Introduction

The CO2 concentrating mechanism in C4 plants generally allows for high rates of net CO2 assimilation and biomass production. During C4 photosynthesis, bicarbonate (HCO3 −) is used to carboxylate phosphoenolpyruvate (PEP) in the mesophyll cells by PEP-carboxylase (PEPc) (Hatch et al. 1967). This reaction generates four-carbon acids, which are subsequently decarboxylated in bundle-sheath cells where ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) and the majority of the C3 cycle is compartmentalized. The CO2 concentrated around Rubisco is typically sufficient to minimize rates of photorespiration; however, the efficiency of this CO2 concentrating mechanism is influenced by the balance between rates of PEPc carboxylation (v p) and Rubisco carboxylation (v c), and the conductance of CO2 between the bundle-sheath and mesophyll cells (g bs). Unfortunately, the efficiency of the CO2 concentrating mechanism is not directly measurable; however, leakiness (ϕ), defined as the fraction of CO2 that is pumped into the bundle-sheath cells that subsequently leaks back out, is often used as a proxy (Farquhar 1983; Hatch et al. 1995).

There have been a number of studies describing how changes in growth conditions (e.g. light, CO2 concentrations, nitrogen, salinity, drought) influence the efficiency of photosynthesis in a variety of C4 species. For example, leakiness was minimal in low-light grown Zea mays, Flaveria bidentis, Bienertia sinuspersici and tended to be lower than high-light grown plants when measured under low photon flux density (PFD) (Kromdijk et al. 2010; Pengelly et al. 2010; Bellasio and Griffiths 2014a, b; Stutz et al. 2014). However, in Amaranthus cruentus, the analysis of leaf stable isotope compositions (δ13C) suggested that the efficiency of C4 photosynthesis was not able to acclimate to the low-light conditions (Tazoe et al. 2006). Additionally, early direct measurements of leaf CO2 isotope exchange (Δ13C) also indicated that leakiness increased under low PFD measurement (Tazoe et al. 2008; Cousins et al. 2008). However, as discussed below, there are questions regarding the use of Δ13C to estimate the efficiency of C4 photosynthesis. Additionally, inaccurate parameterization of bundle-sheath CO2 concentrations and rates of respiration can artificially increase estimates of leakiness using direct measures of Δ13C (Ubierna et al. 2011, 2013; Kromdijk et al. 2014). These recent publications comparing direct measurements of Δ13C and complete models of leaf CO2 isotope exchange suggest that C4 photosynthesis can acclimate to growth under low light to maintain minimal levels of leakiness. However, it still remains unclear how changes in the biochemical capacity of the C4 and C3 cycles and g bs drive this acclimation.

Furthermore, nitrogen limitations may also be important for the ability of C4 photosynthesis to effectively acclimate to low-light growth conditions. For example, nitrogen availability influenced the rates of C4 photosynthesis and leakiness estimated with combined gas exchange and δ13C measurements (Meinzer and Zhu 1998; Tazoe et al. 2006). However, direct measurements of Δ13C and δ13C do not always correlate, and often provide different estimates of leakiness (von Caemmerer et al. 2014; Ellsworth and Cousins 2016). This is in part because δ13C is influenced by post-photosynthetic fractionations that can change over the lifetime of the leaf. Therefore, δ13C is not a direct proxy for instantaneous measurements of Δ13C and estimates of leakiness. As mentioned above, direct measurements of Δ13C generally indicate that leakiness is low in C4 plants; however, this has not been thoroughly tested in response to nitrogen availability and different light growth conditions.

The ability of C4 photosynthesis to acclimate to low light and limited available nitrogen is important for assessing the use of C4 grasses as feedstock for cellulosic biofuel programmes. For example, the C4 grass Miscanthus × giganteus has higher yield than other bio-energy crops (Sims et al. 2006) and significant rates of carbon sequestration into the soil (Hansen et al. 2004; Clifton-Brown et al. 2007). Additionally, Miscanthus is native to Southeast Asia and its cold tolerance makes it suitable for temperate climates (Heaton et al. 2004; Naidu and Long 2004; Farage et al. 2006; Wang et al. 2008). However, in these growth environments and under dense canopies, the low light and limited nitrogen may reduce the efficiency of the C4 concentration mechanism. Additionally, nitrogen limitations in many marginal agricultural lands may also reduce leaf growth and optimum investment in key biochemical steps within the C4 pathway. Therefore, the aim of this study was to test how growth light intensity and nitrogen availability affect leaf anatomy, biochemistry and the efficiency of the CO2 concentrating mechanism in Miscanthus × giganteus.

Materials and methods

Plants

Miscanthus × giganteus (Miscanthus) rhizomes were planted in 6-L pots and grown at Washington State University between May and the end of August 2011. We used a split-plot experimental design where the light treatment was considered the block (replicated four times) and plants within each replicate light treatment were split into the three different nitrogen treatments. The light treatment was either high-light (HL) or low-light (LL) in a temperature-controlled natural light greenhouse with daytime air temperature ranging from 25 to 28 °C and night-time temperature from 20 to 24 °C. The LL conditions (approximately 30 % of incident radiation) were made with four replicated shade structures, each containing plants from the different nitrogen treatments. The location of all pots, including the four shade structures, was frequently changed throughout the greenhouse to avoid potential position effects of each block treatment. Additionally, the nitrogen treatments within a block were randomly arranged and repositioned throughout the experiment. The shade structures were made of 2.1 cm diameter PVC (about 2 cubic meters) with the top and four lateral sides covered with layers of black shade cloth (Polysack Plastic Industries, Nir Yitzhak, Negev, Israel). During clear days, the midday photon flux density (PFD) on top of the pots outside the frame box was approximately 1000 and 300 µmol m−2 s−1 within for HL and LL plants, respectively. The noon daily average of solar radiation (400–1000 nm) for May, June, July and August were 667 ± 54, 680±54, 828±27, 777±17 Watts m−2, respectively. Additionally, the greenhouse had large electric fans to maintain ventilation and air circulation for all plants.

Plants were watered daily and fertilized weekly with 100 ml Hoagland’s solution containing 10 mM KCl, 10 mM CaSO4, 75 µM Iron (Fe-EDTA), 4 mM MgCl2, 0.1 mM H3BO4, 20 µM MnSO4, 20 µM ZnSO4, 4 µM CuSO4, 1 µM MoO3, 1 mM KH2PO4, and a controlled nitrogen amount of either 0.2, 2 or 20 mM NH4NO3. One and half months later, plants were given 100 ml of the same Hoagland’s solution with the corresponding controlled nitrogen content every other day. The plants grown with 0.2, 2, 20 mM NH4NO3 are subsequently referred to as low N (LN), medium N (MN) and high N (HN) treatments, respectively. Approximately, three-month-old Miscanthus plants were used for measurements of leaf gas exchange, photosynthetic discrimination, chlorophyll content, photosynthetic enzyme activity, leaf anatomy, cell wall properties and dry matter N content. Gas exchange and discrimination measurements were made between 9 am and 4 pm local time from July 22nd to August 26th on randomly select plants within a single randomly selected block, this was repeated after all treatments within a block were measured.

CO2 response curves and maximum phosphoenolpyruvate carboxylase activity

Gas exchange measurements were measured on the uppermost fully expanded leaves of four replicate plants per treatment in the 6 cm2 leaf chamber of the LI-6400xt with a red–blue light-emitting diode (LI6400-02B) light source (LI-COR Inc., Lincoln, Nebraska, USA). For all measurements, a leaf temperature of 25°C and a relative humidity between 50 and 70 % were maintained. The CO2 response curves were made at three light intensities (2000, 1000 and 300 μmol quanta m−2 s−1) and the leaf chamber pCO2 was varied in the following sequence: 37, 28, 19, 14, 12, 9, 8, 7, 6, 5.5, 5, 4, 19, 37, 56, 74, 93 Pa.

The initial slope of the linear part of the CO2 response curve was used to estimate the maximum phosphoenolpyruvate carboxylase (PEPc) activity (V pmax) using the following equation (von Caemmerer 2000):

$${\text{Slope}} = \frac{{C_{\text{m}} V_{\text{pmax}} }}{{C_{\text{m}} K_{\text{p}} + V_{\text{pmax}} }},$$
(1)

where C m is mesophyll pCO2, which was assumed to be equal to the pCO2 in the intercellular air space (C i) and K P is the Michaelis–Menten constant of PEPc for CO2 (80 µbar).

Leaf gas exchange and online photosynthetic 13C discrimination

Leaf gas exchange and online photosynthetic discrimination against 13CO2 (∆13C) were measured on the uppermost fully expanded leaves using the LI-6400xt gas exchange analyser with the opaque conifer chamber RGB light source (LI-COR 6400-18, LI-COR Inc., Lincoln, Nebraska, USA) coupled to a tunable-diode-laser absorption spectroscope (TDLAS, TGA 100A, Campbell Scientific, Logan, Utah, USA). The 12CO2 and 13CO2 partial pressure in the LI-COR reference and sample cells were measured by the TDLAS concurrently with a CO2-free tank and two standard tanks (Liquid Technology Corporation, Apopka, FL, USA). The partial pressure of 12CO2 and 13CO2 in the reference and sample lines was calibrated using a gain and offset calculated from the two calibration tanks (Bowling et al. 2003; Ubierna et al. 2011; Sun et al. 2012, 2014). The light response curves were measured by decreasing light intensity from 2000 to 0 µmol photon m−2 s−1 with the leaf chamber at 37 Pa CO2, leaf temperature of 25°C and relative humidity between 50 and 70 %. The simultaneous gas exchange and ∆13C measurements were conducted on four individual plants of Miscanthus from each treatment, and Δ13C was calculated as described by (Evans et al. 1986).

Following the gas exchange and ∆13C measurements, leaves were sampled for the measurements of enzyme activities, chlorophyll content, leaf nitrogen content, leaf carbon and nitrogen isotope composition, specific leaf area (SLA) and monosaccharide composition. Leaf punches for photosynthetic enzyme activities and chlorophyll content measurements were snap-frozen in liquid nitrogen and kept at −80 °C, whereas leaf samples for total nitrogen content, carbon and nitrogen isotope composition were dried in an oven at 80 °C for 48 h. For SLA, leaves were photographed and leaf area was calculated using ImageJ (National Institutes of Health, Bethesda, MA, USA). Dry mass was measured on leaf material oven-dried at 80 °C after 48 h.

Leakiness

Leakiness (ϕ) was estimated by rearranging the equation proposed by Farquhar (1983; Farquhar and Cernusak 2012) and discussed by Ubierna et al. (2013):

$$\phi = \frac{{C_{\text{bs}} - C_{\text{i}} }}{{C_{\text{i}} }} \times \frac{{\Delta {}^{13}C(1 - t)C_{\text{a}} - \bar{a}(C_{\text{a}} - C_{\text{i}} ) - (1 + t)C_{\text{i}} b_{4} }}{{(1 + t)\left[ {b_{3} C_{\text{bs}} - s(C_{\text{bs}} - C_{\text{i}} )} \right] + \bar{a}(C_{\text{a}} - C_{\text{i}} ) - C_{\text{a}} \Delta {}^{13}C(1 - t)}},$$
(2)

where C a, C i and C bs are pCO2 in the atmosphere, in the intercellular air spaces and in the bundle-sheath cells, respectively. Terms \(\bar{a}\), b 3, b 4 and t are defined in Appendix A.

Bundle-sheath conductance (g bs) was solved for each plant by finding the value that minimized the difference between the modelled and observed leaf discrimination for measurements made at 1500 and 2000 μmol quanta m−2 s−1 (Kromdijk et al. 2010; Ubierna et al. 2011). This resulted in g bs values of 0.001 mol m−2 s−1 for plant from all growth conditions.

Enzyme activities and chlorophyll content

The activities of PEPc and Rubisco were measured as previously described by Cousins et al. (2006) with slight modification. For PEPc, 10 μl of leaf extract was combined with 980 μl of assay buffer (50 mM EPPS-NaOH pH 8, 10 mM MgCl2, 0.5 mM EDTA, 0.2 mM NADH, 5 mM glucose-6-phosphate, 1 mM NaHCO3, and 1 U ml−1 malate dehydrogenase) and the reaction was initiated by the addition of 10 μl of 400 mM PEP. For Rubisco, 10 μl of leaf extract was combined with 975 μl of assay buffer (50 mM EPPS-NaOH pH 8, 10 mM MgCl2, 0.5 mM EDTA, 1 mM ATP, 5 mM phosphocreatine, 20 mM NaHCO3, 0.2 mM NADH, 50 U ml−1 creatine phosphokinase, 0.2 mg carbonic anhydrase, 50 U ml−1 3-phosphoglycerate kinase, 40 U ml−1 glyceraldehyde-3-phosphate dehydrogenase, 113 U ml−1 Triose-phosphate isomerise, 39 U ml−1 glycerol 3 phosphate dehydrogenase) and the reaction was initiated by the addition of 15 μl of 34.4 mM ribulose-1,5-biphosphate (RuBP). Enzyme activity was calculated by monitoring the decrease of NADH absorbance at 340 nm with an Evolution 300 UV–VIS spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Additionally, the chlorophyll content was measured according to Porra et al. (1989).

Measurements of leaf anatomical traits

The measurements of leaf anatomy were carried out on mature leaves of similar age used for gas exchange measurements. Samples were fixed in 2 % (v/v) paraformaldehyde and 2 % (v/v) glutaraldehyde in 0.1 M phosphate buffer (pH 7.2), transferred subsequently to 3.5 % paraformaldehyde and 2.5 % glutaraldehyde in 0.1 M sodium cacodylate, 0.12 M sucrose, 10 mM ethylene glycol tetra-acetic acid and 2 mM magnesium chloride for overnight at 4 °C, and postfixed in 2 % (v/v) OsO4 for 2 h at room temperature. Samples were then dehydrated in an acetone series and embedded in Spurr’s epoxy resin. Cross semi-thin (1 µm) and ultra-thin (70 nm) sections were made on a Reichert Ultracut R ultramicrotome (Reichert-Jung GmbH, Heidelberg, Germany). For light microscopy, sections were stained with 1 % (w/v) Toluidine blue O in 1 % (w/v) Na2B4O7, and studied under the Olympus BH-2 (Olympus Optical Co., Ltd.) light microscope with LM Digital Camera & Software (Jenoptik ProgRes Camera, C12plus, Jena, Germany). For transmission electron microscopy (TEM), ultra-thin cross sections were stained with 4 % (w/v) uranyl acetate followed by 2 % (w/v) lead citrate. FEI Tecnai G2 (Field Emission Instruments Company, Hillsboro, OR, USA) equipped with Eagle FP 5271/82 4K HR200KV digital camera transmission electron microscope was used for observation and photography.

Leaf and cell structural traits were characterized from cross sections (avoiding the central vein) of one mature leaf from four different individuals per treatment. Digital images were analysed with Image Analysis (UTHSCSA, version 3.0, University of Texas, San Antonio, TX, USA). The leaf thickness and interveinal distance were measured from semi-thin sections using 5–10 different fields of view for each leaf. Volume fracture of intercellular air spaces (IS) per leaf mesophyll area was calculated as a ratio of area of IS and mesophyll area (leaf section area minus epidermis and vascular tissues). The mesophyll surface area exposed to IS per unit leaf area (S mes) was calculated from measurements of total length of mesophyll cell walls exposed to IS and width of section analysed with curvature correction factor as 1.34 (Evans et al. 1994). The bundle-sheath surface area per unit leaf area (S bs) was measured within the interveinal distance as described by Pengelly et al. (2010) and the percentages of the bundle-sheath perimeter exposed to IS (% BS CW to IS) and percentages of bundle-sheath perimeter not covered by chloroplasts (% BS CW w/o chloroplast) were also calculated. The equations and description of these anatomical calculations are presented in the supplemental Appendix B. The mesophyll and bundle-sheath cell wall thickness was measured from TEM micrographs using at least 10 images for each leaf.

Leaf N content and δ13C

The oven-dried leaves were ground to a consistent powder and dried again at 80 °C for 30 min before a small sample was weighed for nitrogen content and isotope composition measurements. The percentage nitrogen was measured by combustion of samples in an elemental analyser (ECS 4010, Costech Analytical, Valencia, CA) and the isotopic composition of CO2 and N2 was analysed with a continuous flow isotope ratio mass spectrometer (Delta PlusXP, Thermo Finnigan, Bremen) at Washington State University Stable Isotope Facility. Precision of repeated measurements of laboratory standard was <0.1 ‰. δ13C values are reported relative to V-PDB.

Monosaccharide composition of cell wall extracts

Alcohol insoluble residue was prepared as described earlier by Harholt et al. (2006) from the same leaves used for the gas exchange measurements. Samples were hydrolyzed in 2-N-trifluoroacetic acid for 1 h at 120 °C. HPAEC-PAD analysis and quantitation of rhamnose, arabinose, galactose, glucose, xylose, galacturonic acid and glucuronic acid was performed according to ØBro et al. (2004) on an ICS 3000 (Dionex, Sunnyvale, CA) using a CarboPac PA20 (3 × 150 mm, Dionex, Sunnyvale, CA) anion exchange column.

Statistical analysis

A split-plot two-way analysis of variance (ANOVA) was conducted to evaluate differences in SLA, leaf anatomical traits, leaf N content, chlorophyll content, V pmax, and leakiness among nitrogen and light treatments. Simple linear regression analysis was conducted to evaluate relationships between rate of CO2 assimilation and leaf N content, leakiness and leaf N content, chlorophyll content and leaf N content, as well as rate of CO2 assimilation and chlorophyll content. All statistical analyses were carried out using Statistix version 9.0 (Analytical Software, Tallahassee, FL, USA).

Results

Leaf anatomical traits

Leaf thickness was greater in high-light (HL) compared to low-light (LL) plants and generally increased with leaf nitrogen (Fig. 1a, b; Table 1). Additionally, the specific leaf area (SLA; leaf area per unit dry mass) was lower under HL compared to LL plants but was not significantly affected by nitrogen (Table 1). The interveinal distance was greater under HL compared to LL but was not affected by nitrogen (Fig. 1a, b; Table 1). The mesophyll surface area exposed to the intercellular airspace per unit leaf area (S mes; m2 m−2) was greater in HL compared to LL but was not affected by nitrogen (Table 1). Additionally, the bundle-sheath area per unit leaf area (Sbs; m2 m−2) was larger in HL compared to LL plants and increased with nitrogen (Table 1). The percentage of the bundle-sheath perimeter exposed to IS (% BS CW to IS) was not affected by the growth conditions, while the percentages of bundle-sheath perimeter not covered by chloroplasts (% BS CW w/o chloroplast) were lower in HL as a result of sparse chloroplast distribution in bundle-sheath under LL conditions (Table 1). Additionally, the bundle-sheath cell wall thickness was significantly greater in the HL compared to LL grown plants but did not change with nitrogen (Table 1); however, the mesophyll cell wall thickness was not affected by either condition. The leaf cell wall sugar composition varied slightly with rhamnose and xylose content changing with nitrogen, while only galacturonic acid was significantly different between HL and LL plants (Table 1).

Fig. 1
figure 1

Leaf cross sections (a–f) in Miscanthus × giganteus grown at two different irradiances of 1000 µmol quanta m−2 s−1 (high light, a, c, e) and 300 µmol quanta m−2 s−1 (low light, b, d, f), and different nitrogen conditions, high N (a, b), medium N (c, d) and low N (e, f). Scales: 50 µm

Full size image
Table 1 Leaf anatomical properties of Miscanthus × giganteus grown at two different irradiances of 1000 (High light) and 300 µmol quanta m−2 s−1 (Low light), and three different nitrogen conditions (High N, HN; Medium N, MN; Low N, LN)
Full size table

Leaf chlorophyll, nitrogen, carbon isotope composition and enzyme activity

The HL plants had less chlorophyll (mmol m−2) relative to the LL plants and the content increased with nitrogen in both HL and LL (Table 2); however, the chlorophyll a/b was insensitive to light treatments but decreased with nitrogen (Table 2). The total N per unit leaf area was significantly affected by both nitrogen and light treatments, with the leaf N content in HL-HN leaves higher than leaves from all other treatments (Table 2). Under HL, the carbon isotope composition (δ 13C) of dried leaves was more enriched than the LL treatment; however, δ 13C was more depleted with increasing nitrogen availability in both HL and LL plants (Table 2). The in vitro activity of both PEPc and Rubisco was greater in HL compared to LL plants, and increased with nitrogen availability; however, PEPc/Rubisco ratio increased only with light and not nitrogen (Table 2).

Table 2 Leaf parameters of Miscanthus × giganteus grown at two different irradiances of 1000 (High light) and 300 µmol quanta m−2 s−1 (Low light) and three nitrogen conditions (High N, HN; Medium N, MN; Low N, LN)
Full size table

Net CO2 assimilation in response to changing CO2 concentrations

The CO2-saturated rate of net CO2 assimilation (A sat) in the HL-HN compared to LL–HN plants was significantly higher under 2000 µmol quanta m−2 s−1 (30.2 ± 0.7 vs. 21.0 ± 2.2 µmol m−2 s−1) and 1000 µmol quanta m−2 s−1 (23.9 ± 1.1 vs. 16.36 ± 1.8 µmol m−2 s−1) but had only slightly higher rates at 300 µmol quanta m−2 s−1 (12.2 ± 0.5 vs. 10.2 ± 0.8 µmol m−2 s−1) (Fig. 2 a, d). In the MN and LN grown plants, A sat was similar between HL and LL grown plants regardless of the measurement light intensities (Fig. 2b vs. e, c vs. f). Values of A sat decreased with reduced nitrogen availability regardless of growth light conditions; however, the response was greater in the HL compared to LL plants (Fig. 2). The maximum PEPc carboxylase activity (V pmax) determined from the initial slope of the CO2 response curves measured under 2000 µmol quanta m−2 s−1 and using Eq. (1) increased with nitrogen in both HL and LL (Table 2), and was significantly higher under HL compared to LL regardless of nitrogen treatment (Table 2).

Fig. 2
figure 2

Net rate of CO2 assimilation in response to CO2 partial pressure (C i) under different measurement light intensities of 2000 (filled circles), 1000 (open circles) and 300 (inverse triangles) µmol quanta m−2 s−1 in plants grown at 1000 (HL) and 300 µmol quanta m−2 s−1 (LL), and three different nitrogen conditions (High N, HN; Medium N, MN; Low N, LN). Leaf temperature and leaf chamber relative humidity were controlled at 25 °C and 50–70 %, respectively. Data are reported as the arithmetic mean ± 1 standard error (n = 4)

Full size image

Photosynthetic carbon isotope discrimination in response to light

In all plants, the net rate of CO2 assimilation (A net) increased with light intensity (Fig. 3a, b). In the HL plants, A net measured above 500 µmol quanta m−2 s−1 was higher in the HN treatment compared to MN and LN treatments; however, below this measurement light intensity, there was no difference in A net between nitrogen treatments for the HL plants. In the LL grown plants, there was not a significant difference in A net between nitrogen treatments across all the measurement light intensities. Overall, A net was higher in HL compared to LL grown plants regardless of nitrogen treatment (Fig. 3a, b).

Fig. 3
figure 3

Net CO2 assimilation rate, (a) and (b), carbon isotope discrimination (Δ13C), (c) and (d), and bundle-sheath leakiness (ϕ), (e) and (f), as a function of irradiance, in Miscanthus × giganteus grown at two different irradiances of 1000 (high light) and 300 µmol quanta m−2 s−1 (low light), and three different nitrogen conditions (high N, square; medium N, circle; low N, triangle). Measurements were made at an ambient pCO2 of ~35 Pa, and a leaf temperature of 25 °C. Data are reported as the arithmetic mean ± 1 standard error (n = 4)

Full size image

The photosynthetic carbon isotope discrimination (Δ13C) decreased with irradiance for both HL and LL grown plants, regardless of nitrogen treatments (Fig. 3c, d). However, leakiness (ϕ) calculated using Eq. (4) was relatively constant in response to changes in the measurement light intensities in all plants. However, in the LL–MN and LL–LN plants, values of ϕ increased under the lowest measuring condition of 60 µmol quanta m−2 s−1. In general, leakiness across all measured light conditions was slightly lower in HL-HN plants compared to plants from all other treatments (Fig. 3e, f).

Correlations between leaf nitrogen content, A net and leakiness

There was a positive correlation of A net (Fig. 4a; r 2 = 0.93, P = 0.001) and (Fig. 4b; r 2 = 0.72, P = 0.033) with total leaf N. This was primarily driven by the fact that A net (31.2 ± 0.7 µmol m−2 s−1) was significantly higher and ϕ (0.26 ± 0.01) significantly lower in the HL-HN plants compared to plants from the other treatments. The values of A net and ϕ taken from measurements at 2000 µmol quanta m−2 s−1 are presented in Fig. 3.

Fig. 4
figure 4

Relationship between total leaf N (mmol m−2) and a Net CO2 assimilation rate (A net; µmol CO2 m−2 s−1), and b bundle-sheath leakiness (ϕ), in Miscanthus × giganteus grown at two different irradiances of 1000 (High light, open symbols) and 300 µmol quanta m−2 s−1 (Low light, filled symbols), and three different nitrogen conditions (High N, square; Medium N, circle; Low N, triangle). The measurements of A net and ϕ are taken from Fig. 3 at 2000 µmol quanta m−2 s−1. Data are reported as the arithmetic mean ± 1 standard error (n = 4)

Full size image

Discussion

There is limited information on how light and nitrogen availability influence leaf anatomy, biochemistry and the efficiency of the CO2 concentrating mechanism in C4 grasses. Below we discuss how measurements of microscopy, biochemistry and leaf CO2 isotope exchange where used to determine how changes in leaf anatomy and biochemistry influence the efficiency of C4 photosynthesis in Miscanthus grown under different light and nitrogen treatments.

Leaf structure and CO2 movement

In C3 plants, mesophyll conductance to CO2 (g m) restricts substrate availability to Rubisco within the mesophyll chloroplast (Flexas et al. 2014). However, in C4 plants, it is typically assumed that g m has to be large to maintain the high rates of CO2 assimilation, particularly when intercellular CO2 concentrations are low as in many C4 grasses (von Caemmerer et al. 2008, 2014). The mesophyll surface area exposed to the intercellular airspace per unit leaf area (S mes) and cell wall thickness have been implicated in determining g m in both C3 and C4 plants (Evans et al. 1994; Pengelly et al. 2010; von Caemmerer et al. 2014). In fact, changes in these parameters due to growth conditions significantly influence g m in C3 plants; however, there is little information on the response of g m in C4 plants (Gillon and Yakir 2000). The increase in S mes without changes in mesophyll cell wall thickness in the HL compared to the LL grown plants suggests that g m is higher in plants with greater photosynthetic capacity. This would minimize the drawdown of CO2 between the intercellular air space and the mesophyll cytoplasm to enhance HCO -3 availability to PEPc to maintain rates of CO2 assimilation. Unfortunately, unlike in C3 plants, the use of Δ13CO2 is not able to resolve g m in C4 plants.

Similarly, bundle-sheath CO2 conductance (g bs) is in part influenced by the bundle-sheath surface area per unit leaf area (S bs) and the bundle-sheath cell wall thickness (von Caemmerer and Furbank 2003). Values of g bs are not directly measurable and are therefore typically estimated, with certain assumptions, from combined measurements of gas exchange, stable isotope analysis and chlorophyll fluorescence (Bellasio and Griffiths 2014a; Kromdijk et al. 2014; Yin et al. 2016). The value of g bs can have a strong influence on the efficiency of the CO2 concentrating mechanism; therefore, changes in S bs or the bundle-sheath cell wall thickness in response to changes in growth conditions may directly impact leakiness. In the HL plants, the bundle-sheath cell wall thickness and S bs were greater than in LL plant suggesting that any increase in g bs potentially caused by changes in S bs is offset by the increase in cell wall thickness. Additionally, the bundle-sheath cell area (data not shown) was larger in the HL compared to LL plants, which likely influenced the increased interveinal distance in HL plants. However, compared to the HL plants, the percentages of bundle-sheath perimeter not covered by chloroplasts (% BS CW w/o chloroplast) in the LL plants were greater, which may increase the lose of CO2 from the bundle-sheath cells. The nitrogen treatment did not significantly increased cell wall thickness but S bs did significantly increase with nitrogen, which suggests that g bs would be higher in plants with increasing leaf nitrogen. Although there were significant differences in the content of some cell wall monosaccharides (Table 2), the actual percentage change between growth treatments was small. However, the changes in rhamnose, galacturonic acid and galactose are likely to affect the composition of pectin (Scheller et al. 2007), which could influence the diffusion properties of the mesophyll or bundle-sheath cell walls. As discussed below, is influenced by both g bs and the biochemical capacity of the C3 and C4 cycle. Therefore, the potential change in g bs due to changes in the bundle-sheath cells may be compensated by changes in leaf metabolism in order to maintain ϕ under these growth conditions.

C4 Biochemistry

In general, the extractable rates of in vitro PEPc and Rubisco activity were greater in HL versus LL plants and increased with leaf nitrogen increased (Table 2). However, as with rates of net CO2 assimilation (Figs. 2, 3), the difference in PEPc and Rubisco activities was not significantly different between HL and LL plants with low leaf nitrogen (MN and LN). This suggests that limited nitrogen availability can diminish the potential biochemical acclimation of C4 photosynthesis to high-light conditions. Additionally, under low-light conditions, C4 photosynthesis is not able to take advantage of the increased nitrogen availability. Interestingly, the PEPc/Rubisco ratio was not influenced by the nitrogen treatments, regardless of growth light conditions and PEPc/Rubisco was higher in the HL compared to LL plants. This indicates that regardless of leaf nitrogen, the partitioning of nitrogen between PEPc and Rubisco remained constant suggesting a constant potential capacity for these steps of the C4 and the C3 cycle, respectively. However, in the HL plants, the balance of PEPc/Rubisco increased compared to LL plants suggesting a greater capacity of the C4 cycle compared to the Rubisco-driven C3 pathway. If the in vitro activities of these enzymes are proxies for the capacity of the C4 and C3 cycles, respectively, then all else being equal this would suggest a potential over cycling of the CO2 concentrating mechanism and an increase in leakiness. However, the HL plants in the HN treatment had decreased values of leakiness, which as mentioned above is determined by both the capacity of the C4 and C3 cycles as well as g bs. It is therefore likely that the change in PEPc/Rubisco in the HL and HN plants was offset by a decrease in g bs, which would minimize leakiness. However, it should be noted that the extractable maximum activities of PEPc and Rubisco might not accurately depict the in vivo regulated rates.

For example, the in vivo estimates of V pmax determined from the initial slope of the CO2 response curves (see Eq. 1) are significantly less than the in vitro PEPc activity, regardless of growth condition. This suggests either significant regulation of leaf PEPc activity and the enzyme is not 100 % active or the in vivo estimates are under represented. Based on a sensitivity analysis of parameters in Eq. (1), only errors in the measurements of net CO2 assimilation (A net) or the mesophyll CO2 partial pressure (C m) would be sufficient to alter in vivo V pmax to match the in vitro values. While the measurements of A net could be off, it is unlikely they are systematically erroneous. It is more likely that the values of C m which are derived from C m = C i−A/g m are incorrectly estimated due to uncertainties of g m. It is typically assumed that g m is high in C4 plants and tends not to restrict CO2 availability to the mesophyll cells. However, it is difficult to estimate g m in C4 plants and theoretically g m could be relatively low as in C3 plants, which would lower the estimated values of C m. If g m were low then the assumptions used here that C m = C i would be invalid and the derived V pmax from Eq. (1) would be significantly higher. As mentioned before, g m in C4 plants is difficult to measure with Δ13C; however, future analysis of g m in C4 plants could be investigated with combined measurements of Δ18O and the isotopic signature of transpired water (Gillon and Yakir 2000; von Caemmerer et al. 2014; Barbour et al. 2016).

Leaf carbon isotopic signature and CO2 exchange

The leaf carbon isotopic signature (δ13C) was more depleted in LL compared to HL plants and δ13C became more depleted as leaf nitrogen increased (Table 1). In C4 plants, one of the potential driving factors influencing δ13C is the efficiency of the CO2 concentrating mechanism and leakiness. Therefore, the δ13C differences suggest leakiness is greater in LL compared to HL plants and increases with leaf nitrogen. However, recent comparisons of δ13C and estimates of leakiness from measurements of leaf CO2 isotope exchange (Δ13C) suggest that variation in δ13C does not always correspond to changes in Δ13C (Cousins et al. 2008; Kromdijk et al. 2014; von Caemmerer et al. 2014). In fact as discussed below, the Δ13C and the corresponding leakiness were slightly lower in the HL-HN grown plants compared to plants from the other growth conditions; however, there was not a significant difference in leakiness between HL and LL plants across all nitrogen treatments.

Alternatively, δ13C differences could be caused by post-photosynthetic fractionation. However, conservation of mass requires that there must be a loss of carbon from the leaf, for example, via respiration or export from the leaf to influence δ13C. Therefore, differences in the CO2 released during respiration between plants from the different growth conditions could influence leaf δ13C (Ghashghaie et al. 2001). However, the rates of dark respiration were not significantly different between plants from the various growth conditions (data not shown) and the rates were an order of magnitude lower than rates of CO2 assimilation. Therefore, the isotopic fractionation of respiration would have to be unrealistically different between plants. For modelling, δ13C values of respiratory fractionation were not directly measured but estimated as described in Appendix A; however, a sensitivity analysis indicated that even extreme values would not drive the Δ13C shift seen between treatments. Unfortunately, there are limited data in the literature on how respiratory fractionation varies between C4 plants, particularly in response to changing environmental conditions. Therefore, further research is needed to determine the influence of leaf respiration on Δ13C.

Photosynthetic efficiency

The efficiency of the CO2 concentrating mechanism in C4 plants cannot be directly measured; however, through modelling and measurements of Δ13C the efficiency can be estimated from calculations of leakiness (see Eq. 2; Ubierna et al. 2011; Kromdijk et al. 2014; von Caemmerer et al. 2014). There has been a significant discussion in the literature about the sensitivity of leakiness to changes in environmental conditions, particularly under low-light growth and measurement conditions. Early measurements suggested that leakiness increased under low-light measuring conditions estimated from both of Δ13C and combined gas exchange with fluorescence measurements (Henderson et al. 1992; Cousins et al. 2006; Tazoe et al. 2006; Kubásek et al. 2007; Cousins et al. 2008; Kromdijk et al. 2008; Tazoe et al. 2008; Kromdijk et al. 2010). However, more recent estimates using Δ13C and taking into account the full model of C4 leaf CO2 isotope exchange suggest leakiness changes only slightly under the lowest light measurements conditions (Ubierna et al. 2011, 2013; Kromdijk et al. 2014). Additionally, C4 plants grown under low light acclimate to these conditions and have lower leakiness under low light compared to higher light grown plants (Pengelly et al. 2010; Bellasio and Griffiths 2014a, b, c; Kromdijk et al. 2014). It is important to note that leakiness at very low light is difficult to estimate because measurements are confounded by the influence of respiration in the light, especially from the bundle-sheath mitochondria. Therefore, the leakage of CO2 from the bundle-sheath cells under low light may relate to both the CO2 concentrating mechanism and day respiration, the later not directly related to leakiness (Kromdijk et al. 2014; Sage 2014). Here the estimates of leakiness were similar across all growth conditions, except for in the HL and HN plants, which had slightly lower leakiness under all measurement light intensities (Fig. 4). The response of leakiness to decreasing measurement light intensities was relatively small except for in the LN plants at the lowest measurement light intensity. These estimates suggest that the CO2 concentrating mechanism in Miscanthus is robust under these growth conditions, and that the observed changes in leaf anatomy and biochemistry likely help to maintain this efficiency.

Conclusions

The growth of Miscanthus under different light and nitrogen treatments has a strong influence on the leaf anatomy and the levels of extractable PEPc and Rubisco activities. Additionally, the net rates of CO2 assimilation were typically lower under the LL treatment expect in the plants with the lowest leaf nitrogen. Taken together these data suggest that photosynthesis and leaf anatomy in Miscanthus are influenced by an interaction of both light availability and leaf nitrogen. However, the measurements of Δ13C and modelled estimates of leakiness suggest that the efficiency of the CO2 concentrating mechanism in Miscanthus is robust, even when environmental growth conditions drive changes in leaf anatomy and biochemistry.