Siberian Miscanthus sacchariflorus accessions surpass the exceptional chilling tolerance of the most widely cultivated clone of Miscanthus x giganteus

Chilling temperatures (0–15°C) inhibit photosynthesis in most C4 grasses, yet photosynthesis is chilling tolerant in the ‘Illinois’ clone of the C4 grass Miscanthus x giganteus, a candidate cellulosic bioenergy crop. M. x giganteus is a hybrid between Miscanthus sacchariflorus and Miscanthus sinensis; therefore chilling‐tolerant parent lines might produce hybrids superior to the current clone. Recently a collection of M. sacchariflorus from Siberia, the apparent low temperature limit of natural distribution, became available, which may be a source for chilling tolerance. The collection was screened for chilling tolerance of photosynthesis by measuring dark‐adapted maximum quantum yield of PSII photochemistry (Fv/Fm) on plants in the field in cool weather. Superior accessions were selected for further phenotyping: plants were grown at 25°C, transferred to 10°C (chilling) for 15 days, and returned to 25°C for 7 days (recovery). Two experiments assessed: (a) light‐saturated net photosynthetic rate (Asat) and operating quantum yield of PSII photochemistry (ΦPSII), (b) response of net leaf CO2 uptake (A) to intercellular [CO2] (ci). Three accessions showed superior chilling tolerance: RU2012‐069 and RU2012‐114 achieved Asat up to double that of M. x giganteus prior to and during chilling, due to increased ci ‐ saturated photosynthesis (Vmax). RU2012‐069 and RU2012‐114 also maintained greater levels of ΦPSII during chilling, indicating reduced photodamage. Additionally, accession RU2012‐112 maintained a stable Asat throughout the 15‐day chilling period, while Asat continuously declined in other accessions; this suggests RU2012‐112 could outperform others in lengthy chilling periods. Plants were returned to 25°C after the chilling period; M. x giganteus showed the weakest recovery after 1 day, but a strong recovery after 1 week. This study has therefore identified important genetic resources for the synthesis of improved lines of M. x giganteus, which could facilitate the displacement of fossil fuels by cellulosic bioenergy.


| INTRODUCTION
The C 4 grass Miscanthus x giganteus Greef and Deuter ex Hodkinson & Renvoize is exceptionally productive in the US Midwest, yielding 59% greater peak biomass than Zea mays L (Dohleman & Long, 2009). This is explained by its ability to develop and maintain leaves that are photosynthetically competent under the chilling conditions of early Spring and Fall, enabling a longer growing season than Z. mays (Dohleman & Long, 2009). In southern England where low temperatures prevent production of a Z. mays grain crop in most years, M.
x giganteus produced a peak dry biomass yield of 30 t/ha, a record productivity for the UK (Beale & Long, 1995). This correlated with a lack of the chilling damage to photosynthesis observed in Z. mays (Beale, Bint, & Long, 1996).
Superior chilling tolerance and resulting higher yields compared to other candidate bioenergy grasses has garnered strong interest in developing M. x giganteus as a feedstock for cellulosic bioenergy production (Clifton-Brown, Chiang, & Hodkinson, 2008;Clifton-Brown et al., 2018;Heaton, Dohleman, & Miguez, 2010). A recent meta-analysis of over 1,000 published yields, showed that across temperature, fertility and water availability gradients, average yields of M.
x giganteus were more than double those of Panicum virgatum under the same conditions, despite major breeding efforts to improve the latter (LeBauer et al., 2018). As a result Miscanthus could be closer than P. virgatum to commercialscale viability (Clifton-Brown et al., 2018). Unlike P. virgatum, which includes adapted and improved ecotypes, yields of M. x giganteus in the US are largely for the single unimproved 'Illinois' clone (Głowacka, Clark, et al., 2015;Long & Spence, 2013). There is probably no or very little genetic diversity between the 'Illinois' clone and other M. x giganteus legacy clones in the US and Europe (Głowacka, Clark, et al., 2015). Genetic mapping of the parent species, however, finds very significant potential for breeding of high-yielding Miscanthus cultivars (Clark et al., 2014Dong et al., 2018). This raises the likelihood that even more productive forms of M. x giganteus could be developed.
M. x giganteus 'Illinois' is a sterile, triploid hybrid between a tetraploid Miscanthus sacchariflorus (Maxim.) Hack. and the diploid Miscanthus sinensis Andersson (Hodkinson & Renvoize, 2001). M. sacchariflorus and M. sinensis are genetically diverse, distributed over a wide latitudinal and thermal range in eastern Asia from the sub-tropics to Siberia (Clark et al., 2014(Clark et al., , 2016. It was predicted that productivity of M. x giganteus could be improved by up to 25% with the greater chilling tolerance seen in a M. sinensis hybrid (Farrell, Clifton-Brown, Lewandowski, & Jones, 2006). A study of various Miscanthus accessions found that when abruptly exposed to an 11-day 10°C chilling treatment, some Japanese M. sacchariflorus accessions could match the photosynthetic rate of M. x giganteus 'Illinois' (Glowacka et al., 2014). A similar study of Miscanthus found two M. sacchariflorus accessions from Japan capable of surviving night-time frost and achieving light-limited and light-saturated photosynthetic rates >40% greater than M. x giganteus grown at 15°C (Glowacka, Jorgensen, et al., 2015). However, most of these plants originated from temperate regions. Even the highly studied 'Illinois' clone of M. x giganteus appears to originate from a cross of parental accessions from temperate southern Japan (Clark et al., 2014;Głowacka, Clark, et al., 2015). This raises the question of whether higher latitude germplasm, from the coldest regions that support Miscanthus growth, could show still greater chilling tolerance.
One hundred and sixty M. sacchariflorus accessions were recently collected across the eastern region of Russian Siberia (Clark et al., 2016); these represent the northern limit of wild Miscanthus and likely candidates for greater chilling tolerance. The hypothesis that this material would have greater chilling tolerance was tested in three successive physiological experiments, culminating in the discovery of three M. sacchariflorus accessions capable of achieving double the photosynthetic rate of M. x giganteus during chilling.

| MATERIALS AND METHODS
Three experiments were conducted as successive cold-tolerance screens: Experiment 1: measurements of dark-adapted maximum quantum yield of PSII photochemistry (F v /F m ) were used to screen for chilling tolerance in a field-grown population of 92 M. sacchariflorus accessions that originated from 43°N to 49°N in eastern Russia, representing USDA hardiness zones 5 to 3 (warmer to colder) (Clark et al., 2016). These accessions are the northernmost members of the extensive M. sacchariflorus Korea/NE China/Russia diploids genetic group, one of six M. sacchariflorus genetic groups . Seven accessions were chosen for further evaluation based on their field performance and availability of sufficient clonal material for further investigation: RU2012-114 (48.60930°N, 134.21509°E, 42.0 29384°N, 133.17572°E, 35.0 m, hardiness zone 3). Experiment 2: With these seven accessions, a controlled-environment experiment assessed operating light-saturated net rates of leaf CO 2 uptake (A sat ) and operating quantum yield of PSII photochemistry (Φ PSII ) during a 15-day chilling period. Three accessions, RU2012-114, RU2012-112, and RU2012-069, showed high chilling tolerance. Experiment 3: With these three accessions, a controlled-environment experiment analyzed CO 2 -saturated rate of photosynthesis (V max ) and the maximum apparent rate of PEPc carboxylation (V pmax ) during a 15-day chilling period, followed by a 7-day recovery to warm temperature. In all experiments M. x giganteus 'Illinois' was included as a control.  (Clark et al., 2016). Detailed description of the origin of each accession was given previously (Clark et al., 2016). These were then initially planted out on the Aarhus University Farm at Foulum, Denmark, in 2013 (56°30′N, 9°35′E). For this study the accessions were narrowed down to the most chilling-tolerant subset in terms of photosystem II maximum efficiency in the experiments described below. Soil at this site is a sandy loam (typic Fragiudalf; USDA soil taxonomy). Two clones of each accession were planted 150 cm apart in the summer of 2013 together with M. x giganteus 'Illinois' which was propagated from rhizome cuttings. Plants were rainfed and not fertilized. Air temperature at 20 and 150 cm was recorded every 10 min by a meteorological station located within the trial site.

| Dark-adapted chlorophyll fluorescence measurements
Between June 19th and July 2nd, 2014, a 13-day chilling period occurred at the Foulum field site where average daily air temperatures did not exceed 16°C. During this period about 40% of the collection became chlorotic; these accessions were not included in the screen. Ninety-two M. sacchariflorus accessions showing minimal leaf chlorosis were selected: measurements were taken from 11:30 p.m. to 4:00 a.m., on at least two shoots for each of the two plantings of selected accessions, along with M. x giganteus. This timing was to ensure that leaves were dark adapted and that functioning photosystem II (PSII) centers would be fully open (Baker, 2008). The youngest fully expanded leaf, as indicated by ligule emergence, was placed in the cuvette of a portable gas-exchange system incorporating infra-red CO 2 and water vapor analyzers (LI 6400; LI-COR, Inc., Lincoln, NE) and using a combined gas-exchange and pulse-amplitude fluorescence attachment (LI 6400-40; LI-COR). Once chlorophyll fluorescence stabilized (<30 s), dark-adapted maximum quantum yield of PSII photochemistry (F v /F m ) was recorded using a rectangular saturating flash protocol (Baker, 2008).
This method follows that of the rapid screen for chilling tolerance in field grown plants used previously (Glowacka, Jorgensen, et al., 2015).
The mean F v /F m per accession in Experiment 1 was used to select seven accessions that were representative of the full range of variation in F v /F m , and were used for further screening in Experiment 2. These seven were among several accessions that were cloned and shipped for propagation in controlled environments at the University of Illinois. x giganteus, were grown for a controlled-environment experiment. Clonal divisions (n = 4) of each accession were grown from rhizomes in 1.6 L pots containing a peat/bark/perlite-based growing medium (Metro-Mix 900; Sun Gro Horticulture, Agawam, MA). After cloning by rhizome propagation, a slow release fertilizer was added according to the manufacturer's instructions (Osmocote Pro, 8-9 mo 19-5-8 Minors; Everris NA, Inc., Dublin, OH). Plants were watered daily and grown for 53 days in a controlled-environment greenhouse at ~25°C, with high pressure sodium lamps ensuring a minimum photosynthetically active photon flux (Q) of 300 μmol m −2 s −1 and a 14-hr day length. Plants were transferred from the greenhouse to two controlled-environment growth cabinets for 23 days before measurements began (Model PCG20, Conviron, Winnipeg, MB R3H 0R9, Canada). The cabinets maintained a 14 hr /10 hr day/night cycle under 800 photons m -2 s −1 , 25°C daytime/20°C nighttime temperature, with a relative humidity of 75% throughout.

| Gas exchange and chlorophyll fluorescence measurements
A first day of measurement was taken at 25°C (day 0), then cabinet temperature was reduced to 10°C/5°C day/night for 15 days (days 1-15). This experimental timeline has been used previously to mimic the type of chilling (0-12°C) that might affect leaf emergence during spring or expanded leaves in the autumn (Allen & Ort, 2001;Baker, Bradbury, Farage, Ireland, & Long, 1989;Glowacka et al., 2014). Measurements were taken on days 0, 1, 2, 5, 6, 8, 10, 12, and 15. Pictures of plants were taken with a camera (Canon PowerShot SX50 HS, Canon, Tokyo, Japan) before and after the experiment to show differences in any visible damage sustained by plants during chilling.

PIGNON et al.
The youngest fully expanded leaf from each plant's primary tiller was selected for measurement. Leaves were measured with a portable gas-exchange system (LI 6400; LI-COR, Inc.). The manufacturer's combined gas-exchange and fluorescence measurement attachment was used (LI 6400-40; LI-COR). Incident photon flux was set to 2,000 μmol m −2 s −1 , block temperature to either 25 or 10°C on warm and chilling days, respectively, flow rate to 400 µmol/s, reference [CO 2 ] to 400 μmol/mol and leafto-air water vapor pressure deficit maintained at <2 kPa. Light was provided by the integrated red (635 nm wavelength) and blue (465 nm wavelength) light-emitting diodes (LED). Leaves acclimated in the measurement cuvette until the net rate of leaf CO 2 uptake (A) reached a steady state, then measurements were initiated.
Gas-exchange data were recorded and A and g s calculated (von Caemmerer & Farquhar, 1981); followed by modulated fluorescence measurements to derive operating quantum yield of PSII photochemistry (Φ PSII ) using a multiphase flash protocol (Loriaux et al., 2013). Because all measurements were taken under saturating photon flux (Q = 2,000 μmol m −2 s −1 ) A was considered light-saturated and denoted as A sat .
The mean A sat per accession during chilling days in Experiment 2 was used to select three accessions that surpassed M. x giganteus, and were used for further screening in Experiment 3.

| Plant material and growing conditions
Three M. sacchariflorus accessions, and an 'Illinois' clone of M. x giganteus, were grown in another controlled environment experiment. Growing conditions were similar to Experiment 2, except that: n = 6 clonal divisions of each accession were planted in 5.7 L pots containing a soil-free medium (LC1, Sungro Horticulture). Pots were fertilized twice; after cloning by rhizome propagation, and 2 days before the start of measurements with a slow release 17-5-11 fertilizer added according to the manufacturer's instructions (Osmocote Pro, Everris NA, Inc.) and supplemented with iron (Ferrous sulphate heptahydrate, QC Corporation, Girardeau, MO). Plants were grown for 1 month in a controlled-environment cabinet before measurements began (Model PCG20, Conviron) with growing conditions as in Experiment 2.

| Gas exchange measurements
Growth cabinet conditions during the experiment were similar to Experiment 2, except that on day 16 cabinets were returned to 25°C daytime/20°C nighttime for a week.
Measurements were taken on days 0, 1, 3, 5, 7, 9, 11, 13, 15, 16, and 23. A portable gas-exchange system (LI 6400; LI-COR, Inc.) was used for measurement as in Experiment 2, but a larger chamber without fluorescence capability was used to improve measurement accuracy of small gas fluxes (LI 6400-02B; LI-COR). Leaves acclimated in the measurement cuvette until the net rate of leaf CO 2 uptake (A) reached a steady state, then A-c i curves were measured by progressively decreasing [CO 2 ] in the reference cell (400,350,300,250,200,150,100,60, and 0 μmol/mol). Leaves were allowed to acclimate to each step reduction in [CO 2 ] for 2-3 min, as assessed by a resumption of a steady-state A, then gas-exchange data were recorded. Because all measurements were taken under saturating photon flux (Q = 2,000 μmol m −2 s −1 ) A sat was determined as A measured at [CO 2 ] = 400 μmol/mol.
A-c i curves were fit to a nonrectangular hyperbolic function as in (Leakey et al., 2006), and the CO 2 -saturated rate of photosynthesis (V max ) was estimated as the predicted value of each function for c i = 2,000 μmol/mol. The response of A to c i at c i <100 μmol/mol was used to solve for maximum apparent rate of PEPc carboxylation (V pmax ) (von Caemmerer, 2000), with temperature-dependent estimates of the Michaelis-Menten constant of PEPc for [CO 2 ] (Kp) from the C 4 grass Setaria viridis (Boyd, Gandin, & Cousins, 2015). All curve fitting was performed using non-linear regression (PROC NLIN, SAS v8.02; SAS Institute, Cary, NC).

| Statistical analysis
There were three main statistical tests used to compare the three top-performing M. sacchariflorus accessions to M. x giganteus: ANOVA on individual days, repeated measures ANOVA throughout chilling days, and non-linear regression throughout chilling days.
ANOVA (PROC GLM, SAS Institute Inc.) with a onetailed Dunnett's test (significant: p < 0.05 threshold, marginally significant: p < 0.1 threshold) was used to determine whether any of the M. sacchariflorus accessions surpassed M. x giganteus in terms of A sat , Φ PSII , V max , V pmax , and g s on key measurement days: the prechilling day (day 0), the first (day 1) and last (day 15) day of chilling, and the first day of recovery to 25°C (day 16). For the last day of recovery to 25°C (day 23), a similar ANOVA was performed, but with a two-tailed Dunnett's test, to account for the possibility of M. x giganteus surpassing the M. sacchariflorus accessions.
In order to assess photosynthetic traits throughout the chilling period (days 1 through 15), a repeated measures one-tailed ANOVA (significant: p < 0.05 threshold, marginally significant: p < 0.1 threshold) (PROC GLM, SAS Institute Inc.) was used to test whether each M. sacchariflorus accession surpassed M. x giganteus for A sat , Φ PSII , V max , V pmax , and g s .

PIGNON et al.
In different accessions, A sat , Φ PSII , V max , V pmax , and g s would either decrease throughout days 1-15, or decline to a minimum and then increase over several days. To quantify these different patterns, data from days 0-15 were fit by nonlinear regression (PROC NLIN, SAS Institute Inc.) to the equation: This describes an exponential decline of Y during the transition from 25°C to 10°C, followed by a linear increase, decrease, or plateau over the course of days 1-15. Y is either A sat , Φ PSII , V max , V pmax , or g s . m is the key parameter describing the rate of linear increase (m > 0) or decrease (m < 0) of Y in days 1-15. y0 is the Y-intercept, b is the lowest value of Y reached during the chilling period, λ describes the rate of change of Y from y0 to b. Figure S1 is a schematic example describing this fit equation, where Y decreases to a minimum at day 4, then increases until day 15, leading to m > 0. ANOVA (PROC GLM, SAS Institute Inc.) with a one-tailed Dunnett's test (significant: p < 0.05 threshold, marginally significant: p < 0.1 threshold) was used to determine whether any of the M. sacchariflorus accessions surpassed M. x giganteus for m fit to A sat , Φ PSII , V max , V pmax , or g s .
Measurements of A sat , and g s were obtained for M. x giganteus and the three top-performing M. sacchariflorus accessions in Experiments 2 & 3 under comparable experimental conditions, therefore data from both experiments were pooled and a fixed two-factor block for 'Experiment' added when analyzing these variables. In all ANOVAs, homogeneity of variance was verified graphically, and normality of studentized residuals tested by Shapiro-Wilke (PROC UNIVARIATE, SAS Institute) at p = 0.01 threshold.

| Selection of three top-performing accessions
Air temperature varied between 5 and 25°C throughout June of 2014, and dipped at the end of the month as a cool spell set in with three successive days in which average air temperature was 12°C, with dawn air temperatures averaging 8°C ( Figure S2). Field measurements of F v /F m following these 3 days ranged from 0.62 to 0.74, and all but one of the M. sacchariflorus accessions achieved higher F v /F m than M.
x giganteus. (Figure S3). Based on these results and availability of plant material, seven accessions were selected for transfer to Illinois for further analysis in Experiment 2. These were accessions RU2012-114, RU2012-069, RU2012-073, RU2012-083, and RU2012-091, which achieved the highest F v /F m , RU2012-121 as an example of an accession with mid-range F v /F m and RU2012-112 with the third-lowest F v / F m ( Figure S3).
In Experiment 2, average A sat over all chilling days (days 1-15) was lowest in M. x giganteus at 6 µmol m −2 s −1 , and greatest in accessions RU2012-069, RU2012-114 and RU2012-112, ranging from 8.5 to 12 µmol m −2 s −1 ( Figure  1). Therefore, these three accessions were selected for further analysis in Experiment 3, and for statistical comparison to M. x giganteus. All plants were somewhat chlorotic at the end of the chilling period, though damage appeared least pronounced in accession RU2012-069 ( Figure S4).

| Improved chilling A sat in three topperforming M. sacchariflorus accessions
Decline in A sat upon chilling was seen in Experiments 2 and 3, and was matched by decline in Φ PSII , V max , V pmax , and g s (Figures 2-4). A sat was significantly (p < 0.05), and up to 2fold greater than M. x giganteus in accessions RU2012-069 and RU2012-114 prior to chilling (day 0), on the first (day 1) and on the last (day 15) day of chilling (Table 1, Figure 2a), and throughout the entire chilling period (days 1-15) (Table 1, Figure 4a).
In accessions RU2012-069 and RU2012-114, A sat declined continuously over the course of days 1-15 (m < 0 for A sat , Figures 2a, 4b) at a rate of 0.16-0.27 μmol m −2 s −1 per day. This caused A sat to drop by more than half in accession RU2012-069 over this time period (Table 1). In contrast, in RU2012-112, A sat declined from days 1-4, then increased from days 4-15 (m > 0 for A sat , Figures 2a, 4b). m for A sat was significantly greater than in M. x giganteus (p < 0.05),  Figure S1. Arrows identify high-performing accessions, along with the control M. x giganteus, which were selected to test in Experiment 3 indicating RU2012-112 was more successful than M. x giganteus at avoiding continued loss of A sat throughout days 1-15 (Figure 4b).

| Physiological basis for improved A sat
A sat may be affected by Φ PSII , V max , V pmax , and g s , and all of these variables showed similar declines throughout chilling (Figure 2). Throughout the chilling period (days 1-15), Φ PSII was marginally (p < 0.1) or significantly (p < 0.05) lowest in M. x giganteus relative to RU2012-069, RU2012-114 and RU2012-112 (Figure 4c).
Graphical analysis of A-c i curves clearly shows that A sat was not restricted by either g s or V pmax on chilling days (Figure 3). The operating point is the c i found when the external [CO 2 ] is equal to the current atmospheric level: 400 µmol/mol, i.e. the intersection of the response of A to c i and the stomatal supply function described by 1/ g s as shown in Figure 3. In all accessions, the operating point increased during chilling such that control markedly shifted to V max , which is the plateau of the A-c i response (Figure 3; Days 1 & 15). V max was marginally (p < 0.1) or significantly (p < 0.05) greater in RU2012-069 than in M. x giganteus prior to chilling (day 0), on the first (day 1) and on the last (day 15) day of chilling (Table 1, Figure 2d), and throughout the entire chilling period (days 1-15) ( Figure 4a). Further, the decline in V max over the course of days 1-15 was most severe (m < 0 for V max ) in M. x giganteus; in contrast m for V max was significantly greater (p < 0.05) in RU2012-114 and RU2012-112 than in M. x giganteus, indicating improved ability to maintain V max throughout chilling (Figure 4b).
Despite the apparently low sensitivity of A sat to g s ( Figure  3) during chilling, g s was significantly (p < 0.05) greater in accessions RU2012-069 and RU2012-114 than M. x giganteus throughout days 1-15 (Figure 4c).

| Recovery on return to 25°C
RU2012-114 showed the greatest initial recovery (Day 16), achieving A sat significantly greater (p < 0.05) and more than double that of M. x giganteus (Figure 2a, Table 1). However, recovery of M. x giganteus 1 week later (day 23) was markedly superior. Not only was A sat significantly (p < 0.05) , and V pmax (e), for Miscanthus x giganteus and accessions RU2012-114, RU2012-112, and RU2012-069 of Miscanthus sacchariflorus. Plants were kept at 25°C daytime/20°C nighttime on day 0 and days 16 through 23, and at 10°C daytime/5°C nighttime on days 1 through 15. Each point is the mean (±1 SE) of 10, 4, and 6 plants in panels a-b, c, and d-e, respectively. Lines from days 0 to 15 are best-fit curves, as explained in Figure S1 and up to 70% greater in M. x giganteus than any of the M. sacchariflorus accessions, it had also exceeded its own prechilling levels of A sat by 35%. M. x giganteus also showed the significantly (p < 0.05) greatest values for V pmax and g s on day 23 (Figure 2b,e and Table 1). Graphical analysis of A-c i curves on day 23 shows the operating point in all accessions had returned to co-limitation by V max and V pmax on day 23, as in day 0 ( Figure 3); this suggests the improved V pmax and g s of M. x giganteus on day 23 both contributed to its greater A sat .
x giganteus 'Illinois'. RU2012-069 and RU2012-114 both maintained A sat up to two times greater than M. x giganteus throughout chilling. Although RU2012-112 showed a low initial A sat on chilling, consistent with its low F v /F m ( Figure S3) in the field relative to other accessions, A sat improved with time in chilling, as did g s and V max (Figure 2a,b,d). This accession therefore showed a superior ability to acclimate to the chilling conditions. Recovery of A sat to 25°C was lower in M. x giganteus than in these three M. sacchariflorus accessions, but it showed superior recovery 1 week later. However, this required 7 days of 25°C, and such continuous periods of high temperatures would be unlikely at the high latitude limits of Miscanthus spp. cultivation or during the spring and autumn of lower latitude temperate climates. Under these conditions the far more rapid recovery of photosynthesis in the three Siberian accessions, in particular RU2012-114, in the first 24 hr would be far more important (Figure 2a), i.e. allowing these accessions to take advantage of brief periods of warmer weather. This is the first published physiological analysis of Siberian Miscanthus germplasm representing the high latitude extreme of the native range of M. sacchariflorus, which demonstrates an ability to maintain and even increase photosynthetic capacity during chilling. . Lines are average best-fit curves. Responses for V pmax -limited photosynthesis are shown only for ci <200 μmol/mol, whereas average fit responses for V max -limited photosynthesis are shown for ci <1,000 μmol/mol. Achieved photosynthesis (A) at a given ci will be the minimum of these two functions. Stomatal supply functions are traced from the average atmospheric [CO 2 ] to the predicted rate of photosynthesis (A) at the average operating ci for each day and Miscanthus accession. Plants were kept at 25°C daytime/20°C nighttime on day 0 and days 16 through 23, and at 10°C daytime/5°C nighttime on days 1 through 15. Note the Y-axis is adjusted for chilling days to make the graphs more legible. n = 5-6 for each curve 4.1 | Exceptional chilling tolerance found in three Siberian M. sacchariflorus accessions C 4 plants are highly vulnerable to chilling temperatures, i.e. above freezing but below 15°C (Friesen, Peixoto, Busch, Johnson, & Sage, 2014;Long, 1983;Long & Spence, 2013). Field experiments in southern England, near the northern limit of Z. mays production in Europe have shown that the combination of chilling temperatures and high light are particularly damaging to photosynthesis and early growth (Baker et al., 1989). The inter-specific hybrid C 4 grass M. x giganteus was long considered an exception to this rule, being able to avoid such damage and maintain photosynthetic efficiency during and following chilling events. This has been associated with the ability to up-regulate genes coding for key photosynthetic enzymes and chloroplast membrane components (Dohleman & Long, 2009;Friesen & Sage, 2016;Naidu, Moose, Al-Shoaibi, Raines, & Long, 2003;Spence et al., 2014;Wang, Naidu, Portis, Moose, & Long, 2008; and increasing capacity for non-photochemical quenching of PSII excitation energy upon chilling (Farage, Blowers, Long, & Baker, 2006;Spence et al., 2014).

F I G U R E 3 Summary of A-ci curves and stomatal supply functions for
We show the exceptional photosynthetic performance of M. x giganteus at low temperature can be surpassed. Field screening of F v /F m indicated a high potential for chilling tolerance within the Siberian Miscanthus germplasm ( Figure S3).

F I G U R E 4
Means and standard errors of the variables describing the chilling response (days 1-15) of Miscanthus x giganteus and accessions RU2012-114, RU2012-112, and RU2012-069 of Miscanthus sacchariflorus for A sat , g s , V max , V pmax , and Φ PSII . (a, c): The chilling days mean is the mean of a given variable over days 1-15. (b, d): m describes the rate of linear increase (m > 0) or decrease (m < 0) of a given variable during chilling, as explained in Figure S1. For the chilling days mean, # and * identify marginally significant (p < 0.1) and significantly (p < 0.05) greater value of a given variable throughout days 1-15 (one-tailed repeated measures ANOVA) for a M. sacchariflorus accession relative to M. x giganteus. For m, # and * identify marginally significant (p < 0.1) and significantly (p < 0.05) greater value for a M. sacchariflorus accession (RU2012-114, RU2012-112, and RU2012-069) relative to M. x giganteus (Dunnett's one-tailed test). n = 10 for A sat and g s , 6 for V max and V pmax , and 4 for Φ PSII  (Pignon, 2015). A previous study has identified M. sacchariflorus accessions which exceeded photosynthetic rates in M. x giganteus, achieving A sat of approximately 16 μmol m −2 s −1 at 15°C (Glowacka, Jorgensen, et al., 2015). Here we show that M. sacchariflorus accession RU2012-069 achieved a comparable A sat at 13 μmol m −2 s −1 on the first day of chilling (day 1) despite being measured at only 10°C, compared to 15°C in the study of Glowacka, Jorgensen, et al. (2015) (Figure 2a, Table 1). This photosynthetic rate is exceptional for any plant at 10°C, but especially so for a C 4 plant (Long & Spence, 2013;Sage, 2002).
While on recovery, M. x giganteus achieves a higher A sat after 7 days at 25°C (day 23), such temperatures for such a prolonged period would be a rare event in the spring and autumn of the current regions of cultivation of this crop ( Figure 2a, Table 1) (U.S.D.A, 2017). The discovery of accessions with far more rapid recovery on return to warm weather (day 16) than in the widely used 'Illinois' clone of M. x giganteus, suggests important genetic resources for producing new lines of M. x giganteus that can recover rapidly during warm spells interspersed with chilling events. Poor short-term recovery of A sat to 25°C (day 16) in M. x giganteus suggests it would underperform relative to these M. sacchariflorus accessions during transient warm periods early in the growing season, particularly relative to the rapid recovery of RU2012-114 ( Figure 2a, al., 2014). The 'Illinois' clone of M. x giganteus is known to strongly up-regulate expression of genes coding for proteins that are likely limiting to V max and A sat , including PPDK and Rubisco, at chilling and warm temperatures (Spence et al., 2014;. This might explain the higher photosynthetic rates it achieved a week after the end of the chilling treatment (day 23) (Figure 2a, Table 1). While the operating point of photosynthesis was V max -limited at low temperature, control of A sat was shared with g s and V pmax in M. x giganteus on day 23, contributing to the higher A sat (Figure 3) (Glowacka, Jorgensen, et al., 2015).

| Low-temperature A sat primarily driven by V max
For all accessions and prior to chilling (day 0), the operating point of the A/c i response was at the point of inflexion between limitation by the rate of PEP carboxylation (V pmax ) and regeneration (V max ), when measured at 25°C prior to the chilling treatment ( Figure 3). However, during chilling the operating point was on the plateau of the A/c i response for all accessions, showing biochemical limitation by V max and negligible stomatal limitation (Figure 3). Metabolic control analysis indicates that V max is limited by Rubisco, PPDK or the two in combination (Furbank et al., 1997). The higher A sat of RU2012-069 achieved during chilling compared to the other accessions suggests a likely increase in these proteins and/or their activation (Table 1, Figure 2a, d, Figure 3). Given the low stomatal limitation to A in all accessions, reducing g s should in theory reduce evapotranspiratory cooling and slightly warm the leaf, with minimal cost to A (i.e. shifting the operating point to the left of the A-c i curve) (Figure 3). The fact this does not occur in any of the accessions studied here may be an indication of impaired stomatal control at low temperature, a commonly noted feature of C 4 species under chilling conditions (Long & Spence, 2013).
The combination of chilling temperatures and high light are particularly damaging to photosynthesis in C 4 crops resulting in a decline in the maximum efficiency of photosystem II measurable as a decrease in F v /F m and in the operating efficiency of PSII (Φ PSII ) (Baker et al., 1989). At lower photon fluxes, the greater Φ PSII seen in M. sacchariflorus accessions would contribute to improved photosynthesis ( Figure 4c) (Glowacka et al., 2014;Glowacka, Jorgensen, et al., 2015). While greater Φ PSII may be indicative of reduced photoinhibition, and therefore improved chilling tolerance, this by definition will not restrict photosynthesis under the light-saturating conditions used here during the controlledenvironment experiments (Glowacka, Jorgensen, et al., 2015;Pignon, Jaiswal, Mcgrath, & Long, 2017).
sinensis. This is in accordance with the natural distribution of M. sacchariflorus, which extends further north into Asia than M. sinensis (Clark et al., 2014;Clifton-Brown et al., 2008). To maximize the effectiveness of breeding programs involving the M. sacchariflorus accessions described here, it will be necessary to pair them with high performance M. sinensis accessions: crosses with warmer temperate-adapted accessions could produce progeny able to grow under a wide range of climates and surpass both parent's yield potential (Dong et al., 2018;Farrell et al., 2006). This result suggests a means to add to the recent discovery of genetic markers associated with biomass yield in Miscanthus (Dong et al., 2018).
In conclusion, this study, the first to investigate chilling tolerance of photosynthesis in Siberian Miscanthus sacchariflorus has identified accessions with greater tolerance than the widely grown M. x giganteus 'Illinois' clone. These accessions show both higher photosynthetic capacity during a 2-week period of chilling and faster recovery over the first 24 hr of return to non-chilling temperatures. The study has therefore identified important genetic resources for this key trait of low temperature performance for use in the breeding of improved lines of M. x giganteus for cool climates.