Desiccation tolerance in bryophytes relates to elasticity but is independent of cell wall thickness and photosynthesis.

Mosses have been found outliers of the trade-off between photosynthesis and bulk elastic modulus described for vascular plants. Hence, potential trade-offs among physical features of cell walls and desiccation tolerance, water relations, and photosynthesis were assessed in bryophytes and other poikilohydric species. Long-term desiccation tolerance was quantified after variable periods of desiccation/rehydration cycles. Water relations were analyzed by pressure-volume curves. Mesophyll conductance was estimated using both CO2 curve-fitting and anatomical methods. Cell wall elasticity was the parameter that better correlated with the desiccation tolerance index for desiccation tolerant species and was antagonistic to higher absolute values of osmotic potential. Although high values of cell wall effective porosity were estimated compared with the values assumed for vascular plants, the desiccation tolerance index negatively correlated with the porosity in desiccation tolerant bryophytes. Neither cell wall thickness nor photosynthetic capacity were correlated with the desiccation tolerance index of the studied species. The existence of a potential evolutionary trade-off between cell wall thickness and desiccation tolerance is rejected. The photosynthetic capacity reported for bryophytes is independent of elasticity and desiccation tolerance. Furthermore, the role of cell wall thickness in limiting CO2 conductance would be overestimated under a scenario of high cell wall porosity for most bryophytes.

INTRODUCTION

Desiccation tolerance is defined as the capacity to recover normal function after rehydration from a desiccated state of a minimum water potential of −100 MPa, equivalent to the loss of approximately 90% of the intracellular water (Gaff, 1997; Oliver et al., 2020). This has been considered a “primitive” trait (Oliver et al., 2005), because of the high frequency of this feature in the non‐monophyletic descendants from the earliest branching events in the phylogeny of land plants, i.e. the bryophytes (Wood, 2007). It is generally accepted that desiccation tolerance was lost during early evolution of tracheophytes—remaining possibly in seeds, spores, and fern gametophytes (Ballesteros et al., 2020; López‐Pozo, Ballesteros, et al., 2019; Watkins et al., 2007)—and that it was reacquired separately during land plant evolution for all tracheophyte lineages, except for gymnosperms (Farrant & Moore, 2011).

Despite different desiccation tolerance origins, several strategies have been shown to be in common for both bryophyte and tracheophyte desiccation tolerant (DT) species, possibly with the involvement of a seed‐derived strategy for the latter (Farrant et al., 2020; Farrant & Moore, 2011; Oliver et al., 2010). An either constitutive or induced efficient antioxidant system or the expression of Late Embryogenesis Abundant (LEA) proteins in angiosperms and LEA‐like rehydrins in bryophytes are two examples of this (Oliver et al., 2005). The requirements of the cell wall for favoring desiccation tolerance have been also widely studied. Collapsed folded cell walls frequently occur when both “resurrection” plants (DT ferns and angiosperms) and bryophytes desiccate (Moore, Vicré‐Gibouin, et al., 2008; Pressel et al., 2010; Proctor, 2001; Proctor et al., 2007; Shivaraj et al., 2018). This phenomenon has been related with specificities in the chemical composition of the cell wall (Moore, Farrant, & Driouich, 2008; Webb & Arnott, 1982) and it has been linked to the capacity to maintain ultrastructural organization and to avoid mechanical stress on cells during desiccation/rehydration cycles. Consequently, the bulk modulus of elasticity (ε) of leaves/shoots seems to be associated to desiccation tolerance for both bryophytes (Proctor et al., 1998) and resurrection plants (Nadal, Perera‐Castro, et al., 2021). Cell wall thickness (T cw) also seems to play a role in desiccation tolerance, although the mechanisms remain unclear. Hence, Nadal, Brodribb, et al. (2021) recently reported a generalized tendency of resurrection plants to present thicker cell walls compared with desiccation sensitive (DS) species of similar photosynthetic capacity. An intraspecific increase in T cw has also been described as a response to desiccation stress in algae (Hoppert et al., 2004; Morison & Sheath, 1985). However, the relationship between ε or T cw with a quantitative index for the degree of long‐term desiccation tolerance of bryophytes has not been reported yet.

Both ε and T cw have been linked with the photosynthetic capacity. T cw is considered one of the main anatomical traits limiting CO2 diffusion and responsible for the variability of mesophyll conductance of land plants (Evans et al., 2009; Gago et al., 2020; Peguero‐Pina et al., 2017; Veromann‐Jürgenson et al., 2020). Regarding elasticity, Nadal et al. (2018) reported a negative trade‐off between net CO2 assimilation rates (A N) and ε for ferns and angiosperms, which included a few resurrection plants, so that species with a lower ε (higher elasticity) presented also high A N through a still unclear mechanistic/evolutionary trade‐off. In bryophytes, among other specific groups of plants, ε is not associated with their A N, and the capacity to mobilize large amounts of water during dehydration (i.e. higher capacitance at full turgor) is not translated in higher rates of gas exchange (Perera‐Castro, Nadal, & Flexas, 2020). In other words, if elasticity is hypothesized to be involved in the desiccation tolerance of bryophytes, it would not be a constraint for their photosynthetic capacity. On the contrary, very thick cell walls and low exposed chloroplast surface have been related to the low mesophyll conductance of bryophytes (Carriquí, Roig‐Oliver, et al., 2019). In bryophytes, the combination of the thickest cell walls among the land plant phylogeny and the lowest photosynthetic rates (see compiled data of Flexas et al., 2021) have led to the hypothesis that the reduction of T cw through the evolution of land plants is part of a potential trade‐off between photosynthetic capacity and desiccation tolerance (Carriquí et al., 2015; Gago et al., 2019; Hanson et al., 2014).

Beyond this potential trade‐off, the role of cell walls in limiting the photosynthetic capacity of bryophytes in still under debate. The low reported A N of bryophytes seems to be due to CO2 diffusion mainly limiting or co‐limiting the biochemistry of the photosynthetic capacity (Carriquí, Roig‐Oliver, et al., 2019; Perera‐Castro, Waterman, et al., 2020). However, the anatomical modeling applied so far to bryophytes—and to any measured plant—assumes low values of effective porosity of cell walls, either constant or arbitrarily and negatively varying with T cw (Evans et al., 2009; Terashima et al., 2006). Since the relationship between A N and T cw along the land plant's phylogeny is not linear, but asymptotic, with mosses presenting the thickest and lowest (but positive) A N (Flexas et al., 2021), every unit of cell wall seems not to be as restrictive for CO2 diffusion in bryophytes as it is in, e.g. angiosperms. This observation led us to suspect that a larger cell wall porosity could occur in bryophytes as compared with vascular plants. Not knowing the real porosity of the cell walls results in large uncertainties about the percent distribution of estimated anatomical limitations (Tomás et al., 2013, 2014; Tosens, Niinemets, Westoby, & Wright, 2012; Tosens et al., 2016; Lu et al., 2016; Han et al., 2018; Tosens & Laanisto, 2018; Carriquí, Roig‐Oliver, et al., 2019, Carriquí, Douthe, et al., 2019).

The aims of the present study were to describe the physical factors related to the desiccation tolerance of poikilohydric species of bryophytes—mosses, liverworts, and hornworts—lycophytes (Selaginella denticulata) and ferns (Hymenophyllum spp.), especially those regarding cell wall properties, i.e. T cw and ε, and to discuss their role as the potential drivers of an evolutionary trade‐off between photosynthetic capacity and desiccation tolerance. Water relations, gas exchange, and anatomical traits were measured, as well as the weighed recovery of photosynthetic function after short/long‐term desiccation/rehydration cycles as a quantitative index of desiccation tolerance.

MATERIALS AND METHODS

Plant materials and growing conditions

Sixteen species were analyzed in this study (Table 1): one hornwort, seven mosses, five liverworts, one lycophyte, and two filmy ferns (Hymenophyllum spp.). Species were collected in Mallorca (Balearic Islands), Tenerife (Canary Islands), La Rioja (Spain), Effeltrich (Germany), and Parque Katalapi (Puerto Montt, Chile). Specimens of some species (P. purum, T. tamariscinum, P. undulatum, L. cruciata, and S. denticulata) were grown in a moss greenhouse at the University of Balearic Island (UIB, Mallorca, Spain) for more than one year and they were part of the established UIB bryophytes collection before measurements. The rest of the species were measured immediately after collection within the next five days. Native substrate in perforate trays, daily watered with deionized water, was used for the long‐term culturing of bryophytes. For gas exchange measurements, anatomical analysis, pressure–volume curves, and desiccation assays, samples were cleaned, and brown tissues were removed. Green, healthy thalli/shoots were incubated overnight in Petri dishes covered with wet tissues before measurements to avoid possible negative effects of the cutting (Wang et al., 2021). The studied specimen corresponded to some of those listed in Perera‐Castro, Nadal, and Flexas (2020), for which net CO2 assimilation rates were at 400 μmol CO2 mol−1 air and some pressure–volume derived parameters were already reported.

TABLE 1

List of studied species, origin, and ecology/morphology/taxonomy information

Species	Phylum	Family	Morphology	Origin
Anthoceros agrestis (Paton) Damsholt	Anthocerotophyta	Anthocerotaceae	Thalloid	Tenerife
Fissidens serrulatus Brid.	Bryophyta	Fissidentaceae	Leafy	Tenerife
Plagiomnium undulatum (Hedwig) T. J. Koponen	Bryophyta	Mniaceae	Leafy	Mallorca a
Polytrichastrum formosum Hedw.	Bryophyta	Polytrichaceae	Leafy	Effeltrich
Polytrichum juniperinum Hedw.	Bryophyta	Polytrichaceae	Leafy	Tenerife
Pseudoscleropodium purum (Hedw.) M. Fleisch.	Bryophyta	Brachytheciaceae	Leafy	Mallorca a
Sphagnum palustre L. b	Bryophyta	Sphagnaceae	Leafy	La Rioja
Thuidium tamariscinum (Hedw.) Schimp.	Bryophyta	Thuidiaceae	Leafy	Mallorca a
Selaginella denticulata (L.) Spring	Lycophyta	Selaginellaceae	Leafy	Mallorca a
Durmortiera hirsuta (Sw.) Nees	Marchantiophyta	Marchantiaceae	Thalloid	Tenerife
Lunularia cruciata (L.) Dumort. ex Lindb.	Marchantiophyta	Lunulariaceae	Thalloid	Mallorca a
Pellia endiviifolia Dicks. Dum	Marchantiophyta	Pelliaceae	Thalloid	Tenerife
Porella canariensis (F.Weber) Underw.	Marchantiophyta	Porellaceae	Leafy	Tenerife
Saccogyna viticulosa (L.) Dumort.	Marchantiophyta	Geocalycaceae	Leafy	Tenerife
Hymenophyllum caudiculatum Mart.	Pteridophyta	Hymenophyllaceae	Leafy	Concepción (Chile)
Hymenophyllum dicranotrichum (C. Presl) Hook. ex Sadeb.	Pteridophyta	Hymenophyllaceae	Leafy	Concepción (Chile)

Grown for more than 1 year in a moss greenhouse at University of Balearic Islands daily watered with deionized water at ambient temperature.

Correspond to Sphagnum sp. of Perera‐Castro, Nadal, and Flexas (2020).

Long‐term desiccation tolerance assay

In order to test the desiccation tolerance of the studied species, the “Falcon Test” method was modified after López‐Pozo, Flexas, et al. (2019) as described in Perera‐Castro et al. (2021) for bryophytes. Briefly, it is based on the capacity of the specimens to recover maximum quantum yield of PSII (F v/F m) when rehydrated after short/long storage in dried conditions (relative water content at equilibrium <1%–5%) inside 50 ml Falcon tubes with 12 g of silica gel with a relative humidity lower than 10% in equilibrium, which corresponds to a water potential of −310 MPa (Gaff & Oliver, 2013). F v/F m was measured with the fluorometer IMAGING‐PAM (Heinz Walz GmbH) before and after the dehydration/rehydration cycle. The imaging of chlorophyll fluorescence allowed to average the F v/F m values of all pixels corresponding with plant tissue, using minimum fluorescence (F 0) image for selecting the area of interest.

The excess of interstitial water was removed from the full hydrated specimen by gently pressing them against a dry tissue. Hundred milligrams of fresh weight per sample and tube were used, resulting in a rate of drying of 62 mg g−1 dry weight per hour (López‐Pozo, Flexas, et al., 2019). A thin net avoided direct contact of plant material with the silica gel. After 24 h, seven days, 14 days, and 30 days, samples (n = 3–4 independent samples per each period and species) were rehydrated by covering with wet tissues and stored in petri dishes for 24 h. The relative recovered F v/F m at each period of time (rF v/F mi, where i is the total days of desiccation) was calculated as. where F v/F m0 is the F v/F m measured before desiccation. In order to weight up recovery after longer times, the desiccation tolerance index (DTI) was calculated as:

This index was developed by using the strongest desiccant tested by López‐Pozo, Flexas, et al. (2019), the silica gel, varying the length of dry state seeing the high percentage of bryophytes that fully recover F v/F m after 24 h of fast desiccation and rehydration. Thus, only species with values of DTI equal to 0 were considered fast desiccation sensitive (DS), while species with DTI > 0 were attributed qualitatively to fast desiccation tolerant species (DT).

Pressure–volume curves

Six pressure–volume curves per species were performed by slowly air‐drying full hydrated samples and alternately weighing and measuring its water potential by using a psychrometer (model WP4C, Decagon Device Inc.). The excel tool of Sack and Pasquet‐Kok (2011) was used for delimiting the turgor loss point and calculating pressure–volume derived parameters. Turgid weight (TW) of each specimen was estimated as the x‐intercept of leaf water potential (Ψleaf) versus fresh weight at any time on the curve (FW). Notice that the used method for calculating TW allows a reliable differentiation of their maximum internal water content (Perera‐Castro, Nadal, & Flexas, 2020), frequently problematic in bryophytes (e.g. Koster et al., 2010). Thus, relative water content (RWC) was calculated as RWC = 100 (FW − DW)/(TW − DW). DW is the dry constant weight and was obtained after keeping the samples at 70°C for 2–3 days. The point from which the Pressure–volume curve [−1/Ψleaf − (100 − RWC)] became linear was considered the turgor loss point. RWC and Ψleaf measured at that point was obtained (RWCtlp and π tlp, respectively). The saturating water content (SWC) was calculated as SWC = (TW − DW)/DW.

Osmotic potential at full rehydration (π o) was calculated as −1/y‐intercept of the linear regression of the pressure–volume curve under turgor loss point. The bulk modulus of elasticity (ε) was calculated as the slope of pressure potential versus total RWC. Absolute leaf capacitance at full turgor (C ft) was determined from the initial slope of the relationship between Ψleaf and RWC, and was normalized by leaf/shoot/thallus DW: C ft = (TW − DW) ΔRWC/(ΔΨleaf DW). Capacitance at turgor loss point (C tlp) was calculated as C ft but for the relationship between Ψleaf and RWC for values below turgor loss point. The extracellular apoplastic fraction (a f) was considered the fraction of water content when osmotic potential = −∞ and was calculated as x‐intercept of the pressure–volume curve under turgor loss point. For the same samples used in pressure–volume curves and, when possible, gas exchange, shoot/thalli mass area (SMA) was calculated as the ratio of the projected shoot/thalli area and its DW.

Gas exchange and chlorophyll fluorescence measurements

Gas exchange measurements were performed by using a GFS‐3000 system coupled with an IMAGING‐PAM fluorometer (Heinz Walz) and a custom‐made moss cuvette consisting of a gasket affixed to a piece of thin polyester stocking fabric (illustrated as supplementary information in Perera‐Castro, Nadal, & Flexas, 2020). CO2 response curves of net CO2 assimilation rate (A N) were performed for six replicates of each species, except for P. undulatum, Hymenophyllum spp., and S. palustre, where only instantaneous A N at 400 μmol CO2 mol−1 air was determined due to logistic problems related with the handling of the samples during measurements. Relative humidity was kept at 75%–85%, blue irradiance (I) at saturation level (100–1000 μmol m−2 s−1, depending on the species, tested previously with A N–I curves), and air temperature in the gas exchange cuvette at 25°C. The flow rate within the chamber was 750 μmol s−1. The concentration of CO2 surrounding the sample inside chamber (C a) was first set at 400 μmol mol−1, then for CO2‐response curves C a was decreased stepwise down to 50 μmol mol−1 and returned to its original value, followed by a stepwise increase up to 2000 μmol mol−1. After each step of C a, samples were removed from the chamber and placed again in the wet Petri dishes with deionized water for 2 min to avoid their desiccation during measurements, except in the case of S. denticulata, which was measured without the moss cuvette and with the base of the outside shoots and rhizoids under water. Most excessive external water was removed gently with a dry tissue before introducing the sample in the GFS chamber. In order to avoid the errors derived from CO2 leakage in the parameterization of photosynthesis (Flexas et al., 2007), measurements of apparent net CO2 assimilation at 200, 700, 1300, and 2000 μmol CO2 mol−1 air were done for an empty moss cuvette before introducing the specimen. A N and chlorophyll fluorescence (steady‐state and maximum light‐adapted fluorescence, F s and F′m, respectively) were recorded at steady‐state conditions, when diffusional limitations due to external water were considered null and biochemistry was fully light‐adapted (5–20 min). A N was normalized to the measured thallus/shoot area. Light‐adapted yield of PSII (фPSII) was calculated according to Genty et al. (1989): фPSII = (F′m − F s)/F′m. Electron transport rate was calculated from chlorophyll fluorescence (J FLU) according to Krall and Edwards (1992): J FLU = фPSII I αβ, where αβ is the product of absorbed quanta between PSI and PSII and was determined as 4/slope of the relationship between фPSII and фCO2 ([A N + R L]/I) obtained by varying irradiance under non‐photorespiratory conditions in an atmosphere containing <1% O2 (Valentini et al., 1995). Light curves under non‐photorespiratory conditions were measured after CO2 curves and at 400 μmol CO2 mol−1 air. Mitochondrial respiration under light condition (R L) was calculated from the initial light‐limited portion of the low‐O2‐light curves as the negative intercept of the relationship between A N and (фPSII I)/4 according to Yin et al. (2011). Mitochondrial respiration under dark conditions (R D) at 400 μmol CO2 mol−1 air were also measured after 10 min of darkness, except for Hymenophyllum spp.

Anatomical analysis

Twelve out of the nineteen studied species were selected for anatomical analysis: A. agrestis, F. serrulatus, L. cruciata, P. endiviifolia, P. undulatum, P. formosum, P. juniperinum, P. canariensis, P. purum, S. viticulosa, S. denticulata, and T. tamariscinum. Three pieces of leaves/thalli/phyllidia of the same individuals on which gas exchange was performed were taken per species and fixed under vacuum with 4% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer (pH 7.2). Then, fixed fragments were washed three times in 0.1 M sodium cacodylate buffer (pH 7.2) with 5% sucrose, and stored overnight at 4°C in the same buffer. Afterwards, samples were incubated at 30°C for 2 h in a mixture (1:1) of 2% osmium tetroxide and 8% potassium ferrocyanide, followed by dehydration in a graded ethanol series. The dehydrated fragments were embedded in LR White resin (medium grade, London Resin Company Ltd.) and cured in an oven at 60°C for 48 h. Semi‐thin (0.8 μm) and ultrathin (90 nm) transversal sections were obtained with an ultramicrotome and contrasted with 1% toluidine blue or Reynolds lead citrate solution (Reynolds, 1963), respectively. Photographs of the semi‐thin sections were taken at ×200 and ×500 magnification with a light microscope (Olympus) and were photographed with a Moticam 3 (Motic Electric Group Co.). Contrasted ultra‐thin sections were viewed at ×1500 and ×30 000 magnification using a transmission electron microscope (TEM H600, Hitachi). All images were analyzed using the software ImageJ (Wayne Rasband/NIH). Thalli/phyllidia thickness (T leaf), total length of cells and chloroplasts facing intercellular or external air spaces (L m and L c, respectively) and the volume fraction of intercellular air space (f ias, only for S. denticulata and L. cruciata, assuming for the latter that chambers of fibrils were air‐filled during gas exchange measurements) were measured from images obtained by light microscopy. Cell wall thickness (T cw), chloroplast thickness (T chl), and cytosol path length (L cyt, from cell wall to chloroplast) were measured from TEM photographs (5–8 measurements per photo, 2–3 photos of different fields per replicate). Surface area of mesophyll cells or chloroplasts exposed to intercellular airspaces per unit leaf area (S m and S c, respectively) were calculated as described in Evans et al. (1994), applied for bryophytes by Carriquí, Roig‐Oliver, et al. (2019): where W is the width of the measured section (which covered the complete thickness of the thalli/phyllidia) and γ is the curvature correction factors from Thain (1983). γ was calculated to be 1.31–1.36 for most of the studied species (1.43 for S. denticulata), except for P. purum, whose tubular cell morphology was considered to be cylindrical without curvature effect (γ = 1).

Mesophyll conductance calculations and limitations of photosynthesis

Mesophyll conductance (g m) was calculated using two different methods: by curve‐fitting method (g mCF) and by anatomical modeling (g mANAT).

Curve‐fitting method for calculating mesophyll conductance

The curve‐fitting method is based on the calculation of the maximum carboxylation capacity (V cmax), the maximum capacity for electron transport rate (J max), the velocity for triose phosphate utilization (TPU), and g m that best fit CO2 curve data to the model of Farquhar et al. (Farquhar et al., 1980; Van Caemmerer & Evans, 1991, modified by Ethier & Livingston, 2004, and Sharkey et al., 2007, to include g m). The excel tool of Sharkey et al. (Sharkey, 2016) was used for the fitting. The fitting was done for A N–C c curves, being C c the concentration of CO2 at the carboxylation site of Rubisco: C c = C a − A N/g m, in the case of bryophytes which lack stomata and intercellular airspaces, or C c = C i − A N/g m in the case of S. denticulata, where C i is the CO2 concentration in the sub‐stomatal cavity. In vitro Rubisco Michaelis–Menten constants for carboxylation (K c) and oxygenation (K o) and chloroplastic CO2 photocompensation point (ᴦ*) used for the curve fitting were obtained from averaged values for mosses, liverworts and lycophytes reported by Font and Galmés (2016) and Galmés et al. (2014; Table S1).

The relative mesophyll (l m), stomatal (l s), and biochemical (l b) limitations to photosynthesis were calculated according to Grassi and Magnani (2005) with the modifications of Carriquí, Roig‐Oliver, et al. (2019) for bryophytes: where g s is the stomatal conductance to CO2 and g tot is total conductance to CO2 between leaf surface and carboxylation sites (1/g tot = 1/g s + 1/g m). In bryophytes, g tot = g mCF.

Anatomical modeling of mesophyll conductance

Values of g mANAT were obtained by using the one‐dimensional gas diffusion model of Niinemets and Reichstein (2003a, 2003b) as modified by Tomas et al. (2013) and Carriquí, Roig‐Oliver, et al. (2019), where g m is considered to be composited by gas phase conductance (g ias, from substomatal cavities to outer surface of cell walls) and liquid phase conductance (g liq, from outer surface of cell walls to chloroplasts): where R is the gas constant (Pa m3 K−1 mol−1), T is the absolute temperature (K), and H is the Henry's lay constant (Pa m3 mol−1). For most bryophytes and leafy ferns, CO2 resistance of the internal gas phase was considered zero (1/g ias = 0), due to the lack of internal airspaces. The chambers of L. cruciata were assumed to be air‐filled during gas exchange measurement and sample fixation for anatomy analysis, and the concentration of CO2 (C a) was assumed to be equal inside and outside the pore. For L. cruciata and S. denticulata, g ias was calculated by average gas‐phase thickness (ΔL ias) and gas‐phase porosity (f ias): where ϛ is the diffusion path tortuosity (m m−1) and D a is the diffusion coefficient for CO2 in the gas phase (1.51 × 10−5 m2 s−1 at 25°C). ΔL ias was calculated as half the mesophyll thickness. g liq was calculated as: where g cw, g pl, g cyt, g en, and g st are cell wall, plasmatic membrane, cytosol, chloroplast envelope and stroma conductance, respectively. g pl and g en were assumed to be 0.0035 m s−1 (Evans et al., 1994). The rest of determinants of the liquid‐phase diffusion pathway were calculated as: where i stands either for cell wall, cytosol, or stroma conductance. D w is the aqueous phase diffusion coefficient for CO2 (1.79 × 10−9 m2 s−1 at 25°C). The dimensionless factor r f,i accounts for the reduction of D w compared with free diffusion in water, and assumed 1.0 for cell walls (Rondeau‐Mouro et al., 2008) and 0.3 for cytosol and stroma (Niinemets & Reichstein, 2003a; Niinemets & Reichstein, 2003b). ΔL i (m) is the diffusion path length in the corresponding component of the diffusion pathway: p i (m3 m−3) is the effective porosity of the diffusion pathway, assumed to be 1 for cytosol and stroma components. Cell wall porosity (p cw) was assumed to be 0.1 (Tomás et al., 2014), one of the highest constant values assumed for calculating g mANAT. Cell wall porosity was also calculated by allowing g mANAT to match g mCF (then termed p CF). The percentage of contribution to g liq was calculated for each component of the liquid‐phase diffusion pathway as: where i in this case stands either for cell wall, cytosol, stroma, plasmatic membrane or chloroplast envelope. The effect of a possible CO2 concentration mechanism in A. agrestis due to the presence of pyrenoids (Smith & Griffiths, 1996) was not considered since x‐intercept of A N–C a curves was 48.7 ± 6.8 μmol CO2 mol−1, not lower than that of other bryophytes lacking pyrenoids like P. juniperinum (42.4 ± 10.7 μmol CO2 mol−1) or P. formosum (29.6 ± 5.3 μmol CO2 mol−1).

Statistical analysis

All analyses were performed using the R statistical software (R Core Team, 2016). Linear regressions were used to test the relationship between mesophyll conductance parameters calculated with curve‐fitting and anatomical modeling, as well as for testing the relation between desiccation tolerance index and anatomical, pressure–volume and gas exchange‐derived parameters. Non‐lineal model (power or logarithmic function) was also fitted to the relationship between CO2 assimilation rates and anatomical derived parameters, as well as to the relationship among pressure volume derived parameters. In those cases, averaged values per species were used. Differences in pressure–volume derived parameters between DT and DS species and differences between percentages of biochemical versus mesophyll limitation to photosynthesis were tested by Mann–Whitney U test, as well as differences between the measured pressure–volume derived parameters and those reported for ferns and angiosperms (Bartlett et al., 2012; Nadal et al., 2018; Shrestha et al., 2007) and for bryophytes (Cruz de Carvalho et al., 2015; Hájek & Beckett, 2008). The packages used were “plyr” (Wickham, 2011), “ggplot2” (Wickham, 2016), “reshape2” (Wickham, 2007), and “corrplot” (Wei et al., 2017).

RESULTS

Long‐term desiccation tolerance

The long‐term desiccation assay revealed a wide range of tolerances, from a DTI of 80% for the leafy liverwort P. canariensis to a null capacity of recovery of F v/F m for all thalloid species, the leafy liverwort S. viticulosa and the moss S. palustre (Figure 1). The lycophyte S. denticulata and the filmy ferns H. caudiculatum and H. dicranotrichum showed a recovery capacity of 26.5, 34.6, and 58.1% of F v/F m, respectively, if the desiccation period did not exceed 24 h.

FIGURE 1

Recovery of F v/F m (i.e. the ratio between final and initial F v/F m values after and before desiccation, respectively) of the studied species after 24 h of rehydration of dry samples kept in silica gel for 1, 7, 14, or 30 days. Boxplots with defaults settings of “ggplot”

Water relations

All the parameters derived from water relation analyses are shown in Table S2. The water potential at turgor loss point (π tlp) was determined mostly by osmotic potential at full turgor (π o) in both DT and DS species (Figure S2A, P < 0.001, R 2 = 0.645) and, to some extent, by bulk modulus of elasticity (ε; Figure S2B, P = 0.003, R 2 = 0.459). The x‐intercept of the π tlp–π o relationship of bryophytes and S. denticulata was significantly different for the regression reported for ferns and angiosperms (Figure 2A), so that the studied species required higher values of π o (less negative) than the reported vascular plants to achieve the same π tlp (in average, 0.47 MPa of difference). The measured bulk modulus of elasticity for bryophytes and S. denticulata ranged between 0.68 and 6.4 MPa, values significantly lower than compiled data of vascular plants (P < 0.001 for the complete meta‐analysis of Bartlett et al., 2012, plus data from Shrestha et al., 2007, and Nadal et al., 2018). Those low ε values corresponded to low RWCtlp, resulting in a logarithmic RWCtlp–ε relationship for all land plants plotted together (Figure 2B).

FIGURE 2

Contextualization of some pressure–volume derived parameters of bryophytes and S. denticulata of the present study in the angiosperms and ferns reported values of Nadal et al. (2018), Shrestha et al. (2007), and compiled data of Bartlett et al. (2012), as well as in the bryophytes reported values of Hájek and Beckett (2008) and Cruz de Carvalho et al. (2015). (A) Relationship between osmotic potential (π o) and water potential at turgor loss point (π tlp) for tracheophytes and presented data (P = 0.016 for differences in intercept; slope non‐significatively different, P = 0.135). (B) Relationship between bulk modulus of elasticity (ε) and relative water content at turgor loss point (RWCtlp; P < 0.001 for the logarithmic relationship: RWCtlp = 28.9 log10(ε) + 51.8). Data points are mean values of each reported species, except for values of Shrestha et al. (2007) of panel (B), which correspond to undetermined individuals

Diffusional limitation to photosynthesis

Gas exchange‐derived parameters are shown in Table S3. The obtained A N–C c curves are shown for each species in Figure S2. P. juniperinum and P. formosum presented the highest photosynthetic capacity, with the highest values of V cmax (26.8 ± 3.8 and 22.9 ± 1.1 μmol m−2 s−1, respectively, mean ± SE) and J max (39.6 ± 2.7 and 27.6 ± 2.3 μmol m−2 s−1, respectively). Besides Polytrichum species, most of the studied bryophytes (63%) showed values of V cmax and J max below 5.6 and 7.2 μmol m−2 s−1, respectively, being the leafy liverworts S. viticulosa and P. canariensis and the moss F. serrulatus the species with the lowest averaged photosynthetic capacities, i.e. V cmax below 4.5 μmol m−2 s−1 and J max below 5.3 μmol m−2 s−1. The correlation between J max obtained by the curve‐fitting method and J FLU calculated by chlorophyll fluorescence was significant (P = 0.005, R 2 = 0.555, Figure S3), with only two species—S. denticulata and L. cruciata—with J FLU values 2.4‐ and 1.9‐fold higher than J max. Biochemical limitation was significantly larger than diffusional limitation in all studied species (Table S4).

Values for the most relevant parameters measured for the anatomical model are shown in Table S5. A N was highly and linearly correlated with g mCF and S c (Figure 3A,B), as well as with T cw in a non‐linearly way (Figure 3C). S m were lower than 3.7 cm2 cm−2 for most of the studied species, only higher in species with f ias resulted from air‐filled mesophyll or ventilated thallus—S. denticulata and L. cruciata (9.4 ± 1.0 and 6.6 ± 0.3 cm2 cm−2, respectively)—or lamellae—P. formosum and P juniperinum (10.3 ± 0.9 and 13.8 ± 0.3 cm2 cm−2, respectively). S c and S m were highly correlated (P < 0.001, R 2 = 0.816, regression not shown), with chloroplasts occupying on average 45% of the air facing surface. g mCF/S c and T cw were not correlated significantly (P = 0.186, data not shown).

FIGURE 3

Anatomical determinants of net CO2 assimilation rate (A N). Relationship between A N and (A) mesophyll conductance calculated though curve‐fitting method (g mCF, P < 0.001, r 2 = 0.925), (B) surface area of chloroplasts exposed to intercellular airspaces per unit leaf area (S c, P < 0.001, r 2 = 0.877), and (C) thickness of cell wall (T cw, P < 0.001, r 2 = 0.680 for the power function A N = 0.88 T cw −1.13). Data points are mean ± SE

Mesophyll conductance calculated using the anatomical model (g mANAT) was correlated with g mCF (Figure S4, P < 0.0001), being g mCF 1.7‐fold higher than g mANAT, except for the liverwort P. endiviifolia. When effective porosity was calculated so that g mANAT matched g mCF (p CF), values higher than 0.3 m3 m−3 were obtained for most of the studied species (Table S5). g mANAT could match g mCF in A. agrestis and S. denticulata only using unrealistic values of effective porosity higher than 1.

T cw varied between 0.39 (for S. denticulata) and 2.86 μm (for F. serrulatus) and was non‐linearly correlated with T chl (Figure 4A, P = 0.030 for the logarithmic relationship, P = 0.053 for power function), so that species with the thickest cell walls also presented the thinnest chloroplasts (F. serrulatus, P. purum, and P. canariensis). The percentages of limitation to the liquid components of mesophyll conductance of stroma (l st) and cell wall (l cw) were also negatively correlated (Figure 4B). When porosity was estimated by adjusting g mANAT to g mCF—, l cw was lower than 43% in favor of l st, except for the species with thickest cell walls (F. serrulatus and P. purum) or with low p CF (P. endiviifolia, with a p CF = 0.023 ± 0.008 m3 m−3). The variation of the percentages of limitation of all the components of the diffusion pathway considered for the anatomical model of mesophyll conductance are shown in Table S6 for constant and variable porosity.

FIGURE 4

Balance between the morphology of cell wall and chloroplast. (A) Non‐linear relationship between thickness of chloroplast (T chl) and thickness of cell wall (T cw). Line indicates power function: T cw = 1.24 T chl −0.677 (P = 0.053). Inset graph indicate logarithmic T cw–T chl relationship (P = 0.030). (B) Relationship between percentage of anatomical limitation to liquid component of mesophyll conductance of stroma (l st) and cell wall (l cw) with a porosity calculated by adjusting g mANAT to g mCF (P < 0.001 for linear regressions). See Figure 3 for legend

Factors related to desiccation tolerance

For DT species as defined in this study, i.e. species with DTI > 0, that index was negatively correlated with ε (Figure 5A, P < 0.001), that is, the largest long‐term desiccation tolerance was observed in those DT species with more elastic tissues. DS species presented low values of ε (<0.2 MPa, elastic tissues) despite their null capacity to recover from desiccation. Something similar occurred for osmotic potential, which was positively correlated with DTI only for DT species (Figure 5B, P = 0.004). DTI was also negatively correlated with p CF for DT species (Figure 5C, P = 0.032). Neither T cw nor A N or g mCF were correlated with DTI (Figure 5D–F). a f and C ft were not correlated with long‐term desiccation tolerance (data not shown), although their absolute values were significantly different for DT and DS species (Figure S5), so that on average DT species presented a f values 3‐fold higher than DS species and C ft values 48% lower than DS ones.

FIGURE 5

Physiological and morphological determinants of desiccation tolerance. (A) Relationship between desiccation tolerance index (DTI) and bulk modulus of elasticity (ε, P < 0.001), (B) osmotic potential (π o, P = 0.004), (C) porosity calculated from g mANAT adjusted to g mCF (p CF, P = 0.032), (D) thickness of cell wall (T cw), (E) net CO2 assimilation rate at 400 μmol mol−1 of external CO2 (A N), and (F) mesophyll conductance by curve‐fitting method (g mCF). Lines indicate regressions for desiccation tolerant species “sensu this study,” i.e. all species with positive values of DTI. Data points indicate mean ± SE

DISCUSSION

Water relations and cell wall features as promoters of desiccation tolerance

The long‐term fast desiccation/rehydration assay allowed the calculation of a quantitative index of desiccation tolerance (DTI). Our data show that DTI was mostly correlated with the bulk modulus of elasticity and osmotic potential at full turgor only in DT species (Figure 5A,B). All studied thalloid liverworts, the hornwort A. agrestis, the moss S. palustre and the leafy liverwort S. viticulosa experienced a complete inhibition of F v/F m for the shortest tested period of the fast desiccation applied in our study (24 h). For softer desiccation treatments—lower rate and depth of drying, especially in hardened individuals by previous abscisic acid treatments—some of these species have shown some recovery of F v/F m after rehydration (Pence et al., 2005; Pressel et al., 2009). All these sensitive species showed high elasticity, suggesting that elasticity is not the last determinant of tolerance, but still it is a compulsory requirement for bryophytes to tolerate the mechanical stress suffered after desiccation during longer periods. Low elasticity seems to be a generalized characteristic of bryophytes (Beckett, 1997; Proctor et al., 1998; Proctor, 1999; Hájek & Beckett, 2008; Cruz de Carvalho et al., 2015; notice that different methods for calculating ε were used in those studies), only emulated by some grasses and crops (Bartlett et al., 2012; Nadal et al., 2018). On the contrary, the values of π o and π tlp measured for bryophytes in this study (between −0.7 and −3.5 MPa, and −1.7 and 0.3 MPa, respectively) are more frequently found in angiosperms (Figure 2 for the contextualization of π o–π tlp and RWCtlp–ε relationship of the present study in reported values for tracheophytes). In their meta‐analysis, Bartlett et al. (2012) showed the association of low π o or π tlp (more negative values) with arid ecosystems and, therefore, with drought tolerance, placing ε and RWCtlp into a secondary role. Our study demonstrates that the contrary occurs in bryophytes and possibly also in DT lycophytes and filmy ferns, with elasticity (i.e. low ε) being the only parameter highly correlated with a desiccation tolerance index, while absolute values of π o and therefore π tlp, only increase with rigidity. The high elasticity of bryophytes may explain their position in the π tlp–π o relationship and their low values of RWCtlp according to the equations presented by Bartlett et al. (2012) defining the relationships among pressure–volume parameters. The concurrence of a possible mechanism of incompatibility between ε and π o in bryophytes is not clear. The elasticity of cell walls in this group of plants may be mainly dependent of their chemical composition, which includes a higher accumulation of callose (Popper, 2008; Popper et al., 2011; Popper & Fry, 2003). Question arises as to whether the osmolytes that provoke the increase of absolute values of π o are also involved in making the tissues of the studied species more rigid, maybe not so clearly evidenced in tracheophytes (see correlations between ε and π o of Zhang, 1998, Shrestha et al., 2007, and Bartlett et al., 2012) due to the role of additional structures in the determination of elasticity—for instance, thickened cuticles, lignified epidermal cells, fiber bundles, and the presence of sclereids (Salleo & Gullo, 1990).

Curiously, the effective porosity estimated by adjusting g mANAT and g mCF was significantly correlated with the DTI in DT species, but the thickness of cell wall was not (Figure 5C,D). As introduced, the high values of p CF (>0.3 m3 m−3 for most of the studied species) compared with the ones assumed for vascular plants (<0.1 m3 m−3, Tomás et al., 2013, 2014; Tosens, Niinemets, Vislap, et al., 2012, Tosens, Niinemets, Westoby, & Wright, 2012, Tosens et al., 2016; but see also Nobel, 1991) is consistent with the position of bryophytes in the asymptotic A N–T cw relationship. Furthermore, high p CF may explain why the suggested link between T cw and tissue rigidity for tracheophytes (Peguero‐Pina et al., 2017) seems not to apply for the studied bryophytes. Nadal, Brodribb, et al. (2021) have also observed variations in ε associated to DT and DS fronds of the fern Anemia caffrorum independent of the T cw, which was invariable. We recognize that the effective porosity calculated in our study could be overestimated due to the effect of additional biochemical enhancers of mesophyll conductance, i.e. carbonic anhydrases and aquaporins (reviewed by Gago et al., 2020), which could explain p CF values >1 m3 m−3. Assuming that this possible underestimation is constant among studied species, our data suggest that having less porous cell walls seemed to be more relevant than their thickness for long‐term desiccation tolerance in bryophytes and maybe lycophytes. We hypothesize that the less porous cell walls of DT species (p CF = 0.27 m3 m−3 for P. canariensis and 0.18 for P. purum) may evade penetration of digestive enzymes of microorganisms and enhance resistance to mineralization when desiccated. Both low nitrogen concentration and the presence of phenolic compounds have been pointed out as responsible for the low rates of decomposition of bryophytes as compared with tracheophytes (Scheffer et al., 2001; Turetsky, 2003). Perhaps, porosities of vascular plants have less interspecific variability and then an increase of T cw and/or a decrease of nitrogen concentration are the only factors that resurrection plants can do for surviving desiccation periods, explaining their lower A N/T cw ratio (Nadal, Perera‐Castro, et al., 2021).

Beyond T cw, there must exist a direct relation between effective porosity, apoplastic fraction (or volume of apoplastic water), and the volume of total cell walls. In future studies, the combination of pressure volume curves and the deconstruction of 3D cell wall volume though the new X‐ray microscanning technology (Théroux‐Rancourt et al., 2020) would allow direct measurements of cell wall effective porosities and shed light about their role in desiccation tolerance and mineralization.

Role of cell walls in limiting photosynthetic capacity of bryophytes

Independently of the role of effective porosity in determining desiccation tolerance, both A N and T cw were unrelated to DTI for the studied species. Surely, inherent variations in phyllids/thalli structure are defining g m and photosynthetic capacity in bryophytes, as can be evidenced by the correlation between g m modeled from anatomy and gas exchange (Figure S4). Among the measured anatomical traits, the present study evidences a clear dominance of S c above T cw in determining photosynthetic capacity of bryophytes and S. denticulata (Figure 3) in line with data reported by Carriquí, Roig‐Oliver, et al. (2019) for bryophytes and lycophytes but contrary to the meta‐analysis of Flexas et al. (2021) pooling bryophytes and tracheophytes. These discrepancies among anatomical determinants of photosynthesis along phylogeny, together with the null relationship between A N or g mCF and DTI reported in the present study, suggest that the evolutionary trade‐off between photosynthetic capacity and desiccation tolerance proposed by Hanson et al. (2014) and Carriquí et al. (2015) is not based on a direct common mechanistic constrain related with T cw.

Furthermore, the percentage of limitation of the liquid phase of mesophyll conductance due to cell wall (l cw) was below 43% in most species when forcing p cw = p CF, resulting in increased estimated limitation due to the chloroplast stroma (Figure 4B). A trade‐off between chloroplast and cell wall thickness was observed for the studied species (Figure 4A). While a similar relationship has been also established in some vascular plants (Tosens, Niinemets, Westoby, & Wright, 2012), a multispecies compiled dataset has instead shown a positive lineal correlation between them (Ren et al., 2019), which seem to be contradictory with the idea suggested by Tosens, Niinemets, Westoby, and Wright (2012) that species with thicker mesophyll cell wall resistance may be evolutionary constrained to reduce the diffusion pathway in chloroplast and stroma resistance. The advantages of bryophytes with thinner cell walls in presenting thicker chloroplasts are not clear. Since light harvesting capacity is restricted in this group of plants due to their morphology—phyllidia mostly unistratose and with light harvesting and chlorophyll concentration depending mostly on canopy structure (Wang et al., 2016)—, thicker chloroplasts would enhance chlorophyll and Rubisco concentration per unit area (Li et al., 2013), as well as light absorbance. The positive effects of enhancing light harvesting by increasing T chl must be more relevant than the negative effects of increasing stroma resistance in species with the thinnest cell walls and low S c.

In addition to the possible balanced role of cell wall and chloroplast size in determining mesophyll conductance in a scenario of high p cw, photosynthesis was mainly limited by biochemistry rather than mesophyll conductance (Table S4). This is in contrast with previous report by Carriquí, Roig‐Oliver, et al. (2019), who showed that mesophyll conductance limitation was the most limiting for photosynthesis in bryophytes. This apparent discrepancy could be due either to the fact that different species were analyzed in the two studies, but also to the fact that limitation analysis in Carriquí, Roig‐Oliver, et al. (2019) were based on different methodology for estimating g m. Such discrepancies reveal uncertainty as for considering g m as the main ruler of photosynthetic capacity of bryophytes and need to be analyzed in depth in the near future. Despite these specific discrepancies, the results of the present study still confirm some conclusions of Carriquí, Roig‐Oliver, et al. (2019), i.e. (a) gm is very low in bryophytes as compared with tracheophytes, and (b) it significantly contributes to limit their photosynthesis, which results in a very strong linear dependency of A N on g mCF (Figure 3A).

CONCLUSIONS

In conclusion, elasticity is a common feature that favors desiccation tolerance beyond osmotic potential in bryophytes, the lycophyte S. denticulata, and filmy ferns, although it is not exclusive of desiccation tolerant species. In addition, less porous cell walls seem to be more relevant for enhancing desiccation tolerance than the increase of T cw within the studied species, which suggests that the distinctive features of cell wall (generally thick cell walls possibly with high porosity) are not involved in the generalized desiccation tolerance reported for this group of plants. The role of T cw in constraining the evolutionary tendency of decreasing desiccation tolerance along phylogeny of land plants, from the early branched group of bryophytes to the more recently diverged angiosperms, is rejected, since within bryophytes thick cell walls and low A N does not correspond with higher tolerance to desiccation. Furthermore, our data evidence that the contribution of the cell walls to limiting mesophyll conductance is balanced with the thickness of stroma and, therefore, with chloroplast size, in the scenario of a high cell wall porosity estimated here. This observation, together with the high linear correlation between S c and A N suggest that T cw is not only irrelevant for desiccation tolerance, but possibly also less relevant than previously thought for constraining photosynthetic capacity of bryophytes.

AUTHOR CONTRIBUTION

Alicia V. Perera‐Castro and Jaume Flexas designed the study. Alicia V. Perera‐Castro conducted the experiments, performed the analysis and wrote the first draft of the manuscript. Both authors contributed to the following and final versions of the manuscript.

Figure S1. Relationship between water potential at turgor loss point and osmotic potential at full turgor or bulk modulus of elasticity.

Figure S2. A N–C c curves of the studied species.

Figure S3. Relationship between electron transport rate obtained by curve‐fitting method and chlorophyll fluorescence.

Figure S4. Relationship between mesophyll conductance calculated by anatomical modeling and by curve‐fitting method.

Figure S5. Differences in apoplastic fraction and capacitance at full turgor between desiccation and sensitive tolerant studied species.

Table S1. In vitro Rubisco Michaelis–Menten constants used in this study.

Table S2. Pressure–volume derived parameters.

Table S3. Gas exchange‐derived parameters.

Table S4. Percentage of stomatal, mesophyll conductance, and biochemical limitation to photosynthesis.

Table S5. Anatomical derived parameters.

Table S6. Percentage of limitation of liquid component to mesophyll conductance and the contribution of each CO2 diffusion compartment according to the assigned effective porosity.

Click here for additional data file.