- Email: [email protected]
First frost: Effects of single and repeated freezing events on acclimation in Picea abies and other boreal and temperate conifers
Forest Ecology and Management 259 (2010) 1530–1535
Contents lists available at ScienceDirect
Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco
First frost: Effects of single and repeated freezing events on acclimation in Picea abies and other boreal and temperate conifers G. Richard Strimbeck a,*, Trygve D. Kjellsen b a b
Department of Biology, Norwegian University of Science and Technology, Realfagbygget, Høgskoleringen 5, 7491 Trondheim, Norway StatoilHydro Research Centre, 7005 Trondheim, Norway
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 November 2009 Received in revised form 15 January 2010 Accepted 15 January 2010
In experiments with needles of Picea abies, we tested the specific hypothesis that a single night of freezing acts as a signal that triggers a rapid increase in low temperature (LT) tolerance, and the more general hypothesis that repeated or prolonged freezing stimulates increased LT acclimation. In three growth chamber experiments involving acclimation under early- to mid-autumn light and temperature conditions followed by one or more freezing treatments, we found no significant effect of a single night of freezing on LT tolerance, and only limited and inconsistent effects of repeated and prolonged freezing. We also tested the effect of prolonged storage at 5 8C on LT tolerance on samples of three boreal and three temperate conifer species during acclimation under field conditions, and again found no consistent enhancement of LT tolerance attributable to freezing in either group. In agreement with our own and others’ anecdotal observations that some species can attain nearly maximal LT tolerance in the absence of freezing under field conditions, we conclude that freezing is neither required nor a major influence in LT acclimation, at least in well-studied boreal conifer species, while the effects of freezing on temperate conifers are not as well-documented. We conclude that freezing treatment of conifer seedlings to ensure sufficient hardiness for late planting seems to offer little practical advantage. ß 2010 Elsevier B.V. All rights reserved.
Keywords: Conifer Frost Freezing Cold Tolerance Hardiness Acclimation
1. Introduction In temperate and boreal environments, woody plant acclimation to low temperature (LT) begins in late summer in response to decreasing day length, coincident with the development of dormancy in buds (Bigras et al., 2001; Li et al., 2004). Studies of acclimation in a variety of woody plants show that exposure to chilling (0–10 8C) or freezing (below 0 8C) temperatures initiate further acclimation, with some suggesting that freezing temperatures are required for some species to acquire maximum LT tolerance (Sakai, 1966; some early studies of deciduous species reviewed in Weiser, 1970; studies of acclimation in conifers reviewed in Bigras et al., 2001; Beck et al., 2004; Søgaard et al., 2009). Consideration of the biophysics of freezing in plants suggests that an early frost event could act as an unambiguous signal of impending winter. When the environmental temperature falls below 0 8C, water in the plant may supercool by a few degrees, but with sufficiently low temperature ( 2 to 5 8C) freezing will begin in the xylem or other extracellular regions, possibly triggered by nucleators (Lee and Hammel, 1982). This results in dehydration
* Corresponding author. Tel.: +47 73551284; fax: +47 73596100. E-mail addresses: [email protected] (G.R. Strimbeck), [email protected] (T.D. Kjellsen). 0378-1127/$ – see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2010.01.029
of the unfrozen cells as extracellular ice masses grow by drawing water from the unfrozen cytoplasm. Simple calculations using the equation for melting point depression show that a 2 osmolar solution, a typical intracellular value for a conifer during acclimation (Tyree et al., 1978), will lose about 25% of its water in equilibrium with ice at 5 8C, a typical temperature for a ‘‘hard’’ frost event occurring during the autumn months and sufficient to trigger extracellular freezing in plants. Thus, extracellular freezing during a night of frost will result in rapid dehydration of the cells. This could act as a warning of further freezing to come, to which the plant could then respond by initiating the final stages of LT acclimation. These considerations give rise to the specific hypothesis, here called the first frost hypothesis, that a single occurrence of extracellular freezing triggers a rapid phase of acclimation that ensures that the plant is ready to tolerate more extreme freezing events. Here we report on the results of some experiments with Picea abies (L.) Karsten aimed at testing this hypothesis, and additional treatments and experiments to assess the effects of repeated freezing events on LT acclimation in needles of P. abies and other conifer species from both temperate and boreal regions. We also examined the effects of prolonged, continuous freezing on LT tolerance in conifers. Sakai (1966) reported that frozen storage at 3 to 5 8C dramatically improved LT tolerance in boreal Salix, Populus, and Larix twigs and buds. To test for similar effects in conifer foliage, with potential application in improving LT
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
tolerance for late planting, we assessed changes in LT tolerance after dark storage at 5 8C for periods of 2–4 weeks in adult trees of boreal and temperate Abies, Picea, and Pinus species, growing in an arboretum in a relatively mild winter climate in Trondheim, Norway. This experiment was an extension of our monitoring of LT tolerance during a complete acclimation–deacclimation cycle, as reported in a previous paper (Strimbeck et al., 2008). 2. Materials and methods 2.1. Growth chamber experiments In December 2004, 2-year-old dormant nursery-grown P. abies seedlings of a local, low-elevation seed source (Skogplanter MidtNorge AS, Skjerdingstad Nursery, Kva˚l, Norway, www.spmn.no; K2 seed source, Sør-Trøndelag province, 150–249 m elevation, lifted and stored at 2 8C in October, 2004) were planted in 1.5 l pots in a 1:1 mix of potting soil and perlite. These were grown in a greenhouse at 20 8C, under quartz halogen lights at a minimum 200 mmol m 2 s 1 PAR and 20/4 h D/N, augmented by natural light during the daytime hours, and watered three times a week with no additional fertilization for at least 7 weeks before beginning LT acclimation treatments. Programmable environmental chambers (Vo¨tsch Industrietechnik GmbH, Balingen-Frommen, FRG; http:// www.v-it.com/en/vit/start) were used to impose photoperiod and temperature treatments, without replication for chamber effects. In all experiments, light was increased and decreased linearly from darkness to 200 mmol m 2 s 1 PAR over the first and last hour of the day period, respectively, and maintained at the same value through the remainder of the daylight period. Humidity was maintained at or above 60%. Where possible, extra plants were kept in the growth chambers as edge and reserve plants. For freezing treatments, plants were removed from the growth chamber after the end of the daylight period and transferred to a programmable freezing chamber for a freeze/thaw cycle, then returned to the growth chamber before the beginning of the next light period. Plants were transferred in 5 cm thick styrene foam boxes to minimize environmental effects during transfer. We used a minimum temperature of 6 8C to ensure extracellular nucleation and substantial freeze dehydration of the needles, as is assumed in the first frost hypothesis. Although the freezing treatments were not applied independently for individual plants within treatment groups, freezer conditions were closely controlled and monitored and we observed no significant departures from the programmed temperature in any of the experimental freezing treatments, so for the purpose of analysis we treated individual seedlings as experimental replicates. After treatment, all the current-year growth (produced during growth in the greenhouse) of each plant was harvested for LT tolerance assessment. Plants were returned to growth chambers after harvest to maintain normal chamber conditions, and discarded at the end of each experiment. Relative LT tolerance was estimated by controlled freezing of needle sections measurement of relative electrolyte leakage (REL), and nonlinear curve fitting following previously published methods (Strimbeck et al., 2007). We prepared a homogeneous sample of 5 mm needle sections from all the current-year shoots from each plant, and cooled ca. 0.1 g subsamples to each of 10–16 test temperatures on each date, with samples held at 4 8C as the maximum temperature and minimum temperatures ranging from 30 to 60 8C depending on the anticipated level of LT tolerance of the tissue. This approach allowed us to evaluate individual LT tolerance at the time of harvest in each plant used in the experiment. Tm, the midpoint of a sigmoid curve fit to REL data for each plant, corresponds to LT50, the temperature resulting in 50% lethal injury, and was used as the response variable in analyses.
1531
Table 1 Treatment details and results of freezing experiments on low temperature tolerance of Picea abies needles. Values are the mean of eight (Experiments 1 and 2) or six (Experiment 3) 2-year-old plants in each treatment group. Tm is the midpoint of a sigmoid curve fitted to electrolyte leakage data for each plant, equivalent to LT50 when freezing results in complete mortality, as is the case in all of these experiments. Treatment
Tm, 8C
HSD testa
Experiment 1 (12/12 h D/N) 20/15 8C D/N, no frost 20/15 8C D/N, 1 night frost 15/5 8C D/N, no frost 15/5 8C D/N, 1 night frost
9.3 10.4 12.6 12.2
a ab b b
Experiment 2 (8/16 h, 15/5 8C D/N) No frost, day 9 1 night frost, day 9b 2 nights frost, day 9 9 nights frost, day 9 No frost, day 16 1 night frost, day 16b 2 nights frost, day 16 16 nights frost, day 16
14.2 13.0 17.0 21.4 13.6 15.4 14.1 18.4
a a ab b a ab a ab
Experiment 3 (8/16 h 10/2 8C D/N) No frost 1 night frost 7 nights frost 7 days continuous frost
34.7 37.6 34.7 39.3
a a a a
a Within each experiment, means with the same letter are not significantly different at a = 0.05. b Frost treatment on day 7, Tm determined 2 and 9 days after treatment.
We conducted three different acclimation experiments involving different combinations of photoperiod, acclimation temperature, and freezing treatment (Table 1). Experiment 1: This experiment was designed to test the first frost hypothesis, which predicts that a single early autumn frost event will trigger accelerated LT acclimation. After 7 weeks growth in the greenhouse, 60 plants were transferred to a growth chamber for short day-high temperature (SDHT; 12/12 h and 20/15 8C D/N) treatment. After 3 weeks in these conditions, half the plants were transferred to a second chamber for short day-low temperature (SDLT; 12/12 h and 15/5 8C D/N) treatment. These day length and temperature treatments we selected to approximate local conditions at autumnal equinox under warm to cool conditions. Four weeks after the start of SDLT treatment (14 weeks after planting), 8 SDHT and 8 SDLT plants were transferred to the freezer for a single night of frost, then returned to the growth chambers. For night frost, plants were transferred to the freezer at the end of the day period, held at 0 8C for 1 h, cooled to 6 8C over 2 h (3 8C h 1), held at 6 8C for 2 h, warmed to 0 8C over 2 h, and held at 0 8C for 1 h before transfer back to the growth chambers. Unfrozen controls were kept in both growth chambers. Plants were randomly assigned to treatment groups. Frost tolerance was assessed 7 days after night frost treatment in 7 plants from each treatment group. The results were analyzed as a 2 2 factorial with pretreatment temperature and frost treatment as the main effects and Tm as the response variable. The four treatment group means were compared a posteriori using Tukey’s HSD. Experiment 2: This experiment was designed to the compare the effects of single and repeated freezing events on LT acclimation in mid-autumn conditions. Plants were potted at the same time as those in the first experiment, and kept in the greenhouse until after completion of the first experiment. After 17 weeks growth in the greenhouse, 70 plants were transferred to a growth chamber under 8/16 h and 15/5 8C D/N. The 8-h photoperiod corresponds approximately to a calendar date of 1 November at the local latitude. After 4 weeks under these conditions, randomly assigned subsets of plants were exposed to four frost treatments: (1) no
1532
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
frost; (2) one night of frost on the 7th night of frost treatment; (3) two nights of frost on the 1st and 3rd nights of frost treatment; and (4) frost every night over a total of 16 days. For frost treatment, plants were transferred to the freezer 4 h into the dark period, held at 0 8C for 30 min, cooled to 6 8C over 3 h (2 8C h 1), held at 6 8C for 2 h, warmed to 0 8C over 3 h, and held at 0 8C for 3 h before transfer back to the growth chamber. LT tolerance was assessed 9 and 16 days after the beginning of frost treatments, in 8 plants from each treatment group. This design allowed comparison of all four treatments nine days after the first frost treatment in treatments 2–4, comparison of all but treatment 2 sixteen days after the first frost treatment, and assessment of a short-term (2 days after freezing) effect of a single night of frost in treatment 2. Data were analyzed as a one-way analysis of variance with eight treatment groups, with Tukey’s HSD used to compare sample means. We also used an a posteriori linear regression model to test for a linear relationship between nights of frost treatment and Tm in the four treatment groups assessed 9 days after the first night of frost in each group. Experiment 3: This experiment was designed to assess the effects of single and repeated freezing events after acclimation in the same photoperiod but at a lower temperature regime than in Experiment 2. Two-year-old trees of the same provenance were potted as above, but with vermiculite instead of perlite as a soil conditioner. After 4 weeks growth in the greenhouse, 32 plants were transferred to a growth chamber, maintained at 12/12 h and 20/5 8C for 6 weeks, 8/16 h and 15/5 8C for 1 week, and 8/16 h and 10/2 8C for 2 weeks. Six plants were randomly assigned to each of three different frost treatments: (1) one night of frost; (2) 7 nights of frost; and (3) 7 days continuously frozen, with 18 plants maintained in the growth chamber as unfrozen controls. In this experiment, night frost treatment was prolonged to 12 h. To minimize soil freezing during night frost, pots were placed in insulated boxes and packed in loose vermiculite to insulate the pots and trap latent heat of freezing. After transfer, plants receiving night frost treatment were held at 0 8C for 30 min, cooled to 6 8C over 75 min (4.8 8C h 1), held at 6 8C for 12 h, warmed to 0 8C over 75 min, and held at 0 8C for 1 h before transfer back to the growth chamber. After transfer, the freezer was immediately cooled to 6 8C to keep the plants in treatment 4 continuously frozen; these were kept in an insulated box in the freezer to damp temperature increases as the freezer was warmed to 0 8C temporarily to add or remove other plants for frost treatment. LT tolerance was assessed on the 8th day of the experiment (7 days after the first night of frost in treatment 1, the day after the last frost treatment in treatment 2, and after complete thawing of the plants in treatment 3). Data were analyzed in a one-way analysis of variance with Tukey’s HSD used to compare treatment means. 2.2. Storage experiment From August 2006 through May 2007, we monitored LT tolerance of needles of boreal-temperate species pairs of Abies (A. balsamea (L.) Miller and A. alba Miller), Picea (P. obovata Ledeb. and P. sitchensis (Bongard) Carrie´re), and Pinus (P. sylvestris L. and P. jeffreyi Balfour) growing in an arboretum in a relatively mild winter climate in Trondheim, Norway (Strimbeck et al., 2008). In addition to the results already reported, we tested the effect of prolonged storage at 5 8C on LT tolerance during the acclimation period from August through December 2006. On each collection date, a duplicate set of samples, consisting of whole shoots of current-year foliage from three trees of each of the six species, was placed in sealed plastic bags in an insulated styrene foam container, cooled slowly to 5 8C, and held at that temperature until the next collection date. These samples were then slowly thawed and prepared for freeze testing on the same day and using the same procedures as the next round of
fresh samples. Samples were collected at 2–4-week intervals, so the duration of frozen storage varied from 14 to 28 days. Sample preparation and freezing procedures were the same as those in the growth chamber experiments, with further details given previously (Strimbeck et al., 2008). Minimum test temperature ranged from 20 8C for samples collected in August to 65 8C for samples collected in December. In this experiment, we used two response variables to assess treatment effects: Tm, the midpoint of a symmetrical, sigmoid curve fitted to the REL data for each tree, and Ymax, the upper conductivity asymptote of the same curve. Necrosis develops only in samples with Ymax > 0.5; fully acclimated boreal species have Ymax below this threshold and can survive quenching in liquid nitrogen, while temperate species are always killed by freezing to some temperature below Tm (Strimbeck et al., 2007). Thus, both parameters are necessary for an adequate description and comparison of LT tolerance in fully hardy boreal and temperate species. (In the growth chamber experiments, Ymax values were always in the 0.7–0.9 REL range, indicating complete mortality at the lowest test temperatures, so that Tm is equivalent to LT50 and is by itself a meaningful comparative measure of LT tolerance.) We used analysis of variance with a factorial design including all permitted interactions between date, group (temperate versus boreal), freezing storage treatment, and species within group. We included a tree within species and group term and its interaction with freezing treatment as random effects to account for repeated measures effects, and these terms were used in test denominator synthesis as appropriate in testing group, species, and treatment effects. It is of interest to compare LT tolerance after freezing storage both to tolerance of the same trees on the original sample date, here called sample date comparisons, and to that of the same trees that have acclimated under field conditions during the storage interval, called run date comparisons. The latter comparisons were used to assess the effects of freezing on the rate of acclimation compared to that under field conditions. Both Tm and Ymax were analyzed for both sample date and run date comparisons, giving a total of four analyses of variance. Tm cannot be determined for some trees where there is not a clear sigmoid response (16 of 306 cases), and this was the case for all three trees of one species on two dates, so these dates were excluded from the Tm analyses to achieve a balanced design. 3. Results 3.1. Growth chamber experiments In Experiment 1, initial levels of frost tolerance were similar to those in late summer, during the early stages of frost hardening, and sufficient for the plants to survive an early frost event such as that applied in the experimental freezing treatment (Table 1). In the analysis of variance, there was a significant main effect of acclimation temperature (p < 0.0006) but no effect of a single night of frost treatment (p < 0.5903) or its interaction with acclimation temperature (p < 0.2643). In Experiment 2, the groups receiving one or two nights of frost treatment did not differ significantly from the unfrozen controls at any time. The only significant difference among means based on the HSD test was between the group given nightly frost treatments over 9 days and the other three treatment groups on day 9 of the experiment (Table 1). After 16 nights of freezing, mean Tm in the nightly frost group increased so that it no longer differed significantly from the other treatment groups. Similarly, Tm in the group given two nights of frost increased when LT tolerance was measured 13 days after the second night of frost, perhaps due to some dehardening in the relatively warm conditions in the growth chamber. The group given one night of frost showed no
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
1533
Table 2 Mean squares from analyses of variance of low temperature tolerance parameters of conifers during acclimation in the field and after frozen storage for 2–4 weeks at 5 8C. Tm and Ymax are the midpoint and upper asymptote, respectively, of a sigmoid curve fitted to electrolyte leakage data for each plant. Source
Ymax
Tm DF
Sample datea
Run dateb
Sample date
Run date
Date Groupc Spp[Group] Date group Date Spp[Group]
5 1 4 5 20
3746.3 3350.2 157.0 210.2 13.3
2664.6 4157.5 227.4 160.0 24.4
7 1 4 7 28
0.38177 4.61657 0.10605 0.24522 0.02087
0.28243 5.35934 0.14118 0.19518 0.01789
Freezing Date freezing Group freezing Date group freezing Freezing Spp[Group] Date freezing Spp[Group]
1 5 1 5 4 20
389.1 32.4 74.6 18.8 18.6 5.5
10.3 71.3 106.1 3.9 41.0 8.4
1 7 1 7 4 28
0.00536 0.04069 0.000004 0.01785 0.00629 0.00618
0.04395 0.03905 0.02703 0.03476 0.00895 0.00855
Tree[Spp,group](random) Freezing Tree[Spp, Group](random)
12 12
44.1 7.8
49.5 15.6
12 12
0.01397 0.00238
0.01587 0.00212
119 214
7.3
9.6
166 285
0.00263
0.00362
Residual Total
DF
Note: Cell shadings indicate F-test probabilities as follows: white: p > 0.05; light gray: 0.01 < p 0.05; dark gray: 0.001 < p 0.01; black: p 0.0001. a Comparison between samples collected on the same date, before and after freezing treatment; reflects effect of freezing treatment on overall level of low temperature tolerance. b Comparison between frozen samples and samples collected from the same trees at the end of the treatment period; reflects effect of freezing treatment on rate of acclimation. c Boreal group: Abies balsamea, Picea obovata, Pinus sylvestris. Temperate group: Abies alba, Picea sitchensis, Pinus jeffreyi. Three trees per species sampled over 9 dates from 15 August 2006 to 2 January 2007, freezing treatment imposed on duplicate samples collected on the first 8 dates.
increase in LT tolerance 2 days after treatment, but had a lower but not significantly different mean Tm than the unfrozen group 9 days after freezing treatment. Inspection of the data from Experiment 2 suggested a linear trend toward lower Tm with increasing number of freezing cycles, which we tested a posteriori in the four groups in which Tm was measured 9 days after the first night of freezing treatment. While the probability in this test is biased by the a posteriori approach, the relationship is strong enough (p < 0.0003, model R2 = 0.36) to provide some support to a general weak quantitative relationship between cycles of freezing and LT tolerance. However, the results again do not lend much support to the first frost hypothesis, which predicts a large step-change in LT tolerance after a single night of freezing. The slightly lower Tm values in Experiment 2 when compared with Experiment 1 suggest that the reduced day length promoted some LT acclimation, but are generally higher than Tms for trees at an equivalent day length under natural conditions (Strimbeck et al., 2008), suggesting that the growth chamber temperatures may have been high enough to hinder normal acclimation (see also Neuner et al., 1999). Consequently, in Experiment 3 we used lower day and night temperatures to test for frost effects at a more advanced stage of acclimation. The lower temperature acclimation treatment used in Experiment 3 resulted in lower Tm values than the frost two experiments, but there were no significant differences among group means (Table 1), and the means did not rank in order of frost exposure, although mean Tm in the continuously frozen group was almost 5 8C lower than in the unfrozen group. Again, the results of this experiment do not support the first frost hypothesis, in this case in P. abies seedlings in a mid-autumn scenario, and do not offer a clear interpretation of a quantitative frost effect as in the previous experiment. 3.2. Storage experiment The main effects of run date and group and their interaction were highly significant in all four analyses of variance, and the effect of species within group and its interaction with run date were also
consistently significant (Table 2). The effects of freezing treatment and its interactions with other terms were more varied and generally weaker. The date group treatment means, including some that were excluded from analysis due to missing data, are illustrated in Fig. 1. While the sample date comparisons indicate that freezing storage consistently decreased Tm in comparison with the values on the original sample date, the run date comparisons show no consistent improvement in LT tolerance over acclimation under field conditions, mostly in the absence of freezing temperatures, during the storage period. Based on this parameter, neither the boreal nor temperate group acclimated more rapidly in the freezer than trees under generally milder field conditions. Tm in field collected tissue in both groups decreased sharply following a period of several nights and one full day of below freezing temperatures in late October and early November, but it is worth noting that the boreal species were already nearly maximally LT tolerant before this period, as indicated by low Ymax values (Fig. 1) and, in one species, liquid nitrogen quench tolerance (Strimbeck et al., 2008) on 23 October. Over the same interval, the continuously frozen samples in the boreal group showed a similar but slightly smaller decrease in Tm, while further acclimation seems to have been delayed by frozen storage in the temperate group. The significant freezing treatment effect on Ymax in the run date analysis is due to a slightly higher (0.025 relative conductivity units) overall mean in the frozen storage group. There was no significant change in Ymin, the parameter describing the lower asymptote of the fitted curve (data not shown), indicating that there was no increase in baseline levels of electrolyte leakage during storage, so any changes in Ymax can be interpreted as a relative loss (increase in Ymax) or gain (decrease in Ymax) in LT tolerance. In the Ymax data for the boreal group, there are strong departures in both directions on some sample dates, which are probably the main explanation for the significant interaction terms in the analyses. These departures suggest an increase in LT tolerance due to freezing early in the acclimation process, but a relative loss of tolerance or delay in acclimation on later dates. Ymax increased during storage in samples of all three boreal species collected on 23 October, resulting in the strongest departure from the overall pattern. In the temperate group, mean Ymax remained above 0.7, reflecting these species’
1534
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
Fig. 1. Daily maximum, minimum, and 5-day running mean temperature at Værnes International Airport (A and D), and mean Tm (B and E) and Ymax (C and F) of needles of conifer species in boreal (A–C; Abies balsamea, Picea obovata, Pinus sylvestris) and temperate (D–F; Abies alba, Picea sitchensis, Pinus jeffreyi) groups during field acclimation and after storage at 5 8C. Values for stored samples are shown at the end of the storage period, and dashed lines indicate change from pre-storage values. Brackets indicate dates excluded from analyses of variance due to incomplete data. Tm and Ymax are the midpoint and upper asymptote, respectively, of a sigmoid curve fitted to electrolyte leakage data for each plant; Tm is interpreted as LT50 only when Ymax > 0.5.
continued sensitivity to LT stress, but with no consistent directional change due to freezing treatment. The boreal group dehardened slightly during an extended period of mild weather extending from late November through December, as shown by increases in Tm and Ymax (Fig. 1), and in this case frozen storage seems to have delayed or reduced the degree of dehardening. While there is some evidence for accelerated early acclimation, taken as a whole the results offer little support to the hypothesis that prolonged freezing at moderate subzero temperatures enhances LT tolerance or accelerates deep acclimation in either boreal or temperate conifer foliage. 4. Discussion Our results confirm those of numerous other studies showing increased frost tolerance with lower but above-freezing environmental temperatures, but they do not support the first frost hypothesis. We have produced no positive evidence that a single round of freezing has any real effect on the acclimation process in P. abies, and even repeated or prolonged freezing seems to produce inconsistent and relatively minor differences in LT tolerance. Beck et al. (2004) asserted that one or a few frost events can induce rapid acclimation in P. sylvestris, a boreal species, but this conclusion seems to be based primarily on observation of changes in LT tolerance before and after natural frost events in the field, without an unfrozen control. The experimental studies reported in the same paper show an effect of freezing only after numerous cycles of freezing, and also lack an unfrozen control. In similar combined field and experimental study, van Huystee et al. (1967) observed accelerated acclimation in Cornus sericea after a few nights of frost under field conditions, but used experimental treatments involving repeated cycles of freezing to increasingly lower temperatures to achieve full acclimation, so their experiments did not isolate the effects of one or a few nights of frost. These species grow primarily in boreal environments and develop a high degree of LT tolerance;
given the limitations of these earlier studies and our own results, it seems unlikely that freezing acts as an environmental signal in boreal species, at least as a general rule. Similar experiments with relatively frost sensitive temperate species might help clarify whether the first frost hypothesis applies to these less LT tolerant species. Even if freezing does not have a signaling function, various studies have suggested that repeated or prolonged freezing is important or necessary in attaining full acclimation. This idea seems to date to a canonical paper by Weiser (1970), asserting that ‘‘frost often appears be the triggering stimulus’’ for acclimation beyond that promoted by photoperiod alone, citing the work with C. sericea noted above (van Huystee et al., 1967), along with two other studies that also do not clearly isolate the effects of freezing treatment (Tumnanov and Krasavtsev, 1959; Howell and Weiser, 1970). In a review of acclimation in conifers, Bigras et al. (2001) cite several articles that support the view that freezing is necessary for full acclimation. In many cases these involve complex acclimation treatments with simultaneous changes in day length and chilling and freezing temperatures that do not completely isolate the effect of freezing. However, two recent studies found that repeated cycles of freezing over as much as 10 weeks, a much longer time frame than in the present work, may enhance frost tolerance by 5–10 8C in P. abies (Søgaard et al., 2009) and P. sylvestris (Beck et al., 2004), with unfrozen control groups or groups receiving only a few nights of frost still tolerant of temperatures in the 30 to 40 8C range. While our growth chamber experiments show some increases in LT tolerance associated with freezing, these are inconsistent and generally limited to few degrees. Thus the results of the relatively few studies that experimentally isolate freezing indicate that its effects are rather moderate and not necessary to establish sufficient LT tolerance for survival under natural conditions. Bigras et al. (2001) cite a few articles showing extensive acclimation in the absence of freezing temperatures, supporting
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
the general conclusion that freezing enhances acclimation in temperate species, but that boreal species acclimate fully under short photoperiod and chilling temperatures alone (Silim and Lavender, 1994). Our own study of acclimation of boreal and temperate conifers under relatively mild autumn conditions supports this interpretation; the boreal species acclimated to below 40 8C, and in some cases became liquid nitrogen quenchtolerant, before the first frost, while under the same conditions the temperate species acclimated to 10 to 20 8C, then acclimated to 20 to 30 8C after several nights of frost (Fig. 1; Strimbeck et al., 2008). Thus the boreal species seem to follow a fixed program for rapid acclimation in response to decreasing day length and chilling temperature, while temperate species may maintain some flexibility, perhaps related to an ability to extend the photosynthetic season (Schaberg, 2000). The results of the storage experiment confirm what we have already found in our previous analysis of the field acclimation data (Strimbeck et al., 2008), that there are strong and consistent changes in both LT tolerance parameters during acclimation, and that there are also consistent differences in the way the parameters in boreal versus temperate groups and the species within them change during acclimation. This experiment was partly inspired by the work of Sakai (1966), who found that storage at 3 to 5 8C for 10 days dramatically enhanced LT tolerance in stems of boreal Salix, Populus, and Larix species, while storage at 0 or 10 8C was less effective. However, in our case the results show no consistent effect of prolonged freezing on either boreal or temperate conifer foliage, and certainly not to the degree observed by Sakai. While Sakai provides evidence of sugar accumulation in continuously frozen tissue, the combined effects of low temperature and partial dehydration probably severely limit the rates of most metabolic processes, including those associated with acclimation. In practical terms, there seems to be little or no advantage to using freezing treatment in hardening protocols for seedling before late planting, at least for P. abies and likely other boreal species. While the results of Søgaard et al. (2009) and Beck et al. (2004) indicate some enhancement of LT tolerance by repeated freezing treatment over several weeks, chilling treatment alone resulted in sufficient tolerance to survive not only early frost, but temperatures as low 30 to 40 8C. The results of all these studies concern 2-year old to adult plants, so it remains a possibility that there is some advantage to freezing treatment of first-year plants, but this can only be assessed by further experimentation. Our results may also have some bearing on planting and management practice with respect to climate change. The boreal species’ lack of response to autumn frost indicates that LT acclimation in these species follows a rigid program driven by decreasing photoperiod and chilling temperatures. Trade-offs between the acclimated state and the ability to fix carbon during late autumn and winter (Schaberg et al., 1998; Schaberg, 2000) may affect growth and yield of these species as compared to species that maintain some flexibility in temperature responses, perhaps by
1535
using freezing events as a cue to initiate deeper hardening. Changes in the timing and frequency of autumn frost events, as generally predicted especially in boreal regions, may be one factor that affects productivity and competitiveness of these frost insensitive species in plantations and naturally regenerated forests. Acknowledgements This work was funded by internal sources at the Department of Biology, Norwegian University of Science and Technology. P. abies seedlings were donated by Skogplanter Midt-Norge AS, Skjerdingstad, Norway. We thank Paul Schaberg and three anonymous reviewers for helpful comments on an earlier version of the manuscript. References Beck, E.H., Heim, R., Hansen, J., 2004. Plant resistance to cold stress: mechanisms and environmental signals triggering frost hardening and dehardening. J. Biosci. 29, 449–459. Bigras, F.J., Ryyppo¨, A., Lindstro¨m, A., Stattin, E., 2001. Cold acclimation and deacclimation of shoots and roots of conifer seedlings. In: Bigras, F.J., Colombo, S.J. (Eds.), Conifer Cold Hardiness. Kluwer Academic, Dordrecht, pp. 57–88. Howell, G.S., Weiser, C.J., 1970. Environmental control of cold acclimation in apple. Plant Physiol. 45, 390. Lee, M.T., Hammel, H.T., 1982. Ice nucleating agents in xylem sap. Cryo-Letters 3, 314. Li, C.Y., Junttila, O., Palva, E.T., 2004. Environmental regulation and physiological basis of freezing tolerance in woody plants. Acta Physiol. Plant. 26, 213–222. Neuner, G., Ambach, D., Buchner, O., 1999. Readiness to frost harden during the dehardening period measured in situ in leaves of Rhododendron ferrugineum L. at the alpine timberline. Flora 194, 289–296. Sakai, A., 1966. Studies of frost hardiness in woody plants. II. Effect of temperature on hardening. Plant Physiol. 41, 353. Schaberg, P.G., 2000. Winter photosynthesis in red spruce (Picea rubens Sarg.): limitations, potential benefits, and risks. Arct. Antarct. Alp. Res. 32, 375–380. Schaberg, P.G., Shane, J.B., Cali, P.F., Donnely, J.R., Strimbeck, G.R., 1998. Photosynthetic capacity of red spruce during winter. Tree Physiol. 18, 271–276. Silim, S., Lavender, D., 1994. Seasonal patterns and environmental regulation of frost hardiness in shoots of seedlings of Thuja plicata, Chamaecyparis nootkatensis, and Picea glauca. Can. J. Bot. 72, 309–316. Søgaard, G., Granhus, A., Johnsen, O., 2009. Effect of frost nights and day and night temperature during dormancy induction on frost hardiness, tolerance to cold storage and bud burst in seedlings of Norway spruce. Trees (Berl.) 23, 1295– 1307. Strimbeck, G.R., Kjellsen, T.D., Schaberg, P.G., Murakami, P.F., 2007. Cold in the common garden: comparative low-temperature tolerance of boreal and temperate conifer foliage. Trees: Struct. Funct. 21, 557–567. Strimbeck, G.R., Kjellsen, T.D., Schaberg, P.G., Murakami, P.F., 2008. Dynamics of low-temperature acclimation in temperate and boreal conifer foliage in a mild winter climate. Tree Physiol. 28, 1365–1374. Tumnanov, I.I., Krasavtsev, O.A., 1959. Hardening of northern woody plants by temperatures below zero. Sov. Plant Physiol. 6, 663–673. Tyree, M.T., Cheung, N.S., MacGregor, M.E., Talbot, A.J.B., 1978. The characteristics of seasonal and ontogenetic changes in the tissue–water relations of Acer, Populus, Tsuga, and Picea. Can. J. Bot. 56, 635–647. van Huystee, R.B., Weiser, C.J., Li, P.H., 1967. Cold acclimation in Cornus Stolonifera under natural and controlled photoperiod and temperature. Bot. Gaz. 128, 200– 205. Weiser, C.J., 1970. Cold resistance and injury in woody plants. Science 169, 1269– 1278.
Contents lists available at ScienceDirect
Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco
First frost: Effects of single and repeated freezing events on acclimation in Picea abies and other boreal and temperate conifers G. Richard Strimbeck a,*, Trygve D. Kjellsen b a b
Department of Biology, Norwegian University of Science and Technology, Realfagbygget, Høgskoleringen 5, 7491 Trondheim, Norway StatoilHydro Research Centre, 7005 Trondheim, Norway
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 November 2009 Received in revised form 15 January 2010 Accepted 15 January 2010
In experiments with needles of Picea abies, we tested the specific hypothesis that a single night of freezing acts as a signal that triggers a rapid increase in low temperature (LT) tolerance, and the more general hypothesis that repeated or prolonged freezing stimulates increased LT acclimation. In three growth chamber experiments involving acclimation under early- to mid-autumn light and temperature conditions followed by one or more freezing treatments, we found no significant effect of a single night of freezing on LT tolerance, and only limited and inconsistent effects of repeated and prolonged freezing. We also tested the effect of prolonged storage at 5 8C on LT tolerance on samples of three boreal and three temperate conifer species during acclimation under field conditions, and again found no consistent enhancement of LT tolerance attributable to freezing in either group. In agreement with our own and others’ anecdotal observations that some species can attain nearly maximal LT tolerance in the absence of freezing under field conditions, we conclude that freezing is neither required nor a major influence in LT acclimation, at least in well-studied boreal conifer species, while the effects of freezing on temperate conifers are not as well-documented. We conclude that freezing treatment of conifer seedlings to ensure sufficient hardiness for late planting seems to offer little practical advantage. ß 2010 Elsevier B.V. All rights reserved.
Keywords: Conifer Frost Freezing Cold Tolerance Hardiness Acclimation
1. Introduction In temperate and boreal environments, woody plant acclimation to low temperature (LT) begins in late summer in response to decreasing day length, coincident with the development of dormancy in buds (Bigras et al., 2001; Li et al., 2004). Studies of acclimation in a variety of woody plants show that exposure to chilling (0–10 8C) or freezing (below 0 8C) temperatures initiate further acclimation, with some suggesting that freezing temperatures are required for some species to acquire maximum LT tolerance (Sakai, 1966; some early studies of deciduous species reviewed in Weiser, 1970; studies of acclimation in conifers reviewed in Bigras et al., 2001; Beck et al., 2004; Søgaard et al., 2009). Consideration of the biophysics of freezing in plants suggests that an early frost event could act as an unambiguous signal of impending winter. When the environmental temperature falls below 0 8C, water in the plant may supercool by a few degrees, but with sufficiently low temperature ( 2 to 5 8C) freezing will begin in the xylem or other extracellular regions, possibly triggered by nucleators (Lee and Hammel, 1982). This results in dehydration
* Corresponding author. Tel.: +47 73551284; fax: +47 73596100. E-mail addresses: [email protected] (G.R. Strimbeck), [email protected] (T.D. Kjellsen). 0378-1127/$ – see front matter ß 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2010.01.029
of the unfrozen cells as extracellular ice masses grow by drawing water from the unfrozen cytoplasm. Simple calculations using the equation for melting point depression show that a 2 osmolar solution, a typical intracellular value for a conifer during acclimation (Tyree et al., 1978), will lose about 25% of its water in equilibrium with ice at 5 8C, a typical temperature for a ‘‘hard’’ frost event occurring during the autumn months and sufficient to trigger extracellular freezing in plants. Thus, extracellular freezing during a night of frost will result in rapid dehydration of the cells. This could act as a warning of further freezing to come, to which the plant could then respond by initiating the final stages of LT acclimation. These considerations give rise to the specific hypothesis, here called the first frost hypothesis, that a single occurrence of extracellular freezing triggers a rapid phase of acclimation that ensures that the plant is ready to tolerate more extreme freezing events. Here we report on the results of some experiments with Picea abies (L.) Karsten aimed at testing this hypothesis, and additional treatments and experiments to assess the effects of repeated freezing events on LT acclimation in needles of P. abies and other conifer species from both temperate and boreal regions. We also examined the effects of prolonged, continuous freezing on LT tolerance in conifers. Sakai (1966) reported that frozen storage at 3 to 5 8C dramatically improved LT tolerance in boreal Salix, Populus, and Larix twigs and buds. To test for similar effects in conifer foliage, with potential application in improving LT
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
tolerance for late planting, we assessed changes in LT tolerance after dark storage at 5 8C for periods of 2–4 weeks in adult trees of boreal and temperate Abies, Picea, and Pinus species, growing in an arboretum in a relatively mild winter climate in Trondheim, Norway. This experiment was an extension of our monitoring of LT tolerance during a complete acclimation–deacclimation cycle, as reported in a previous paper (Strimbeck et al., 2008). 2. Materials and methods 2.1. Growth chamber experiments In December 2004, 2-year-old dormant nursery-grown P. abies seedlings of a local, low-elevation seed source (Skogplanter MidtNorge AS, Skjerdingstad Nursery, Kva˚l, Norway, www.spmn.no; K2 seed source, Sør-Trøndelag province, 150–249 m elevation, lifted and stored at 2 8C in October, 2004) were planted in 1.5 l pots in a 1:1 mix of potting soil and perlite. These were grown in a greenhouse at 20 8C, under quartz halogen lights at a minimum 200 mmol m 2 s 1 PAR and 20/4 h D/N, augmented by natural light during the daytime hours, and watered three times a week with no additional fertilization for at least 7 weeks before beginning LT acclimation treatments. Programmable environmental chambers (Vo¨tsch Industrietechnik GmbH, Balingen-Frommen, FRG; http:// www.v-it.com/en/vit/start) were used to impose photoperiod and temperature treatments, without replication for chamber effects. In all experiments, light was increased and decreased linearly from darkness to 200 mmol m 2 s 1 PAR over the first and last hour of the day period, respectively, and maintained at the same value through the remainder of the daylight period. Humidity was maintained at or above 60%. Where possible, extra plants were kept in the growth chambers as edge and reserve plants. For freezing treatments, plants were removed from the growth chamber after the end of the daylight period and transferred to a programmable freezing chamber for a freeze/thaw cycle, then returned to the growth chamber before the beginning of the next light period. Plants were transferred in 5 cm thick styrene foam boxes to minimize environmental effects during transfer. We used a minimum temperature of 6 8C to ensure extracellular nucleation and substantial freeze dehydration of the needles, as is assumed in the first frost hypothesis. Although the freezing treatments were not applied independently for individual plants within treatment groups, freezer conditions were closely controlled and monitored and we observed no significant departures from the programmed temperature in any of the experimental freezing treatments, so for the purpose of analysis we treated individual seedlings as experimental replicates. After treatment, all the current-year growth (produced during growth in the greenhouse) of each plant was harvested for LT tolerance assessment. Plants were returned to growth chambers after harvest to maintain normal chamber conditions, and discarded at the end of each experiment. Relative LT tolerance was estimated by controlled freezing of needle sections measurement of relative electrolyte leakage (REL), and nonlinear curve fitting following previously published methods (Strimbeck et al., 2007). We prepared a homogeneous sample of 5 mm needle sections from all the current-year shoots from each plant, and cooled ca. 0.1 g subsamples to each of 10–16 test temperatures on each date, with samples held at 4 8C as the maximum temperature and minimum temperatures ranging from 30 to 60 8C depending on the anticipated level of LT tolerance of the tissue. This approach allowed us to evaluate individual LT tolerance at the time of harvest in each plant used in the experiment. Tm, the midpoint of a sigmoid curve fit to REL data for each plant, corresponds to LT50, the temperature resulting in 50% lethal injury, and was used as the response variable in analyses.
1531
Table 1 Treatment details and results of freezing experiments on low temperature tolerance of Picea abies needles. Values are the mean of eight (Experiments 1 and 2) or six (Experiment 3) 2-year-old plants in each treatment group. Tm is the midpoint of a sigmoid curve fitted to electrolyte leakage data for each plant, equivalent to LT50 when freezing results in complete mortality, as is the case in all of these experiments. Treatment
Tm, 8C
HSD testa
Experiment 1 (12/12 h D/N) 20/15 8C D/N, no frost 20/15 8C D/N, 1 night frost 15/5 8C D/N, no frost 15/5 8C D/N, 1 night frost
9.3 10.4 12.6 12.2
a ab b b
Experiment 2 (8/16 h, 15/5 8C D/N) No frost, day 9 1 night frost, day 9b 2 nights frost, day 9 9 nights frost, day 9 No frost, day 16 1 night frost, day 16b 2 nights frost, day 16 16 nights frost, day 16
14.2 13.0 17.0 21.4 13.6 15.4 14.1 18.4
a a ab b a ab a ab
Experiment 3 (8/16 h 10/2 8C D/N) No frost 1 night frost 7 nights frost 7 days continuous frost
34.7 37.6 34.7 39.3
a a a a
a Within each experiment, means with the same letter are not significantly different at a = 0.05. b Frost treatment on day 7, Tm determined 2 and 9 days after treatment.
We conducted three different acclimation experiments involving different combinations of photoperiod, acclimation temperature, and freezing treatment (Table 1). Experiment 1: This experiment was designed to test the first frost hypothesis, which predicts that a single early autumn frost event will trigger accelerated LT acclimation. After 7 weeks growth in the greenhouse, 60 plants were transferred to a growth chamber for short day-high temperature (SDHT; 12/12 h and 20/15 8C D/N) treatment. After 3 weeks in these conditions, half the plants were transferred to a second chamber for short day-low temperature (SDLT; 12/12 h and 15/5 8C D/N) treatment. These day length and temperature treatments we selected to approximate local conditions at autumnal equinox under warm to cool conditions. Four weeks after the start of SDLT treatment (14 weeks after planting), 8 SDHT and 8 SDLT plants were transferred to the freezer for a single night of frost, then returned to the growth chambers. For night frost, plants were transferred to the freezer at the end of the day period, held at 0 8C for 1 h, cooled to 6 8C over 2 h (3 8C h 1), held at 6 8C for 2 h, warmed to 0 8C over 2 h, and held at 0 8C for 1 h before transfer back to the growth chambers. Unfrozen controls were kept in both growth chambers. Plants were randomly assigned to treatment groups. Frost tolerance was assessed 7 days after night frost treatment in 7 plants from each treatment group. The results were analyzed as a 2 2 factorial with pretreatment temperature and frost treatment as the main effects and Tm as the response variable. The four treatment group means were compared a posteriori using Tukey’s HSD. Experiment 2: This experiment was designed to the compare the effects of single and repeated freezing events on LT acclimation in mid-autumn conditions. Plants were potted at the same time as those in the first experiment, and kept in the greenhouse until after completion of the first experiment. After 17 weeks growth in the greenhouse, 70 plants were transferred to a growth chamber under 8/16 h and 15/5 8C D/N. The 8-h photoperiod corresponds approximately to a calendar date of 1 November at the local latitude. After 4 weeks under these conditions, randomly assigned subsets of plants were exposed to four frost treatments: (1) no
1532
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
frost; (2) one night of frost on the 7th night of frost treatment; (3) two nights of frost on the 1st and 3rd nights of frost treatment; and (4) frost every night over a total of 16 days. For frost treatment, plants were transferred to the freezer 4 h into the dark period, held at 0 8C for 30 min, cooled to 6 8C over 3 h (2 8C h 1), held at 6 8C for 2 h, warmed to 0 8C over 3 h, and held at 0 8C for 3 h before transfer back to the growth chamber. LT tolerance was assessed 9 and 16 days after the beginning of frost treatments, in 8 plants from each treatment group. This design allowed comparison of all four treatments nine days after the first frost treatment in treatments 2–4, comparison of all but treatment 2 sixteen days after the first frost treatment, and assessment of a short-term (2 days after freezing) effect of a single night of frost in treatment 2. Data were analyzed as a one-way analysis of variance with eight treatment groups, with Tukey’s HSD used to compare sample means. We also used an a posteriori linear regression model to test for a linear relationship between nights of frost treatment and Tm in the four treatment groups assessed 9 days after the first night of frost in each group. Experiment 3: This experiment was designed to assess the effects of single and repeated freezing events after acclimation in the same photoperiod but at a lower temperature regime than in Experiment 2. Two-year-old trees of the same provenance were potted as above, but with vermiculite instead of perlite as a soil conditioner. After 4 weeks growth in the greenhouse, 32 plants were transferred to a growth chamber, maintained at 12/12 h and 20/5 8C for 6 weeks, 8/16 h and 15/5 8C for 1 week, and 8/16 h and 10/2 8C for 2 weeks. Six plants were randomly assigned to each of three different frost treatments: (1) one night of frost; (2) 7 nights of frost; and (3) 7 days continuously frozen, with 18 plants maintained in the growth chamber as unfrozen controls. In this experiment, night frost treatment was prolonged to 12 h. To minimize soil freezing during night frost, pots were placed in insulated boxes and packed in loose vermiculite to insulate the pots and trap latent heat of freezing. After transfer, plants receiving night frost treatment were held at 0 8C for 30 min, cooled to 6 8C over 75 min (4.8 8C h 1), held at 6 8C for 12 h, warmed to 0 8C over 75 min, and held at 0 8C for 1 h before transfer back to the growth chamber. After transfer, the freezer was immediately cooled to 6 8C to keep the plants in treatment 4 continuously frozen; these were kept in an insulated box in the freezer to damp temperature increases as the freezer was warmed to 0 8C temporarily to add or remove other plants for frost treatment. LT tolerance was assessed on the 8th day of the experiment (7 days after the first night of frost in treatment 1, the day after the last frost treatment in treatment 2, and after complete thawing of the plants in treatment 3). Data were analyzed in a one-way analysis of variance with Tukey’s HSD used to compare treatment means. 2.2. Storage experiment From August 2006 through May 2007, we monitored LT tolerance of needles of boreal-temperate species pairs of Abies (A. balsamea (L.) Miller and A. alba Miller), Picea (P. obovata Ledeb. and P. sitchensis (Bongard) Carrie´re), and Pinus (P. sylvestris L. and P. jeffreyi Balfour) growing in an arboretum in a relatively mild winter climate in Trondheim, Norway (Strimbeck et al., 2008). In addition to the results already reported, we tested the effect of prolonged storage at 5 8C on LT tolerance during the acclimation period from August through December 2006. On each collection date, a duplicate set of samples, consisting of whole shoots of current-year foliage from three trees of each of the six species, was placed in sealed plastic bags in an insulated styrene foam container, cooled slowly to 5 8C, and held at that temperature until the next collection date. These samples were then slowly thawed and prepared for freeze testing on the same day and using the same procedures as the next round of
fresh samples. Samples were collected at 2–4-week intervals, so the duration of frozen storage varied from 14 to 28 days. Sample preparation and freezing procedures were the same as those in the growth chamber experiments, with further details given previously (Strimbeck et al., 2008). Minimum test temperature ranged from 20 8C for samples collected in August to 65 8C for samples collected in December. In this experiment, we used two response variables to assess treatment effects: Tm, the midpoint of a symmetrical, sigmoid curve fitted to the REL data for each tree, and Ymax, the upper conductivity asymptote of the same curve. Necrosis develops only in samples with Ymax > 0.5; fully acclimated boreal species have Ymax below this threshold and can survive quenching in liquid nitrogen, while temperate species are always killed by freezing to some temperature below Tm (Strimbeck et al., 2007). Thus, both parameters are necessary for an adequate description and comparison of LT tolerance in fully hardy boreal and temperate species. (In the growth chamber experiments, Ymax values were always in the 0.7–0.9 REL range, indicating complete mortality at the lowest test temperatures, so that Tm is equivalent to LT50 and is by itself a meaningful comparative measure of LT tolerance.) We used analysis of variance with a factorial design including all permitted interactions between date, group (temperate versus boreal), freezing storage treatment, and species within group. We included a tree within species and group term and its interaction with freezing treatment as random effects to account for repeated measures effects, and these terms were used in test denominator synthesis as appropriate in testing group, species, and treatment effects. It is of interest to compare LT tolerance after freezing storage both to tolerance of the same trees on the original sample date, here called sample date comparisons, and to that of the same trees that have acclimated under field conditions during the storage interval, called run date comparisons. The latter comparisons were used to assess the effects of freezing on the rate of acclimation compared to that under field conditions. Both Tm and Ymax were analyzed for both sample date and run date comparisons, giving a total of four analyses of variance. Tm cannot be determined for some trees where there is not a clear sigmoid response (16 of 306 cases), and this was the case for all three trees of one species on two dates, so these dates were excluded from the Tm analyses to achieve a balanced design. 3. Results 3.1. Growth chamber experiments In Experiment 1, initial levels of frost tolerance were similar to those in late summer, during the early stages of frost hardening, and sufficient for the plants to survive an early frost event such as that applied in the experimental freezing treatment (Table 1). In the analysis of variance, there was a significant main effect of acclimation temperature (p < 0.0006) but no effect of a single night of frost treatment (p < 0.5903) or its interaction with acclimation temperature (p < 0.2643). In Experiment 2, the groups receiving one or two nights of frost treatment did not differ significantly from the unfrozen controls at any time. The only significant difference among means based on the HSD test was between the group given nightly frost treatments over 9 days and the other three treatment groups on day 9 of the experiment (Table 1). After 16 nights of freezing, mean Tm in the nightly frost group increased so that it no longer differed significantly from the other treatment groups. Similarly, Tm in the group given two nights of frost increased when LT tolerance was measured 13 days after the second night of frost, perhaps due to some dehardening in the relatively warm conditions in the growth chamber. The group given one night of frost showed no
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
1533
Table 2 Mean squares from analyses of variance of low temperature tolerance parameters of conifers during acclimation in the field and after frozen storage for 2–4 weeks at 5 8C. Tm and Ymax are the midpoint and upper asymptote, respectively, of a sigmoid curve fitted to electrolyte leakage data for each plant. Source
Ymax
Tm DF
Sample datea
Run dateb
Sample date
Run date
Date Groupc Spp[Group] Date group Date Spp[Group]
5 1 4 5 20
3746.3 3350.2 157.0 210.2 13.3
2664.6 4157.5 227.4 160.0 24.4
7 1 4 7 28
0.38177 4.61657 0.10605 0.24522 0.02087
0.28243 5.35934 0.14118 0.19518 0.01789
Freezing Date freezing Group freezing Date group freezing Freezing Spp[Group] Date freezing Spp[Group]
1 5 1 5 4 20
389.1 32.4 74.6 18.8 18.6 5.5
10.3 71.3 106.1 3.9 41.0 8.4
1 7 1 7 4 28
0.00536 0.04069 0.000004 0.01785 0.00629 0.00618
0.04395 0.03905 0.02703 0.03476 0.00895 0.00855
Tree[Spp,group](random) Freezing Tree[Spp, Group](random)
12 12
44.1 7.8
49.5 15.6
12 12
0.01397 0.00238
0.01587 0.00212
119 214
7.3
9.6
166 285
0.00263
0.00362
Residual Total
DF
Note: Cell shadings indicate F-test probabilities as follows: white: p > 0.05; light gray: 0.01 < p 0.05; dark gray: 0.001 < p 0.01; black: p 0.0001. a Comparison between samples collected on the same date, before and after freezing treatment; reflects effect of freezing treatment on overall level of low temperature tolerance. b Comparison between frozen samples and samples collected from the same trees at the end of the treatment period; reflects effect of freezing treatment on rate of acclimation. c Boreal group: Abies balsamea, Picea obovata, Pinus sylvestris. Temperate group: Abies alba, Picea sitchensis, Pinus jeffreyi. Three trees per species sampled over 9 dates from 15 August 2006 to 2 January 2007, freezing treatment imposed on duplicate samples collected on the first 8 dates.
increase in LT tolerance 2 days after treatment, but had a lower but not significantly different mean Tm than the unfrozen group 9 days after freezing treatment. Inspection of the data from Experiment 2 suggested a linear trend toward lower Tm with increasing number of freezing cycles, which we tested a posteriori in the four groups in which Tm was measured 9 days after the first night of freezing treatment. While the probability in this test is biased by the a posteriori approach, the relationship is strong enough (p < 0.0003, model R2 = 0.36) to provide some support to a general weak quantitative relationship between cycles of freezing and LT tolerance. However, the results again do not lend much support to the first frost hypothesis, which predicts a large step-change in LT tolerance after a single night of freezing. The slightly lower Tm values in Experiment 2 when compared with Experiment 1 suggest that the reduced day length promoted some LT acclimation, but are generally higher than Tms for trees at an equivalent day length under natural conditions (Strimbeck et al., 2008), suggesting that the growth chamber temperatures may have been high enough to hinder normal acclimation (see also Neuner et al., 1999). Consequently, in Experiment 3 we used lower day and night temperatures to test for frost effects at a more advanced stage of acclimation. The lower temperature acclimation treatment used in Experiment 3 resulted in lower Tm values than the frost two experiments, but there were no significant differences among group means (Table 1), and the means did not rank in order of frost exposure, although mean Tm in the continuously frozen group was almost 5 8C lower than in the unfrozen group. Again, the results of this experiment do not support the first frost hypothesis, in this case in P. abies seedlings in a mid-autumn scenario, and do not offer a clear interpretation of a quantitative frost effect as in the previous experiment. 3.2. Storage experiment The main effects of run date and group and their interaction were highly significant in all four analyses of variance, and the effect of species within group and its interaction with run date were also
consistently significant (Table 2). The effects of freezing treatment and its interactions with other terms were more varied and generally weaker. The date group treatment means, including some that were excluded from analysis due to missing data, are illustrated in Fig. 1. While the sample date comparisons indicate that freezing storage consistently decreased Tm in comparison with the values on the original sample date, the run date comparisons show no consistent improvement in LT tolerance over acclimation under field conditions, mostly in the absence of freezing temperatures, during the storage period. Based on this parameter, neither the boreal nor temperate group acclimated more rapidly in the freezer than trees under generally milder field conditions. Tm in field collected tissue in both groups decreased sharply following a period of several nights and one full day of below freezing temperatures in late October and early November, but it is worth noting that the boreal species were already nearly maximally LT tolerant before this period, as indicated by low Ymax values (Fig. 1) and, in one species, liquid nitrogen quench tolerance (Strimbeck et al., 2008) on 23 October. Over the same interval, the continuously frozen samples in the boreal group showed a similar but slightly smaller decrease in Tm, while further acclimation seems to have been delayed by frozen storage in the temperate group. The significant freezing treatment effect on Ymax in the run date analysis is due to a slightly higher (0.025 relative conductivity units) overall mean in the frozen storage group. There was no significant change in Ymin, the parameter describing the lower asymptote of the fitted curve (data not shown), indicating that there was no increase in baseline levels of electrolyte leakage during storage, so any changes in Ymax can be interpreted as a relative loss (increase in Ymax) or gain (decrease in Ymax) in LT tolerance. In the Ymax data for the boreal group, there are strong departures in both directions on some sample dates, which are probably the main explanation for the significant interaction terms in the analyses. These departures suggest an increase in LT tolerance due to freezing early in the acclimation process, but a relative loss of tolerance or delay in acclimation on later dates. Ymax increased during storage in samples of all three boreal species collected on 23 October, resulting in the strongest departure from the overall pattern. In the temperate group, mean Ymax remained above 0.7, reflecting these species’
1534
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
Fig. 1. Daily maximum, minimum, and 5-day running mean temperature at Værnes International Airport (A and D), and mean Tm (B and E) and Ymax (C and F) of needles of conifer species in boreal (A–C; Abies balsamea, Picea obovata, Pinus sylvestris) and temperate (D–F; Abies alba, Picea sitchensis, Pinus jeffreyi) groups during field acclimation and after storage at 5 8C. Values for stored samples are shown at the end of the storage period, and dashed lines indicate change from pre-storage values. Brackets indicate dates excluded from analyses of variance due to incomplete data. Tm and Ymax are the midpoint and upper asymptote, respectively, of a sigmoid curve fitted to electrolyte leakage data for each plant; Tm is interpreted as LT50 only when Ymax > 0.5.
continued sensitivity to LT stress, but with no consistent directional change due to freezing treatment. The boreal group dehardened slightly during an extended period of mild weather extending from late November through December, as shown by increases in Tm and Ymax (Fig. 1), and in this case frozen storage seems to have delayed or reduced the degree of dehardening. While there is some evidence for accelerated early acclimation, taken as a whole the results offer little support to the hypothesis that prolonged freezing at moderate subzero temperatures enhances LT tolerance or accelerates deep acclimation in either boreal or temperate conifer foliage. 4. Discussion Our results confirm those of numerous other studies showing increased frost tolerance with lower but above-freezing environmental temperatures, but they do not support the first frost hypothesis. We have produced no positive evidence that a single round of freezing has any real effect on the acclimation process in P. abies, and even repeated or prolonged freezing seems to produce inconsistent and relatively minor differences in LT tolerance. Beck et al. (2004) asserted that one or a few frost events can induce rapid acclimation in P. sylvestris, a boreal species, but this conclusion seems to be based primarily on observation of changes in LT tolerance before and after natural frost events in the field, without an unfrozen control. The experimental studies reported in the same paper show an effect of freezing only after numerous cycles of freezing, and also lack an unfrozen control. In similar combined field and experimental study, van Huystee et al. (1967) observed accelerated acclimation in Cornus sericea after a few nights of frost under field conditions, but used experimental treatments involving repeated cycles of freezing to increasingly lower temperatures to achieve full acclimation, so their experiments did not isolate the effects of one or a few nights of frost. These species grow primarily in boreal environments and develop a high degree of LT tolerance;
given the limitations of these earlier studies and our own results, it seems unlikely that freezing acts as an environmental signal in boreal species, at least as a general rule. Similar experiments with relatively frost sensitive temperate species might help clarify whether the first frost hypothesis applies to these less LT tolerant species. Even if freezing does not have a signaling function, various studies have suggested that repeated or prolonged freezing is important or necessary in attaining full acclimation. This idea seems to date to a canonical paper by Weiser (1970), asserting that ‘‘frost often appears be the triggering stimulus’’ for acclimation beyond that promoted by photoperiod alone, citing the work with C. sericea noted above (van Huystee et al., 1967), along with two other studies that also do not clearly isolate the effects of freezing treatment (Tumnanov and Krasavtsev, 1959; Howell and Weiser, 1970). In a review of acclimation in conifers, Bigras et al. (2001) cite several articles that support the view that freezing is necessary for full acclimation. In many cases these involve complex acclimation treatments with simultaneous changes in day length and chilling and freezing temperatures that do not completely isolate the effect of freezing. However, two recent studies found that repeated cycles of freezing over as much as 10 weeks, a much longer time frame than in the present work, may enhance frost tolerance by 5–10 8C in P. abies (Søgaard et al., 2009) and P. sylvestris (Beck et al., 2004), with unfrozen control groups or groups receiving only a few nights of frost still tolerant of temperatures in the 30 to 40 8C range. While our growth chamber experiments show some increases in LT tolerance associated with freezing, these are inconsistent and generally limited to few degrees. Thus the results of the relatively few studies that experimentally isolate freezing indicate that its effects are rather moderate and not necessary to establish sufficient LT tolerance for survival under natural conditions. Bigras et al. (2001) cite a few articles showing extensive acclimation in the absence of freezing temperatures, supporting
G.R. Strimbeck, T.D. Kjellsen / Forest Ecology and Management 259 (2010) 1530–1535
the general conclusion that freezing enhances acclimation in temperate species, but that boreal species acclimate fully under short photoperiod and chilling temperatures alone (Silim and Lavender, 1994). Our own study of acclimation of boreal and temperate conifers under relatively mild autumn conditions supports this interpretation; the boreal species acclimated to below 40 8C, and in some cases became liquid nitrogen quenchtolerant, before the first frost, while under the same conditions the temperate species acclimated to 10 to 20 8C, then acclimated to 20 to 30 8C after several nights of frost (Fig. 1; Strimbeck et al., 2008). Thus the boreal species seem to follow a fixed program for rapid acclimation in response to decreasing day length and chilling temperature, while temperate species may maintain some flexibility, perhaps related to an ability to extend the photosynthetic season (Schaberg, 2000). The results of the storage experiment confirm what we have already found in our previous analysis of the field acclimation data (Strimbeck et al., 2008), that there are strong and consistent changes in both LT tolerance parameters during acclimation, and that there are also consistent differences in the way the parameters in boreal versus temperate groups and the species within them change during acclimation. This experiment was partly inspired by the work of Sakai (1966), who found that storage at 3 to 5 8C for 10 days dramatically enhanced LT tolerance in stems of boreal Salix, Populus, and Larix species, while storage at 0 or 10 8C was less effective. However, in our case the results show no consistent effect of prolonged freezing on either boreal or temperate conifer foliage, and certainly not to the degree observed by Sakai. While Sakai provides evidence of sugar accumulation in continuously frozen tissue, the combined effects of low temperature and partial dehydration probably severely limit the rates of most metabolic processes, including those associated with acclimation. In practical terms, there seems to be little or no advantage to using freezing treatment in hardening protocols for seedling before late planting, at least for P. abies and likely other boreal species. While the results of Søgaard et al. (2009) and Beck et al. (2004) indicate some enhancement of LT tolerance by repeated freezing treatment over several weeks, chilling treatment alone resulted in sufficient tolerance to survive not only early frost, but temperatures as low 30 to 40 8C. The results of all these studies concern 2-year old to adult plants, so it remains a possibility that there is some advantage to freezing treatment of first-year plants, but this can only be assessed by further experimentation. Our results may also have some bearing on planting and management practice with respect to climate change. The boreal species’ lack of response to autumn frost indicates that LT acclimation in these species follows a rigid program driven by decreasing photoperiod and chilling temperatures. Trade-offs between the acclimated state and the ability to fix carbon during late autumn and winter (Schaberg et al., 1998; Schaberg, 2000) may affect growth and yield of these species as compared to species that maintain some flexibility in temperature responses, perhaps by
1535
using freezing events as a cue to initiate deeper hardening. Changes in the timing and frequency of autumn frost events, as generally predicted especially in boreal regions, may be one factor that affects productivity and competitiveness of these frost insensitive species in plantations and naturally regenerated forests. Acknowledgements This work was funded by internal sources at the Department of Biology, Norwegian University of Science and Technology. P. abies seedlings were donated by Skogplanter Midt-Norge AS, Skjerdingstad, Norway. We thank Paul Schaberg and three anonymous reviewers for helpful comments on an earlier version of the manuscript. References Beck, E.H., Heim, R., Hansen, J., 2004. Plant resistance to cold stress: mechanisms and environmental signals triggering frost hardening and dehardening. J. Biosci. 29, 449–459. Bigras, F.J., Ryyppo¨, A., Lindstro¨m, A., Stattin, E., 2001. Cold acclimation and deacclimation of shoots and roots of conifer seedlings. In: Bigras, F.J., Colombo, S.J. (Eds.), Conifer Cold Hardiness. Kluwer Academic, Dordrecht, pp. 57–88. Howell, G.S., Weiser, C.J., 1970. Environmental control of cold acclimation in apple. Plant Physiol. 45, 390. Lee, M.T., Hammel, H.T., 1982. Ice nucleating agents in xylem sap. Cryo-Letters 3, 314. Li, C.Y., Junttila, O., Palva, E.T., 2004. Environmental regulation and physiological basis of freezing tolerance in woody plants. Acta Physiol. Plant. 26, 213–222. Neuner, G., Ambach, D., Buchner, O., 1999. Readiness to frost harden during the dehardening period measured in situ in leaves of Rhododendron ferrugineum L. at the alpine timberline. Flora 194, 289–296. Sakai, A., 1966. Studies of frost hardiness in woody plants. II. Effect of temperature on hardening. Plant Physiol. 41, 353. Schaberg, P.G., 2000. Winter photosynthesis in red spruce (Picea rubens Sarg.): limitations, potential benefits, and risks. Arct. Antarct. Alp. Res. 32, 375–380. Schaberg, P.G., Shane, J.B., Cali, P.F., Donnely, J.R., Strimbeck, G.R., 1998. Photosynthetic capacity of red spruce during winter. Tree Physiol. 18, 271–276. Silim, S., Lavender, D., 1994. Seasonal patterns and environmental regulation of frost hardiness in shoots of seedlings of Thuja plicata, Chamaecyparis nootkatensis, and Picea glauca. Can. J. Bot. 72, 309–316. Søgaard, G., Granhus, A., Johnsen, O., 2009. Effect of frost nights and day and night temperature during dormancy induction on frost hardiness, tolerance to cold storage and bud burst in seedlings of Norway spruce. Trees (Berl.) 23, 1295– 1307. Strimbeck, G.R., Kjellsen, T.D., Schaberg, P.G., Murakami, P.F., 2007. Cold in the common garden: comparative low-temperature tolerance of boreal and temperate conifer foliage. Trees: Struct. Funct. 21, 557–567. Strimbeck, G.R., Kjellsen, T.D., Schaberg, P.G., Murakami, P.F., 2008. Dynamics of low-temperature acclimation in temperate and boreal conifer foliage in a mild winter climate. Tree Physiol. 28, 1365–1374. Tumnanov, I.I., Krasavtsev, O.A., 1959. Hardening of northern woody plants by temperatures below zero. Sov. Plant Physiol. 6, 663–673. Tyree, M.T., Cheung, N.S., MacGregor, M.E., Talbot, A.J.B., 1978. The characteristics of seasonal and ontogenetic changes in the tissue–water relations of Acer, Populus, Tsuga, and Picea. Can. J. Bot. 56, 635–647. van Huystee, R.B., Weiser, C.J., Li, P.H., 1967. Cold acclimation in Cornus Stolonifera under natural and controlled photoperiod and temperature. Bot. Gaz. 128, 200– 205. Weiser, C.J., 1970. Cold resistance and injury in woody plants. Science 169, 1269– 1278.