Research Report

Massive Tandem Proliferation of ELIPs SupportsConvergent Evolution of Desiccation Tolerance acrossLand Plants1[OPEN]

Robert VanBuren,a,b,2,3 Jeremy Pardo,b,c Ching Man Wai,a,b Sterling Evans,a and Dorothea Bartelsd

aDepartment of Horticulture, Michigan State University, East Lansing, Michigan 48824bPlant Resilience Institute, Michigan State University, East Lansing, Michigan 48824cDepartment of Plant Biology, Michigan State University, East Lansing, Michigan 48824dIMBIO, University of Bonn, Kirschallee 1, D-53115 Bonn, Germany

ORCID IDs: 0000-0003-2133-2760 (R.V.); 0000-0003-3419-095X (J.P.); 0000-0002-2913-5721 (C.M.W.); 0000-0001-6375-078X (S.E.);0000-0003-0505-3083 (D.B.).

Desiccation tolerance was a critical adaptation for the colonization of land by early nonvascular plants. Resurrection plants havemaintained or rewired these ancestral protective mechanisms, and desiccation-tolerant species are dispersed across the landplant phylogeny. Although common physiological, biochemical, and molecular signatures are observed across resurrection plantlineages, features underlying the recurrent evolution of desiccation tolerance are unknown. Here we used a comparativeapproach to identify patterns of genome evolution and gene duplication associated with desiccation tolerance. We identifieda single gene family with dramatic expansion in all sequenced resurrection plant genomes and no expansion in desiccation-sensitive species. This gene family of early light-induced proteins (ELIPs) expanded in resurrection plants convergent throughrepeated tandem gene duplication. ELIPs are universally highly expressed during desiccation in all surveyed resurrection plantsand may play a role in protecting against photooxidative damage of the photosynthetic apparatus during prolongeddehydration. Photosynthesis is particularly sensitive to dehydration, and the increased abundance of ELIPs may helpfacilitate the rapid recovery observed for most resurrection plants. Together, these observations support convergent evolutionof desiccation tolerance in land plants through tandem gene duplication.

The ability to survive near complete anhydrobiosis isubiquitous, and desiccation tolerance is observed inprokaryotes, protists, fungi, plants, and even animals.The origin of desiccation tolerance in plants is un-known, but its prevalence throughout land plants andgreen algae suggests it is ancestral. Desiccation toler-ance was essential during the transition from water toland, and these ancestral protective mechanisms areretained in some bryophytes, lycophytes, and ferns(Proctor, 1990; Oliver et al., 2005; Porembski, 2011).Early vegetative desiccation tolerance mechanisms

likely formed the basis of pollen and seed desiccationpathways in angiosperms, because core abscisic acid(ABA)–regulated pathways have conserved functionsin seed and nonseed plants (Eklund et al., 2018). Plantswith vegetative desiccation tolerance are broadlytermed ‘resurrection plants’ because of their revivalfrom typically lethal, prolonged dehydration. Vegeta-tive desiccation tolerance in angiosperms arose morerecently, and mounting evidence suggests it evolvedthrough rewiring pathways leading to desiccation tol-erance in seeds (Oliver et al., 2000; Illing et al., 2005;Farrant and Moore, 2011; Costa et al., 2017; VanBurenet al., 2017). Desiccation tolerance evolved indepen-dently in at least 13 lineages of angiosperms, and over300 diverse resurrection plant species have been iden-tified to date across all major plant taxa outside ofgymnosperms (Bewley and Krochko, 1982; Oliver et al.,2000, 2005). This is likely an underestimation, becausemany desiccation-tolerant fern and fern allies areuncharacterized (Porembski, 2011).

The successful induction of desiccation tolerance re-quires the coordinated expression of complex pathwayswith thousands of genes, and the deployment ofphysiological and biochemical safeguards to protectcellular macromolecules and machinery. The photo-synthetic machinery is particularly sensitive to dehy-dration and must be maintained intact or repairedquickly after rehydration (Challabathula et al., 2016).

1This work was supported in part by funding from the NSF | BIO| Division of Molecular and Cellular Biosciences (MCB) (NSF-MCB1817347 to R.V.). S.E. was a participant in the Plant Genomics Re-search Experience for Undergraduates Program funded by NSF |BIO | Division of Biological Infrastructure (NSF-DBI 1358474). Theauthors declare no competing financial interests.

2Author for contact: bobvanburen@gmail.com.3Senior author.The author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Robert VanBuren (bobvanburen@gmail.com).

R.V. designed and conceived the research, and wrote the article; allauthors analyzed the data, and read and approved the finalmanuscript.

[OPEN]Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.18.01420

1040 Plant Physiology�, March 2019, Vol. 179, pp. 1040–1049, www.plantphysiol.org � 2019 American Society of Plant Biologists. All Rights Reserved. www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from

Copyright © 2019 American Society of Plant Biologists. All rights reserved.

Desiccation-associated responses are well character-ized, and a core set are conserved across all resurrectionplants. These include accumulation of osmoprotectants,reactive oxygen species (ROS) scavengers, and late-embryogenesis abundant proteins as well as changesin cell structure and other physical properties (Illinget al., 2005; Zhang and Bartels, 2018). These conservedmolecular signatures suggest desiccation toleranceevolved convergently through rewiring similar, preex-isting pathways across independent lineages (Bewley,1979; Gaff, 1997). This hypothesis is now testable giventhe wealth of genomic resources for desiccation-tolerant species. The genomes of several model resur-rection plants have been sequenced including Boeahygrometrica (Xiao et al., 2015), Oropetium thomaeum(VanBuren et al., 2015), Xerophyta viscosa (Costa et al.,2017), Lindernia brevidens (VanBuren et al., 2018b), Se-laginella lepidophylla (VanBuren et al., 2018c), and Se-laginella tamariscina (Xu et al., 2018). Genomes are alsoavailable for the ABA-inducible desiccation-tolerantbryophytes Physcomitrella patens (Rensing et al., 2008)and Marchantia polymorpha (Bowman et al., 2017). De-spite abundant genomic resources, large-scale com-parisons of desiccation-related genes and pathwaysbetween these species are limited.Gene duplications drive innovation and facilitate

rapid adaptation to abiotic stresses (Hanada et al.,2008). Tandem gene duplicates in O. thomaeum(VanBuren et al., 2017), S. lepidophylla (VanBuren et al.,2018c), and X. viscosa (Costa et al., 2017) are enriched infunctions related to dehydration stress and desiccation.Although similar enrichment patterns are observedacross these taxa, overlap of duplicated orthologousgenes has not been analyzed between independentresurrection plant lineages. Here, we identified patternsof gene duplication and gene family expansion associ-ated with the evolution of desiccation tolerance. Weidentified a single gene family uniquely expanded in allsequenced resurrection plants. This gene family ex-pansion is driven primarily by tandem duplication, andgenes in this family are universally highly expressed inall surveyed resurrection plants. Similar duplicationpatterns despite ;475 million years of separation(Smith et al., 2010) support convergent evolution ofdesiccation tolerance.

RESULTS

Patterns of Gene Family Dynamics in ResurrectionPlant Genomes

The evolution of desiccation tolerance may occurthrough recurrent duplication of shared genes andpathways across independent lineages of resurrectionplants. Duplication of these common features couldincrease the abundance of important osmoprotectantsand end point metabolites, or facilitate high-levelpathway rewiring through regulatory neo or sub-functionalization. To test whether conserved gene

duplication events are contributing to desiccation tol-erance, we surveyed changes in gene family composi-tion across diverse desiccation-tolerant and desiccation-sensitive (DS) land plant genomes. Orthogroups wereidentified using OrthoFinder (Emms and Kelly, 2015)for nine monocot, six eudicot, two bryophyte, and twolycophyte genomes (Supplemental Table S1). Threedesiccation-tolerant angiosperms (L. brevidens, O. tho-maeum, and X. viscosa) and two bryophytes with ABA-induced desiccation tolerance (P. patens and M. poly-morpha) were included for analysis. In total, 26,406orthogroups were identified across the 19 species and456,776 (78.8%) of the input genes were assigned toorthogroups. Of these, 4,625 orthogroups had repre-sentatives from all 19 species. This subset of conservedorthogroups was used to identify gene families thatwere expanded in all resurrection plant genomes butshowed no expansion in DS plants. Three orthogroupshave expanded in all resurrection plants comparedwith DS species: OG0212, OG1697, and OG2512. TheArabidopsis (Arabidopsis thaliana) orthologs in OG2512include the Gly-rich domain protein AtGRDP1, whichhas a role in seed germination and ABA signaling(Rodríguez-Hernández et al., 2014). The AtGRDP1null mutant has higher sensitivity to salt and osmoticstress, and overexpression increases stress tolerance(Rodríguez-Hernández et al., 2014). OG1697 containsorthologs related to the rice (Oryza sativa) domesticationgene SH4, which is involved in seed shattering anddevelopment of the seed abscission layer (Li et al.,2006). OG0212 contains the early light-induced pro-teins (ELIPs), which have a well-defined role in pro-tection against photooxidative damage under high lightconditions (Hutin et al., 2003). Together, this supportsgene duplication-mediated rewiring of seed-relatedprocesses in resurrection plants, consistent with previ-ous hypotheses (Bewley, 1979; Illing et al., 2005; Costaet al., 2017; VanBuren et al., 2017). P. patens and M.polymorpha require exogenous ABA to induce desicca-tion tolerance and are not classified as true resurrectionplants. Excluding these two lineages from enrichmenttests identified a single expanded orthogroup (OG0212;ELIPs) common to all surveyed resurrection plants. Thenumber of ELIPs in this orthogroup ranged from 10 to26 for resurrection plants and from 1 to 8 for DS line-ages. Gene family dynamics, evolutionary history, andexpression patterns of ELIPs were further characterizedacross land plants and green algae.

ELIP Gene Family Dynamics across Land Plants andGreen Algae

We expanded the analysis of ELIP composition to 75sequenced land plant and Chlorophyta genomes to testwhether the dramatic expansion we observed is uni-versally unique to desiccation-tolerant species. ELIPswere identified using a homology-based approach withBLAST and classified as singletons or tandem dupli-cates based on their physical proximity in the genomes

Plant Physiol. Vol. 179, 2019 1041

VanBuren et al.

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

(see Materials and Methods). All DS land plants haveless than ten ELIPs with an average of 3.1 per genome,compared with an average of 20.7 in desiccation-tolerant species (Fig. 1). DS monocots tend to havemore ELIPs than DS eudicots (avg. 3.7 vs 3.0), but thedifference is not significant (Wilcoxon rank-sum,P = 0.09). DS grasses (family Poaceae) have an enrich-ment of ELIPs compared with all angiosperms, with anaverage of 5.0 per genome (Wilcoxon rank-sum, P ,0.05). The only Charophyte genome used in this study(Chara braunii) has one ELIP. Chlorophyta have con-siderably more ELIPs than DS land plants (avg. 7.5) butfewer than resurrection plants. The high copy number ofELIPs may help combat rapid changes in light intensityand the increased ultraviolet radiation-mediated photo-bleaching that green algae experience in dynamic aquaticor exposed environments.

Expansion of ELIPs was previously reported in thegenomes of Boea hydrometrica (Xiao et al., 2015), S. lep-idophylla (VanBuren et al., 2018c), and L. brevidens(VanBuren et al., 2018b). No such enrichment of ELIPswas reported in the resurrection plants O. thomaeum(VanBuren et al., 2015), X. viscosa (Costa et al., 2017), orS. tamariscina (Xu et al., 2018), suggesting a role forELIPs in desiccation tolerance is not universal. The O.thomaeum genome was recently updated and reas-sembled, and the O. thomaeum V2 assembly has 22newly annotated ELIPs (VanBuren et al., 2018a). Onlyone ELIP was reported in S. tamariscina (Xu et al., 2018),but BLAST against the genome identified 74 completeor partial unannotated ELIPs (Supplemental Table S2).Because these ELIPs were not annotated in the S. tam-ariscina genome, this species was excluded from theexpression analysis. Numerous clusters of desiccation-associated genes were described in the X. viscosa ge-nome (Costa et al., 2017), but the ten annotated ELIPswere not classified as clusters of desiccation-associatedgenes. This is surprising given the tandemly duplicatednature of the ELIPs, and their high expression duringdesiccation in X. viscosa.

Gene duplications can arise in mass during wholegenome duplication events, or at the single-gene levelthrough tandem or retrotransposon-mediated dupli-cation. We further characterized the origin of ELIPsto identify the mechanism driving the large-scale du-plication observed in resurrection plants. Tandemlyduplicated ELIPs were identified using a syntenic ap-proach based on their physical proximity within thegenome. Most ELIPS in DS species are dispersed ran-domly across the genome as singletons (Fig. 1). AllELIPs from plants within the Brassicales are singletons,and species from other angiosperm orders have amix ofsingleton and tandem gene copies. In contrast, mostELIPs in resurrection plants (74%; 111 out of 149) arefound in large tandem arrays. All ELIPs inO. thomaeumand X. viscosa are tandemly duplicated, and other res-urrection plants have a mix of singletons and tandemduplicates. B. hydrometrica has the lowest number oftandemly duplicated ELIPs among resurrection plants,but this may be an artifact of the fragmented nature of

the genome assembly (Xiao et al., 2015). Desiccation-induced accumulation of ELIP transcripts was alsoobserved in the model moss Syntrichia (Zeng et al.,2002; Oliver et al., 2004), but the number of ELIPs inthe genome is unknown.

O. thomaeum has the largest tandem array of ELIPs,with 22 copies on chromosome 8 (Fig. 2). Despite a highdegree of gene-level collinearity between O. thomaeumand the C4 panicoid grasses sorghum (Sorghum bicolor)and Setaria (Setaria italica), ELIPs from the O. thomaeumtandem array have no syntenic orthologs (Fig. 2A). TheELIP array inO. thomaeum is divided into two segmentsof eleven copies split by a 60-kb region spanning ninegenes (Fig. 2, A and B). O. thomaeum genes that flankand bisect the ELIP tandem array are collinear withSorghumand Setaria. SeveralO. thomaeumELIPs containfragments of, or are flanked by, intact retrotransposons(Fig. 2B). These closely associated Ty3-gypsy ret-rotransposons may have facilitated repeated geneduplication.

Expression of ELIP Genes during Desiccationand Rehydration

The repeated duplication of ELIPs across resurrectionplants suggests a conserved role in desiccation toler-ance. To test for desiccation-related expression, wereanalyzed available RNAseq data from O. thomaeum(VanBuren et al., 2017), B. hydrometrica (Xiao et al.,2015), S. lepidophylla (VanBuren et al., 2018c), L. brevi-dens (VanBuren et al., 2018b), X. viscosa (Costa et al.,2017), and the moss Bryum argenteum (Gao et al.,2015). Expression data were processed using the samepipeline for each species, to more accurately comparepatterns across experiments. A genome for B. argenteumis not available; therefore, a set of representative tran-scripts was assembled using the RNAseq data withTrinity (Haas et al., 2013). Eight ELIP transcripts wereidentified in B. argenteum, but thismay not represent thetotal number of ELIPs in the genome.

ELIPs are consistently among the highest expressedgenes in resurrection plants during varying dehydra-tion and rehydration timecourses (Fig. 3). Each of thesurveyed dehydration and rehydration timecourseswere sampled at different relative water contents,different length and severity of water deficit, and dif-ferent times after rehydration. Because of this, webroadly classified time points into three groups to facil-itate cross-species comparisons: well-watered (green),dehydrating/desiccating (yellow), and rehydration(blue). ELIPs have low or undetectable expression un-der well-watered conditions in all surveyed tolerantand sensitive species. ELIPs have the highest expres-sion in desiccated tissues, and expression increasesthroughout the progression of dehydration stress. Thehigh expression of ELIPs is generally maintained dur-ing early rehydration (.24 h), with expression de-creasing as plants return to normal relative watercontent. ELIP down-regulation during rehydration is

1042 Plant Physiol. Vol. 179, 2019

[REMOVED MACROBUTTON FIELD]Desiccation Tolerance Arose via Gene Duplication

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

Figure 1. ELIP composition in sequenced land plant and Chlorophyta genomes. The number of ELIPs are plotted for 74 genomes.Tandemly duplicated ELIPs are plotted in red, and single copy or interspersed ELIPs are plotted in blue. Desiccation-tolerantspecies are highlighted in orange.

Plant Physiol. Vol. 179, 2019 1043

VanBuren et al.

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

reflected by the rate of recovery. B. argenteum is able torecover in ,1 h after rehydration, and ELIP expressionreturns to basal levels in less than 24 h. Recovery in X.viscosa and L. brevidens occurs more slowly, and ELIPsare highly expressed during all of the sampled rehy-dration time points.

Drought-induced expression of ELIPs is not uniqueto resurrection plants, and ELIPs are upregulated underwater deficit in DS species (Fig. 4). Tandem duplicationin resurrection plants may increase the absolute tran-script abundance of ELIPs and improve photoprotectivecapacity. We surveyed the combined expression of ELIPsunder themost desiccated time point for each species andcompared this with drought samples from Arabidopsis(9-d drought, relative water content 60%; Crisp et al.,2017), maize (Zea mays, soil water content 40%), andLindernia subracemosa. ELIPs have, on average, 622-foldhigher expression abundance in resurrection plants thanDS species under water deficit (Fig. 4; Wilcoxon ran sumP = 0.023). O. thomaeum and L. brevidens have the highestabsolute expression of ELIPs, andX. viscosahas the lowestexpression among resurrection plants.

DISCUSSION

Resurrection plants span all major taxa outsideof gymnosperms and independent desiccation-tolerant lineages diverged up to 475 million yearsago (Smith et al., 2010). Despite this ancient diver-gence, all resurrection plants seem to use a set ofconserved molecular mechanisms to combat thedamages associated with desiccation. This phenotypicconvergence could arise through independent co-option of selection-constrained pathways or through

rewiring of nonoverlapping (independent) pathways.Based on genome-scale comparisons of desiccation-tolerant and DS species, we identified a single genefamily that has expanded in all sequenced resurrectionplant genomes. Given the wide taxonomic sampling,we can infer expansion of the ELIP gene family oc-curred independently in each desiccation-tolerant lin-eage. The rapid and dramatic expansion arose throughrepeated tandem gene duplication, and ELIPs arefound in large arrays ranging in size from 10 to 22copies. DS species typically have 1 to 5 ELIPs with highcollinearity across plant genomes. The ELIP tandemarray inO. thomaeum has recently been translocated to adifferent chromosome and is nonsyntenic with orthol-ogous ELIPs in other grass genomes. Several ELIPs inO. thomaeum are nested around intact or fragmentedretrotransposons, providing a source for their duplica-tion and translocation. Taken together, repeated ELIPduplication supports convergent evolution of desicca-tion tolerance across land plants.

Desiccation induces tremendous stress on the mac-romolecules and molecular machinery of the cell. Pho-tosynthesis is inhibited in desiccated tissues, andhomoiochlorophyllous (chlorophyll retaining) resur-rection plants are susceptible to increased photo-inhibition and oxidative damage from unboundchlorophyll or unstable light harvesting complexes.ELIPs are induced under high light stress and protectagainst photooxidative damage via chlorophyll bind-ing and stabilization of photosynthetic complexes(Hutin et al., 2003). ELIPs accumulate in photosynthetictissue under other abiotic stresses including cold,drought, heat, and low nutrients (Bartels et al., 1992;Beator et al., 1992; Levy et al., 1993; Adamska andKloppstech, 1994; Montané et al., 1997), although their

Figure 2. Unique tandem array of ELIPs inO. thomaeum. A, Microsynteny plot of thesyntenic region flanking the ELIP array inO. thomaeum, Sorghum, and Setaria.Genes in the forward orientation areshown in yellow, and blue genes are inreverse orientation. Syntenic orthologsbetween the three genomes are denoted byconnecting gray lines. The region con-taining ELIPs is highlighted in brown. B,Expanded view of the ELIP tandem array inO. thomaeum. ELIPs are shown in blue,and other genes are shown in gray. Longterminal repeat (LTR) retrotransposons areshown in red.

1044 Plant Physiol. Vol. 179, 2019

[REMOVED MACROBUTTON FIELD]Desiccation Tolerance Arose via Gene Duplication

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

role in some of these stresses is unknown. ELIPs arealso induced under other conditions where chlorophyllis degraded such as chromoplast biogenesis in tomato(Solanum lycopersicum), suggesting a broader role ofphotoprotection during thylakoid restructuring (Brunoand Wetzel, 2004). Overexpression of a Medicago trun-catula ELIP inNicotiana benthamiana increased resistanceto freezing, chilling, osmotic stress, and high light(Araújo et al., 2013).ELIPs were among the earliest desiccation-related

proteins identified, and ELIPs have a conserved roleduring desiccation in Craterostigma plantagineum(Bartels et al., 1992; Alamillo and Bartels, 2001) andSyntrichia ruralis (Zeng et al., 2002). ELIPs are alsohighly expressed during desiccation in Haberlea rho-densis (Gechev et al., 2013) and Sporobolus stapfianus(Yobi et al., 2017), but these lineages lack reference ge-nomes; therefore, they were not included in this study.The two ELIPs in Arabidopsis are functionally redun-dant, and overexpression of either is sufficient to rescuethe photosensitivity in the pleiotropic chaos mutant(Hutin et al., 2003). ELIPs in resurrection plants likely

maintain their ancestral photoprotective role, but therepeated tandem duplication increases their relative a-bundance. ELIPs are among the most highly expressedgenes in desiccating leaf tissues, and the proteinsaccumulate in the thylakoid of desiccating C. plan-tagineum (Bartels et al., 1992; Alamillo and Bartels,2001). ELIP accumulation in C. plantagineum washypothesized to bind free chlorophyll and stabilizethe thylakoid to reduce photooxidative damageduring prolonged desiccation. Outside of this singlestudy in C. plantagineum, ELIP abundance and lo-calization has not been surveyed in resurrectionplants, and the link between duplication, transcriptaccumulation, and protein production remains to betested. Overexpression of a C. plantagineum ELIP inM. truncatula increased drought tolerance and re-covery, supporting a potential role in desiccationtolerance (Araújo et al., 2013).Excess light is a pervasive challenge in desiccated leaf

tissue. Poikilochlorophyllous resurrection plants dis-mantle their photosynthetic apparatus during desicca-tion to mitigate light- associated damage. In contrast,

Figure 3. Desiccation-related expression of ELIPs across divergent resurrection plants. Log2 transformed expression ofELIPs is plotted using RNAseq data from six resurrection plants. Colors in violin plots correspond to hydration state:green are well-watered, yellow are dehydrating/desiccated, and blue are rehydrating. Stages are defined in SupplementalTable S3.

Plant Physiol. Vol. 179, 2019 1045

VanBuren et al.

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

homoiochlorophyllous resurrection plants retain thy-lakoids and chlorophyll during desiccation, but recovermore quickly upon rehydration. PSII has negligibleactivity in desiccated tissues of Craterostigma pumilum,but near normal maximum quantum yield of PSII(Fv/Fm) is observed only 18 h after rehydration(Charuvi et al., 2015). Recovery in mosses is even faster,and Fv/Fm reaches two-thirds the well-watered ratioin less than 40 min after rehydration (Proctor andSmirnoff, 2000). Homiochlorophyllous resurrectionplants have considerably more ELIPs (avg. 21) thanpoikilochlorophyllous species, and this is likely reflec-tive of the increased photooxidative damage resultingfrom maintaining the photosynthetic apparatus duringdesiccation. The increased relative abundance of ELIPsin the thylakoid of homiochlorophyllous resurrectionplants may protect and stabilize PSII complexes andfacilitate more rapid neutralization and degradation ofunbound chlorophyll.

X. viscosa has the lowest number of ELIPs (10)among resurrection plants, and it is the only poikilo-chlorophyllous species with a sequenced genome. Thecomparatively low number of ELIPs may reflect thealternative strategy of dismantling photosynthetic ma-chinery then degrading chlorophyll to reduce photo-oxidative damage. High expression of ELIPs in X. viscosais maintained throughout rehydration. This species takesseveral days to fully recover and to resynthesize chloro-phyll; therefore, ELIPs may have a protective role duringchlorophyll synthesis, similar to greening seedlings.Sequences of additional resurrection plant genomes

are needed to further test this link between ELIP copynumber and strategies for mitigating damage fromexcess light.

ELIP duplication supports convergent evolution.ELIPs, together with other abundantly expressed pro-tective proteins such as late-embryogenesis abundantproteins, are likely central components of the desiccationresponse. Desiccation pathways typically associatedwith seeds are induced during vegetative desiccation,suggesting this trait evolved through pathway rewiring(Illing et al., 2005; Farrant and Moore, 2011; Costa et al.,2017; VanBuren et al., 2017). Transcription factors areunder strict dosage constraints, and no conserved du-plications of genes encoding seed-related transcriptionfactors such as ABI3 and ABI5 (Lopez-Molina et al.,2001) were identified in this study. This suggests eitherindependent lineages duplicated and neofunctionalizeddifferent genes to induce seed-associated desiccationpathways, or existing genes were rewired via cis-regulatory elements as previously shown (Giarolaet al., 2018). This remains to be tested but will bepossible with additional genomes and genome-scaleanalysis of cis-elements and transcription factor bind-ing sites. Taken together, we hypothesize rewired-desiccation pathways enabled plants to withstandprolonged drying and the duplication of ELIPs allowedplants to withstand excessive light in the absenceof water.

METHODS

Orthogroup Identification and Enrichment Patterns

Orthologous genes across a subset of 19 land plantspecies were analyzed to identify gene families that areexpanded in desiccation tolerant lineages. The genefamily analysis included nine monocot (Ananas como-sus, Brachypodium distachyon, Oryza sativa, Oropetiumthomaeum, Panicum virgatum, Sorghum bicolor, Setariaitalic, Xerophyta viscosa, Zostera marina), six eudicot(Lindernia brevidens, Lindernia subracemosa, Arabidopsis[Arabidopsis thaliana], Medicago truncatula, Solanumlycopersicum, Vitis vinifera), two bryophyte (Marchantiapolymorpha and Physcomitrella patens), and two lyco-phyte genomes (Selaginella lepidophylla and Selaginellamoellendorffii). This includes three desiccation-tolerantangiosperms and two bryophytes with ABA-induceddesiccation tolerance. The predicted protein sequencesfor each species were first clustered into orthologousgroups using Orthofinder (v2.2.6; Emms and Kelly,2015). Diamond (v0.9.24; Buchfink et al., 2015) wasused to conduct pairwise protein alignments, and allother parameters were set to defaults. The resulting26,406 orthogroups were filtered to include only the4,625 groups with at least one ortholog in all 19 species.For each orthogroup, the following values were calcu-lated: the proportion of genes in that orthogroupamong the 7 desiccation-tolerant plants, the proportionof genes in that orthogroup among the 14 DS plants,

Figure 4. Total expression of ELIPs in desiccation-tolerant and DSspecies. The Log2 transformed, summed TPM of ELIPs from eachspecies are plotted for drought/desiccation time points (yellow) andcomparable well-watered or rehydrated time points (green and blue,respectively).

1046 Plant Physiol. Vol. 179, 2019

[REMOVED MACROBUTTON FIELD]Desiccation Tolerance Arose via Gene Duplication

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

and the combined proportion using the followingformulas:

pdt ¼∑dt species 1

dt spesies i # genes in orthogroup

∑dt species 1dt species i total # of genes in any orthogroup

pds ¼∑ds species 1

ds species i # genes in orthogroup

∑ds species 1ds species i total # of genes in any orthogroup

p ¼∑species 1

species i # genes in orthogroup

∑species 1species i total # of genes in any orthogroup

The two proportions were then compared using az-score calculated as follows:

Z ¼ ðpdt 2 pdsÞffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffipð12 pÞ

�1ndt

þ 1nds

�s

A 1-sided hypothesis test was calculated by com-paring the Z-score to a normal distribution. Theresulting p-value was then adjusted using the Benja-mini and Hochberg procedure (Benjamini andHochberg, 1995) to obtain q-values. Orthogroups witha q-value of less than 0.05 were considered to be sig-nificantly enriched.

Identification of ELIPs

Based on orthogroup enrichment, the analysis ofELIPs was extended to 72 land plant and green algaegenomes downloaded from Phytozome (V12; https://phytozome.jgi.doe.gov/pz/portal.html). ELIPs werealso annotated in the Charophyte Chara braunii(Nishiyama et al., 2018). ELIPs were identified usingBLASTp with the Arabidopsis ELIP1 (AT3G22840.1)as a query and an e-value cutoff of 1e-15 for land plantsand 1e-6 for green algae. ELIPs were further classifiedas singleton or tandemly duplicated based on theirphysical proximity in the genome (,5 genes separat-ing ELIPs).

Expression Analysis

Illumina RNAseq data from O. thomaeum (VanBurenet al., 2017), Boea hydrometrica (Xiao et al., 2015),S. lepidophylla (VanBuren et al., 2018c), L. brevidens(VanBuren et al., 2018b), L. subracemosa (VanBurenet al., 2018b), Xerophyta viscosa (Costa et al., 2017),Arabidopsis (Crisp et al., 2017), maize (PRJNA419326),and the moss Bryum argenteum (Gao et al., 2015) weredownloaded from the National Center for Biotechnol-ogy Information sequence read archive and reanalyzed.Raw Illumina reads were trimmed using Trimmomatic

(v0.33; Bolger et al., 2014) with default parameters toremove adapters and low-quality bases. A summary ofthe RNAseq data used in this study along with therelative water content and designation for each samplecan be found in Supplemental Table S3. A set of rep-resentative transcripts for B. argenteum was assembledusing Trinity (Haas et al., 2013) with the availableRNAseq data (Gao et al., 2015). Expression levels werequantified using Kallisto (v0.44.0; Bray et al., 2016)against the respective set of gene models for each spe-cies and the Trinity-based transcripts for B. argenteum.Parameters for Kallisto were left as default, and 100bootstraps were run per sample. Expression wasquantified in transcript per million (TPM), and a meanacross all of the replicates was used to compare ex-pression of ELIPs in each species. Log2 transformedTPM expression values of ELIPs were plotted. Sampleswere clustered into three groups (well-watered, desic-cating, and rehydrating) to compare expression be-tween species.

Grass Comparative Genomics

Syntenic gene pairs between the O. thomaeum(VanBuren et al., 2015, 2018a), S. italica (Zhang et al.,2012), and Sorghum bicolor (Paterson et al., 2009) ge-nomes were identified using the python version ofMCSCAN toolkit (V1.1; Wang et al., 2012). Gene modelsfrom the three genomes were aligned using LAST, andhits were filtered using default parameters to find thebest syntenic blocks. A microsyntenic dotplot of theregion containing the ELIP tandem array in Oropetiumwas constructed using MCScan.

Accession Numbers

Genomes and gene models were downloaded fromPhytozome (V12; https://phytozome.jgi.doe.gov/pz/portal.html). Illumina RNAseq data were downloadedfrom the National Center for Biotechnology Informa-tion sequence read archive under the following Bio-Projects: O. thomaeum (PRJNA286116), B. hydrometrica(PRJNA277046), S. lepidophylla (PRJNA420971), L. brevi-dens (PRJNA488068), L. subracemosa (PRJNA488068), X.viscosa (PRJNA295811), Arabidopsis (PRJNA391262),maize (PRJNA419326), and the moss B. argenteum(PRJNA272646).

Supplemental Data

The following supplemental materials are available.

Supplemental Table S1. Summary of orthologous gene family clustering.

Supplemental Table S2. Annotation of ELIPs in the Selaginella tamariscinagenome.

Supplemental Table S3. Summary of expression and physiology data usedin this study.

Received November 13, 2018; accepted December 21, 2018; published January2, 2019.

Plant Physiol. Vol. 179, 2019 1047

VanBuren et al.

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

LITERATURE CITED

Adamska I, Kloppstech K (1994) Low temperature increases the abun-dance of early light-inducible transcript under light stress conditions.J Biol Chem 269: 30221–30226

Alamillo JM, Bartels D (2001) Effects of desiccation on photosynthesispigments and the ELIP-like dsp 22 protein complexes in the resurrectionplant Craterostigma plantagineum. Plant Sci 160: 1161–1170

Araújo S, Duque A, Silva J, Santos D, Silva A, Fevereiro P (2013) Waterdeficit and recovery response of Medicago truncatula plants expressingthe ELIP-like DSP22. Biol Plant 57: 159–163

Bartels D, Hanke C, Schneider K, Michel D, Salamini F (1992) Adesiccation-related Elip-like gene from the resurrection plant Cra-terostigma plantagineum is regulated by light and ABA. EMBO J 11:2771–2778

Beator J, Pötter E, Kloppstech K (1992) The effect of heat shock on mor-phogenesis in barley: Coordinated circadian regulation of mRNA levelsfor light-regulated genes and of the capacity for accumulation of chlo-rophyll protein complexes. Plant Physiol 100: 1780–1786

Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: Apractical and powerful approach to multiple testing. J R Stat Soc Series BStat Methodol 57: 289–300

Bewley JD (1979) Physiological aspects of desiccation tolerance. Ann RevPlant Physiol 30: 195–238

Bewley J, Krochko J (1982) Desiccation-tolerance. In O Lange, P Nobel, COsmond, H Ziegler, eds, Physiological Plant Ecology II. Encyclopedia ofPlant Physiology (New Series), vol 12 / B. Springer, Berlin, Heidelberg,pp 325–378

Bolger AM, Lohse M, Usadel B (2014) Trimmomatic: A flexible trimmer forIllumina sequence data. Bioinformatics 30: 2114–2120

Bowman JL, Kohchi T, Yamato KT, Jenkins J, Shu S, Ishizaki K, YamaokaS, Nishihama R, Nakamura Y, Berger F et al (2017). Insights into landplant evolution garnered from the Marchantia polymorpha genome. Cell171: 287–304. e215.

Bray NL, Pimentel H, Melsted P, Pachter L (2016) Near-optimal probabi-listic RNA-seq quantification. Nat Biotechnol 34: 525–527

Bruno AK, Wetzel CM (2004) The early light-inducible protein (ELIP) geneis expressed during the chloroplast-to-chromoplast transition in ripen-ing tomato fruit. J Exp Bot 55: 2541–2548

Buchfink B, Xie C, Huson DH (2015) Fast and sensitive protein alignmentusing DIAMOND. Nat Methods 12: 59–60

Challabathula D, Puthur JT, Bartels D (2016) Surviving metabolic arrest:Photosynthesis during desiccation and rehydration in resurrectionplants. Ann N Y Acad Sci 1365: 89–99

Charuvi D, Nevo R, Shimoni E, Naveh L, Zia A, Adam Z, Farrant JM,Kirchhoff H, Reich Z (2015). Photoprotection conferred by changes inphotosynthetic protein levels and organization during dehydration of ahomoiochlorophyllous resurrection plant. Plant Physiol 167:1554-65

Costa MD, Artur MA, Maia J, Jonkheer E, Derks MF, Nijveen H,Williams B, Mundree SG, Jiménez-Gómez JM, Hesselink T, et al(2017) A footprint of desiccation tolerance in the genome of Xerophytaviscosa. Nat Plants 3: 17038

Crisp PA, Ganguly DR, Smith AB, Murray KD, Estavillo GM, Searle I,Ford E, Bogdanovic O, Lister R, Borevitz JO, et al (2017) Rapid re-covery gene downregulation during excess-light stress and recovery inArabidopsis. Plant Cell 29: 1836–1863

Eklund DM, Kanei M, Flores-Sandoval E, Ishizaki K, Nishihama R,Kohchi T, Lagercrantz U, Bhalerao RP, Sakata Y, Bowman JL (2018) Anevolutionarily conserved abscisic acid signaling pathway regulatesdormancy in the liverwort Marchantia polymorpha. Curr Biol 28:3691–3699.e3

Emms DM, Kelly S (2015) OrthoFinder: Solving fundamental biases inwhole genome comparisons dramatically improves orthogroup infer-ence accuracy. Genome Biol 16: 157

Farrant JM, Moore JP (2011) Programming desiccation-tolerance: Fromplants to seeds to resurrection plants. Curr Opin Plant Biol 14: 340–345

Gaff D (1997) Mechanisms of desiccation tolerance in resurrection vascularplants. In Mechanisms of environmental stress resistance in plants.Harwood Academic Publishers, Amsterdam, Netherlands, pp 43–58

Gao B, Zhang D, Li X, Yang H, Zhang Y, Wood AJ (2015) De novo tran-scriptome characterization and gene expression profiling of the desic-cation tolerant moss Bryum argenteum following rehydration. BMCGenomics 16: 416

Gechev TS, Benina M, Obata T, Tohge T, Sujeeth N, Minkov I, Hille J,Temanni M-R, Marriott AS, Bergström E, et al (2013) Molecularmechanisms of desiccation tolerance in the resurrection glacial relicHaberlea rhodopensis. Cell Mol Life Sci 70: 689–709

Giarola V, Jung NU, Singh A, Satpathy P, Bartels D (2018) Analysis ofpcC13-62 promoters predicts a link between cis-element variations anddesiccation tolerance in Linderniaceae. J Exp Bot 69: 3773–3784

Haas BJ, Papanicolaou A, Yassour M, Grabherr M, Blood PD, Bowden J,Couger MB, Eccles D, Li B, Lieber M, et al (2013) De novo transcriptsequence reconstruction from RNA-seq using the Trinity platform forreference generation and analysis. Nat Protoc 8: 1494–1512

Hanada K, Zou C, Lehti-Shiu MD, Shinozaki K, Shiu S-H (2008) Impor-tance of lineage-specific expansion of plant tandem duplicates in theadaptive response to environmental stimuli. Plant Physiol 148: 993–1003

Hutin C, Nussaume L, Moise N, Moya I, Kloppstech K, Havaux M (2003)Early light-induced proteins protect Arabidopsis from photooxidativestress. Proc Natl Acad Sci USA 100: 4921–4926

Illing N, Denby KJ, Collett H, Shen A, Farrant JM (2005) The signature ofseeds in resurrection plants: A molecular and physiological comparisonof desiccation tolerance in seeds and vegetative tissues. Integr CompBiol 45: 771–787

Levy H, Tal T, Shaish A, Zamir A (1993) Cbr, an algal homolog of plantearly light-induced proteins, is a putative zeaxanthin binding protein.J Biol Chem 268: 20892–20896

Li C, Zhou A, Sang T (2006) Rice domestication by reducing shattering.Science 311: 1936–1939

Lopez-Molina L, Mongrand S, Chua N-H (2001) A postgermination de-velopmental arrest checkpoint is mediated by abscisic acid and requiresthe ABI5 transcription factor in Arabidopsis. Proc Natl Acad Sci USA 98:4782–4787

Montané M-H, Dreyer S, Triantaphylides C, Kloppstech K (1997) Earlylight-inducible proteins during long-term acclimation of barley to pho-tooxidative stress caused by light and cold: High level of accumulationby posttranscriptional regulation. Planta 202: 293–302

Nishiyama, T, Sakayama H, de Vries J, Buschmann H, Saint-Marcoux D,Ullrich, KK, Haas FB, Vanderstraeten L, Becker D., Lang, D. (2018).The chara genome: Secondary complexity and implications for plantterrestrialization. Cell 174: 448–464.e424.

Oliver MJ, Tuba Z, Mishler BD (2000) The evolution of vegetative desic-cation tolerance in land plants. Plant Ecol 151: 85–100

Oliver MJ, Dowd SE, Zaragoza J, Mauget SA, Payton PR (2004) The re-hydration transcriptome of the desiccation-tolerant bryophyte Tortularuralis: Transcript classification and analysis. BMC Genomics 5: 89

Oliver MJ, Velten J, Mishler BD (2005) Desiccation tolerance in bryo-phytes: A reflection of the primitive strategy for plant survival in de-hydrating habitats? Integr Comp Biol 45: 788–799

Paterson AH, Bowers JE, Bruggmann R, Dubchak I, Grimwood J,Gundlach H, Haberer G, Hellsten U, Mitros T, Poliakov A, et al (2009)The Sorghum bicolor genome and the diversification of grasses. Nature457: 551–556

Porembski S (2011) Evolution, diversity, and habitats of poikilohydrousvascular plants. In Plant Desiccation Tolerance. Springer, pp 139–156

Proctor M (1990) The physiological basis of bryophyte production. BotJ Linn Soc 104: 61–77

Proctor MC, Smirnoff N (2000) Rapid recovery of photosystems on re-wetting desiccation-tolerant mosses: Chlorophyll fluorescence and in-hibitor experiments. J Exp Bot 51: 1695–1704

Rensing SA, Lang D, Zimmer AD, Terry A, Salamov A, Shapiro H,Nishiyama T, Perroud P-F, Lindquist EA, Kamisugi Y, et al (2008) ThePhyscomitrella genome reveals evolutionary insights into the conquest ofland by plants. Science 319: 64–69

Rodríguez-Hernández AA, Ortega-Amaro MA, Delgado-Sánchez P,Salinas J, Jiménez-Bremont JF (2014) AtGRDP1 gene encoding aglycine-rich domain protein is involved in germination and responds toABA signalling. Plant Mol Biol Report 32: 1187–1202

Smith SA, Beaulieu JM, Donoghue MJ (2010) An uncorrelated relaxed-clock analysis suggests an earlier origin for flowering plants. Proc NatlAcad Sci USA 107: 5897–5902

VanBuren R, Bryant D, Edger PP, Tang H, Burgess D, Challabathula D,Spittle K, Hall R, Gu J, Lyons E, et al (2015) Single-molecule sequencingof the desiccation-tolerant grass Oropetium thomaeum. Nature 527:508–511

1048 Plant Physiol. Vol. 179, 2019

[REMOVED MACROBUTTON FIELD]Desiccation Tolerance Arose via Gene Duplication

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.

VanBuren R, Wai CM, Zhang Q, Song X, Edger PP, Bryant D, Michael TP,Mockler TC, Bartels D (2017) Seed desiccation mechanisms co-opted forvegetative desiccation in the resurrection grass Oropetium thomaeum.Plant Cell Environ 40: 2292–2306

VanBuren R, Wai CM, Keilwagen J, Pardo J (2018a) A chromosome scaleassembly of the model desiccation tolerant grass Oropetium thomaeum.Plant Direct. 2: 1–9

VanBuren R, Wai CM, Pardo J, Giarola V, Ambrosini S, Song X, BartelsD. (2018b). Desiccation tolerance evolved through gene duplication andnetwork rewiring in Lindernia. Plant Cell

VanBuren R, Wai CM, Ou S, Pardo J, Bryant D, Jiang N, Mockler TC,Edger P, Michael TP (2018c) Extreme haplotype variation in thedesiccation-tolerant clubmoss Selaginella lepidophylla. Nat Commun 9: 13

Wang Y, TangH, Debarry JD, Tan X, Li J, Wang X, Lee TH, Jin H,Marler B, GuoH, et al (2012) MCScanX: A toolkit for detection and evolutionary analysis ofgene synteny and collinearity. Nucleic Acids Res 40: e49–e49

Xiao L, Yang G, Zhang L, Yang X, Zhao S, Ji Z, Zhou Q, Hu M, Wang Y,Chen M, et al (2015) The resurrection genome of Boea hygrometrica: Ablueprint for survival of dehydration. Proc Natl Acad Sci USA 112:5833–5837

Xu Z, Xin T, Bartels D, Li Y, Gu W, Yao H, Liu S, Yu H, Pu X, Zhou J, et al(2018) Genome analysis of the ancient tracheophyte Selaginella tamar-iscina reveals evolutionary features relevant to the acquisition of desic-cation tolerance. Mol Plant 11: 983–994

Yobi A, Schlauch KA, Tillett RL, Yim WC, Espinoza C, Wone BW,Cushman JC, Oliver MJ (2017) Sporobolus stapfianus: Insights into des-iccation tolerance in the resurrection grasses from linking tran-scriptomics to metabolomics. BMC Plant Biol 17: 67

Zeng Q, Chen X, Wood AJ (2002) Two early light-inducible protein (ELIP)cDNAs from the resurrection plant Tortula ruralis are differentially ex-pressed in response to desiccation, rehydration, salinity, and high light.J Exp Bot 53: 1197–1205

Zhang Q, Bartels D (2018) Molecular responses to dehydration and des-iccation in desiccation-tolerant angiosperm plants. J Exp Bot 69:3211–3222

Zhang G, Liu X, Quan Z, Cheng S, Xu X, Pan S, Xie M, Zeng P, Yue Z,Wang W, et al (2012) Genome sequence of foxtail millet (Setaria italica)provides insights into grass evolution and biofuel potential. Nat Bio-technol 30: 549–554

Plant Physiol. Vol. 179, 2019 1049

VanBuren et al.

www.plantphysiol.orgon March 4, 2020 - Published by Downloaded from Copyright © 2019 American Society of Plant Biologists. All rights reserved.