Functional analysis of corn husk photosynthesis.

The husk surrounding the ear of corn/maize (Zea mays) has widely spaced veins with a number of interveinal mesophyll (M) cells and has been described as operating a partial C(3) photosynthetic pathway, in contrast to its leaves, which use the C(4) photosynthetic pathway. Here, we characterized photosynthesis in maize husk and leaf by measuring combined gas exchange and carbon isotope discrimination, the oxygen dependence of the CO(2) compensation point, and photosynthetic enzyme activity and localization together with anatomy. The CO(2) assimilation rate in the husk was less than that in the leaves and did not saturate at high CO(2), indicating CO(2) diffusion limitations. However, maximal photosynthetic rates were similar between the leaf and husk when expressed on a chlorophyll basis. The CO(2) compensation points of the husk were high compared with the leaf but did not vary with oxygen concentration. This and the low carbon isotope discrimination measured concurrently with gas exchange in the husk and leaf suggested C(4)-like photosynthesis in the husk. However, both Rubisco activity and the ratio of phosphoenolpyruvate carboxylase to Rubisco activity were reduced in the husk. Immunolocalization studies showed that phosphoenolpyruvate carboxylase is specifically localized in the layer of M cells surrounding the bundle sheath cells, while Rubisco and glycine decarboxylase were enriched in bundle sheath cells but also present in M cells. We conclude that maize husk operates C(4) photosynthesis dispersed around the widely spaced veins (analogous to leaves) in a diffusion-limited manner due to low M surface area exposed to intercellular air space, with the functional role of Rubisco and glycine decarboxylase in distant M yet to be explained.

The maize (Zea mays) leaf utilizes CO 2 to make sugars using the C 4 photosynthetic pathway, in which carbon assimilation is essentially split into two distinct cycles within the leaf. The partitioning of these two cycles is facilitated by two specialized photosynthetic cell types within the leaf: the bundle sheath (BS) cells, which are clustered around vascular bundles (VB) and are surrounded by mesophyll (M) cells, forming a wreath-like formation known as Kranz anatomy (Dengler and Nelson, 1999). Initially in the C 4 cycle, CO 2 is fixed by phosphoenolpyruvate carboxylase (PEPC) in M cells, which leads to the formation of the C 4 dicarboxylic acids malate and Asp. These acids then diffuse to BS cells via plasmodesmata, where they are decarboxylated. The released CO 2 is subsequently fixed by Rubisco in the C 3 cycle (Edwards and Walker, 1983;Hatch, 1987). This cellular partitioning of the two cycles enables Rubisco to operate in a CO 2 -rich envi-ronment, limiting photorespiration and maximizing photosynthetic CO 2 assimilation. C 4 photosynthesis is suggested to have more than 50 independent evolutionary origins (Sage, 2004;Muhaidat et al., 2007). Biochemical and structural diversity within different C 4 pathways include variation in the primary decarboxylating enzyme, the position of chloroplasts in BS cells, the degree of M and BS chloroplast granal development, and the size, number, and structure of BS mitochondria (Edwards and Voznesenskaya, 2011). Maize is a well-studied monocot example of the NADP-malic enzyme (ME) biochemical C 4 subtype, characterized anatomically by the absence of a mestome sheath, centrifugal position of organelles in chlorenchymatous BS, grana-deficient chloroplasts in BS cells, and chloroplasts with well-developed grana in M cells (Gutierrez et al., 1974).
A feature considered essential in C 4 photosynthesis for maintaining efficient transport of C 4 acids from M to BS cells is high vein density or fewer cells between VB (Sage, 2004;McKown and Dengler, 2007). Typically, there are very few M cells between adjacent BS cells in most C 4 grasses, including maize (Hattersley and Watson, 1975;Dengler et al., 1994). However, in maize, the photosynthetically active husk that covers the ear exhibits a very low vein density, similar to C 3 plants, with up to 20 M cells between VB (Langdale et al., 1988). A study by Langdale et al. (1988) looking at in situ localization of C 3 and C 4 enzymes in maize husk hypothesized that M cells distant from the VB were in fact performing C 3 photosynthesis. They identified Rubisco mRNA transcripts present throughout husk M cells and also observed a slight oxygen dependence on the CO 2 assimilation rate, a feature typically absent in C 4 plants. Further evidence for this theory was reported by Yakir et al. (1991), who measured the natural abundance of carbon isotopes ( 13 C and 12 C) in maize husk and leaf dry matter, calculating that there was significant C 3 fixation of CO 2 in the husk, contributing to the production of husk cellulose (16%).
A later study used real-time PCR analysis of photosynthetic enzyme transcripts, including Rubisco, PEPC, NADP-ME, and pyruvate phosphate dikinase, in maize leaves and husk, which again showed Rubisco transcripts in husk M cells distant from the BS clusters surrounding the VB (Hahnen et al., 2003). Results from these studies have contributed to the theory that vein spacing influences the pattern and degree of photosynthetic gene expression and that the accumulation of C 4 enzymes is regulated locally around individual veins.
At present, an international focus on engineering new C 4 crop plants from existing C 3 plants to supply future food demand has led to a considerable interest in understanding the evolutionary path from C 3 to C 4 photosynthesis (Sheehy et al., 2007;Westhoff and Gowik, 2010). This understanding is crucial to deciding which structural and genetic features of photosynthesis must be adapted or developed to make a C 4 plant and those that are common to both systems. Using maize as a model of possible coexistence of C 3 and C 4 photosynthetic tissues, the aim of this study was to compare the form of photosynthesis carried out in maize husk with that in leaves using the standard physiological, anatomical, and biochemical techniques employed to distinguish between C 3 and C 4 photosynthesis. Here, we present to our knowledge the first detailed gasexchange perspective of husk photosynthesis, including measurements of compensation points and carbon isotope discrimination, and link it to anatomical fea-tures, photosynthetic enzyme activities, and immunolocalization of Rubisco, PEPC, and Gly decarboxylase (GDC).

Anatomical Measurements of Leaf and Husk
In the C 4 monocot maize, there are two types of photosynthetic organs, the leaf and the leaf sheath. In the ear, the husk (which is an expanded sheath) has a leaf at the terminal end in some genotypes. The leaf emerging from the husk is analogous to a leaf emerging from the sheath of the stem. Two varieties of corn/ maize, B73 and Sweet Corn, Kelvedon Glory, F1 (SWC), were used to analyze the anatomy of the husk and leaf. In both maize varieties, we use "husk" to refer to the outer layer husk that covers the ear. We use the term "leaf" to refer to a standard leaf originating at the stem in B73 and for the leaf protruding from the terminal end of the husk in SWC.
The classic Kranz-type anatomy, with characteristic centrifugal position of chloroplasts in BS cells, is clearly present in all maize leaves (Fig. 1, A, C, and E), but there is a highly modified version in the husk ( Fig. 1, B, D, and F). In the husk, a clear layer of BS cells containing a few mostly centrifugally located chloroplasts ( Fig. 1, D and F) surround the veins, with large, rounded adjacent M cells (M1) and more distant M cells between the veins. In leaf cross-sections, the M:BS cell ratio was 2:3, with the total adjacent M:BS area ratio of 1.6 6 0.1. In the husk, the BS cells were similar in size to those in the leaves but were surrounded by large rounded M cells (Table I; Fig. 1, B and D). Thus, from cross-sections of the husk, the M:BS cell ratio (0.5:1) is lower than in the leaves and the M:BS area ratio is much higher (5.6 6 0.3). Chloroplasts appear in both BS and M cells in the husk, with an obviously lower density than in the leaves. In the husk, there is also a higher density of chloroplasts toward the abaxial side (the side exposed to the atmosphere) and an Figure 1. Light microscopy images of B73 leaf (A) and husk (B) and SWC leaf (C and E) and husk (D and F). BS cells and the first (M1) and second (M2) layer of M cells from BS are labeled. Centrifugal positioning of chloroplasts in BS cells is shown at higher magnification (arrows in E and F). Bars = 100 mm for A, C, and D, 200 mm for B, and 50 mm for E and F. apparent higher density of chloroplasts around the veins than in the more distant M cells (Supplemental Fig. S1).
Anatomical measurements on the surfaces of the leaves and husks included measurements of vein density and stomatal number per area. Stomatal numbers were measured for both the upper (adaxial) and lower (abaxial) leaf surfaces and the inner (morphologically adaxial) and outer (abaxial) husk surfaces. The stomatal frequency on the abaxial side of the husks (mean of 98 mm 22 for the two varieties) is similar to that on the abaxial and adaxial surfaces of the leaves (mean of 94 mm 22 ; Table I). However, the husk had four to five times more stomata per mm 2 on the abaxial than on the adaxial epidermis in B73 and SWC. Vein density in the husk was four to five times lower than in the leaf in both SWC and B73 (Table I). Anatomical measurements taken from cross-sectional light microscope images of B73 included the average BS and M cell areas and VB area (including BS and vein tissue), the interveinal distance, and the surface area of M cells exposed to intercellular air space (S m ) and the BS surface area (S b ), both expressed per unit of leaf area as defined previously (Pengelly et al., 2010). Although no significant difference was observed between the husk and leaf BS cell area (Table I), 10-fold and 6-fold increases were observed in the husk M cell area and VB area, respectively, in comparison with the leaf ( Table I). The values of S m and S b were considerably less in the husk compared with the leaf (by 65%), indicating a potential for limitations to CO 2 diffusion.

In Situ Immunolocalization of Photosynthetic Enzymes
Immunolocalization of Rubisco, PEPC, and GDC in the leaf and husk was performed using the immunogold technique in combination with both confocal microscopy ( Fig. 2) and transmission electron microscopy (TEM; Fig. 3). Using confocal microscopy, Rubisco was clearly labeled in BS chloroplasts of SWC leaf ( Fig. 2A), while PEPC was found primarily in M cells (Fig. 2B). Occasional bright labeling was also observed in the epidermis due to the antibody binding nonspecifically to epidermal cell walls (Voznesenskaya et al., 2001). In the husk, immunolabeling of Rubisco Table I. Leaf and husk anatomical parameters of maize Values represent means 6 SE of five or more replicate observations. Asterisks indicate a significant difference between leaf and husk in each maize variety (P , 0.05). S b , BS surface area/leaf area as defined previously ; S m , M surface area exposed to intercellular air space/leaf area.

Chlorophyll Content and in Vitro Activity of PEPC and Rubisco
Chlorophyll content and the activity of Rubisco and PEPC were measured on extracts from the leaf and husk discs from B73 and SWC on which gas-exchange measurements had been made. The husk in both B73 and SWC contained much less chlorophyll than the leaf (Table II). PEPC activity expressed on an area basis was significantly reduced in the husk, approximately 10-fold and 5-fold less than that in the leaf in B73 and SWC, respectively (Table II). Rubisco activity (expressed on an area basis) in the husk was only half that in the leaf. The PEPC-Rubisco ratio was much lower in the husk compared with the leaf in both SWC and B73 (Table II).

Gas Exchange and Carbon Isotope Discrimination
The CO 2 assimilation rate (A) at ambient CO 2 partial pressure (pCO 2 ) in the husk was 15% of that in the leaf on a leaf area basis in B73 and 30% in SWC (Fig. 4, A and E; Table II). On a chlorophyll basis, rates of CO 2 fixation under high irradiance and pCO 2 did not differ greatly ( Fig. 4, B, D, and F; Table II). Dark respiration rates were lower in the husk than in the leaf for both B73 and SWC, whereas stomatal conductance did not differ significantly between the leaf and husk (Table II). CO 2 assimilation rate in response to increasing irradiance was substantially less in the husk than that of the leaf in B73 on a leaf area basis and saturated at lower irradiance than CO 2 assimilation rate in the leaf (Fig. 4C).
The response of CO 2 assimilation rate to increasing intercellular pCO 2 (C i ) in the leaves was that typically observed for C 4 species, exhibiting a low compensation point and saturation of CO 2 assimilation rates below ambient pCO 2 . In contrast, the CO 2 response of the husk exhibited a high CO 2 compensation point (G) and nonsaturating kinetics, even well above ambient pCO 2 (Fig. 4, A, B, E, and F). Although the G was higher in the husk compared with the leaf for both B73 and SWC, it did not increase with increasing oxygen partial pressure (pO 2 ), as is commonly observed for C 3 species (Fig. 5).
Carbon isotope discrimination (D, defined as D = R air / R p 2 1, where R p is the ratio 13 C/ 12 C in the photosynthetic product), measured in real time concurrently with gas exchange on SWC, was similar between the leaf and husk (Table II; Fig. 6B). However, the intercellular-to-ambient CO 2 ratio (C i /C a ) in the husk was double that in the leaf over a range of C i values (Fig.  6C). When plotted against C i /C a , D measurements for both the leaf and husk formed discrete clusters lying around a theoretical C 4 line estimating the relationship between D and C i /C a , using a leakiness (w) of 0.25 and assuming saturating amounts of carbonic anhydrase such that the reversible conversion of CO 2 and HCO 3 2 is at isotopic equilibrium (Cousins et al., 2006;Fig. 7). Carbon isotope discrimination measured from dry matter (d 13 C, relative to the standard V-Pee Dee Belemnite) was slightly more negative (depleted in 13 C) in the husk compared with the leaf (Table II).

Corn Husk Displays Unusual Gas-Exchange Characteristics
Two of the photosynthetic organ types found in the C 4 monocot maize, the standard foliar leaf (originating either from the stem or the terminal end of the husk) and the husk surrounding the ear, have been previ- Table II. Photosynthetic and biochemical properties of maize leaf and husk Measurements represent averages 6 SE of three or four replicate observations. Gas-exchange measurements were made at 25°C, an irradiance of 1,500 mmol quanta m 22 s 21 , and ambient CO 2 between 360 and 380 mbar. Asterisks indicate a significant difference between leaf and husk in each maize variety (P , 0.05). A, CO 2 assimilation rate. 0.76 6 0.04 0.12 6 0.01* 0.51 6 0.01 0.14 6 0.01* Chlorophyll a/b ratio 4.81 6 0.17 5.14 6 0.14 4.79 6 0. . A and B, CO 2 assimilation rate as a function of C i in maize (B73) leaf (black circles) and husk (white circles). Gas-exchange measurements were made in the glasshouse at 1,500 mmol quanta m 22 s 21 and a leaf temperature of 28°C. C and D, CO 2 assimilation rate as a function of irradiance in maize (B73) leaf and husk. Gasexchange measurements were made at ambient CO 2 between 360 and 380 mbar and a leaf temperature of 28°C. E and F, CO 2 assimilation rate as a function of C i in maize (SWC) leaf and husk. Gas-exchange measurements were made either in the laboratory or the glasshouse at 1,500 mmol quanta m 22 s 21 and a leaf temperature of 25°C.
Functional Analysis of Corn Husk Photosynthesis ously thought to manage carbon assimilation in different ways, with the suggestion that in the husk, photosynthesis may have C 3 -like characteristics (Langdale et al., 1988;Yakir et al., 1991;Hahnen et al., 2003). Maize leaf is known to be C 4 with archetypal Kranz anatomy, differentiation between cells with the complement of C 3 and C 4 enzymes, and a typical C 4 photosynthetic rate at ambient CO 2 . Our physiological measurements confirm typical CO 2 response curves for these leaves (Fig.  4). In contrast, the husk exhibits an unusual response of CO 2 assimilation rate to CO 2 that does not saturate in the measured range of pCO 2 . This is unlike the response of the leaves of C 4 species or responses commonly observed for leaves of C 3 species, which show marked change in the CO 2 response above ambient pCO 2 , where CO 2 assimilation rate becomes limited by the rate of ribulose 1,5-bisphosphate regeneration (von Caemmerer and Farquhar, 1981). Together with the saturation of CO 2 assimilation rate at low irradiance, this suggests that CO 2 assimilation rate in the husk could be severely limited by CO 2 diffusion from intercellular air space to the M cells. This conclusion is supported by the substantially lower values of M surface area exposed to intercellular air space (S m ) in husk compared with the leaf (see below). Insufficient carbonic anhydrase in M cytosol to facilitate the conversion of CO 2 to HCO 3 2 for PEP carboxylation could also contribute to the diffusion-limited phenotype (von Caemmerer et al., 2004).

Corn Husks Have High Compensation Points That Lack Oxygen Sensitivity
The oxygen dependence of G is usually an excellent indicator of the nature of the photosynthetic pathway of a plant and has frequently been used to distinguish between C 3 , C 3 -C 4 intermediate, and C 4 species (Holaday and Chollet, 1983;Ku et al., 1991;Furbank et al., 2009). The G is consistently higher in C 3 compared with C 4 species at ambient pO 2 and is linearly dependent on pO 2 due to increasing photorespiratory CO 2 release. In C 4 species, there is no discernible oxygen dependence of G (von Caemmerer, 2000). C 3 -C 4 intermediate species frequently display a nonlinear dependence of G on pO 2 , with low G values at low pO 2 but marked increases at higher pO 2 (Holaday and Chollet, 1983;Ku et al., 1991;von Caemmerer, 2000). Compensation points of the husk for both SWC and B73 were high at ambient pO 2 compared with that for the leaves, and this is most readily explained by the Figure 5. CO 2 compensation point (G) as a function of oxygen partial pressure in maize B73 leaf (black circles) and husk (white circles), maize SWC leaf (black squares) and husk (white squares), and rice (Oryza sativa; stars). Measurements were made in the laboratory at 1,500 mmol quanta m 22 s 21 and a leaf temperature of 25°C. Figure 6. Concurrent measurements of CO 2 assimilation rate (A) and carbon isotope discrimination (B) in leaf (black symbols) and husk (white symbols) of SWC as a function of C i . The ratio of intercellular to ambient CO 2 (C i /C a ) is also shown (C). Different symbols denote replicate leaf and husk samples. Measurements were made at 1,500 mmol quanta m 22 s 21 and a leaf temperature of 28°C.
high ratios of respiration rate to CO 2 assimilation rate ( Table II). The fact that, like leaves, the husk did not display an oxygen dependence of G suggests that the husk primarily fixes CO 2 via the C 4 photosynthetic pathway and that the high level of Rubisco evident from immunolabeling in distant M cells plays a minor role in husk photosynthesis. However, if C 3 photosynthesis was operational in those cells at a low level (less than 10%), it would generate only a small rise in G with pO 2 , which would be difficult to detect experimentally (von Caemmerer, 2000).

Carbon Isotope Discrimination of Maize Husk Photosynthesis Is C 4 Like
Photosynthetic carbon isotope discrimination (D) is determined by fractionation occurring during CO 2 diffusion from the atmosphere to the site of CO 2 fixation and the discrimination factors associated with the carboxylation steps. Carbon isotope discrimination is much larger in C 3 compared with C 4 species because of Rubisco's preference for 12 CO 2 . In C 4 species, both the lower discrimination of the PEP carboxylation step and the fact that Rubisco's ability to discriminate against 13 CO 2 is reduced by being compartmentalized in the gas-tight BS mean that D is much less than in C 3 species (Farquhar et al., 1989). Carbon isotope discrimination measurements made on corn dry matter were consistent with measurements reported by Yakir et al. (1991) indicating a slight difference in discrimination between the leaf and husk. Dry matter carbon isotope discrimination values are confounded by the fact that the source of the carbon is not necessarily derived from local photosynthesis, and postphotosynthetic fractionation can also affect the values (Yakir et al., 1991;Henderson et al., 1992;Badeck et al., 2005). Therefore, we made D measurements in real time concurrently with measurements of CO 2 assimilation rates. Our measurements showed that both the leaf and husk displayed C 4 -like carbon isotope discrimination similar to D measurements made previously for C 4 monocot species, 3‰ to 4‰ (Henderson et al., 1992;von Caemmerer et al., 2007), and in contrast to typical C 3 species measurements of approximately 20‰ (Evans et al., 1986;von Caemmerer and Evans, 1991).
Equations predicting carbon isotope discrimination during C 3 or C 4 photosynthesis (Farquhar, 1983;Cousins et al., 2006;Tazoe et al., 2009) can be used to infer various parameters from the average D and C i /C a measured at ambient CO 2 (Table II). In C 3 plants, D increases with an increase in the C i /C a ratio, while in C 4 plants, the response depends on other factors. This includes the degree of leakiness (w, the ratio of the rate of CO 2 leakage from the BS to the rate of CO 2 supply to this compartment). For example, with increasing C i / C a , the modeled response for C 4 plants shows that D values decrease with increasing C i /C a at the low w values that are typical for C 4 plants. In this case, the husk (at high C i /C a ) could have a similar D value to the leaf (at the lower C i /C a ) by having greater leakage in the C 4 system, by an increase in the ratio of PEPC carboxylation to CO 2 hydrations due to a lack of carbonic anhydrase, or by the occurrence of 5% to 10% of C 3 photosynthesis depending on assumptions made for CO 2 diffusion conductance in the M. The results with isotope discrimination do not support a significant contribution by C 3 photosynthesis in the husk, although it is not possible to come up with an estimate of any C 3 contribution due to the above uncertainties. However, we made measurements at several ambient CO 2 concentrations that we hypothesized might bring about different ratios of C 3 to C 4 photosynthesis. Since we observed no discernible difference in measured D values (Fig. 6), this is consistent with little contribution from C 3 photosynthesis in the M cells compared with C 4 photosynthesis in the husk.

Maize Husk Displays Kranz-Like Anatomy at Low Vein Density
Kranz anatomical features were present in all maize leaves (distinctive and classical NADP-ME), while in the husk, the anatomy was altered, having a weak Kranz-like structure around the veins and many interveinal M cells. Maize leaves (Fig. 1, A and C) exhibited large BS cells and adjacent, more or less radially arranged, palisade-shaped M cells. M cells adjacent to BS cells (M1) in the husk do not differ in shape from the distant M cells but were smaller. The M-BS ratio in the husk, which is 3.5 times higher than in the leaves, lies between the numbers characteristic of C 4 NADP-ME and C 3 species (Hattersley, 1984). Light microscopy of the husk showed that chloroplasts are more concentrated in chlorenchyma cells around the VB than in the   Fig. S1).
Most notably, the husk has a significantly lower vein density than the leaves, observed previously by Langdale et al. (1988) and others, which was quantified in this study as 4.5-fold higher than the husk. This distance was bridged by M cells roughly 10-fold bigger than those found between VB in leaves. In C 4 species having Kranz anatomy around individual veins, a small physical distance between veins is often thought of as a necessity for C 4 function, due to the need for C 4 acids to move from M to BS cells. Analysis of 119 grasses (Hattersley and Watson, 1975) suggested that in C 4 monocot species, no M cell is separated from the nearest BS cell by more than one other M cell (the maximum cell distant count). Yet, recent work has shown that in the C 4 dicot Flaveria bidentis, this variable is surprisingly plastic (Araus et al., 1991;Sage and McKown, 2006;Pengelly et al., 2010). The M conductance to CO 2 diffusion from the intercellular air space to the M cytosol has been suggested in C 4 plants to correlate with the M surface area exposed to intercellular air space (S m ). The fact that S m is significantly less in the husk compared with the leaf suggests a low conductance to CO 2 diffusion from intercellular air space to M cells. Since the measurements of S m included distal M cells that lack PEPC (Table II), the actual conductance is likely to be even less, supporting the conclusions reached from gas-exchange measurements that in the husk, photosynthesis is diffusion limited. In C 4 plants, the BS conductance to CO 2 diffusion can be estimated in part by measurement of the BS surface area per unit of leaf area (S b ;von Caemmerer and Furbank, 2003;Pengelly et al., 2010). Our 3-fold lower estimates of S b in the husk compared with the leaves suggest lower BS conductance in the husk compared with the leaf.

Corn Husk Displays Unusual Expression Patterns of Rubisco, PEPC, and GDC
The in vitro activity of Rubisco and PEPC on a leaf area basis was substantially less in the husk compared with the leaf, in accordance with its low chlorophyll content and low photosynthetic capacity. However, the ratio of PEPC to Rubisco activity was also substantially less, demonstrating that the difference between the husk and leaf is not just one of scale. Immunolocalization showed in the leaf and husk PEPC selectively expressed in the cytosol of M cells around the VB but not in BS or distal M cells, similar to the gene expression pattern of PEPC observed by Hahnen et al. (2003). The labeling of Rubisco in the leaf was typical of a standard C 4 expression pattern shown previously (Langdale et al., 1987), with Rubisco selectively expressed in BS chloroplasts. In the husk, there was not only significant labeling for Rubisco in BS chloroplasts but also in the M cells adjacent to the BS cells and the second M cell layer (Table II; Figs. 2 and 3). Detection of the protein here in distant M cells confirms previous reports based on the localization of Rubisco transcripts (Langdale et al., 1988;Hahnen et al., 2003). The presence of Rubisco in M cells was accompanied by the presence of GDC, suggesting that these M cells could perhaps operate C 3 photosynthesis. However, the lack of an oxygen dependence of G and the low carbon isotope discrimination argue against substantial CO 2 fixation occurring via the C 3 photosynthetic pathway in these cells. The expression of Rubisco and the lack of PEPC expression in these distant husk M cells are puzzling, but there are examples of similar phenotypes in C 4 -like species such as Flaveria brownii (Bauwe, 1984;Reed and Chollet, 1985;Cheng et al., 1988). F. brownii had originally been considered a typical C 4 species based on anatomical, physiological, and biochemical criteria. It exhibits low G values that lack oxygen sensitivity and has C 4 -like carbon isotope values (Holaday et al., 1984;Monson et al., 1987Monson et al., , 1988. However, immunolocalization studies have shown the presence of Rubisco in M as well as BS cells (Bauwe, 1984;Reed and Chollet, 1985). Furthermore, the activity of Rubisco and of a number of other Calvin cycle enzymes was detected in isolated M protoplasts (Cheng et al., 1988). It is not clear whether a complete complement of Calvin cycle enzymes is expressed in distal M cells in maize husk, and this deserves further investigation.

CONCLUSION
Our observations and measurements have shown that in maize, the outer husk surrounding the ear operates a C 4 -like photosynthetic pathway. Both carbon isotope analysis and the lack of oxygen sensitivity in the compensation point show that the majority of Rubisco actively participating in CO 2 assimilation is that compartmentalized within the chloroplasts of BS cells, which is supported by the C 4 cycle via PEPC in the adjacent M. In C 4 species having Kranz anatomy around individual veins, high vein density is seen as a requirement for active C 4 photosynthesis. Maize husk operates C 4 photosynthesis, but as it is dispersed around widely spaced veins, photosynthesis is diffusion limited in part due to low M surface area exposed to intercellular air space. The role of Rubisco in distant M cells in husk photosynthesis remains unresolved. We suggest that transcriptome analysis in distal M cells may help establish whether all enzymes required for C 3 photosynthesis are expressed in this cell type. Understanding husk photosynthesis in maize, a species where the genome has been sequenced, may provide another route to understanding the path from C 3 to C 4 photosynthesis. Our study has also highlighted the need for a multidisciplinary approach in correctly labeling the photosynthetic pathway of a plant, as neither physiological, biochemical, nor anatomical measurements alone are enough.

Growth Conditions and Sampling of Maize Photosynthetic Organs
Two varieties of corn/maize (Zea mays), B73 and Sweet Corn, Kelvedon Glory, F1 (SWC), were grown during the summer months in a glasshouse under natural light conditions (28°C day and 18°C night temperatures). Plants were grown in 30-L pots in a garden soil mix with fertilizer (Osmocote; Scotts Australia) and watered daily. Experiments and sampling were done on both the husk and leaf tissues. Here, "husk" refers to the outer layer of thick ribbed sheath surrounding the ear and "leaf" refers to a standard leaf originating at the stem in B73 and a leaf protruding from the husk in SWC.

Gas-Exchange Measurements
Gas-exchange measurements were made with a 6-cm 2 leaf chamber of the LI-6400 with a red-blue light-emitting diode light source (Li-Cor) either in the glasshouse or the laboratory.
In the glasshouse, leaves were first equilibrated for 30 min at an ambient CO 2 of 400 mmol mol 21 , 1,500 mmol quanta m 22 s 21 irradiance, and a leaf temperature of 28°C before measurements were taken. To measure the CO 2 response, CO 2 concentrations were changed every 2 min, with values of 0, 30, 50, 75, 100, 150, 200, 300, 400, 500, 600, 800, and 1,200 mmol mol 21 in the reference cell of the LI-6400. Following this, leaves were acclimated at 400 mmol mol 21 for 30 min, and then irradiance was reduced in steps at 2-min intervals.
In the laboratory, measurements of CO 2 response were made as described above but at 25°C. This was followed by measurements of CO 2 compensation point at different oxygen levels. For this, air entering the LI-6400 was prepared by using two mass-flow controllers (MKS Instruments) to mix different concentrations of N 2 and oxygen. Compensation points were calculated from measurement of CO 2 response curves at low CO 2 concentrations. The average atmospheric pressure in Canberra, Australia, is 950 mbar.

Measurements of Carbon Isotope Discrimination and Gas Exchange
Dry matter carbon isotope discrimination measurements were made as reported previously (Pengelly et al., 2010). For online measurements, two LI-6400 systems were coupled to a tunable diode laser (model TGA100; Campbell Scientific) for concurrent measurements of carbon isotope discrimination and gas exchange (Bowling et al., 2003;Griffis et al., 2004;Tazoe et al., 2011). Input gases (N 2 and oxygen) were mixed using mass-flow controllers (Omega Engineering), and measurements were first made at 2% oxygen and at reference CO 2 of 400 mmol mol 21 , 1,500 mmol quanta m 22 s 21 , and leaf temperature of 25°C for 30 min. Then, the reference CO 2 was changed stepwise to 1,000, 700, 500, 400, 300, 200, and 100 mmol mol 21 and measurements were made for 30 min at each CO 2 concentration.
Following gas-exchange measurements, 0.5-cm 2 discs were removed from tested leaves and husks, snap frozen in liquid nitrogen, and stored at 280°C for later measurement of photosynthetic enzyme activities and chlorophyll content.

Measurement of Rubisco and PEPC Activity and Chlorophyll Content
Activities of the photosynthetic enzymes Rubisco and PEPC were measured as described previously Pengelly et al., 2010). Chlorophyll was extracted from frozen leaf or husk discs in a Tissuelyser II ball mill (Retsch) with 80% acetone. Chlorophyll a and b contents were spectrophotometrically quantified (Porra et al., 1989).

Determination of Stomatal Numbers
Stomatal numbers were measured from the same or similar leaf and husk as used for gas-exchange measurements from silicone rubber impressions taken from both sides of the leaves or husks (von Caemmerer et al., 2004). Stomata and epidermal cells were counted from positives made from the impressions with nail polish, in 10 different fields of view per leaf, with a compound microscope using a magnification of 200-fold. Three leaves per plant were measured from three individual plants. Digital photographs of each field were taken and cells counted with the publicly available ImageJ software (http://rsb.info.nih.gov/ij/).

Leaf Vein Density Analysis
Vein density was measured on leaf and husk sections taken from three individual plants. Leaves were initially cleared by immersion in a 95% ethanol, 5% NaOH solution for 4 d and rehydrated in water for 1 h. Sections were cut from areas used for gas-exchange measurements avoiding major veins. Digital images were taken at 503 magnification. Vein density was determined from each image by measuring total length of veins within a 2-cm 2 quadrant using ImageJ quantification software.

Embedding of Leaf and Husk Sections for Light and Electron Microscopy
Leaf sections measuring approximately 2 mm 3 5 mm were taken from fully developed husk and leaves of three B73 and three SWC maize plants. Fresh sections were fixed in buffer containing 1.25% glutaraldehyde and 2% formaldehyde in 50 mM PIPES buffer, pH 7.2, under a vacuum for 4 h. Fixed sections were dehydrated in an alcohol dilution series and then embedded in either Araldite resin (Electron Microscopy Sciences) for light microscopy or London Resin White acrylic resin (Electron Microscopy Sciences) for use in immunolocalization.

Anatomical Measurements by Light Microscopy
Semithin sections of 0.5 mm thickness were cut from embedded husk and leaf pieces using glass knives on a Reichert ultramicrotome (Reichert Technologies), stained with toluidine blue, and heat fixed to glass slides. Slides were viewed using a Zeiss Axioskop light microscope (Carl Zeiss) at 4003 magnification. Images from each cross-section were analyzed using ImageJ software for a variety of anatomical measurements, including BS cell area, M cell area, VB area, interveinal distance, the M surface area exposed to intercellular air space/leaf area (S m ), and the BS surface area/leaf area (S b ), as described previously (Pengelly et al., 2010). Both leaf and husk S m were calculated using an approximate curvature correction factor of 1.43 (Evans et al., 1994;von Caemmerer et al., 2007).

In Situ Immunolocalization
Antibodies used (all raised in rabbit) were anti-Spinacea oleracea Rubisco (large subunit specific) IgG (courtesy of Dr. Bruce McFadden), commercially available anti-maize PEPC IgG (Chemicon), and anti-Pisum sativum mitochondrial GDC IgG (courtesy of Dr. David Oliver). Preimmune serum was used in all cases as a control.
Leaf cross-sections (0.8-1 mm thick) were dried from a drop of water onto gelatin-coated slides and blocked for 1 h with Tris-buffered saline plus Tween 20 + bovine serum albumin (TBST + BSA; 10 mM Tris-HCl, 150 mM NaCl, 0.1% [v/v] Tween 20, and 1% [w/v] BSA, pH 7.2). They were then incubated for 3 h with preimmune serum diluted in TBST + BSA (1:100 dilution), anti-Rubisco (1:50), or anti-PEPC (1:50). The slides were washed with TBST + BSA and then treated for 1 h with 5-nm Protein A-gold (diluted 1:100 with TBST + BSA). After washing, the sections were exposed to a silver enhancement reagent for 20 min according to the manufacturer's directions (Amersham), stained with 0.5% (w/v) Safranin O, and imaged in a reflected/transmitted mode using a Zeiss Confocal LSM 510 Meta Laser Scanning Microscope (Carl Zeiss). The background labeling with preimmune serum was very low, although occasional labeling occurred in areas where the sections were wrinkled due to trapping of antibodies/label (data not shown).
For TEM immunolabeling, thin sections (approximately 70-90 nm) on Formvar-coated nickel grids were incubated for 1 h in TBST + BSA to block nonspecific protein binding on the sections. They were then incubated for 3 h with the preimmune serum diluted in TBST + BSA, anti-Rubisco (1:50), anti-PEPC (1:50), or anti-GDC (1:10) antibodies. After washing with TBST + BSA, the sections were incubated for 1 h with Protein A-gold (5 and 15 nm) diluted 1:100 with TBST + BSA. The sections were washed sequentially with TBST + BSA, TBST, and distilled water. The sections with 5-nm gold labeling were treated with the gold enhancement kit GoldEnchance (Nanoprobes) to obtain higher density of labels and better contrast for observation under low magnification and then poststained with a 1:3 dilution of 0.5% (w/v) potassium permanganate and 2% (w/v) uranyl acetate. The sections with 15-nm gold labeling were poststained with the same solution and used for statistical analysis of label density. Images were collected using a Philips CM200 UT transmission electron microscope (FEI Co.). The density of labeling was determined by counting the gold particles on electron micrographs and calculating the number of particles per unit of area (mm 2 ) in mitochondria (GDC), chloroplasts (Rubisco), or cytosol (PEPC) and nonspecific background labeling. For each cell type, replicate measurements were made on parts of cell sections (n = 10-12). Cell area was measured using ImageJ software.

Statistical Analysis
The relationship between mean values of photosynthetic, anatomical, and biochemical data obtained throughout this study was tested using Student's t test (P , 0.05). Significant differences are marked with asterisks.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Volume fraction of chloroplasts per cell in husk versus leaves in maize (SWC).