Stomata: the holey grail of plant evolution

Abstract

The greatest cost associated with terrestrial photosynthesis is maintaining hydration in the presence of phenomenal evaporative forces from the atmosphere (Wong et al., 1979). Without the capacity to maintain internal water reserves, vascular plants (tracheophytes) would never have escaped the soil boundary layer (Raven, 1977). Two key adaptations enable homoiohydry in vascular land plants: (1) a means to rapidly conduct water over long distances via xylem and (2) the ability to regulate water use by stomata (Raven, 1977). Xylem alone has long been credited for the evolutionary success of tracheophytes. Trees are only found in this clade, with most “nonvascular” land plants (bryophytes) confined to the soil boundary layer and relying on vegetative desiccation tolerance to survive drought (Proctor et al., 2007). In contrast, stomata, which predate xylem in the fossil record and are found in most extant land plant clades, are often relegated to a level of lesser importance for driving the evolution of homoiohydric land plants (Raven, 2002). We would argue that physiological data, particularly from bryophytes, challenge this conventional wisdom rooted in morphological observation and suggest that the evolution of stomatal function was a critical innovation for the evolution of large plants. Across biomes, taller growth maximizes the capture of light, substantially increasing individual productivity. In vascular plants, there is abundant evidence that growing tall confers a selective benefit, from the adaptive advantage of a fast growth rate in forest tree seedlings (Walters and Reich, 2000), convergent evolution of woody growth in all major lineages (Stewart and Rothwell, 1983; Lens et al., 2013), and competition for light explaining the evolution of canopy structures in forests (Falster and Westoby, 2003). A perplexing question then is why bryophytes have remained apparently immune to this competition, with most species growing within the substrate boundary layer and no evidence of extinct bryophyte trees. Today, a majority of bryophyte species are highly adapted to ecological niches devoid of or with minimal competition from vascular plants or indeed any other photosynthetic organism (Shaw and Renzaglia, 2004), making the idea of them competing with vascular plants for light moot. This argument is somewhat problematic considering that the bryophyte ancestor likely emerged before the appearance of vascular plants (Wellman et al., 2003) (although there are no fossils of bryophytes in the Rhynie Chert, one of the oldest known macrofossil land plant assemblages), yet did not evolve to fill the ecological niche rapidly occupied by early vascular plants. Reproductive limitations may play a role, with dominant bryophyte gametophytes relying on liquid water to transport motile spermatozoids to the female egg (Glibert, 2000). Yet similar reproductive limitations have been overcome in vascular plants (e.g., the evolution of pollen), and in bryophytes spermatozoids can travel vast distances in water (Pressel and Duckett, 2019) or by insect dispersal (Gibson and Miller-Brown, 1927). Furthermore, flagellate sperm have not limited the ecological diversity of other vascular plant groups, including ferns (Watkins et al., 2007). While the ecological specialization of most extant bryophytes renders solving the absence of tall bryophytes intractable, the Polytrichales provide a notable exception as the tallest bryophytes (species of Dawsonia, the tallest in this group, reach 0.7 m) (van Zanten, 1973) (Fig. 1). Recent work in these giants of the bryophyte world has found that the internal water-conducting hydroids in the moss Polytrichum, while of completely independent origin, are functionally analogous to xylem, being capable of transporting considerable volumes of water under tension in the vegetative gametophyte (Brodribb et al., 2020; Duckett and Pressel, 2020). This observation raises an intriguing conundrum, why are these tall mosses, that are capable of transporting water though a vascular system, still so small compared to vascular plants? In tracheophytes, stomata are found on the surface of leaves; in mosses, stomata, when present, are confined to the solitary spore capsule in the unbranched sporophyte (Paton and Pearce, 1957). Polytrichum leaves with lamellae only one cell thick means that, while hydraulic conductivity is the same as many tracheophytes, evaporation is poorly regulated. Unregulated transpiration poses no problem under humid conditions; however, when vapor pressure deficit (VPD) increases, the excessive water loss, despite a thick cuticular and wax investiture, results in a negative water potential sufficient to cause embolism, ending water transport (Brodribb et al., 2020). These observations suggest that stomata on leaves were indeed essential for the evolution of homoiohydric land plants, with stomatal closure at high VPD in vascular plants able to reduce significant declines in water potential and thereby prevent embolism (Brodribb et al., 2017). If the greatest limitation to Polytrichum competing with vascular plants is simply a lack of stomata on vegetative organs, then why do the leaves of Polytrichales not have stomata? No extant gametophytes have stomata, yet stomata are found on stems below reproductive structures in both sporophyte and gametophyte generations of the extinct pre-vascular plant Aglaophyton (Edwards et al., 1998), suggesting that the dominant life history stage of bryophytes is not in itself a limitation. We argue that while stomata are superficially similar structurally across all land plants, typically taking the form of two guard cells surrounding a pore; considerable evolution in stomatal function across land plant lineages is the reason why, although some bryophytes have highly elaborate vascular tissue, they do not utilize stomata to regulate leaf water loss. In contrast to tracheophytes that bear stomata on anatomically complex leaves and stems, bryophyte stomata are exclusively located on sporangia and contribute to a coordinated process that results in spore production and dispersal rather than to general assimilation (Renzaglia et al., 2017; Duckett and Pressel, 2018; Kubásek et al., 2021). Among bryophytes, stomata are absent in all extant liverworts (Renzaglia et al., 2007, 2020; Duckett and Pressel, 2018), an observation consistent with the maturation of the sporophyte within gametophyte protective tissue. Stomata on sporangia of mosses and hornworts, in contrast to tracheophytes, play an important role in promoting water loss for spore maturation and release (Merced and Renzaglia, 2013, 2017; Duckett et al., 2009; Pressel et al., 2014; Field et al., 2015; Chater et al., 2016; Renzaglia et al., 2017). Once open, thickened cell walls of mature bryophyte stomata render them physically incapable of closing and thus useless for mitigating excessive water loss. The capsules of bryophytes are relatively short-lived compared to the subtending gametophytes, consequently the selective pressures to maintain water relations during the growing season that drove the evolution of complex stomatal opening and closing capacity and signals in tracheophytes did not play a role in bryophyte diversification. Despite these compelling data, there still remains a pervasive alternative view that when stomata first appeared, they were already in possession of the full suite of signaling and molecular operating machinery found in modern angiosperms and thus stomatal function was the same as in modern angiosperms (Chater et al., 2011; Harris et al., 2020). In a recent example, Zhao et al. (2019) claimed that the colonization of land was enabled by an omnipresent chloroplast retrograde signal that closes all stomata during water stress. This paper is similar in conclusion to a body of literature dating back more than a decade professing that all stomata respond to the hormone abscisic acid (ABA) (Chater et al., 2011; Ruszala et al., 2011; Cai et al., 2017). Levels of this hormone increase in angiosperms when water status declines, triggering a signaling cascade that actively closes stomata (Geiger et al., 2009; Lee et al., 2009; Ma et al., 2009; Park et al., 2009; McAdam and Brodribb, 2015). Arguments in support of universal stomatal function across all land plants deserve close scrutiny, as they imply stomata were irrelevant for plant adaptation, diversification or massive ecological transitions over the past 400 million years, and cannot explain why mosses with efficient hydroids such as Polytrichum have not capitalized on them to regulate leaf water loss. Observations of stomatal aperture responses in the moss Sphagnum fallax are central to the theory of Zhao et al. (2019), that a proposed chloroplast retrograde signal, 3′-phosphoadenosine 5′-phosphate (PAP), has closed stomata in response to water deficit for the past 500 million years. These observations are perplexing given that the stomata of Sphagnum species are highly distinct from those of other land plants and have been described as pseudostomata (Duckett et al., 2009; Merced, 2015). Sphagnum pseudostomata lack pores and subtending intercellular air spaces and are covered by a calyptra throughout capsule development (Fig. 2). The guard cells of Sphagnum never separate to form a discrete pore; they simply collapse when cell volume and turgor decline (Fig. 2). Consequently, pseudostomata do not function in the dynamic regulation of gas exchange, because guard cell collapse is irreversible (Duckett et al., 2009; Merced, 2015). Even if PAP drives guard cell rejoining in Sphagnum, then the mechanism must have facilitated guard cell inflation, a converse function to that in the Zhao et al. (2019) model. It should also be noted that water-conducting cells are absent in Sphagnum. While questions might arise surrounding the taxonomic validity of the moss used in the study by Zhao et al. (2019), even if another moss species, such as the most likely candidate Funaria (based on the single stomatal image provided), was used in their study, major differences in stomatal function between bryophytes and angiosperms further preclude any conclusion of universal mechanistic homology (Kubásek et al., 2021). Consistent with a role in sporophyte maturation and desiccation, a function that is antithetic to that of tracheophyte stomata, hornwort and moss stomata including those of Funaria (Fig. 2F), open and become locked in that state due to guard cell wall chemistry and architecture, preventing subsequent closure (Merced and Renzaglia, 2014, 2019; Renzaglia et al., 2017). Whereas mature stomata in angiosperms are responsive to a variety of environmental and endogenous cues including light intensity, water status, ABA, plasmolysis, and physical damage, those of bryophytes remain unchanged (Duckett and Pressel, 2018; Pressel et al., 2018; Kubásek et al., 2021). Also contrary to functional congruence across land plants are considerable differences in stomatal numbers and sizes in bryophytes that are unrelated to taxonomy, ecology, genome sizes, and atmospheric CO2 levels (Field et al., 2015; Duckett and Pressel, 2018). Indeed, the loss of stomata in two hornwort clades and at least 60 times in mosses indicates that they are essentially disposable in bryophytes unlike their near universality in vascular plants (Renzaglia et al., 2020). In mosses and hornworts, ion changes in the guard cells have been found to occur concurrently with similar ion changes in epidermal cells (Lucas and Renzaglia, 2002; Duckett et al., 2009; Duckett and Pressel, 2018). Consequently, we cannot conclude that the ion flux data presented by Zhao et al. (2019) were guard-cell-specific without epidermal cell controls. Furthermore, the Zhao et al. (2019) model for universal stomatal closure by PAP does not consider evolution in ion channels or their guard cell specificity (Sussmilch et al., 2019a). These evolutionary transitions have occurred in ion channels that play a critical role in angiosperm guard cell movements: such as the absence of outward- and inward-rectifying Shaker potassium channel genes in bryophyte and lycophyte genomes (Gomez-Porras et al., 2012; Sussmilch et al., 2019a; Dreyer et al., 2021), respectively, and major differences in the activation of S-type anion channels across tracheophytes (McAdam et al., 2016). Importantly, it is yet to be shown if chloroplast signals specifically change guard cell gene expression outside of angiosperms. While the behavior of bryophyte stomata is undoubtedly divergent from the behavior of angiosperm stomata, it has recently been suggested that the ancestor of land plants possessed stomata that functioned like those of the annual, angiosperm herb Arabidopsis and that bryophyte stomatal function is highly derived (Harris et al., 2020; Rich and Delaux, 2020). We would argue that evolution of stomatal responses across tracheophyte lineages challenges this view as well as the concept of a universal stomatal closure model by PAP. The in situ stomata of lycophytes and ferns respond to changes in leaf water status as highly predictable passive-hydraulic valves (Brodribb and McAdam, 2011). The stomata of angiosperms do not respond in this way (Buckley, 2019). Contrary to some reports that extremely high levels of exogenous ABA slightly reduce stomatal aperture in some fern and lycophyte species (Ruszala et al., 2011; Cai et al., 2017; Hõrak et al., 2017), there is no evidence that endogenous ABA produced by a plant during drought, or any other endogenous metabolic signal like PAP, drives functional stomatal closure under drought stress in species from these lineages (Brodribb and McAdam, 2011; McAdam and Brodribb, 2012; Cardoso et al., 2019, 2020). These results suggest that the stomata of the ancestor of vascular land plants responded to leaf water status as passive hydraulic valves and that the evolution of a functional stomatal response to ABA (driven by evolution in the interaction of key signaling proteins; Sussmilch et al., 2019b) arose in the common ancestor of the seeds plants and was instrumental in the evolutionary success of this lineage of plants. Retrograde signaling may be ancient, but like most plant hormone signaling pathways, neofunctionalization, diversity, and cell specificity (e.g., action in guard cells) are likely to have evolved gradually through time (Sussmilch et al., 2019b; Blázquez et al., 2020; Cannell et al., 2020; McAdam and Sussmilch, 2021), not in a single event 500 million years ago. The importance of PAP signals in regulating Arabidopsis stomatal response to water stress was established using mutants (Pornsiriwong et al., 2017); based on current data, it is far from parsimonious to conclude that this signal closes the stomata of all land plants. Nevertheless, this recent work highlights the critical need to study how diversity in stomatal function has influenced the macroevolution of land plant lineages. This future endeavor is indeed critical because there is evidence that evolution in these simple structures was instrumental in the evolution of homoiohydry and tall stature (Brodribb et al., 2020), anatomical adaptations that enable survival during drought (Cardoso et al., 2020), facilitation of spore dispersal (Chater et al., 2016; Duckett and Pressel, 2018), the ability of trees to survive in seasonally dry environments (Brodribb et al., 2014) and leaves to attain high rates of photosynthesis (Rockwell and Holbrook, 2017). Furthermore, differences in stomatal function underlie differences in ecological strategies across tracheophytes, particularly with regards to light environment (Doi et al., 2015) and soil water availability (Martínez-Vilalta and Garcia-Forner, 2017). While it is an impactful claim to state a single signal has ruled stomata for all of time (Zhao et al., 2019) or that Arabidopsis physiological function reflects a land plant ancestral state (Rich and Delaux, 2020), such approaches to physiological evolution will never reveal why, for instance, with very similar xylem physiology (Brodribb et al., 2020) and a selective pressure to grow tall (McNickle et al., 2016), Polytrichum does not overtop Sequoia. We thank the two reviewers and Jarmila Pittermann for helpful comments. F.S. acknowledges funding from the Australian Research Council (DE200101133). R.H. acknowledges support from the German Research Foundation (DFG) for the project ‘Evolution of molecular mechanisms that control stomatal closure’ (HE1640/40-1). S.A.M.M., J.G.D., and F.C.S. drafted the manuscript; all authors contributed ideas and revisions to the text; J.G.D., K.S.R., S.P., and A.M. contributed images for figures.