Evidence of stress imprinting with population‐level differences in two moss species

Abstract Plants are often repeatedly exposed to stresses during their lives and have a mechanism called stress imprinting that provides “memories” of stresses they experience and increases their ability to cope with later stresses. To test hypotheses that primed bryophytes can preserve their stress imprinting after 6 days of recovery and induce higher levels of osmolytes and ROS‐scavenging activities upon later stress exposure, and there exist population‐level differentiation in their desiccation defenses, we transplanted samples of two populations of each of two moss species, Hypnum plumaeforme and Pogonatum cirratum, in a nature reserve in southern China. After 16 months of acclimation, sets of each population were subjected to control, one‐time desiccation stress, duplicated desiccation stress and cross‐stress (low temperature stress followed by desiccation stress) treatments. Levels of oxidant enzymes, osmolytes, and phytohormones in the samples were then determined. The desiccation stress generally led to increases in activities or contents of superoxide dismutase, guaiacol peroxidase, catalase, proline, soluble sugars, soluble proteins, and stress hormones including abscisic acid (ABA), jasmonates (JA), and salicylic acid (SA), with differences between both species and populations. After a 6‐day recovery period, contents of phytohormones (including ABA, JA, SA, and cytokinins) in stressed H. plumaeforme had substantially fallen toward control levels. The duplicated and cross‐stress treatments generally led to further accumulation of proline, soluble sugars, and soluble proteins, with further increases in activities of antioxidant enzymes in some cases. Furthermore, significant differences between allochthonous and native populations were found in contents of malondialdehyde and osmolytes, as well as antioxidant enzyme activities. Our results confirm the hypotheses and highlight the importance of osmolytes in mosses' stress responses.

For bryophytes, plenty of studies has demonstrated that experienced desiccation can make them adapt to desiccation stress better (Beckett, 1999;Beckett, Marschall, & Laufer, 2005;Bopp & Werner, 1993;Dilks & Proctor, 1976;Schonbeck & Bewley, 1981a, 1981bWerner, Ros Espín, Bopp, & Atzorn, 1991); however, this stress imprinting tends to wear off with the extension of the recovery time, and slower desiccation rates produce longer imprinting times (Brinda, Stark, Clark, & Greenwood, 2016). For naturally growing mosses in the field, water between dense individuals generally leads to relatively slow desiccation rates and the intervals between dry events varied. It has been reported that a period of more than 7 days is required to remove the acclimation experience (Hájek & Vicherová, 2014;Hellwege, Dietz, Volk, & Hartung, 1994;Stark, Greenwood, Brinda, & Oliver, 2014), and studies have also demonstrated that levels of osmolytes and antioxidants may increase in mosses upon exposure to low temperature or water stress, and may remain relatively high even after alleviation of stresses for 10 days (Liu et al., 2016;. So, we hypothesized that the priming information can be stored in mosses for a certain time (e.g., 6-10 days) and induce higher tolerance to a later stress. A complicating factor is that populations within a species' distribution are often exposed to different environmental conditions and thus develop local adaptations (Briggs, 1972;Chambers & Emery, 2016;Lázaro-Nogal et al., 2016). As both selective environmental pressures and gene flow limitations drive differentiation, between-population distances and dispersal potentials both affect the extent of population-level differentiation (Chambers & Emery, 2016;Korpelainen, Pohjamo, & Laaka-Lindberg, 2005). Moreover, due to the limited dispersal potential of bryophytes, significant population differentiation can be found within quite a small area (Briggs, 1972;Snäll, Ribeiro, & Rydin, 2003). Although population differentiation of bryophytes has received some attention, studies have generally focused on their morphological and life-history traits (Hedderson & Longton, 2008) or direct determination of their genetic structure (Pohjamo, Korpelainen, & Kalinauskaitė, 2008;Wang, Zhu, & Wang, 2012).
Between-population differences in their physiological traits have received much less attention despite their importance in species' fitness. However, it is clearly important to characterize and distinguish such differentiation and stress imprinting mechanisms (if present) in attempts to clarify bryophytes' suites of stress responses.

Hypnum plumaeforme is a widespread species in China, and
Pogonatum cirratum is also widely distributed but mainly in tropical and subtropical China. The two mentioned species (Figure 1) were selected as test plants because of their wide distribution and good survival after transplanting, which make them suitable for population-level differentiation study. In the study presented here, samples of two populations of each species at sites with substantially different environments were collected and transplanted in a subtropical nature reserve in southern China. After 16 months of acclimation, they were subjected to several stress regimes. We then determined and compared their antioxidant enzyme activities and levels of both osmolytes and phytohormones. The following hypotheses were tested as follows: Mosses have stress imprinting mechanisms; antioxidant enzymes, osmolytes, and phytohormones are involved in the mechanisms; and there are significant between-species and population-level differences in the mosses' stress tolerance. We also expected to find greater differentiation between the two H. plumaeformae populations than between the two P. cirratum populations, because the latter are geographically closer to each other.  Liu, Liu, Lei, Zhang, & Guo, 2011). Gametophytes of each population were planted in 12 trays (30 × 50 cm) assigned (in triplicate) to four treatment regimes: control (CK), one-time desiccation stress (OD), duplicated desiccation stress (DD), and cross-stress (low temperature followed by desiccation stress, TD). The planting density was similar to that of their natural communities. The treatments (each including a 6-day recovery period between the C, D, or T phase) are graphically illustrated and explained in Figure 3. The mosses were acclimated in the field until February 2017, during which they were watered every 3 days.

| Plants and treatment design
According to the records of a weather station located in the nature reserve, the mean annual temperature in 2016 is 19.5°C, with an average temperature of the hottest month (August) and coldest month (February) of 26.3°C and 10.3°C, respectively. The annual rainfall in the study area is 2096 mm, with the highest monthly rainfall in August (355 mm) and lowest in December (2.3 mm). In February 2017, trays assigned to the CK, OD, and DD treatments were moved to a wire cage near the valley and watered according to the schedule shown in Figure 3. Six days after the start of these treatments, the TD trays were moved to a climate chamber with the temperature set to 1 and 3°C in 12-hr light (50 μmol m −2 s −1 ) and 12-hr dark periods, respectively, and the relative humidity (RH) set to 85%, and 3 days later, they were also moved to the wire cage and exposed to the same conditions as the OD samples ( Figure 3). During the treatment period (from February 27 to March 30), the mean temperature is 16.2°C and the total rainfall is 121.2 mm, with the RH in the wire cage generally higher than 80% as aperiodically measured using a hygrothermograph (TES 1361, Taiwan). After the first desiccation treatment, both species showed significant morphological difference with the control group ( Figure 1) and the water content in desiccated and controlled samples was 0.24 ± 0.03 g/g DW and 1.34 ± 0.24 g/g DW for P. cirratum, and 0.27 ± 0.04 g/g DW and 2.42 ± 0.32 g/g DW for H. plumaeforme.
Samples of the top 2 cm of gametophytes (ca. 2 g) representing each triplicated species, population, and treatment combination were collected for determination of phytohormones and F I G U R E 1 Photographs of (a) controlled Hypnum plumaeforme, (b) desiccated Hypnum plumaeforme, (c) controlled Pogonatum cirratum, and (d) desiccated Pogonatum cirratum collected from Heishiding Nature Reserve. Water was supplied every 2 days for the controlled samples and was withheld for 12 days for the desiccated samples F I G U R E 2 Geographic locations of the sampled populations other physiological indices immediately after the CK, OD, DD, and TD treatments. Samples of H. plumaeforme assigned to DD and TD treatments were also collected to determine their phytohormone levels immediately after the 6-day recovery period. Samples were washed with distilled water for three times to remove debris and sand, and then excess water was blotted with paper sheets.
Samples for phytohormone analyses were stored at −80°C prior to determination. Other samples were stored at −20°C and analyzed within 48 hr.

| Proline contents
Proline was extracted using sulfosalicylic acid, as previously described (Wang & Huang, 2015). About 0.5 g of each sample was ground with 5 ml extraction solution (3% sulfosalicylic acid) and transferred to a test tube, then boiled for 10 min. After cooling to room temperature and centrifugation at 6,000 rpm for 10 min, the supernatant was used to determine proline contents following Troll and Lindsley (1955), with minor modification. Briefly, a 2 ml portion of each extract was mixed with 2 ml glacial acetic acid and 2 ml of 2% ninhydrin in ethanol-glacial acetic-phosphoric acid solution (25:60:15, by volume) then boiled for 40 min. After cooling, 3.5 ml of toluene was added, and the mixture was shaken vigorously and left to stand until it had delaminated. The toluene phase was collected and added to 0.5 ml methanol then analyzed spectrophotometrically at 520 nm. Toluene was used for blanks.

| Soluble sugar contents
The samples' soluble sugar contents were determined according to SIPP (Shanghai Institute of Plant Physiology Chinese Academy of Sciences (SIPP), 1999). Samples of about 0.3 g were weighed and ground in liquid nitrogen, then 20 ml of ice-cold 80% ethanol was added, and the samples were ground again for 3-5 min. Each mixture was transferred to a conical flask, extracted at 80°C for 30 min, and then (after cooling) subjected to negative-pressure filtration. The filtrate was diluted to 50 ml and transferred to a new conical flask, 0.5 ml of saturated lead subacetate solution was added, and the mixture was shaken vigorously. Then, 0.2 g sodium oxalate was added, and the mixture was refiltered. A portion (0.5 ml) of the resulting filtrate was added to 0.5 ml 80% ethanol and 5 ml anthrone reagent (containing 0.2% anthrone and 76% H 2 SO 4 ), boiled for 2 min, and after cooling analyzed spectrophotometrically at 630 nm. Sucrose was used as a calibration standard.

| Phytohormone contents
ABA, JA, SA, cis-zeatin (cis-Z), cis-zeatin riboside (cis-ZR), transzeatin (trans-Z), and IAA contents were extracted, purified, F I G U R E 3 Overview of the treatment regimes. In D treatment periods, no water was supplied to the mosses, but during the other treatment periods, water was supplied every 2 days. The mosses were exposed to ambient temperatures (7-26°C) during C, D, and recovery periods. During T treatment periods, the mosses were placed in a climate chamber providing 12-hr light/12-hr dark cycles, with temperatures of 3 and 1°C during the light and dark phases, respectively, and 50 μmol m −2 s −1 illumination during the light phases

| Data analysis
Means and standard errors of each measured variable were calculated. Differences between treatments within each population were analyzed using one-way ANOVA, and Tukey's test (or Dunnett's test when homogeneity of variance requirements was not met) was used to identify differences significant at the 0.05 probability level. Log-transformation was applied when necessary. In addition, between-treatment and between-population differences were explored by two-way ANOVA. Differences in variables are regarded as significant if p < 0.05, and only mentioned if they meet this criterion.
Increases or decreases in variables refer to measured differences between stressed samples and corresponding controls after the treatments unless otherwise stated.

| Oxidative stress
The Differences in SOD activities between the two P. cirratum populations were not significant, but CAT and POX activities were significantly higher in P-Jinggangshan samples than in P-Heishiding samples ( Figure 4, Table 2).

| Osmolytes
The stress treatments caused significant accumulation of proline, soluble sugars, and soluble proteins in both mosses. In H. plumaeforme, DD and TD treatments caused higher accumulation of proline than OD, while in P. cirratum, contents of soluble sugars and soluble proteins were higher in DD and TD samples (except for soluble proteins in TD samples of the P-Heishiding population) than in OD samples ( Figure 5). Results of two-way ANOVA showed that contents of proline, soluble sugars, and soluble proteins were generally higher in allochthonous populations than in native populations (except for soluble sugars in P. cirratum) ( Table 1).

| Phytohormones
The stresses caused sharp increases in ABA contents in H. plumaeforme: up to 11.4-and 16.7-fold in H-Heishiding and H-Lushan samples, respectively ( Figure 6). The DD and TD treatments F I G U R E 4 Malondialdehyde (MDA) contents and activities of superoxide dismutase (SOD), catalase (CAT), and guaiacol peroxidase (POX) in Hypnum plumaeforme and Pogonatum cirratum gametophytes after indicated stress treatments (mean + 1 SE, n = 3). CK, OD, DD, and TD refer to control, one-time desiccation stress, duplicated desiccation stress, and low temperature followed by desiccation stress, respectively. Different lowercase letters above columns denote significant difference between treatments within each population (one-way ANOVA with Tukey's post hoc test). H-Heishiding and H-Lushan refer to Heishiding and Lushan populations of H. plumaeforme, respectively, while P-Heishiding and P-Jinggangshan refer to Heishiding and Jinggangshan populations of P. cirratum, respectively. Note the difference in scales for POX activities in the two moss species    Figures 6 and 7).

The stress treatments caused increases in contents of cis-Z, cis-
ZR, and trans-Z in H. plumaeforme. After the 6-day recovery period, the cis-ZR and trans-Z contents of DR and TR samples were lower than in OD, DD, and TD samples, but cis-Z levels were only significantly lower in H-Heishiding samples, in which levels of cis-Z and trans-Z were higher in DR samples than in TR samples ( Figure 6).
TA B L E 2 Effects of population (P-Heishiding and P-Jinggangshan) and treatment (control, one-time desiccation stress, duplicated desiccation stress, and cross-stress) on oxidative and osmotic indices in Pogonatum cirratum as shown by   In previous studies, we found that after a 10-day recovery period stressed mosses still maintained relatively high levels of SOD and osmolytes (Liu et al., 2016;. In the present study, we found that proline, soluble sugar, and soluble protein contents were higher after duplicated and cross-stress treatments than after one-time stress treatments. These results indicate that the first stress primed accumulation of osmolytes in the mosses when subjected to the second stress, and improved their adaptation to the later stress. Thus, they confirm two of our hypotheses: that mosses have stress imprinting mechanisms, and osmolytes are involved in them. These results also indicated that the stress imprinting can be maintained in the tested mosses for 6 days. Further increases in SOD and CAT activities in H-Lushan, SOD in P-Heishiding, and POX F I G U R E 6 ABA, JA, SA, IAA, cis-Z, cis-ZR, and trans-Z contents in gametophytes of Hypnum plumaeforme subjected to indicated stress treatments (mean + 1 SE, n = 3). CK, OD, DD, and TD refer to control, one-time desiccation stress, duplicated desiccation stress, and low temperature followed by desiccation stress. DR and TR indicate contents measured after 6 days of recovery from the first desiccation or low temperature treatment. Different lowercase letters above columns denote significant difference between treatments within each population (one-way ANOVA with Tukey's post hoc test). Data were logtransformed before further analysis in P-Jinggangshan samples subjected to duplicated or cross-stresses indicated that the first stress also induced stronger increases in ROS-scavenging activities in the mosses during exposure to the later stress (Ye & Gressel, 2000). On the other hand, POX activities were lower in TD samples of the Heishiding populations of both species, but not the allochthonous populations, than in corresponding OD and DD samples, highlighting the differences in low temperature responses between populations from different latitudes.

Moreover, even after a year of acclimation in Heishiding Nature
Reserve, levels and activities of MDA, antioxidant enzymes, and F I G U R E 7 ABA, JA, SA, IAA, cis-Z, cis-ZR, and trans-Z contents in gametophytes of Pogonatum cirratum subjected to indicated stress treatments (mean + 1 SE, n = 3). CK, OD, DD, and TD refer to control, one-time desiccation stress, duplicated desiccation stress, and low temperature followed by desiccation stress. Different lowercase letters above columns denote significant differences between treatments within each population (one-way ANOVA with Tukey's post hoc test, in all cases except JA in P-Jinggangshan, for which Dunnett's test was applied). Data were log-transformed before further analysis

Treatment
F I G U R E 8 Mean (+1 SE, n = 3) IAA contents in gametophytes of Hypnum plumaeforme and Pogonatum cirratum subjected to indicated stress treatments. CK, OD, DD, and TD refer to control, one-time desiccation stress, duplicated desiccation stress, and low temperature followed by desiccation stress. DR and TR indicate contents measured after 6 days of recovery from the first desiccation or low temperature treatment. Different lowercase letters above columns denote significant difference between treatments within each population (one-way ANOVA with Tukey's post hoc test). Data were log-transformed before further analysis. H-Heishiding and H-Lushan refer to Heishiding and Lushan populations of H. plumaeforme, respectively, while P-Heishiding and P-Jinggangshan refer to Heishiding and Jinggangshan populations of P. cirratum, respectively population-level differentiation is present in both species, but more strongly in H. plumaeforme than in P. cirratum. This is in accordance with our third hypothesis and suggests that geographical distance might be one of the drivers of population-level differentiation (Chambers & Emery, 2016;Korpelainen et al., 2005), though further studies are needed to illustrate the role of geological distance in their population differentiation.
We found no previous studies on responses of JA and SA to stress in mosses (literature screening procedures not described). However, our study demonstrates that stresses induced increases in JA and SA levels in both populations of H. plumaeforme, as observed in vascular plants, but only in P-Jinggangshan samples of P. cirratum.
Cytokinins are a large group of phytohormones, mostly N 6 -substituted adenine derivatives, which are strongly associated with growth and development, and generally considered to be negative regulators of abiotic stress signaling (Nishiyama et al., 2011;Verslues, 2016 After the 6-day recovery period, the elevated JA, SA, cis-Z, cis-ZR, and trans-Z contents in stressed H. plumaeforme samples had decreased significantly toward control levels. Such reductions in levels of stress response substances after recovery, and enhancement of defense mechanisms only when stress recurs, avoid costs of constitutive activation of stress responses (Brinda et al., 2016;Bruce et al., 2007;Cruz de Carvalho et al., 2014). However, their ABA levels were still higher than in control samples after recovery, which might promote faster and better responses to later stresses as most stress responses are ABA-mediated (Bopp & Werner, 1993;Hellwege et al., 1994;Kohli et al., 2013).
Upon exposure to duplicated or cross-stresses, we detected further accumulations of trans-Z in samples of both H. plumaeforme populations, and of cis-Z in H-Heishiding samples, which might be associated with stress imprinting. Moreover, we detected substantial between-hormone, between-species, and between-population variations in responses, reflecting the sensitivity and complexity of phytohormone regulation in stress responses. Due to the complex crosstalk among phytohormones (Kohli et al., 2013;Verma et al., 2016) as well as the involvement of epigenetic changes in stress responses, more detailed and specific studies are needed to elucidate the underlying mechanisms of hormone regulation in the mosses' stress imprinting.

| CON CLUS ION
Desiccation stress induced increases in antioxidant enzyme activities and contents of osmolytes and ABA in the mosses. Though with substantial between-species and population-level differences, the first desiccation or low temperature stress led to further increases in levels or activities of osmolytes and antioxidant enzymes upon exposure to the later desiccation stress after 6 days of recovery, indicating that the stress imprinting in the mosses can be maintained for at least 6 days. Significantly higher levels or activities of MDA, SOD, CAT, POX, and osmolytes were detected in the allochthonous population of H. plumaeforme than in the native population even after one year of acclimation, indicating distinct population differentiation, but the differentiation between the populations of the P. cirratum was weaker, which might be related to their closer geographical distance.

ACK N OWLED G M ENTS
This research was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
WQ Liu was responsible for the design and implementation of the entire study. JQ Xu and WQ Liu carried out the study, including fieldwork, laboratory experiments, data analysis, and drafting the manuscript. W Fu, XY Wang, and YF Chen participated in the laboratory experiments, data analysis, and drafting the manuscript. CY Lei collected the weather data during the study period and participated in the fieldwork and laboratory experiments. WQ Liu, YF Chen, and CY Lei critically revise the manuscript. All authors gave the final approval of the version to be published.