Resurrection plants are a diverse group of angiosperms with a unique ability to tolerate near total desiccation – sometimes for years – appearing to be dead, and then to “spring back to life”, turning green and robust in a few hours.
Unlike “most” plants, resurrection plants are primed to accept drying soil, altering their gene expression and protein and metabolite complements even to the point at which soil moisture is totally depleted.2 Within a few hours or days, they are as dry as the air and prepared to stay that way indefinitely, or to repeat the cycle a dozen times in a year.
Clearly, from the standpoint of natural history, these plants are interesting as a phenomenon. Today, however, they are also recognized as potentially critical models because of their potential as resources for the manipulation of economically important species. The physiology, biochemistry, proteomic, metabolomic and transcriptomic level aspects of this have been very capably reviewed in recent years.1–3 This introduction to the group relies heavily on those reviews.
Based on 8000 years of agricultural experience, dry plants are dead plants; the exceptions are fascinating. A mere century ago, Dinter described the first ‘resurrection’ bush, Myrothamnus flabellifolia, demonstrating that it could go from green, to powdery dry, to green again in very short times (Dinter 1919, cited in 3).
In recent years, the number of known resurrection plants has exploded, from just 10 fifty years ago, to ca. 300 now.4 Even at that, they constitute only about 0.1% of all angiosperms.
Acknowledgements – This introduction was produced by John Cheeseman to whom comments, suggestions, and other words should be address. Of course, it was not accomplished with out help. Many thanks are due to Maria-Cecilia Costa (University of Cape Town), Mel Oliver (USDA/University of Missouri), Don Gaff (Monash University) and Dorothea Bartels (University of Bonn) for their input and assistance in developing this and the resurrection plants species pages.
Resurrection plants are desiccation tolerant
All seed plants, desiccation tolerant or not, have some degree of drought tolerance; they all have a leaf (and stem) cuticle that retards water loss, stomata to regulate evapotranspiration and allow CO2 entry, and xylem for long distance water transport between the soil and leaves. Together, these have led to an evolutionary movement toward drought tolerance through avoidance (homoiohydry). This can be a good technique. and it works in many – but not all – environments.
More drought tolerant plants often also have xeromorphy (thickened, heavily cuticularized leaves), leaves that roll or curl when drying, stomata buried in grooves, very hairy, reflective leaves. Still, they can’t survive dehydration to less than 40% relative water content. This is still very far from being air-dry.
Resurrection plants, in contrast, are truly desiccation tolerant. They occur where this is a prerequisite for existence, i.e. predominantly in shallow soils (e.g. 1 cm) on rocky slabs or shallow depressions in outcroppings in semitropical and tropical regions of Africa, America and Australia (although one, Lindernia brevidans, is native to east African montane rainforests).1 Most extreme are the chasmophytes; they grow in shallow cracks in rocks.
95% of plants are “DT-ready”
The ability of plants to tolerate dehydration requires that they be able to suspend growth and development, surviving in a dry state until water is available again. In fact, this is not a rare capacity; for seeds, it is common, representative of ca. 90% of modern angiosperms and gymnosperms, and it is equally widespread in pollen. Seed desiccation tolerance allows the formation of soil seed banks, and long range and long duration dispersal. In turn, this, along with the desiccation tolerance of pollen, potentiates expansion of a species’ range. Together with seed longevity (lifespan of seeds in dry storage), seed DT is the basis for ex situ conservation using seeds.
This is an exceptionally important point: in fact, nearly all seed plants – angiosperms and gymnosperms alike – have the genetic information for desiccation tolerance! Other than in seeds, vegetative DT is represented in algae, lichens, mosses, liverworts, and ferns as well as in angiosperm resurrection plants.
It has been proposed that DT arose initially with the transition from aquatic to terrestrial life forms. Then, both the probability of experiencing adverse conditions and the survival benefits associated with DT were high.3 However, as plants expanded in the terrestrial habitat and more complex ecosystems arose, a drawback of non-seed DT became evident: desiccation tolerance is expensive and DT plants grow slowly. Shifting and then shutting down metabolism during drying and fully recovering upon rehydration requires large amounts of energy.
Instead, to avoid desiccation, most plants use the strategies of stress avoidance (completing their life cycles in times of low stress) or stress amelioration (development of structures such as deep root systems so that the lack of water apparent at the soil surface is not limiting).
As a result, DT in vascular plants largely disappeared. Today, DT remains common in pteridophytes, but is rare in angiosperms and absent in gymnosperms. In total, it is estimated that only about 1,300 species of vascular plants are characterized by DT in any vegetative tissues, mostly in corms, rhizomes and tubers.
The 300 exceptional angiosperms which are tolerant at the organismal level are, thus, not a taxonomically defined group going back to the time of the first colonization of land. Rather, extreme desiccation tolerance is found in 13 unrelated, mostly herbaceous lineages, implying it has evolved at least 13 times.3 In each case, DT has been regained by repurposing processes otherwise limited to seeds for foliage desiccation tolerance.
Most importantly, there is huge diversity in the life-history and physiological strategies that have evolved along with the DT phenomenon. Resurrection plants occur in both monocots and dicots, and in each case, there are those species which lose their chlorophyll on desiccation (poikilochlorphyllous), and those which retain it (homoiochlorophyllous). Most monocots are in the former group, and most dicots are in the latter.4 There is considerable variation in the tolerance (and sensitivity) in different life stages and tissues. Only one species is known to have leaves that are constitutively desiccation tolerant: Myrothamnus flabellifolia. The most variable genus is probably Borya. The genus includes both summer deciduous and non-deciduous DT species. In some, e.g. Borya constricta, desiccation tolerance is not constitutive but requires about two days to induce during drying. In B. scirpoides tissue tolerance is limited to the bases of young leaves. B. laciniata is likewise only tissue tolerant in the newest leaves (but not just at the bases). Both of those species are summer deciduous. Several non-deciduous species demonstrate tolerance in leaves in a range of developmental positions – in some cases only younger leaves, and in others, to the fully expanded and oldest.3
Desiccation tolerance at the molecular level
Although only a small subset of the full resurrection plant species has been studied in detail, the variations among them highlight the extreme complexity involved. To understand DT at the mechanistic level from the onset of drying to complete dehydration to rehydration and recovery, a molecular approach is clearly desirable, even if most, or even all, relevant processes lack simple explanations. A number of recent reviews have detailed the evolutionary themes and variations of resurrection plants3,5, and comparative behavior at the transcriptome level.6762,5 Over the last few years, rapid progress has been made in genome sequencing, and combining it with comparative transcriptomics. 5–7
As noted by Gaff & Oliver3, DT does not occur in an evolutionary vacuum, nor need it be studied that way; based on molecular evidence to date, desiccation-tolerant life forms and the acquisition of dormancy or a dormancy-like stage accompanied one and another. One goal of molecular-level studies on this point has been the identification of key genes and metabolites associated with both processes.2 In large part, all the machinery associated with DT must be induced; constitutively expressing it would be extremely expensive and serve no purpose. For example, cell walls, membranes, macromolecules of all sorts and cellular components must be protected from the denaturation, mechanical strains and oxidative stresses that accompany dehydration. A large part of this is accomplished by vitrifying the cytosol, i.e. by replacing water with other polar and much less fluid molecules, especially sucrose and Late Embryogenesis Abundant (LEA) proteins.1
To get beyond that true but simplistic statement, one approach is to mine published data sets from different species to identify those genes, or at least those which are shared among organisms.5 While theoretically a good approach, this suffers from the lack of deep sequencing data from experiments with comparable designs and the fact that many of the genes involved seem to be taxon restricted, or orphan, genes.1
A second approach has been to use suspected DT-related genes from resurrection plants in transformation of other model systems. The objective, here, is to transfer a desirable, but usually complex, trait to the target. Success with this approach has been mixed and is limited by the large number of possible single genes to test and by the very large number of networked genes to be coordinated.
Finally, the elucidation of regulatory networks controlling DT and their key genes is a necessary and promising approach whether the goal is to understand DT at the molecular level, or to manipulate other species. Unfortunately, most resurrection plants are polyploid with large genomes and difficult or impossible to transform. With few, very recent, examples, their genomes have not yet been sequenced.
One of the immediate conclusions from transcriptome studies (without genomes) is that different species have very different approaches, but all leading to DT. Craterostigma plantaganeum and Haberlea rhodopensis are good examples of this contrast.2The expression patterns of LEAs in Haberlea include a number that are constitutively expressed in fully hydrated plants, but are induced to higher levels during desiccation. This implies that under “control” conditions, this species is already primed for desiccation tolerance. Transcripts encoding LEAs in Craterostigma, on the other hand, are abundant only in dehydrated leaves. The massive expression of stress-protective transcripts during dehydration and rehydration implies the existence of finely tuned regulatory networks, that may have evolved by selection of promoter cis elements. These, too, are species, or at least genus, specific.
The problem is made more complicated by the fact that as many as 40% of the contigs (but in Xerophyta viscosa, only 5%) from transcriptome sequencing efforts have no matches in protein databases and encode productions of unknown function. Their changing expression during dehydration clearly indicates that they are critical to understanding the mechanisms of desiccation tolerance. This population also includes long non-coding RNAs (ncRNA), which unlike short ncRNA and siRNA, are poorly conserved and regulate gene expression at the epigenetic, transcriptional or post-transcriptional levels by as yet unknown mechanisms. They may control RNA stability as well as protein synthesis rates in the cytosol.2
The future is in genomes
Looking to the future, the most rapid progress in understanding DT will result from studies involving both a high quality genome and transcriptomes representing different stages of the desiccation and recovery process. Three recent papers have highlighted this potential. It is unfortunately not possible in this space to detail all the results and conclusions of those studies, but only to mention them briefly. If you have gotten this far, go read them.
First, Xiao et al. 7 sequenced the 1.7Gb genome of Boea hygrometrica. Boea is homoiochlorophyllous, and the sequencing revealed expansion of Early Light Induced Proteins (ELIPs) and 5S rRNA genes important for protection of the photosynthetic machinery during drying and rehydration. Interestingly, the Boea genome showed no DT-specific genome organizational features, implying that the resurrection capability largely reflects alteration in the control of dehydration-associated genes otherwise associated with seed maturation and drying. The dispersal of dehydration DEGs throughout the genome suggests DT was not acquired recently, but rather, resulted from repurposing of the DT phenotype from seeds to vegetative tissues.
The expansion of the rRNA gene complement was also notable, and the authors suggested it may be associated with increased DNA stability. The more than 1100 rRNAs is 25-50 times the number in the most closely related asterid genomes thus far sequenced (potato and tomato).
With the genome in hand, Xiao et al. then followed the progress of dehydration and rehydration using analysis of GO classifications of the DEGs.
It is particularly interesting to note that as desiccation was approached, mRNA surveillance genes appeared and accumulated, indicating emphasis on cleaning up damage in the transcript pool. At the same time, transcripts for pathways involved in photosynthesis and N metabolism were depleted even though the plants remained green.
Overall, this study demonstrated a central core of genes involved in DT, including those for ABA metabolism and signaling, phospholipid signaling, LEAs, ROS scavengers and ELIPs.
In the second paper, VanBuren et al.6pioneered application of single-molecule sequencing to the study of desiccation responses, generating a high-quality complete genome sequence (as well as chloroplast, mitochondrial genomes) for the homoiochlorophyllous grass, Oropetium thomaeum. This species, in addition to being a resurrection plant, is diploid, and has the smallest known grass genome. The single-molecule technique included both the gene space and intergenic regions (telomeres, centromeres, transposable elements and rRNA clusters), a major advance over other plant genomes. Such information will be essential for understanding the multi-dimensional gene interactions involved in many phenotypic responses, not just those involving desiccation.
Finally, Costa et al.8 produced a high quality genome of Xerophyta viscosa and assessed changes in the transcriptome during dehydration. Although this species is octoploid, the chromosomes are small and the overall haploid genome size is only 295Mb. The fraction of the genes that are TRG (orphans) is also very low for a plant genome, as is the percentage of transposable elements.
As we might now consider “expected”, transcriptional changes during dehydration reflected induction of genes associated with desiccation and maturation in seeds, and especially those controlled by ABA. This species is interesting in that young seedlings are desiccation sensitive, but tolerance can be induced by application of ABA. Unlike in Boea, however, dehydration responsive changes reflected transcript abundance in “clusters of desiccation-associated genes” (CoDAGs). With these, Costa et al. drew attention to LEA proteins, the ABI3 regulon, and the complex processes of autophagy. The locations of parologous genes in the CoDAGs was such that the authors concluded they were not simply different alleles for the same gene.
This study also took the approach of gene network analysis, identifying four “self-organizing maps” (SOMs) involved in metabolic redirection during drying and rehydration. Although ABA signaling is clearly involved, as in all other studies of DT, about 20% of the nodes in the SOMs were not differentially expressed in response to ABA application.
Further analysis of these complex networks, both in X. viscosa and in other plants can be expected to point to central controlling (if not master controlling) genes. If manipulation of drought tolerance in crops is the goal, these may well be the keys to solving the problem.