Intracellular position of mitochondria and chloroplasts in bundle sheath and mesophyll cells of C3 grasses in relation to photorespiratory CO2 loss

Abstract In C3 plants, photosynthetic efficiency is reduced by photorespiration. A part of CO2 fixed during photosynthesis in chloroplasts is lost from mitochondria during photorespiration by decarboxylation of glycine by glycine decarboxylase (GDC). Thus, the intracellular position of mitochondria in photosynthetic cells is critical to the rate of photorespiratory CO2 loss. We investigated the intracellular position of mitochondria in parenchyma sheath (PS) and mesophyll cells of 10 C3 grasses from 3 subfamilies (Ehrhartoideae, Panicoideae, and Pooideae) by immunostaining for GDC and light and electron microscopic observation. Immunostaining suggested that many mitochondria were located in the inner half of PS cells and on the vacuole side of chloroplasts in mesophyll cells. Organelle quantification showed that 62–75% of PS mitochondria were located in the inner half of cells, and 62–78% of PS chloroplasts were in the outer half. In mesophyll cells, 61–92% of mitochondria were positioned on the vacuole side of chloroplasts and stromules. In PS cells, such location would reduce the loss of photorespiratory CO2 by lengthening the path of CO2 diffusion and allow more efficient fixation of CO2 from intercellular spaces. In mesophyll cells, it would facilitate scavenging by chloroplasts of photorespiratory CO2 released from mitochondria. Our data suggest that the PS cells of C3 grasses have already acquired an initial structure leading to proto-Kranz and further C3–C4 intermediate anatomy. We also found that in the Pooideae, organelle positioning in PS cells on the phloem side resembles that in mesophyll cells.


Introduction
Photorespiration reduces photosynthetic efficiency of C 3 plants, particularly at high temperature with low CO 2 concentration. In rice, a staple crop worldwide, reduction of photosynthesis by photorespiration is estimated to be 25-35% at 30-35 °C at current CO 2 levels (Sage, 2001). Therefore, the suppression of photorespiration is an important target to improve crop productivity. The first reaction of photorespiration is oxidation of ribulose 1,5-bisphosphate by ribulose 1,5-bisphosphate carboxylase/oxygenase (rubisco); phosphoglycolate generated by this reaction is metabolized in the glycolate pathway, which requires cooperation between chloroplasts, peroxisomes, and mitochondria. Glycine generated in peroxisomes is transported to mitochondria, where it is decarboxylated by glycine decarboxylase (GDC) (Bauwe, 2011;Douce et al., 2001).
In higher plants, the C 4 photosynthetic pathway is the most efficient mechanism that reduces photorespiration; C 4 plants are thought to have evolved from C 3 plants primarily in response to a decline in atmospheric CO 2 concentrations that began during the Cretaceous and continued until the Miocene (Ehleringer et al., 1991). The leaves of most C 4 plants show Kranz type leaf anatomy characterized by two types of photosynthetic cells -mesophyll and bundle sheath (BS) cells -and photorespiration is repressed by concentrating CO 2 around rubisco in chloroplasts of BS cells (Hatch, 1987). As a result, C 4 plants retain high photosynthetic capacity even in environments that cause high photorespiration (Ehleringer et al., 1991).
There are plants having photosynthetic characteristics intermediate between those of C 3 and C 4 plants, which are called C 3 -C 4 intermediate plants (Edwards & Ku, 1987). Their leaf anatomy and CO 2 exchange characteristics such as CO 2 compensation points and O 2 inhibition of photosynthesis are in between those of C 3 and C 4 plants. Low rates of photorespiration in C 3 -C 4 intermediates result mainly from the intercellular localization of GDC (Monson & Rawsthorne, 2000;Rawsthorne, 1992). C 3 -C 4 intermediate plants are considered to be an intermediate stage in Phylogeny Working Group, 2001;Edwards & Smith, 2010). Our results reveal that positioning of the organelles that leads to proto-Kranz anatomy is already present in PS cells of these C 3 grasses. In addition, we confirmed that many mitochondria in mesophyll cells are positioned on the vacuole side of chloroplasts and suggest that the mechanism of scavenging photorespiratory CO 2 proposed for rice and wheat (Sage & Sage, 2009) also exists in other C 3 grasses.

Plant materials and growth
We investigated 10 C 3 species of 8 genera from 3 grass subfamilies: Oryza alta Swallen (W0017)

Visualization of mitochondria by immunostaining of the P-protein of GDC
The photorespiratory enzyme GDC is a major protein in the mitochondria of photosynthetic cells of C 3 plants (Douce et al., 2001). Therefore, we visualized the mitochondria in mesophyll and BS cells of C 3 grass leaves by the evolution of C 4 plants from C 3 ancestors (Edwards & Ku, 1987;Monson & Rawsthorne, 2000;Sage et al., 2014). The repositioning and enhancement of mitochondria and chloroplasts in BS cells and the BS-specific expression of GDC are important for the establishment of C 3 -C 4 intermediate photosynthesis. However, the C 3 cycle operates in both M and BS cells of C 3 -C 4 intermediates (Edwards & Ku, 1987;Monson & Rawsthorne, 2000).
Recently, the early stage of evolution of C 3 -C 4 intermediate plants from C 3 plants was termed proto-Kranz (Muhaidat et al., 2011;Sage et al., 2013Sage et al., , 2014. Leaf anatomy of proto-Kranz plants resembles that of C 3 -C 4 intermediates, but GDC accumulates in both mesophyll and BS mitochondria. Proto-Kranz plants still perform C 3 photosynthesis, but many mitochondria in their BS cells are located along the inner tangential walls adjacent to the vascular bundle. Some chloroplasts cover these mitochondria and therefore may capture some of photorespiratory CO 2 released from BS mitochondria. Proto-Kranz species have been found in some genera such as Panicum and Steinchisma in the Poaceae (Brown et al., 1983;Sage et al., 2014), Flaveria in the Asteraceae , Heliotropium in the Boraginaceae (Muhaidat et al., 2011;Vogan et al., 2007), and Salsola in the Chenopodiaceae (Voznesenskaya et al., 2013). However, it is unclear whether proto-Kranz species are common in other groups. It is also unknown whether organelles are scattered randomly or are positioned according to some rules in the BS cells of C 3 plants. This information is important for understanding early steps of evolution of photosynthetic types.
Some C 3 plants restrict photorespiratory CO 2 efflux without alterations in their biochemistry Moser et al., 2015;Sage & Sage, 2009). Sage and Sage (2009) observed that in rice mesophyll cells, much of the cell periphery is occupied by chloroplasts and their protrusions (stromules) filled with stroma containing rubisco (Bourett et al., 1999;Gray et al., 2001). Stromules may efficiently refix some CO 2 released from mitochondria, and it is important for the mitochondria in mesophyll cells to be located on the vacuole side of the chloroplasts and stromules. However, information about the position of mitochondria in mesophyll cells of C 3 grasses is scarce.
In this study, we investigated intracellular positioning of mitochondria and chloroplasts in parenchyma sheath (PS) and mesophyll cells of 10 C 3 species from 3 subfamilies of the grass family (Ehrhartoideae, Panicoideae, and Pooideae) using in situ immunostaining of GDC and light and electron microscopic observation. The Ehrhartoideae (formerly Oryzoideae) is a C 3 subfamily occurring mainly in tropical and subtropical regions, whereas the Pooideae is a C 3 subfamily thriving mainly in cool temperate regions. The Panicoideae includes both C 3 and C 4 species and is distributed in tropical and subtropical regions (Grass immunohistochemical staining of the P-protein of GDC and observed them under a light microscope. Fully expanded uppermost leaves were sampled from three plants of each species at 09:00 on a sunny day and fixed within 30 min. Segments from the middle of the leaves were fixed with 3% (w/v) paraformaldehyde / 0.2% (v/v) glutaraldehyde in 50 mM sodium phosphate (pH 6.8) for 10 h at 4 °C and then washed in phosphate buffer (pH 6.8). Fixed tissues were dehydrated through an ethanol/butanol series, infiltrated with 100% butanol overnight at room temperature, and incubated in a 1:1 mixture of butanol : Paraplast (Paraplast X-TRA, Fisher Scientific Co., Houston, TX, USA) for 1 day at 56 °C. Samples were then infiltrated with pure Paraplast for 8 h at 56 °C and in fresh Paraplast overnight at 56 °C. Finally, samples were embedded in fresh Paraplast. Transverse sections (8 μm thick) were cut on a microtome (Yamato Koki Co.,Ltd., Saitama, Japan), mounted on slides coated with poly-l-lysine (Sigma-Aldrich, Inc., St. Louis, MO, USA) and dried overnight at 46 °C.
Slides were immersed in xylene to remove the Paraplast and the sections on the slides were rehydrated through an ethanol series. The sections were incubated with 0.1% (w/v) trypsin for 30 min and then with 5% (v/v) hydrogen peroxide for 1 h, rinsed with distilled water and 10 mM phosphate buffer (pH 7.2) containing 150 mM naCl (PBS), and incubated with 5% (w/v) bovine serum albumin (BSA) in PBS for 10 min. The sections were then incubated with antiserum against the P-protein of GDC of pea leaf mitochondria diluted 1:2000 with PBS containing 1% (w/v) BSA for 1 h at room temperature. The antiserum was provided by Prof. D. J. Oliver (University of Idaho, Moscow, ID, USA); the same antiserum was used in our previous studies of grasses (Yoshimura et al., 2004) and crucifers (Ueno, 2011;Ueno et al., 2003). The sections were rinsed with distilled water and PBS, incubated with a secondary antibody (anti-rabbit goat antibody conjugated with horseradish peroxidase; American Qualex, San Clemente, CA, USA) diluted 1:100 with PBS containing 1% (w/v) BSA for 30 min, and rinsed with distilled water and PBS. Finally, the P-protein in the sections was visualized by use of Peroxidase-Stain-DAB (3,3-diaminobenzidine tetrahydrochloride) Kit (nacalai Tesque, Inc., Kyoto, Japan). The sections were observed under a light microscope.

Leaf anatomy and ultrastructure
Segments from the middle of fully expanded uppermost leaves collected from three plants of each species on a sunny day were fixed in 3% (v/v) glutaraldehyde in 50 mM sodium phosphate buffer (pH 6.8) at room temperature for 2 h, washed with phosphate buffer, postfixed with 2% (v/v) OsO 4 in 50 mM phosphate (pH 6.8) buffer for 2 h, dehydrated through an acetone series, and embedded in Spurr's resin. For light microscopy, semi-thin sections were cut with a glass knife and stained with 1% (w/v) toluidine blue O. For electron microscopy, ultra-thin sections were cut with a diamond knife and stained with lead citrate. They were observed under an electron microscope (JEM-100CX II K, JEOL Ltd., Tokyo, Japan) at 75 kV.

Quantification of mitochondria and chloroplasts
We recorded the numbers and intracellular positions of mitochondria and chloroplasts in 28 to 40 mesophyll cells and PS cells (outer BS cells surrounding the vascular bundle) of small vascular bundles (Chonan et al., 1974) selected from leaves of three plants under the electron microscope. In PS cells, we counted the numbers of organelles positioned in the inner half of cells (i.e. in the inner tangential walls and the inner half of the radial walls of cells) and in the outer half of cells (i.e. in the outer tangential walls and the outer half of the radial walls of cells) ( Figure 6). In general, PS cells are round in transverse section. However, those of irregular shape were excluded from measurement. In mesophyll cells, we counted the mitochondria on the vacuole side of chloroplasts (inner position) and on the cell wall side of chloroplasts including isolated mitochondria not associated with chloroplasts but adjacent to the cell wall (outer position; Figure 6); equal numbers of mesophyll cells were selected from abaxial and adaxial sides. Simultaneously, we counted the chloroplasts in mesophyll cells. We also evaluated the position of mitochondria in L-type PS cells (Williams et al., 1989 and see Results for details) of the Pooideae as for mesophyll cells.

Statistical analysis
Student's t-test was used to test the significance (p < 0.05) of the differences between the proportions of mitochondria or chloroplasts located in the centripetal and centrifugal positions in PS cells and the differences between the proportions of mitochondria located in the inner and outer positions in mesophyll cells and L-type PS cells.

Light microscopic observation of leaf structure
In leaves of C 3 grasses, the vascular bundle is generally surrounded by two layers of BS: outer PS and inner mestome sheath (MS) (Brown, 1975). In small vascular bundles of rice, however, the MS is not obvious (Kaneko et al., 1980). We found similar features in the C 3 species of the Ehrhartoideae and Panicoideae (Figure 1(A)-(D)). In the C 3 species of the Pooideae, vascular bundles were surrounded by both PS and MS (Figure 1(E)-(H)). The mesophyll cells them from 'ordinary' PS cells (according to the terminology of Williams et al., 1989). Ordinary PS cells contained fewer chloroplasts than mesophyll cells, whereas L-type PS cells had similar numbers of chloroplasts.

Light microscopic observation of mitochondria by GDC immunostaining
In all grasses examined, immunohistochemical staining of the P-protein of GDC revealed the mitochondria in PS and mesophyll cells as reddish brown granules under a light microscope (Figure 2). However, some mesophyll cells of the Ehrhartoideae showed almost no staining of Oryza species had many lobes, which is a characteristic of armed cells (Chonan, 1970;Sage & Sage, 2009). In the mesophyll of Oryza, cells were densely arranged and intercellular spaces were small (Figure 1(A)). Intercellular spaces were larger in the Panicoideae and Pooideae than in the Ehrhartoideae (Figure 1(C), (E) and (G)). The PS cells of the Ehrhartoideae contained few chloroplasts ( Figure  1(B)), whereas those of the Panicoideae and Pooideae contained many (Figure 1(D), (F) and (H)). In the Pooideae, the PS cells on the phloem side of the bundle differed structurally from those on the xylem side and were similar in their appearance to mesophyll cells (Figure 1(E)-(H)). We call the PS cells on the phloem side 'L-type' PS cells and distinguish In the mesophyll cells of all C 3 grasses examined, many mitochondria were positioned on the vacuole side of chloroplasts (Figure 2(B), (D), (F) and (H)). In the L-type PS cells of the Pooideae, the position of mitochondria was similar to that in mesophyll cells, and no centripetal positioning of mitochondria were found, unlike in ordinary PS cells (Figure 2(E) and (G)).

Ultrastructural observation and quantification of mitochondria and chloroplasts
In the PS cells of all C 3 grasses except A. fatua, a significantly higher proportion of mitochondria (64-75%) was found in (Figure 2(B)), probably because the cell walls of mesophyll cells in Oryza species are thicker than those of other grasses, which may have impeded the diffusion of the antibody, color development solution, or both into mesophyll cells.
In the PS cells, the number of mitochondria differed between species. The PS cells of the Panicoideae and ordinary PS cells of the Pooideae contained relatively many mitochondria, which were positioned along the inner tangential cell walls, forming a band around the vascular bundle (Figure 2(C), (E) and (G). In the Ehrhartoideae, PS cells contained only few mitochondria, which were scattered within the cell (Figure 2(A)). mitochondria located in the inner position was higher in the former (92%) than in the latter (68%). This was because the mesophyll chloroplasts of O. undulatifolius (Suppl. data 2B) have more-developed stromules than those of P. bisulcatum (Figure 4(C) and (D)).
The intracellular positioning of mitochondria in L-type PS cells of the Pooideae was evaluated as for mesophyll cells, because these cell types are structurally similar ( Figure 5). In the L-type PS cells of all pooid grasses except for B. rigidus, 68-88% of mitochondria were positioned on the vacuole side of chloroplasts ( Figure 5; Table 3). In B. rigidus, the proportion of mitochondria in the outer and inner positions was almost the same (Table 3). L-type PS cells contained fewer mitochondria and chloroplasts per cell (Table 3) than mesophyll cells (Table 2).

Discussion
This study investigated the intracellular positions of mitochondria and chloroplasts in the PS and mesophyll cells of leaves of 10 C 3 grass species from 3 different subfamilies. These organelles showed unique positioning related to photosynthetic and photorespiratory functions.

Intracellular position of mitochondria and chloroplasts in PS cells
In the PS cells (except the L-type) of C 3 grasses, 62-75% of mitochondria and 22-38% of chloroplasts were located in the inner half of cells (Table 1). These values were lower than those reported for the proto-Kranz species of Heliotropium (Muhaidat et al., 2011) and Flaveria , which are considered to represent an early stage of evolution of C 3 -C 4 intermediates from C 3 plants. Our data show that the organelle positioning found in the PS cells is common to C 3 grasses. The evolutionary origin of the proto-Kranz type would result from the structural the centripetal than in the centrifugal position (Table 1). In A. fatua, 62% of mitochondria were found in the centripetal position, although the difference was not significant (Table 1). In contrast, a significantly higher proportion of chloroplasts (63-78%) was found in the centrifugal than in the centripetal position in the PS cells of all C 3 grasses except O. alta and B. catharticus (Table 1). In the latter two species, 62-69% of chloroplasts were in the centrifugal position, but the difference was not significant. Centrifugally positioned mitochondria generally occurred near chloroplasts (Figure 3(C), (E), and (G)), whereas some centripetal mitochondria were not associated with chloroplasts (Figure 3(B), (D), (F), and (H)). In P. bisulcatum, B. catharticus, and T. aestivum, mitochondria covered by chloroplasts were occasionally observed along the inner tangential walls of PS cells (Suppl. data 1). In B. catharticus and T. aestivum, mitochondria covered by chloroplasts occurred only in the vicinity of plasmodesmata (Suppl. data 1B, C).
In the mesophyll cells of all grasses except for B. rigidus, 68-92% of mitochondria were positioned on the vacuole side of chloroplasts ( Figure 4, Table 2). (In B. rigidus, this value was 61%, although the difference was not significant; Table 2). Mitochondria located close to cell walls and not associated with chloroplasts were rare, although some mitochondria were located in the gaps between chloroplasts (Figure 4(D)); these mitochondria were counted as located in the outer position. In B. rigidus, the gaps between chloroplasts were large because chloroplasts were sparse, and thus many mitochondria were located in the gaps, being adjacent to the cell wall (Suppl. data 2A). Stromules were observed in the mesophyll cells of all C 3 grasses (Figure 4(B), (F) and (H)), but rarely in the PS cells (Figure 3(H)). In the mesophyll cells of the three Oryza species, stromules were well developed (Figure 4(B)). When O. undulatifolius (Panicoideae) was compared with P. bisulcatum belonging to the same subfamily, the proportion of Table 1. intracellular position and number of mitochondria and chloroplasts in parenchyma sheath cells of c 3 grasses.
Notes. Values are means ± Sd of three plants. See Figure 6 for the position of organelles in the outer and inner half of cell. For the Pooideae, data for ordinary parenchyma sheath cells are shown. * indicate significant differences between the two positions of organelles (p < 0.05). 29.3 ± 8.1 70.7 ± 8.1* 6.7 ± 0.8 62.9 ± 3.5 37.1 ± 3.5* 4.3 ± 0.5 mitochondria (Leegood, 2008;Yoshimura et al., 2004). Yet the PS cells of C 3 grasses appear to position organelles to fix CO 2 most effectively ( Figure 6). Centrifugal chloroplast location may be advantageous for efficient fixation of CO 2 from intercellular spaces Ueno, 2011;Ueno et al., 2003). In terms of carbon economy, it may be better to fix intercellular CO 2 rather than CO 2 generated in vascular tissue. Because CO 2 is produced in mitochondria, the centripetal location of mitochondria in PS cells may reduce the loss of photorespiratory (and respiratory) CO 2 by lengthening the path of CO 2 diffusion ( Figure 6). In the PS cells of some C 3 grasses, mitochondria were covered by chloroplasts, especially in the vicinity of plasmodesmata feature existing in the PS cells of C 3 grasses. Although there are several studies on the ultrastructure of PS cells in C 3 grasses (Brown et al., 1983;Carolin et al., 1973;Kaneko et al., 1980;Sage & Sage, 2009), they made no mention of the organelle positioning. We consider that it would be difficult to find the particular positioning of organelles without the immunostaining of GDC and the quantification of organelles. An interesting question is what is the physiological significance of this organelle distribution in the PS cells of C 3 grasses. Photosynthetic and photorespiratory activities of these cells are much lower than those of mesophyll cells, as estimated from the numbers of chloroplasts and  Notes. iS, intercellular space; c, chloroplast; mt, mitochondrion; p, peroxisome; v, vacuole; *, chloroplast protrusions (stromules). Bars = 1 μm.

Table 2. intracellular position and number of mitochondria and chloroplasts in mesophyll cells of c 3 grasses.
Notes. Values are means ± Sd of three plants. *indicate significant differences between the two positions of mitochondria (p < 0.05).
In the PS of the Pooideae and Panicoideae, a band around the vascular bundle representing the accumulation of GDC was more prominent than in the PS of the Ehrhartoideae. In light of the results of electron microscopy, the difference in the intensity of this band is likely caused by a higher total number of mitochondria rather than a higher proportion of centripetal mitochondria. The PS cells of Ehrhartoideae appear to contain fewer organelles than those of the Pooideae and Panicoideae, although the difference in the size of PS cells between species should also be considered. Whether the quantitative difference in organelles of PS cells reflects some physiological characteristics of each subfamily is unknown.
In the Pooideae, the PS cells on the xylem side were ordinary, whereas those on the phloem side were structurally similar to mesophyll cells. Williams et al. (1989) named these cells in barley the L-type PS cells. In our study, the number and intracellular position of organelles in L-type PS cells were similar to those in mesophyll cells, indicating an alteration of PS cells to mesophyll cells on the phloem side in the C 3 species of the Pooideae. The BS in C 3 plants is thought to have various physiological roles (Griffiths et al., 2013;Leegood, 2008;Sage, 2001), and C 3 species of the Pooideae would be an interesting model to study BS function in C 3 plants.
between PS cells and vascular bundle cells. These mitochondria may have high respiratory activity to supply energy for metabolite transport (O'Brien & Carr, 1970). Currently, it is unclear whether the positioning of organelles found in the PS cells of C 3 grasses is widespread in other C 3 species. However, at least the proto-Kranz grass species such as Steinchisma laxa (formerly Panicum laxum) and Panicum hylaeicum (Brown et al., 1983;Sage et al., 2014) have increased proportions of centripetal mitochondria and chloroplasts in the PS cells. Subsequently, C 3 -C 4 intermediate species such as Steinchisma hians (formerly Notes. iS, intercellular space; lPS, l-type parenchyma sheath cell; M, mesophyll cell; MS, mestome sheath cell; PS, ordinary parenchyma sheath cell; c, chloroplast; mt, mitochondrion. Bars = 1 μm. Table 3. intracellular position and number of mitochondria and chloroplasts in l-type parenchyma sheath cells of c 3 species of the Pooideae. Notes. Values are means ± Sd of three plants. *indicate significant differences between the two positions of mitochondria (p < 0.05).  Notes. c 2 indicates a two-carbon photorespiratory metabolite. PS, parenchyma sheath.

Species
glutamine synthetase-glutamate synthase cycle in chloroplasts, but some may be lost from leaves (Kumagai et al., 2011). Intracellular positioning of mitochondria may also reduce NH + 4 emission. In this study, we examined plants grown under optimum thermal conditions; the C 3 species of the Ehrhartoideae and Panicoideae were grown under summer conditions, whereas those of the Pooideae were under spring conditions. However, plant mitochondria can move in response to environmental stimuli, including light (Islam et al., 2009;Logan & Leaver, 2000). The development of stromules is also affected by environmental factors Moser et al., 2015;Yamane et al., 2012). Whether the intracellular positioning of mitochondria found in this study is changed by environmental factors that affect photorespiratory activity, such as light intensity, temperature, and CO 2 concentration, will be the topic of our future study.

Disclosure statement
no potential conflict of interest was reported by the authors.

Funding
This study was supported by Japan Society for the Promotion of Science KAKEnHI [grant number 15K14638] to O.U. The wild rice accession used in this study was obtained from the national Institute of Genetics supported by the national Bioresource Project, MEXT, Japan.

Intracellular position of mitochondria and chloroplasts in mesophyll cells
Many mitochondria were positioned on the vacuole side of chloroplasts ( Figure 6), although there was no significant difference in the distribution of mitochondria between the outer and inner positions in the mesophyll cells of B. rigidus, where the gaps between chloroplasts were relatively large and held many mitochondria. In mesophyll cells of Oryza species and O. undulatifolius, stromules extended along the cell wall, and many mitochondria were positioned on the vacuole side of chloroplasts and stromules; this position might help rubisco in chloroplasts including stromules to refix photorespiratory (and respiratory) CO 2 released from mitochondria before it is lost outside mesophyll cells (Bourett et al., 1999;Buchner et al., 2015;Sage & Sage, 2009). Previous studies also examined the ultrastructure of mesophyll cells of C 3 grasses (Bourett et al., 1999;Gielwanowska et al., 2005;Lütz & Engel, 2007;Sage & Sage, 2009;Stata et al., 2014;Winter et al., 1993). In their electron micrographs, we can find the inner positioning of mitochondria as reported in our study. However, all the studies excluding the study on rice by Sage and Sage (2009) made no mention of this positioning.
In rice and wheat, Busch et al. (2013) reported that 24-38% of photorespired and respired CO 2 is reassimilated in mesophyll cells, the periphery of which is covered by chloroplasts and stromules; this reassimilation boosts photosynthesis by 8-11%. Busch et al. (2013) measured S c /S m , which was calculated as the length of cell periphery facing the intercellular air space covered by chloroplasts (S c ) divided by the total length of cell periphery facing the intercellular air space (S m ) in mesophyll cells and used as a structural parameter that affects CO 2 flux in chloroplasts. The value of intracellular positioning of mitochondria measured in our study could be used as a structural parameter indicating CO 2 loss from mitochondria. Our data suggest that the mechanism of photorespiratory (and respiratory) CO 2 scavenging proposed for rice and wheat Sage & Sage, 2009) may be extended to other C 3 grasses. The proportion of mitochondria located in the inner position in mesophyll cells did not largely differ between the C 3 grasses of the Ehrhartoideae and Panicoideae (68-92%), which occur mainly in tropical and subtropical regions, and those of the Pooideae (61-91%), which occur in cold temperate regions. Therefore, intracellular positioning of mitochondria in mesophyll cells is a likely mechanism to scavenge photorespiratory (and respiratory) CO 2 that is widespread in C 3 grasses and is not restricted to certain subfamilies. In the glycolate pathway, NH + 4 is also released during glycine decarboxylation in mitochondria (Keys et al., 1978;Leegood et al., 1995). Most NH + 4 is incorporated in the