Endophytes alleviate the elevated CO2-dependent decrease in photosynthesis in rice, particularly under nitrogen limitation.

The positive effects of high atmospheric CO2 concentrations [CO2] decrease over time in most C3 plants because of down-regulation of photosynthesis. A notable exception to this trend is plants hosting N-fixing bacteria. The decrease in photosynthetic capacity associated with an extended exposure to high [CO2] was therefore studied in non-nodulating rice that can establish endophytic interactions. Rice plants were inoculated with diazotrophic endophytes isolated from the Salicaceae and CO2 response curves of photosynthesis were determined in the absence or presence of endophytes at the panicle initiation stage. Non-inoculated plants grown under elevated [CO2] showed a down-regulation of photosynthesis compared to those grown under ambient [CO2]. In contrast, the endophyte-inoculated plants did not show a decrease in photosynthesis associated with high [CO2], and they exhibited higher photosynthetic electron transport and mesophyll conductance rates than non-inoculated plants under high [CO2]. The endophyte-dependent alleviation of decreases in photosynthesis under high [CO2] led to an increase in water-use efficiency. These effects were most pronounced when the N supply was limited. The results suggest that inoculation with N-fixing endophytes could be an effective means of improving plant growth under high [CO2] by alleviating N limitations.

Introduction

Climate change poses a complex set of problems that cannot be addressed by single countries or individuals alone. We are now seeing compelling evidence of multifarious environmental changes, just as climatologists have predicted (IPCC, 2013). The main driver of global climate change is an increase in atmospheric concentrations of carbon dioxide, [CO2], 13.5% of which are attributed to current unsustainable agricultural practices, such as the use of excessive amounts of chemical fertilizers and resource-intensive farming techniques (IPCC, 2007). Manufacturing nitrogen (N) fertilizers through the standard Haber–Bosch process emits greenhouse gases into the atmosphere, which negatively affect environmental processes by exacerbating global warming (Vance, 2001). Moreover, excessive applications of N fertilizer have disrupted cropland ecological N cycles, further disturbing the nutrient balances of rivers and oceans in a process known as eutrophication, which has become a serious environmental concern (Mueller ).

A large body of literature exists from empirical studies that report that increases in [CO2] positively affect the performance of C3 crop species. When this effect was initially discovered and documented, it was accompanied by an optimistic projection that CO2 enrichment would increase crop biomass (Drake ), a response that was often referred to as a CO2 fertilization effect (Thornton ). However, overall increases in C3 plant biomass have fallen short of what was previously anticipated in theoretical models (Moore ; Long ). This is partially attributed to down-regulation of photosynthesis, which has been observed in at least 25 C3 species studied in 227 free-air CO2 enrichment (FACE) experiments (Ainsworth and Long, 2005). This down-regulation is expected to partially offset the positive effects of CO2 fertilization (Long ) when plants are exposed to a long-term elevated [CO2] (Long ). There is now a consensus around the inability of CO2 enrichment to provide a solution to food security issues in many regions, and it appears that a serious deficit in food production to support an increasing human population will remain a global challenge (Tilman ).

Given that many experimental approaches have demonstrated that the down-regulation of C3 photosynthesis is heavily dependent on N supplies and subsequent N metabolism (Stitt and Krapp, 1999; Reich ; Shimono and Bunce, 2009; Ruiz-Vera ), the impacts of elevated [CO2] on crop yield then become a subject that is closely associated with N acquisition from the rhizosphere and N-use efficiency in plants (Ainsworth and Rogers, 2007; Zhu ). The mechanistic explanation for this is that the turnover and activities of Rubisco and other enzymes associated with the Calvin–Benson cycle are decreased because excess glucose production acts as a transduction signal (Stitt and Krapp, 1999). Under elevated [CO2], the production of photosynthates by C3 plants exceeds their sink capacity, which in turn blocks the transcriptional or translational control of the Rubisco small subunit, resulting in decreases in Rubisco and in photosynthetic capacity (Ainsworth and Rogers, 2007). As Rubisco constitutes almost 50% of the total soluble N in leaves and its turnover is N intensive, this process is accelerated when N is limited (Gamage ).

In contrast, legumes are reported to be a group of the C3 plants that are unlikely to experience significant photosynthetic down-regulation associated with a long-term exposure to high [CO2] (Ainsworth and Rogers, 2007; Rogers ). The alleviation of down-regulation in legumes can be attributed to their capability to reset the balance of C and N metabolic processes in a way that stimulates biological N fixation (BNF) under nutrient-limited conditions (Ainsworth and Rogers, 2007; Rogers ).

Increases in N supply due to nodule-forming rhizobium bacteria are also known to help increase the carbon sink strength in legumes under elevated [CO2] conditions, thus contributing to the alleviation of down-regulation of photosynthesis (Ainsworth and Rogers, 2007). With a sufficient N pool and plentiful carbohydrates from increased photosynthesis, the plants can produce more sink tissue and thus reduce carbohydrate levels in the phloem. This in turn stimulates an increase in the rate of carbohydrate export in the source tissues and hence allows them to carry out more photosynthesis (Ainsworth and Bush, 2011). In addition, the symbiotic nodulating bacteria serve as an active ‘biological sink’ (Ainsworth ), costing their host plants about 14–25% of daily photosynthates (Lambers ; Kaschuk ).

Endophytes are also microbial symbionts (bacteria, fungi, or yeast), but they differ from nodulating bacteria in that they live inside plants, typically in the intercellular spaces (apoplast) or vascular tissues. They provide their hosts with various benefits for growth and development under a variety of stressful environmental conditions (Rho ).

Doty isolated a series of beneficial bacteria and yeast strains from poplar and willow trees (Salicaceae) native to Washington State, USA, some of which were subsequently characterized to contain nifH genes and to increase N-fixation levels in inoculated plants (Knoth ). A follow-up in vitro study found that the amount of N fixed by these Salicaceae endophytes was comparable to that of the free-living N2-fixing Azotobacter species (Kandel ). Moreover, an in planta study with inoculation of the endophytes confirmed their capacity to perform BNF in leaves (Knoth ). Diazotrophic endophytes of the Salicaceae have been found not only to result in biomass increases in various plant hosts of different species (Khan ; Knoth ), but also to increase drought tolerance in poplar cuttings (Khan ) and in conifers (Aghai ), and water-use efficiency (WUE) in rice (Rho ). However, their capacity to enhance photosynthesis and to mitigate the down-regulation of photosynthesis in C3 plant hosts under elevated [CO2] has not been examined.

We hypothesized that these Salicaceae endophytes, like rhizobium bacteria in legumes, would mitigate the down-regulation of photosynthesis in plants when acting as diazotrophic symbionts. To test this, we used rice (Oryza sativa) as a model C3 crop species and investigated its photosynthetic biochemistry by simultaneous measurements of leaf gas exchange and chlorophyll fluorescence using plants grown under elevated [CO2]. This study is thus the first to investigate endophyte effects on C3 photosynthesis under elevated [CO2], and provides insights for comparison with our knowledge of nodulating mutualists.

Materials and methods

We conducted a series of greenhouse CO2-enrichment studies at the University of Washington Center for Urban Horticulture, USA (47°39′27ʺ N, 122°17′21ʺ W; 10 m above sea level). The studies were conducted over two growing seasons (spring to summer) in 2013 and 2014. The first experiment (Expt 1, 2013) was carried out under high-N conditions (HN), whilst the second (Expt 2, 2014) was carried out under low-N conditions (LN). Details of the experiments are summarized in Table 1.

Table 1.

Experimental details of the study

	Experiment 1	Experiment 2
Growing season	Spring to Summer (8 April–22 July, 2013)	Spring to Summer (21 March–2 September, 2014)
Growing duration	105 d	165 d
Experimental settings	2×2 (INOC×CO2)	2×2 (INOC×CO2)
Sample size	4 pots of 4 plants, one pair of chambers	8 pots of 4 plants, two pairs of chambers
Nitrogen fertilization*	High N (HN), 1×	Low N (LN), ¼×
Mean AMB [CO2]/ELE [CO2]	407/843 μmol mol–1	437/886 μmol mol–1
Mean day/night air temperature	23/19 °C	23/19 °C
Mean day/night relative humidity	57/64%	60/66%
Mean instantaneous light intensity	216.7 µmol m–2 s–1	176.3 µmol m–2 s–1
Mean daily light integral	11.6 mol m–2 d–1	9.1 mol m–2 d–1
Endophyte strains used**	PTD1 (Rhizobium sp.)	WP5 (Rahnella sp.)

INOC, endophyte inoculation treatment; CO2, CO2 enrichment treatment; AMB, ambient CO2 treatment; ELE, elevated CO2 treatment.

* Compared to the standard dose of Hoagland’s solution (Hoagland and Arnon, 1950).

** Closest 16S rRNA match identified through the BLAST NCBI database (Doty ).

CO2 enrichment system and environmental conditions

For the CO2 treatment, we used four sunlit, closed chambers. Fresh atmospheric (outdoor) air was supplied to the chambers from outside of the greenhouse using flexible dryer aluminum ducting (15.24×609.6 cm) with variable speed inline fans (Model FR, Fantech, Sarasota, FL, USA). Together with the fresh air, CO2 gas balanced with N2 was supplied through clear Tygon tubing (Saint-Gobain Performance Plastics, Akron, OH, USA) from a 22.7-kg tank (Praxair, Seattle, WA, USA). The CO2 gas was provided to two of the four chambers to provide the elevated [CO2] treatment (ELE, ~800 ppm). The other two chambers remained at ambient [CO2] (AMB, ~400 ppm). [CO2] in the chambers was monitored every 30 min by diffusion-type CO2 probes (GMP343, Vaisala Oyj., Vantaa, Finland). The mean [CO2] of the AMB and ELE chambers was 407 and 843 ppm, respectively, for Expt 1, and 437 and 886 ppm, respectively, for Expt 2.

A temperature and relative humidity probe (CS215, Campbell Scientific) was installed 20 cm from the top of each chamber and a quantum sensor (SQ-100, Apogee Instruments, Logan, UT, USA) was set up around each probe. Data from these instruments were collected every 15 min by a datalogger (CR1000, Campbell Scientific). Supplementary artificial lights (high-pressure sodium 400 W single-phase bulbs, Phillips Electronics) were used throughout both experiments to maintain consistent light intensity in the greenhouse during the daytime from 06.00 to 20.00 (14/10 h photoperiod). Mean day/night temperatures and relative humidity in Expt 1 were 23/19 °C and 57/64%, and in Expt 2 were 23/19 °C and 60/66%. Mean light intensity and daily light integral in Expt 1 were 216.7 μmol m–2 s–1 and 11.6 mol m–2 d–1 photosynthetic photon flux density (PPFD), and in Expt 2 were 176.3 μmol m–2 s–1 and 9.1 mol m–2 d–1 PPFD.

Specifications for the CO2 chamber system used in the current study can be also found in Kinmonth-Schultz and Kim (2011) and Nackley .

Preparation of microbial material

The endophyte strains PTD1 (Rhizobium tropici) and WP5 (Rahnella sp.), were used in Expts 1 and 2, respectively. PTD1 was isolated from hybrid poplar trees (Populus trichocarpa × deltoides) whilst WP5 was isolated from wild poplar trees (P. trichocarpa). The isolation processes and in vitro characteristics of the two strains, including production of indole-3-acetic acid and N2 fixation abilities conferred by the nifH gene, have previously been reported in Doty and Doty .

The isolates were stored in a 30% glycerol buffered solution at –80 °C prior to the beginning of both experiments. Before inoculation, the frozen endophytes were propagated on MG/L media plates (5.0 g l–1 tryptone, 2.5 g l–1 yeast extract, 5.2 g l–1 NaCl, 10.0 g l–1 mannitol, 1.32 g l–1 sodium glutamate, 0.50 g l–1 KH2PO4, 0.2 g l–1 MgSO4.7H2O, and 2 μg biotin at pH 7.0) (Cangelosi ), and then cultured overnight in N-limited combined carbon media solution (NL-CCM) (Solution 1: 5.0 g l–1 sucrose, 5.0 g l–1 mannitol, 0.5 ml l–1 sodium lactate, 0.8 g l–1 K2HPO4, 0.2 g l–1 KH2PO4, 0.1 g l–1 NaCl, 0.025 g l–1 Na2MoO4.2H2O, 0.028 g l–1 Na2FeEDTA, 0.1 g l–1 yeast extract; Solution 2: 0.2 g l–1 MgSO4.7H2O, 0.06 g l–1 CaCl2; a 9:1 mixture of solutions 1 and 2 was used after they had been autoclaved) (Rennie, 1981). The bacterial density (OD600) of the culture was measured using a spectrophotometer (UV-1700, Shimazu) and adjusted to 0.1 (equivalent to ~1×107 cells ml–1) to create the inoculum by diluting the culture with sterile deionized water and N-free media solution (NFM) (all g l–1: 69.9 KH2PO4, 19.84 K2HPO4, 174.7 K2SO4, 119.7 MgSO4.7H2O, 100 MgCl2.6H2O, 219.8 CaCl2.H2O, 3.38 MnSO4.H2O, 0.5 CuSO4.5H2O, 0.55 ZnSO4.7H2O, 3.83 H3BO3, 0.24 NaMoO4.2H2O, 0.11 CoSO4.6.5H2O, and 35 Fe sequestrine, at pH 6.5) (Doty ).

Preparation of plant material and inoculation with endophytes

Oryza sativa subsp. japonica M-206, which is very-early to early-maturing variety, was chosen as the host plant species based on our previous results (Kandel ). The seeds were surface-sterilized by soaking in 3% (v/v) NaOCl solution for 4 h, followed by rinsing five times with sterile water under aseptic conditions. The seeds were then incubated on a thin, solidified layer of 1% (w/v) water agar in sterile and sealed Petri dishes. There was 100% germination of the seeds within 2 d in both experiments.

At 7 d after germination (DAG), the seedlings were transferred to 3.8-l pots filled with horticultural media Sunshine Mix #4 and 11.4-l pots with Mix #2 (SunGro, Bellevue, WA, USA) in Expts 1 and 2, respectively. The volumes of the pots were sufficient to avoid possible sink-limitation effects from contained roots that have been observed in CO2 studies (Arp, 1991) and did not restrict the growth of roots: no differences associated with the ELE treatment were found in the root/shoot ratio of plants at harvest (data not shown, P=0.829; P=0.908 in Expt 1, and P=0.807 in Expt 2). The pots were cleaned by soaking in commercial bleach for 30 min. Four seedlings were transplanted to each pot. Immediately after transplanting, 2 ml of the endophyte inoculum was applied directly to the tip of the youngest emerging leaf of seedlings in half of the pots using a micropipette, and these were designated as the endophyte-inoculated treatment (E+). The same amount of endophyte-free NFM solution was applied to the remaining seedlings, which were designated as the mock-inoculated control (E–). After inoculation, the pots were moved to the growth chambers where they were placed in plastic buckets (5.7-l and 19-l for Expts 1 and 2, respectively) for easier supply of water and fertilizers, which were poured into the buckets for absorption through drainage holes in the pots. Four E– pots and four E+ pots (each containing four plants) were placed in each treatment chamber. For Expt 1, one AMB and one ELE chamber were used, giving a total of four replicate pots per treatment. For Expt 2, two AMB and two ELE chambers were used, giving a total of eight replicate pots per treatment.

From 14 DAG, 200 ml of NFM fertilizer solution was supplied to the plastic buckets every week. In Expt 1 the plants received a high-N (HN) treatment of 0.640 g l–1 and in Expt 2 they received a low-N (LN) treatment of 0.160 g l–1, equal to full- and quarter-strength Hoagland’s solution, respectively (Hoagland and Arnon, 1950).In addition to the fertilizer supply, the buckets were fully refilled with water each week to maintain a constant supply.

Verification of endophyte colonization

The inoculation method used in this study has previously been shown to be effective (Knoth et al., 2013, 2014; Kandel ; Khan ; Rho ; reviewed in Kandel ). To verify the colonization, a cultivation-based bacterial estimation method was used with semi-selective growing media. Approximately 100 mg of fresh leaf and root tissue was collected on 52 DAG in Expt 1 and 199 DAG in Expt 2. The samples were surface-sterilized by submerging them in 0.6% NaOCl for 3 min and 8 min for leaf and root tissues, respectively. The samples were then rinsed three times with sterile de-ionized water, and ground with a mortar and pestle in 400 µl of NL-CCM medium under aseptic conditions. The plant extracts were transferred to 25 ml of NL-CCM medium and grown for 48 h. The N-limited medium were used to serve as a semi-selective system to separate the culturable N-fixing endophytes in the extracts. The bacterial density of the extracts was quantified using a spectrophotometer by determining OD600 of 1-ml aliquots and converting into cell counts. The results showed consistent differences in the bacterial cell counts between the E– and E+ treatments, which verified the inoculation method (Supplementary Fig. S1 at JXB online; P=0.081: P=0.187 within HN and P<0.001 within LN).

Simultaneous leaf gas-exchange and chlorophyll fluorescence measurements

Leaf gas exchange and chlorophyll fluorescence parameters were determined on the second youngest fully expanded leaves at the panicle initiation stage on 65 DAG and 100 DAG in Expts 1 and 2, respectively. The following parameters were recorded: net CO2 assimilation rate (A), atmospheric [CO2] (Ca), [CO2] in the intercellular space (Ci), [CO2] in the chloroplasts (Cc), stomatal conductance (gs), mesophyll conductance (gm), transpiration rate (E), and electron transport rate (ETR). To measure these parameters simultaneously, LI-6400XT portable infrared gas analysers (Li-Cor Inc.) equipped with 2-cm2 leaf chamber fluorometers (LI-6400–40, Li-Cor Inc.) were used. The measurements were further used to calculate the intrinsic and extrinsic water-use efficiencies (iWUE and eWUE), which are defined by the ratios of A/gs and A/E.

To examine the biochemical properties of photosynthesis, CO2-response curves (A/Ci curves) were constructed using a built-in program function of the gas-exchange and fluorescence measurement systems. Net CO2 assimilation rates at different Ca values ranging from 50–1400 μmol CO2 mol–1 were recorded together with the other gas-exchange and fluorescence parameters. Leaf temperature was maintained at 25 °C and RH was kept at 40–60% during measurements. Light intensity was set at a saturating level of 1500 μmol m–2 s–1. Ca was first decreased stepwise from 400 μmol m–2 to 50 μmol mol–1, and then increased stepwise from 400 μmol mol–1 to 1400 μmol mol–1 under ambient [O2] conditions (21%). At each step, photosynthesis was allowed to stabilize to the new Ca level for 1–2 min before readings were taken, except for the increase from 50 μmol mol–1 to 400 μmol m–2 when an extra 2 min was allowed. The same protocol was then used to construct A/Ci curves under low [O2] conditions (2%), using an O2 cylinder balanced with N2 gas.

To estimate the internal conductance (gm) of leaves, the low-[O2] method developed by Bunce (2009) was used, combining the two A/Ci curves constructed at the different [O2] levels. First, to calculate the mitochondrial respiration rate (Rd) and the CO2 compensation point in the absence of Rd (Γ*), three A/Ci curves at three different light intensities (100, 200, and 800 μmol m–2 s–1 PPFD) were constructed at Ca values ranging from 50–250 μmol mol–1 following the method of Laisk (1977). Second, using the photosynthetic parameterization established by Sharkey (2016), the Farquhar–von Caemmerer–Berry (FvCB) biochemical properties of photosynthesis of the A/Ci curves were estimated at 21% [O2], namely maximum carboxylation rate (Vc,max), maximum ETR (Jmax), triose phosphate utilization rate (TPU), and the Michaelis–Menten constants for carboxylation (Kc) and oxygenation (Ko) of Rubisco (Farquhar ). Finally, the variables Rd, Γ*, Vc,max, Jmax, Kc, and Ko were used in the model parameterization of the low-[O2] method to calculate gm under the atmospheric [CO2] (Bunce, 2009).

Measurement of chlorophyll content

Measurements of chlorophyll content were taken on the same leaves that were used for gas-exchange and chlorophyll fluorescence, and were obtained using a hand-held chlorophyll meter (SPAD-502Plus, Konica Minolta Sensing Inc.). The measurements were taken immediately after the photosynthesis measurements.

Gene expression analysis using dye-based qPCR

In Expt 2, the leaf blades that had been used in the gas exchange measurements were harvested at 110 DAG between 5.5–6.5 h after the beginning of the photoperiod. The leaf samples were placed in 50-ml centrifuge tubes and immediately stored at –80 °C until further analysis.

Total leaf mRNAs were extracted using a Spectrum Plant Total RNA Kit (Sigma-Aldrich). The isolated mRNAs were mixed with a master reaction mixture containing fluorescent dyes (iQ SyBR Green Supermix, Bio-Rad Laboratories) prior to cDNA synthesis using an iScript cDNA Synthesis Kit (Bio-Rad Laboratories) and were then loaded into a real-time qPCR thermocycler (Chromo 4 System, Bio-Rad Laboratories). The relative gene expression of rbcS was calculated according to Pfaffl (2001). The oligonucleotide primer pair used was 5´-TGAAGCAGATCGAGTACCTGC-3´ and 5´-CTCCTCGAGCTCCTTGAGCA-3´. This pair of primers is known to have a highly homologous 185-bp part in their sequences for multiple genes from the rice rbcS gene family, such as OsRBCS2, OsRBC3, OsRBC4, and OsRBC5, and this can be utilized in quantifying total rbsS gene copy number (Suzuki ).

Statistical analysis

Each year the experiment was run under one N regime in order to test the other treatment effects, and therefore the N treatment effect was not subject to statistical comparisons because of the uneven conditions in Expts 1 and 2 (Table 1). Instead, the N treatment was set as a random effect in a mixed-effect model. Under a given N condition, the experiment followed a split-plot design. The inoculation treatment (INOC) and the [CO2] treatment (CO2) were regarded as fixed effects. For Expt 2, since two pairs of chambers were used for the CO2 treatment, the chamber effect on variability was tested as a random effect, and was determined to be insignificant at the α=0.05 level. Therefore, all parameters from the gas-exchange, fluorescence, and chlorophyll-content measurements were subjected to two-way ANOVA by fitting a linear mixed-effect model regression, using R version 3.2.2 (http://www.r-project.org). Within-group separation was assessed using Tukey’s HSD method.

Results

Alleviation of photosynthetic down-regulation by endophytes under elevated CO2 conditions

Analysis of the A/Ci curves indicated that photosynthetic down-regulation at the panicle initiation stage occurred as a result of the ELE treatment under both HN and LN conditions in E– plants (Fig. 1). However, the A/Ci curves of E+ plants had higher asymptotes than those of the E– plants.

Fig. 1.

CO2 response (A/Ci) curves of rice leaves at the panicle initiation stage following growth under ambient (AMB, ~400 ppm) and elevated [CO2] (ELE, ~800 ppm). The top panels show the responses of plants grown under high-N conditions (HN, n=4) and the bottom panels show plants grown under low-N conditions (LN, n=8). E–, mock-inoculated control plants; E+, endophyte-inoculated plants. Data are means (±SE). The solid horizontal lines indicate the asymptotes of the curves for E– plants under AMB conditions.

For the E– plants, all the FvCB photosynthetic biochemistry parameters of the A/Ci curves (Vc,max, Jmax, and TPU) were decreased by ELE compared to the E+ plants under AMB conditions (Table 2). Decreases in Vcmax, Jmax, and TPU of 10%, 3%, and 21%, respectively, were observed in E– plants under HN conditions in the ELE treatment, and Jmax and TPU were decreased by 16% and 2%, respectively, under LN conditions. In contrast, reductions in the parameters were not observed for E+ plants. Under ELE conditions, E+ plants had 5% and 14% increases in Vc,max and Jmax, respectively, compared to E– plants under HN conditions. The extent of the increases were higher under LN conditions, with Vc,max, Jmax, and TPU being increased by 33%, 7%, and 22%, respectively. The increases in these parameters in response to endophyte inoculation were all significant (Table 2, P<0.01). Significant INOC×CO2 interaction effects on Jmax and TPU were also detected (P=0.031 and P=0.010, respectively).

Table 2.

Photosynthetic parameters derived from the CO2 response curves (Fig. 1)

Treatments			n	Parameters
N	CO2	INOC		A max (μmol CO2 m–2 s–1)	ETR (μmol CO2 m–2 s–1)	V cmax (μmol CO2 m–2 s–1)	J max (μmol CO2 m–2 s–1)	TPU (μmol CO2 m–2 s–1)	g s (mol H2O m–2 s–1)	g m (mol CO2 m–2 s–1)	C i/Ca	C c/Ci	E (mmol H2O m–2 s–1)	iWUE (mol CO2 mol–1 H2O)	eWUE (mol CO2 mol–1 H2O)	SPAD units
HN	AMB	E–	4	25.46 (0.476)	128.1 (5.246)	103.3 (2.570)	135.6 (10.94)	11.77 (0.851)	0.593 (0.091)	0.190 (0.023)	0.783 (0.032)	0.543 (0.028)	6.517 (0.329)	0.0046 (0.0007)	0.393 (0.017)	41.57 (1.157)
		E+	4	25.17 (0.029)	127.8 (5.017)	113.9 (7.645)	149.9 (5.545)	10.75 (0.321)	0.636 (0.047)	0.160 (0.007)	0.799 (0.016)	0.461 (0.028)	7.143 (0.335)	0.0042 (0.0004)	0.369 (0.018)	42.77 (0.510)
	ELE	E–	4	26.36 (0.481)	103.8 (2.431)	93.11 (3.539)	131.5 (3.155)	9.308 (0.364)	0.612 (0.095)	0.135 (0.009)	0.883 (0.020)	0.713 (0.020)	5.851 (0.702)	0.0047 (0.0009)	0.473 (0.061)	40.75 (1.016)
		E+	4	32.23 (2.437)	124.9 (9.380)	108.12 (6.902)	155.2 (10.90)	11.35 (0.816)	0.627 (0.041)	0.188 (0.027)	0.873 (0.009)	0.742 (0.013)	5.783 (0.044)	0.0052 (0.0004)	0.558 (0.044)	43.35 (1.187)
LN	AMB	E–	8	10.85 (0.797)	84.66 (7.882)	35.65 (3.059)	89.90 (4.460)	5.979 (0.326)	0.191 (0.016)	0.166 (0.031)	0.744 (0.014)	0.732 (0.066)	2.259 (0.191)	0.0058 (0.0003)	0.484 (0.022)	38.42 (1.465)
		E+	8	11.14 (0.940)	85.53 (5.339)	43.96 (5.147)	86.40 (5.405)	6.525 (0.362)	0.177 (0.021)	0.150 (0.049)	0.717 (0.011)	0.642 (0.056)	2.069 (0.139)	0.0065 (0.0003)	0.540 (0.028)	38.62 (1.071)
	ELE	E–	8	14.23 (0.663)	91.77 (6.745)	36.50 (2.529)	75.50 (2.553)	5.854 (0.162)	0.164 (0.026)	0.050 (0.014)	0.784 (0.022)	0.433 (0.076)	1.709 (0.116)	0.0097 (0.0011)	0.844 (0.033)	37.36 (1.515)
		E+	8	17.71 (1.348)	117.3 (7.909)	47.58 (5.825)	96.13 (5.819)	7.284 (0.392)	0.155 (0.018)	0.121 (0.022)	0.736 (0.020)	0.684 (0.069)	1.815 (0.204)	0.0121 (0.0010)	1.013 (0.070)	41.85 (2.562)
Treatment effects
INOC, df=1				9.441** (0.004)	4.910* (0.032)	9.780** (0.003)	8.923** (0.005)	6.989* (0.011)	0.006 (0.939)	1.134 (0.294)	2.898° (0.096)	1.006 (0.323)	0.133 (0.717)	2.627 (0.112)	5.176* (0.028)	3.294° (0.076)
CO2, df=1				34.29*** (< 0.001)	2.308 (0.136)	0.118 (0.733)	0.060 (0.807)	0.101 (0.752)	0.315 (0.577)	5.456* (0.025)	12.09** (0.001)	0.032 (0.857)	11.63** (0.001)	29.25*** (< 0.001)	74.17*** (< 0.001)	0.315 (0.577)
INOC×CO2, df=1				6.310* (0.016)	4.518* (0.039)	0.233 (0.632)	5.022* (0.030)	6.562* (0.014)	0.012 (0.912)	3.988° (0.053)	0.664 (0.420)	5.594* (0.024)	0.009 (0.926)	1.254 (0.269)	2.325 (0.135)	1.928 (0.172)

INOC, endophyte inoculation treatment; CO2, CO2 enrichment treatment; AMB, ambient CO2 treatment; ELE, elevated CO2 treatment (see Table 1)

Data are means (±SE) of four (HN) or eight (LN) replicates. F-values from two-way ANOVA are given together with P-values in brackets. °, P<0.1 (highlighted in italics); *, P<0.05; ** P<0.01; *** P<0.001 (highlighted in bold).

For E– plants, the Aele/Aamb ratios were below 1.0 over the range of Ci values tested, indicating down-regulation of photosynthesis (Fig. 2). However, endophyte inoculation shifted the ratios to above 1.0, indicating mitigation of down-regulation and further increases in the photosynthetic capacity of E+ plants. This alleviation was observed under both HN and LN conditions over the range of Ca values from 400–800 μmol CO2 mol–1 (P<0.05).

Fig. 2.

The ratio of CO2 assimilation rates of rice leaves grown under elevated (Aele, ~800 ppm) and ambient [CO2] (Aamb, ~400 ppm) over the range of [CO2] used in the gas-exchange measurements for the A/Ci curves shown in Fig. 1. E–, mock-inoculated control plants; E+, endophyte-inoculated plants. The left panel shows the responses of plants grown under high-N conditions (HN, n=4) and the right panel shows plants grown under low-N conditions (LN, n=8). Data are means (±SE). HN: INOC*, significant inoculation effect on the Aele/Aamb ratio at P<0.05.

Increases in ETR and gm under elevated CO2 with endophyte inoculation

The ETR was greater in E+ plants than in E– plants only under the ELE treatment(Fig. 3), and endophyte inoculation increased ETR by 20% and 28% of in HN and LN conditions, respectively (Table 2). No changes were observed under AMB conditions in either N regime. There was a significant INOC×CO2 interaction effect (Fig. 3).

Fig. 3.

Electron transport rate (ETR) of rice leaves at the panicle initiation stage following growth under ambient (AMB, ~400 ppm) and elevated [CO2] (ELE, ~800 ppm). The data are derived from the A/Ci curves shown in Fig. 1. The top panels show the responses of plants grown under high-N conditions (HN, n=4) and the bottom panels show plants grown under low-N conditions (LN, n=8). E–, mock-inoculated control plants; E+, endophyte-inoculated plants. Data are means (±SE). The results of two-way ANOVA are shown below the figure. Within the HN treatment there was a significant effect of CO2 (*P<0.05). Within the LN treatment there was a significant effect of CO2 (*P<0.05), and effects of inoculation (INOC, °P<0.1) and an INOC×CO2 interaction (°P<0.1). INOC effects within the CO2 treatment were determined using Tukey’s HSD.

Mesophyll conductance showed contrasting responses to endophyte inoculation depending on [CO2] (Fig. 4, Table 2), and the INOC×CO2 interaction effect was marginally significant (P=0.053). Compared with AMB conditions, the ELE treatment significantly reduced gm in E– plants, with 29% and 70% decreases under HN and LN conditions, respectively, whilst E+ plants showed an increase of 18% with HN and a decrease of 27% with LN. Under ELE conditions, in response to endophyte inoculation plants showed increases of 39% and 142% with HN and LN, respectively. This was associated with increases in Cc/Ci of 4% and 58% with HN and LN, respectively (Table 2).

Fig. 4.

Mesophyll conductance (gm) of rice leaves at the panicle initiation stage following growth under ambient (AMB, ~400 ppm) and elevated [CO2] (ELE, ~800 ppm). The data are derived from the A/Ci curves shown in Fig. 1. The top panels show the responses of plants grown under high-N conditions (HN, n=4) and the bottom panels show plants grown under low-N conditions (LN, n=8). E–, mock-inoculated control plants; E+, endophyte-inoculated plants. Data are means (±SE). The results of two-way ANOVA are shown below the figure. Within the HN treatment there was a significant interaction between inoculation and CO2 (INOC×CO2, *P<0.05), and within the LN treatment there was a significant effect of CO2 (*P<0.05). INOC effects within the CO2 treatment were determined using Tukey’s HSD.

Increases in WUE with endophyte inoculation

Overall, no significant changes in response to the INOC treatment were found in iWUE in E+ plants but a significant increase was observed in eWUE (P=0.028, Table 2). The overall increase in Cc/Ci by 58% with endophyte inoculation was in accordance with this increase in eWUE (P=0.096).

The INOC treatment did not change eWUE of the plants under HN conditions (Fig. 5). Under LN conditions, however, a significant increase of 20% in eWUE was observed in E+ plants (P=0.045).

Fig. 5.

Extrinsic water use efficiency (eWUE) of rice leaves at the panicle initiation stage following growth under ambient (AMB, ~400 ppm) and elevated [CO2] (ELE, ~800 ppm). eWUE is defined by net CO2 assimilation rate (A) divided by transpiration rate (E). The data are derived from the A/Ci curves shown in Fig. 1. The top panels show the responses of plants grown under high-N conditions (HN, n=4) and the bottom panels show plants grown under low-N conditions (LN, n=8). E–, mock-inoculated control plants; E+, endophyte-inoculated plants. Data are means (±SE). The results of two-way ANOVA are shown below the figure. Within the HN treatment there was a significant effect of CO2 (**P<0.01), and within the LN treatment there were significant effects of CO2 (*P<0.05) and inoculation (INOC, *P<0.05). INOC effects within the CO2 treatment were determined using Tukey’s HSD.

Increases in chlorophyll content with endophyte inoculation

Overall, a small increase in chlorophyll content as represented by SPAD value was observed for the INOC treatment (5%, P=0.076, Table 2). Within the individual treatments, this small increase was only observed under ELE conditions, where it was seen with both HN and LN (P=0.079 and P=0.076, respectively, Fig. 6).

Fig. 6.

Chlorophyll contents of rice leaves at the panicle initiation stage following growth under ambient (AMB, ~400 ppm) and elevated [CO2] (ELE, ~800 ppm). Chlorophyll contents are presented as SPAD units. The top panels show the responses of plants grown under high-N conditions (HN, n=4) and the bottom panels show plants grown under low-N conditions (LN, n=8). E–, mock-inoculated control plants; E+, endophyte-inoculated plants. Data are means (±SE). The results of two-way ANOVA are shown below the figure. Within the HN treatment there was an effect of inoculation (INOC, °P<0.1). There were no effects within the LN treatment. INOC effects within the CO2 treatment were determined by orthogonal contrasts.

Increases in rbcS expression with endophyte inoculation

Endophyte-inoculated plants grown under LN conditions showed a higher level of rbcS expression; however, the difference was only marginal (P=0.059; Fig. 7). Under AMB conditions, E+ plants had 66% higher expression compared to E– plants, but under ELE conditions the increase was only 7%. No CO2 and INOC×CO2 effects were observed).

Fig. 7.

Expression levels of the rbcS gene in rice leaves at the panicle initiation stage following growth under ambient (AMB, ~400 ppm) and elevated [CO2] (ELE, ~800 ppm) conditions with low N supply. The actual copy numbers of rbcS were calculated by the method of Pfaffl (2001). E–, mock-inoculated control plants; E+, endophyte-inoculated plants. Data are means (±SE) of n=6–8 replicates. An effect of the inoculation treatment under ambient [CO2] was determined (°P<0.1).

Discussion

Our results showed that the presence of symbiotic endophytes in rice mitigated the down-regulation of C3 photosynthesis in response to long-term exposure to elevated [CO2]. Notably, this phenomenon appeared to be derived from improvements on both the demand and supply sides of the photosynthetic biophysical and biochemical processes. On the demand side, electron transport in the light reaction complexes was up-regulated, resulting in increases in ETR upon inoculation with endophytes. Given that C3 photosynthesis is mostly limited by the provision of NADPH and ATP from the light reactions under conditions of elevated CO2 (ELE), the increase in ETR indicates the importance of the impact of endophytes on photosynthesis under future climate conditions. On the supply side, the velocity of CO2 diffusion to the site of carboxylation was increased upon endophyte inoculation under ELE conditions. Under such conditions, even though atmospheric CO2 is sufficient for photosynthetic assimilation, biophysical and biochemical barriers such as stomatal, cell wall, and liquid-phase resistances in leaves hinder CO2 diffusion and further fixation. The rapid delivery of the substrate to the reaction site is therefore integral in improving the efficiency of the assimilation process. In this respect, the increases in gm triggered by endophytes are especially important. These photosynthetic adaptation responses in the symbiotic plants were consistent over the different N levels that we examined.

Endophytes ameliorate photosynthetic down-regulation under elevated [CO2]

Typical responses to elevated [CO2] in the down-regulation of C3 photosynthesis are decreases in the initial slope of the A/Ci curve (Vc,max) and decreases in the asymptote (Jmax), as observed in our results (Fig. 1). The down-regulation of photosynthesis begins with accumulation of non-structural carbohydrates (NSCs), which results in starch accumulation in source tissue chloroplasts that are exposed to a long-term elevated CO2. This in turn causes a reduction in Rubisco turnover rates, and ultimately results in decreases in Rubisco and associated enzymes that operate CO2 assimilation in the Calvin–Benson Cycle as a consequence of negative feedback (Drake ). Interestingly, however, E+ plants appeared to have higher values of FvCB C3 photosynthetic parameters, with Vc,max, Jmax, and TPU being higher than those of E– plants under ELE conditions, regardless of N levels (Table 2, Fig. 1). This pattern of the A/Ci curves resembled previously observed photosynthetic responses of legumes to ELE conditions reported by Ainsworth and Rogers (2007). These authors found that fewer down-regulation symptoms were reported in legumes than in other C3 crop species in 31 studies using FACE facilities across the world, as shown in A/Ci curves and their parameterizations. The two main reasons behind this were considered to be that N derived from biological N2 fixation (BNF) could be used to sustain Rubisco content and capacity under ELE conditions, and that the nodules created by symbiotic rhizobium bacteria in the legume roots provided an increased sink strength (also reviewed in Leakey , with a focus on physiological CO2 responses of legumes). Ainsworth empirically tested this source–sink relationship between plant hosts and symbiotic bacteria under elevated CO2 levels by genetically modifying the sink strength of the plant. They found that a decrease in sink capacity in response to losing nodules in the root tissues resulted in substantial down-regulation of photosynthesis compared to non-genetically modified controls. Furthermore, extra N fixed by endophytes could be utilized to create sink tissues as building blocks of new biomass (Kim ).

In this regard, a mechanistic explanation for the mitigation of down-regulation in the plants that we inoculated with diazotrophic endophytes can be found by analogy with the legume–rhizobium symbiosis. First, BNF by endophytes is a well-established functional trait of plant–endophyte interactions (Doty, 2017), as shown in our previous in planta study with Salicaceae endophytes (Knoth ). In addition, although the amount of N in the leaves that originated from the endophytes was not estimated in our present study, the chlorophyll content of the leaves was higher in E+ plants than in E– (Table 2, Fig. 6). There is a strong positive linear correlation between leaf chlorophyll content and leaf N content in rice, and SPAD units are used as an indicator of leaf N status (Zhu ; Yang ). Hence, the increase in the chlorophyll content could be the explanation for the overall improvement of photosynthetic performance in the presence of the N-fixing endophytes. Second, although the biological sink strength of endophytes has never been quantified, they are active symbionts that consume carbohydrates supplied by the plant host. Given that other symbiotic associations (e.g. rhizobium bacteria and mycorrhizal fungi) cost the host plant around 5–20% of total carbohydrates (Lambers ; Jones ; Kaschuk ), endophytes should drain a substantial amount of carbohydrates from the host, thereby serving as active biological sinks. Host plants become able to create and provide larger C food reserves when there are abundant environmental substrates, especially under ELE conditions.

The demand side: endophytes facilitate photosynthetic electron transfer under elevated CO2 conditions

Although the underlying mechanisms that triggered increases in ETR in response to endophyte inoculation under ELE conditions (Fig. 3) are difficult to elucidate with our data, the response was consistent for both N-sufficient and N-limited conditions. Woodward found that although photosynthetic rates of symbiotic and non-symbiotic plants were not significantly different, photochemical efficiency (Φ PSII) increased in symbiotic tomato plants with fungal endophytes. Increases in Φ PSII can be translated into the increases observed in ETR in the present study since the relationship between the two is positively linear (Baker, 2008). This implies that more ATP and NADPH were produced in the light reaction of photosynthesis with the same number of quanta than in the non-symbiotic plants, and they are consumed in the process of the CO2 fixation, as seen in the increases in Amax (Table 2, Fig. 1).

Given that plants under elevated [CO2] will experience more limitation of photosynthesis as a result of decreased regeneration of RuBP (Ainsworth and Rogers, 2007), the increases in ETR resulting from endophyte inoculation together with increases in other relevant PSII activities (e.g. photochemical and non-photochemical quenching; data not shown) suggest that the endophyte symbiosis could be an effective means of mitigating the impacts of elevated [CO2] on photosynthesis..

The carboxylation process by Rubisco was also slightly improved by endophyte inoculation as indicated by the increases in rbcS expression in E+ plants (Fig. 7), which was in accordance with the steeper initial slopes of the A/Ci curves and the corresponding increases in Vc,max in E+ plants (Fig. 1, Table 2). A greater abundance of Rubisco in the presence of endophytes probably enhanced the carboxylation capacity of the Calvin cycle. These improvements in Rubisco capacity would help plants hosting endophytes to alleviate photosynthetic down-regulation under elevated CO2 conditions. This alleviation would surpass that found in legume–rhizobium symbiosis under elevated CO2 conditions since legumes do not seem to show increases in Rubisco capacity. Instead, symbiotic legumes maintain the same Vc,max and Jmax under elevated CO2 conditions (Ainsworth and Rogers, 2007).

The supply side: endophytes promote internal CO2 diffusion under elevated CO2 conditions

Higher gm facilitates the diffusion of CO2 through the chloroplast walls and through other layers on the pathway of molecules from the atmosphere to the site of carboxylation. Therefore, plants with higher gm should have a better supply of CO2 (Flexas ).

Increases in gm were only observed under ELE conditions and this coincided with increases in ETR (Figs 3, 4). The coordinated mechanisms of photosynthetic electron transport of the thylakoid membrane in PSII together with the supply of bicarbonate to the thylakoid space are described and reviewed by Govindjee and van Rensen (1993) and van Rensen and Klimov (2005). CO2 in the apoplastic spaces of leaves exists in the form of either bicarbonate (HCO3–) or carbonic acid (CO32–). Under ELE conditions, plants produce more carbohydrates due to increases in the supply of carboxylation substrate (Ainsworth and Long, 2005). Assimilated C in the form of NSCs can be used by microorganisms that live in the intercellular spaces of the host plants. As endophytes consume photoassimilates in the process of respiration, they release CO2, which can be readily dissolved to bicarbonate ions (Rho ). This may promote the light-harvesting process in PSII by increasing ETR in coordination as well. Up-regulating the activities of the PSII complex also suppresses stomatal opening while not affecting CO2 assimilation, leading to increasing WUE of the plants (Głowacka ), which we observed in our results (Table 2).With more CO2 and NSCs under ELE, and therefore more microbial release of respiratory CO2, there may be a better chance of symbiotic plants having more internal CO2 available for assimilation. As a consequence, gm could be increased by this signal and further stimulate the entire assimilation process, as indicated by the increases in Amax in our data.

A more efficient internal supply of CO2 with endophytes under ELE conditions led to increases in WUE, as shown by Ci/Ca and Cc/Ci (Table 2). Ci/Ca decreased while Cc/Ci increased when N was limited under ELE conditions. This indicated that the increases in WUE under ELE conditions with endophytes were attributable to a more rapid supply of CO2 inside the leaf, rather than to reactions of the stomata. Flexas argued that gm is explicitly related to water movement in the leaves and it should be set as a target for breeding to increase WUE in crops. Internal CO2 diffusion in leaves is catalysed by carbonic anhydrases, which may be represented by gm. For example, overexpression of carbonic anhydrases in Arabidopsis substantially increases WUE (Hu ). Although the overexpressing lines also showed decreases in gs, which we did not observe in our results, our data suggest that using symbionts could be a means to increase WUE.

In addition, we found that N supply affected this response, with increases in WUE in E+ plants being more significant under LN conditions than under HN (Fig. 5). In contrast, Rho showed that Salicaceae endophytes increased WUE by reducing stomatal aperture during the afternoon while maintaining photosynthetic capacity under AMB conditions. Under our ELE conditions, endophytes instead appeared to modulate internal components of the leaf with more carbohydrates available in the host plant. More fundamental approaches at the molecular scale are needed to provide a mechanistic understanding of the responses.

The beneficial effects of endophytes on WUE are greater under elevated CO2 conditions

Physiological characteristics at the leaf level were not noticeably altered with endophyte inoculation under AMB conditions (Table 2). This is in agreement with the results of Rogers who found that hardwood cuttings of Populus deltoides inoculated with Enterobacter species showed increased biomass but no effects on photosynthetic parameters such as A, gs, and photosynthetic WUE (i.e. either iWUE or eWUE). The authors suggested that the increases in productivity were more related to increases in leaf area at the whole-plant physiological scale. Although varying effects were observed in some parameters, E+ plants in our study displayed similar responses to E– plants under AMB conditions.

Nevertheless, many photosynthetic parameters examined in this study showed significant INOC×CO2 interactions, including Amax, ETR, Jmax, TPU, gm, and Cc/Ci (Table 2). This indicated that endophyte inoculation enhanced these photosynthetic properties, which was more efficient under ELE conditions.

Conclusions

In this study, selected diazotrophic endophytes originally isolated from the Salicaceae family were inoculated onto rice plants and enabled them to benefit from higher capacities for various processes in photosynthetic CO2 assimilation. The down-regulation of C3 photosynthesis in response to elevated [CO2] was ameliorated with endophyte inoculation, as observed by monitoring A/Ci responses. Electron transport in the light reactions and CO2 diffusion in the intercellular spaces were facilitated by endophyte inoculation, resulting in increases in the ETR and gm under both high- and low-N conditions, but this was only distinguished under elevated CO2 conditions where the plants experienced down-regulation of photosynthesis in response to high [CO2]. The overall improvement of photosynthetic responses led to increases in the WUE of the plants. In the presence of endophytes, the leaf gas-exchange properties of rice were similar to those of legume species grown under elevated [CO2] as found in previous studies (Ainsworth and Rogers, 2007; Leakey ; Rogers ). Limited numbers of legume species exist and biological N2 fixation in them is reported to be declining with increasing ambient [CO2] (Hungate ). Our findings suggest that there is potential for using endophytic microbial partners in non-legume host crops to promote plant growth under future climatic conditions. Further research is required to determine the molecular mechanisms that underlie the alleviation of photosynthetic down-regulation under elevated [CO2].

Supplementary Data

Supplementary data are available at JXB online.

Fig. S1. Verification of endophyte colonization as determined by bacterial cell counts.

Click here for additional data file.