Sink–Source Balance and Down-Regulation of Photosynthesis in Raphanus sativus: Effects of Grafting, N and CO₂

Abstract

To clarify whether excessive accumulation of total non-structural carbohydrate (TNC) causes down-regulation of photosynthesis in Raphanus sativus, we manipulated sink–source balance to alter TNC levels in source leaves and examined its effects on photosynthetic characteristics, whole-plant biomass allocation and anatomical characteristics of leaves and petioles. Comet and Leafy varieties with large and small hypocotyls were reciprocally grafted to change hypocotyl sink strength. They were grown at high or low nitrogen (N) availability and at elevated or ambient CO₂. Maximum photosynthetic rate, which was highly correlated with Rubisco and leaf N contents, was hardly correlated with TNC across the grafting combinations and growth conditions. Biomass allocation to petioles and hypocotyls and accumulation of TNC in each organ were significantly higher at low N. TNC and structural carbohydrates such as cellulose and hemicellulose were higher and the proportion of intercellular air space in source leaves was lower at low N and elevated CO₂. We conclude that excess TNC does not cause severe down-regulation of photosynthesis, and cell walls and petioles are also major carbohydrate sinks responding to changes in sink–source and carbon-nitrogen balances, which contribute to alleviating further accumulation of TNC to avoid its negative effects in source leaves.

Introduction

The sink–source balance of the plant changes depending on its environment. Accordingly, the environment can change the biomass of leaves, stems and roots via changing the sink and source strength of these organs. The photosynthetic capacity of source leaves also changes with changes in biomass allocation. For example, the leaf-to-root ratio of the plant differs depending on irradiance and/or soil nitrogen (N) availability, which lead to an optimal photosynthetic rate to maximize relative growth rate (RGR) of the plant in that environment (Hilbert 1990, Osone and Tateno 2003, Sugiura and Tateno 2011).

When non-structural carbohydrates such as soluble sugars and starch accumulate in the leaves, photosynthesis may be suppressed. This suppression is known as down-regulation or feedback regulation of photosynthesis (Sheen 1994, Stitt and Krapp 1999). The accumulation of total non-structural carbohydrate (TNC, sum of soluble sugars and starch) in leaves and down-regulation of photosynthesis are often observed concurrently, especially under elevated CO₂ conditions. However, the extent of down-regulation differs markedly among species or cultivars (Goldschmidt and Huber 1992, Ainsworth and Rogers 2007). Since it is predicted that atmospheric CO₂ will continue to increase towards the end of this century (IPCC 2014), it is very important to reveal the causal relationships among CO₂ concentration during growth, accumulation of TNC and the down-regulation of photosynthesis from the viewpoint of the whole-plant sink–source balance.

It is widely believed that down-regulation of photosynthesis occurs when the source activity of leaves exceeds the sink strength (Paul and Driscoll 1997, Nakamura et al. 1999, Ainsworth et al. 2004). A situation of elevated CO₂ and low N availability, which results in excessive accumulation of carbohydrates in source leaves, would cause severe down-regulation of photosynthesis (Krapp and Stitt 1995, Stitt and Krapp 1999). Plants with large sink organs, such as radish (Raphanus sativus) (Usuda and Shimogawara 1998) and potato (Solanum tuberosum L.) (Sage et al. 1989) show little down-regulation even at elevated CO₂, probably because TNC are preferentially translocated to the storage sink organs, keeping the TNC level in the leaves low.

Even at elevated CO₂, if N availability was high, neither suppression of photosynthesis nor marked accumulation of TNC was observed (Stitt and Krapp 1999, Theobald et al. 1998, Makino and Mae 1999, Sanz-Sáez et al. 2010). It is thus possible that down-regulation of photosynthesis at elevated CO₂ is simply attributable to N depletion, which would occur due to growth enhancement by elevated CO₂, particularly in plants grown in small pots (Arp 1991). However, Araya et al. (2010) conducted sugar-feeding experiments with Phaseolus vulgaris and proposed that the increase in TNC decreases leaf photosynthetic capacity even at high N availability. It was also found that the Rubisco content is selectively decreased, whereas the N concentration is little affected at elevated CO₂ in some plant species, such as Brassica oleracea, Chenopodium album (Sage et al. 1989) and P. vulgaris (Nakano et al. 1998), suggesting that accumulation of TNC causes the selective decrease in Rubisco in these plant species. The importance of C/N balance has also been noted by Paul and Driscoll (1997). They showed that, in tobacco, simultaneous decreases in sink and source activities after N withdrawal and shading did not cause down-regulation of photosynthesis, even though starch accumulated in the source leaves. From the lines of evidence mentioned above, accumulation of TNC and sink–source and carbon-to-nitrogen (C/N) balances are involved in the down-regulation of photosynthesis, but the mechanisms and interspecific differences underlying this down-regulation are not fully understood.

To investigate how sink–source imbalance affects the down-regulation of photosynthesis through changes in TNC in source leaves, we conducted reciprocal grafting experiments using two Raphanus sativus varieties with large and small hypocotyl sink strengths at medium N and ambient CO₂ conditions (Sugiura et al. 2015). Although there are alternative ways, such as girdling (Goldschmidt and Huber 1992), to manipulate TNC levels, the advantage of this method is that grafted plants can live longer and would not show injury responses once the grafting was completed. Our previous study raised the possibility that excess TNC can be converted to some structural carbohydrates or exported to other sink organs to alleviate its negative effects on the photosynthetic apparatus in radish plants. This is because excess accumulation of sugars and starch could lead to photoinhibition (Layne and Flore 1993, Urban and Alphonsout 2007), damage the photosynthetic reaction center and distort chloroplast ultrastructure. Furthermore, toxic metabolites derived from photosynthates, such as reactive carbonyls, also damage proteins and DNA (Shimakawa et al. 2013, Takagi et al. 2014), and excessive sugar accumulation can also cause decreases in osmotic potential.

For further understanding of the relationships among photosynthetic capacity, TNC, sink–source and C/N balances and the accumulation of structural carbohydrates, we propose three working hypotheses to be tested using our radish grafting system: (i) the effect of accumulation of TNC on the down-regulation of photosynthesis is minor in radish plants because the capacity for TNC storage in mesophyll cells is high, and excess TNC can be efficiently allocated from leaves to other sink organs, such as petioles and hypocotyls; (ii) excess TNC in source leaves would be converted to structural carbohydrates, which would increase cell wall components and change leaf anatomical traits; and (iii) the capacity for sugar transport in source leaves would be enhanced by increasing the size of the vascular bundles in the petioles.

To test these hypotheses, we conducted reciprocal grafting experiments not only at high N and ambient CO₂, as in Sugiura et al. (2015), but also under low N and elevated CO₂ conditions, where TNC level would increase greatly. Anatomical characteristics, C-related components (starch, soluble sugars and cell wall materials) and N-related components (Rubisco and total N) of the source leaves were examined to evaluate how these factors were interrelated. Whether the down-regulation of photosynthesis occurred and whether it was accompanied by a selective decrease in the Rubisco content were carefully evaluated. Biomass and TNC levels were measured for leaves, petioles, hypocotyls and fine roots to obtain insight into translocation of excess TNC. Moreover, anatomical examinations of petiolar vascular bundles were made because vascular structure would be related to TNC export capacity (Adams et al. 2013a, 2013b). Based on the results obtained, we discuss the mechanisms underlying the regulation of photosynthetic capacity in response to changes in sink–source balance and C/N ratio in radish.

Results

Characteristics of biomass allocation and whole-plant growth

Although CL plants [scions of Leafy (L) grafted on stocks of comet (C)] and LL plants (L scions grafted on L stocks) needed more time than CC (C scions grafted on C stocks) and LC (L scions grafted on C stocks) plants until scions and stocks were completely united, successfully grafted plants grew well during the experimental period (Fig. 1). There was no difference in dry mass and biomass allocation to each organ between pre-grafted C and L plants (Fig. 2A), whereas these traits differed markedly among the grafting combinations and growth conditions after grafting (Fig. 2B). Leaf-to-root ratio (L/R), leaf area-to-root ratio (L_A/R), leaf mass ratio (LMR) and petiole mass ratio (PMR) were higher at high N (HN-aCO₂ and HN-eCO₂), whereas hypocotyl mass ratio (HMR) was higher at low N (LN-aCO₂ and LN-eCO₂) (Table 1). LMR and PMR were higher in CL and LL with small hypocotyls, whereas HMR was higher in CC and LC with large hypocotyls.

Table 1

Biomass allocation patterns of grafted plants

L/R (g g−1)			ANOVA			N**			${\text{CO}}_2^{\text{ns}}$		N × ${\text{CO}}_2^{\text{ns}}$
	CC				LC				CL				LL
LN-aCO2	3.90	ab	±	0.61	5.73	a	±	2.10	3.31	b	±	0.76	5.31	a	±	0.85
LN-eCO2	4.64	b	±	0.61	6.19	a	±	1.65	4.45	b	±	0.45	5.88	ab	±	0.98
HN-aCO2	8.47	a	±	2.32	8.86	a	±	1.87	7.71	a	±	1.16	7.64	a	±	0.81
HN-eCO2	6.78	b	±	0.80	10.43	a	±	2.54	6.85	ab	±	1.82	8.03	ab	±	1.61
LA/R (m2 g−1)			ANOVA			N**			CO2*		N × ${\text{CO}}_2^{\text{ns}}$
	CC				LC				CL				LL
LN-aCO2	0.118	a	±	0.014	0.107	ab	±	0.041	0.070	c	±	0.015	0.081	bc	±	0.015
LN-eCO2	0.106	a	±	0.026	0.087	ab	±	0.024	0.061	b	±	0.004	0.084	ab	±	0.011
HN-aCO2	0.352	a	±	0.101	0.349	a	±	0.061	0.306	a	±	0.080	0.300	a	±	0.042
HN-eCO2	0.283	ab	±	0.036	0.372	a	±	0.095	0.253	b	±	0.068	0.251	b	±	0.057
LMR (g g−1)			ANOVA			N**			${\text{CO}}_2^{\text{ns}}$		N × ${\text{CO}}_2^{\text{ns}}$
	CC				LC				CL				LL
LN-aCO2	0.228	d	±	0.026	0.299	c	±	0.023	0.436	b	±	0.040	0.573	a	±	0.035
LN-eCO2	0.249	d	±	0.059	0.363	c	±	0.057	0.482	b	±	0.030	0.546	a	±	0.064
HN-aCO2	0.397	b	±	0.032	0.423	b	±	0.076	0.586	a	±	0.041	0.654	a	±	0.031
HN-eCO2	0.345	c	±	0.056	0.368	c	±	0.064	0.559	b	±	0.062	0.665	a	±	0.025
PMR (g g−1)			ANOVA			N**			${\text{CO}}_2^{\text{ns}}$		N × ${\text{CO}}_2^{\text{ns}}$
	CC				LC				CL				LL
LN-aCO2	0.051	a	±	0.011	0.061	a	±	0.011	0.131	b	±	0.019	0.146	b	±	0.021
LN-eCO2	0.043	a	±	0.009	0.075	b	±	0.019	0.130	c	±	0.016	0.152	c	±	0.026
HN-aCO2	0.101	a	±	0.017	0.089	a	±	0.019	0.181	b	±	0.020	0.187	b	±	0.020
HN-eCO2	0.094	a	±	0.014	0.073	a	±	0.014	0.151	b	±	0.020	0.182	c	±	0.024
HMR (g g−1)			ANOVA			N**			${\text{CO}}_2^{\text{ns}}$		N × ${\text{CO}}_2^{\text{ns}}$
	CC				LC				CL				LL
LN-aCO2	0.662	a	±	0.041	0.582	b	±	0.048	0.297	c	±	0.042	0.172	d	±	0.042
LN-eCO2	0.653	a	±	0.080	0.500	b	±	0.081	0.279	c	±	0.044	0.207	c	±	0.040
HN-aCO2	0.453	a	±	0.047	0.439	a	±	0.099	0.155	b	±	0.031	0.073	b	±	0.021
HN-eCO2	0.510	a	±	0.070	0.521	a	±	0.086	0.206	b	±	0.060	0.067	c	±	0.014

L/R, L_A/R, LMR and HMR were determined 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂.

Values are means ± SD (n = 5–7).

Effects of N (low and high) and CO₂ (ambient and elevated) were evaluated by analysis of variance (ANOVA, **P < 0.01, *P < 0.05; ns, P > 0.05).

Different lower case letters indicate significant differences among CC, LC, CL and LL (Tukey’s test, P < 0.05).

Fig. 1

Harvested grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. Each scale bar indicates 30 cm.

Fig. 2

Dry mass of leaves, hypocotyls and fine roots obtained by growth analysis of radish plants before grafting 12–14 days after sowing (A) and those of grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively (B). Bars from top to bottom represent values for leaves (white), petioles (light gray), hypocotyls (dark gray) and fine roots (black). Values for hypocotyls of the scions (upper dark gray bars) and those of hypocotyls of stocks (lower dark gray bars) are represented separately for CL. Values are mean + SD (n=5–7). n.s, non-significant difference between C and L (Student's t-test, P > 0.01). Effects of N (low N and high N) and CO₂ (ambient CO₂ and elevated CO₂) were evaluated by analysis of variance (ANOVA, **P < 0.01; *P < 0.05; ns, P > 0.05). Different lower-case letters indicate significant differences among CC, LC, CL and LL (Tukey’s test, P < 0.05).

Relative growth rate (RGR) was significantly higher at high N (HN-aCO₂ and HN-eCO₂) and at elevated CO₂ (LN-eCO₂ and HN-eCO₂) (Table S1). Among the four grafting combinations, LL with small hypocotyls showed the highest growth and RGR only at high N (HN-aCO₂ and HN-eCO₂), whereas CL showed the lowest growth and RGR under all growth conditions. Net assimilation rate (NAR) was higher at elevated CO₂ (LN-eCO₂ and HN-eCO₂), and the differences among the grafting combinations were very small. Leaf area ratio (LAR) was significantly higher at high N (HN-aCO₂ and HN-eCO₂) and lower at elevated CO₂ (LN-eCO₂ and HN-eCO₂). LAR was higher in CL and LL with small hypocotyls than CC and LC with large hypocotyls. Specific absorption rate of N (SAR) was significantly higher in CC and LC than in CL and LL under all growth conditions.

Accumulation of TNC and photosynthetic characteristics of source leaves

Reciprocal grafting with C and L caused marked changes in sink–source balance, as shown in the change in biomass allocation (Fig. 2), which led to differences in TNC on a leaf area basis among grafting combinations and growth conditions (Fig. 3A). TNC was significantly affected by both N and CO₂. TNC was lowest in CC with large hypocotyls under all growth conditions, and it was higher in LC and LL regardless of sink strength.

Fig. 3

TNC and photosynthetic characteristics of leaves of grafted radish plants. TNC, leaf nitrogen content per area (N_area) and Rubisco content were obtained 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively (A, C, D). Dark respiration rate (R_d), photosynthetic rate under growth conditions (A_growth) and maximum photosynthetic rate (A_max) were obtained 16–17 and 11–12 days after transfer to LN-aCO₂ and LN-eCO₂, and HN-aCO₂ and HN-eCO₂, respectively (B). White, gray and black bars represent A_max, A_growth and R_d, respectively, in (B). Values are mean + SD (n=5–7). Effects of N (low and high) and CO₂ (ambient and elevated) were evaluated by analysis of variance (ANOVA, **P < 0.01; *P < 0.05; ns, P > 0.05). Different lower-case letters indicate significant differences among CC, LC, CL and LL (Tukey’s test, P<0.05).

Photosynthetic characteristics of source leaves differed depending on growth conditions (Fig. 3B). Photosynthetic rate under growth conditions (A_growth) and maximum photosynthetic rate (A_max) were higher at high N than at low N at both ambient and elevated CO₂. The difference in A_max among the grafting combinations was small. Dark respiration rate (R_d) was lower at elevated CO₂, and was lowest in CC except at LN-eCO₂. Leaf nitrogen content per area (N_area) was significantly higher at high N (HN-aCO₂ and HN-eCO₂), while it was not affected by CO₂ levels (Fig. 3C). Differences in N_area and Rubisco content (Fig. 3D) were almost the same as those in A_max.

To check whether the changes in TNC and soluble sugars due to sink–source imbalance caused the down-regulation of photosynthesis, relationships among A_max, Rubisco content, TNC and soluble sugars were examined for each growth condition. Among the grafted plants, there were weak negative correlations between A_max and TNC at low N (LN-aCO₂ and LN-eCO₂) (Fig. 4A, B) but no correlations were found at high N (HN-aCO₂ and HN-eCO₂) (Fig. 4C, D). Irrespective of growth conditions and grafting combinations, Rubisco content was not correlated with TNC (Fig. 4E–H). A_max and Rubisco content were not correlated with soluble sugars under any of the growth conditions (Fig. S2), whereas there were negative correlations between stomatal conductance and TNC at elevated CO₂ (LN-eCO₂ and HN-eCO₂) (Fig. S3B, D).

Fig. 4

Relationships between maximum photosynthetic rate (A_max) and TNC (A–D) and between Rubisco content and TNC (E–H) for leaves of grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. Open circles, filled triangles, open triangles and filled circles represent values for CC, LC, CL and LL, respectively. Solid lines are regression lines. Values of R² are (A) 0.40, (B) 0.30, (C) 0.03, (d) 0.01 (E) 0.04, (F) 0.10, (G) 0.05 and (H) 0.00.

Among all the grafted plants, A_max was highly correlated with N_area (Fig. 5A, B) and Rubisco content (Fig. 5C, D) at both ambient CO₂ (LN-aCO₂ and HN-aCO₂) and elevated CO₂ (LN-eCO₂ and HN-eCO₂). There was also a strong correlation between Rubisco content and N_area among all the grafting combinations and growth conditions (Fig. 5E).

Fig. 5

Relationships between maximum photosynthetic rate (A_max) and leaf nitrogen content per area (N_area) (A, B), between A_max and Rubisco content (C, D) and between Rubisco content and N_area (E) for leaves of the grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. Values obtained for all growth conditions are shown in (E). Open circles, filled triangles, open triangles and filled circles represent values obtained from CC, LC, CL and LL, respectively. Solid lines are regression lines. Values of R² are (A) 0.72, (B) 0.68, (C) 0.66, (D) 0.73 and (E) 0.87.

TNC and N_mass in each organ

TNC was significantly higher at low N (LN-aCO₂ and LN-eCO₂) in all organs, and was higher at elevated CO₂ (LN-eCO₂ and HN-eCO₂) only in leaves (Fig. 6). Starch accounted for most of the TNC in leaves, whereas soluble sugars were major components of TNC in hypocotyls and fine roots. Among the grafted plants, TNC in leaves and that in petioles were lowest in CC (Fig. 6A, B). There was no clear tendency for TNC in hypocotyls and fine roots among grafting combinations (Fig. 6C, D). Both N and CO₂ affected N_mass significantly. N_mass was higher at high N (HN-aCO₂ and HN-eCO₂) and lower at elevated CO₂ (LN-eCO₂ and HN-eCO₂) in leaves, petioles and hypocotyls (Table S2). N_mass values in leaves and petioles were lowest in LL among the grafting combinations under all growth conditions.

Fig. 6

TNC in leaves (A), petioles (B), hypocotyls (C) and fine roots (D) in grafted plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. Bars from top to bottom represent values for glucose (white), fructose (light grey), sucrose (dark gray) and starch (black), respectively. Values are mean + SD (n=5–7). Effects of N (low and high) and CO₂ (ambient and elevated) were evaluated by analysis of variance (ANOVA; **P < 0.01; *P < 0.05; ns, P > 0.05). Different lower-case letters indicate significant differences among CC, LC, CL and LL (Tukey’s test, P < 0.05).

Morphological and anatomical traits in source leaves

Leaf mass per area (LMA), structural leaf mass per area (sLMA), cellulose mass per area (CMA) and C/N ratio were significantly higher at low N and elevated CO₂ (Fig. 7A–D). At low N (LN-aCO₂ and LN-eCO₂), LMA, sLMA, and CMA were lower in CC and LC with large hypocotyls than in CL and LL with small hypocotyls. On the other hand, these traits did not differ as much among the grafting combinations at high N (HN-aCO₂ and HN-eCO₂). Changes in C/N ratio in response to N and/or CO₂ were like those in TNC (Fig. 3A). Among all grafted plants, CMA was more correlated with sLMA (Fig. S4B) than with LMA (Fig. S4A).

Fig. 7

Leaf traits of grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. (A) Leaf mass per area (LMA), (B) structural leaf mass per area (sLMA), (C) cell wall mass per area (CMA) and D) C/N ratio. Values are means + SD (n=5–7). Effects of N (low and high) and CO₂ (ambient and elevated) were evaluated by analysis of variance (ANOVA; **P < 0.01; ns, P > 0.05). Different lower-case letters indicate significant differences among CC, LC, CL and LL (Tukey’s test, P < 0.05).

Leaf anatomical traits were analyzed using leaf transverse sections (Fig. S1). In leaves with high TNC, large chloroplasts with starch grains were clearly observed (Fig. S5). Leaf thickness was not affected by N or CO₂ levels (Fig. 8A), whereas leaf density was significantly affected by both N and CO₂ levels (Fig. 8B). The proportion of intercellular air space was significantly greater at high N (HN-aCO₂ and HN-eCO₂), while it was lower at elevated CO₂ (LN-eCO₂ and HN-eCO₂) (Fig. 8C). Differences in the proportion of intercellular air space among grafting combinations were small. The size of mesophyll cells was also affected by N, and it was smallest in CL under all growth conditions (Fig. 8D).

Fig. 8

Leaf anatomical traits of grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. (A) Leaf thickness, (B) leaf density, (C) intercellular airspace and (D) mesophyll cell size. Values are mean + SD (n=5–7). Effects of N (low and high) and CO₂ (ambient and elevated) were evaluated by analysis of variance (ANOVA; **P < 0.01; *P < 0.05; ns, P > 0.05). Different lower-case letters indicate significant differences among CC, LC, CL and LL (Tukey’s test, P < 0.05).

Relationships between C/N ratio, TNC and CMA

There were strong correlations between TNC and C/N ratio and between CMA and TNC of leaves. The relationships between TNC and C/N ratio were consistent irrespective of the type of graft (Fig. 9A). The relationships between CMA and TNC were significantly different between at low N (LN-aCO₂ and LN-eCO₂) (Fig. 9B) and at high N (HN-aCO₂ and HN-eCO₂) (Fig. 9C). The regression lines for CMA–TNC relationships for CL and LL lay above those for CC and LC. In other words, for any TNC level CMA was significantly lower in CC and LC with large hypocotyls than in CL and LL with small hypocotyls.

Fig. 9

Relationships between TNC and C/N ratio under all growth conditions (A) and those between cell wall mass per area (CMA) and TNC at low N (B) and at high N (C) for leaves of grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. Open circles, filled triangles, open triangles and filled circles represent values for CC, LC, CL and LL, respectively. Solid lines are regression lines for all plants (A) and for CC, LC, CL and LL (B, C). The value of R² is 0.86 (A). Values of slope, intercept, R² and results of statistical analysis for each line (B, C) are shown in Table S5.

Leaf intercellular air space was negatively correlated with both C/N ratio and TNC in all grafted plants (Fig. S6A, B), whereas mesophyll cell size was less correlated with them (Fig. S6C, D). However, mesophyll cell size was positively correlated with leaf intercellular air space in all grafted plants (Fig. S6E). Among grafted plants, CL showed a relationship that was significantly different from others.

Relationships between leaf traits and petiole anatomy

Anatomical traits of the petiole—transverse-sectional area of petiole (S_pet+), transverse-sectional area of vascular bundle (S_vb), transverse-sectional area of xylem (S_xy) and transverse-sectional area of phloem (S_ph)—differed depending on both growth conditions and grafting combinations (Figs. S7, S8). These values, expressing the size and vascular functions of the petiole, were greater at high N (HN-aCO₂ and HN-eCO₂) and highest in LL among the grafting combinations (Table S3). On the other hand, S_pet, S_vb, S_xy, and S_ph per leaf area were higher at low N or elevated CO₂. At low N (LN-aCO₂ and LN-eCO₂), these were higher in CL and LL than in CC and LC.

Single leaf area, single leaf mass, and single petiole mass were also much greater at high N (HN-aCO₂ and HN-eCO₂) (Fig. S9). Among grafted plants, these were greater in CL and LL with small hypocotyls than in CC and LC with large hypocotyls. Single leaf-to-petiole mass ratio was not affected by N or CO₂ level (Table S4). However, single leaf area-to-petiole mass ratio was much higher at high N (HN-aCO₂ and HN-eCO₂), and was lower in CL and LL with small hypocotyls than in CC and LC with large hypocotyls.

We also found close relationships between petiole anatomical traits and TNC. Petiole mass per leaf area (Fig. 10A), S_pet per leaf area (Fig. 10B), S_vb per leaf area (Fig. S10A), S_xy per leaf area (Fig. S10B), and S_ph per leaf area (Fig. S10C) were all positively correlated with TNC in the leaves. The regression lines of petiole mass per leaf area, S_vb per leaf area and S_ph per leaf area were significantly different among grafting combinations, whereas S_pet per leaf area or S_xy per leaf area were not (Table S5).

Fig. 10

Relationship between single petiole mass per leaf area (A), transverse-sectional area of petiole (S_pet) per leaf area (B) and TNC in leaves of grafted radish plants 18 days after transfer to LN-aCO₂ and LN-eCO₂ and 13 days after transfer to HN-aCO₂ and HN-eCO₂, respectively. Values for all growth conditions are shown. Open circles, filled triangles, open triangles and filled circles represent values for CC, LC, CL and LL, respectively. Solid lines are regression lines for CC, LC, CL and LL. Values of slope, intercept, R² and results of statistical analysis for each line are shown in Table S5.

Discussion

Effect of accumulation of TNC on down-regulation of photosynthesis is minor in radish plants

The results support our hypothesis that the accumulation of TNC does not cause or only slightly causes the down-regulation of photosynthesis in radish plants. As hypothesized, this is partly due to the high capacity of the source leaves for storage of TNC (Fig. 3A). Large chloroplasts with starch grains were clearly observed in source leaves, especially in the mesophyll cells of LC and LL plants at low N (Fig. S5), which reflected higher TNC levels in them even when a large sink was grafted on. This may be due to a difference in TNC storage capacity between L and C varieties.

We found negative relationships between A_max and TNC only at LN-aCO₂ and LN-eCO₂ (Fig. 4A, B). Contrary to previous findings (Krapp and Stitt 1995, Nakano et al. 1998), Rubisco content did not decrease with increasing TNC (Fig. 4E, F) or increasing soluble sugars (Fig. S2E, F). The selective decrease in Rubisco content that was observed in certain plant species (Sage et al. 1989, Nakano et al. 1998) was not observed either, since the ratio of Rubisco content to N_area was highly constant under all growth conditions across the grafting combinations (Fig. 5E). As there were strong correlations between A_max, N_area and Rubisco under all growth conditions (Fig. 5A–D), the observed slight decrease in A_max (Fig. 3B) was largely due to N deficiency and partly due to decreases in stomatal conductance (Fig. S3B, D). These results strongly support the idea that the down-regulation of photosynthesis caused by TNC does not occur in R. sativus.

The present results are consistent with some previous studies showing that the down-regulation of photosynthesis would not occur even at elevated CO₂ as long as sufficient N is supplied (Theobald et al. 1998, Stitt and Krapp 1999, Sanz-Sáez et al. 2010). Several studies, including our previous one using several cultivars of radish plants, also indicated that neither the accumulation of TNC nor elevated CO₂ caused the down-regulation of photosynthesis (Usuda and Shimogawara 1998, Usuda 2004, Choi et al. 2011, Sugiura et al. 2015).

The present results confirm the marked interspecific differences in the response of photosynthetic traits to TNC among plant species. For example, Spinacia oleracea (Krapp et al. 1991), Chenopodium album, Solanum tuberosum (Krapp et al. 1993) and Phaseolus vulgaris (Araya et al. 2006) were considered to be sugar-sensitive species, decreasing Rubisco levels and transcript levels of photosynthetic genes such as rbcS and Rubisco content in response to sugar feeding or cold-girdling. These photosynthetic traits responded less to the accumulation of starch and sugars, which was caused by phosphorous deficiency in tobacco (Paul and Stitt 1993). Paul and Driscoll (1997) also showed that the increase in hexose content, which had been thought to be the main player in sugar repression, was only modest in tobacco. It is suggested that systemic signals may be involved in the regulation of leaf photosynthetic capacity. Sims et al. (1998) showed that, in soybean, even if individual leaves were exposed to low CO₂ and the TNC level was low, down-regulation of photosynthesis occurred when the whole plant was exposed to elevated CO₂. Another important viewpoint is that leaf developmental stage is also involved in the extent of down-regulation. Krapp et al. (1991) and Araya et al. (2006) showed that the extent of photosynthetic down-regulation with sugar feeding was prominent in mature source leaves compared with young sink leaves. To clarify the causality and mechanisms underlying interspecific differences in the sugar repression of photosynthesis, therefore, it is necessary to perform further analysis considering these points in combination with analysis of metabolomes and transcriptomes at the whole-plant level.

Effects of hypocotyl sink strength, N availability and CO₂ conditions on whole-plant growth

L/R, L_A/R and LMR increased with N availability, whereas these traits did not change markedly with CO₂ (Table 1). These are typical responses of plants to N and CO₂ (Poorter et al. 2000, Poorter et al. 2012). On the other hand, interestingly, HMR was lower at high N (HN-aCO₂ and HN-eCO₂). We showed that higher allocation of photosynthates to hypocotyls is promoted at low N, which supports the hypothesis that excess TNC can be allocated more from leaves to other organs.

Since there is a trade-off between growth and storage, we expected that grafted plants with large hypocotyls would show a lower RGR due to the decrease in LAR (Chapin et al. 1990) at both high and low N. However, at low N, CL and LL with small hypocotyls showed lower total dry mass and RGR than CC and LC (Fig. 2 and Table S1). This was partly because CL and LL resumed growth later than CC and LC plants after nursery treatment (Sugiura et al. 2015). Lower nitrogen absorption rate of roots (SAR) in CL and LL than CC and LC would be responsible for the lower NAR in these plants. Although CL showed the highest A_max (Fig. 3B), N_area (Fig. 3C) and Rubisco (Fig. 3D) at low N (LN-aCO₂ and LN-eCO₂), this would be realized by translocation of nitrogenous compounds from the old leaves to the new leaves that were used for the photosynthesis measurements. Rapid aging of the old leaves is apparent from the photographs of CL and LL (Fig. 1). The decrease in LAR caused by accumulation of TNC (Fig. 3A) and structural carbohydrates such as cell wall materials (Fig. 7C) also lowered RGR in CL and LL plants with small hypocotyls (Table S1). Thus, the lower sink strength would not necessarily increase whole-plant growth at low N (LN-aCO₂ and LN-eCO₂). On the other hand, higher LMR and lower LMA led to the highest RGR in LL at high N (HN-aCO₂ and HN-eCO₂). The present experiment showed clear effects of the trade-off between growth and storage at high N.

Capacity to store starch and soluble sugars is prominent in all organs under low N and elevated CO₂ conditions

TNC in the leaves was lower in the grafted plants with large hypocotyls at low N, whereas it was still higher in L than in C leaves regardless of hypocotyl size, even at high N (Figs. 3A, 6A). It is well known that TNC accumulates in the leaves in response to manipulations such as fruit removal (Bertin et al. 1999), girdling (Urban and Alphonsout 2007) and cold-girdling (Krapp and Stitt 1995). TNC increased especially at LN-eCO₂, but the extent was highly variable among species and cultivars (Wong 1990, Ahmed et al. 1993, Poorter et al. 1997, Scheible et al. 1997, Ainsworth et al. 2004). The present results show that large amounts of TNC can be accumulated in every organ in grafted radish plants (Fig. 6). It is possible that the capacity for storage of TNC in organs other than leaves is also related to interspecific differences in the down-regulation of photosynthesis since it can alleviate the accumulation of TNC in the source leaves. Therefore, even when investigating leaf traits, it is important to consider whole-plant growth.

Coordinated changes in leaf morphological traits with C/N ratio and TNC

We found closely related changes in C/N ratio, TNC and structural carbohydrates such as cell wall materials (Fig. 7). It is well known that low N and elevated CO₂ conditions lead to low leaf N, which further causes accumulation of TNC, whereas only a few studies reported effects of C/N and/or TNC on structural carbohydrates in leaves. Previous studies showed that CMA slightly increased in Plantago major at low N (Onoda et al. 2008), and cell wall thickness increased by 20% at elevated CO₂ in Arabidopsis thaliana (Teng et al. 2006) and rice (Zhu et al. 2012), which is consistent with the present results.

CMA correlated more strongly with sLMA than LMA. sLMA mostly consists of structural carbohydrates, where cell wall mass account for 25% of sLMA (Fig. S4). Therefore, leaves with a larger amount of cell wall also contained other structural carbohydrates, such as pectin. Teng et al. (2006) reported that not only cellulose but also pectin increased in A. thaliana at elevated CO₂.

Although these structural components increased by up to 20% in preceding studies, CMA increased 4-fold in the present study. This would be explained by the drastic changes in leaf anatomy and cell wall thickness in the leaves of grafted radish plants. Since leaf thickness was almost the same among the grafting combinations under all growth conditions (Fig. 8A), changes in the proportion of intercellular air space (Fig. 8C), mesophyll cell size (Fig. 8D) and probably cell wall thickness would contribute to the increase in CMA, especially at LN-eCO₂.

The cell wall as a sink for excess carbohydrate

Consistent with our previous study, accumulation of TNC was dependent on changes in C/N ratio (Fig. 9A), which were induced by changes in sink–source ratio. In the present study, we found that accumulation of TNC due to the increase in C/N ratio further caused an increase in CMA (Fig. 9b, C) and a decrease in intercellular air space (Fig. S6A, B), which suggests the importance of the cell wall as a carbohydrate sink, especially at LN-eCO₂ (Fig. 7C). We also found that CMA increased less in CC and LC plants than in CL and LL plants, which indicates that the existence of large storage organs alleviates not only the deposition of TNC but also that of structural carbohydrates. These findings support the hypothesis that excess TNC is partly converted to cell wall materials.

The increase in cell wall mass and coordinated decreases in mesophyll cell size and the proportion of intercellular air space (Fig. S6E) could influence the diffusivity of CO₂ from the substomatal cavities to sites of carboxylation through the cell wall, which is called mesophyll conductance (Terashima et al. 2006, Evans et al. 2009, Terashima et al. 2011). Therefore, differences in mesophyll conductance would also contribute to the difference in A_max among growth conditions, since CMA and leaf anatomy differed markedly among the grafting combinations. Accordingly, it is possible that these observed changes would be effective in alleviating the excessive accumulation of TNC by decreasing mesophyll conductance and consequently photosynthetic rate. Although the down-regulation of photosynthesis has been investigated from the viewpoint of biochemistry, our present results suggest that CO₂ assimilation rate could be regulated through anatomical changes in response to changes in sink–source ratio. Since excess amounts of photosynthetic products could lead to photoinhibition by damaging the photosynthetic apparatus (Roden and Ball 1996, Urban and Alphonsout 2007), the ecological significance of the down-regulation of photosynthesis is possibly that it protects the photosynthetic apparatus from TNC (Demmig-Adams et al. 2014).

Petioles are an important carbohydrate sink

While there have been many studies on changes in biomass allocation to leaves, stems and roots, there are few studies on changes in biomass allocation to petioles in rosette plants. Meta-analysis studies revealed that stem mass ratio does not change greatly in response to N or CO₂ conditions (Poorter et al. 2000, Poorter et al. 2012). This idea may apply to the petioles of our studied plants since the ratio of single leaf mass to single petiole mass did not change greatly among the grafting combinations under a given growth condition (Table S4). On the other hand, the lower ratio of leaf area to petiole mass at low N suggests that the petiole biomass was in excess with regard to leaf water demand, which depends on leaf area. The higher S_pet, S_vb and S_xy per leaf area at LN-eCO₂ support this idea (Table S2). The higher S_ph per leaf area further suggests that sugar transport capacity of the phloem was also enhanced, which would also contribute to the alleviation of excess TNC in the leaves, especially at LN-eCO₂. Furthermore, these parameters increased coordinately with the increase in TNC (Figs. 10, S11). It is also interesting that single petiole mass per leaf area (Fig. 10A), S_vb per leaf area (Fig. S10A) and S_ph per leaf area (Fig. S10C) increased significantly more with increasing TNC in CL and LL plants with small sink organs than CC and LC plants with large sink organs (Table S5). Therefore, it is possible that the activity of the vascular cambium was promoted in response to accumulation of TNC, and its activity was also affected by the existence of large sink organs. As we expected, it was revealed that petioles are also important sink organs for excess TNC, especially under low N and elevated CO₂ conditions, which can cause lower sink–source ratios.

Conclusion

Our present study demonstrates that the effects of TNC on the down-regulation of photosynthesis were minor in R. sativus, since only a slight decrease in A_max and no decrease in Rubisco content with increasing TNC were observed, which is different from other plant species. On the other hand, TNC levels in respective organs and morphological traits of source leaves were both greatly affected by the sink–source balance. It is quite notable that radish plants can store large amounts of TNC in every organ; the proportion of TNC was up to 35% of dry mass. We also found that not only hypocotyls but also petioles and source leaves themselves are important carbon sinks. Furthermore, the observed changes in morphological and anatomical traits, such as CMA, leaf intercellular air space and petiole thickening, with increasing TNC suggest that radish plants have an ability to consume excess TNC accumulated in response to sink–source imbalance in a structural way. This alleviates further accumulation of TNC and its negative effects on the photosynthetic apparatus in radish plants. Therefore, we can further hypothesize that the capacity to store TNC at the whole-plant level, the sink strength of each organ in importing and/or consuming excessive TNC, and the ability to reduce carbon assimilation by changing leaf anatomy would be closely involved in the interspecific differences in the down-regulation of photosynthesis caused by TNC.

Materials and Methods

Plant materials

We used two varieties of radish, Raphanus sativus var. radicula ‘Comet’ and R. sativus var. longipinnatus ‘Leafy’ (Hadaikon, a direct translation would be ‘leafy radish’), which are popular commercial cultivars in Japan. Comet and Leafy varieties are hereafter called C and L, respectively. Seeds of C (Takii seed Co., Ltd., Japan) and L (Atariya Noen Co., Ltd., Japan) were sown in autoclaved vermiculite in 200-ml plastic pots. They were grown in a controlled growth room, where photosynthetically active photon flux density (PFD) was about 200 μmol m⁻² s⁻¹during the14-h light period, mean air temperature was 25 °C, and relative humidity was 80%. After germination, a nutrient solution (N:P:K, 6:10:5, Hyponex Japan, Osaka, Japan) containing 8 mM N and other essential nutrients was applied at 1/500 strength until grafting.

Reciprocal grafting

Five each of pre-grafted C and L were harvested 12–14 days after sowing and used for growth analysis (see below). On the same day, seedlings of C and L with two cotyledons and first and second true leaves were used for grafting following Sugiura et al. (2015). Briefly, scions of C were grafted on stocks of both C and L, and scions of L were also grafted on stocks of C and L, and they were named CC, CL, LC and LL in the order of scion and stock. After grafting, they were nursed in plastic containers where air humidity was almost 100% and PFD was <100 μmol m⁻² s⁻¹ for 11−12 days. We considered that grafting was completed when guttation was observed at the leaf edge. After the nursery treatment, successfully-grafted plants were transplanted into 500-ml plastic pots and used for experiments.

Growth conditions

The grafted plants were grown under four constant growth conditions in growth chambers: low N and ambient CO₂ (LN-aCO₂), low N and elevated CO₂ (LN-eCO₂), high N and ambient CO₂ (HN-aCO₂) and high N and elevated CO₂ (HN-eCO₂). Nutrient solutions containing 16 mM and 2 mM N were used at 1/250 and 1/2000 strengths for high N and low N treatments, respectively. We used two CO₂-controlled growth chambers (LPH-0.5PSH; Nippon Medical & Chemical Instrument) where PFD was 250 μmol m⁻² s⁻¹ (measured with a quantum sensor, LI-1000; Li-Cor, Lincoln, NE, USA) during the14-h light period. In each chamber, air temperature and relative humidity were set to 25 °C and 60%, respectively, and ambient and elevated CO₂ concentrations were set to 400 and 800 ppm, respectively. These parameters were recorded on each of two thermo-hygro-CO₂ meters in the chambers throughout the growth period (TR-76Ui; T&D Corporation, Nagano, Japan).

Gas exchange measurements and harvests

At 16−17 and 11–12 days after transfer to growth chambers for LN-aCO₂ and LN-eCO₂ plants and HN-aCO₂and HN-eCO₂ plants, respectively, photosynthetic characteristics of the mature leaves were determined with a portable gas-exchange system (LI-6400; Li-Cor, Lincoln, NE, USA). The measurements were conducted in five to seven leaves for each grafting combination. The rate of photosynthesis under growth conditions (A_growth, µmol m⁻² s⁻¹) was measured at a PFD of 250 µmol m⁻² s⁻¹ in 400 ppm CO₂ for the LN-aCO₂ and HN-aCO₂ plants, and in 800 ppm CO₂ for LN-eCO₂ and HN-eCO₂ plants. Maximum photosynthetic rates (A_max, µmol m⁻² s⁻¹) were measured at a PFD of 1000 µmol m⁻² s⁻¹ and 400 ppm CO₂ for LN-aCO₂ and HN-aCO₂ plants and in 800 ppm CO₂ for LN-eCO₂ and HN-eCO₂ plants. Dark respiration rate (R_d) was also measured. The leaves used for photosynthesis measurements were also used for measurements of leaf thickness, TNC and Rubisco contents (see below).

At 18 and 13 days after transfer to growth chambers for LN-aCO₂ and LN-eCO₂ plants and for HN-aCO₂ and HN-eCO₂ plants, respectively, five to seven plants were harvested for each grafting combination. At the end of the dark period, six leaf discs (1.22 cm in diameter) were sampled from the one or two mature leaves that were used for gas exchange measurements. Three leaf discs were immediately oven-dried at 80 °C to determine LMA (leaf mass per area, g m⁻²), leaf N content and TNC. The other three leaf discs were immediately frozen in liquid N₂ and stored at −80 °C to determine the Rubisco content and cell wall content (see below).

The leaves used for these measurements were divided into lamina and petiole to evaluate biomass allocation between the leaves and petioles. The rest of the plants were further divided into leaves, petioles, hypocotyls of the scions, those of the stocks, and fine roots. They were oven-dried at 80 °C for >3 days to determine their dry weights.

Morphological and physiological traits

Oven-dried leaf discs obtained from the leaves used for photosynthesis measurements were weighed to determine LMA. Leaf area of each single leaf used for photosynthesis measurements was determined by dividing the dry mass of each leaf by its LMA. Dried leaves used for the photosynthesis measurements, the rest of the leaves, petioles, hypocotyls and fine roots were weighed, and leaf-to-root mass ratio (L/R, g g⁻¹), leaf area-to-root mass ratio (L_A/R, m² g⁻¹), leaf area ratio (LAR, leaf area to total dry mass, m² g⁻¹), leaf mass ratio (LMR, leaf dry mass to total dry mass, g g⁻¹), petiole mass ratio (PMR, petiole dry mass to total dry mass, g g⁻¹) and hypocotyl mass ratio (HMR, hypocotyl mass to total dry mass, g g⁻¹) were calculated. Ratios of single leaf mass to petiole mass (g g⁻¹) and single leaf area to petiole mass (mm² mg⁻¹) were also calculated.

All the dried samples were ground with a Multi-beads Shocker (Yasui Kikai, Osaka, Japan). The C and N contents of the ground samples were measured with a CN analyzer (Vario EL, Elementar Analyzensysteme GmbH, Hanau, Germany). The N content per mass (N_mass, g N g⁻¹) and C content per mass (C_mass, g C g⁻¹) were calculated for each of the single leaves used for the measurements, the rest of the leaves, hypocotyls and fine roots. Leaf N content per area (N_area, g N m⁻²) and leaf C content per area (C_area, g C m⁻²) of the leaves used for photosynthesis measurements were calculated as a product of N_mass and LMA, and C_mass and LMA, respectively. The C/N ratio (g C g N⁻¹) of these leaves was calculated considering the amount of C derived from TNC:

$$C/N=\frac{C_{area}-0.4\times TNC}{N_{area}},$$

(1)

where 0.4 is the concentration of carbon based on the assumption that TNC consists solely of C₆H₁₂O₆, and TNC is expressed on a leaf area basis (g m⁻², see below). The aim of this calculation was to exclude the effect of the increase in TNC that accompanies the increase in C/N ratio.

Net assimilation rate (NAR, g m⁻² d⁻¹), specific absorption rate of N by roots (SAR, g N g⁻¹ d⁻¹) and relative growth rate (RGR, g g⁻¹ d⁻¹) were determined between harvests of the pre-grafted and grafted plants:

$$NAR=\frac{\left(W_2-W_1\right)\left(\text{ln}{L}_{A2}-\text{ln}{L}_{A1}\right)}{\left(L_{A2}-L_{A1}\right)\left(t_2-t_1\right)},$$

(2)

$$SAR=\frac{\left(N_2-N_1\right)\left(\text{ln}{R}_2-\text{ln}{R}_1\right)}{\left(R_2-R_1\right)\left(t_2-t_1\right)},$$

(3)

and

$$RGR=\frac{\text{ln}{W}_2-\text{ln}{W}_1}{t_2-t_1},$$

(4)

where W, L_A, N and R are total dry mass (g), leaf area (m²), total N content (g N) and fine root biomass (g) at the harvest dates of the grafted plants (t₂) and pre-grafted plants (t₁).

Leaf transverse section

From each of the leaves used for the photosynthesis measurements, small lamina segments and petiole slices were excised and fixed in FAA (5% formaldehyde, 5% acetic acid and 45% ethanol). After fixation, they were dehydrated in a series of ethanol solutions and embedded in Technovit 7100 (Heraus Holding, Hanau, Germany). Leaf transverse sections of 1 µm thickness and petiole transverse sections of 2 µm thickness were cut with an ultramicrotome and stained with a solution containing 0.1% (w/v) toluidine blue and 1% (w/v) sodium borate.

Anatomical traits of leaves and petioles were analyzed using Image-J software (National Institutes of Health freeware) (Fig. S1). Leaf thickness (T), proportion of intercellular air space (P_int) and average mesophyll cell size (M) were calculated as:

$$T=\frac{S_{tot}}{W},$$

(5)

$$P_{\text{int}}=\frac{S_{\text{int}}}{S_{\text{mes}}+S_{\text{int}}},$$

(6)

$$M=\frac{S_{mes}}{N_{mes}},$$

(7)

where S_tot, S_mes, S_int and N_mes are transverse-sectional areas of total leaf, mesophyll cells and intercellular air space, and number of mesophyll cells. Transverse-sectional areas of petiole (S_pet), vascular bundles (S_vb), xylem (S_xy) and phloem (S_ph) were also measured to evaluate anatomical changes of petioles in the grafted plants.

Total non-structural carbohydrates and structural LMA

Ground samples of leaves used for the photosynthesis measurements; the rest of the leaves, petioles, hypocotyls and the fine roots were used for the determinations of glucose, sucrose, fructose and starch contents. Soluble sugars were extracted from about 5–10 mg of ground samples with ethanol and dissolved in distilled water to determine glucose and sucrose contents, and the precipitate was used for the determination of starch as described in Sugiura et al. (2015). The fructose content was determined using the water solution containing soluble sugars (Hachiya et al. 2014). The solutions obtained from leaves, petioles and roots were diluted 10 times and those obtained from hypocotyls were diluted 20 times with distilled water; 0.2 ml of the diluted solutions were mixed with 0.2 ml of 0.1% resorcinol and 0.6 ml of 30% HCl, and they were heated at 80°C for 20 min. After cooling, absorbance was read at 540 nm. Then, TNC was calculated on both leaf area and mass bases.

Structural LMA (sLMA, g m⁻²) was calculated by subtracting TNC from LMA (Bertin et al. 1999, Sugiura et al. 2015). Leaf density was calculated by dividing LMA by leaf thickness.

Rubisco content and cell wall content

Rubisco was extracted from three frozen leaf discs and the Rubisco content on an area basis was determined by SDS–PAGE (Makino et al. 1986). The gel was stained with Coomassie Brilliant Blue (CBB), and the band of the Rubisco large subunit was cut out from the gel. CBB was extracted from the gel segment with formamide, and absorbance of the extract was determined spectrophotometrically. The residual pellet after extraction was used for the determination of cell wall content, defined as the sum of cellulose and hemi-cellulose, following Onoda et al. (2008) with some modifications. The pellet was washed with 1 ml of 20 mM HEPES buffer (pH 7.0) containing 1% SDS, heated at 95 °C for 10 min. then centrifuged (12,000g for 10 min). The pellet was heated with RO water at 98 °C for 1 h and treated with amyloglucosidase (Sigma Aldrich, A-9228) in 50 mM Na-acetate buffer (pH 4.5) at 55 °C for 1 h in order to remove starch. The pellet was washed with 1 M NaCl to remove cytoplasmic proteins. The remaining cell wall material was washed with RO water and ethanol and oven-dried at 80 °C for >1 day. After drying, the pellets were weighed and cell wall mass per leaf area (CMA, g m⁻²) was determined.

Statistical analysis

Statistical tests were performed with Systat13 (Hulinks Inc., Tokyo, Japan). Total dry mass was compared between the ungrafted combinations with Student’s t-test. The effects of N (low and high) and CO₂ (ambient and elevated) on total dry mass and morphological and physiological traits were evaluated by analysis of variance (ANOVA). The traits were also compared among CC, LC, CL and LL by multiple pairwise Tukey’s test comparisons.

Differences in the regression lines for the relationships between single petiole mass per leaf area, S_pet per leaf area, S_vb per leaf area, S_xy per leaf area, S_ph per leaf area and TNC were tested by analysis of covariance (ANCOVA) followed by multiple comparison (Tukey’s test, P < 0.05). Relationships between CMA and TNC and those between mesophyll cell size and the proportion of intercellular air space were also tested by ANCOVA followed by multiple comparison (Tukey’s test, P < 0.05).

Funding

CREST (Creation of essential technologies to utilize carbon dioxide as a resource through the enhancement of plant productivity and the exploitation of plant products to I.T.); Grant-in-Aid for JSPS Fellows (No. 14J07443 to D.S.).

Supplementary Data

Supplementary data are available at PCP Online.

Acknowledgements

We greatly thank the members of Laboratory of Plant Ecology of The University of Tokyo for valuable comments on an early draft and for technical assistance.

Abbreviations

Abbreviations
 
• C

  ungrafted Comet plants

 
• L

  ungrafted Leafy plants

 
• CC

  grafted plants: scions of C grafted on stocks of C

 
• CL

  grafted plants: scions of C grafted on stocks of L

 
• LC

  grafted plants: scions of L grafted on stocks of C

 
• LL

  grafted plants: scions of L grafted on stocks of L

 
• LN-aCO₂

  low nitrogen and ambient CO₂ condition

 
• LN-eCO₂

  low nitrogen and elevated CO₂ condition

 
• HN-aCO₂

  high nitrogen and ambient CO₂ condition

 
• HN-eCO₂

  high nitrogen and elevated CO₂ condition

 
• L/R

  leaf-to-root ratio (g g⁻¹)

 
• L_A/R

  leaf area-to-root ratio (m² g⁻¹)

 
• LMR

  leaf mass ratio (g g⁻¹)

 
• PMR

  petiole mass ratio (g g⁻¹)

 
• HMR

  hypocotyl mass ratio (g g⁻¹)

 
• RGR

  relative growth rate (g g⁻¹ d⁻¹)

 
• NAR

  net assimilation rate (g m⁻² d⁻¹)

 
• LAR

  leaf area ratio (m² g⁻¹)

 
• SAR

  specific absorption rate of nitrogen (g N g⁻¹ d⁻¹)

 
• R_d

  dark respiration rate (μmol m⁻² s⁻¹)

 
• A_growth

  photosynthetic rate under growth conditions (μmol m⁻² s⁻¹)

 
• A_max

  maximum photosynthetic rate (μmol m⁻² s⁻¹)

 
• PFD

  photon flux density

 
• TNC

  total nonstructural carbohydrate (mg g⁻¹ or g m⁻²)

 
• LMA

  leaf mass per area (g m⁻²)

 
• sLMA

  structural leaf mass per area (g m⁻²)

 
• CMA

  cellulose mass per area (g m⁻²)

 
• C_area

  leaf carbon content per area (g N m^–2)

 
• C_mass

  carbon content per mass (g N g⁻¹)

 
• N_area

  leaf nitrogen content per area (g N m⁻²)

 
• N_mass

  nitrogen content per mass (g N g⁻¹)

 
• C/N ratio

  carbon–nitrogen ratio (g C g N⁻¹)

 
• T

  leaf thickness (μm)

 
• P_int

  proportion of intercellular air space

 
• M

  mesophyll cell size (μm²)

 
• S_tot

  transverse-sectional area of total leaf (μm²)

 
• S_mes

  transverse-sectional area of mesophyll cell (μm²)

 
• S_int

  transverse-sectional area of intercellular air space (μm²)

 
• N_mes

  number of mesophyll cells

 
• S_pet

  transverse-sectional area of petiole (mm²)

 
• S_vb

  transverse-sectional area of vascular bundle (mm²)

 
• S_xy

  transverse-sectional area of xylem (mm²)

 
• S_ph

  transverse-sectional area of phloem (mm²)