Can chilling tolerance of C4 photosynthesis in Miscanthus be transferred to sugarcane?

The goal of this study was to investigate whether chilling tolerance of C4 photosynthesis in Miscanthus can be transferred to sugarcane by hybridization. Net leaf CO2 uptake (Asat) and the maximum operating efficiency of photosystem II (ФPSII) were measured in warm conditions (25 °C/20 °C), and then during and following a chilling treatment of 10 °C/5 °C for 11 day in controlled environment chambers. Two of three hybrids (miscanes), ‘US 84‐1058’ and ‘US 87‐1019’, did not differ significantly from the chilling tolerant M. ×giganteus ‘Illinois’ (Mxg), for Asat, and ΦPSII measured during chilling. For Mxg grown at 10 °C/5 °C for 11 days, Asat was 4.4 μmol m−2 s−1, while for miscane ‘US 84‐1058’ and ‘US 87‐1019’, Asat was 5.7 and 3.5 μmol m−2 s−1, respectively. Miscanes ‘US 84‐1058’ and ‘US 87‐1019’ and Mxg had significantly higher rates of Asat during chilling than three tested sugarcanes. A third miscane showed lower rates than Mxg during chilling, but recovered to higher rates than sugarcane upon return to warm conditions. Chilling tolerance of ‘US 84‐1058’ was further confirmed under autumn field conditions in southern Illinois. The selected chilling tolerant miscanes have particular value for biomass feedstock and biofuel production and at the same time they can be a starting point for extending sugarcane's range to colder climates.


Introduction
Sugarcane (Saccharum spp. hybrid) is one of the world's most important crops. In 2013, sugarcane produced 1.9 billion tonnes of biomass, more than any other single crop, for sugar and bioenergy via ethanol and electricity (Botha & Moore, 2014;FAOSTAT, 2014). Sugarcane is grown commercially in over 100 countries on a total of 26.5 million hectares (data for 2013; FAOSTAT, 2014). However, commercial sugarcane production is limited to tropical and subtropical environments, due to the crop's limited tolerance to cold; southern Louisiana, USA, is perhaps where commercial production is the most challenged by cold. Agronomic success of modern sugarcane varieties can be explained by effective introgression of genes from wild germplasm, particularly from S. spontaneum into S. officinarum, starting in the early 1900s (Daniels & Roach, 1987;D'Hont et al., 1996;Hoarau et al., 2001;Piperidis et al., 2010;Andru et al., 2011). Additional genetic contributions from S. robustum, S. sinense and S. barberi are most likely present in modern sugarcane varieties (Daniels & Roach, 1987;Lima et al., 2002;Brown et al., 2007). The introgressed genes provide sources of disease resistance, vigor, ability to ratoon, and better yields under abiotic stresses (Mangelsdorf, 1960;Chen & Lo, 1988;Chen, 1993).
Miscanthus is a potentially valuable genetic resource for improving sugarcane. Particularly, Miscanthus is a source of resistance to downy mildew (Peronosclerospora sacchari), culmicolus smut (Sporisorium scitamineum; Chen & Lo, 1988), lesion nematodes (Pratylenchus spp.; E. Sacks, personal communication), as well as tolerance to drought and cold (Lo et al., 1978). Previously, the Taiwan Sugarcane Institute's collection of over 120 Miscanthus clones was evaluated to select parents for resistance to culmicolus smut and downy mildew and then to introduce the resistance into sugarcane by intergeneric hybridization with sugarcane and subsequent backcrossing to sugarcane (Chen & Lo, 1988). In the second backcross of the intergeneric hybrids to sugarcane (BC2), the downy mildew resistance inherited from Miscanthus was maintained and at the same time sugar content similar to the sugarcane parent's was restored.
C 4 is potentially more efficient than C 3 photosynthesis in its use of light, nitrogen, and water (Long, 1983;Long & Spence, 2013). However, in cooler environments, peak yields of most C 4 plants are markedly reduced, and only a few C 4 species (e.g., Miscanthus 9giganteus, Spartina anglica, and Spartina pectinata) can match the production of C 3 crops under cooler temperatures (Heaton et al., 2008;Long & Spence, 2013;Sage et al., 2014). In stark contrast to Miscanthus, sugarcane is noted to be one of the most chilling sensitive crop species. When grown at 10°C/5°C (day/night) for 11 days, sugarcane exhibited a >98% reduction in CO 2 assimilation relative to control plants grown at 25°C/20°C (day/night) . At 8-12°C, sugarcane CO 2 assimilation extrapolates to zero (Nose et al., 1994) while its close relative M. 9giganteus retains most of its photosynthetic capacity (Farage et al., 2006). Below 20°C, leaves of sugarcane grow slowly (Allison et al., 2007) and when grown below 10-15°C leaf elongation is negligible (1.4 mm day À1 )  or absent (Allison et al., 2007). Studies of the effect of cool temperatures on field-grown sugarcane in Hawaii revealed seasonal differences in chilling injury. In winter, minimum leaf temperatures of ca. 14°C were associated with inhibition, and in summer, minimum temperatures as high as ca. 20°C were sufficiently cool to reduce the maximal photosynthesis capacity (Grantz, 1989). For these reasons, sugarcane is usually grown between latitude 30N and 35S, but its northern range limit has not been firmly established. In contrast to sugarcane, the exceptional chilling tolerance of Miscanthus allows it to be grown with success in the cooler climates of NW Europe (Beale & Long, 1995;Lewandowski et al., 2000;Clifton-Brown et al., 2001;Je _ zowski et al., 2011) and the Midwest USA (Heaton et al., 2008). Thus, hybridization of sugarcane with chilling tolerant Miscanthus germplasm could theoretically provide a means to develop more chilling tolerant sugarcane.
For perennial plants adapted to temperate environments, overwintering requires survival at temperatures that are not conducive to growth, and especially tolerance to freezing. For Miscanthus, overwintering in temperate environments is facilitated by dormancy. After a perennial crop survives the winter, its next challenge is to establish photosynthetically competent leaves as early in the growing season as possible and maintain photosynthesis as late into the growing season as temperatures will allow, thereby maximizing carbon assimilation over the season (Long & Spence, 2013). As demonstrated by Farrell et al.'s (2006), productivity model, extending the growing season for Miscanthus by 30 days, would result in up to 25% higher yield. However, earlier canopy development will only result in higher yield if early growth and low temperature tolerance are combined. Leaf necrosis resulting from late frosts during the beginning of spring can greatly retard canopy establishment because few nutrients will remain in the rhizomes after initial growth for a second cohort of shoots (Kaiser, 2014;K. Głowacka data not published). For these reasons, early emergence of leaves that are photosynthetically competent at chilling temperatures is the crucial feature of highly productive perennial grasses in temperate climate.
For the present study, we chose three Saccharum sp. 9 Miscanthus sp. hybrids that had been previously observed to overwinter as far north as Booneville, Arkansas (35°05 0 N, 93°59 0 W), with a minimum winter air temperature of À14°C and an average monthly temperature of À0.3°C in the coldest month of 2000 (Burner et al., 2009). Although rhizomes of these three selected miscanes showed cold tolerance for overwintering, their ability to maintain photosynthetic capacity in aboveground tissues under chilling conditions (≤14°C) has not been determined.
This study examines whether (i) the chilling tolerance of C 4 photosynthesis in Miscanthus is apparent in the hybrids under controlled and field conditions; (ii) the hybrids show improved recovery of photosynthesis upon return to warm conditions relative to sugarcane; and (iii) the hybrids retain the high photosynthetic capacity of sugarcane under warm conditions, that is, is chilling tolerance achieved at the expense of capacity under warm conditions?

Propagation of plant material
The sugarcanes and miscanes were propagated from 10 to 15 cm stem sections with mature buds at the nodes. With the sheathing leaves removed, bare stem pieces were planted vertically in cell trays (38-cell star trays; T.O. Plastics) with one stem piece per cell containing a peat-, bark-, and perlite-based growing medium (Metro-Mix 900; Sun Gro Horticulture, Agawam, MA, USA). Cells were kept initially in greenhouse under mist (VibroNet Mister Nozzle, 20.1 l h À1 ; Netafime; Tel Aviv, Israel) for 10s every 10 min during daylight hours. The day/night cycle followed natural light with the temperatures 25°C/21°C. After the new shoots appeared, clonal divisions were transferred to 1 l pots of the same soil mix (mini-treepot # MT38; Stuewe & Sons, Tangent, OR, USA) for subsequent use in controlled environment chambers. Mxg was propagated from 3-cm-long rhizome pieces with visible roots and nodes, then grown in 1 l pots as described above. Z. mays seeds were sown directly into 1 l pots. When the plants were transferred to minitreepots, an all-purpose slow release fertilizer was added following the manufacturer's instructions (Osmocote Classic, 8-9 mo 13-13-13; Everris NA, Inc., Dublin, OH, USA), and one teaspoon of additional ferrous sulfate heptahydrate per pot was added (QC Corporation, Girardeau, MO, USA). Prior to transfer to controlled environment cabinets, plants were grown in a greenhouse at~25°C. Throughout, soil moisture content was maintained by watering to field capacity daily.

Gas exchange and chlorophyll fluorescence in controlled environment chambers
To mimic the type of chilling that might develop during spring after leaf emergence or to expanded leaves in the autumn, the plants were grown at: 25°C day/20°C night (warm) for 10 days, followed by 11 days at 10°C/5°C (chilling), and then returned to 25°C/20°C for one day. From three to six replicate, plants for each of the eight accessions were grown in two controlled environment chambers (Conviron PGC20; Controlled Environments, Winnipeg, Manitoba, Canada) equipped with an opened counter-balanced light canopy with ten high output dimmable metal halide bulbs (Mastercolor CDM_TP MV; PHILLIPS).  *Tai & Miller, 1988;Tai et al., 1991;Burner et al., 2009. Leaf photosynthetic gas exchange and modulated chlorophyll fluorescence were measured on the most recently emerged leaf on the main stem, as judged by ligule appearance. The positions of plants within the chambers were changed every 2 days to avoid confounding any undetected variation in environment within the chamber with accessions. In both chambers, a 14-h-day/10-h-night cycle with 1000 lmol photons m À2 s À1 and relative humidity of 65% was maintained. Leaf photosynthetic gas exchange and modulated chlorophyll fluorescence were measured in situ on the most recent fully expanded attached leaves, with an open gas exchange system incorporating differential infrared CO 2 and water vapor analyzers (LI-6400, LI-COR, Lincoln, NE, USA). With this system, the leaf was enclosed in a controlled environment cuvette, which tracked the light, temperature, and humidity in the controlled environment chamber. Chlorophyll pulse amplitude modulated fluorescence was measured simultaneously with a fluorometer incorporated into the cuvette lid (LI-6400-40; LI-COR, Inc.). Measurements were conducted in ambient air (210 mmol mol À1 O 2 and 390 lmol mol À1 CO 2 ), 1000 lmol m À2 s À1 photon flux and 65% relative humidity. Leaf temperature was maintained at the growth temperature for each accession and treatment. Actinic light was supplied by light-emitting diodes (90% red light, 630 nm; 10% blue light, 470 nm). To maximize the fluorescence emissions, the fluorometer parameters (e.g., flash intensity and duration) were adjusted and the multiphase protocol was used (Genty et al., 1989). These measurements were taken in warm conditions (25°C) just prior to the chilling treatment, immediately after the temperature was reduced to 10°C (day 0), during each of the 11 days at 10°C (except days 6, 8, and 10), and finally one day after transferring the plants back to 25°C (12th day of the experimentrecovery). All measurements were taken during the daylight hours on light-adapted leaves when a steady state CO 2 and water vapor flux was obtained in the cuvette (20-50 min). For each accession, from three to six replicate plants were measured for each treatment. From these procedures, measurements of lightsaturated leaf net CO 2 uptake per unit leaf area (A sat ), quantum yield of photosystem II (Ф PSII ), stomatal conductance to water vapor (g s ), and intercellular CO 2 concentration (c i ) were obtained as described previously (Bernacchi et al., 2003).

Field experiment
A field experiment was established on May 22, 2013 at the Dixon Springs Agricultural Center (37°26 0 18″N, 88°39 0 56″ W; USDA hardiness zone 6/7 border) from plugs propagated in cells, as detailed above. Plots were single rows of eight plants (ramets), spaced on 0.9 m centers. The trial was a randomized complete block design with four replicates. The soil was a silt loam (fine-silty, mixed, active, mesic, Oxyaquic, Fragiudalfs, 1-3% organic matter). Air temperature at a height of 2 m above ground was recorded every 10 s by the meteorological station which was located 400 m away from the planting (WARM, 2014).
At the end of first growing season on October 23, 2013 and October 24, 2013, photosynthetic leaf CO 2 uptake was measured, as described above, in ambient air (210 mmol mol À1 O 2 and 400 lmol mol À1 CO 2 ), at 1500 lmol m À2 s À1 of photon flux and 65% relative humidity. Leaf temperature was maintained in the cuvette at the average ambient temperature of 13°C.

Data analysis
Statistical analyses were performed with SAS PROCEDURE GLM (SAS v. 9.3, SAS Institute, Cary, NC, USA). For data from the growth chamber experiment, two-way analyses of variance were conducted to determine whether genotype and/or treatment had significant effects on A sat , Ф PSII , g s , and c i /c a . Analyses were conducted for each day of the experiment. Dunnett's multiple comparison tests were used to compare each genotype with the chilling tolerant control, Mxg (Fig. 1), or with the chilling sensitive control, Saccharum sp. 'L79-1002' (Fig. 2). For the field experiment, in which data were collected during two consecutive days, two-way analyses of variance were performed to assess whether there were significant differences between days and tested genotypes. As the date of measurement in the field experiment did not significantly affect the measured parameters A sat , g s , and c i /c a (P = 0.08; P = 0.29; P = 0.65, respectively), the date factor was omitted in the final analyses. For comparisons of means between the Mxg control and other genotypes, Dunnett's was used (Fig. 3c-e). Fig. 1 (a-b) Light-saturated leaf net CO 2 uptake rate (A sat ), (c-d) quantum yield of photosystem II (Ф PSII ), (e-f) stomatal conductance to water vapor (g s ), and (g-h) ratio of intercellular to atmospheric CO 2 concentration (c i /c a ) for warm conditions prior to chilling treatment, after transfer of plants to chilling (day 0), on 11th day of chilling treatment and one day after transfer of plants back to the warm conditions (12th day of experimentrecovery). Left panels (a, c, e, and g) are absolute values; right panels (b, d, f, and h) indicate responses to treatments expressed as a percentage of rates observed in warm conditions before the chilling treatment (i.e., percentage of control, white bars in the adjacent left panels). Plants were grown at 25°C/20°C (warm) day/night, with 14-h-day/10-hnight cycle under 1000 lmol photons m À2 s À1 , excepting the 11 days at 10°C/5°C (chilling). In all panels, accessions are ordered according to A sat on day 12 of the experiment (panel a; from highest to lowest; fourth bar (black fill) for each genotype).

Chilling experiment in controlled environment chambers
Two of three miscanes, 'US 84-1058' and 'US 87-1019', did not differ significantly for A sat from the chilling tolerant control, Mxg, after being subjected to chilling conditions (10°C/5°C) for 11 days, and also after subsequently being returned to warm temperatures for one day (25°C/20°C) (Fig. 1a). For Mxg grown at 10°C/5°C for 11 days, A sat was 4.4 lmol m À2 s À1 , while for miscane 'US 84-1058' and 'US 87-1019', A sat was 5.7 and 3.5 lmol m À2 s À1 , respectively, which were not significantly different from that of Mxg. Moreover, 'US 87-1019' and 'US 84-1058' showed the highest rates of leaf CO 2 uptake prior to chilling treatment (Fig. 1a), indicating that their improved chilling tolerance was not at the expense of photosynthetic capacity under warm conditions. Additionally, at the beginning of the chilling treatment (day 0 of experiment when plants were transferred from 25°C to 10°C), miscane 'US 84-1058' had A sat and Ф PSII values twice (15.5 lmol m À2 s À1 and 0.14, respectively) and miscane 'US 87-1019' over 1.3 times (9.9 lmol m À2 s À1 and 0.08, respectively) as large as those of Mxg (6.3 lmol m À2 s À1 and 0.06, respectively) (Fig. 1a,c). In contrast, the remaining accessions after 11 day at 10°C/5°C exhibited up to 103% reduction in A sat (Saccharum sp. 'L79-1002') and up to 90% reduction in Ф PSII (Saccharum sp. 'L79-1002' and S. officinarum 'Louisiana Purple') relative to Mxg. Recovery values, one day after transfer back to 25°C, were similar for the two miscanes 'US 87-1019', 'US 84-1058' and Mxg, with A sat between 11.9 and 14.8 lmol m À2 s À1 and Ф PSII between 0.20 and 0.24. Relative to the prechilling conditions, Mxg, miscane 'US 87-1019' and 'US 84-1058' showed 79%, 58%, and 41% recovery of A sat , respectively (Fig. 1b), and 107%, 84%, and 54% recovery of Ф PSII , respectively (Fig 1d). The lowest recovery of photosynthesis on return to warm conditions was observed for the sugarcane S. officinarum 'Louisiana Purple' (4% of the prechilling A sat and 21% of prechilling Ф PSII ), which was significantly lower than the Mxg control (P < 0.001).
For the chilling tolerant control, Mxg, A sat , and Ф PSII declined for the first 2-3 day at 10°C/5°C, followed by a rebound over the following days to stabilize at 69% and 91% of the rates on initial transfer to 10°C/5°C, respectively (Fig. 2a,d). Two miscanes 'US 87-1019' and 'US-1058' also showed a rebound in photosynthesis starting from day 5 or 6, with the A sat and Φ PSII on the final day of 10°C/5°C treatment ending at~39-44% of the rates recorded on day 0 of chill treatment (Fig. 2a,  d). Only miscanes 'US 87-1019', 'US 84-1058', and the chilling tolerant control, Mxg, had significantly higher level of stabilization of A sat than the cane Saccharum sp. 'L79-1002', which failed to stabilize readings of A sat during 11 days of chilling treatment and ended at À2% of the day 0 rates, respectively (Fig. 2a-c). For Φ PSII, four accessions, miscane 'US 87-1019', miscane 'US-1058', Saccharum hybr. 'NCo310', and Mxg, stabilized at significantly higher levels than Saccharum sp. 'L79-1002' (Fig. 2d-f). However, the commercial cultivar Saccharum hybr. 'NCo310' was not significantly different from energycane 'L79-1002' for A sat , and c i /c a at the end of chilling period (after 11 day in 10°C/5°C). There were no significant differences between energycane 'L79-1002', S. officinarum 'Louisiana Purple' and Z. mays 'FR1064' for A sat , Φ PSII , g s , and c i /c a after 11 days in 10°C/5°C (Fig. 2c,f,i,l).
All lines increased in c i /c a on transfer to 10°C/5°C (Fig. 1g-h). Over the 11 days of chilling, all genotypes except miscane 'US-1058' had significantly higher c i /c a than the Mxg control. In contrast to the observations for c i /c a , g s of all genotypes decreased over the 11 days of chilling ( Fig. 1e-f). For three of eight genotypes, g s increased with the onset of chilling on day 0 and then decreased during the subsequent days of chilling treatment; the other five genotypes decreased in g s after only 20 min in chilling. One day after, the plants were transferred from chilling to warm conditions, miscanes 'US 87-1019' and 'US-1058' had a similar recovery of g s as Mxg, but the other genotypes did not (Fig. 1f). Values are expressed as a percentage of rates on day 0 measured immediately after transfer of plants from warm (25°C/20°C day/night) to chilling conditions. Light-saturated leaf net CO 2 uptake rate (A sat ; a-c), quantum yield of photosystem II (Ф PSII ; d-f), stomatal conductance to water vapor (g s ; g-i), and ratio of intercellular to atmospheric CO 2 concentration (c i /c a ; j-l). Panels a, d, g, and j are miscanes 'US 84-1028', 'US 84-1058' and positive control Mxg 'Illinois', b, e, h, and k are miscane 'US 87-1019' and its cane parent Saccharum hyb. 'NCo310', c, f, i, and l are negative controls. Data are means AE SE (n = from 3 to 6, as indicted below Fig. 1g). An asterisk indicates a significantly higher value, and a cross indicates a significantly lower value in comparison with Saccharum sp.  When miscane 'US 87-1019' and its sugarcane parent 'NCo310' were grown at 25°C (warm), no significant differences in A sat and Ф PSII between parent and progeny were observed. Specifically, 'US 87-1019' grown at 25°C/20°C for 10 days had A sat and Ф PSII values of 25 lmol m À2 s À1 and 0.28, respectively, and similarly for Saccharum 'NCo310' the values were only slightly lower at 22 lmol m À2 s À1 and 0.27, respectively (Fig. 1a,  c). However, on the 11th day of chilling treatment, miscane 'US 87-1019' had a leaf CO 2 uptake rate that was three times that of its sugarcane parent. Additionally, miscane 'US 87-1019' had 2.8 times higher CO 2 assimilation than its sugarcane parent, 'NCo310', after transfer back to warm conditions (recovery).

Field experiment
Plants grown in the field at Dixon Springs, Illinois, in the autumn of 2013 experienced chilling temperatures below 10°C during the 17 days (4 days in September and 13 days in October) prior to measurements of gas exchange (Fig. 3a). The average air temperature in first 22 days of October was 15.5°C with the minimum at 0.1°C and maximum at 29.7°C. On 23 and 24 October during the part of day when measurements were taken, the average, low and high temperatures were as follows: 9.2°C (7.9-10.0°C), 7.4°C (4.9-9.1°C), respectively (Fig. 3b). Of three miscanes examined in late October in the field, one accession, 'US 84-1058', had comparable A sat , g s , and c i /c a to the chilling tolerant control, Mxg (Fig. 3c-e). A sat for Mxg and miscane 'US 84-1058' was 13.1 lmol m À2 s À1 (Fig. 3c). The A sat of Mxg was three time higher than A sat of to two other miscanes, 'US 87-1019' and 'US 84-1928' and six times higher than A sat of S. officinarum 'Louisiana Purple'.

Discussion
Can chilling tolerance of C 4 photosynthesis in Miscanthus be transferred to sugarcane?
To answer the question whether chilling tolerance of C 4 photosynthesis in Miscanthus can be transferred to sugarcane, we compared the gas exchange readings for three miscanes with results obtained for Mxg when grown in the same chilling conditions. As the Miscanthus parents of the miscane hybrids were not available, we chose for a chilling tolerant control the previously studied Miscanthus genotype, Mxg, which has been shown to have exceptionally efficient photosynthesis at low temperature for a high yielding C 4 plant (Long & Spence, 2013). Unexpectedly, we identified two Saccharum 9 Miscanthus hybrids (miscanes 'US 84-1058' and 'US 87-1019') that were not significantly different from the chilling tolerant Mxg when gas exchanges values were compared for plants grown at 10°C/5°C for 11 days in controlled environment chambers (Fig. 1). Furthermore, we were able to confirm photosynthetic chilling tolerance of miscane 'US 84-1058' under field conditions during the autumn in southern Illinois (Fig. 3), indicating that the chilling tolerance of C 4 photosynthesis in Miscanthus could be transferred to sugarcane. However, not all of the miscane genotypes that we tested had chilling tolerance. For example, miscane 'US 84-1028' lacked chilling tolerant photosynthesis, with A sat after 11 days of growth at 10°C/5°C that was a fraction of that of the best miscane in our study, 'US 84-1058' (Fig. 1). Differences among miscane genotypes for chilling tolerance could be due to different levels of chilling tolerance contributed by the Miscanthus parents and/or interactions between genes from the Miscanthus and Saccharum parents. Moreover, the initial A sat before cold treatment was higher for the miscanes 'US 84-1058' and 'US 87-1019' than for Mxg (Fig. 1a). Thus, even though the relative responses (compared to their initial A sat values, Fig. 1b) of 'US 84-1058' and'US 87-1019' to chilling were intermediate to Mxg (high) and the sugarcane (low) controls, because they started with higher per se values, their final A sat values were similar to Mxg even after 11 days of chilling stress. Because the Miscanthus parents were not available for our current study, the comparison of F1 progeny with their non-Saccharum parents was unfortunately not possible. However, in a previous screen of Miscanthus germplasm, we found variation among species, and among genotypes within species for leaf extension rates and photosynthesis at low temperature . Thus, selection of parents and early generation miscane hybrids for chilling tolerant photosynthesis is advisable to successfully breed sugarcane for this trait.
The miscanes evaluated for photosynthesis at low temperature in the current study, 'US84-1028', 'US84-1058', and 'US87-1019', had previously been shown to resprout after being cut in the autumn and then allowed to over-winter in a field (overwintering ability when cut to 15 cm in the autumn) at Booneville, Arkansas (Burner et al., 2009). Interestingly, in the previous study 'US84-1028' and 'US84-1058' had acceptable overwintering ability and vigor in Arkansas, but 'US87-1019' lacked vigor after overwintering (Burner et al., 2009), whereas in the current study, photosynthesis of 'US84-1028' had poor tolerance to chilling, but 'US84-1058' and 'US87-1019' had good tolerance to chilling. Thus, chilling tolerant photosynthesis did not necessarily ensure good overwintering ability and vice versa. Cold tolerance is a complex set of component traits, and it will likely be advantageous to select for the different components when breeding sugarcane for adaptation to more temperate environments than where this crop is currently grown commercially.
The primary source of germplasm to improve sugarcane for tolerance to biotic and abiotic stresses has been S. spontaneum. However, the potential of S. spontaneum for breeding cold tolerant sugarcane may be more limited than Miscanthus because the latter has a more northern and temperate natural distribution (to hardiness zone 3 in eastern Russia). Additionally, as S. spontaneum is listed on the federal noxious weed list, from the United States Department of Agriculture Plants Database it cannot be evaluated under natural field conditions; only tests under controlled conditions are allowed. Although the natural distribution of S. spontaneum does not extend as far north in Asia as Miscanthus, S. spontaneum populations occur in environments as diverse as tropical lowlands in South-East Asia to the temperate midlatitudes of Honshu, Japan. Thus, the cold tolerance of sugarcane progeny derived from S. spontaneum is expected to be strongly associated with a given S. spontaneum accession's adaptation to the environment in which it originated. Brandes (1940) reported selecting cold tolerant S. spontaneum clones able to survive 18 days of below freezing temperatures. In contrast, Breaux & Irvine (1976) observed that when actively growing young seedlings of S. spontaneum were frozen for 8 h, only 10% survived the test; moreover, under natural freezing conditions, these selected survivors were no more resistant to low temperatures than unselected populations. Recently, Hale et al. (2014) evaluated 41 S. spontaneum accessions for survivability of below ground (stubble) buds following exposure to freezing temperatures of À7°C and identified four accessions that had more ratoon cold tolerance than the most tolerant commercial sugarcane variety tested (HoCP 96-540). However, when progeny of ten sugarcane 9 S. spontaneum hybrids was examined for stubble cold tolerance in Arkansas, none of the progeny survived (Burner et al., 2009). When the energycane variety 'Ho 02-113', a hybrid of S. spontaneum 'SES 234' from the Himalayan foothills of northern India and a leading commercial sugarcane variety LCP 85-384 (S. officinarum 9 S. spontaneum 9 S. barberi 9 S. sinense) (Milligan et al., 1994;Hale et al., 2013), was grown for six days at 12°C/5°C and then moved back to 25°C/20°C for one day, its ability to recover net CO 2 assimilation rate was 63% of the initial values at 25°C on day 0 (Friesen et al., 2014); however, this was 21% less than the recovery of Mxg when grown in the same conditions. In our experiment, the difference between the recovery level of Mxg and the best miscane tested, 'US87-1019', was 20.5%. Thus, both S. spontaneum and Miscanthus accessions have potential to improve chilling tolerance in sugarcane. Choice of S. spontaneum and Miscanthus parents with exceptional levels of cold tolerance will be key to further improving sugarcane for this trait. To date, crosses between Miscanthus and sugarcane have used genotypes of Miscanthus that grew well in subtropical or tropical environments, where sugarcane crossing was routinely conducted. Thus, there is an opportunity to make additional genetic gains in cold tolerance of sugarcane by selecting donor Miscanthus parents with greater cold adaptation, such as those that originate from northern China, northern Japan, and eastern Russia.

Physiological mechanisms of chilling tolerance
All eight accessions studied showed decreases in A sat , and Φ PSII and increases in c i /c a , after transfer from warm to chilling (Fig. 1). In chilling, stomata reacted by reducing their aperture when the concentration of CO 2 increased in the intercellular compartment as a consequence of decreased of CO 2 fixation. The time needed for the stomata to compensate for new internal environmental conditions associated with chilling differed among genotypes but all eventually did compensate. For three of eight genotypes, g s increased with the onset of chilling on day 0 and then decreased during the following chilling treatment; the other five genotypes decreased in g s after only 20 min in chilling. Thus, differences among accessions for relative decreases in A sat could not be explained by deficiency in CO 2 or by loss of stomatal function. On the other hand, comparisons of A sat readings with values of Φ PSII indicated that lightinduced chilling damage of PSII was the primary reason for low CO 2 assimilation in chilling susceptible accessions. Similar patterns of change for physiological parameters in response to low temperature were previously observed in Miscanthus and sugarcane (Farage et al., 2006;Głowacka et al., 2014).
Typically, in warm-temperate environments at the beginning and end of the growing season significant fluctuations of temperatures are recorded (e.g., USDA hardiness zone 8 in the southern USA). At these times of the year, after chilling or frost events at night, days with bright sun and relatively high temperatures can occur, that causes a potential challenge for photosynthetic apparatus in C 4 plants mainly because of the risk of photodamage to PSII (Long, 1983;Long et al., 1994;Allen & Ort, 2001).
In our study, in comparison with the chilling sensitive sugarcane lines, two miscanes ('US 84-1058' and 'US 87-1019'), were better in recovery of A sat , Φ PSII, and g s when returned to warm conditions for <24 h (Fig. 1). Additionally, 'US 84-1058' exhibited relatively high A sat when grown in natural cycles of chilling and warm temperatures during autumn in a field trial in southern Illinois. Thus, some miscanes appear to have an ability to protect the photosynthetic apparatus, permitting relatively high rates of photosynthesis at low temperature per se, and perhaps more importantly, enabling these undamaged plants to assimilate even more carbon on subsequent warm days.
Will chilling tolerant miscanes be less photosynthetically efficient in warm conditions than sugarcane?
A key challenge for plant breeders is to introgress desirable traits from wild or distantly related species into domesticated crops but at the same retain the favorable traits of the crop. When crossing a temperate-adapted species such as M. sinensis to a tropical crop such as sugarcane, there is the potential for tradeoffs associated with adaptation to different temperatures. Fortunately, however, we did not observe such a tradeoff for photosynthesis in the miscanes. For example, miscane 'US 87-1019' did not differ significantly from its sugarcane parent, 'NCo310', for A sat and Ф PSII when grown at 25°C/ 20°C but had significantly higher leaf photosynthetic gas exchange on the 11th day of chilling treatment ( Fig. 1). Previously, studies on Saccharum 9 Miscanthus hybrids showed that values of agronomic traits for miscane were intermediate between the two parents (Chen, 1953;Chen et al., 1983). Whether backcrossing can be employed without losing chilling tolerant photosynthesis will depend on the number of genes that confer the trait, their interaction with sugarcane genes and the identification of marker-trait associations for markerassisted selection.
We have shown that the chilling tolerance of C 4 photosynthesis in Miscanthus can be successfully transferred to sugarcane. The selected chilling tolerant miscanes, 'US 84-1058' and 'US 87-1019', have particular value for biomass feedstock and biofuel production, and at the same time they can be a starting point for extending sugarcane's range to higher latitudes and altitudes than current production regions. Although previous efforts to improve sugarcane with genes from Miscanthus have been few, these initial efforts point the way toward a bright future. In the last 30 years, our understanding of chilling tolerance in Miscanthus (Long & Spence, 2013;Głowacka et al., 2014), as well as the genetic diversity of this genus, has increased greatly (Hodkinson et al., 2002;Clark et al., 2014;Głowacka et al., 2015). This knowledge will facilitate improvement of sugarcane with genes from Miscanthus.