Comparative transcriptomics reveals that a novel form of phenotypic plasticity evolved via lineage‐specific changes in gene expression

Abstract

Novel forms of phenotypic plasticity may evolve by lineage‐specific changes or by co‐opting mechanisms from more general forms of plasticity. Here, we evaluated whether a novel resource polyphenism in New World spadefoot toads (genus Spea ) evolved by co‐opting mechanisms from an ancestral form of plasticity common in anurans—accelerating larval development rate in response to pond drying. We compared overlap in differentially expressed genes between alternative trophic morphs constituting the polyphenism in Spea versus those found between tadpoles of Old World spadefoot toads (genus Pelobates ) when experiencing different pond‐drying regimes. Specifically, we (1) generated a de novo transcriptome and conducted differential gene expression analysis in Spea multiplicata , (2) utilized existing gene expression data and a recently published transcriptome for Pelobates cultripes when exposed to different drying regimes, and (3) identified unique and overlapping differentially expressed transcripts. We found thousands of differentially expressed genes between S . multiplicata morphs that were involved in major developmental reorganization, but the vast majority of these were not differentially expressed in P . cultripes . Thus, S . multiplicata 's novel polyphenism appears to have arisen primarily through lineage‐specific changes in gene expression and not by co‐opting existing patterns of gene expression involved in pond‐drying plasticity. Therefore, although ancestral stress responses might jump‐start evolutionary innovation, substantial lineage‐specific modification might be needed to refine these responses into more complex forms of plasticity.

Full text

Ecology and Evolution. 2023;13:e10646.  
|
1 of 18
https://doi.org/10.1002/ece3.10646
www.ecolevol.org
Received:25September2023
|
Accepted :9Octob er2023
DOI:10.1002 /ece3.10646
RESEARCH ARTICLE
Comparative transcriptomics reveals that a novel form of
phenotypic plasticity evolved via lineage-specific changes in
gene expression
Andrew J. Isdaner1 | Nicholas A. Levis1,2 | David W. Pfennig1
This is an op en access arti cle under the ter ms of the CreativeCommonsAttribution License, which permits use, distribution and reproduction in any medium,
provide d the original wor k is properly cited.
©2023TheAuthors .Ecolog y and Evoluti on published by Jo hn Wiley & S ons Ltd.
AndrewJ .Isdan erandNi cholasA .Levi sshoul dbeconsi deredj ointfir staut hors.
1Department of Biology, University
ofNorthCarolina,ChapelHill,North
Carolina,USA
2Depar tmentofBiolog y,Indiana
University,Bloomington,Indiana,USA
Correspondence
David W. Pfen nig, De part ment of Biology,
UniversityofNo rthC arolin a,Chap elHill,
NC27599,USA.
Email: dpfennig@unc.edu
Funding information
Nationa lScienceFoundation,Grant/
AwardNumber:DEB-1753865
Abstract
Novel forms of phe notypic plas ticity may evolve by lin eage-specific ch anges or by
co-opting mechanisms from more general forms of plasticity. Here, we evaluated
whether a novel resourcepolyphenism in New World spadefoot toads (genusSpea)
evolved by co-optin g mechanisms from an ance stral form of plas ticity common in
anurans—accelerating larval development rate in response to pond drying. We com-
pared overlap in differentially expressed genes between alternative trophic morphs
constituting the polyphenism in Speaversus those found between tadpoles of Old
Worldspadefoottoads(genusPelobates)whenexperiencingdifferentpond-dryingre-
gimes.Specifically,we(1)generatedadenovotranscriptomeandconducteddifferen-
tial gene expression analysis in Spea multiplicata,(2)utilizedexistinggeneexpression
data and a recently published transcriptome for Pelobates cultripes when exposed to
differentdryingregimes,and(3)identifieduniqueandoverlappingdifferentiallyex-
pressed transcripts. We found thousands of differentially expressed genes between
S. multiplicatamorphsthatwereinvolvedinmajordevelopmentalreorganization,but
thevastmajorityofthesewerenotdifferentiallyexpressedinP. cultripes. Thus, S. mul-
tiplicata'snovelpolyphenismappearstohavearisenprimarilythroughlineage-specific
changes in geneexpression and notbyco-optingexistingpatterns of gene expres-
sioninvolvedinpond-dryingplasticity.Therefore,althoughancestralstressresponses
might jump-start evolutionary innovation, substantial lineage-specific modification
might be needed to refine these responses into more complex forms of plasticity.
KEYWORDS
developmental plasticity, gene expression, novelty, phenotypic plasticity, spadefoot,
transcriptomics
TAXONOMY CLASSIFICATION
Evolutionaryecology,Genetics,Genomics

2 of 18
|
ISDANER et al.
1 | INTRODUC TION
Phenotypic plasticityisan intrinsic propert yoflife(Nijhout, 2003;
Pfennig, 2021a; Sultan, 2021). Indeed ,a ll major groups of or gan-
isms—from bacteria to mammals—can respond to environmental
variation by undergoing reversible or irreversible shifts in some
aspects of theirphenot ype, including(atthe molecular level) gene
expression(reviewedinSultan,2021). Moreover, plasticity is critical
toecology and evolution(Pfennig, 2021b), having been implicated
in mediati ng species i nteracti ons and coexi stence (A grawal, 20 01;
Hend r y, 2016;Hessetal.,2022; Pfennig & Pfennig, 2012; Turcotte
& Levine, 2016); evolutionar y innovation (Levis & P fennig, 2021;
Moczek et al., 2011); speciation and adaptive radiation (Pfennig
et al., 2010; Schneider & Meyer, 2017; Susoy et al., 2016; West-
Eberhard, 1989, 2003; Wund et al., 2008); and macroevolutionary
transitionsinindividuality(Davison&Michod,2021).
Amongthemostspectacularformsofplasticit yarepolyphenisms
(sensuMichener,1961), the occurrence of multiple, discrete environ-
mentally induced phenotypes in a single population. The evolution
ofapolyphenism haslongbeen viewed asacritical phasein major,
lineage-specific innovations (Levis & Pfennig, 2021; Mayr, 1963;
Moczek et a l., 2011; Nij hout, 2003; West-Eberh ard, 1989, 2003).
Polyphenism promotes innovation by facilitating the accumulation
ofcrypticgeneticvariation(Falconer&Mackay,1996;Gianola,1982 ;
Reid & Acker, 2022; Roff, 1996). Cr yptic genetic variation, in turn,
fuelsplasticity-ledevolution,whichoccurswhenselectionpromotes
evolutionary change by acting on quantitative genetic variation
expose d to selecti on by environme ntal changes a nd plastic ity (re-
viewed in Levis & Pfennig, 2021).
In contrast to these well-characterized evolutionary conse-
quencesofpolyphenism,theoriginsofpolyphenismneedtobebet-
terunderstood.Generally,polyphenismsarethoughttoarisewhen
disruptiveselectionactsoncontinuouslyvarying plasticity (a reac-
tion norm) andmolds it into discrete phenotypes(Pfennig,2021a).
However, the developmentaland genetic processes that promote
the evolutionary refinement of ancestral plasticity into an adaptive
polyph enism require g reater expla nation (Levis & Ra gsdale, 2022;
Sommer, 2020). A l eading hypoth esis is that a novel p olyphenism
evolvesbyredeployingexistingdevelopmentalmachinery(Abouheif
& Wray, 2002; Bhardwaj et al., 2020; Hanna & Abouheif, 2022;
Projecto-Garciaetal.,2017; Sommer, 2020;Suzuki&Nijhout,2006).
Forexample,thegeneticanddevelopmentalunderpinnings ofare-
source polyphenism in the nematode, Pristionchus pacificus, partially
overlap with both the mechanisms controlling an ancestral form of
facultative diapause in which larvae develop into an environmen-
tally resistant “dauer”form(Bentoetal.,2010; Casasa et al., 2021;
Ogawaetal.,2009)and with conserved starvation-responsegenes
(Casasaetal.,2021).Thus,theevolutionofapolyphenismmightco-
opt mechanisms underlying ancestral plastic responses to stressful
environmental conditions (for areviewofstress-induced co-option
driving novelty, see Love & Wagner, 2022).
Alongside such shared mechanisms, unique (i.e., lineage-spe-
cific) evolutionarychange also contributes to novel forms(Babonis
et al., 2016;Cabrales-Orona&Délano-Frier,2021; Jasper et al. , 2015;
Johnson, 2018; Khalturin et al., 2009). For exampl e, a novel loco-
motory trait in water striders, Rhagovelia spp., that permitted them
tofillanunoccupiednicheinvolvedlineage-specificmolecularevo-
lution (Sa ntos et al., 2017). Similarly, the resource polyphenism in
diplogastridnematodes mentioned aboveinvolveslineage-specific
evolution ary changes i n key regulatory ge nes (Biddle & Rag sdale,
2020; Ragsdale et al., 2013).
Co-optionandnon-shared,lineage-specificevolutionmostlikely
work together to shape the evolution of complex phenot ypes, in-
cludingthoseassociatedwithpolyphenisms.However,moreworkis
needed to understandbetter the extenttowhichco-optionversus
lineage-s pecific ch anges under lie the evoluti on of novel plast icity.
Moreover, given that plasticity may also facilitate the origins of novel
complex traits(seeabove),such studiespromisetoprovideimport-
antinsightsintothefactorsthatpromoteevolutionaryinnovation.A
first stepinansweringthis question is to identify patterns of gene
expression that are unique to a derived form of plasticity and not
shared withmore general forms of plasticity.Such lineage-specific
expression patterns could suggest either the broad elaboration of
existing forms of plasticit y or the evolution of novel forms of plas-
ticity.Futureinvestigationswouldthenbeneededtodistinguishbe-
tween these two possibilities.
Tobegintoaddressthisneed,wesoughttocharacterizetheex-
tent to which a derived resource polyphenism is mediated at the mo-
lecularlevelbylineage-specificchangesversus mechanismsshared
with ancestral plastic responses. To do so, we evaluated whether
derived and ancestral forms of plasticity overlap in gene expression
patterns. We focused on gene expressionfor three reasons. First,
nearly all forms of plasticit y are underlain by differences in gene
expression(Goldstein&Ehrenreich,2021; Renn & Schumer, 2013).
Second,geneexpression dataprovide abundant information (Price
et al., 2022), which can offer additional insights into underlying
mechani sms. Finally, the grow ing body of tra nscriptomic dat a en-
ablescomparativeapproaches needed to examine lineage-specific
versusco-optedevolution.Indeed,asdescribedbelow,akeyfeature
ofourstudyutilizedexistinggeneexpressiondata.
Ourfocalspecies,theMexicanspadefoottoad,Spea multiplicata,
has evolved a novel form of plasticit y: a lar val resource polyphen-
ism(Ledón-Rettig&Pfennig,2 011; Pfennig, 1992 a; Figure 1a). Spea
tadpoles typically develop into an “omnivore” morph, which eats
detritus, algae, and plankton. However, if theyare exposed to live
prey earl y in life (such as fai ry shrimp o r other tadpo les; Harmon
et al., 2023; Levis et al., 2015; Pfennig, 1990), some individuals
express an alternative “carnivore” morph (Figure 1a). This novel
phenotype—which has evolved only in the genus Spea (Ledón-
Rettig et al., 2008)—developsfaster than the omnivore morph (de
laSerna Buzon et al.,2020; P fennig, 1992a) and appears to be the
analog to developmentally accelerated forms found in other anurans
(Pfennig,199 2b). Moreover, the carnivore morph is thought to have
arisenwhen pre-existing (ancestral withinScaphopodidae) trophic
plasticity was refined by selection into an adaptive phenotype as
part of a polyphenism (reviewed in Levis&Pfennig, 2019). Recent

|
3 of 18
ISDANER et al.
studies have found that this polyphenism entails changes to lipid me-
tabolism, cholesterol and steroid biosynthesis, and peroxisome form
andfunction(Leviset al.,2020, 2021, 2022).Interestingly,manyof
these same processes mediate another, much more common form of
plasticity in anurans: the ability to facultatively accelerate develop-
mentinresponsetoponddrying(Figure 1b).
Ina sh rin ki ngpond, th et adp olesofmanyanura ns pecie sc anfac-
ultatively initiate metamorphosis and thereby escape the stressful
conditions of higher competition and desiccation. Such developmen-
talac celeratio noccursth ro ug houtthean ur anphy lo geny(Figure 1c),
suggesting it is an ancestralform of plasticity.Of relevance to our
study, another research team recently investigated the transcrip-
tomic bases of this plasticity in Pelobates cultripes, a European spade-
foot that is among the closest relatives of Spea(Figure 1c). Notably,
this team found that this plasticity involves changes to lipid metab-
olism,cholesterol,andsteroidbiosynthesis(Liedtkeetal.,2021)—all
of which were implicated in mediating Spea's resource polyphenism
(Levisetal.,2020, 2021, 2022).
Based on this overlap in mechanisms between the two forms of
plasticity, and the fact that the carnivore morph develops faster than
theomnivoremorph,wehypothesizedthatbeingabletoaccelerate
developm ent (an ancest ral form of plas ticity) cont ributed, at le ast
in part, to the evolution of Spea'sresourcepolyphenism(aderived
form of plasticity; Figure 1c).Ifthisresourcepolyphenismdidindeed
evolveusingsharedmechanismsfromthemoreancestralpond-dry-
ing plasticity, we predicted that we would find significant overlap
in differentially expressed genes between these t wo forms of plas-
ticity. To test this p rediction, we us ed comparative tr anscriptomics to
FIGURE 1 Twoformsofplasticit yinanurantadpoles.(a)Speatadpoles(liketheseS. multiplicata) have evolved a resource polyphenism, in
whichtheydevelopintoeitheranomnivoremorph(left)or,iftheyareexposedtoliveprey,adistinctivecarnivoremorph(right).(b)Tadpoles
ofmanyanuranspeciescanalsofacultativelyacceleratedevelopmentinresponsetoponddrying.Here,aS. multiplicata metamorph
escapesadryingpond.(c)Althoughcarnivore-omnivoreplasticityoccursonlyinSpea(familyScaphiopodidae;open circle),development-rate
plasticityhasbeenreportedinatleast11anuranfamilies(filled squares),suggestingitmayhaveprecededcarnivore-omnivoreplasticity
(namesareshownonlyforanuranfamiliesinwhicheitherformofplasticityhasbeenreported).ThisstudyfocusesonSpea multiplicata
(Scaphiopodidae)andPelobates cultripes(Pelobatidae;boldfont).PhylogenyofanuranfamiliesfromAmphibiaWeb(2016); Phylogeny of
spadefoottoadsfromZengetal.(2 014;note:thephylogenyshownhereshowsonlyonespeciesofOldWorldspadefootsinthefamily
Pelobatidae);phylogeneticdistributionofcarnivore-omnivoreplasticityfromLedón-Rettigetal.(2008); phylogenetic distribution of
development-rateplasticityfromRichter-Boixetal.(2011),withadditionalrecordsfromFanetal.(2014),Székelyetal.(2017), and Venturelli
etal.(2022).

4 of 18
|
ISDANER et al.
determine the extent to which gene expression differences between
S. multiplicata carnivores and omnivores overlap with gene expres-
sion responses to pond drying in P. cultripes. We did so by making
use of S. multiplicata carnivores and omnivores generated for a previ-
ouslypublishedtranscriptomicstudy(Levisetal.,2021) and recently
published transcriptomic data from P. cultripes(Liedtkeetal.,2021).
Inthisway,weleveragedexistingdatatoevaluatewhetheranovel
formofphenotypicplasticityevolvedbylineage-specificchangesor
byco-optingmechanismsfromancestralformsofplasticity.
2 | METHODS
2.1 | Acquisition of experimental tadpoles
Spea multiplicata carnivores and omnivores were generated for a
previous ly published tr anscriptomic s tudy (Levis et al ., 2021).For
that stu dy,t hree pairs of M exican spa defoot toads (S. multiplicata)
were collected in amplexus from a newly formed, temporary pond
nearPortalArizona(“PO2-N Pond”)andtransported to the nearby
SouthwesternResearchStationtobreed.Foreachsibship,tadpoles
weredivided into fiveboxesof 80 tadpoles eachandfedfishfood
(10 mgdaily to mimic ponddetritus;P fennig et al., 1991) as well as
live fair y shrimp and live Scaphiopus couchii tadpoles. Competition,
shrimp consumption, and tadpole consumption contribute to the
developmentof carnivores(Leviset al., 2015, 2017; Pfennig, 1990,
1992b), but not all individuals that experience these cues develop
into a carnivore; some remain omnivores even after experiencing
carnivore-inducing conditions (Pfennig, 1990). When the tadpoles
were 10d old, five omnivores and five carnivores per sibship were
randomlysampled,euthanizedwitha0.8%aqueous tricaine meth-
anesulfonate (MS-222) solution, and placed in a microcentrifuge
tube fil led with RNAlate r. These sam ples remained at r oom tem-
peraturefor 24 htoallowRNAlaterpenetrationandthenwere fro-
zenat−20°CuntilbeingshippedtotheUniversityofNorthCarolina
overnight ondry ice. Samples wereheld at−80°Cuntil use in the
present study.
2.2 | RNA extraction, library
preparation, and sequencing
Weextractedwhole-bodytot alRNAfromthreecarnivoretadpoles
and three omnivore tadpoles from each of three sibships, for a total
of nine carnivores and nine omnivores. We used whole-tadpole
samples to match the approach used in Liedtke et al. (2021) for
P. cultripesascloselyaspossible.TotalRNAwasextractedusingthe
TRIzol Plus RNA Purification Kit (Invitrogen,#12183555),followed
bytre atmentwithDNase.Wedetermine dRNApurityforeachs am-
pleusingaNanoDrop2000(ThermoScientific)andquantifiedtotal
RNAona Qubit4 using the RNA HS AssayKit(ThermoScientific).
The RNA s amples were shi pped to Novogene, w here sample Q C,
librarypreparation,andsequencingwereperformed.Wegenerated
150-PE re ads using a NovaS eq 6000 s equencer (Ill umina). Of the
initial18samples,14passedqualitycontrolandweresequenced.All
foursamplesthatwerenotsequencedwerecarnivores, with three
from a single family. That family was included in generating the de
novo transcriptome but excluded from all differential expression
analyses(seebelow).
2.3 | Generation of de novo transcriptome and
quality assessment
The sequence data was examined for quality using “FastQC”
(Andrews,2010).Aftercombiningallreads,weutilized“Trinity”v2.8.6
(Haasetal.,2013)totrimreads,performinsiliconormalization,and
then generate a draft assembly of the S. multiplicatatadpolewhole-
body transcriptome (“Trinity” flags used: --trimmomatic --normal-
ize_max_read_cov50).Trimmingwasperformedwithinthe“Trinity”
call using default Trimmomatic settings: SLIDINGWINDOW:4:5
LEADING:5TRAILING:5MINLEN:25(Bolgeretal.,20 14).
Weexaminedthequa lit yofthetrans criptomeforb othreadre p-
resentation and completeness of gene content. To investigate read
representation, we mapped normalized read pairs back onto the
transc riptome using Bo wtie2 v2.4. 5 (Langmea d & Salzberg, 2012)
to determine the percentage of all paired reads represented in the
transcriptome assembly. To examine gene content completeness, we
first used “BUSCO” v.5.2.2 (Manni et al., 2021) with the “tetrapo-
da-odb10 databa se” as our refere nce, which all ows us to examine
whether highly conserved tetrapod genes are present in the assem-
bly. We also ran “blastx” against both the SwissProt database and the
Xenopus tropicalisproteome,usinganEvaluethresholdof≤1e−20 , to
identif ysequencesthathighlymatchotherrelatedtranscriptomes.
2.4 | Functional annotation of transcriptome
Functionalannotationofthetranscriptomeassemblywasperformed
using “Trinotate”v.3.2.1(Bryant et al., 20 17). “Trinotate” combines
various annotations into a single output; each annotation is per-
formed in dividually. We first i dentified tra nscript sequen ces with
similarities to known proteins using “blastx” (Evalue cutoff ≤1e−5)
against the SwissProt database and a subset of the SwissProt data-
base consisting of only human genes. We next sought likely coding
regions using “TransDecoder” (https:// github. com/ Trans Decoder).
The resulting putative coding regions were queried against the
complete SwissProt database and a subset of the SwissProt data-
base consisting of only vertebrates using “blastp” (Evalue cutof f
≤1e−5). We additionally searched for conserved protein domains
using“HMMER”(http:// hmmer. org)againstthePfamdatabase(Finn
et al., 2015).Weused“SignalP ”v4.1(Petersenetal.,2011) to predic t
signalpe pt idesand“TmHMM ”v 2. 0(https:// servi ces. healt htech. dtu.
d k / s e r v i c e . p h p ? T M H M M - 2 . 0 )topredicttransmembraneregions.As
afinalstep,we appliedgeneontology (GO)terms, aswellasKEGG
(KyotoEnc yclopediaofGenes andGenomes; http:// www. genome.

|
5 of 18
ISDANER et al.
jp/kegg/) and EggNOG (Huerta-Cepas etal., 2015) annotations as
provided by “Trinotate,” to each transcript in the assembly. The re-
sulting annotation database produced by “Trinotate” was examined
using the R package “TrinotateR” (ht tps:// github. com/ cstub ben/
trino tateR ).
We next estimated transcript abundance using “kallisto” v0.46.1
(Bray et al. , 2016) and subseq uently exclude d all transcript s with
less than o ne transcript p er million (<1 TPM) from the transcrip-
tome assembly since transcripts with very low expression levels are
of dubious biological relevance. The assembly was next evaluated
using NCBI'sVecScreentofilter outpossiblevector, adapter,and/
or primer contamination of the transcriptome. We additionally used
the“MC SC”decontaminationmethod(Lafond-Lapalmeetal.,2016)
with the target phylum Chordata to attempt to filter out any in-
ferred non-chordate transcripts. Further, the resulting transcripts
were compa red to the “nt” dat abase using “blas tx” (Evalue cu toff
≤1e−5), and all transcripts with best matches outside Chordata were
removed. Transcripts that hadnomatchwereretained. Finally,the
assemblywasblaste dagainsttheNCBIUniVecdatabaseusingstan-
dard VecScreen parameters, filtering out transcripts with a match
below an Evalue threshold of 1e−7.Thus,weappliedextensivequal-
ity control filters to produce our final transcriptome.
2.5 | Analysis of differential gene expression
Forourdifferentialexpressionanalyses,weonlyincludedS. multipli-
catasamplesfromfamilieswithomnivoreandcarnivoresequencing
data(i.e.,families5and11).Wefirstestimatedtranscriptabundance
attheTrinity“gene”levelusing“kallisto”v0.46.1(Brayetal.,2016).
Utilizing “edgeR” v3.38.1 (McCarthy et al., 2012; Robinson
et al., 2010), we examined the clustering of individuals by morph
using multi-dimensional scaling of log2 counts per million. “edgeR”
wasthenusedtoidentifydifferentiallyexpressedgenes(DEGs)be-
tween carnivores and omnivores, with family as a covariate. We con-
sidered all genes with a false discovery rate of q < 0.05significantly
differentially expressed. This set of differentially expressed genes
likely constitutes a set of downstream “effectors” that maint ain or
allow func tioning of the alternative phenotypes, as opposed to up-
stream master regulatory genes.
We implemented the same procedure to identify differentially
expressed genes in response to pond dr ying in P. cultripes, which
undergoes plastic developmental acceleration in response to drying
pondconditions.TherawsequencedataforP. cultripes was accessed
from the NCBISequence Read Archive (SRA; SRP161446) and the
transcriptomefromtheNCBITranscriptomeShotgunAssemblyda-
tabase (TSA; GHBH01000000) under BioProject PRJNA490256.
Previous analysisof this data (Liedtkeet al., 2021) identified differ-
entially e xpressed gene s between a high -water control and t hree
different time points in a low-water treatment. We re-analyzed
differentialgeneexpressionforeachpairof high-watercontroland
low-water tre atment tim e points ind ividuall y.Doi ng so allowed us
to evaluate how each timepoint cor responds (in terms of shared
differentially expressed genes) to differential expression between
carnivoresandomnivoreswhileanalyzingeachdatasetidentically.
2.6 | Functional annotation of differentially
expressed genes
We examined each species' differentially expressed genes for func-
tionalenrichmentusingg:Profilerinits web-basedimplementation
(Raudve re et al., 2019). We conducted this analysis using annota-
tionsforthehumanproteomeasthebackgrounddomain.ForP. cul-
tripes, this analysis was performed for differentially expressed genes
ineachpair wisesetof high-watercontrolandlow-watertreatment
time points. We corrected for multiple testing using g:Profiler's
g:SCS algorithm. We examined ontologies and pathways from the
GO:BiologicalProcess,KEGG,andReactomedatabases.
2.7 | Analysis of overlap in differentially
expressed genes and functional annotations between
Spea and Pelobates
Becau se sequence dif ferences across t he two species mi ght lead
tosimilarsequencesmatching differentannotations,weper formed
areciprocal best-hit annotation using “blastn” to generate a list of
matching sequences between S. multiplicata and P. cultripes(asop-
posed to comparing best-hit annotations to oneanother,since the
best match may be a different or tholog or in a different reference
species across the two spadefoot species). We then performed a
second differential expression analysis using “edgeR” for each spe-
ciesusingthisdirectcross-speciesannotation.
To address the question of whether S. multiplicata utilizes an
existin g plastic res ponse to deser t conditions , we queried th e list
resulting from the differentially expressed gene analysis in S. multipli-
cata against the corresponding list from each pairwise comparison in
P. cultripes(i.e.,bet weeneachlow-watertimepointandthehigh-wa-
ter control) to determine the number of genes overlapping between
the two species contrasts. We performed permutation tests at
eachtimepointtoevaluateifthenumberofoverlappingDEGs was
greater than expected by random chance. To conduct these tests,
werandomlysampledgenesfromtheexpression-filteredtranscrip-
tomeofeachspeciescorrespondingto:(1)thenumberofgenesdif-
ferentially expressed in S. multiplicata,and (2) the numberofgenes
differentially expressed in P. cultripes. We then examined the num-
ber of overlapping genes from each permutation on a pairwise basis
corresponding to the original analyses and determined the number
ofpermutationsthatequaledorexceededtheequivalentvaluefrom
the actual data to calculate a measure of statistical significance.
We next identified the differentially expressed genes in S. mul-
tiplicata that were not significantly differentially expressed at each
timepoint in P. cultripes or that did not align to genes in P. cultripes.
Theseanalysesexaminewhether(1)constitutivelyexpressedgenes
in P. cultripeshave acquired newdifferential expressionpatternsin

6 of 18
|
ISDANER et al.
S. multiplicataasacomponentofit sp ol yp he ni sm ,and( 2)theS. multi-
plicatapolyp he ni smpossessesuniquege ne snotf ou nd inP. cultripes,
respectively.Findinggenesineitherofthesetwoclasses wouldbe
consistent with lineage-specific evolution of gene expression in
S. multiplicata.
Finally,weanalyzedoverlappingfunctionalannotationusingg:-
Profiler on the overlapping gene sets in each time period and com-
parison (DEGs vs.DEGsorDEGsvs.non-significant genes)andfor
the set of gen es unique to S. multiplicata's polyphenism using an-
notatio ns from the GO: Biol ogical Process , KEGG, and Reac tome
datasets.
3 | RESULTS
3.1 | Transcriptomics of Spea morphs
Conducting comparative geneexpressionrequired thedenovoas-
sembly of a transcriptome for S. multiplicata tadpoles. To do so, we
generate d between 16.4 and 24. 5 million 150-PE re ads (mean of
20.8millionreads),foratotalof582.0million150-PEreads, across
the 14 sequen ced samples . After in sili co normalizat ion, 20.4 mil-
lionpair-end reads (3.5%of thetotal reads) ser ved as input foras-
sembling the S. multiplicata transcriptome. The output from “Trinity”
consisted of 457,153 transcript contigs (median length = 430 bp),
whichclusteredinto310,955“genes”(thatis,clustersoftranscripts
withsharedsequencecontent).Wemapped97.3%ofallpairedreads
ontothe transcriptome de novo assembly (Table 1).BUSCOanaly-
sisindicatesnear-completesequenceinformation for 93.1% of the
genes in the “tetrapoda_odb10” database, with just 2.0% of genes
fragme nted and 4.9% missin g (Table 2).Ali gning the S. multiplicata
transcriptome to the SwissProt database using “blastx” resulted in
15,00 4 SwissProt prote ins represente d by nearly full- length tra n-
scripts(>80%alignmentcoverage),andasimilaranalysiscomparing
to the X. tropicalisproteomerevealed15,882 proteins represented
bynearlyfull-lengthtranscripts,outofthe22,282genesand37,716
proteins found in the X. tropicalis proteome. These values compare
favorably to the number of nearly full-length transcripts aligned
to each protein database in the recently assembled P. cultripes
transcriptome (13,645 and 12,715 proteins, respectively; Liedtke
et al., 2019). Theseresults indicate that we havegeneratedahigh-
quality transcriptome for whole-tadpole S. multiplicata, at least as
complete as those previously assembled for other species of spade-
foottoads(Liedtkeetal.,2019).
Multiple functional annotations of the S. multiplicata transcrip-
tomeservedasinputforTrinotate(completeannotationinDataS1).
A comparison of the transcriptome assembly to the SwissProt
database using “blastx” provided a best-match annotation for
216,650 transcripts (Table 3). When these annotations were sub-
jectedto GO analysis, we matched 21,251 uniqueGO terms (out
of2,042,040 total terms).Predictionof coding regions(CDS)with
“TransDecoder”identified159,127CDS,representing51.2%ofthe
Trinity “genes” in the assembly. Comparison of the TransDecoder
results against the SwissProt database using “blastp” annotated
115,297 CDS, and a second comparison to the vertebrate-only
subsetofSwissProtannotated113,254CDS. Otherannotationsof
TransDecoder-predictedCDSincluded99,382hitsagainstthePfam
database, 11,935 signalP-predicted peptides, 27,573 TmHMM-
predicted transmembrane proteins, and 152,816 KEGG terms
(Table 3). Among sequences thatwere annotated with vertebrate
genes,26,281uniqueproteinsfrom the vertebrate-onlysubsetof
SwissProt were recovered in S. multiplicata. Parallel analysis of the
P. cultripestranscriptomeidentified 25,029 unique vertebrate pro-
teins in that species' transcriptome. Between the two species, there
were32,853uniqueproteinsrecovered,with18,457(56.2%ofthe
total) shared between the two species, 7824 (23.8%) unique to
S. multiplicata,and6572(20.0%)uniquetoP. cultripes.Afterfiltering
toremovetranscriptswithlowexpression(<1TPM), 70.8%ofthe
TAB LE 1 TranscriptomeassemblystatisticsforSpea multiplicata.
S. multiplicata transcriptome assembly
Total raw reads 582,031,892
Insiliconormalizedreads 20,401,976
Trinity transcripts in assembly 547,15 3
Trinity “genes” in assembly 310,955
Read pair s aligned to the assembly 97. 3 %
Proper pair reads aligned to the assembly 92.1%
N50oftranscript s 2676 bp
N50oflongestisoformper“gene” 855 bp
Sizeoftotaltranscriptome 500,4 33,975 bp
Sizeoftranscriptomeonlyincl.longestisoform
per “gene”
195,537,278 bp
Mediansizeoftranscripts 430 bp
Mediansizeoflongestisoformper“gene” 342 bp
Averagesizeoftranscripts 1094.7 bp
Averagesizeoflongestisoformper“gene” 628.8 bp
Note: Trinity outputs are provided at both the tr anscript and “gene”
levels.
TAB LE 2 GenecontentcompletenessassessmentoftheSpea
multiplicata transcriptome assembly.
S. multiplicata transcriptome gene content
Proteinsrepresentedbynearlyfull-lengthtranscript sa compared to
SwissProt 15,004
Xenopus tropicalis proteome 15,822
BUSCOresults
Complete 93.1%
Fragmented 2.0%
Missing 4.9%
Note:BUSCOwasperformedusingthe“tetrapoda-odb10”database.
a>80%alignmentcoverage;basedongroupedhighscoringsegment
pairs(HSPs)toaccountformultiplefragment spertranscriptaligningto
asinglesequence.

|
7 of 18
ISDANER et al.
transcripts in the transcriptome were retained. VecScreen filtering
furtherreduced the size of the S. multiplicata transcriptome by95
transcripts. Afterconducting “MCSC” decontamination to remove
inferred non-chordate transcripts and manual filtration using the
UniVec database, the transcriptome consisted of 288,112 tran-
scripts.Insummary,our densely annotated,filteredtranscriptome
allows for robust and informative downstream analyses of gene ex-
pressionpatternsandothertranscriptomicinquiries.
With our transcriptome assembled, we next analyzed gene ex-
pression patterns between carnivores and omnivores in S. multipli-
cata. Multidimensional scaling analysis on standardized count data
(log2 CPM) for S. multiplicata revealed distinct clusters for carnivores
andomnivoresalong the first dimension, accounting for 46% ofthe
variationbetween the twomorphs(Figure 2a).Ofthe12,676genes
withexpressiondataacrossthesamples,2177hadsignificantlyhigher
expression in omnivores, while 2203 had significantly higher expres-
sionincarnivores(FDR < 0.05;Figure 2b, Data S2).
WhenDEGsareclusteredusinghierarchicalclusteringwithcom-
plete linkagebased on expression pattern (Figure 2c), there are six
majorclustersofDEGsbetweenS. multiplicata omnivores and carni-
vores.Intotal,34.6%ofthetotalgenesweredifferentiallyexpressed
betweenthemorphs.Functionalenrichmentanalysisofthesegenes
resulted in many terms, reflecting many DEGs between morphs
(DataS3). Thus, our de novo transcriptome enables the detection of
uniquegeneexpressionprofilesforcarnivores andomnivoresthat
differ in functions related to protein metabolism, developmental
processes, regulation of cellular processes, cell differentiation, sig-
nal transduction, cellular response to chemical stress, and cardiac
muscle contraction. The large number of DEGs and wide range of
functional categories support the idea that divergence between
Annotation summary of the Spea multiplicata transcriptome assembly
TransDecoder-predictedcodingregions(ORFs) 15 9,1 27
SwissProtproteinhits(blastp) 76 , 4 0 7//1 15 , 2 97
SwissProtvertebratesonlyproteinhits(blastp) 7 4, 5 3 6// 11 3 , 2 5 4
Pfamhits(HMMER) 64, 573//99,382
Predictedpeptides(signalP) 38 19 //1 1 ,93 5
Predictedtransmembraneproteins(tmHMM) 16,875//27,573
GOPfam 26 2 8 //61,836
KEGG 38,663//152,816
Transcript sannotatedagainstSwissProt(blastx) 216,650
blastxGOterms(unique//total) 21,251//2,042,040
TABLE 3 SummaryofTrinotateresults
indicating the number of annotations for
unique//totalTransDecoder-predicted
candidate genes identified with various
tools and databases.
FIGURE 2 Geneexpressionpatterns
within Spea multiplicata tadpoles.
(a)Multidimensionalscalingplotsof
log2-counts-per-millionalongthefirst
dimensions.(b)VolcanoplotofRNA-seq
data at the Trinit y “gene” level, where
differentially expressed genes with
q < 0.05arestatisticallysignificant.
(c)Heatmapoflog2counts-per-millions
fortranscriptsthatshowstatistically-
significant differential expression
between carnivores and omnivores.
Carnivore samples are labeled with C1
through C5, omnivore samples are labeled
fromO1throughO6.

8 of 18
|
ISDANER et al.
these two morphs encompasses a complex suite of developmental
differences.
3.2 | Transcriptomics of P. cultripes developmental
acceleration
To compare gene expression between ancestral and derived forms
of plasticity, we next performed differential gene expression anal-
ysis in P. cultripes in the same way as in S. multiplicata (usingraw
read data previously published in Liedtke et al., 2 019). Pairwise
differential expression betweenthe 24-h high watercontrol and
each low wate r treatment (24, 48 , or 72 h) in P. cultripes results
in the following (Data S2): (1) in the 24-h treatment, 79 tran-
script s were differen tially expres sed; (2) in the 48-h tr eatment,
337transcripts were differentiallyexpressed;and(3)in the 72-h
treatment,208transcriptsweredif ferentiallyexpressed.Intotal,
492uniquetranscriptsweredifferentiallyexpressedbetweenthe
controlandalltreatments .Functionalenrichmentanalysisofeach
pairwisecomparison(DataS3)revealed:(1)notermsforthe24-h
treatment; (2) cholesterol metabolic processes, lipid metabolic
processes, steroid metabolic processes, and steroid biosynthetic
process es for the 48-h t reatment; a nd (3) cholest erol metab olic
processes, alcohol metabolic processes, steroid metabolic pro-
cesses, lipid biosynthetic processes, steroid biosynthetic pro-
cesses, and regulation of mast cell cytokine production for the
72-htreatm en t.Th us ,d espite th en um be ro fDEGs be tweenwater
level treatments in P. cultripes being roughly an order of magni-
tude lower than the number bet ween carnivore and omnivore S.
multiplicata,theDEGsinP. cultripes are still enriched for particular
functional(e.g.,geneontology)categories.
3.3 | Shared responses between resource
polyphenism and developmental acceleration
Comparing the results of the differential gene expression analy-
sis acrossboth species based on thereciprocal-best-match an-
notation, we identified a ver y limited number of genes that were
differentially expressed both between carnivores and omnivores
in S. multiplicataandbetweenhigh-watercontrolandlow-water
treatments in P. cultripes (see Table 4; Figure 3). There were five
overlapping genes between S. multiplicata and P. cultripes when
comparedto the 24-h low-water treatment (Figure 4a), 35 over-
lappinggenes when compared to the 48-h low-watertreatment
(Figure 4c),and13overlappinggeneswhencomparedtothe72-h
low-water treatment (Figure 4e). In tot al, there were 4 6 unique
overlapping genes between those differentially expressed in
S. multiplicata and those in P. cultripes. Permutation tests indicate
that each result is not significantly dif ferent from random expec-
tations (24-h treatment: p = .80, 48-h treatment: p = .09; 72-h
treatment: p = .90; Figure 3b,d,f), suggesting that there is not a
greater than expected number of differentially expressed genes
shared between these forms of plasticity.
When we queriedwhether there were shared processesamong
the overlapping genes between carnivores and omnivores and
for each time period (i.e., the set ofshared DEGs between species
comparisons), we found that these genes were enriched for particu-
larfunctional terms at the latter twotimepoints(Figure 5; Table 5).
Specifically, when comparing S. multiplicata with P. cultripes using the
48-h low-water treatment, shared processes based on the shared
DEGs include terms related to steroidmetabolic processes, carbon
metabolic processes, pyruvate metabolic processes, steroid biosyn-
thesis,cholesterolbiosynthesis,andtRNAaminoacylation.Comparing
S. multiplicata with P. cultripes using the 72-h low-water treatment
yielded some of the same (and similar)functionally enriched terms,
including steroid metabolic processes, steroid biosynthesis, and cho-
lesterol biosynthesis. Additionally, terms for cholesterol metabolic
processes, lipid biosynthesis, and lipid metabolism were enriched at
thistimepoint.Thus,althoughthenumberofoverlappingDEGsisnot
greater than expected, those that overlap are functionally enriched for
putatively important biological processes.
3.4 | Lineage-specific gene expression plasticity in
S. multiplicata
Whencomparing the DEGs in S. multiplicata to the genes that are
not significantly differentially expressed in P. cultripes, we found
that2860genesoverlappedforthe24-hlow-watertreatmentcom-
parison,2829genesoverlapped for the 48-h low-watertreatment
comparison, and 2855 genes overlapped for the 72-h low-water
treatment(Figure 3).Additionally,anumberofDEGsinS. multiplicata
do not align to any gene in P. cultripesafter reciprocal-best-match
annotation. These number approximately 1620 at each of the three
timepoints(Figure 3). Together, this suggests that a large number
of genes insensitive to pond drying/developmental acceleration in
Pelobates are condition dependent in the context of Spea's resource
polyphenism.
Functionalenrichmentanalysisoftheset ofDEGsin S. multiplicata
overlapping with genes not significantly differentially expressed in
P. cultripesreturnedmanyhigh-levelfunctionalterms(DataS3), includ-
ing terms for organismal, head, brain, and nervous system development;
protein metabolism; and response to endogenous stimuli, commensu-
ratewiththelarge-scalechangesinvolvedintheresourcepolyphenism.
Likewise, the genes showing plasticity in S. multiplicata but that did not
align to genes in P. cultripes were enriched for diverse terms, including
brain,head,andnervoussystemdevelopment(DataS3). Together, this
suggeststhatmajordevelopmentalreorganizationisinvolvedinthere-
source polyphenism, but genes underlying these changes were either
not plastic or did not align to transcripts in the ancestral pond drying
response of P. cultripes.Generally,thesefindingsdonotsupportthehy-
pothesisofco-option,butsuggestthatlineage-specificchangestogene
expression dominate the Spea plastic response.

|
9 of 18
ISDANER et al.
TAB LE 4 DifferentiallyexpressedgenessharedbetweenSpea multiplicata(S.m. below) resource polyphenism and Pelobates cultripes(P.c. below) developmental acceleration
(logFC = log2(fold-change),HW = highwater,LW = lowwater,C = carnivore,O = omnivore).
Shared differentially expressed genes
Human gene ID Gene name
P.c.
treatment P.c. logFC P.c. p-value S.m. logFC S.m. p-value P.c. higher expression S.m. higher expression
SPSB3 splA/ryanodine receptor doma in and SOCS box protein 3 24 h 1.41 5.77e-05 −1 .1 2 3.05e-0 4 HW C
HPX Hemopexin 24 h −2. 02 1.54e-05 2.60 3.94e-06 LW O
PTP4A2 Proteint yrosinephosphatase4A2 24 h −1. 5 4 1.56e-07 −1.19 1.70e-05 LW C
ADM2 Adrenomedullin 2 24 h 2.04 3.23e-08 −2 . 67 4.68e-08 HW C
TMEM67 Transmembraneprotein67 24 h −1. 3 5 2.00e- 06 1.45 9. 8 8e- 0 5 LW O
CYTL1 Cytokine like 1 48 h −1.9 7 4.78e-06 −2. 33 3.80e-08 LW C
FGL1 Fibrinogenlike1 48 h −0.97 2.00e-04 0.83 1.51e-02 LW O
RCN1 Reticulocalbin 1 48 h 1.15 6.08e-05 −1 . 39 4.68e-03 HW C
MARS1 Methionyl-tRNAsynthetase1 48 h 1.22 2.35e-04 −1. 0 0 2 .42e-05 HW C
ESCO2 Est abl.ofsisterchromatidcohesionN-acetyltrans ferase2 48 h 1.18 3.40e-04 −1. 2 8 1.95e-03 HW C
ARHGEF3 Rho guanine nucleotide exchange factor 3 48 h −1.36 8 .41e- 05 − 0.97 3 .59e-05 LW C
PKLR Pyruvate kinase L/R 48 h 3.80 2 .40 e -1 2 −0.83 9.02e-03 HW C
FARSA Phenylalanyl-tRNAsynthetasesubunitalpha 48 h 1.47 1.33e-04 −1 . 6 6 2.92e-0 9 HW C
CA RS1 Cysteinyl-tRNAsynthetase1 48 h 1.15 8 . 07e-0 5 −1 . 2 9 9.68 e - 0 8 HW C
SPSB3 splA/ryanodine receptor doma in and SOCS box protein 3 48 h 2.05 3.89e-05 −1 .1 2 3.05e- 04 HW C
FCHSD1 FCHanddoubleSH3domains1 48 h −0.82 4.34e-04 1.40 6.91e-05 LW O
PMM1 Phosphomannomutase 1 48 h 0.99 2.25e-04 1.57 9.36 e - 0 5 HW O
ADM2 Adrenomedullin 2 48 h 1.63 1.53e-04 −2. 67 4.68e-08 HW C
GP2 Glycoprotein2 48 h −1 .4 7 6.76 e- 0 6 1.35 1.09e-03 LW O
GART Phosphoribosylglycinamide formyltransferase 48 h 0.97 3. 65e-0 4 −1 .01 7.71e-03 HW C
DHCR24 24-dehydrocholestrol reductase 48 h 3.39 4.81e-09 0.91 4.60e-03 HW O
TM7SF2 Transmembrane7super familymember2 48 h 1. 51 1.49e-04 1.78 1.40e-07 HW O
CY P51A 1 Cytochrome P450 family 51 subfamily A member 1 48 h 3. 59 3. 7 9e -15 1.97 4.01e-05 HW O
TYR Tyrosinase 48 h 1.10 1.54e-05 0.91 3.33e-03 HW O
SCD Stearoyl-CoAdesaturase 48 h 1.96 8.41e- 05 −1 . 15 5.54e-03 HW C
NARS1 Asparaginyl-tRNAsynthetase1 48 h 1.57 7. 0 1e- 0 6 −0.93 1.02e-03 HW C
MSMO1 Methylsterol monooxygenase 1 48 h 3.13 3.00e-09 1.95 8.59e-04 HW O
ACSS2 Acyl-CoA synthetase short chain family member 2 48 h 2.14 8 . 51e -05 2.08 4. 99 e -11 HW O
TP53RK TP53 regulating kinase 48 h 1.78 4. 21e- 07 −0.81 4.36e-03 HW C
(Continues)

10 of 18
|
ISDANER et al.
Shared differentially expressed genes
Human gene ID Gene name
P.c.
treatment P.c. logFC P.c. p-value S.m. logFC S.m. p-value P.c. higher expression S.m. higher expression
PHGDH Phosphoglycerate dehydrogenase 48 h 1.62 2.55 e-07 −2. 3 2 4.66 e -10 HW C
C9ORF64 Q-nucleotideN-glycosylase1 48 h 2 .74 3.11e-0 4 1.06 2 .74 e-0 3 HW O
PS AT1 Phosphoserine aminotransferase 1 48 h 1.92 1.66e-06 0.93 7. 3 2 e - 0 3 HW O
CYP4F22 CytochromeP450family4subf amilyFmember22 48 h 2.01 1.22e-09 0.92 2 .41e -03 HW O
IPO4 Importin4 48 h 1.19 3.70e-04 −1 . 0 4 3.51 e- 0 4 HW C
MAD2L1BP MAD2L1bindingprotein 48 h 1.94 4.32e- 06 −1 . 55 4.14e- 03 HW C
AC AT2 Acetyl-CoAacetyltransferase2 48 h 1.25 2.45e-0 4 0.92 7. 0 2 e - 0 3 HW O
STOML 2 Stomatin like 2 48 h 1.85 1.44e-05 −1 . 36 1.67e-06 HW C
FEN1 Flapstructure-specificendonuclease1 48 h 1.20 3.67e- 0 4 −0.95 6.95e-03 HW C
CRYGD Crystallin gamma D 48 h 1.57 7. 3 5 e - 0 5 0.81 1.12e-02 HW O
ADK Adenosinekinase 48 h 1.46 7.77e - 0 5 −0.78 1.64e-02 HW C
PKLR Pyruvate kinase L/R 72 h 3.00 4.50e-07 −0.83 9.02e-03 HW C
TMEM97 Transmembraneprotein97 72 h 1.23 2.22e-0 4 1.50 2.96e-06 HW O
CALU Calumenin 72 h −1 . 94 1.88e-04 − 0.92 1.68e-02 LW C
HSD3B7 Hydroxy-delta-5-steroiddehydrogenase 72 h 1.52 2.76e-04 1.66 8.26e-05 HW O
AOC1 Amineoxidasecoppercontaining1 72 h −2.0 5 8.82e-06 1.56 3.35e- 05 LW O
CY P51A 1 Cytochrome P450 family 51 subfamily A member 1 72 h 3.03 5.3 3 e -10 1.97 4.01e-05 HW O
DHCR24 24-dehydrocholestrol reductase 72 h 2.84 5.94e-07 0 .91 4.60e-03 HW O
MSMO1 Methylsterol monooxygenase 1 72 h 2.61 4.18 e-07 1.95 8. 59e- 04 HW O
PISD Phosphatidylserine decarboxylase 72 h 1.72 2.20e-04 −1 .02 9. 5 7e - 05 HW C
ACSS2 Acyl-CoA synthetase short chain family member 2 72 h 2.29 4 .74 e -05 2.08 4. 99 e -11 HW O
MARCHF8 Membraneassociatedring-CH-typefinger8 72 h 4.59 5.64e-08 1.37 4.23e-04 HW O
HMGCR 3-hydroxy-3-methylglutaryl-CoAreductase 72 h 2.00 1.17e-04 1.43 1.80e-05 HW O
ATL 2 AtlastinGTPase2 72 h 3.45 4.02e-06 0.62 8.75e-0 3 HW O
Note:GenesthatappearinmorethanoneP.c. treatment are labeled in bold.
TAB LE 4 (Continued)

|
11 of 18
ISDANER et al.
4 | DISCUSSION
Using a de novo transcriptome for S. multiplicata and a previously
published transcriptome and data for P. cultripes, we investigated
the origins of gene expression plasticity associated with a novel
larval resource polyphenism in S. multiplicata(Figure 1a,c). We
found that this derived form of plasticity appears to have evolved
primari ly via lineage-s pecific chan ges in gene expre ssion as op-
posed to co- opting mech anisms from an a ncestral fo rm of plas-
ticity—accelerating larval development rate in response to pond
drying(Figure 1b,c).
Specifically, we found that these two forms of plasticity share a
minimal set of differentially expressed genes and that most genes
showing mo rph-biased e xpression in S. multiplicata were not asso-
ciated with the pond drying response in P. cultripes(Figures 3 and
4).Ontheonehand,thisfindingwasunexpected:thepolyphenism
in S. multiplicataischaracterizedbythedevelopmentallyaccelerated
carnivo re morph (de la Ser na Buzon et al., 2020; Pfennig, 19 92 a,
1992b). On th e other hand , this polyp henism involve s much more
thanjustdevelopmentalacceleration.Indeed,wefoundthattheset
of genes showing plasticity in S. multiplicata, but not showing it in
P. cultripes,is enrichedformajor organismal, head,andbrain devel-
opment terms. These data are therefore consistent with previous
studies, which have shown that carnivores differ from omnivores
behavior ally (Pfenn ig, 1999; Pfennig et al., 19 93; Pomeroy, 1981),
morphologically (Levis et al., 2018; Martin & Pfennig, 2009;
Pfennig, 1992b; Pfennig & Murphy, 2000, 2002), and physiologically
(Ledón-Rettig,2021;Ledón-Rettig et al.,2008, 2009, 2023). Thus,
itmakessensethatextensivelineage-specificgeneexpressionplas-
ticity has evolved in Spea's polyphenism when compared to the rel-
atively simple plasticity of developmental acceleration in Pelobates.
Given our f inding of few shared r esponses and ex tensive lin-
eage-specific responses, we speculate that the evolution of this
polyphenism may have expanded from a limited set of shared plas-
tic responses that are functionally enriched for having roles in lipid
metabolism(especiallycholesterolbiosynthesis),steroidbiosynthe-
si s , andt R N Aamin o a c y l atio n ( Figure 5; Table 5).Subsequently,previ-
ouslynon-plasticgenesmayhavebeenrecruitedasthepolyphenism
underwentelaborationandrefinement(Casasaetal.,2020;Foquet
et al., 2021; Morris et al., 2014). Such a process may be especially
likely to occu r when, as suggeste d elsewhere (Levis et al., 2021,
2022), the shared responses constitute a core set of genes that
promote a tadpole'sdevelopment into alternativetrajectories, and
when the lineage-specific plasticity genes constitute those that
maintain,elaborate,andrefinethealternativephenotypes(Lafuente
& Beldade, 2019). In deed, the evol ution of polyp henisms in oth er
taxa involves bringing other developmental processes into a con-
ditionally expressed context (Abouheif & Wray, 2002; Bhardwaj
et al., 2020; Hanna & Abouheif,2022;Projecto-Garcia et al., 2017;
Sommer, 2020; Suzuki & N ijhout, 2006). Thus, we speculate that
plasticity in a small set of genes and processes might set a lineage on
thepathtoevolvingapolyphenism,butsubstantial lineage-specific
alterations are needed for a polyphenism to actually arise.
Our results come with caveat s. First, using whole tadpoles
might obscure additional responses at individual tissue levels. We
used whole tadpoles to ensure that our de novo transcriptome and
analyses were similar to those of the previous study we were using
asa reference (Liedtke et al.,2021).Additionally,thepolyphenism
in S. multiplicata involves a mosaic of tissues throughout the body,
includi ng the gut, jaw musc les, and brain (se e above). Yet, future
workwould benefitfromtakinga tissue-specific look at the devel-
opment of both forms of plasticity, especially given the evidence
from this system (Levis et al., 2022) and other systems (Mateus
et al., 2014; Oo stra et al., 2018; Suzuk i & Nijhout, 2006; van der
Burg & Reed, 2021) that tissues differ in how they respond to in-
ternalandexternalenvironmentalchange.Anothercaveatconcerns
thelimited temporalsampling.Ifthe omnivore-to-carnivoretransi-
tion was assayed sooner (or later), orthe response to pond drying
wasassayedsooner(orlater),theremayhavebeenmoresimilarities
betweenthetwoformsof plasticity.As wehaveno apriori reason
FIGURE 3 Differentiallyexpressed
genes(DEGs)inSpea multiplicata
in relation to Pelobates cultripes.
Differentially expressed genes in
S. multiplicatacategorizedbywhether
they:(1)overlapwithdifferentially
expressed genes in P. cultripes(red),
(2)overlapwithgenesthatarenot
significantly differentially expressed in
P. cultripes(gray),(3)donotaligntoany
genes in P. cultripes(lightblue).

12 of 18
|
ISDANER et al.
to believe any particular timepoint in the P. cultripes data is more
similar to the S. multiplicata data, we compared all timepoints here,
but future studies would benefit from more precise matching of the
timeframeof development. Finally,given that Spea (like Pelobates)
exhibits pond-drying plasticit y (Figure 1b,c), future studies should
replicate the Pelobates experiment in Spea and determine the pat-
ternstheirdevelopmentalrateplasticitygenerated.Indoingso,one
could identify which differentially expressed genes are related to de-
velopmental speed per se and not Pelobates-specificplasticity.Thus,
future studies could benefit from fine-grained tissue, temporal,
andlineage-specificsamplingto characterizefurtherthedegree to
which these two forms of plasticity share transcriptomic bases.
With such future analyses in mind, the transcriptome assembled
here provides a significant resource for Spea. For examp le, it will
facilitate analyses of splicing and regulatory differences between
morphs, investigating expression differences related to sexual se-
lection and hybridization (Chen & Pfennig, 2020; P fennig, 2007;
Seidl et al., 2019), and the transcriptional bases of other aspects
of Spea biolo gy (Levis et al ., 2021, 2022).Additionally,as demon-
strated here, the transcriptome will allow for comparative studies
of plasticity not only across other spadefoot species, but also more
widelyamongAnuraandhighertaxa.Thistranscriptomeprovidesa
significant addition to the growing genomic resources available for
S. multiplicata,whichto datehadnofull-lengthtranscriptome-wide
annotationtoaccompanyitsassembledgenome(Seidletal.,2 019).
Moreover, it helps fulfill calls for more such resources in anurans
generally(Koschetal.,2023).
Inconclusion,ourresultsprovideimpor tantinsightsintoanovel
polyphenism's evolutionary and developmental origins. The number
of genes shared between an ancestral plastic response to pond dr y-
ing via developmental acceleration in P. cultripes and the more com-
plex polyphenism in Spea is dwarfed by the much greater number of
genes gaining plasticity in Spea.Theselineage-specificgeneexpres-
sionpatternsareinvolvedinmajordevelopmentalshift sthatsupport
the complex whole organism changes involved in carnivore produc-
tion. Consistent with gene expression plasticity evolution in other
systems(Casasaetal.,2020;Foquetetal.,2021), we also found that
Spea'spolyphenismrequiresmoregeneexpressionchangesthanthe
pond drying response in Pelobates. Together, this suggests that more
general ancestral stress responses might be a springboard for subse-
quentevolutionary innovation, but that substantial lineage-specific
FIGURE 4 Examiningtheoverlapindifferentialgeneexpression
between Spea multiplicata and Pelobates cultripes. Each row
depict s a Euler plot of the total number of differentially expressed
genesineachspecies(blue = S. multiplicata,yellow = P. cultripes)
and a histogram of the number of overlapping genes in 1000
permutations, in each of which the actual number of differentially
expressedgeneswasselectedfromtheexpression-filteredlistof
genes found in the transcript data from each species, respectively.
The observed number of overlaps is marked on each histogram,
as is the calculated p-value(theproportionofpermutationswith
more overlapping genes than the obser ved value). The number of
differentially expressed genes between carnivores and omnivores
in S. multiplicata is the same in all three Euler plots. The control for
eachrowisthe24-hhigh-watersamples,whilethetreatmentsare
thesamplesfrom24-hlowwater(a,b),48-hlowwater(c,d),and
72-hlowwater(e,f).
FIGURE 5 OverlappingfunctionalannotationbetweenSpea
multiplicata and Pelobates cultripes. Stacked bar plot of overlapping
functional annotation of the set of overlapping differentially
expressedgenes,includingannotationsfromtheGO:Biologic al
Process,KEGG,andReactomedatabases.Eachbarisbasedon
the list of overlapping genes bet ween carnivores and omnivores in
S. multiplicataandthegenesfromacomparisonbet weenthe24-h
high-watercontrolandthelow-watertreatmentstatedalongthe
x-axis(24-,48-,and72-hlow-watertreatments,fromlefttoright).

|
13 of 18
ISDANER et al.
TABLE 5 SummaryofGO:BiologicalProcess(GO:BP),KEGG,andReactome(REAC)termsidentifiedthroughfunctionalannot ationanalysisoftheoverlappinggenesbetweencarnivoresand
omnivores in Spea multiplicataandhighandlowwaterateachtime-pointtreatmentinPelobates cultripes.
P. cultripes treatment Source GO term Term name
Corrected
p-value
48 hlowwater GO:BP GO:0044281 Small molecule metabolic process 9.56 E-10
GO:BP GO: 0019752 Carboxylic acid metabolic process 1.60E-07
GO:BP GO:0043436 Oxoacidmetabolicprocess 2.31E-07
GO:BP GO:0006082 Organicacidmetabolicprocess 2. 55E-07
GO:BP GO:0006520 Cellular amino acid metabolic process 4.18E- 05
GO:BP GO:0044283 Small molecule biosynthetic process 4.41 E-04
GO:BP GO:1901566 Organonitrogencompoundbiosyntheticprocess 6.79E-04
GO:BP GO:0006418 tRNAaminoacylationforproteintranslation 1.39E-03
GO:BP GO:0006399 tRNAmetabolicprocess 1.6 6E-03
GO:BP GO:00 43039 tRNAaminoacylation 1.99E- 03
GO:BP GO:0043038 Aminoacidactivation 2.17E-03
GO:BP GO:1902653 Secondary alcohol biosynthetic process 3.25E-0 3
GO:BP GO:0006695 Cholesterol biosynthetic process 3.25E-0 3
GO:BP GO :1901617 Or ganichydroxycompoundbiosyntheticprocess 4.22E-03
GO:BP GO :0 016126 Sterol biosynthetic process 5.37E- 03
KEGG KEGG:00100 Steroid biosynthesis 1.02E-05
KEGG KEGG :01100 Metabolic pathways 1.59E- 05
KEGG KEGG:00970 Aminoacyl-tRNAbiosynthesis 2.75E- 04
KEGG KEG G:01200 Carbon metabolism 5.85E-0 4
KEGG KEGG:00620 Pyruvate metabolism 1.46 E-02
REAC RE AC:R-HSA-191273 Cholesterol biosynthesis 1.01E- 06
REAC RE AC:R-HSA-379716 CytosolictRNAaminoacylation 1.15E- 04
REAC RE AC:R-HSA-8957322 Metabolism of steroids 5.78E- 04
REAC RE AC:R-HSA-379724 tRNAAminoacylation 1.17 E-0 3
REAC REAC:R-HSA-2426168 ActivationofgeneexpressionbySREBF(SREBP) 4.53E-02
(Continues)

14 of 18
|
ISDANER et al.
P. cultripes treatment Source GO term Term name
Corrected
p-value
72 hlowwater GO:BP GO:0006695 Cholesterol biosynthetic process 3.38E-05
GO:BP GO:1902653 Secondary alcohol biosynthetic process 3.38E-05
GO:BP GO :0 016126 Sterol biosynthetic process 5.61E-05
GO:BP GO:0006694 Steroid biosynthetic process 5.78E-05
GO:BP GO:0008610 Lipid biosynthetic process 1.03E-04
GO:BP GO :1901617 Or ganichydroxycompoundbiosyntheticprocess 2.72E-04
GO:BP GO :1901615 Organichydroxycompoun dmetabolicprocess 6.27E- 04
GO:BP GO:0044283 Small molecule biosynthetic process 6.82E-0 4
GO:BP GO:0008202 Steroid metabolic process 1.15E- 03
GO:BP GO:0008203 Cholesterol metabolic process 1.53E-0 3
GO:BP GO:0046165 Alcoholbiosyntheticprocess 1 .57E- 03
GO:BP GO:1902652 Secondary alcohol metabolic process 2.03E-03
GO:BP GO:0006066 Alcoholmet abolicprocess 2.2 2E-03
GO:BP GO :0 01612 5 Sterol metabolic process 2.38 E-03
GO:BP GO:0006629 Lipid metabolic process 9.4 8 E- 0 3
GO:BP GO:1900222 Negativeregulationofamyloid-betaclearance 1.70E-02
KEGG KEGG :01100 Metabolic pathways 1.70E- 05
KEGG KEGG:00100 Steroid biosynthesis 5.74E-05
REAC RE AC:R-HSA-191273 Cholesterol biosynthesis 2.37E-04
REAC RE AC:R-HSA-8957322 Metabolism of steroids 1.30E-03
REAC RE AC:R-HSA-211945 PhaseI—Functionalizationofcompounds 2.01E-02
REAC REAC :R-HSA-556833 Metabolism of lipids 4.36E-02
Note: The t reatment in P. cultripesusedforeachsetofannotationsisinthefirstcolumn(notethattherearenosignific anttermswhenusingthe24-hlow-watertreatmentinP. cultripes).
TABLE 5 (Continued)

|
15 of 18
ISDANER et al.
modification is needed to craft such responses into an adaptive
polyphenism. More generally, our work suggests that the evolution
ofcomplexformsofplasticity(likeresourcepolyphenism)mayhave
little reliance on simpler forms of ancestral plasticity, which could
explain why polyphenisms are relatively rare across the tree of life.
AUTHOR CONTRIBUTIONS
Andrew J. Isdaner:Datacuration(lead);formalanalysis(lead);in-
vestigation(lead);methodology(equal);resources(equal);validation
(lead);visualization(equal); writing–originaldraft(equal);writing–
review and editing (equal). Nicholas A. Levis: Conceptualization
(lead); formal analysis (supporting); investigation (suppor ting);
methodology (equal); project administration (supporting);supervi-
sion(suppor ting);visualization (supporting);writing– original draft
(equal); writing – review and editing (equal). David W. Pfennig:
Formalanalysis (supporting);investigation(supporting);methodol-
ogy (equa l); projec t administr ation (lea d); resource s (equal); supe r-
vision(lead);validation(supporting); visualization (equal);writing –
originaldraft(equal);writing–reviewandediting(equal).
ACKNOWLEDGMENTS
WethankEmilyHarmonandKarinPfennigforlaboratoryassistance
and discussion that improved the manuscript and two anonymous
referee s for helpful co mments. A g rant from th e National S cience
Foundation(DEB-1753865)toD.P.fundedthework.UNC's IACUC
approved all procedures.
DATA AVAIL AB ILI T Y STAT EME N T
AllrawsequencesreadsareavailableintheNCBISRA(SRP339994).
The transcriptome assembly has been deposited at DDBJ/EMBL/
GenBan k under the accessi on GKIA00000000. Both the raw se-
quence reads and the transcriptome assembly can be found in
BioProject PRJNA768487. Data S1–S3 may be found published
alongside this paper.
ORCID
Andrew J. Isdaner https://orcid.org/0000-0002-7944-0927
REFERENCES
Abouheif,E., & Wray, G. A . (2002).Evolutionofthe gene networkun-
derlying wing polyphenism in ant s. Science, 297(5579), 249–252.
h t t p s : / / d o i . o r g / 1 0 . 1 1 2 6 / s c i e n c e . 1 0 7 1 4 6 8
Agrawal,A.A.(2001).Phenotypicplasticityintheinteractionsandevo-
lution of species. Science, 294,321–326.
AmphibiaWeb.(2016).Informationonamphibianbiologyandconserva-
tion(webapplicationhttp:// amphi biaweb. org/ ).
Andrews,S. (2010). FastQC: Aqualit ycontroltoolfor high throughput
sequence data. h t t p : // w w w . b i o i n f o r m a t i c s . b a b r a h a m . a c . u k / p r o j e 
c t s / f a s t q c
Babonis, L. S., Mar tindale, M. Q., & Ryan , J. F. (2016). Do novel genes
drive morphological novelty? An investigation of the nemato-
somes in the sea anemone Nematostella vectensis. BMC Evolutionary
Biology, 16(1),114.h t t p s : / / d o i . o r g / 1 0 . 1 1 8 6 / s 1 2 8 6 2 - 0 1 6 - 0 6 8 3 - 3
Bento, G ., Ogawa, A., & So mmer, R. J. (2010). Co-op tion of the hor-
mone-signalling module dafachronic acid-DAF-12 in nematode
evolution. Nature, 466(7305), 494–497. h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / 
n a t u r e 0 9 1 6 4
Bhardw aj, S., Jolan der, L. S.-H., Wen k, M. R., Ol iver, J. C., Nijho ut, H.
F.,& Monteiro,A. (2020).Origin of themechanismofphenotypic
plasticity in satyrid butterfly eyespot s. eLife, 9,e49544.h t t p s : // d o i .
o r g / 1 0 . 7 5 5 4 /e L i f e . 4 9 5 4 4 
Biddle,J.F.,&Ragsdale,E.J.(2020).Regulatorsofanancientpolyphen-
ism evolved through episodic protein divergence and par allel gene
radiations. Proceedings of the Royal Society B, 287(19 21),20192595.
Bolger, A. M ., Lohse, M ., & Usadel, B . (2014). Trimmomatic : A flexible
trimmer for Illumina sequence data. Bioinformatics, 30(15),2114–
2120. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / b i o i n f o r m a t i c s / b t u 1 7 0
Bray,N. L.,Pimentel, H., Mels ted, P.,&Pachter,L .(2016).Near-optimal
probabilistic RNA-seqquantification. Nature Biotechnology, 34(5),
525–527.h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / n b t . 3 5 1 9
Bryant, D. M., Johnson, K., DiTommaso, T., Tickle, T., Couger, M. B.,
Payzin- Dogru, D., L ee, T.J ., Leigh, N. D. , Kuo, T.H ., Davis, F. G.,
Bateman,J., Bryant,S.,G uzikowski,A .R.,Tsai,S.L .,Coyne,S.,Ye,
W. W.,Fre eman, R . M., Jr., Peshkin , L., Tabin, C. J. , … Whited, J.
L. (2017). A tissue-mapped axolotl de novo transcriptome enables
identification of limb regeneration factors. Cell Reports, 18(3),762–
776 . h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . c e l r e p . 2 0 1 6 . 1 2 . 0 6 3
Cabrales-Orona, G.,& Délano-Frier,J. P.(2021). Searchingforaniden-
tity: Functional characterization of taxonomically restricted
genes in gr ain amaran th. In D. Adhika ry, M. K. Dey holos, & J. P.
Délano-Frier (Eds.), The Amaranth genome (pp. 97–124). Springer
InternationalPublishing.
Casas a, S., Biddl e, J. F.,Kou tsovoulos , G. D., & Ragsd ale, E. J. (2021).
Polyphenism of a novel trait integrated rapidly evolving genes into
ancestrally plastic networks. Molecular Biology and Evolution, 38(2),
33 1–3 4 3 . h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / m o l b e v / m s a a 2 3 5
Casasa, S., Zattara, E. E., & Moczek, A . P.(2020). Nutrition-responsive
gene expression and the developmental evolution of insect poly-
phenism . Nature Ecolog y & Evolution, 4(7),970–978.https:// doi. org/
1 0 . 1 0 3 8 / s 4 1 5 5 9 - 0 2 0 - 1 2 0 2 - x
Chen,C.,&P fennig,K.S.(2020).Femaletoadsengaginginadaptivehy-
bridization prefer high-quality heterospecifics as mates. Science,
367(6484),1377–1379.h t t p s : / / d o i . o r g / 1 0 . 1 1 2 6 / s c i e n c e . a a z 5 1 0 9
Davison,D.R .,&Michod,R.E.(2021).Phenotypicplasticity andevolu-
tionar ytransitionsinindividuality.InD.W.Pfennig(Ed.),Phenotypic
plasticity and evol ution: Causes, consequen ces, controversies (pp.
241–266).CRCPress.
delaSernaBuzon,S.M.,Martin,R.A.,&Pfennig,D.W.(2020).Carryover
effec ts and the evolution of polyphenism. Biological Journal of th e
Linnean Society, 131(3),622–631.
Falconer,D.S., &Mackay,T.F.C.(1996).Introduction to qua ntitative ge-
netics(4thed.).LongmanGroupLtd.
Fan, X.-L ., Lin, Z.-H., & Wei, J. (2014). Effectsof hydroperiod duration
ondevelopmentalplasticityintigerfrog(Hoplobatrachus chinensis)
tadpoles. Zoological Research, 35(2),124–131.
Finn,R .D.,Coggill,P.,Eberhardt,R.Y.,Eddy,S.R.,Mistry,J.,Mitchell,A.
L.,Potter,S.C.,Punta,M.,Qureshi,M.,Sangrador-Vegas,A.,Salazar,
G.A.,Tate,J.,&Bateman,A .(2015).ThePfamproteinfamiliesda-
tabase: Towards a more sustainable future. Nucleic Acids Research,
44(D1),D279–D285.h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / n a r / g k v 1 3 4 4
Foquet,B ., Castellanos, A. A ., &Song, H.(2021). Comparative analysis
of phenot ypic plasticity sheds light on the evolution and molecu-
lar underpinnings of locust phase polyphenism. Scientific Reports,
11(1),11925.h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 5 9 8 - 0 2 1 - 9 1 3 1 7 - w
Gianola,D. (1982).Theory and anal ysisofthresholdcharacters. Journal
of Animal Science, 54(5),107 9–1096.h t t p s : / / d o i . o r g / 1 0 . 2 5 2 7 / j a s 1 9 
8 2 . 5 4 5 1 0 7 9 x
Goldstein, I., & Ehrenreich,I. M. (2021). Genetic variation in phenotypic
plasticity.In D. W. Pfennig (Ed.),Phenotypic plasticity and evolution:
Causes, consequences, controversies(pp.91–111).CRCPress.

16 of 18
|
ISDANER et al.
Haas, B. J ., Papanicolao u, A., Yassour, M., Gr abherr, M., Bloo d, P. D.,
Bowden, J., Couger, M. B., Eccles, D., Li, B., Lieber, M., MacManes,
M.,Ott ,M.,Orvis,J.,Pochet,N.,Strozzi,F.,Weeks,N.,Westerman,
R.,William,T.,Dewey,C.N.,…Regev,A.(2013).Denovotranscript
sequencereconstruction from RNA-sequsing the trinity plat form
for reference generation and analysis. Nature Protocols, 8(8),1494–
1512. h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / n p r o t . 2 0 1 3 . 0 8 4
Hanna, L ., & Abou heif, E. (2022 ). Deep cons ervati on and co-op tion of
programmed cell death facilitates evolution of alternative pheno-
types at multiple biological levels. Seminars in Cell & Developmental
Biology, 145, 28–41. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . s e m c d b . 2 0 2 2 . 0 5 . 
024
Harmon , E. A., Eva ns, B., & Pfe nnig, D. W. (2023). Frog ha tchings use
earlyenvironmentalcuestoproduceananticipatoryresource-use
phenotype. Biology Letters, 19, 20220613. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 8 / 
rsbl. 2022. 0613
Hendr y,A.P.(2016).Keyquestionsontheroleofphenotypicplasticityin
eco-evolutionarydynamics.Journal of Heredity, 107,25–41.
Hess,C.,Levine,J.M.,Turcotte,M. M.,&Har t,S.P.(2022).Phenot ypic
plasticity promotes species coexistence. Nature Ecolog y & Evolution,
6(9),1256–1261.h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 5 5 9 - 0 2 2 - 0 1 8 2 6 - 8
Huerta-Cepas,J.,Szklarczyk,D.,Forslund,K.,Cook,H.,Heller,D.,Walter,
M. C., Rattei, T., Mende, D. R., Sunagawa, S., Kuhn, M., Jensen, L .
J., von Mer ing, C., & B ork, P. (2015). eggN OG 4.5: A hier archic al
orthology framework with improved functional annotations for eu-
karyotic,prokaryoticandviralsequences.Nucleic Acids Research,
44(D1),D286–D293.h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / n a r / g k v 1 2 4 8
Jasper, W. C., Li nksvayer, T. A., At allah, J., Fri edman, D., C hiu, J. C., &
Johnson,B. R. (2015).Large-scalecodingsequence change under-
lies the evolution of postdevelopmental novelty in honey bees.
Molecular Biolog y and Evolution, 32(2),334–346.https:// doi. org/ 10.
1 0 9 3 / m o l b e v / m s u 2 9 2
Johnson,B.R.(2018).Taxonomicallyrestrictedgenesarefundamentalto
biology and evolution. Frontiers in Genetics, 9,407.https:// doi. org/
1 0 . 3 3 8 9 / f g e n e . 2 0 1 8 . 0 0 4 0 7 
Khalturin, K., Hemmrich, G., Fraune,S., Augustin, R ., & Bosch, T.C . G.
(2009).Morethanjustorphans:Aretaxonomically-restrictedgenes
important in evolution? Trends in Genetics, 25(9),404–413.ht tp s://
d o i . o r g / 1 0 . 1 0 1 6 / j . t i g . 2 0 0 9 . 0 7 . 0 0 6
Kosch, T. A., C rawford, A . J., Muell er,R . L., Wollen berg Valero, K . C.,
Power, M. L., Rodríguez, A., O'Connell, L. A., Young, N. D., &
Skerrat t,L .F.(2023).Comparativeanalysisofamphibiangenomes:
An emerging resource for basic and applied research. bioRxiv,
2023-2002.
Lafond-Lapalme,J.,Duceppe, M.-O., Wang,S.,Moffet t,P.,&Mimee,B.
(2016).Anew method for decontaminationofde novo transcrip-
tomes using a hierarchical clustering algorithm. Bioinformatics,
33(9),1293–1300.h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / b i o i n f o r m a t i c s / b t w 7 9 3
Lafuente,E.,&Beldade,P.(2019).Genomicsofdevelopmentalplasticity
in animals. Frontiers in Genetics, 10, 720.h t t p s : / / d o i . o r g / 1 0 . 3 3 8 9 / 
f g e n e . 2 0 1 9 . 0 0 7 2 0 
Langmead,B.,&Salzberg,S.L.(2012).Fastgapped-readalignmentwith
bowtie 2. Nature Methods, 9(4),357–359.h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / 
nmeth.1923
Ledón-Rettig,C.,Pfennig,D.W.,&Crespi,E.J.(2009).Stresshormones
and the f itness con sequence s associate d with the tr ansition t o a
novel diet in larval amphibians. Journal of Experimental Biology, 212,
3743–3750.
Ledón-Rettig, C. C. (2021). Novel br ain gene-expression pat terns
are associated with a novel predaceous behaviour in tadpoles.
Proceedi ngs of the Royal Society B, 288 ,20210079.https:// doi. org/
1 0 . 1 0 9 8 / r s p b . 2 0 2 1 . 0 0 7 9
Ledón-Rettig,C.C.,Lo,K.M .,&Lagon,S.R .(2023).Baselinecorticoste-
rone levels in spadefoot toads reflec t alternate larval diets one year
later. General and Comparative Endocrinology, 339,114291.h t t p s : //
d o i . o r g / 1 0 . 1 0 1 6 / j . y g c e n . 2 0 2 3 . 1 1 4 2 9 1
Ledón-Rettig, C. C., & Pfennig, D. W. (2011). Emerging model systems
ineco-evo-devo:Theenvironmentallyresponsivespadefoottoad.
Evolution and Developm ent, 13,391–400.
Ledón-Rettig, C. C., Pfennig, D. W., & Nascone-Yoder, N. (2008).
Ancestral variationand the potential for geneticaccommodation
inlarvalamphibians:Implic ationsfortheevolutionofnovelfeeding
strategies. Evolution and Development, 10(3),316–325.
Levis,N.A.,delaSernaBuzon,S.,&P fennig,D.W.(2015).Aninducible
offense: Carnivore mor ph tadpoles induced by tadpole carnivory.
Ecology and Evolution, 5(7), 1405–1411. https:// doi. org/ 10. 10 02/
ece3.1448
Levis,N.A .,Isdaner,A.,&P fennig,D.W.(2018).Morphologicalnovelty
emergesfrompre-existingphenotypic plasticity.Nature Ecology &
Evolution, 2,1289–1297.
Levis, N. A.,Kelly,P.W., Harmon, E. A.,Ehrenreich,I.M., McKay, D.J.,
& Pfenn ig, D. W. (2021). Transcripto mic bases of a p olypheni sm.
Journal of Experimental Zoology Part B: Molecular and Developm ental
Evolution, 336,482–485.h t t p s : / / d o i . o r g / 1 0 . 1 0 0 2 / j e z . b . 2 3 0 6 6 
Levis, N . A., Mart in, R. A. , O'Donnel l, K. A., & P fennig, D. W. (2017).
Intraspecific adaptive radiation: Competition, ecological oppor-
tunity, and phenot ypic diversification within species. Evolution,
71(10 ) , 24 96 –2 509.
Levis, N . A., McKay, D. J., & P fennig, D. W. (2022). D isentangl ing the
developmental origins of a novel phenotype: Enhancement versus
reversal of environmentally induced gene expression. Proceedings
of the Royal Society B, 289, 20221764. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 8 / 
r s p b . 2 0 2 2 . 1 7 6 4
Levis,N.A.,&Pfennig,D.W.(2019).Phenotypicplasticity,canalization,
and the or igins of novelty: Evidence and mechanisms from a mphibi-
ans. Seminars in Cell & Developmental Biology, 88,80–90.ht tps:// d o i .
o r g / 1 0 . 1 0 1 6 / j . s e m c d b . 2 0 1 8 . 0 1 . 0 1 2
Levis, N. A.,& Pfennig,D. W.(2021).Innovationand diversification via
plasticity-ledevolution.InD.W.Pfennig(Ed.),Phenotypic plasticity
and evolution: Cause s, consequences, controversies (pp. 211–240).
CRC Press.
Levis,N.A.,&Ragsdale,E.J.(2022).Link ingmolecularmechanismsand
evolutionary consequences of resource polyphenism. Frontie rs
in Integrative Neuroscience, 16, 8 05061. h t t p s : / / d o i . o r g / 1 0 . 3 3 8 9 / 
fnint.2022.805061
Levis, N . A., Reed , E. M. X., P fennig, D. W., & Buf ord Reiskind , M. O.
(2020). Identification of candidate loci for adaptive phenotypic
plasticity in natural populations of spadefoot toads. Ecology and
Evolution, 10(16),8976–8988.https:// doi. org/ 10. 1002/ ece3. 6602
Liedtke , H. C., Gar rido, J. G., E steve-Codi na, A., Gu t, M., Aliot o, T.,&
Gomez-Mestre,I.(2019). De novo assembly and annotation of the
larval transcriptome of two spadefoot toads widely divergent in de-
velopmental rate. G3: Genes, Genomes, Genetics, 9(8),26 47–2655.
h t t p s : / / d o i . o r g / 1 0 . 1 5 3 4 / g 3 . 1 1 9. 4 0 0 3 8 9
Lied tk e, H. C .,Ha rn ey,E.,&Gom ez -M es t re ,I .( 20 21 ).Cr os s- sp eciestr an -
scriptomics uncovers genes underlying genetic accommodation of
developmental plasticity in spadefoot toads. Molecular Ecology, 30,
2220–2234.h t t p s : / / d o i . o r g / 1 0 . 1 1 1 1 / m e c . 1 5 8 8 3 
Love, A. C ., & Wagner, G. P. (2022). Co-opt ion of stress me chanisms
in the origin of evolutionar y novelties. Evolution, 76 (3), 394–413.
https :// doi. or g/ 10. 1111/ evo. 14 421
Manni, M.,Berkeley,M. R ., Sepp ey,M.,Simao, F.A., & Zdobnov,E. M.
(2021). BUSCO update: Nove l and streamlin ed workflows al ong
with broader and deeper phylogenetic coverage for scoring of
eukaryotic, prokaryotic, and viral genomes. Molecular Biology
and Evolution, 38, 4647–4654. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / m o l b e v / 
msab19 9
Marti n, R. A. , & Pfennig , D. W.(2 009). Disru ptive sele ction in nat ural
populations: The roles of ecologic al specialization and resource
competition. The American Naturalist, 174 ,268–281.
Mateus, A . R., Marques-Pita, M.,O ostra, V.,Lafuente,E., Brakefield,
P.M.,Zwaan,B.J.,&Beldade,P.(2014).Adaptivedevelopment al

|
17 of 18
ISDANER et al.
plasticity: Compartmentalized responses to environmental
cues and to corresponding internal signals provide phenotypic
flexibility. BMC Biology, 12, 97. h t t p s : / / d o i . o r g / 1 0 . 1 1 8 6 / s 1 2 9 1 
5 - 0 1 4 - 0 0 9 7 - x
Mayr,E.(1963).Animal sp ecies an d evolution.HarvardUniversityPress.
McCar thy, D. J., Chen, Y., & Smyt h, G. K. (201 2). Differen tial expres-
sion analysis of multifactor RNA-se q experiments with respect
to biological variation. Nucleic Acids Research, 40(10), 4288–4297.
h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / n a r / g k s 0 4 2
Michener,C.D.(1961).SocialpolymorphisminHymenoptera.Symposium
of the Royal Entomological Society of London, 1,43–56.
Moczek, A. P., Sultan, S. E., Foster, S., Ledón-Ret tig, C., Dworkin, I.,
Nijhout , H. F., Abouhei f,E ., & Pfennig , D. W.(2 011).T he role of
developmental plasticity in evolutionary innovation. Proceedings of
the Royal Society B: Biological Sciences, 278,2705–2713.ht t p s : // doi.
o r g / 1 0 . 1 0 9 8 / r s p b . 2 0 1 1 . 0 9 7 1
Morris,M.R.J.,Richard,R .,Leder,E.H.,Barrett,R .D.H.,Aubin-Hor th,
N., & Roger s, S. M. (2014). Gen e expressi on plastic ity evolves i n
respon se to colonizat ion of freshwate r lakes in three spine stick-
leback. Molecular Ecology, 23(13), 3226–3240 . https:// doi. org/ 10.
1111/mec.12820
Nijhout ,H.F.(2003).Developmentandevolutionofadaptivepolyphen-
isms. Evolution and Development, 5(1),9–18.
Ogawa, A ., Streit , A., Ant ebi, A., & S ommer, R. J. (20 09). A conse rved
endocrine mechanism controls the formation of dauer and infective
larvae in nematodes. Current Biology, 19(1),67–71.https:// doi. org/
1 0 . 1 0 1 6 / j . c u b . 2 0 0 8 . 1 1 . 0 6 3
Oostra,V.,Saastamoinen,M.,Zwaan,B.J.,&Wheat,C.W.(2018).Strong
phenot ypic plasticity limits potential for evolutionary responses to
climate change. Nature Communications, 9(1),1005.https:// doi. org/
1 0 . 1 0 3 8 / s 4 1 4 6 7 - 0 1 8 - 0 3 3 8 4 - 9
Petersen,T.N .,Brunak,S.,von Heijne,G.,& Nielsen,H.(2011).SignalP
4.0: Discriminating signal peptides from transmembrane regions.
Nature Methods, 8(10), 785–786. h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / n m e t h . 
1701
Pfenn ig, D. W. (1990). The adapti ve signific ance of an envi ronmenta l-
ly-cued development alswitchinananurant adpole.Oecologia, 85,
101–107.
Pfennig, D. W. (1992a). Polyphenism in spadefoot toad tadpoles as
a locally adjusted evolutionarily stable str ategy. Evolution, 46,
140 8–1420.
Pfennig,D.W.(1992b).Proximateandfunctionalcausesofpolyphenism
in an anuran tadpole. Functional Ecology, 6,167–174.
Pfennig, D. W. (1999). Cannibalistic tadpoles that pose the greatest
threat to kin are most likely to discriminate kin. Proceedings of the
Royal Societ y of London. Serie s B, 266 ,57–81.
Pfennig,D.W.(2021a).Keyquestionsaboutphenotypicplasticity.InD.
W.Pfennig (Ed.), Phenotypic plasticit y and evolution: Causes, conse-
quences, controversies(pp.55–88).CRCPress.
Pfennig,D.W.(2021b).Phenot ypic plasticity and evolution: Causes, conse-
quences, controversies. CRC Press.
Pfennig,D.W.,Mabry,A.,&Orange,D.(1991).Environmentalcausesof
correlationsbetweenageandsizeatmetamorphosisinScaphiopus
multiplicatus. Ecology, 72,2240–2248.
Pfennig,D. W.,&Mur phy,P.J. (200 0). Character displacementin poly-
phenic tadpoles. Evolution, 54,1738–1749.
Pfennig,D.W.,&Murphy,P.J.(2002).Howfluc tuatingcompetitionand
phenotypic plasticity mediate species divergence. Evolution, 56(6),
1217–1228.
Pfennig,D.W.,&Pfennig,K.S.(2012).Evolution's wedge: Comp etition and
the origins of diversity. University of California Press.
Pfennig, D. W., Reeve, H.K., & Sherman,P.W.(1993).K in recognition
and cannibalism in spadefoot toad tadpoles. Animal Behaviour,
46(1),87–94.
Pfennig,D.W.,Wund,M.A.,Snell-Rood,E.C.,Cruickshank,T.,Schlichting,
C. D., & Moc zek, A. P. (2010). Phe notypic p lastici ty's imp acts on
diversification and speciation. Trends in Ecolog y & Evolution, 25,
45 9– 4 6 7.
Pfennig,K.S.(2007).Facultativematechoicedrivesadaptivehybridiza-
tion. Science, 318,965–967.
Pomeroy,L .V.(1981).Developmental polymorphism in the tadpoles of the
spadefoot toad Scaphiopus multiplicatus.(Ph.D.diss.), Universit y of
California,Riverside,CA .
Price, P. D., Palmer D roguet t, D. H., Taylor, J. A. , Kim, D. W., Place, E .
S.,Rogers, T.F.,Mank, J.E.,Cooney,C .R .,&Wright,A. E. (2022).
Detecting signatures of selection on gene expression. Nature
Ecology & Evolution, 6,1035–1045.h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 5 5 
9 - 0 2 2 - 0 1 7 6 1 - 8
Pro ject o-Garcia, J., Biddle, J.F.,&Ragsdale ,E .J.(20 17) .Deco di ngthear-
chitecture and origins of mechanisms for developmental polyphen-
ism. Current Opinion in Genetics & Development, 47,1–8.ht t p s : // d o i .
o r g / 1 0 . 1 0 1 6 / j . g d e . 2 0 1 7 . 0 7 . 0 1 5
Ragsdale, E. J., Müller,M. R., Rödelsperger,C., & Sommer,R . J. (2013).
Adevelopmentalswitchcoupledto theevolutionofplasticit yac ts
through a sulfatase. Cell, 155(4),922–933.
Raudvere, U., Kolberg, L., Kuzmin, I., Ar ak, T.,Adler, P.,Peter son, H.,&
Vilo, J. (2019). G:Profiler:A webserverforfunctional enrichment
analysisandconversions of gene list s(2019update).Nucleic Acids
Research, 47(W1), W191–W198. h t t p s : / / d o i . o r g / 1 0 . 1 0 9 3 / n a r / 
gkz369
Reid,J.M.,&Acker,P.(2022).Propertiesofphenotypicplasticityindis-
crete threshold traits. Evolution, 76(2),190–206.https:// doi. org/ 10.
1111/evo.144 08
Renn, S. C . P., & Schum er,M . E. (2013). Geneti c accommodat ion and
behavioural evolution: Insights from genomic studies. Animal
Behaviour, 85(5), 1012–1022. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / j . a n b e h a v . 
2013. 02. 012
Richter-Boix,A.,Tejedo,M.,&Rezende,E.L.(2011).Evolutionandplas-
ticit y of anuran la rval de velopme nt in respon se to desicc ation. A
comparative analysis. Ecology and Evolutio n, 1(1),15–25.ht t p s : // d o i .
org / 10. 1002/ ec e3. 2
Robinson, M. D., McCarthy, D. J., & Smyth , G. K. (2010). edgeR : A
Bioconductor package for differential expression analysis of digital
gene expression data. Bioinformatics, 26(1), 139–140. h t t p s : // d o i .
o r g / 1 0 . 1 0 9 3 / b i o i n f o r m a t i c s / b t p 6 1 6
Roff,D.A .(1996).Theevolutionofthreshold trait sinanimals.Quarterly
Review of Biology, 71,3–35.
Santo s,M.E.,Le Bo uq ui n,A. ,C ru mière,A .J.J .,& Kh il a,A.(2017 ).Ta xon-
restricted genes at the origin of a novel trait allowing access to a
new environment. Science, 358,386–390.
Schneider,R.F.,&Meyer,A .(2017). Howplasticity,geneticassimilation
and cryptic genetic variation may contribute to adaptive radiations.
Molecular Ecology, 26(1), 330–350. https:// doi . org/ 10. 1111/ mec.
13880
Seidl, F., Levis , N. A., Jones , C. D., Monroy-Eklu nd, A., Ehren reich, I.
M., & Pfe nnig, K. S. (2 019).Vari ation in hybrid g ene expressi on:
Implicationsfortheevolutionof geneticincompatibilitiesininter-
breeding species. Molecular Ecology, 28(20), 4 667–4679. ht tp s://
doi . org/ 10. 1111/ mec. 15246
Sommer,R.J.(2020).Phenotypicplasticity :Fromtheoryandgeneticsto
current and future challenges. Genetics, 215(1), 1–13. h t t p s : // d o i .
org/ 10. 1534/ genet ics. 120. 303163
Sultan,S.E. (2021). Phenotypicplasticity asanintrinsic propert yof or-
ganisms.InD. W.Pfennig (Ed.),Phenotypic plasticity and evolution:
Causes, consequences, controversies(pp.3–24).CRCPress.
Susoy, V., Herrmann, M., Kanzaki, N., Kruger, M., Nguyen, C. N .,
Rödelsperger, C., Röseler, W., Weiler, C., Giblin-Davis, R. M.,
Ragsdale, E. J., & Sommer,R .J.(2016).L arge-scale diversification
without genetic isolation in nematode symbionts of figs. Science
Advances, 2(1),e15010 31.
Suzuki,Y.,&Nijhout,H.F.(2006).Evolutionofapolyphenismbygenetic
accommodation. Science, 311,650–652.

18 of 18
|
ISDANER et al.
Székely,D.,Denoël,M.,Székely,P.,&Cogălniceanu,D.(2017).Ponddry-
ing cues and their effects on growth and metamorphosis in a fast
developing amphibian. Journal of Zoolog y, 303(2),129–135.ht tps://
d o i . o r g / 1 0 . 1 1 1 1 / j z o . 1 2 4 6 8 
Turcotte,M.M.,&Levine,J.M.(2016).Phenot ypicplasticit yandspecies
coexistence. Tre nds in Ecology & Evolution, 31(10),803–813.ht tp s : //
d o i . o r g / 1 0 . 1 0 1 6 / j . t r e e . 2 0 1 6 . 0 7 . 0 1 3
vanderBurg,K.R. L.,&Reed,R.D.(2021).Seasonalplasticity: Howdo
butterfly wing pattern traits evolve environmental responsiveness?
Current Opinion in Genetics & Development, 69,82–87.h t t p s : // doi.
o r g / 1 0 . 1 0 1 6 / j . g d e . 2 0 2 1 . 0 2 . 0 0 9
Venturelli,D. P.,Lofeu, L.,Kohlsdorf,T.,& Klein, W.(2022). Responses
to dehydration in tadpoles of Physalaemus nattereri (Anura:
Leptodactylidae). Hydrobiologia, 849(7 ), 1613–1624. h t t p s : // d o i .
o r g / 1 0 . 1 0 0 7 / s 1 0 7 5 0 - 0 2 2 - 0 4 8 0 5 - z
West-Eberhard,M.J.(1989).Phenotypicplasticit yandtheoriginsofdi-
ver si t y. Ann ual Review of Ecology a nd Systematics, 20,249–278.
West-Eberhard, M. J. (2003). Developmental plasticity an d evolution.
OxfordUniversityPress.
Wund,M .A .,Baker,J.A .,Clancy,B.,Golub,J.L.,&Foster,S.A.(20 08).
A test of th e "flexible s tem" model of evol ution: Ances tral plas-
ticity, genetic accommodation, and morphological divergence in
the threespine stickleback radiation. American Naturalist, 172,
449– 462.
Zeng,C.,Gomez-Mestre,I.,&Wiens, J.J. (2014).Evolution ofrapidde-
velopment in spadefoot toads is unrelated to arid environments.
PLoS One, 9(5), e96637. h t t p s : / / d o i . o r g / 1 0 . 1 3 7 1 / j o u r n a l . p o n e . 
0096637
SUPPORTING INFORMATION
Additional supporting information can be found online in the
Suppor tingInformationsectionattheendofthisarticle.
How to cite this article: Isdaner,A.J.,Levis,N.A.,&Pfennig,
D.W.(2023).Comparativetranscriptomicsrevealsthata
novelformofphenotypicplasticityevolvedvialineage-
specific changes in gene expression. Ecology and Evolution,
13, e106 46. ht tps://doi.org/10.10 02 /ece3.106 46