• The nineteenth century air-gun explanation for explosive spore discharge in Sphagnum has never been tested experimentally. Similarly, the function of the numerous stomata ubiquitous in the capsule walls has never been investigated.
• Both intact and pricked Sphagnum capsules, that were allowed to dry out, all dehisced over an 8–12 h period during which time the stomatal guard cells gradually collapsed and their potassium content, measured by X-ray microanalysis in a cryoscanning electron microscope, gradually increased. By contrast, guard cell potassium fell in water-stressed Arabidopsis.
• The pricking experiments demonstrate that the air-gun notion for explosive spore discharge in Sphagnum is inaccurate; differential shrinkage of the capsule walls causes popping off the rigid operculum. The absence of evidence for a potassium-regulating mechanism in the stomatal guard cells and their gradual collapse before spore discharge indicates that their sole role is facilitation of sporophyte desiccation that ultimately leads to capsule dehiscence.
• Our novel functional data on Sphagnum, when considered in relation to bryophyte phylogeny, suggest the possibility that stomata first appeared in land plants as structures that facilitated sporophyte drying out before spore discharge and only subsequently acquired their role in the regulation of gaseous exchange.
Among all the explosive propagule-dispersing mechanisms found in the plant and fungal kingdoms, that of Sphagnum is unique. Here alone are the spores discharged by an air-gun mechanism which has been described as the result of pressure as the capsules dry out and shrink ultimately popping off of the lids (operculum). Indeed, the noise and sight of popping Sphagnum capsules, was well known to Linnaeus and is familiar to those who walk on bogs in the summer months (Goebel, 1905; Sundberg, 2005). Explosive spore discharge in other groups involves either the generation of hydrostatic pressure, such as in brown algae (Toth, 1976) and fungi (Ingold, 1965; Trail, 2007) or the sudden release of tension via cavitation, such as the spore catapult in leptosporangiate ferns (Moran, 2004).
The air-gun explanation for the discharge of Sphagnum spores has been generally accepted unchallenged in the literature for over a century based almost entirely on two early works by Nawaschin (1896) and Cavers (1911). From calculations of capsule volume Nawaschin concluded that shrinkage, as the capsules dry out, builds up air pressure to 4–6 atmospheres whereupon the opercula are forcibly ejected. The volume of air liberated when capsules dehisced in absolute alcohol produced a similar figure of 5 atmospheres at the point of dehiscence. Nawaschin's drawings clearly show that although the stomatal guard cells in the capsule wall shrink and partially separate as the capsules dry out they never open internally. In fact, intercellular air spaces are conspicuously absent from Sphagnum capsules until they begin to dry out. Cavers (1911) elaborated Nawaschin's conclusions but without any experimental evidence; as the spores mature, shrinkage of the subjacent columella creates a spherical space into which air diffuses through the walls between the stomatal guard cells as long as these remain moist. When the capsules begin to dry out the air is trapped. The change in capsule shape from spherical to cylindrical is the result of buckling of the thin periclinal walls of the epidermal cells with the much thicker anticlinal walls remaining largely unchanged. Two subsequent anatomical studies (Maier, 1973; Boudier, 1988) confirmed these observations, except for Cavers’ (1911) assertion of inward diffusion of air since substomatal air chambers are absent in Sphagnum. These two later authors demonstrated that dehiscence involves the rupture of the thin-walled cells around the operculum. However, both accepted without question the air-gun notion of explosive spore discharge but failed to explain the precise mechanism.
Reference to a possible role for the stomata in spore discharge in Sphagnum raises the question as to the real role of these enigmatic structures. The fact that they are ubiquitous in the genus with numbers per capsule ranging between 100 and 1000 (Paton, 1957; Paton & Pearce, 1957; Andrews, 1961) points to some crucial function. This is particularly intriguing as in other mosses some of the highest recorded numbers include 250 in Polytrichastrum alpinum, 240 in Philonotis and 185 in Bryum. In addition, numbers vary widely between closely related taxa (e.g. 200 in Funaria hygrometrica compared with only 14 in Aphanoregma (Physcomitrella) patens) as does presence and absence between closely related genera; for example the absence in Atrichum and Pogonatum in the Polytrichales (Paton, 1957; Duckett & Ligrone, 2004).
The first illustrations of Sphagnum stomata (Schimper, 1858) demonstrate that they cannot function in the exchange of oxygen and carbon dioxide or act as conduits of water loss because they are covered by the calyptra throughout sporophyte ontogeny (Duckett & Ligrone, 2004), which in turn is encased in perichaetial leaves until the pseudopodium begins to elongate. This is reminiscent of liverwort sporophyte development within the confines of gametophytic tissues and where the capsules are devoid of stomata. The capsules in Sphagnum are exposed to the external environment only when the calyptra layer withers as the pseudopodium begins to elongate (Boudier, 1988). Wiildner (1887), Haberlandt (1886) and Cavers (1911) were aware that, in Sphagnum, the stomata do not behave normally but doubts about their function remain. Some authors consider them as rudimentary (Paton & Pearce, 1957; Parihar, 1972; Crandall-Stotler, 1984; Schofield, 1985; Smith, 2004), whereas others (Crum & Anderson, 1981; Ireland, 1982, Boudier, 1988) refer to them as pseudostomata to reflect the functional uncertainties and Raven (2002) argues in favour of homoplasy of stomata across land plants. Perhaps the most salient observations are by Boudier (1988): stomata develop very late in relation to sporogenesis, lose their contents, which are indistinguishable from the other epidermal cells even when the calyptra breaks down (cf. the guard cells with starch-filled plastids in Funaria; Sack & Paolillo, 1983), and they lack both pores and substomatal air spaces.
Recent moss phylogenies based on molecular and morphological data (Newton et al., 2000; Cox et al., 2004; Renzaglia et al., 2007) suggest that the stomata in Sphagnum are not homologous with those in other mosses, which is a reasonable conclusion given the suite of other sporophytic differences between these two groups.
The aims of the study described here were: to test experimentally the veracity of the air-gun hypothesis; and to establish whether or not these structures have a key role in spore discharge. Our findings are discussed in the broader context of stomatal evolution in land plants.
Materials and Methods
Various Sphagnum species (S. cuspidatum Ehrh. ex Hoffm. S. palustre L., S. squarrosum Crome, S. subnitens Russow & Warnst. and S. tenellum (Brid.) Bory) with ripe and nearly ripe capsules were collected in the summer months (June–August) from sites in southern England. The wild-collected specimens were maintained at 4°C in a growth chamber.
For air drying, the sporophytes were allowed to dry using the laboratory atmosphere (approx. 20–22°C and 60% RH) as a standard drying environment. Samples of 100 or more excised sporophytes were placed on a microbalance either intact or following pricking with a fine needle and weighed at regular intervals together with recording the numbers that had dehisced. Three replicates were done for each treatment. The relative water loss (RWL%) was calculated as 100 − [(FWx − DW) × 100/(FW0 − DW)] where FW0 and FWx indicate the fresh weight at the beginning and end of the experimental time, respectively. Dry weight was determined after oven drying the capsules at 80°C overnight.
The same weight data were collected from sporophytes of mature undehisced capsules of Bartramia pomiformis Hedw. that are approximately the same size and shape as those of S. palustre and for Atrichum undulatum as a typical moss exemplar lacking stomata (Paton & Pearce, 1957). The effects of capsule immersion in 2 m NaCl and sucrose were also noted. The presence of air in capsules at different stages of drying out was monitored by squashing them in immersion oil.
For routine scanning electron microscopy (SEM) sporophytes were dehydrated in a 1 : 1 ethanol–acetone series, critical-point dried using CO2 as the transfusion fluid, sputter-coated with 390 nm palladium-gold and viewed using a Hitachi S570 scanning electron microscope (Pressel & Duckett, 2006).
For cryo-SEM, sporophytes were first mounted on an aluminium stub using Tissue-Tek (Sakura Finetek, Zoeterwoude, the Netherlands) and plunged in liquid nitrogen slush to preserve their hydrated and partially dehydrated states in a frozen condition in an Alto 2500 (Gatan, Abingdon, UK). The frozen sporophytes were vacuum-transferred to a high vacuum cryogenic preparation chamber to prevent contamination and the build-up of ice. Ice was sublimed off the surface by raising the temperature up to –90°C for 5 min. The samples were then cooled to –130°C and AuPd sputter coated with a cold magnetron sputter coater. The coated sporophytes were then inserted directly into the scanning electron microscope via an airlock to avoid ice build-up and to maintain their frozen state. Inside the scanning electron microscope, the samples rested on a cold stage with the temperature maintained at −130°C. A FEI Quanta 3D FEG dual beam microscope (FEI Company, OR, USA) with an Oxford Instruments (Abingdon, UK) energy dispersive spectroscopy (EDX) attachment was used to image the sample at 30 kV and beam current of 0.11 nA. Focused ion beam milling at 30 kV and 0.1 nA was done on 5 × 5 µm areas of the sporophytes and an Oxford Instruments EDX attachment was used for elemental analysis.
Elemental spectra sets were obtained for five stomata plus adjacent capsule epidermal cells for fully turgid, partially dehydrated and dehisced capsules of Sphagnum subnitens with epidermal cells of hydrated pseudopodia providing controls for potassium concentrations in typical somatic cells.
For comparison with Sphagnum, elemental spectra sets were obtained from five open stomata and adjacent epidermal cells immediately following removal of leaves from well-watered Arabidopsis plants. Other detached leaves were allowed to dry out in the laboratory and five further spectra sets collected as soon as the stomata closed (after c. 1 h and 10% water loss).
Entire, mature capsules of Sphagnum tenellum lost in excess of 50% (RWL) within the first 5–6 h of drying and up to 98% after 10–12 h at 20–22°C. Capsules began to dehisce after 5–6 h and 50% of the capsules had dehisced after 8–9 h (90% RWL). After 10–12 h all the capsules had dehisced (Fig. 1a). At 28–30°C rates of water loss were faster with all the capsules dehiscing after 8 h (data not shown). Rates of water loss and dehiscence of S. palustre and S. subnitens followed the same pattern (data not shown). Rates of water loss of mature Bartramia and Atrichum capsules were much slower (20% after 5–6 h) and only reached 30–50% RWL after 24 h of drying (Fig. 1b). In addition the lids remained in situ. Rates of water loss similar to, or even slower than, those reported in Bartramia were recorded for a range of other mosses taxa, for example Polytrichum formosum, Brachyhecium rutabulum and Isothecium myosuroides (data not shown).
Pricking of the Sphagnum capsules led to an increase in the rate of water loss and, in parallel, faster capsule dehiscence. The figures for all three species followed exactly the same pattern; after 1.5 h of drying 50% RWL and 30% dehiscence, after 4–5 h 90% RWL and 95% dehiscence and 100% dehiscence after 6–7 h. Our interpretation of these differences is that pricking speeded up evaporation via the holes created. Immersion of capsules in 2 m NaCl and sucrose had no discernable effects on their morphology and did not lead to dehiscence. Fig. 2 illustrates the changes and gradual accumulation of air in Sphagnum capsules leading up to dehiscence. When surrounded by perichaetial leaves the pale-brown spherical capsules containing mature spores do not contain any air. The same is true until the pseudopodia are fully elongated. Only after the rupture of the calyptra, when the capsules begin to dry out, does air gradually begin to build up. Initially, this is not associated with any appreciable change in the volume of the capsules but, by the time they are shrunken and on the point of dehiscence, air completely fills the internal chamber both in the columellar region and around the now dry spore mass. The desiccation-induced changes in capsule shape are fully reversible. When dried-out dehisced capsules, even those from old herbarium specimens, are placed in water, within 10 min they regain their spherical shape. Immersion of undehisced capsules in 2 m NaCl or 2 m sucrose had no effect on their shape.
Figure 3 shows capsules at different stages in the dehiscence process. Immediately after the pseudopodium elongates the calyptra, with the shrivelled remains of the archegonial neck at its apex, remains intact (Fig. 3a) and the cells, with scattered peripheral chloroplasts around a central vacuole, remain fully hydrated (Boudier, 1988) thus acting as a continuous barrier to any significant diffusion of gases into or from the capsules. Breakage of this tissue signals the onset of drying out (Fig. 3b,c) until at dehiscence the formerly smooth capsule walls have become deeply corrugated (Fig. 3e). Within 5 min of immersion in water such desiccated capsules regain their predehiscence spherical shape (Fig. 3d). The stomata (Fig. 3f–h) before drying out are overlain by the calyptra; they have swollen guard cells but lack a central pore and intercellular air spaces internally. The outer surface of rehydrated dehisced capsules becomes less deeply grooved but the guard cells remain collapsed (Fig. 3i).
Figure 4 illustrates the sequential changes to the surface of drying out capsules. After 2–3 h of drying out the initially highly swollen guard cells (Fig. 4a) develop central depressions while the other epidermal cells remain turgid and unchanged (Fig. 4b,c). Approaching dehiscence the capsule surface is highly undulate and the guard cells deeply shrunken (Fig. 4d,e). At or immediately post-dehiscence the capsule walls remain deeply grooved and the outer walls of the guard cells are often ruptured (Fig. 4f,g). It is therefore unsurprising that the guard cells do not recover their original shape when the capsules are rehydrated (Fig. 3i). Capsules continue to dry out post-dehiscence (Fig. 4h,i) with slit-like depressions marking the positions of the epidermal cell lumina. Very low water content also leads to the breakdown of the common wall between the guard cells – a feature usually found in herbarium specimens.
Cryoscanning electron microscopy and X-ray microanalysis
Figure 5 shows cryo-SEM images of Sphagnum and Arabidopsis after focused ion beam milling before X-ray microanalysis. The milling holes are clearly located over the swollen hydrated (Fig. 5a) and collapsed guard cells (Fig. 5b), with holes in adjacent cells (Fig. 5a) and the pseudopodium (Fig. 5c) acting as control sites. The milling holes are similarly precisely located over the smaller guard cells in Arabidopsis (Fig. 5d,e). The amounts of potassium, expressed as weights detected by the probe in the various cells in Sphagnum capsule walls and in Arabidopsis leaves are summarized in Table 1. Typical spectra that underpin the numerical data are illustrated in Fig. 6.
Table 1. Mean from 8–10 readings of the percentage weights of potassium from X-ray microanalysis
N/A, not applicable.
3.36 ± 0.36
3.24 ± 0.45
1.56 ± 0.12
3.91 ± 0.64
1.06 ± 0.09
1.14 ± 0.08
1.48 ± 0.06
Partially dehydrated capsule
1.62 ± 0.23
1.53 ± 0.27
1.99 ± 0.09
1.97 ± 0.06
Broken guard cells
0.76 ± 0.06
Sporophyte surface wet
0.32 ± 0.03
Sporophyte surface dry
0.95 ± 0.09
The quantitative data and spectra show similar peaks, and potassium concentrations between the guard cells and adjacent epidermal cells in the leaves of Arabidopsis in the turgid state with open stomata (Fig. 6a,b) contrast markedly with the much lower potassium peaks and concentrations in the guard cells of slightly wilted leaves (Fig. 6c) with closed stomata.
In Sphagnum, however, drying out of the sporophytes produces very different results. From similar concentrations of potassium in the epidermal cells of the pseudopodium and fully hydrated stomatal guard cells (Fig. 6e), and higher values for the sporophyte epidermal cells (Fig. 6f), drying out sees an initial increase in the potassium in the guard cells to about the same amount as in the epidermal cells (Fig. 6g). At the point of dehiscence potassium concentration is the same in both cell types and significantly higher than in fully hydrated capsules before the onset of drying out (Table 1). We interpret the increase in potassium on the surface of desiccated capsules as the result of leakage from broken guard cells (Fig. 4f–i).
Comparison of the cryo-specimens and specimens prepared via critical point drying reveals one further interesting feature of Sphagnum capsules: the complete absence of any cuticle-like covering. This is in striking contrast to most other moss capsules, not to mention taxa, with wax-covered leafy gametophores (Proctor, 1979).
This study clearly demonstrates that the air-gun hypothesis for the explosive discharge of Sphagnum spores is inaccurate; the build-up of air pressure in punctured capsules is impossible. The mechanism of spore discharge can now be much more simply explained solely in terms of differential shrinkage of the capsule walls – namely buckling of the thin periclinal walls (Cavers, 1911; Maier, 1973; Boudier, 1988) that eventually causes popping off of the rigid operculum whose diameter scarcely changes.
Together the microscopic and weight data clearly demonstrate that capsule drying out involves water loss principally through the thin-walled guard cells compared with the other cells in the capsule walls that have much thicker walls (Boudier, 1988). Absence of visible effects with 2 m sucrose and 2 m NaCl indicate that osmotic forces have no role in the process. The reversibility of the shape changes, which also occur in rehydrated herbarium specimens many decades old, confirms that the mechanism of capsule shrinkage resides solely in the properties of the unevenly-thickened epidermal walls. At the moment of dehiscence sporophyte water loss is over 70% of the fresh weight and only 15% above the dry weight. By contrast, in Bartramia and Atrichum, exemplars typical of other mosses with stomate and estomate capsules, respectively (Paton & Pearce, 1957), the sporophytes lose water much more slowly. These rates of sporophytic water loss, even in Sphagnum, are considerably slower than the exponential rates of water loss from moss gametophores under the same experimental conditions; for example, Polytrichum formosum has a half-desiccation time of just 17 min (Proctor et al., 2007). Sphagnum shoots also dry out at similar rates to gametophores of other mosses (Hájek & Beckett, 2008).
The X-ray microanalytical data reveal major differences in the behaviour of Sphagnum stomata compared with those in Arabidopsis, a typical example from vascular plants (Table 1). Our data show that in the latter the potassium content of the guard cells in the turgid state is much the same as in adjacent epidermal cells (Fig. 6a,b) and falls to much lower amounts than the epidermal cells when the plants are under water stress (Fig. 6c,d). By contrast, in fully hydrated Sphagnum capsules, the guard cells have a lower potassium content than the epidermal cells, which increases in both as the capsules dry out. Thus the stomata in Sphagnum do not respond to desiccation in the same way as those in vascular plants and there is no evidence of potassium regulation between the sporophyte guard cells and the adjacent epidermal cells. In addition, the guard cells of Sphagnum lack the xylans found in vascular plants, a feature also absent from hornwort guard cells (Carafa et al., 2005), and xyloglucans differ between mosses and vascular plants (Pena et al., 2008). Sphagnum also lacks the unevenly thickened walls with ledges subtending the pores characteristic of typical stomata, including those in other mosses and hornworts (Paton, 1957), and substomatal air spaces. Our study thus points to drying out and capsule shrinkage leading to rapid dehiscence as the sole role for the stomata in Sphagnum rather than metabolic gaseous exchange as found in other groups. Our data on the different guard cell potassium contents of open and closed stomata of Arabidopsis are closely in line with previous X-ray microanalytical studies of potassium fluxes in vascular plant stomata (Dayanandan & Kaufman, 1975; Willmer et al., 1983; Langer et al., 2004; Hayles et al., 2007). However, it should be noted that the actual numerical values between these different studies are not directly comparable because of differences in the equipment used and the different beam diameters, currents and take-off angles. It would now be of considerable interest to explore the relative potassium contents of guard cells and epidermal cells in other mosses and in hornworts, the likely sister group to vascular plants (Renzaglia et al., 2007), in response to cues that induce stomata closure. Unlike Sphagnum, the stomata of other bryophytes have substomatal chambers and an intercellular network, not found in their estomate counterparts.
The present confirmation that the sporophytes of Sphagnum, unlike those of other mosses, remain enclosed by a continuous layer of calyptral cells raises questions about their carbon acquisition. Physiological studies on the sporophytes in a range of mosses but not including Sphagnum, show these to be clearly capable of photosynthesis from early in their development even when completely encased by the calyptra (e.g. in Polytrichum), although translocation of carbohydrates across the placenta from the gametophytes also makes a significant contribution to sporophytic biomass (see review by Ligrone & Gambardella, 1988). The same is also likely to be the case in Sphagnum since, as in other mosses, chlorophyll is present from the earliest stages in sporophyte development together with plastids with well-developed thylakoid systems up until the capsules change colour from green to brown and begin to dry out (Ligrone et al., 1993).
Our findings add a new level for interpretation for the evolution of stomata. All recent phylogenies (Newton et al., 2000; Cox et al., 2004; Shaw & Renzaglia, 2004), molecular plus or minus morphology, place Sphagnum as sister or near-sister to other mosses. Whatever the precise configuration of the trees there is always at least one group of mosses (Andreaea, Andreaobryum, Takakia) lacking stomata between Sphagnum and those with unequivocal stomata (Polytrichales, Oedipodiales, Tetraphidales) subtended by photosynthetic tissues with intercellular spaces. From these trees alone, regardless of any structural or functional considerations, it is clearly evident that the stomata of mosses, other than Sphagnum, must be either a reacquisition or are completely different structures, as argued by Cox et al. (2004) from Boudier's (1988) morphological data. The present demonstration of a key role only in capsule dehiscence supports the view of Cox et al. (2004) and is consistent with the hypothesis that true stomata evolved once in mosses at the base of the Polytrichales. Following the present clarification of stomatal function in Sphagnum, does this resolve the questions of homology and homoplasy? Not quite; in a living world simply littered with examples of transference of function (Corner, 1958; Baum & Donahue, 2002), including the molecular characterization of an increasing catalogue of organismic structures (Jaramillo & Kramer, 2004), it can be argued equally well that stomata first appeared as structures facilitating drying out and only subsequently assumed a role in gaseous exchange following the evolution of gas-filled intercellular spaces. Equally plausible is a scenario whereby stomata evolved in early land plants as a site of gas exchange and were subsequently co-opted in Sphagnum capsules to facilitate drying out and dehiscence. Without determination of stomatal homology across plant groups and with no commonality between tracheophyte stomatal anatomy and physiology with those of Sphagnum, it would seem that the pairs of cells in the capsule epidermis in Sphagnum are best referred to as pseudostomata. The key to drying out and subsequently capsule dehiscence through increased tension, not pressure, lies in these tiny structures. Although in true mosses, stomata appear to have a clear-cut role in gaseous exchange as they are exposed to the atmosphere, form pores when turgid and are subtended by spongy photosynthetic tissue, it has yet to be determined if they undergo diurnal cycles of opening and closing as in tracheophytes, and evidence for a role of abscisic acid in opening and closing rests in but a single study on Funaria (Garner & Paolillo, 1973; Sack & Paolillo, 1983). So, in peristomate mosses, duality of function cannot be ruled out: gaseous exchange during capsule expansion followed by maturational water loss before dehiscence. The ultimate proof will lie in comparative functional studies, such as those described here, in conjunction with using the rapidly expanding molecular tool kit following the publication of the Physcomitrella genome (Rensing et al., 2008). Stomatal genes are now prime candidates for bioinformatic probing across mosses, hornworts and tracheophytes and unveiling of the secrets of millions of years of stomatal evolution (Webb & Baker, 2002), one of the key features of plant survival on land, now lie within our grasp. Whatever the postgenomic future holds, the present functional study has shown that a description of a biological process that has gone unchallenged and permeated the literature, and been taught to thousands of botanists for more than a century, was ultimately a myth that could be overturned by simple experimentation.
Steve Schmitt and Dee Gates in the Imaging Centre at Southern Illinois University for skilled technical assistance. S.P. acknowledges the financial support from a Leverhulme Trust Early Career Fellowship.