Running title: Inducible responses to multiple predators
Lei Gu, Shanshan Qin, Yunfei Sun, Jing Huang, Siddiq Akbar, Lu Zhang,
Zhou Yang*
Jiangsu Province Key Laboratory for Biodiversity and Biotechnology,
School of Biological Sciences, Nanjing Normal University, 1 Wenyuan
Road, Nanjing 210023, China
* Corresponding author: Zhou Yang
TEL: +86-25-85891671.
E-mail: yangzhou@njnu.edu.cn
Jiangsu Province Key Laboratory for Biodiversity and Biotechnology,
School of Biological Sciences, Nanjing Normal University, 1 Wenyuan
Road, Nanjing 210023, China
Abstract
Inducible defenses of prey are evolved under diverse and variable
predation risks. However, during the co-evolution of prey and multiple
predators, the responses of prey to antagonistic predation risks, which
may put the prey into a dilemma of responding to predators, remain
unclear. Based on antagonistic predation pressure from an invertebrate
(Chaoborus larvae) and a vertebrate (Rhodeus ocellatus )
predator, we studied the responses of multiple traits and transcriptomes
of the freshwater crustacean Ceriodaphnia cornuta under multiple
predation risks. Chaoborus predation risk altered the expression
of genes encoding cuticle proteins and modulated the biosynthesis of
steroid hormones, cutin, suberine, and wax, leading to the development
of horns and increase in size at the late developmental stage.
Meanwhile, fish predation risk primarily triggered genes encoding
ribosomes and those involved in unsaturated fatty acid biosynthesis and
cysteine and methionine metabolism, resulting in smaller individual size
and earlier reproduction. Inducible responses of both transcriptome and
individual traits revealed that predator-dependent unique responses were
dominant and the dilemma of antagonistic responses was relatively
limited. However, the unique individual traits in response to
invertebrate predation could be significantly impaired by vertebrate
predation risk, even though the unique responses to different predators
were extremely weakly correlated and could be elicited simultaneously.
These results indicate that diverse predator-dependent unique responses
are favored by Ceriodaphnia during its co-evolution with multiple
predators. Nonetheless, Ceriodaphnia is not a generalist that can
fully adopt all predator-dependent unique responses simultaneously under
multiple predation risks.
Keywords: Cladoceran; Chaoborus ; Fish; Inducible
defense; Predation risk
Introduction
In the co-evolution of predators and prey, defense is critical for prey
survival. Under variable predation risks, inducible defenses triggered
by predation cues are favored by prey (Tollrian & Harvell, 1999). For
successful defense against predators, prey organisms adopt various
inducible protective characteristics, including behavioral (De Meester,
1993), morphological (Gu et al., 2021), chemical defense (Selander et
al., 2015), and life history (Kvile, Altin, Thommesen, & Titelman,
2021) traits. Since diverse predation risks can prevent the stable
expression of an inducible defensive trait (Steiner & Auld, 2012), the
present study sought to understand the responses of prey to multiple
predation risks, particularly to predators exerting antagonistic
selection pressures.
Inducible defenses are common in aquatic organisms, such as
phytoplankton (Lürling, 2020), zooplankton (Diel, Kiene,
Martin-Creuzburg, & Laforsch, 2020), amphibians (Mitchell,
Bairos-Novak, & Ferrari, 2017), and fish (Brönmark & Miner, 1992).
Through the integration of inducible defense research, we classified
responses to multiple predators into two major types. The first type is
the general response, which evolves through diffusion co-evolution and
represents reciprocal adaptation in response to similar predators; for
instance, mayflies adopt the same avoidance behavior under predation
risks from different fish (Alvarez, Landeira-Dabarca, & Peckarsky,
2014). The second type is the specific response, which evolves through
pairwise co-evolution between specific predators and prey, such as the
immune responses to pathogens (Westra et al., 2015) and the inducible
crests of Daphnia in response to Notonecta predation
(Grant & Bayly, 1981). Moreover, the specific responses of different
traits can be further subdivided into antagonistic responses to the same
trait and unique responses to separate traits. Under antagonistic
selection pressures, if unique responses to separate traits are
dominant, the complex defense responses may incur maintenance cost, that
is, energetic costs of the sensory and regulatory mechanisms. Moreover,
if the prey primarily exhibits antagonistic responses of the same
traits, inducible responses to a predator may incur an environmental
cost, that is, vulnerability to other predators (Auld, Agrawal, &
Relyea, 2010); this environmental cost is further linked to selection
and plays a dual role in the evolution of inducible defenses
(Decaestecker, De Meester, & Ebert, 2002). Therefore, we hypothesized
that predator-dependent unique responses are dominant, helping, at least
in part, avoid the dilemma of the prey for responding to predators under
multiple predation risks.
In aquatic ecosystems, cladocerans are at the middle of the food chain,
acting as a food resource for insects and fishes (Miner, De Meester,
Pfrender, Lampert, & Hairston, 2012). These invertebrate and vertebrate
predators constitute antagonistic selection pressures on the size or
habitat selection of waterfleas; for instance, larger plankton are
vulnerable to large visual predators (e.g., fish) but less vulnerable to
small ambush predators (e.g., Chaoborus larvae) (Swift, 1992).
Presumably, inducible responses of Daphnia to visual predators
primarily include life history changes (Effertz & Von Elert, 2014),
which are rather different from the defensive traits triggered by
invertebrate predators, such as the development of “twist” (Herzog,
Rabus, Ribeiro, & Laforsch, 2016), neck teeth (Tollrian, 1993), and
horns (Gu, Qin, Zhu, et al., 2020). These reports support our
hypothesis.
Antagonistic inducible traits are commonly expressed by Daphnia .
For instance, Daphnia hyalina shows completely opposite responses
of size and reproduction under predation pressures by vertebrates and
invertebrates (Stibor & Lüning, 1994). Daphnia galeata prefers
deeper habitats under fish predation risk, while inhabits upper water
layers under Chaoborus predation risk (Dodson, 1988). In
addition, general responses, such as the development of an elongated
tail spine, are observed in Daphnia in response to fish,Triops , and Notonecta (Gu, Qin, Lu, et al., 2020;
Ritschar, Rabus, & Laforsch, 2020). Consistent with inducible defensive
traits, both general and specific responses appear at the molecular
level; for instance, in Daphnia magna , actin and tubulin
expression is decreased under Chaoborus larvae or fish predation
risks (Pijanowska & Kloc, 2004), ribosomal protein and vitellogenin
expression is increased under fish predation risk (Effertz, Mueller, &
Von Elert, 2015), and cuticle protein expression is increased but
vitellogenin expression is decreased under Triops predation risk
(Otte, Fröhlich, Arnold, & Laforsch, 2014). Overall, in Daphnia ,
a given species exhibits various types of responses under antagonistic
predation risks. However, from these sporadic studies on differentDaphnia species and clones, we cannot determine the precise type
of response preferred by the prey.
Ceriodaphnia cornuta is a widely distributed species with
sensitive inducible defensive traits (Gu et al., 2021; Qin et al.,
2021), providing a suitable model for testing our hypothesis. Since some
inducible traits are hidden (Laforsch, Ngwa, Grill, & Tollrian, 2004),
research on a few traits is insufficient. In recent years, omics
technologies have furthered our understanding of the mechanisms of
inducible defense (Hales et al., 2017; Zhang et al., 2021). Therefore,
to test our hypothesis that predator-dependent unique responses are
dominant, we assessed multiple traits and transcriptomes of C.
cornuta in response to predation pressures from Chaoborus larvae
and fish, respectively. Additionally, to test the hypothesis that
predator-dependent unique responses help prey avoid the dilemma of
responding to predators, we examined the expression of individual traits
under joint predation risks and explored the correlations between
various inducible responses.
Materials and methods
Predation risks
Predation risks were simulated using different predator-conditioned
medium prepared following the methodology described by Gu, Qin, Zhu, et
al. (2020). We cultured four Rhodeus ocellatus or 100Chaoborus sp. larvae in aged tap water, fed them C.
cornuta for 6 h, and then transferred them into 2 L of COMBO medium
(Kilham, Kreeger, Lynn, Goulden, & Herrera, 1998) for 18 h. The
predator-conditioned medium stock containing different predator
kairomones (Hahn, Effertz, Bigler, & von Elert, 2019; Weiss et al.,
2018) was passed through a 0.22 μm glass fiber filter (Millipore), and
the filtrate was stored in a refrigerator until experiments. The control
(C) included COMBO medium. The fish (F) and Chaoborus (CH)
predation risk treatments included respectively 20- (i.e., 1 fish per 10
L) and 2.5-fold (i.e., 20 Chaoborus larvae per liter) diluted
filtered medium stocks. Finally, the combination treatment (CH + F),
representing joint predation risks, included both diluted medium
stocks.
Life history experiment
The C. cornuta clone used in this experiment was sampled from
Lake Taihu (31°22′13.548″N, 120°0′16″E), China. C. cornuta was
cultured in COMBO medium at 25°C and 500 lx fluorescent light intensity
under a 14:10 h light/dark cycle and fed Chlorella pyrenoidosa(1.5 mg C·L-1). Synchronous C. cornuta at a
density of 1 individual per 10 mL, was adapted to the above conditions
for at least two generations. To test the type of inducible responses of
individual traits as well as the response strategy under joint predation
risks, we set up a full factor experiment containing C, F, CH, and CH +
F. We randomly divided newborn individuals into different treatments
within 12 h. Each individual was cultured in 10 mL of medium, with 10
replicates per treatment. The media in different treatments were
refreshed daily. Experimental workflow is presented in Figure S1.
Body size and horns were measured at maturity and at the late
developmental stage (i.e., at 16th day). Horns were scored following the
method described by Gu et al. (2021): absent (score 0), small (score 5),
and large (score 10). Individual scores were normalized by a maximum
score to define the induction level (0%–100%). In addition, time to
the first brood, neonate size, brood number, total offspring number, and
average brood size were recorded.
RNA extraction and
sequencing
To accurately evaluate the type of response at the transcriptional
level, we sequenced the transcriptomes of C. cornuta under C, F,
and CH treatments. Groups of 250 newborn individuals were cultured in
2.5 L of medium with three replicates per treatment. During cultivation,
responses triggered by different predation risks were verified through
inducible traits, that is, horns and body size at maturity. The medium
was refreshed daily, and samples were collected within 12 h after the
first brood. The samples were frozen in liquid nitrogen and homogenized
in TransZol Up. Total RNA was extracted using the TransZol Up Plus RNA
Kit (ER501, TRANS, China), following the manufacturer’s instructions.
RNA quality was assessed using the Agilent 2100 Bioanalyzer (Agilent
Technologies, USA) and verified using agarose gel electrophoresis. The
RNA integrity number of all samples was > 7.0.
The total mRNA of Ceriodaphnia was enriched by Oligo (dT) beads.
The enriched mRNA was fragmented into short fragments using
fragmentation buffer and reverse-transcribed into cDNA using random
primers. The cDNA fragments were purified, end-repaired, poly(A)-added,
and ligated to Illumina sequencing adapters. The ligation products were
size-selected and PCR-amplified to develop a cDNA library. The cDNA
library was sequenced using Illumina HiSeqTM 4000 by
Gene Denovo Biotechnology Co. (Guangzhou, China).
Transcript assembly and
annotation
Since genomic sequencing of Ceriodaphnia has not been conducted
to date, we adopted de novo RNA-Seq to analyze the C.
cornuta transcriptome. To obtain high-quality clean reads, sequenced
reads were cleaned by removing reads containing adapters, >
10% unknown nucleotides (N), and low-quality reads (Q-value ≤ 20).
Clean reads were assembled into unigenes using Trinity (Grabherr et al.,
2011). To annotate the unigenes, we used the BLASTx program with an
E-value < 10-5 in the NCBI non-redundant
protein (Nr), SWISS-PROT protein, Kyoto Encyclopedia of Genes and
Genomes (KEGG), and COG/KOG databases. Functional protein annotations
were obtained according to the best alignment results.
Gene expression analysis and
RT-qPCR
Unigene expression was quantified and normalized to reads per kb per
million (RPKM) (Mortazavi, Williams, McCue, Schaeffer, & Wold, 2008).
Differential transcript expression analysis between the control and
different predation risk treatments was performed using DESeq2. Genes
with a false discovery rate (FDR) < 0.05 and absolute fold
change ≥ 1 were considered differentially expressed genes (DEGs). We
categorized the DEGs into different types according to our
classification and then used KEGG pathways for functional analysis.
Pathways with P ≤ 0.05 were considered significantly enriched.
The representative DEGs of various significantly enriched pathways were
selected according to the following precedence conditions: stable
expression, large fold change, and pathways annotated in a closely
related species, that is, D. magna and Daphnia pulex .
To validate the RNA-Seq data in the C. cornuta transcriptome, we
quantified the expression of 18 random DEGs using RT-qPCR.
ddH2O was used as the negative control. cDNA was
synthesized from mRNA using Reverse Transcriptase SuperMix (R233,
Vazyme, China), and RT-qPCR was performed using ChamQ Universal SYBR
qPCR Master Mix (Q711, Vazyme, China). All primer sequences are listed
in Table S1. We obtained expression data for four alternative reference
genes (Scoville & Pfrender, 2010) and calculated their average gene
expression stability using geNorm. Glyceraldehyde-3-phosphate
dehydrogenase (G3PD), RNA polymerase II (RNAP II), and elongation factor
1-alpha genes (EF) were determined to be stably expressed and
geometrically averaged to calculate the gene expression normalization
factor for each sample. Gene expression was quantified using the
2-ΔΔt method. The concordance between RNA-Seq and
RT-qPCR data was assessed using regression analysis.
Statistical analysis
To test the effects of different predation risks, a MANOVA was performed
on individual traits. When the residuals were normally distributed
(Shapiro–Wilk test) and variances were homogeneous (Levene’s test), the
data for each trait were evaluated by two-way ANOVA, followed by
Bonferroni comparisons between different groups. When the normality test
failed, the Scheirer–Ray–Hare test was used, followed by the Wilcoxon
rank-sum test to assess significant differences among the different
treatments. In the present study, statistical significance was set at P
< 0.05. To simplify the effects of joint predation risks on
different types of responses, we considered our results based on the
conceptual diagram of possible prey responses in the combined predator
treatment (Fig. S2). To test the correlations between different types of
responses, we determined Pearson’s correlation coefficients between
traits as well as between representative DEGs in different categories.
The above statistical tests were performed in R (version 3.6.2). In
addition, correlation, principal component, DEGs, and pathway enrichment
analyses were conducted using OmicShare Tools
(https://www.omicshare.com/tools/).
Results
Morphology and life history
traits
The different predation risks significantly triggered various responses
of morphology and life history traits (Table 1). Compared with the
control, the responses induced by fish and Chaoborus predation
risks could be classified into the following four categories (Fig. 1).
(1) Unique responses to Chaoborus larvae [horn expression (at
maturity and 16th day) and total offspring number]: C. cornutadeveloped horns (maturity: P < 0.001; 16th day:P < 0.001) and produced more offspring (P =
0.002) under Chaoborus predation risk, although these responses
were not significant under fish predation risk. (2) Unique responses to
fish (time to first brood and neonate size): the neonate size (P= 0.006) and time to first brood (P < 0.001) ofC. cornuta were significantly decreased under fish predation
risk, although these responses were not significant underChaoborus predation risk. (3) General responses (size at maturity
and brood number): the size (CH vs. C: P = 0.007; F vs. C:P < 0.001) and brood number (CH vs. C: P =
0.006; F vs. C: P = 0.017) were significantly decreased under
both fish and Chaoborus predation risks. (4) Antagonistic
responses (size at 16th day): C. cornuta size increased underChaoborus predation risk (P = 0.034) but decreased under
fish predation risk (P < 0.001). Additionally, no
significant differences were observed in average brood size. Thus, the
unique responses in individual traits of C. cornuta to different
predators were dominant: unique responses (five traits) >
general responses (two traits) > antagonistic responses
(one trait).