EXG-09628; No of Pages 8 Experimental Gerontology xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Experimental Gerontology journal homepage: www.elsevier.com/locate/expgero
How perceived predation risk shapes patterns of aging in water fleas
2Q2
Barbara Pietrzak a,b,⁎, Piotr Dawidowicz b, Piotr Prędki b, Maciej J. Dańko a,⁎⁎
3 4
a
5
a r t i c l e
6 7 8 9 10 11 12
Article history: Received 15 January 2015 Received in revised form 19 April 2015 Accepted 14 May 2015 Available online xxxx
13 14 15 16 17 18
Keywords: Kairomone Extrinsic mortality Senescence Resource allocation Rate of aging
a b s t r a c t
R O
i n f o
O
Laboratory of Evolutionary Biodemography, Max Planck Institute for Demographic Research, Konrad-Zuse-Straße 1, 18057 Rostock, Germany Department of Hydrobiology, Institute of Zoology, Faculty of Biology, Biological and Chemical Research Centre, University of Warsaw, ul. Żwirki i Wigury 101, 02-089 Warsaw, Poland
Predation is an important selection pressure which shapes aging patterns in natural populations, and it is also a significant factor in the life history decisions of individuals. Exposure to the perceived threat of size-dependent fish predation has been shown to trigger adaptive responses in animal life history including an increase in early reproductive output. In water fleas, this response to perceived predation risk appears to have a cost, as a lifespan in an environment free of predation cues is 20% longer. The aim of this study is to establish the biodemographic basis of phenotypic differences in the water flea lifespan which are induced by the cues of fish predation. We examined mortality by fitting the Gompertz–Makeham model of mortality to large cohorts of two cladoceran species, Daphnia longispina and Diaphanosoma brachyurum. Our findings indicate that perceived exposure to the threat of fish predation (induced through chemical cues) only accelerated the rate of aging in Diaphanosoma, and not in Daphnia where the treatment led to an earlier onset of aging. The second of these two phenotypic responses is consistent with the genetically based differences between Daphnia from habitats that differ with respect to predation risk. In contrast, the response of Diaphanosoma demonstrates that the cue of extrinsic mortality—in this case, fish predation—is a key factor in shaping these cladoceran life histories in the wild, and is one of the few interventions which has been shown to induce a plastic change in the rate of aging. © 2015 Published by Elsevier Inc.
P
b
T
E
D
Section Editor: Diana Van Heemst
37 35 34
E
36
1. Introduction
39
Predation is one of the most frequently cited sources of extrinsic mortality shaping aging. In an early prediction, Williams (1957) stated that “low adult death rates should be associated with low rates of senescence, and high adult death rates with high rates of senescence.” Yet since the publication of a highly influential paper on this topic by Hamilton (1966), the validity of this simple prediction has been repeatedly challenged and theoretically debated (Abrams, 1993; Cichoń, 1997; Williams and Day, 2003; Reznick et al., 2004; Caswell, 2007; Dańko et al., 2012; Dańko and Kozłowski, 2012). Hamilton formalized the idea that the force of selection declines with age, which constitutes the core for three main theories of aging: (i) mutation accumulation theory (MA) assuming that senescence is a result of late acting deleterious mutations (Medawar, 1952, but see e.g. Dańko et al., 2012), (ii) antagonistic pleiotropy theory (AP) assuming that senescence is a result of genes with some early benefits, but later costs (Williams, 1957), and (iii) disposable soma theory (DS) assuming that senescence is a
48 49 50 51 52 53 54
R
N C O
46 47
U
44 45
R
38
42 43
19 20 21 22 23 24 25 26 27 28 29 30 31 32
C
33
40 41
F
1Q1
⁎ Correspondence to: B. Pietrzak, Department of Hydrobiology, Institute of Zoology, Faculty of Biology, Biological and Chemical Research Centre, University of Warsaw, ul. Żwirki i Wigury 101, 02-089 Warsaw, Poland. ⁎⁎ Corresponding author. E-mail addresses: [email protected] (B. Pietrzak), [email protected] (P. Dawidowicz), [email protected] (M.J. Dańko).
negative byproduct of an adaptive process constrained by tradeoffs (Kirkwood, 1977). The DS theory is sometimes considered as complementary to the AP, because it can model physiological basis of AP. The formalization proposed by Hamilton gives clear predictions about the effect of extrinsic mortality. Generally, in density independent population (or when density acts equally on survival of all ages) age independent extrinsic mortality should not affect the life histories and thus patterns of senescence (Abrams, 1993; Caswell, 2007). However, these two conditions are seldom met in nature — we should not expect that a population can grow to infinity or that the extrinsic mortality can only be age-independent. In light of recent theoretical advances (see e.g., Caswell, 2007) we expect that senescence should be promoted at least if extrinsic mortality increases with age even in density independent populations. Hence, according to life history theory, investments in later-life fitness components are futile when the probability of surviving to higher ages is low due to extrinsic mortality. Resources should then be allocated to enhance early rather than late reproduction and survival (Kirkwood, 1977; Stearns, 1992; Cichoń, 1997). Especially in an environment in which the risk of dying from predation increases with age, large investments in early reproduction and a short lifespan are expected to evolve. These shifts in life history have attracted both considerable theoretical attention (reviewed in Kozłowski, 2006), and have been studied on an evolutionary scale. The predicted differences in resource allocation patterns and the resulting life histories have been observed in natural populations in habitats along predation
http://dx.doi.org/10.1016/j.exger.2015.05.008 0531-5565/© 2015 Published by Elsevier Inc.
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
F
O
R O
101 102
P
99 100
D
97 98
T
95 96
C
93 94
E
91 92
R
89 90
R
87 88
O
86
C
84 85
later aging, and with changes in b as faster/more rapid/accelerated or slower aging. As the aging-related mortality pattern is subject to natural selection, we can expect to see tradeoffs emerge between survival at different ages and between other life history traits. For instance, resources spent on current reproduction limit potential future reproduction due to inadequate resources allocated to maintenance and repairs. Therefore, in an environment in which the chances of survival decrease with age, we would expect to see intensive, but short early investments in soma (Cichoń, 1997), followed by earlier (higher a) or accelerated (higher b) aging (Fig. 1). The aim of the present study is to examine the biodemographic basis of the previously observed lifespan reduction in water fleas (Cladocera: Crustacea) which were exposed to a perceived (but not real) fish predation risk (Dawidowicz et al., 2010). Fish are visual predators, and it is well documented that they exert in positively size-dependent predation in water flea populations (e.g. Ivlev, 1961; Brooks and Dodson, 1965). Thus, fish-related extrinsic mortality risk of these indeterminate growers increases with age (Gliwicz and Pijanowska, 1989). We hypothesized that among water fleas kept under the perceived threat of size-dependent fish predation, this perceived risk would have a little effect on the water fleas' mortality early in life, but would induce either an earlier (without a change in b) or/and a more rapid increase (with a change in b) in their mortality at later ages. This hypothesis was based on the assumption that unlike survival at later ages, early survival is under strong selection pressure. We also expected to see shifts in the mean age at reproduction to earlier ages. It would be adaptive for the fish-threatened cladocerans to maximize both fitness components—i.e., fecundity and survival—early in life, at the cost of their later decline. Here, we test these predictions in two cladoceran species, and investigate their patterns of aging with respect to a perceived threat of size-dependent predation, which we introduced by olfactory cues (comparisons 1 and 2, Table 1). In addition, we relate these phenotypic differences in aging patterns to previously published data by Dawidowicz et al. (2013) and Dudycha and Tessier (1999) on genotypic differences between individuals from populations experiencing different levels of extrinsic mortality in their habitats. These were Daphnia from deep dark strata (hypolimnion), where fish predation risk is low, compared to Daphnia dwelling in well-lit surface layers (epilimnion), where risk from these visual predators is high (comparison 3, Table 1); and Daphnia from permanent lakes, experiencing relatively low extrinsic mortality levels in their habitat, compared to Daphnia
N
82 83
gradients (e.g., Dudycha and Tessier, 1999; Rennie et al., 2010). However, these differences have rarely been investigated as a resource allocation problem at the level of the individual sensing the risk of predation. In previous research, the maximum, the median, or the average lifespan have been used to compare cohorts between experimental treatments, as the lifespan is the most easily obtainable proxy for the aging process (e.g. Martin et al., 2002). Yet, it documents only the final outcome of aging, while telling us little about the trajectory of aging itself (Kowald, 2002; Bonsall, 2006). There are different views on how demographic aging should be measured (see, e.g., Partridge and Barton, 1996; Bronikowski and Promislow, 2005). The most common approach is to fit a parametric model, whereby the parametric estimates of the model are used as tools for mortality comparisons (e.g., Finch, 1990; Pletcher, 1999). The most widely used class of parametric models is based on the Gompertz (1825) distribution. The basic Gompertz model assumes that mortality is an exponential function of age, and describes only the impact of age-dependent sources of mortality risk. The model is thus often extended by an age-independent term typically related to background mortality (Makeham, 1867), which is relieved from the impact of aging (e.g., Makeham, 1867; Golubev, 2004). Furthermore, Golubev (2004) has suggested that the Makeham term should not be ignored in the data analysis, not only because of its possible biological significance, but also because this could bias estimates of Gompertz parameters. The Gompertz–Makeham model assumes that there is an exponential increase in age-dependent mortality μ(t). Rather than employing the notation of many theoretical papers, we have chosen to use a representation typical of many statistical packages (e.g., Colchero et al., 2012), which takes the form: μ(t) = e a + bt + c; where t denotes age, the exponent of the parameter a represents the initial intrinsic agedependent mortality, b represents the Gompertz exponential coefficient (defined hereafter as the rate of aging), and parameter c is the abovementioned Makeham term. Such a representation reduces the bias in the estimation of confidence intervals for a parameter (personal observation). The Gompertz–Makeham model describes mortality in many populations quite well. But while this model is both popular and useful, it is important to note its limitations, including the fact that it is empirically driven rather than strongly rooted in the biological theory (but see, e.g., Strehler and Mildvan, 1960; Golubev, 2009). In the Gompertz–Makeham setting, a shorter lifespan can result from higher initial intrinsic mortality (exp a), more rapid aging (b), higher ageindependent mortality (c), or a combination thereof. Throughout the text we will refer to changes associated with changes in a as earlier or
U
80 81
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
E
2
Fig. 1. Gompertz–Makeham hazard functions illustrating: (A) earlier/later emergence of aging modeled by changes in a; (B) more rapid/slower aging modeled by changes in b; (C) higher/ lower levels of age-independent background mortality modeled by changes in c.
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
Table 1 List of comparisons made. In each of the four comparisons, either earlier or faster aging is expected in the second group in pair. Comparisons 1 and 2 deal with phenotypic change in response to olfactory cue of predation, comparisons 3 and 4 deal with genetic differences between animals from habitats of different predation regimes. References in text.
Daphnia kairomone 2 3 4
Diaphanosoma control Diaphanosoma kairomone Daphnia from hypolimnion Daphnia from epilimnion Daphnia pulicaria Daphnia pulex
Phenotypic plasticity, direct effect of environment Intraspecies differences, indirect effect of environment (same experimental conditions) Interspecies differences, indirect effect of environment (same experimental conditions)
168
2. Material and methods
169
2.1. Animals and life table experiment
170
188 189
Daphnia longispina and Diaphanosoma brachyurum (Diaphanosoma hereafter), cladocerans of two common temperate lake species, were used in the experiments. We used standard, constant laboratory conditions, according to a well-established protocol in Daphnia life table experiments, i.e. summer photoperiod (16L:8D), temperature of 20 °C, green alga Scenedesmus obliquus as food source, filtered aged lake water as culture medium, which was pre-exposed to single fish (crucian carp Carassius auratus) for 24 h and refiltered before use to obtain olfactory cues of predation (kairomone). We followed 210 individuals of each species either in the absence or in the presence of the chemical cues of fish predation (5 individuals × 6 experimental vessels × 7 strains × 2 species × 2 treatments = 840 individuals). Offspring were counted daily. As the experiment was originally designed to test for differences in median lifespan, the individuals were followed from their births until more than one-half of the individuals in an experimental vessel had died; thus, the numbers of censored individuals varied between vessels. Each of the four datasets was obtained by pooling mortality data for seven homogenous clonal cohorts. The origin of the animals used and the experimental protocol were described in detail in Dawidowicz et al. (2010).
190
2.2. Statistical analysis
191
The survival data were analyzed using a Kaplan–Meier estimator and were fitted with Gompertz and Gompertz–Makeham models using a maximum likelihood estimation accounting for type II rightcensored data where needed. To determine which of these two models was more likely, we used a likelihood ratio test. Typically, under a likelihood ratio test a pair of models is compared, with one model being nested within the other. In our case, the Gompertz was a null (nested, c = 0) model, while the Gompertz–Makeham was an alternative (unrestricted, c ≥ 0) model. Because the null hypothesis c = 0 lies at the boundary of parameter space, the asymptotic null distribution of the likelihood ratio test statistic in this case is a mixture of χ20 (zero degrees of freedom) and χ21 (one degree of freedom) distributions with equal weights of 0.5. In this case the p-value is calculated according to the equation:
185 186 187
192 193 194 195 196 197 198 199 200 201 202 203 204
C
E
183 184
R
181 182
R
179 180
N C O
177 178
U
175 176
p‐value ¼ 0:5P χ 20 N R þ 0:5P χ 21 N R ; 206 207 208
Later reproduction and at larger size, longer lifespan Earlier reproduction and at smaller size, shorter lifespan Later reproduction and longer lifespan Earlier reproduction and shorter lifespan Later reproduction and longer lifespan Earlier reproduction and shorter lifespan Later reproduction and longer lifespan Earlier reproduction and shorter lifespan
No predation cue Size dependent (fish) predation olfactory cue Safe from fish predation Exposed to fish predation Permanent; low extrinsic mortality Temporary; high extrinsic mortality
T
from temporary ponds, experiencing high extrinsic mortality levels (Dudycha, 2004) (comparison 4, Table 1).
173 174
Known differences
No predation cue Size dependent (fish) predation olfactory cue
166 167
171 172
Environment
F
t2:7 t2:8 t2:9 t2:10 t2:11 t2:12
Phenotypic plasticity, direct effect of environment
where P denotes probability and R is a likelihood ratio test statistic (see, e.g., Verbeke and Molenberghs, 2000 for details). We avoided comparisons of parameter estimates of populations fitted with two different models by choosing the same model for the two compared populations.
We followed a simple rule: the Gompertz model was chosen only if a likelihood ratio test supported the null hypothesis in both populations. This approach allowed us to compare differences in c even when the Gompertz model was more likely for one population. This can also be seen as a conservative approach because the Gompertz–Makeham model is considered the primary model (Golubev, 2004) in most situations. Golubev (2004) showed that ignoring the Makeham (c) term may lead to a bias in the estimation of Gompertz parameters (a and b). Furthermore, Pletcher (1999) showed that the likelihood ratio test may be more likely to choose the overly simplistic Gompertz model for small and average sample sizes. In light of these considerations, we decided to set significance threshold α for all of the likelihood ratio tests to 0.1, which made it easier to reject the null hypothesis (i.e., Gompertz) and accept the alternative hypothesis (i.e., Gompertz–Makeham). The statistical inference on the differences between the parameter estimates, the survivorship, and the ln-mortality curves was based on specific confidence intervals (CI). The confidence intervals of the estimated Gompertz/Gompertz–Makeham parameters were calculated using a BCa method (e.g., Efron and Tibshirani, 1993, p. 184–188) from a sample of 250000 parametric bootstrap replicates, and by taking into account multi-vessel type II censored data (e.g., Kateri and Balakrishnan, 2008). The confidence intervals of the estimated survivorship and the ln-mortality curves were calculated analytically based on the sandwich estimator (e.g., Efron and Tibshirani, 1993, p. 310–311). The significance threshold for all tests, except the likelihood ratio test, was set to α = 0.05. To test for age-specific differences in fecundity between treatments, we performed ANOVA on the mean age at reproduction for replicate vessels. The mean age at reproduction was calculated as: TM = Σ m(t) × t / Σ m(t), where t is age and m(t) is age-specific fecundity.
209 210
2.3. Comparisons
239
First, we determined whether the Gompertz model or the Gompertz–Makeham model better describes the mortality patterns in the pairs of compared populations using the procedure described above (Section 2.2). We started with comparisons of the responses of D. longispina and of Diaphanosoma to a kairomone treatment and to a control treatment. Next, we compared the aging patterns in these two species, as Diaphanosoma inhabits the risky surface layers, or epilimnion; while D. longispina migrates to the hypolimnetic refuge during the daytime. We then added comparisons of epilimnetic to hypolimnetic D. longispina, and Daphnia pulex to Daphnia pulicaria from earlier studies. Finally, we compared the phenotypic life history shifts which emerge when D. longispina perceived cues of predation with the previously published data on the differences observed on an evolutionary scale between Daphnia spp. originating from high and low risk habitats (Dudycha and Tessier, 1999; Dawidowicz et al., 2013). Dudycha and Tessier (1999) compared the aging profiles of D. pulicaria, a species from lakes, which are permanent habitats and where Daphnia mortality rates are relatively low; and of D. pulex, a species from temporary ponds,
240
O
t2:6
Daphnia control
R O
1
P
Comparison
t2:5
D
t2:4
E
t2:1 t2:2 t2:3
3
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238
241 242 243 244 245 246 247 248 249 Q3 250 251 252 253 254 255 256 257
3.1. A comparison of the perceived predation and the control treatments
269
We found that the Gompertz model was strongly supported only in D. brachyurum: the p-values of the likelihood ratio test for both treatments were much greater than the assumed significance level of 0.1 (both were close to 0.5), and the estimated c parameters under the Gompertz–Makeham model equaled zero. For the D. longispina, the Gompertz model was strongly supported only for the kairomone treatment (p-value close to 0.5, c = 0), but for the control group the Gompertz–Makeham model performed better (p-value = 0.057). Consequently, we chose to use the Gompertz–Makeham model for both treatments of D. longispina. The survival analysis showed that the survivorship patterns differed significantly between the treatments in later life in both of the species studied. Survivorship was lower in animals kept in kairomone-treated water beginning on day 29 in D. longispina, and beginning on day 17 in D. brachyurum (Fig. 2A & B).
T C E
282 283
R
280 281
R
278 279
O
276 277
C
274 275
N
272 273
U
270 271
F
268
264
O
3. Results
262 263
R O
267
260 261
Consequently, the predation cue affected the mortality trajectories in both species. Under the anticipated predation threat, the mortality hazard was equal to or lower than that of the control in early life, but it was significantly higher beginning on day 20 in D. longispina and on day 12 in Diaphanosoma. Fig. 3A and B illustrates the mortality hazard rate of the studied cladocerans. The estimated parameters of the fitted Gompertz–Makeham model and their confidence intervals are presented in Fig. 4 and summarized in Appendix Table 1. The comparison of the confidence intervals for the estimated parameters revealed different patterns of mortality response in the two studied species. In D. longispina, the phenotypic change in age-specific mortality could not be clearly attributed to a specific and significant change in any combination of fitted parameters (Fig. 4). In Diaphanosoma, the change in the mortality pattern was clearly related to the lower values of a and the higher values of b in the kairomone treatment than in the control treatment (intrinsic mortality rate exp a was three times lower, and the rate of aging b was 2.5 higher in kairomone than in control treatment), which indicates more rapid aging (Fig. 4). Through kairomone exposure, the average age of first reproduction was shifted 20% earlier in D. longispina and 27% earlier in Diaphanosoma (Dawidowicz et al., 2010). During the first 10 days of life, D. longispina produced on average 1.8 times fewer offspring in the control than in the kairomone treatment, and Diaphanosoma produced 3.2 times fewer offspring (Dawidowicz et al., 2010). Thus, in the fish treatment overall fertility was shifted toward earlier ages in both species, with the mean age at reproduction TM being decreased by 29% in D. longispina (from 21.2 ± 0.6 days to 14.9 ± 0.6 days (mean ± SD), ANOVA: F1, 76 = 114.1, p b 0.0001) and by 18% in Diaphanosoma (from 13.5 ± 0.5 days to 11.1 ± 0.5 days, F1, 76 = 20.37, p b 0.0001).
E
265 266
where environmental conditions vary greatly and where Daphnia experience high mortality levels (Dudycha, 2004). Dawidowicz et al. (2013) compared the aging profiles of D. longispina which are able to find a refuge from visual predators such as fish by migrating during the day to the dark deep hypolimnion, with the profiles of D. longispina which live constantly in the well-lit epilimnion. Thus, to the best of our knowledge, we analyzed all of the published water flea survival data dealing with the indirect effects of predation, that have sufficient age spans and resolution.
P
258 259
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
D
4
Fig. 2. Phenotypic (top panels) and genetic (bottom panels) differences in cladoceran survivorships. Top panels: (A) Daphnia longispina, and (B) Diaphanosoma brachyurum reared in control and predation cue treatment — kairomone. Bottom panels: (C) Daphnia longispina originating from low (hypolimnion) and high (epilimnion) fish predation risk habitats (data from Dawidowicz et al., 2010), and (D) D. pulicaria from lakes and D. pulex from ponds (data from Dudycha and Tessier, 1999). Blue symbols — control treatment or animals originating from low risk habitats; Red symbols — predation cue treatment or animals originating from high risk habitat. Dots — Kaplan–Meier estimator; lines — survivorship curves based on fitted Gompertz– Makeham model with CI calculated from Fischer information matrix. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
5
D
P
R O
O
F
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
U
N C O
R
R
E
C
T
E
Fig. 3. Phenotypic (top panels) and genetic (bottom panels) differences in cladoceran mortality. Panels and color symbols as in Fig. 2. Dots — empirical ln-mortality; lines — ln-hazard curves based on fitted Gompertz–Makeham model with CI calculated from Fischer information matrix. Parameters of the best fit models are given in Fig. 4 and Appendix Table 1.
Fig. 4. Estimated parameters of the Gompertz–Makeham model for different experiments. Confidence intervals calculated by BCa method from a sample of 250 000 parametric bootstrap estimates.
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
356
In this study we showed that the anticipated risk of death can trigger phenotypic changes in aging. More specifically, we showed how olfactory cues of age-dependent predation affect the trajectory of the aging process in an organism capable of differential resource allocation. Two freshwater microcrustaceans Daphnia and Diaphanosoma were chosen for our analysis, as predation is one of the strongest forces shaping water flea life histories in the wild (Lynch, 1980). Positively sizedependent fish predation is believed to be the major source of mortality in large-bodied cladocerans in lakes ahead of starvation, parasitism, and aging (Gliwicz and Pijanowska, 1989). When we look at populations from habitats with varying risks of fish predation on an evolutionary scale, we can see that the lifespan of an individual Daphnia is shorter and its reproduction is more likely to be shifted toward earlier ages in a high-risk habitat than in a low-risk habitat (Dudycha and Tessier, 1999; Dawidowicz et al., 2013). Our study showed for two species of freshwater cladocerans how the perceived predation risk modulates the phenotypic aging profiles. Olfactory cues of size-dependent predation triggered either earlier or accelerated aging, respective of species. This was coupled with shifts in the mean age at reproduction to earlier ages, all in accordance with present theories and our predictions.
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
C
333 334
E
331 332
R
329 330
R
327 328
O
325 326
C
323 324
N
321 322
U
319 320
F
4. Discussion
318
O
355
316 317
R O
354
We found that the Gompertz model was preferable only in the hypo/ epi-limnion group: the p-values for the likelihood ratio tests were high (0.355 and 0.449, respectively). In the Daphnia pulex/pulicaria group, the effect of parameter c was significant for the D. pulicaria population (the p-value of the likelihood ratio test was close to zero). Thus, because it was found to be the better model, the Gompertz–Makeham model was chosen for the whole group. Between-population comparisons revealed that the estimated late life survivorship was always lower in animals from habitats with a higher extrinsic mortality risk. It was significantly lower in nonmigrating D. brachyurum than in D. longispina beginning on day 19 in the control treatment, and on day 15 in the kairomone treatment (Fig. 2A–B) (it should be noted, however, that the populations were fitted with two different models: the Gompertz and the Gompertz– Makeham, respectively). Beginning on day 5, survivorship was significantly lower in the epilimnion than in the hypolimnion population of D. longispina studied by Dawidowicz et al. (2013) (Fig. 2C); and beginning on day 41, survivorship was lower among the pond D. pulex than among the lake D. pulicaria studied by Dudycha and Tessier (1999) (Fig. 2D). Beginning at age five, mortality in the epilimnion population of D. longispina was higher than in the hypolimnion population at all ages (Fig. 3C–D). A comparison of the confidence intervals for the estimated parameters shows that there were significant differences in both Gompertz parameters. Epilimnetic Daphnia had exp a that was twice as high, as well as a b parameter that was twice as high as that of Daphnia from hypolimnion (Fig. 4). The differences in the mortality patterns found among the pond D. pulex and the lake D. pulicaria could be attributed to the higher values of the a parameter and the lower values of the c parameter in D. pulex than in D. pulicaria. The postponement of aging (without a significant change in b) in D. pulicaria was clearly observed in this case. This significant delay appears to have been driven by an initial intrinsic mortality exp a which was found to be about 40 000 times lower in the species from the less risky habitat (Figs. 3 and 4). The average age of first reproduction was about 20% earlier in the epilimnetic than in the hypolimnetic D. longispina. Moreover, during the first 10 days of their lives, the epilimnetic Daphnia produced 2.1 times more offspring per capita than their hypolimnetic conspecifics. Both the age of first reproduction and the mean age at reproduction were also earlier in D. pulex than in D. pulicaria.
P
314 315
Water flea mortality caused by fish predation increases with size, and with age. Under such selection, the optimal strategy is to allocate resources into early reproduction and defense at the cost of somatic repair and later maintenance. Such shifts toward the earlier onset of reproduction and higher early fecundity are well documented under conditions in which Daphnia is exposed to chemical cues of fish presence in the absence of actual predation events (reviewed in Lampert, 2011). How these strategy shifts might be associated with late life costs has only recently been demonstrated by Dawidowicz et al. (2010). The authors found that the median lifespan of cladocerans of the two genera was shortened by about 20% under the conditions of perceived fish predation. Here, using the Gompertz/Gompertz–Makeham model as a method to study the changes in mortality schedules in response to a predator cue, we showed that the shortened lifespan of Daphnia could be attributed to an increased Gompertz's a rather than to an increased rate of aging b, although none of these results were significant, probably due to colinearities between the Gompertz parameters and the Makeham term. This finding is intuitively and visually confirmed in Fig. 3, in which both ln-mortality curves are parallel at the oldest ages. It is therefore clear that we are observing earlier and later onsets of aging, rather than accelerated/decelerated aging. In contrast, in Diaphanosoma, the shortening of the lifespan was found to be due to a more than twofold increase in the rate of aging (b), despite the lower a. Here, both mortality curves clearly cross and then diverge with age (Fig. 3). The threatened Diaphanosoma thus appears to provide a clear example of accelerated aging. Accelerated aging has been previously observed under some commonly used environmental interventions, yet to our knowledge it has not been observed in association with perceived predation risk. The modulation of aging trajectories realized within the phenotypic plasticity of a genotype has been well documented in some organisms, like Caenorhabditis elegans. For example, dietary restriction has been shown to increase lifespan by lowering the slope of the mortality increase with age (Lenaerts et al., 2007); and mild heat shocks have been found to have similar effects (Wu et al., 2009). The nematodes have also been found to have more rapid increases in mortality rates when exposed to pathogens (Baeriswyl et al., 2010). Moreover, sensory perception has been shown to regulate the lifespan of this animal (Apfeld and Kenyon, 1999), and in Drosophila melanogaster exposure to olfactory cues of food presence shortened fly lifespan (Libert et al., 2007). This is comparable to our findings on water flea earlier or accelerated aging under olfactory cues of predation risk. Again in agreement, visual and chemical predator cues were previously shown to cause oxidative damage in prey damselfly larvae (Janssens and Stoks, 2013). Aging is often understood as the interaction between the accumulation of damage, and maintenance or repair (e.g. Strehler and Mildvan, 1960; Cichoń, 1997). Some damage can be repaired, other is irreversible, and remains in for the rest of life of the individual (e.g. Rübe et al., 2011). The irreversible damage in particular can compromise the repair mechanism itself and contribute to the increase in the rate of damage accumulation in the future. There is still a debate on how the process of damage accumulation converts to mortality. Theoretically, two factors seem important: (i) the organism's ability to reduce the rate of accumulation of irreversible damage, and (ii) its physiological response to already accumulated damage. The first depends on the effectiveness of repair mechanisms, as well as on the deterioration of these mechanisms with damage or age. Hereby, the rate at which deterioration occurs is determined by resources spent on damage protection, and resources that are invested into repair. The amount of resources available depends on the specific resource allocation strategy of the organism, and physiological constraints of the body. This should clearly affect the rate of aging (e.g. Gompertz's b). The second factor describes an organism's vulnerability, or the physiological ability to tolerate the damage. Specifically, it can be understood as the rate at which accumulated damage is translated into mortality. This rate is constant throughout the lifespan, and can be interpreted as Gompertz's a. In consequence, different levels of
D
3.2. Comparisons of populations from high and low risk habitats
T
313
E
6
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
501 502 503 504 505
C
461 462
E
459 460
R
512
The research was partially supported by Max Planck Institute for Demographic Research to Maciej J. Dańko. The experimental part of the study was supported by a grant from the Polish Ministry of Science and Higher Education No. N304 005 32/0647 to Barbara Pietrzak. Jutta Gampe, Alexander Scheuerlein, Jan Kozłowski, Hal Caswell, Daniel Levitis, reviewer Jeff Dudycha and another anonymous reviewer provided valuable comments which helped to improve the manuscript.
513 Q4
References
520
F
457 458
R
455 456
N C O
453 454
U
451 452
Acknowledgments
O
The present findings offer new insights into the effects on the aging process of the perceived risk of earlier death associated with predation threat. Cladocerans of two genera respond to the perceived risk of size-dependent predation with either more rapid or earlier aging. This phenotypic effect is consistent with differences observed between populations from risky and safe habitats. Finally, as the studied species
449 450
R O
500
448
506 507
Abrams, P.A., 1993. Does increased mortality favor the evolution of more rapid senescence? Evolution 47, 877–887. Apfeld, J., Kenyon, C., 1999. Regulation of lifespan by sensory perception in Caenorhabditis elegans. Nature 402, 804–809. Baeriswyl, S., Diard, M., Mosser, T., Leroy, M., Manière, X., Taddei, F., Matic, I., 2010. Modulation of aging profiles in isogenic populations of Caenorhabditis elegans by bacteria causing different extrinsic mortality rates. Biogerontology 11, 53–65. Bonsall, M.B., 2006. Longevity and ageing: appraising the evolutionary consequences of growing old. Philos. Trans. R. Soc. Lond. B Biol. Sci. 361, 119–135. Bronikowski, A.M., Promislow, D.E.L., 2005. Testing evolutionary theories of aging in wild populations. Trends Ecol. Evol. 20, 271–273. Brooks, J.L., Dodson, S.I., 1965. Predation, body size, and composition of plankton. Science 150, 28–35. Caswell, H., 2007. Extrinsic mortality and the evolution of senescence. Trends Ecol. Evol. 22, 173–174. Cichoń, M., 1997. Evolution of longevity through optimal resource allocation. Proc. R. Soc. B 264, 1383–1388. Colchero, F., Jones, O.R., Rebke, M., 2012. BaSTA: an R package for Bayesian estimation of age-specific survival from incomplete mark-recapture/recovery data with covariates. Methods Ecol. Evol. 3, 466–470. Dańko, M.J., Kozłowski, J., 2012. Mutation accumulation may only be a minor force in shaping life-history traits, even when reproduction is sexual. PLoS ONE 7 (4), e48302. Dańko, M.J., Kozłowski, J., Vaupel, J.W., Baudisch, A., 2012. Mutation accumulation may be a minor force in shaping life history traits. PLoS ONE 7 (4), e34146. Dawidowicz, P., Prędki, P., Pietrzak, B., 2010. Shortened lifespan — another cost of predator avoidance in cladocerans? Hydrobiologia 643, 27–32. Dawidowicz, P., Prędki, P., Pietrzak, B., 2013. Depth selection behavior and longevity in Daphnia: an evolutionary test for the predation-avoidance hypothesis. Hydrobiologia 715, 87–91. Dudycha, J.L., 2004. Mortality dynamics of Daphnia in contrasting habitats and their role in ecological divergence. Freshw. Biol. 49, 505–514. Dudycha, J.L., Tessier, A.J., 1999. Natural genetic variation of life span, reproduction, and juvenile growth in Daphnia. Evolution 53, 1744–1756. Efron, B., Tibshirani, R.J., 1993. An Introduction to the Bootstrap. Chapman and Hall, New York. Finch, C.E., 1990. Longevity, Senescence, and the Genome. University of Chicago Press, Chicago, London. Gliwicz, Z.M., Pijanowska, J., 1989. The role of predation in zooplankton succession. In: Sommer, U. (Ed.), Plankton Ecology: Succession in Plankton Communities. Springer-Verlag, New York, pp. 253–295. Golubev, A., 2004. Does Makeham make sense? Biogerontology 5, 159–167. http://dx.doi. org/10.1023/B:BGEN.0000031153.63563.58. Golubev, A., 2009. How could the Gompertz–Makeham law evolve. J. Theor. Biol. 238, 1–17. http://dx.doi.org/10.1016/j.jtbi.2009.01.009. Gompertz, B., 1825. On the nature of the function expressive of the law of human mortality, and on the mode of determining the value of life contingencies. Philos. Trans. R. Soc. 115, 513–585. Hamilton, W.D., 1966. The moulding of senescence by natural selection. J. Theor. Biol. 12, 12–45. Ivlev, V.S., 1961. Experimental Ecology of the Feeding of Fishes. Yale University Press, New Haven. Janssens, L., Stoks, R., 2013. Predation risk causes oxidative damage in prey. Biol. Lett. 9, 20130350. Kateri, M., Balakrishnan, N., 2008. Inference for a simple step-stress model with Type-II censoring and Weibull distributed lifetimes. IEEE Trans. Reliab. 57, 616–626. Kirkwood, T.B.L., 1977. Evolution of ageing. Nature 270, 301–304. Kowald, A., 2002. Lifespan does not measure ageing. Biogerontology 3, 187–190. Kozłowski, J., 2006. Why life histories are diverse. Pol. J. Ecol. 54, 585–605. Lampert, W., 2011. Daphnia: development of a model organism in ecology and evolution. In: Kinne, O. (Ed.), Excellence in Ecology. Book 21. International Ecology Institute, Oldendorf/Luhe.
P
5. Conclusions
446 447
normally experience the presence of fish and their chemical cues in natural habitats, the problem at hand should be studied in the reverse scenario: i.e., whether senescence slows down or is delayed when there is no perceived fish predation threat. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.exger.2015.05.008.
D
499
444 445
T
498
vulnerability to damage should generate different levels of mortality even at the same level of damage (see also Partridge et al., 2005 and Simons et al., 2013). Ecology of a species is expected to lead to evolution of different mechanisms of resource allocation and physiology of damage accumulation, repair and resistance. Cladocerans from the two studied genera differ in terms of their life histories, their behavioral strategies, and their habitats. D. brachyurum stays in surface waters that are well lit during the day, where the risk associated with visual predators, such as fish, is higher. These surface waters (epilimnion) are also warm and abundant in food, and thus enable faster growth and development. In contrast, D. longispina is mostly a migratory species. It usually stays in the high-risk, high-food surface waters during the night only, when fish predation risk is low, while migrating to relatively safe deeper strata during the day. Not surprisingly, Diaphanosoma, the species which on average faces a higher predation risk, was shown here to have a higher rate of aging than D. longispina when compared in the lab under a lakeimitating fish cue presence; i.e., in the kairomone treatment (a more than twofold difference in b was found between the species, but it is important to note that they were fitted with different models; namely, the Gompertz and the Gompertz–Makeham models, respectively). The extraordinary shifts in the Diaphanosoma's life history toward earlier reproduction and more rapid aging could be interpreted as follows: the earlier reproduction is an anti-predator defense alternative to escape behavior, while the accelerated aging is the cost thereof, as the strength of the life history response seems to correlate negatively with the efficiency of escape behavior (Pietrzak et al., unpublished data; see also Thys and Hoffmann, 2005; Vijverberg et al., 2006). Thus, the lower rates of fitted mortality observed early in life in the kairomone-treated Diaphanosoma may be a sign that a trade-off takes place between early and late fitness components. The enhancement of early survival under the perceived threat would surely be adaptive if more females survived until the first reproduction. In addition, it could be explained within the framework of the “maintenance gap” hypothesis if the observed early life response with greater reproductive output increases the maintenance requirements of the animal (Wensink et al., 2012). Our between-habitat comparisons revealed that late-life survivorship was always lower in animals from mortality-riskier habitats: specifically, it was lower in the non-migrating D. brachyurum as compared to the migrating D. longispina; it was lower in the population of that latter species from the well-lit epilimnion than in the population from the deep, dark hypolimnion; and it was lower in D. pulex from temporary ponds than in D. pulicaria from lakes. Moreover, shifts in agespecific fecundity—measured as the age at first reproduction, the early reproductive effort, or the mean age at reproduction—were consistent across comparisons. Finally, we found that the presence or the lack of kairomone may have had an effect on age-independent mortality (c). The significance of this mortality parameter was only clear in the D. pulicaria cohorts when it was compared to that of the D. pulex cohorts (Dudycha and Tessier, 1999), and when it masked the lower degree of frailty of D. pulicaria, which delayed senescence in this species. The perception of risk might result in decreased vulnerability to external ageindependent causes of death due to, for example, risk avoidance behavior, which may be expressed even in protected environments. Thus, the level of age-independent mortality could be under the control of the organism to some extent; this is an issue which deserves further study.
E
442 443
7
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
508 509 510 511
514 515 516 517 518 519
521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581
F
Simons, M.J.P., Koch, W., Verhulst, S., 2013. Dietary restriction of rodents decreases aging rate without affecting initial mortality rate — a meta-analysis. Aging Cell 12, 410–414. Stearns, S.C., 1992. The Evolution of Life Histories. Oxford University Press, New York. Strehler, B.L., Mildvan, A.S., 1960. General theory of mortality and aging. Science 132, 14–21. Thys, I., Hoffmann, L., 2005. Diverse responses of planktonic crustaceans to fish predation by shifts in depth selection and size at maturity. Hydrobiologia 551, 87–98. Verbeke, G., Molenberghs, G., 2000. Linear mixed models for longitudinal data. Springer Series in Statistics. Springer, New York. Vijverberg, J., Doksæter, A., Van Donk, E., 2006. Contrasting life history responses to fish released infochemicals of two co-occurring Daphnia species that show different migration behaviour. Arch. Hydrobiol. 167, 89–100. Wensink, M.J., van Heemst, D., Rozing, M.P., Westendorp, R.G., 2012. The maintenance gap: a new theoretical perspective on the evolution of aging. Biogerontology 13, 197–201. Williams, G.C., 1957. Pleiotropy, natural selection and the evolution of senescence. Evolution 11, 398–411. Williams, P.D., Day, T., 2003. Antagonistic pleiotropy, mortality source interactions, and the evolutionary theory of senescence. Evolution 57, 1478–1488. Wu, D., Rea, S.L., Cypser, J.R., Johnson, T.E., 2009. Mortality shifts in Caenorhabditis elegans: remembrance of conditions past. Aging Cell 8, 666–675.
N
C
O
R
R
E
C
T
E
D
P
R O
Lenaerts, I., van Eigen, S., van Fleteren, J., 2007. Adult-limited dietary restriction slows gompertzian aging in Caenorhabditis elegans. Ann. N. Y. Acad. Sci. 1100, 442–448. Libert, S., Zwiener, J., Chu, X., VanVoorhies, W., Roman, G., Pletcher, S.D., 2007. Regulation of Drosophila life span by olfaction and food-derived odors. Science 315, 1133–1137. Lynch, M., 1980. The evolution of cladoceran life histories. Q. Rev. Biol. 55, 23–42. Makeham, W.M., 1867. On the law of mortality. J. Inst. Actuaries 13, 325–358. Martin, L.J., Mahaney, M.C., Bronikowski, A.M., Dee Carey, K., Dyke, B., Comuzzie, A.G., 2002. Lifespan in captive baboons is heritable. Mech. Ageing Dev. 123, 1461–1467. Medawar, P.B., 1952. An Unsolved Problem in Biology. HK Lewis, London. Partridge, L., Barton, N.H., 1996. On measuring the rate of ageing. Proc. R. Soc. B 263, 1365–1371. Partridge, L., Pletcher, S.D., Mair, W., 2005. Dietary restriction, mortality trajectories, risk and damage. Mech. Ageing Dev. 126, 35-4. Pletcher, S.D., 1999. Model fitting and hypothesis testing for age-specific mortality data. J. Evol. Biol. 12, 430–439. Rennie, M.D., Purchase, C.F., Shuter, B.J., Collins, N.C., Abrams, P.A., Morgan, G.E., 2010. Prey life-history and bioenergetic responses across a predation gradient. J. Fish Biol. 77, 1230–1251. Reznick, D.N., Bryant, M.J., Roff, D., Ghalambor, C.K., Ghalambor, D.E., 2004. Effect of extrinsic mortality on the evolution of senescence in guppies. Nature 431, 1095–1099. Rübe, C.E., Fricke, A., Widmann, Y.A., Fürst, T., Madry, H., Pfreundschuh, M., Rübe, C., 2011. Accumulation of DNA damage in hematopoietic stem and progenitor cells during human aging. PLoS ONE 6.
U
582 583 584 585 586 587 588 589 590 591 592 593 Q5 594 595 596 597 598 599 600 601 602 603 604
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
O
8
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625
Contents lists available at ScienceDirect
Experimental Gerontology journal homepage: www.elsevier.com/locate/expgero
How perceived predation risk shapes patterns of aging in water fleas
2Q2
Barbara Pietrzak a,b,⁎, Piotr Dawidowicz b, Piotr Prędki b, Maciej J. Dańko a,⁎⁎
3 4
a
5
a r t i c l e
6 7 8 9 10 11 12
Article history: Received 15 January 2015 Received in revised form 19 April 2015 Accepted 14 May 2015 Available online xxxx
13 14 15 16 17 18
Keywords: Kairomone Extrinsic mortality Senescence Resource allocation Rate of aging
a b s t r a c t
R O
i n f o
O
Laboratory of Evolutionary Biodemography, Max Planck Institute for Demographic Research, Konrad-Zuse-Straße 1, 18057 Rostock, Germany Department of Hydrobiology, Institute of Zoology, Faculty of Biology, Biological and Chemical Research Centre, University of Warsaw, ul. Żwirki i Wigury 101, 02-089 Warsaw, Poland
Predation is an important selection pressure which shapes aging patterns in natural populations, and it is also a significant factor in the life history decisions of individuals. Exposure to the perceived threat of size-dependent fish predation has been shown to trigger adaptive responses in animal life history including an increase in early reproductive output. In water fleas, this response to perceived predation risk appears to have a cost, as a lifespan in an environment free of predation cues is 20% longer. The aim of this study is to establish the biodemographic basis of phenotypic differences in the water flea lifespan which are induced by the cues of fish predation. We examined mortality by fitting the Gompertz–Makeham model of mortality to large cohorts of two cladoceran species, Daphnia longispina and Diaphanosoma brachyurum. Our findings indicate that perceived exposure to the threat of fish predation (induced through chemical cues) only accelerated the rate of aging in Diaphanosoma, and not in Daphnia where the treatment led to an earlier onset of aging. The second of these two phenotypic responses is consistent with the genetically based differences between Daphnia from habitats that differ with respect to predation risk. In contrast, the response of Diaphanosoma demonstrates that the cue of extrinsic mortality—in this case, fish predation—is a key factor in shaping these cladoceran life histories in the wild, and is one of the few interventions which has been shown to induce a plastic change in the rate of aging. © 2015 Published by Elsevier Inc.
P
b
T
E
D
Section Editor: Diana Van Heemst
37 35 34
E
36
1. Introduction
39
Predation is one of the most frequently cited sources of extrinsic mortality shaping aging. In an early prediction, Williams (1957) stated that “low adult death rates should be associated with low rates of senescence, and high adult death rates with high rates of senescence.” Yet since the publication of a highly influential paper on this topic by Hamilton (1966), the validity of this simple prediction has been repeatedly challenged and theoretically debated (Abrams, 1993; Cichoń, 1997; Williams and Day, 2003; Reznick et al., 2004; Caswell, 2007; Dańko et al., 2012; Dańko and Kozłowski, 2012). Hamilton formalized the idea that the force of selection declines with age, which constitutes the core for three main theories of aging: (i) mutation accumulation theory (MA) assuming that senescence is a result of late acting deleterious mutations (Medawar, 1952, but see e.g. Dańko et al., 2012), (ii) antagonistic pleiotropy theory (AP) assuming that senescence is a result of genes with some early benefits, but later costs (Williams, 1957), and (iii) disposable soma theory (DS) assuming that senescence is a
48 49 50 51 52 53 54
R
N C O
46 47
U
44 45
R
38
42 43
19 20 21 22 23 24 25 26 27 28 29 30 31 32
C
33
40 41
F
1Q1
⁎ Correspondence to: B. Pietrzak, Department of Hydrobiology, Institute of Zoology, Faculty of Biology, Biological and Chemical Research Centre, University of Warsaw, ul. Żwirki i Wigury 101, 02-089 Warsaw, Poland. ⁎⁎ Corresponding author. E-mail addresses: [email protected] (B. Pietrzak), [email protected] (P. Dawidowicz), [email protected] (M.J. Dańko).
negative byproduct of an adaptive process constrained by tradeoffs (Kirkwood, 1977). The DS theory is sometimes considered as complementary to the AP, because it can model physiological basis of AP. The formalization proposed by Hamilton gives clear predictions about the effect of extrinsic mortality. Generally, in density independent population (or when density acts equally on survival of all ages) age independent extrinsic mortality should not affect the life histories and thus patterns of senescence (Abrams, 1993; Caswell, 2007). However, these two conditions are seldom met in nature — we should not expect that a population can grow to infinity or that the extrinsic mortality can only be age-independent. In light of recent theoretical advances (see e.g., Caswell, 2007) we expect that senescence should be promoted at least if extrinsic mortality increases with age even in density independent populations. Hence, according to life history theory, investments in later-life fitness components are futile when the probability of surviving to higher ages is low due to extrinsic mortality. Resources should then be allocated to enhance early rather than late reproduction and survival (Kirkwood, 1977; Stearns, 1992; Cichoń, 1997). Especially in an environment in which the risk of dying from predation increases with age, large investments in early reproduction and a short lifespan are expected to evolve. These shifts in life history have attracted both considerable theoretical attention (reviewed in Kozłowski, 2006), and have been studied on an evolutionary scale. The predicted differences in resource allocation patterns and the resulting life histories have been observed in natural populations in habitats along predation
http://dx.doi.org/10.1016/j.exger.2015.05.008 0531-5565/© 2015 Published by Elsevier Inc.
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
F
O
R O
101 102
P
99 100
D
97 98
T
95 96
C
93 94
E
91 92
R
89 90
R
87 88
O
86
C
84 85
later aging, and with changes in b as faster/more rapid/accelerated or slower aging. As the aging-related mortality pattern is subject to natural selection, we can expect to see tradeoffs emerge between survival at different ages and between other life history traits. For instance, resources spent on current reproduction limit potential future reproduction due to inadequate resources allocated to maintenance and repairs. Therefore, in an environment in which the chances of survival decrease with age, we would expect to see intensive, but short early investments in soma (Cichoń, 1997), followed by earlier (higher a) or accelerated (higher b) aging (Fig. 1). The aim of the present study is to examine the biodemographic basis of the previously observed lifespan reduction in water fleas (Cladocera: Crustacea) which were exposed to a perceived (but not real) fish predation risk (Dawidowicz et al., 2010). Fish are visual predators, and it is well documented that they exert in positively size-dependent predation in water flea populations (e.g. Ivlev, 1961; Brooks and Dodson, 1965). Thus, fish-related extrinsic mortality risk of these indeterminate growers increases with age (Gliwicz and Pijanowska, 1989). We hypothesized that among water fleas kept under the perceived threat of size-dependent fish predation, this perceived risk would have a little effect on the water fleas' mortality early in life, but would induce either an earlier (without a change in b) or/and a more rapid increase (with a change in b) in their mortality at later ages. This hypothesis was based on the assumption that unlike survival at later ages, early survival is under strong selection pressure. We also expected to see shifts in the mean age at reproduction to earlier ages. It would be adaptive for the fish-threatened cladocerans to maximize both fitness components—i.e., fecundity and survival—early in life, at the cost of their later decline. Here, we test these predictions in two cladoceran species, and investigate their patterns of aging with respect to a perceived threat of size-dependent predation, which we introduced by olfactory cues (comparisons 1 and 2, Table 1). In addition, we relate these phenotypic differences in aging patterns to previously published data by Dawidowicz et al. (2013) and Dudycha and Tessier (1999) on genotypic differences between individuals from populations experiencing different levels of extrinsic mortality in their habitats. These were Daphnia from deep dark strata (hypolimnion), where fish predation risk is low, compared to Daphnia dwelling in well-lit surface layers (epilimnion), where risk from these visual predators is high (comparison 3, Table 1); and Daphnia from permanent lakes, experiencing relatively low extrinsic mortality levels in their habitat, compared to Daphnia
N
82 83
gradients (e.g., Dudycha and Tessier, 1999; Rennie et al., 2010). However, these differences have rarely been investigated as a resource allocation problem at the level of the individual sensing the risk of predation. In previous research, the maximum, the median, or the average lifespan have been used to compare cohorts between experimental treatments, as the lifespan is the most easily obtainable proxy for the aging process (e.g. Martin et al., 2002). Yet, it documents only the final outcome of aging, while telling us little about the trajectory of aging itself (Kowald, 2002; Bonsall, 2006). There are different views on how demographic aging should be measured (see, e.g., Partridge and Barton, 1996; Bronikowski and Promislow, 2005). The most common approach is to fit a parametric model, whereby the parametric estimates of the model are used as tools for mortality comparisons (e.g., Finch, 1990; Pletcher, 1999). The most widely used class of parametric models is based on the Gompertz (1825) distribution. The basic Gompertz model assumes that mortality is an exponential function of age, and describes only the impact of age-dependent sources of mortality risk. The model is thus often extended by an age-independent term typically related to background mortality (Makeham, 1867), which is relieved from the impact of aging (e.g., Makeham, 1867; Golubev, 2004). Furthermore, Golubev (2004) has suggested that the Makeham term should not be ignored in the data analysis, not only because of its possible biological significance, but also because this could bias estimates of Gompertz parameters. The Gompertz–Makeham model assumes that there is an exponential increase in age-dependent mortality μ(t). Rather than employing the notation of many theoretical papers, we have chosen to use a representation typical of many statistical packages (e.g., Colchero et al., 2012), which takes the form: μ(t) = e a + bt + c; where t denotes age, the exponent of the parameter a represents the initial intrinsic agedependent mortality, b represents the Gompertz exponential coefficient (defined hereafter as the rate of aging), and parameter c is the abovementioned Makeham term. Such a representation reduces the bias in the estimation of confidence intervals for a parameter (personal observation). The Gompertz–Makeham model describes mortality in many populations quite well. But while this model is both popular and useful, it is important to note its limitations, including the fact that it is empirically driven rather than strongly rooted in the biological theory (but see, e.g., Strehler and Mildvan, 1960; Golubev, 2009). In the Gompertz–Makeham setting, a shorter lifespan can result from higher initial intrinsic mortality (exp a), more rapid aging (b), higher ageindependent mortality (c), or a combination thereof. Throughout the text we will refer to changes associated with changes in a as earlier or
U
80 81
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
E
2
Fig. 1. Gompertz–Makeham hazard functions illustrating: (A) earlier/later emergence of aging modeled by changes in a; (B) more rapid/slower aging modeled by changes in b; (C) higher/ lower levels of age-independent background mortality modeled by changes in c.
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
Table 1 List of comparisons made. In each of the four comparisons, either earlier or faster aging is expected in the second group in pair. Comparisons 1 and 2 deal with phenotypic change in response to olfactory cue of predation, comparisons 3 and 4 deal with genetic differences between animals from habitats of different predation regimes. References in text.
Daphnia kairomone 2 3 4
Diaphanosoma control Diaphanosoma kairomone Daphnia from hypolimnion Daphnia from epilimnion Daphnia pulicaria Daphnia pulex
Phenotypic plasticity, direct effect of environment Intraspecies differences, indirect effect of environment (same experimental conditions) Interspecies differences, indirect effect of environment (same experimental conditions)
168
2. Material and methods
169
2.1. Animals and life table experiment
170
188 189
Daphnia longispina and Diaphanosoma brachyurum (Diaphanosoma hereafter), cladocerans of two common temperate lake species, were used in the experiments. We used standard, constant laboratory conditions, according to a well-established protocol in Daphnia life table experiments, i.e. summer photoperiod (16L:8D), temperature of 20 °C, green alga Scenedesmus obliquus as food source, filtered aged lake water as culture medium, which was pre-exposed to single fish (crucian carp Carassius auratus) for 24 h and refiltered before use to obtain olfactory cues of predation (kairomone). We followed 210 individuals of each species either in the absence or in the presence of the chemical cues of fish predation (5 individuals × 6 experimental vessels × 7 strains × 2 species × 2 treatments = 840 individuals). Offspring were counted daily. As the experiment was originally designed to test for differences in median lifespan, the individuals were followed from their births until more than one-half of the individuals in an experimental vessel had died; thus, the numbers of censored individuals varied between vessels. Each of the four datasets was obtained by pooling mortality data for seven homogenous clonal cohorts. The origin of the animals used and the experimental protocol were described in detail in Dawidowicz et al. (2010).
190
2.2. Statistical analysis
191
The survival data were analyzed using a Kaplan–Meier estimator and were fitted with Gompertz and Gompertz–Makeham models using a maximum likelihood estimation accounting for type II rightcensored data where needed. To determine which of these two models was more likely, we used a likelihood ratio test. Typically, under a likelihood ratio test a pair of models is compared, with one model being nested within the other. In our case, the Gompertz was a null (nested, c = 0) model, while the Gompertz–Makeham was an alternative (unrestricted, c ≥ 0) model. Because the null hypothesis c = 0 lies at the boundary of parameter space, the asymptotic null distribution of the likelihood ratio test statistic in this case is a mixture of χ20 (zero degrees of freedom) and χ21 (one degree of freedom) distributions with equal weights of 0.5. In this case the p-value is calculated according to the equation:
185 186 187
192 193 194 195 196 197 198 199 200 201 202 203 204
C
E
183 184
R
181 182
R
179 180
N C O
177 178
U
175 176
p‐value ¼ 0:5P χ 20 N R þ 0:5P χ 21 N R ; 206 207 208
Later reproduction and at larger size, longer lifespan Earlier reproduction and at smaller size, shorter lifespan Later reproduction and longer lifespan Earlier reproduction and shorter lifespan Later reproduction and longer lifespan Earlier reproduction and shorter lifespan Later reproduction and longer lifespan Earlier reproduction and shorter lifespan
No predation cue Size dependent (fish) predation olfactory cue Safe from fish predation Exposed to fish predation Permanent; low extrinsic mortality Temporary; high extrinsic mortality
T
from temporary ponds, experiencing high extrinsic mortality levels (Dudycha, 2004) (comparison 4, Table 1).
173 174
Known differences
No predation cue Size dependent (fish) predation olfactory cue
166 167
171 172
Environment
F
t2:7 t2:8 t2:9 t2:10 t2:11 t2:12
Phenotypic plasticity, direct effect of environment
where P denotes probability and R is a likelihood ratio test statistic (see, e.g., Verbeke and Molenberghs, 2000 for details). We avoided comparisons of parameter estimates of populations fitted with two different models by choosing the same model for the two compared populations.
We followed a simple rule: the Gompertz model was chosen only if a likelihood ratio test supported the null hypothesis in both populations. This approach allowed us to compare differences in c even when the Gompertz model was more likely for one population. This can also be seen as a conservative approach because the Gompertz–Makeham model is considered the primary model (Golubev, 2004) in most situations. Golubev (2004) showed that ignoring the Makeham (c) term may lead to a bias in the estimation of Gompertz parameters (a and b). Furthermore, Pletcher (1999) showed that the likelihood ratio test may be more likely to choose the overly simplistic Gompertz model for small and average sample sizes. In light of these considerations, we decided to set significance threshold α for all of the likelihood ratio tests to 0.1, which made it easier to reject the null hypothesis (i.e., Gompertz) and accept the alternative hypothesis (i.e., Gompertz–Makeham). The statistical inference on the differences between the parameter estimates, the survivorship, and the ln-mortality curves was based on specific confidence intervals (CI). The confidence intervals of the estimated Gompertz/Gompertz–Makeham parameters were calculated using a BCa method (e.g., Efron and Tibshirani, 1993, p. 184–188) from a sample of 250000 parametric bootstrap replicates, and by taking into account multi-vessel type II censored data (e.g., Kateri and Balakrishnan, 2008). The confidence intervals of the estimated survivorship and the ln-mortality curves were calculated analytically based on the sandwich estimator (e.g., Efron and Tibshirani, 1993, p. 310–311). The significance threshold for all tests, except the likelihood ratio test, was set to α = 0.05. To test for age-specific differences in fecundity between treatments, we performed ANOVA on the mean age at reproduction for replicate vessels. The mean age at reproduction was calculated as: TM = Σ m(t) × t / Σ m(t), where t is age and m(t) is age-specific fecundity.
209 210
2.3. Comparisons
239
First, we determined whether the Gompertz model or the Gompertz–Makeham model better describes the mortality patterns in the pairs of compared populations using the procedure described above (Section 2.2). We started with comparisons of the responses of D. longispina and of Diaphanosoma to a kairomone treatment and to a control treatment. Next, we compared the aging patterns in these two species, as Diaphanosoma inhabits the risky surface layers, or epilimnion; while D. longispina migrates to the hypolimnetic refuge during the daytime. We then added comparisons of epilimnetic to hypolimnetic D. longispina, and Daphnia pulex to Daphnia pulicaria from earlier studies. Finally, we compared the phenotypic life history shifts which emerge when D. longispina perceived cues of predation with the previously published data on the differences observed on an evolutionary scale between Daphnia spp. originating from high and low risk habitats (Dudycha and Tessier, 1999; Dawidowicz et al., 2013). Dudycha and Tessier (1999) compared the aging profiles of D. pulicaria, a species from lakes, which are permanent habitats and where Daphnia mortality rates are relatively low; and of D. pulex, a species from temporary ponds,
240
O
t2:6
Daphnia control
R O
1
P
Comparison
t2:5
D
t2:4
E
t2:1 t2:2 t2:3
3
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238
241 242 243 244 245 246 247 248 249 Q3 250 251 252 253 254 255 256 257
3.1. A comparison of the perceived predation and the control treatments
269
We found that the Gompertz model was strongly supported only in D. brachyurum: the p-values of the likelihood ratio test for both treatments were much greater than the assumed significance level of 0.1 (both were close to 0.5), and the estimated c parameters under the Gompertz–Makeham model equaled zero. For the D. longispina, the Gompertz model was strongly supported only for the kairomone treatment (p-value close to 0.5, c = 0), but for the control group the Gompertz–Makeham model performed better (p-value = 0.057). Consequently, we chose to use the Gompertz–Makeham model for both treatments of D. longispina. The survival analysis showed that the survivorship patterns differed significantly between the treatments in later life in both of the species studied. Survivorship was lower in animals kept in kairomone-treated water beginning on day 29 in D. longispina, and beginning on day 17 in D. brachyurum (Fig. 2A & B).
T C E
282 283
R
280 281
R
278 279
O
276 277
C
274 275
N
272 273
U
270 271
F
268
264
O
3. Results
262 263
R O
267
260 261
Consequently, the predation cue affected the mortality trajectories in both species. Under the anticipated predation threat, the mortality hazard was equal to or lower than that of the control in early life, but it was significantly higher beginning on day 20 in D. longispina and on day 12 in Diaphanosoma. Fig. 3A and B illustrates the mortality hazard rate of the studied cladocerans. The estimated parameters of the fitted Gompertz–Makeham model and their confidence intervals are presented in Fig. 4 and summarized in Appendix Table 1. The comparison of the confidence intervals for the estimated parameters revealed different patterns of mortality response in the two studied species. In D. longispina, the phenotypic change in age-specific mortality could not be clearly attributed to a specific and significant change in any combination of fitted parameters (Fig. 4). In Diaphanosoma, the change in the mortality pattern was clearly related to the lower values of a and the higher values of b in the kairomone treatment than in the control treatment (intrinsic mortality rate exp a was three times lower, and the rate of aging b was 2.5 higher in kairomone than in control treatment), which indicates more rapid aging (Fig. 4). Through kairomone exposure, the average age of first reproduction was shifted 20% earlier in D. longispina and 27% earlier in Diaphanosoma (Dawidowicz et al., 2010). During the first 10 days of life, D. longispina produced on average 1.8 times fewer offspring in the control than in the kairomone treatment, and Diaphanosoma produced 3.2 times fewer offspring (Dawidowicz et al., 2010). Thus, in the fish treatment overall fertility was shifted toward earlier ages in both species, with the mean age at reproduction TM being decreased by 29% in D. longispina (from 21.2 ± 0.6 days to 14.9 ± 0.6 days (mean ± SD), ANOVA: F1, 76 = 114.1, p b 0.0001) and by 18% in Diaphanosoma (from 13.5 ± 0.5 days to 11.1 ± 0.5 days, F1, 76 = 20.37, p b 0.0001).
E
265 266
where environmental conditions vary greatly and where Daphnia experience high mortality levels (Dudycha, 2004). Dawidowicz et al. (2013) compared the aging profiles of D. longispina which are able to find a refuge from visual predators such as fish by migrating during the day to the dark deep hypolimnion, with the profiles of D. longispina which live constantly in the well-lit epilimnion. Thus, to the best of our knowledge, we analyzed all of the published water flea survival data dealing with the indirect effects of predation, that have sufficient age spans and resolution.
P
258 259
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
D
4
Fig. 2. Phenotypic (top panels) and genetic (bottom panels) differences in cladoceran survivorships. Top panels: (A) Daphnia longispina, and (B) Diaphanosoma brachyurum reared in control and predation cue treatment — kairomone. Bottom panels: (C) Daphnia longispina originating from low (hypolimnion) and high (epilimnion) fish predation risk habitats (data from Dawidowicz et al., 2010), and (D) D. pulicaria from lakes and D. pulex from ponds (data from Dudycha and Tessier, 1999). Blue symbols — control treatment or animals originating from low risk habitats; Red symbols — predation cue treatment or animals originating from high risk habitat. Dots — Kaplan–Meier estimator; lines — survivorship curves based on fitted Gompertz– Makeham model with CI calculated from Fischer information matrix. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
5
D
P
R O
O
F
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
U
N C O
R
R
E
C
T
E
Fig. 3. Phenotypic (top panels) and genetic (bottom panels) differences in cladoceran mortality. Panels and color symbols as in Fig. 2. Dots — empirical ln-mortality; lines — ln-hazard curves based on fitted Gompertz–Makeham model with CI calculated from Fischer information matrix. Parameters of the best fit models are given in Fig. 4 and Appendix Table 1.
Fig. 4. Estimated parameters of the Gompertz–Makeham model for different experiments. Confidence intervals calculated by BCa method from a sample of 250 000 parametric bootstrap estimates.
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
356
In this study we showed that the anticipated risk of death can trigger phenotypic changes in aging. More specifically, we showed how olfactory cues of age-dependent predation affect the trajectory of the aging process in an organism capable of differential resource allocation. Two freshwater microcrustaceans Daphnia and Diaphanosoma were chosen for our analysis, as predation is one of the strongest forces shaping water flea life histories in the wild (Lynch, 1980). Positively sizedependent fish predation is believed to be the major source of mortality in large-bodied cladocerans in lakes ahead of starvation, parasitism, and aging (Gliwicz and Pijanowska, 1989). When we look at populations from habitats with varying risks of fish predation on an evolutionary scale, we can see that the lifespan of an individual Daphnia is shorter and its reproduction is more likely to be shifted toward earlier ages in a high-risk habitat than in a low-risk habitat (Dudycha and Tessier, 1999; Dawidowicz et al., 2013). Our study showed for two species of freshwater cladocerans how the perceived predation risk modulates the phenotypic aging profiles. Olfactory cues of size-dependent predation triggered either earlier or accelerated aging, respective of species. This was coupled with shifts in the mean age at reproduction to earlier ages, all in accordance with present theories and our predictions.
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
C
333 334
E
331 332
R
329 330
R
327 328
O
325 326
C
323 324
N
321 322
U
319 320
F
4. Discussion
318
O
355
316 317
R O
354
We found that the Gompertz model was preferable only in the hypo/ epi-limnion group: the p-values for the likelihood ratio tests were high (0.355 and 0.449, respectively). In the Daphnia pulex/pulicaria group, the effect of parameter c was significant for the D. pulicaria population (the p-value of the likelihood ratio test was close to zero). Thus, because it was found to be the better model, the Gompertz–Makeham model was chosen for the whole group. Between-population comparisons revealed that the estimated late life survivorship was always lower in animals from habitats with a higher extrinsic mortality risk. It was significantly lower in nonmigrating D. brachyurum than in D. longispina beginning on day 19 in the control treatment, and on day 15 in the kairomone treatment (Fig. 2A–B) (it should be noted, however, that the populations were fitted with two different models: the Gompertz and the Gompertz– Makeham, respectively). Beginning on day 5, survivorship was significantly lower in the epilimnion than in the hypolimnion population of D. longispina studied by Dawidowicz et al. (2013) (Fig. 2C); and beginning on day 41, survivorship was lower among the pond D. pulex than among the lake D. pulicaria studied by Dudycha and Tessier (1999) (Fig. 2D). Beginning at age five, mortality in the epilimnion population of D. longispina was higher than in the hypolimnion population at all ages (Fig. 3C–D). A comparison of the confidence intervals for the estimated parameters shows that there were significant differences in both Gompertz parameters. Epilimnetic Daphnia had exp a that was twice as high, as well as a b parameter that was twice as high as that of Daphnia from hypolimnion (Fig. 4). The differences in the mortality patterns found among the pond D. pulex and the lake D. pulicaria could be attributed to the higher values of the a parameter and the lower values of the c parameter in D. pulex than in D. pulicaria. The postponement of aging (without a significant change in b) in D. pulicaria was clearly observed in this case. This significant delay appears to have been driven by an initial intrinsic mortality exp a which was found to be about 40 000 times lower in the species from the less risky habitat (Figs. 3 and 4). The average age of first reproduction was about 20% earlier in the epilimnetic than in the hypolimnetic D. longispina. Moreover, during the first 10 days of their lives, the epilimnetic Daphnia produced 2.1 times more offspring per capita than their hypolimnetic conspecifics. Both the age of first reproduction and the mean age at reproduction were also earlier in D. pulex than in D. pulicaria.
P
314 315
Water flea mortality caused by fish predation increases with size, and with age. Under such selection, the optimal strategy is to allocate resources into early reproduction and defense at the cost of somatic repair and later maintenance. Such shifts toward the earlier onset of reproduction and higher early fecundity are well documented under conditions in which Daphnia is exposed to chemical cues of fish presence in the absence of actual predation events (reviewed in Lampert, 2011). How these strategy shifts might be associated with late life costs has only recently been demonstrated by Dawidowicz et al. (2010). The authors found that the median lifespan of cladocerans of the two genera was shortened by about 20% under the conditions of perceived fish predation. Here, using the Gompertz/Gompertz–Makeham model as a method to study the changes in mortality schedules in response to a predator cue, we showed that the shortened lifespan of Daphnia could be attributed to an increased Gompertz's a rather than to an increased rate of aging b, although none of these results were significant, probably due to colinearities between the Gompertz parameters and the Makeham term. This finding is intuitively and visually confirmed in Fig. 3, in which both ln-mortality curves are parallel at the oldest ages. It is therefore clear that we are observing earlier and later onsets of aging, rather than accelerated/decelerated aging. In contrast, in Diaphanosoma, the shortening of the lifespan was found to be due to a more than twofold increase in the rate of aging (b), despite the lower a. Here, both mortality curves clearly cross and then diverge with age (Fig. 3). The threatened Diaphanosoma thus appears to provide a clear example of accelerated aging. Accelerated aging has been previously observed under some commonly used environmental interventions, yet to our knowledge it has not been observed in association with perceived predation risk. The modulation of aging trajectories realized within the phenotypic plasticity of a genotype has been well documented in some organisms, like Caenorhabditis elegans. For example, dietary restriction has been shown to increase lifespan by lowering the slope of the mortality increase with age (Lenaerts et al., 2007); and mild heat shocks have been found to have similar effects (Wu et al., 2009). The nematodes have also been found to have more rapid increases in mortality rates when exposed to pathogens (Baeriswyl et al., 2010). Moreover, sensory perception has been shown to regulate the lifespan of this animal (Apfeld and Kenyon, 1999), and in Drosophila melanogaster exposure to olfactory cues of food presence shortened fly lifespan (Libert et al., 2007). This is comparable to our findings on water flea earlier or accelerated aging under olfactory cues of predation risk. Again in agreement, visual and chemical predator cues were previously shown to cause oxidative damage in prey damselfly larvae (Janssens and Stoks, 2013). Aging is often understood as the interaction between the accumulation of damage, and maintenance or repair (e.g. Strehler and Mildvan, 1960; Cichoń, 1997). Some damage can be repaired, other is irreversible, and remains in for the rest of life of the individual (e.g. Rübe et al., 2011). The irreversible damage in particular can compromise the repair mechanism itself and contribute to the increase in the rate of damage accumulation in the future. There is still a debate on how the process of damage accumulation converts to mortality. Theoretically, two factors seem important: (i) the organism's ability to reduce the rate of accumulation of irreversible damage, and (ii) its physiological response to already accumulated damage. The first depends on the effectiveness of repair mechanisms, as well as on the deterioration of these mechanisms with damage or age. Hereby, the rate at which deterioration occurs is determined by resources spent on damage protection, and resources that are invested into repair. The amount of resources available depends on the specific resource allocation strategy of the organism, and physiological constraints of the body. This should clearly affect the rate of aging (e.g. Gompertz's b). The second factor describes an organism's vulnerability, or the physiological ability to tolerate the damage. Specifically, it can be understood as the rate at which accumulated damage is translated into mortality. This rate is constant throughout the lifespan, and can be interpreted as Gompertz's a. In consequence, different levels of
D
3.2. Comparisons of populations from high and low risk habitats
T
313
E
6
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
501 502 503 504 505
C
461 462
E
459 460
R
512
The research was partially supported by Max Planck Institute for Demographic Research to Maciej J. Dańko. The experimental part of the study was supported by a grant from the Polish Ministry of Science and Higher Education No. N304 005 32/0647 to Barbara Pietrzak. Jutta Gampe, Alexander Scheuerlein, Jan Kozłowski, Hal Caswell, Daniel Levitis, reviewer Jeff Dudycha and another anonymous reviewer provided valuable comments which helped to improve the manuscript.
513 Q4
References
520
F
457 458
R
455 456
N C O
453 454
U
451 452
Acknowledgments
O
The present findings offer new insights into the effects on the aging process of the perceived risk of earlier death associated with predation threat. Cladocerans of two genera respond to the perceived risk of size-dependent predation with either more rapid or earlier aging. This phenotypic effect is consistent with differences observed between populations from risky and safe habitats. Finally, as the studied species
449 450
R O
500
448
506 507
Abrams, P.A., 1993. Does increased mortality favor the evolution of more rapid senescence? Evolution 47, 877–887. Apfeld, J., Kenyon, C., 1999. Regulation of lifespan by sensory perception in Caenorhabditis elegans. Nature 402, 804–809. Baeriswyl, S., Diard, M., Mosser, T., Leroy, M., Manière, X., Taddei, F., Matic, I., 2010. Modulation of aging profiles in isogenic populations of Caenorhabditis elegans by bacteria causing different extrinsic mortality rates. Biogerontology 11, 53–65. Bonsall, M.B., 2006. Longevity and ageing: appraising the evolutionary consequences of growing old. Philos. Trans. R. Soc. Lond. B Biol. Sci. 361, 119–135. Bronikowski, A.M., Promislow, D.E.L., 2005. Testing evolutionary theories of aging in wild populations. Trends Ecol. Evol. 20, 271–273. Brooks, J.L., Dodson, S.I., 1965. Predation, body size, and composition of plankton. Science 150, 28–35. Caswell, H., 2007. Extrinsic mortality and the evolution of senescence. Trends Ecol. Evol. 22, 173–174. Cichoń, M., 1997. Evolution of longevity through optimal resource allocation. Proc. R. Soc. B 264, 1383–1388. Colchero, F., Jones, O.R., Rebke, M., 2012. BaSTA: an R package for Bayesian estimation of age-specific survival from incomplete mark-recapture/recovery data with covariates. Methods Ecol. Evol. 3, 466–470. Dańko, M.J., Kozłowski, J., 2012. Mutation accumulation may only be a minor force in shaping life-history traits, even when reproduction is sexual. PLoS ONE 7 (4), e48302. Dańko, M.J., Kozłowski, J., Vaupel, J.W., Baudisch, A., 2012. Mutation accumulation may be a minor force in shaping life history traits. PLoS ONE 7 (4), e34146. Dawidowicz, P., Prędki, P., Pietrzak, B., 2010. Shortened lifespan — another cost of predator avoidance in cladocerans? Hydrobiologia 643, 27–32. Dawidowicz, P., Prędki, P., Pietrzak, B., 2013. Depth selection behavior and longevity in Daphnia: an evolutionary test for the predation-avoidance hypothesis. Hydrobiologia 715, 87–91. Dudycha, J.L., 2004. Mortality dynamics of Daphnia in contrasting habitats and their role in ecological divergence. Freshw. Biol. 49, 505–514. Dudycha, J.L., Tessier, A.J., 1999. Natural genetic variation of life span, reproduction, and juvenile growth in Daphnia. Evolution 53, 1744–1756. Efron, B., Tibshirani, R.J., 1993. An Introduction to the Bootstrap. Chapman and Hall, New York. Finch, C.E., 1990. Longevity, Senescence, and the Genome. University of Chicago Press, Chicago, London. Gliwicz, Z.M., Pijanowska, J., 1989. The role of predation in zooplankton succession. In: Sommer, U. (Ed.), Plankton Ecology: Succession in Plankton Communities. Springer-Verlag, New York, pp. 253–295. Golubev, A., 2004. Does Makeham make sense? Biogerontology 5, 159–167. http://dx.doi. org/10.1023/B:BGEN.0000031153.63563.58. Golubev, A., 2009. How could the Gompertz–Makeham law evolve. J. Theor. Biol. 238, 1–17. http://dx.doi.org/10.1016/j.jtbi.2009.01.009. Gompertz, B., 1825. On the nature of the function expressive of the law of human mortality, and on the mode of determining the value of life contingencies. Philos. Trans. R. Soc. 115, 513–585. Hamilton, W.D., 1966. The moulding of senescence by natural selection. J. Theor. Biol. 12, 12–45. Ivlev, V.S., 1961. Experimental Ecology of the Feeding of Fishes. Yale University Press, New Haven. Janssens, L., Stoks, R., 2013. Predation risk causes oxidative damage in prey. Biol. Lett. 9, 20130350. Kateri, M., Balakrishnan, N., 2008. Inference for a simple step-stress model with Type-II censoring and Weibull distributed lifetimes. IEEE Trans. Reliab. 57, 616–626. Kirkwood, T.B.L., 1977. Evolution of ageing. Nature 270, 301–304. Kowald, A., 2002. Lifespan does not measure ageing. Biogerontology 3, 187–190. Kozłowski, J., 2006. Why life histories are diverse. Pol. J. Ecol. 54, 585–605. Lampert, W., 2011. Daphnia: development of a model organism in ecology and evolution. In: Kinne, O. (Ed.), Excellence in Ecology. Book 21. International Ecology Institute, Oldendorf/Luhe.
P
5. Conclusions
446 447
normally experience the presence of fish and their chemical cues in natural habitats, the problem at hand should be studied in the reverse scenario: i.e., whether senescence slows down or is delayed when there is no perceived fish predation threat. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.exger.2015.05.008.
D
499
444 445
T
498
vulnerability to damage should generate different levels of mortality even at the same level of damage (see also Partridge et al., 2005 and Simons et al., 2013). Ecology of a species is expected to lead to evolution of different mechanisms of resource allocation and physiology of damage accumulation, repair and resistance. Cladocerans from the two studied genera differ in terms of their life histories, their behavioral strategies, and their habitats. D. brachyurum stays in surface waters that are well lit during the day, where the risk associated with visual predators, such as fish, is higher. These surface waters (epilimnion) are also warm and abundant in food, and thus enable faster growth and development. In contrast, D. longispina is mostly a migratory species. It usually stays in the high-risk, high-food surface waters during the night only, when fish predation risk is low, while migrating to relatively safe deeper strata during the day. Not surprisingly, Diaphanosoma, the species which on average faces a higher predation risk, was shown here to have a higher rate of aging than D. longispina when compared in the lab under a lakeimitating fish cue presence; i.e., in the kairomone treatment (a more than twofold difference in b was found between the species, but it is important to note that they were fitted with different models; namely, the Gompertz and the Gompertz–Makeham models, respectively). The extraordinary shifts in the Diaphanosoma's life history toward earlier reproduction and more rapid aging could be interpreted as follows: the earlier reproduction is an anti-predator defense alternative to escape behavior, while the accelerated aging is the cost thereof, as the strength of the life history response seems to correlate negatively with the efficiency of escape behavior (Pietrzak et al., unpublished data; see also Thys and Hoffmann, 2005; Vijverberg et al., 2006). Thus, the lower rates of fitted mortality observed early in life in the kairomone-treated Diaphanosoma may be a sign that a trade-off takes place between early and late fitness components. The enhancement of early survival under the perceived threat would surely be adaptive if more females survived until the first reproduction. In addition, it could be explained within the framework of the “maintenance gap” hypothesis if the observed early life response with greater reproductive output increases the maintenance requirements of the animal (Wensink et al., 2012). Our between-habitat comparisons revealed that late-life survivorship was always lower in animals from mortality-riskier habitats: specifically, it was lower in the non-migrating D. brachyurum as compared to the migrating D. longispina; it was lower in the population of that latter species from the well-lit epilimnion than in the population from the deep, dark hypolimnion; and it was lower in D. pulex from temporary ponds than in D. pulicaria from lakes. Moreover, shifts in agespecific fecundity—measured as the age at first reproduction, the early reproductive effort, or the mean age at reproduction—were consistent across comparisons. Finally, we found that the presence or the lack of kairomone may have had an effect on age-independent mortality (c). The significance of this mortality parameter was only clear in the D. pulicaria cohorts when it was compared to that of the D. pulex cohorts (Dudycha and Tessier, 1999), and when it masked the lower degree of frailty of D. pulicaria, which delayed senescence in this species. The perception of risk might result in decreased vulnerability to external ageindependent causes of death due to, for example, risk avoidance behavior, which may be expressed even in protected environments. Thus, the level of age-independent mortality could be under the control of the organism to some extent; this is an issue which deserves further study.
E
442 443
7
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
508 509 510 511
514 515 516 517 518 519
521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581
F
Simons, M.J.P., Koch, W., Verhulst, S., 2013. Dietary restriction of rodents decreases aging rate without affecting initial mortality rate — a meta-analysis. Aging Cell 12, 410–414. Stearns, S.C., 1992. The Evolution of Life Histories. Oxford University Press, New York. Strehler, B.L., Mildvan, A.S., 1960. General theory of mortality and aging. Science 132, 14–21. Thys, I., Hoffmann, L., 2005. Diverse responses of planktonic crustaceans to fish predation by shifts in depth selection and size at maturity. Hydrobiologia 551, 87–98. Verbeke, G., Molenberghs, G., 2000. Linear mixed models for longitudinal data. Springer Series in Statistics. Springer, New York. Vijverberg, J., Doksæter, A., Van Donk, E., 2006. Contrasting life history responses to fish released infochemicals of two co-occurring Daphnia species that show different migration behaviour. Arch. Hydrobiol. 167, 89–100. Wensink, M.J., van Heemst, D., Rozing, M.P., Westendorp, R.G., 2012. The maintenance gap: a new theoretical perspective on the evolution of aging. Biogerontology 13, 197–201. Williams, G.C., 1957. Pleiotropy, natural selection and the evolution of senescence. Evolution 11, 398–411. Williams, P.D., Day, T., 2003. Antagonistic pleiotropy, mortality source interactions, and the evolutionary theory of senescence. Evolution 57, 1478–1488. Wu, D., Rea, S.L., Cypser, J.R., Johnson, T.E., 2009. Mortality shifts in Caenorhabditis elegans: remembrance of conditions past. Aging Cell 8, 666–675.
N
C
O
R
R
E
C
T
E
D
P
R O
Lenaerts, I., van Eigen, S., van Fleteren, J., 2007. Adult-limited dietary restriction slows gompertzian aging in Caenorhabditis elegans. Ann. N. Y. Acad. Sci. 1100, 442–448. Libert, S., Zwiener, J., Chu, X., VanVoorhies, W., Roman, G., Pletcher, S.D., 2007. Regulation of Drosophila life span by olfaction and food-derived odors. Science 315, 1133–1137. Lynch, M., 1980. The evolution of cladoceran life histories. Q. Rev. Biol. 55, 23–42. Makeham, W.M., 1867. On the law of mortality. J. Inst. Actuaries 13, 325–358. Martin, L.J., Mahaney, M.C., Bronikowski, A.M., Dee Carey, K., Dyke, B., Comuzzie, A.G., 2002. Lifespan in captive baboons is heritable. Mech. Ageing Dev. 123, 1461–1467. Medawar, P.B., 1952. An Unsolved Problem in Biology. HK Lewis, London. Partridge, L., Barton, N.H., 1996. On measuring the rate of ageing. Proc. R. Soc. B 263, 1365–1371. Partridge, L., Pletcher, S.D., Mair, W., 2005. Dietary restriction, mortality trajectories, risk and damage. Mech. Ageing Dev. 126, 35-4. Pletcher, S.D., 1999. Model fitting and hypothesis testing for age-specific mortality data. J. Evol. Biol. 12, 430–439. Rennie, M.D., Purchase, C.F., Shuter, B.J., Collins, N.C., Abrams, P.A., Morgan, G.E., 2010. Prey life-history and bioenergetic responses across a predation gradient. J. Fish Biol. 77, 1230–1251. Reznick, D.N., Bryant, M.J., Roff, D., Ghalambor, C.K., Ghalambor, D.E., 2004. Effect of extrinsic mortality on the evolution of senescence in guppies. Nature 431, 1095–1099. Rübe, C.E., Fricke, A., Widmann, Y.A., Fürst, T., Madry, H., Pfreundschuh, M., Rübe, C., 2011. Accumulation of DNA damage in hematopoietic stem and progenitor cells during human aging. PLoS ONE 6.
U
582 583 584 585 586 587 588 589 590 591 592 593 Q5 594 595 596 597 598 599 600 601 602 603 604
B. Pietrzak et al. / Experimental Gerontology xxx (2015) xxx–xxx
O
8
Please cite this article as: Pietrzak, B., et al., How perceived predation risk shapes patterns of aging in water fleas, Exp. Gerontol. (2015), http:// dx.doi.org/10.1016/j.exger.2015.05.008
605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625
