Cold acclimation and deacclimation of Ptelea and Zanthoxylum (Rutaceae)

Sample Collection

We conducted assessments every 4 to 6 wk between August and April during the overwintering cycles of 2018–2019, 2019–2020, and 2020–2021. During the first 2 cycles, we collected samples of common hoptree from 4 plants (1 female and 3 males) in a wooded area in Ames, Iowa (42.0141 N, 93.3920 W) and, with permission, from 4 plants (2 females and 2 males) in Shield Prairie Wildlife Area near Muscatine, Iowa (41.4835N, 91.1246W). In 2020–2021, we obtained samples from 4 female and 4 male plants of common hoptree from the same wooded area in Ames, Iowa, and from 8 plants of common pricklyash from a wooded area in Boone, Iowa (42.0512 N, 93.7179 W). Sex of source plants was considered only for common hoptree. During the first 2 cycles, samples of western hoptree were collected with permission from 2 locations in California, Black Diamond Mines Regional Preserve in Antioch (37.9553 N, 121.8641 W) and Mount Diablo State Park in Contra Costa County (37.9047 N, 121.9288 W). We sampled 4 to 6 plants of western hoptree for each freezing experiment. For common pricklyash, which colonizes asexually, stems arising from the ground within 4 m (13 ft) of one another were considered 1 genetic entity. On 19 January 2021, in addition to sampling common hoptree and common pricklyash, samples from 3 trees each of Amur corktree (Phellodendron amurense Rupr. [Rutaceae]), silver maple (Acer saccharinum L. [Aceraceae]), tree of heaven (Ailanthus altissima (Mill.) Swingle [Simaroubaceae]), and staghorn sumac (Rhus typhina L. [Anacardiaceae]) were collected within a 1-km (0.62-m) radius of 42.0253 N, 93.6461 W.

Male flower of the dioecious species common hoptree (Ptelea trifoliata).

Female flower of the dioecious species common hoptree (Ptelea trifoliata).

Freeze–Thaw Procedure

We collected samples and stored them in moist paper towels inside plastic bags on ice from 14 to 24 h before freezing tests began. We considered an experimental unit to be a 1.5-cm (0.6-in) section of the most recent year’s growth, excluding 5 cm (2 in) at the tip of the branch. One stem section for each temperature was taken from each tree, with the total number of trees being 8 as detailed under the Sample Collection section above. Each tree was considered a replication. Each unit was put into a test tube with 150 μl (0.005 oz) of deionized water to facilitate ice nucleation. Freezing tests were conducted in a programmable freezer (Tenney Chamber, Thermal Product Solutions, New Columbia, Pennsylvania) as described by Wang and others (2008). Freeze–thaw tests were conducted on the day following collection. In 2019–2020, we collected samples of common hoptree on 2 September, 3 October, 7 November, 17 December, 23 January, 4 March, 13 April, and 18 May. We collected samples of western hoptree on 4 November, 27 January, and 8 March. In 2020–2021, we collected samples of common hoptree and common pricklyash on 20 August, 24 September, 12 November, 15 December, 19 January, 22 February, 23 March, and 27 April. Units from each plant were gradually frozen to various temperatures that were determined based on the previous month’s results and expected climate temperatures. We programmed the freezer to decrease 1 °C (1.8 °F) every 40 min for temperatures between 0 and –5 °C (32 and 23 °F) and 4 °C per h (7.2 °F/h) for lower temperatures. Samples were held at each target temperature for 10 min. To ensure a gradual thaw, all units were kept on ice after freezing from 12 to 22 h, followed by 1 h each at 4 °C (39.2 °F) and at room temperature, respectively.

Freeze–Thaw Injury Methods

Freeze–thaw injury was measured by visual discoloration, electrolyte leakage, and TTC reduction. The discoloration method was used only in 2018–2019; TTC reduction and electrolyte leakage were used in the last 2 overwintering cycles.

Units analyzed with discoloration were kept damp for 2 wk after thawing, cut longitudinally, and then visually analyzed on a scale from 1 to 4. Light green samples, indicating little damage, were rated 1; darker green or yellow samples, suggesting the most damage, were rated 4 (Zimmerman and others 2005).

Electrolyte Leakage Method

With 7 ml (0.24 oz) of deionized water added, thawed units to be measured for electrolyte leakage were vacuum-infiltrated for 3 min and shaken on a platform shaker (Innova 2300; New Brunswick Scientific Company, Edison, New Jersey) for 3 h at 250 rpm. We measured initial and final electrical conductivity with a conductivity meter (model 3100; YSI, Yellow Springs, Ohio) before and after autoclaving, respectively. Lethal temperature was defined as the temperature at which 50% injury occurred (LT₅₀). Gompertz function was used to determine LT₅₀, which was fitted using SigmaPlot 10.0 (Systat Software, San Jose, California). We compared statistical differences by using the Jackknife method and the ANOVA test in RStudio (bootstrap package, v. 4.0.3, R Development Core Team, Vienna, Austria; Lim and others 1998). The injuries to samples of Amur corktree, silver maple, tree of heaven, staghorn sumac, common hoptree, and common pricklyash, collected on 19 January 2021 and subjected to 0, –28, –40 °C (32, –18, and –40 °F) in the programmable freezer and to –80 °C (–112 °F) in an ultra-low freezer, were compared with ANOVA and Tukey’s honestly significant difference (HSD) test (JMP Pro 15 software; version 15, SAS Institute, Cary, North Carolina).

New growth at the end of a stem of common hoptree (Ptelea trifoliata).

TTC Reduction Method

After an additional 3 h of thawing, we sliced units in half longitudinally and placed the cut-side downward in 0.05-M phosphate buffer (pH 7.4) with 0.7% (for the 2 hoptree species) or 0.9% (for common pricklyash) TTC for 18 to 22 h in darkness. We determined in preliminary experiments that the 2 concentrations of TTC were the most effective for the respective species. All stem segments were photographed with consistent artificial light. Images were processed with TurfAnalyzer (http://turfanalyzer.com) to determine red stain cover (hue, 0–25; saturation, 30–100; brightness, 0–100) (Figure 1). Percentage of red cover was calculated as the proportion of red pixels to total pixels in each image (Karcher and others 2017). Percentage of red covers was averaged across replicates for each corresponding temperature and fitted with a Gompertz sigmoidal curve (Lim and others 1998) using SigmaPlot 10.0 (Systat Software, San Jose, California). We defined the LT₅₀ as the temperature on the Gompertz curve at which the percentage red cover was halfway between the cover of the no-injury and dead controls (at 0 °C and –80 °C [32 and –112 °F], respectively) within experiments.

Comparison of Freeze–Thaw Injury Methods

Tissue of common and western hoptree showed only subtle discoloration, which rendered the discoloration method ineffective. In contrast, electrolyte leakage (Figure 2, panel A) and TTC reduction (Figure 2, panel B) indicated increasing damage as temperatures were lowered in a way that fitted a Gompertz sigmoidal curve; however, in December and January the red-staining formazan used in the TTC reduction method appeared very similar (light pink) and produced similar percentages of red cover for common hoptree at all sampled temperatures except –80° (–112 °F) (Figure 2, panel C). Consequently, a Gompertz sigmoidal curve could not be fitted, and LT₅₀ values could not be calculated for those 3 mo (Figure 2, panel C). December and January data therefore are not included in Figure 3. In May, for common hoptree, only samples treated with the highest 2 temperatures stained at all, and LT₅₀ could not be calculated because the Gompertz curve could not be fitted to red cover percentages (Figure 3).

Figure 1

Stem segments of common hoptree treated with a 0.5-M phosphate buffer with 0.7% triphenyl tetrazolium chloride (TTC) and subjected to complete freeze stress, –80 °C (–112 °F) (A) and no freeze stress, 0 °C (32 °F) (B). In panel A, the left stem segment shows no red stain, indicating no reduction of TTC. The stem to its right shows the same segment when analyzed by TurfAnalyzer (A). In panel B, the left stem segment shows tissues stained dark red, indicating reduction of TTC to formazan by living tissue. To its right, the neon-green color depicts the red pixelized area when analyzed by TurfAnalyzer.

Despite these omissions, LT₅₀ values obtained from TTC reduction assays presented for common hoptree in Figure 3 follow the pattern shown in the more complete LT₅₀ values obtained from electrolyte leakage assays, which grew steadily colder until late January followed by a deacclimation first noted in early March. TTC reduction was not fully effective for common pricklyash either (Figure 4). In August, September, and April, the samples gradually decreased in stain as freezing stress increased (Figure 5, panel A), while the samples stained strongly at low temperatures from November to March, even showing red staining in samples subjected to –80 °C (–112 °F) in December, January, and February (Figure 5, panel B). With such similar red cover percentages for each sample temperature, it was impossible to fit a Gompertz curve and calculate an LT₅₀ for any months except August, September, and April (Figure 4). These 3 LT₅₀ values obtained from the TTC reduction assay do align with the more complete pattern of LT₅₀ values obtained from the electrolyte leakage assays. Notably, LT₅₀ values obtained from TTC reduction assays were consistently colder than those derived from the electrolyte leakage method for both common hoptree (Figure 3) and common pricklyash (Figure 4).

Both TTC reduction and electrolyte leakage methods can be used to document cold acclimation and deacclimation of western hoptree, but TTC reduction became unreliable in December and January for common hoptree, because most samples stained lightly and similarly (Figure 2, panel C). It may be that living tissue that is deeply dormant does not stain properly. Hay (1962) found that dormant tissues absorbed but were unable to reduce TTC after 24 h, but non-dormant tissues stained readily. By contrast, in May, only the samples subjected to the 2 warmest temperatures stained at all, suggesting low freezing tolerance. In this case, a Gompertz curve could not be fitted with samples at only 2 temperatures containing any red cover percentage, nor could an LT₅₀ be calculated. Visual discoloration, the third method used to assess freeze injury, cannot be recommended for common hoptree, because differences in injury resulted in only subtle changes in color.

Figure 2

Example of electrolyte leakage results for common hoptree on 13 April 2020. A Gompertz sigmoidal curve is fitted to the percentage injury at various temperatures. Lethal temperature (LT₅₀) is marked at 50% injury (A). Example of reduction of 2,3,5-triphenyltetrazolium chloride results for western hoptree on 27 January 2020. A Gompertz sigmoidal curve is fitted to the percentage of red cover detected by TurfAnalzyer at various temperatures. Lethal temperature (LT₅₀) is marked at the point where the red cover percentage is halfway between the red cover percentage of the 2 control samples, subjected to either 0 °C (32 °F) (no injury) or –80 °C (–112 °F) (freeze-killed) (B).Example of reduction of 2,3,5-triphenyltetrazolium chloride results for common hoptree on 17 December 2019. The sigmoidal curve could not be fitted because no trend was observed between percentage of red cover detected by TurfAnalzyer and temperature. Because it was impossible to fit a Gompertz sigmoidal curve to the results, no lethal temperature (LT₅₀) could be calculated (C).

Figure 3

Mean lethal temperature (LT₅₀) of western hoptree and 2 populations of common hoptree during autumn, winter, and spring of 2019–2020, determined by using 2 methods: electrolyte leakage and reduction of 2,3,5-triphenyltetrazolium chloride (TTC). Samples of common hoptree collected on 2 September, 3 October, 7 November, 17 December, 23 January, 4 March, 13 April, and 18 May. Samples of western hoptree collected on 4 November, 27 January, and 8 March. Error bars from jackknife data are too small to be seen. Percentage of red cover was detected by TurfAnalzyer, though results from December, January, and May showed no pattern, and no LT₅₀ could be calculated.

Figure 4

Lethal temperature (LT₅₀) of common hoptree and common pricklyash in central Iowa throughout the autumn, winter, and spring of 2020–2021, determined by using 2 methods, electrolyte leakage and reduction of 2,3,5-triphenyltetrazolium chloride (TTC). Error bars from jackknife data are too small to be seen. Percentage of red cover was detected by TurfAnalzyer, though results from November to March showed no pattern, and no LT₅₀ could be calculated.

TTC reduction was not as effective as electrolyte leakage for reliably calculating LT₅₀ values for common pricklyash. Even at low temperatures, samples stained strongly through most of the winter, and samples subjected to –80 °C showed red staining in December, January, and February (Figure 5, panel B). This suggests strong cold hardiness in common pricklyash, but lack of a clear change in red cover percentage across sampled temperatures prevented Gompertz curves from being fitted and LT₅₀ values from being calculated in all months except August, September, and April (Figure 4).

Quantifying the production of red-staining formazan by measuring percentage of red pixel coverage with TurfAnalyzer allowed for a more objective interpretation of TTC reduction measurements. While past researchers have limited their classification of damage to a rating scale or to noting alive or dead (Takeda and others 1993; Pellett and Heleba 1998; Nesbitt and others 2002), TurfAnalyzer can provide a percentage of damage. Use of TurfAnalyzer may also increase the effectiveness of this method for species that result in staining difficult to classify with the naked eye (Nesbitt and others 2002).

Electrolyte leakage results generated an LT₅₀ consistently warmer than the LT₅₀ calculated from results of TTC reduction (Figure 3). TTC reduction measures mitochondrial respiration, while ion leakage measures membrane integrity. Because a lack of or inefficient TTC reduction reflects damage to mitochondrial respiration whereas enhanced electrolyte leakage indicates damage to the plasma membrane, it is not surprising that LT₅₀ values calculated from these methods would differ. Other studies have found that mitochondria and plasma membrane are differentially sensitive to an identical freeze–thaw stress (Steffen and others 1989; Vyse and others 2020). Vyse and others (2020) observed a colder LT₅₀ for respiration than they did for electrolyte leakage in leaves of mouseear cress (Arabidopsis thaliana (L.) Heynh. [Brassicaceae]). In a species of plants from the nightshade family, Solanum acaule Bitter (Solanaceae), Steffen and others (1989) found that leaf respiration began to decline at –4 and –4.5 °C (24.8 and 23.9 °F), when 50% of ion leakage had already occurred. Our data indicating a relatively greater tissue injury as assessed by electrolyte leakage than that by TTC reduction assays are consistent with these prior reports.

Freeze Tolerance Comparison of Ptelea and Zanthoxylum

Western hoptree had begun deacclimating by 27 January 2020, and by 8 March 2020 it had a warmer LT₅₀ than common hoptree had until April (Figure 3). This result corresponds with air temperatures where plants were sampled (Figure 6). Common pricklyash was near its maximum cold hardiness in November, while common hoptree did not reach maximum cold hardiness until January (Figure 4). Common hoptree maintained a colder LT₅₀ in January and March than did common pricklyash (Figure 4). February LT₅₀ values for common hoptree and common pricklyash are suspect because they are calculated based on the assumption that samples subjected to –80 °C (–112 °F) demonstrate 100% damage; however, considering the extremely cold temperatures recorded in the 2 wk immediately before the February freezing (Figure 6) and the corresponding red-staining TTC samples of common pricklyash subjected to 80 °C (–112 °F) (Figure 5, panel B), this assumption does not seem sound (Figure 4).

Figure 5

Samples of common pricklyash treated with 2,3,5-triphenyltetrazolium chloride (TTC) to show freezing injury at various temperatures in September, when percentage red cover decreased as temperature decreased (A), and in February, when percentage red cover remained similar at all temperatures (B).

Decreasing photoperiods initiate cold acclimation, which is then increased by exposure to colder temperatures (Arora and Taulavuori 2016). By contrast, warm temperatures are the most important factor causing deacclimation (Arora and Taulavuori 2016). Bud dormancy can also affect deacclimation, because endodormant buds resist deacclimation until a chilling requirement has been met, which transitions buds to an ecodormant state that then responds to spring warming by undergoing deacclimation (Kalberer and others 2006; Arora and Taulavuori 2016). Slow deacclimation can be a key factor in overall cold hardiness, because of the possibility of unseasonably early warm temperatures or unusually late frosts (Arora and Taulavuori 2016).

The Mediterranean climate where western hoptree grows requires limited cold hardiness, and the species had begun deacclimating by 27 January 2020 (Figure 3). By contrast, common hoptree must withstand consistently lower temperatures throughout the winter (Figure 6), which was reflected in its greater cold acclimation (Figure 3). In addition to differences in climate between California and Iowa, genetic differences may also play a part in how cold hardiness differs between these species (Taulavuori and others 2004). Common pricklyash acclimated more quickly than did common hoptree, but common hoptree achieved a lower LT₅₀ in January and March than did common pricklyash. Its maximal cold hardiness and more gradual deacclimation indicate that common hoptree is the more cold-hardy of the 2 species; Kalberer and others (2007) also noted that several species of Rhododendron L. (Ericaceae) with higher midwinter hardiness were faster deacclimators than those with relatively lower midwinter hardiness. Note, however, that we did not sample throughout the native ranges of common pricklyash, which extends to northern Minnesota and eastern North Dakota, and common hoptree, which extends across the Upper Midwest. Additional research could sample extensively to assess the full extent of genetic potential for cold acclimation and deacclimation kinetics in the 2 species. Research shows that the timing and extent of cold acclimation as well as growth cessation (dormancy) is influenced by latitudinal differences in daylength and temperature, and the induction and release of dormancy modulate cold acclimation and deacclimation (Kalberer and others 2006; Arora and Taulavuori 2016).

While common pricklyash and common hoptree are distributed primarily in areas as cold as Zones 3 and 5, respectively, factors other than winter temperature may have prevented the northward expansion of common hoptree (USDA 2012; Kartesz 2015). Observations of common hoptree near Minneapolis (Dirr and Warren 2019) substantiate that the species is capable of surviving in areas with harsher winters than those of Zone 5. Although the maximal hardiness of common hoptree might suggest that it could be planted in Zone 2b, we encourage caution in extrapolating our results to planting recommendations (Figure 4) (USDA 2012). We subjected our samples to target temperatures for 10 min, whereas trees in nature could be exposed to cold temperatures for much longer periods of time. Plants could also occasionally experience colder temperatures than the hardiness zones suggest, which record the average coldest temperatures for each year over the past 30 y (USDA 2012). Our study also does not account for differences in cold hardiness among plants of varying ages. Evidence suggests that cold hardiness increases with age in seedlings of Sakhalin corktree (Phellodendron sachalinense (F. Schmidt) Sarg. [Rutaceae]) and species of Rhododendron (McNamara and Pellett 2000; Lim and others 2014).

Figure 6

Daily maximum and minimum temperatures in Muscatine, Iowa, from August 2019 to May 2020 (A). Daily maximum and minimum temperatures in Ames, Iowa, from August 2019 to May 2020 (B). Daily maximum and minimum temperatures in Antioch, California, from October 2019 to March 2020. Please note the difference in the y-axis scale compared with the other 3 plots (C). Daily maximum and minimum temperatures in Ames, Iowa, from August 2020 to April 2021 (D).

Freeze Tolerance Comparison of Sexes

While female and male plants of common hoptree had different LT₅₀ values from month to month, no consistent sex-based pattern emerged over time. Male plants had LT₅₀ values ranging from –3.0 to –29.7 °C (26.6 to –21.5 °F), and female plants had LT₅₀ values of –3.0 to –35.4 –C (26.6 to –31.7 °F). Although freezing temperatures have resulted in physiological responses that differ among females and males of some plant species, our results do not indicate a consistent difference in common hoptree. Others measured sex-based differences by comparing electrolyte leakage in leaves over 24 h or 2 wk of exposure to cold temperatures (Li and others 2005; Gupta and others 2012). Our comparison of acclimation and deacclimation curves throughout the winter may have been insufficient to capture subtle differences between sexes that could be observed on a shorter time scale in a different tissue.

Freeze Tolerance Comparison of Species in the Order Sapindales

Percentage injury rather than LT₅₀ was determined because only 4 freezing stress temperatures were used because of limited quantities of samples. On 19 January 2021, we observed no difference among 6 species—common hoptree, common pricklyash, Amur corktree, silver maple, tree of heaven, and staghorn sumac—when comparing mean injury at –40 °C (–40 °F). Replicates varied considerably. The mean injury ± standard error at –40 °C (–40 °F) for tree of heaven, staghorn sumac, silver maple, common pricklyash, Amur corktree, and common hoptree was 87.3% ± 31.2%, 72.4% ± 11.9%, 71.3% ± 8.7%, 56.7% ± 12.6%, 52.3% ± 21.3%, and 46.8% ± 8.7%, respectively.