle sheath ratio for FR, and thus the greater becomes our error parameter.

Figure 3.

Light transmission images of leaf cross-sections from Z. mays (a) and A. rosea (b).The figure shows micrographs of a leaf exhibiting similar absorptances, and largely different absorptances in the bundle sheath and the mesophyll compartment, in (a) and (b), respectively (compare Fig. 2a). Areas which were subjected to absorptance analysis in transmission images were defined from parallel fluorescence images (see Material and methods).

Since the relation in Fig. 2b suggests that optical heterogeneities in leaf sections influence the FR signal, we corrected our microscopy data (the ordinates of Fig. 2a) by the bundle sheath/mesophyll absorptance ratio (the abscissa of Fig. 2b), and plotted the corrected values against the data from flow cytometry ( Fig. 2c). Clearly, the absorptance- corrected data from microscopy are significantly better correlated with the flow cytometric results than the non- corrected values. In addition, microscopy fluorescence ratios for NAD-ME species became smaller than 1, and hence agreed with the published differences in PSI/PSII ratios between bundle sheath and mesophyll cells. Most importantly, the (1,1) point is now included in the 95% confidence interval of the first-order regression, as postulated above. In summary, Fig. 2c strongly suggests that the absorptance-corrected FR fluorescence corresponds to the true FR emission in leaf sections. Therefore, we will assess the relative concentrations of PSI and of PSII in C4 leaves from original FFR and from absorptance-corrected FR fluorescence values.

Besides photosystem concentrations, variations in fluorescence emission spectra of either of the two photosystems or fluorescence quenching processes could influence the fluorescence signal from leaf sections. Because the leaf sections were pretreated with DCMU, which inhibits PSII reaction centre quenching, and because earlier data support comparable photosystem fluorescence characteristics in bundle sheath and mesophyll chloroplasts ( Pfündel et al. 1996 ), our results will be discussed only in terms of photosystem concentrations. Clearly, the FFR and FR values do not translate directly into photosystem concentrations since the fluorescence emission spectrum of PSI exhibits a shoulder in the FR spectral window, and that of PSII shows vibrational satellite bands in the FFR spectral window ( Govindjee 1995; Croce et al. 1996 ). Within such limitations, the fluorescence analysis presented here provides direct information on the distribution of individual photosystems in the leaf, and hence differs markedly from conventional analyses that yield information only on the ratio of PSI/PSII by analysing isolated bundle sheath and mesophyll fractions from C4 leaves.

Figure 4a shows the bundle sheath to mesophyll ratio for FR of our 12 C4 species where the species are sorted according to their FR ratio. In NAD-ME species, the FR ratios indicate higher concentrations of PSII in bundle sheath than in mesophyll cells. There was no statistically significant difference in FFR between the two photosynthetic compartments, with the exception of P. oleracea ( Fig. 4b).

Figure 4.

Relation between the chlorophyll fluorescence from bundle sheath cells and mesophyll cells. All results represent data from bundle sheath relative to mesophyll cells. Means of the absorptance-corrected ratio for FR and of the ratio for FFR are shown in (a) and (b), respectively. Data from NAD-ME C4 species and from NADP-ME C4 species are drawn as black and white bars, respectively. The 95% confidence intervals are given as vertical bars.

Hence, our limited number of species suggests that in NAD-ME C4 plants an increased PSI/PSII ratio in mesophyll cells is primarily achieved by concentrating PSII in bundle sheath cells.

Among the NADP-ME C4 species, the members of the genus Flaveria exhibited the smallest differences in FR and FFR between bundle sheath and mesophyll cells ( Fig. 4). These modest differences can be most simply explained by the concomitant action of NADP-ME and NAD-ME C4 biochemistry as suggested by the data of Meister, Agostino & Hatch (1996). The remaining NADP-ME C4 species displayed a relatively uniform fluorescence behaviour: the low bundle sheath/mesophyll ratios for FR indicated that the PSII concentrations in bundle sheath chloroplasts are markedly reduced ( Fig. 4a), whereas the FFR ratios suggest similar PSI concentrations in the two cell types ( Fig. 4b). Because all the data in Fig. 4 indicate that the variations in the FR ratios can be much greater than in the FFR ratios, it is tempting to speculate that the PSII concentration can be efficiently regulated, while PSI is constitutively expressed.

However, this speculation is severely challenged by data from the NADP-ME species, A. hirta, which exhibited much higher FFR values in bundle sheath than in mesophyll cells ( Fig. 1). The similar values for the absorptance-corrected FR ( Fig. 5) suggest that the higher PSI/PSII ratio in bundle sheath cells compared to mesophyll cells is solely produced by increasing the PSI concentration in the bundle sheath cells. In A. hirta, the normal PSII concentration in the bundle sheath chloroplasts suggests that linear electron transport through both photosystems is present to some extent, and produces redox equivalents to drive CO2 reduction. A simple explanation for increased NADPH demands could be that the dihydroxyacetone phosphate/3-phosphoglycerate shuttle transporting redox equivalents from the mesophyll to the bundle sheath compartment in many other NADP-ME species ( Hatch 1987), does not operate in A. hirta.

Figure 5.

Bundle sheath/mesophyll chlorophyll fluorescence ratio in A. hirta in comparison to typical data from NADP-ME C4 species. Means of the absorptance-corrected ratio for FR, of the ratio for FFR, and of the absorptance-corrected bundle sheath/mesophyll ratio for FFR/FR are shown in (a), (b) and (c), respectively. Data from A. hirta (hatched bars) are compared with the mean of our NADP-ME species excluding Flaveria species (open bars). 95% confidence intervals are given as vertical bars.

In the above fluorescence analyses, we have only considered the average fluorescence intensities from cells and disregarded possible intensity gradients within a cell. Such gradients, however, were not obvious except in bundle sheath cells from E. maculata, which exhibited a radial gradient of FFR; in contrast, the FR fluorescence was rather homogeneously distributed in these cells ( Fig. 1). The images resulting from the combination of FFR and FR, or from calculating the quotient of FFR and FR, show a radial gradient of the FFR/FR ratio and hence suggest a gradient of the PSI/PSII ratio. Fluorescence intensity profiles of the microscopy images in Fig. 1i,j illustrate that the FR, after absorptance correction, remained relatively constant across bundle sheath cells ( Fig. 6); therefore, the gradient of FFR/FR is probably not caused by re-absorption phenomena.

Figure 6.

Fluorescence intensity profiles across a Kranz structure in E. maculata. An intensity profile of the absorptance-corrected FR fluorescence (FR× absorptance) is shown as a white line, and a profile of the original FR fluorescence is shown as a black line. Profiles were taken from Fig. 1 and are normalized to the same maximum value. As shown from left to right, the profile crosses mesophyll cells, a bundle sheath cell, the vascular bundle (VB), a bundle sheath cell, and mesophyll cells.

In summary, the fluorescence gradient in E. maculata, and also the different ways by which NADP-ME species increase the FFR/FR ratio in bundle sheath cells, stresses the heterogeneity within the group of NADP-ME C4 species. Thus, much more information appears to be needed to understand the complex regulation of cell-specific expression of photosystems in C4 plants.


We conclude that absorptance-corrected confocal laser scanning microscopy yields reliable information on the true fluorescence from PSI and PSII in cross-sections of C4 leaves. From our fluorescence data we have derived new insights into the photosystem distribution in C4 leaves which, to our knowledge, have not been reported previously: firstly, NAD-ME C4 species concentrate PSII in the bundle sheath chloroplasts; secondly, in sharp contrast to other NADP-ME C4 species, a high PSI concentration occurs in bundle sheath chloroplasts of A. hirta, and alone causes the high PSI/PSII ratio which is typically found in bundle sheath chloroplasts of NADP-ME C4 species; and, thirdly, in bundle sheath cells from E. maculata, a gradient for PSI exists. These results from C4 plants encourage future investigations that focus on photosystem gradients in C3 leaves. The use of cryogenic temperatures, which increase fluorescence yields and reduce the spectral overlap of the PSI and PSII emission spectra, will improve the performance of chlorophyll fluorescence microscopy.


We thank Dr Armin Meister for stimulating discussions during the course of this work, and Dr Bob Porra for help in preparing the manuscript. We are grateful to Dr Nancy Dengler for providing us with seeds from Arundinella hirta. We thank Jens Tiedemann for introducing us to cryo-microtomy. This work was supported by grants from the Deutsche Forschungsgemeinschaft and from the state of Sachsen-Anhalt.