More stressful event does not always depress subsequent life performance

CHEN Ying-ying, ZHANG Wei, MA Gang, MA Chun-sen

State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing 100193, P.R.China

Abstract Climate change has led to a substantial increase in intensity and duration of heat waves worldwide. Predicting the ecological impacts of hot events should incorporate both immediate and potential carry-over effects in different intensities of heat waves.Previous studies suggested that higher heat dose in early life stage of insect generally decreased immediate survival and depressed adult reproduction through carry-over effects, or unchanged adult performance through recovery effects. However,our previous study showed a different pattern, in which longer heat exposures in larval stage did not always decrease but sometimes increase the subsequent adult maturation success in the diamondback moth. We speculated that it might be another important pattern in the carry-over effects vs. heat dose, and conducted experiments using a global pest, Plutella xylostella. Our present results suggested that heat exposures in early life stage reduced the immediate survival and produced general declines with significant zigzag fluctuating patterns in subsequent body size and reproduction as exposure durations increased. The similar patterns were also validated in other insect taxa and other stresses by reanalyzing the experiment data from literatures. The finding highlights the importance for differentiating the biological effects and consequences of changes in heat dose at fine scales; daily exposure hours of a hot day should be considered to predict population dynamic under climate change.

Keywords: climate change, Plutella xylostella, extreme temperature, reproduction, carry-over effect

1. lntroduction

Climate change has led to a substantial increase in intensity and duration of heat waves worldwide (IPCC 2012), and has caused significant impacts on biological systems across various levels, ranging from development responses of individuals, population dynamics, to ecosystem structure and functions (Teskey et al. 2015; Thomson et al.2015; Van Dievel et al. 2017; Vázquez et al. 2017). The biological effects of heat waves depend essentially on heat dose, i.e., the combination of temperature amplitudes and exposure durations. Ecological consequences of heat waves include not only the immediate effects but also the potential carry-over effects on later life stages, importantly on the reproductive adult stage (Weinig and Delph 2001;Harrison et al. 2011; Zhang et al. 2015). Immediate effects have been well documented and generally concluded that animals’ immediate survival continuously decreases as heat dose increases (Denlinger and Yocum 1998). Carryover consequences caused by stress in early stages have attracted scientists’ attention in recent years (Harrison et al.2011; O’Connor et al. 2014). For animals with incremental development such as bird (Callebaut 1991), reptiles (Bull 1980) and fish (Jonsson and Jonsson 2014), stress in early life stage can be carried over to adult stage and depress adult reproduction. For insects, especially holometabolous species with complex life cycles, this issue is still under arguments.

For insects, previous studies have proposed two viewpoints concerning carry-over effects. One major point suggests that subsequent performance will linearly or nonlinearly decrease with increasing heat exposure durations(Zhang et al. 2015). Higher development temperatures will lead to non-reversible phenotype changes, such as a decline in number of ovarioles (Cohet and David 1978;Hodin and Riddiford 2000; Bader and Williams 2012), a shrink of body size (French et al. 1998; Angilletta et al. 2004;Gardner et al. 2011) that mostly irreversibly decrease adult reproduction. Larger amount of stress dose always leads to more disruption of cell microenvironment or tissue injury(Denlinger and Yocum 1998; Rohmer et al. 2004). While the other point, i.e., lifecycle modularity hypothesis (Potter et al. 2011) or adaptive decoupling hypothesis (Moran 1994; Stoks and Córdoba-Aguilar 2012), states that early stress has no long-term effects on later adult performance.Insects have multiple disjunct life stages separated by molting or/and metamorphosis (Wilbur 1980; Moran 1994),which allow repair and even reconstruction of morphology(i.e., exoskeleton), hormone regulation systems, and physiological metabolism (Consoulas 2000, 2002; Seifert et al. 2012). Thus, insects’ modular life-cycle allows them to recover and uncouple body injury in one life stage to the later life stage (Potter et al. 2011; Xing et al. 2014). However, in a previous study, we recognized an unusual phenomenon that a subsequent performance, i.e., adult maturation success firstly decreased as exposure durations increased when the diamondback moth larvae were heated under a series of exposure durations, but it contrarily increased dramatically to the control level after a near-lethal dose (Appendix A;Zhang et al. 2015). We speculated that it would be another important pattern in carry-over effects vs. heat dose.

We used the diamondback moth, Plutella xylostella, a global pest as our model organism, which is well known for the strong adaptation to stress, e.g., pesticides (Ferré and Van Rie 2002; Furlong et al. 2013) and extreme temperatures (Gu 2009; Nguyen et al. 2014). We examined how different exposure durations at 40°C for 0-5 h (with a 0.5-h interval) during 1st-instar larvae stage affect the immediate survival and survivors’ subsequent development,survival rate, body size and adult reproduction. To ensure the reliability of our conclusion, we used multiple replications for each treatment (24-96 replications and 10 individuals for each replication). We found that increasing heat exposure duration in early life stages did generate a general decline trend but with significant increasing episodes in pupal weight and fecundity of female survivors.

2. Materials and methods

2.1. lnsects and experimental manipulation

We collected diamondback moth larvae from Brassica fields in Wuhan, Hubei Province, China, in May 2010. Since the collection, the diamondback moth has been reared using the artificial diet followed (Zhang et al. 2013) at (25±1)°C and 15 h L:9 h D. We picked ～4 000 newly hatched larvae(age less than 12 h) from the stock population as tested insects. The larvae were randomly divided into 11 groups,with one for control (25°C) and 10 for heat treatments (40°C for 0.5 to 5 h with an increment of 0.5 h, respectively). The number of replications for each treatment depended on the heat exposure durations. More replications were applied for heat treatments with longer exposures to ensure enough individuals remained after stronger heat stresses (Table 1).

lmmediate survival testBefore heat stress, 10 individuals for each replication were placed into a plastic tube (15 mm diameter×5.5 mm height) and were fed with fresh artificial food (Southland Products Incorporated, USA) to avoid larvae dehydration. Every 24 tubes from the same treatment were inserted to a 24-well plate. Following heat experiment details in Appendix B, the plates were immerged into a water bath,where water was continuously stirred by a pump to ensure homogeneity of water temperature to maintain (40±0.5)°C.Tubes of the control group were moved to the (25±0.5)°C.After heat exposure, the plates with tested individuals were taken out of the water bath and were maintained in the climate chamber ((25±0.5)°C and 60-80% RH) for 24 h.Then the number of 1st-instar larval survivors in each tube was checked and recorded.

Survivors’ subsequent performance testTo balance the sample size for testing the subsequent performance,65-70 1st-instar larval survivors were used for 0-1.5 h treatments, while all survivors (about 30-90 individuals)were used for 2.0-5.0 h treatments (Table 1). The 1st-instar larval survivors were transferred into glass tubes (15 mm diameter×10 mm length), one larva was reared in one tube.Larvae were supplied with fresh artificial food and food was renewed every 3 days to ensure that larval development and growth. Once adults eclosed, all adults (age less than24 h) from the same treatment were paired (one female and one male) in a glass bottle (30 mm diameter×15 mm length). A cotton ball soaked with 5% honey solution and a piece of “egg card” made of Parafilm (Zhang et al. 2013)was supplied for egg laying. The cotton balls and egg cards were replaced daily. The number of different instar larvae and pupae was observed twice per day (08:00 and 20:00).After pupation for 1.5 day, pupae from each treatment were weighed individually by using a balance (Sartorius BP221S,Germany). Emerged adults were counted twice per day(08:00 and 20:00). After adult pairing, we checked adult survival and counted eggs (on the wall of the glass tube and egg card) daily until adults died.

Table 1 Number of replications and individuals used to test the immediate survival and subsequent performance of Plutella individuals

2.2. Statistical analysis

We used generalized linear models (GLM) in the “glm”function in R, and the “car” package with type II statistics,to determine how different heat exposures affect immediate survival of 1st-instar larvae, development rate of 1st-instar larva and 2nd-instar larva to adult, pupal weight and the longevity of both sex, female fecundity. The mean differences of the traits between treatments were analyzed using Tukey’s post hoc test in the multcomp package(“glht” function). We used a contingency analysis to compare the maturation success between treatments by computing the Chi-square statistic. Maturation success was measured as the proportion of emerged adults from 1st-instar larvae survived after heat treatments. We analyzed the relationships between female pupal weight and female fecundity by fitting a linear regression model.To determine whether differences of fecundity were driven by differences in female pupal weight which was generated by 1st-instar larval treatments, we run the GLM analysis with heat exposures, pupal weight and their interactions as fixed effects. All statistics were done in R (R Development Core Team 2006)

3. Results

3.1. lmmediate and subsequent survival rate responding to heat exposure duration

The immediate survival of 1st-instar larvae decreased from 100.0 to 6.3% as heat exposure durations increased from 0 to 5.0 h (Table 2; Fig. 1-A). The subsequent survival, i.e., the maturation success of 1st-instar larval survivors significantly decreased from 0.82 at the control to 0.66 at 4.0-h treatment(Table 2), and other treatments had a similar maturation success with the control (Fig. 1-A). Heat exposures did not influence the longevity of female or male adults (Table 2).

3.2. lmmediate and subsequent development rate responding to heat exposure duration

Development rate of 1st-instar larvae generally declined with an increase of exposure durations at 40°C (Table 2;Fig. 1-B). Heat exposures did not influence the subsequent development from 2nd-instar larva to adult (Table 2;Fig. 1-B).

3.3. Subsequent body size responding to heat exposure duration

Heat treatments on 1st-instar larval stage affected the female’s body size (measured in pupal weight), but not the male’s (Table 2). Female’s pupal weight also displayed a general decline as exposure duration increased, but with a detailed exposure duration dependent zigzag pattern(Fig. 1-C). The pupal weight decreased from 8.4 mg of the controls to 7.3 mg (n=23) when the exposure time increased from 0 to 2.0 h; then increased to 7.9 mg (n=25)at 2.5 h; again decreased to the lowest 7.2 mg at 4.0 h;further increased to 7.5 mg (n＞10) at 4.5 or 5.0 h exposures(Fig. 1-C).

Table 2 Effect of heat exposure duration on different performance of Plutella individuals.

Fig. 1 Mean performance after exposing 1st-instar larvae of the diamondback moth at 40°C for different exposure durations. A,immediate survival and subsequent survival rate (measured as immature succes). B, development rate of 1st-instar and 2nd-instar larva to adult. C, female pupal weight. D, adult fecundity. Different letters above each bar (SE) indicate significant differences between exposure durations at 40°C (P＜0.05) based on post hoc tests (Tukey).

3.4. Subsequent adult reproduction responding to heat exposure duration

Larval heat exposures significantly influenced adult fecundity(Table 2). The egg production performed a general decrease but with a significant zigzag pattern responding to heat exposure duration (Fig. 1-D). Egg production of adults (n＞22 for each treatment) decreased by 16% (from 270 to 227 eggs female-1) when exposure increased from 0 to 2.0 h;then increased to 271 eggs female-1(n=25) at 2.5 h; again decreased to the lowest (230 eggs female-1; n＞15) at 4.0 h;finally slightly increased at 4.5 h (Fig. 1-D).

Adult fecundity is linearly correlated with pupal weight(R2=0.312; P＜0.001) (Fig. 2). 1st-instar larval heat treatments had no more significant effects on the adult fecundity after considering differences in pupal weight(X2=12.10, df=10; P=0.279). This suggests that these complex effects by larval heat treatments on fecundity were largely driven by the changes in female body size.

4. Discussion

Fig. 2 Relationship between female pupal weight and lifetime fecundity. The positive linear regression analysis showed that larger female pupae always had more lifetime fecundity.

Predicting the ecological impacts of hot events needs to consider both immediate and potential carry-over effects of different dose of heat waves. Previous studies suggested that heat stress in early life stage decreased immediate survival (Denlinger and Yocum 1998), depressed adult performance through carry-over effects or unchanged the adult performance through adaptive decoupling mechanisms(Moran 1994; Stoks and Córdoba-Aguilar 2012). Our results both in our previous study (Zhang et al. 2015) and present detailed experiments suggested that early heat stress also reduced immediate survival, but surprisingly produced a general decline with significant zigzag fluctuating patterns in subsequent body size and reproduction as exposure durations increased.

Although other authors did not state the similar conclusion, their data derived from different insect taxa and different heat conditions clearly showed the similar pattern with ours. The immediate survival generally decreased as increasing heat dose in early stage, while subsequent adult fecundity in Harmonia axyridis (Zhang et al. 2014), Bemisia tabaci (Cong et al. 2010; Chen et al. 2011), puptal weight in Ostrinia furnacalis (Han 2017) also bounced up after a certain decrease (Fig. 3). Such patterns also occurred in other traits, such as adult emergence rate, maturation success and adult longevity (Fig. 3). In addition, this pattern was also found in other stress study, for example,adults from γ-ray stressed eggs in Drosophila melanogaster contrarily increased body weight or fecundity when ray intensity became stronger (Vaiserman et al. 2004). Thus,we speculated that higher heat dose not always depress subsequent life performance more.

Fig. 3 Subsequent performance curve with increasing heat dose. Mean performance and heat dose data were derived from published paper listed. Heat dose from different literatures were transformed by the formula: (Heat treatment temperature-Control temperature)×Exposure duration (degree-hours). Min-Max normalization process was used to rescale original data of performance and heat dose, so that all values are within the range of 0 and 1.

The zigzag pattern of adult reproduction corresponding to the heat dose increase in early life stage could be explained very well by the synchronized zigzag pattern of pupal weight (Figs. 1-C and 3). We know that body size affects organism’s physiology and consequent reproduction(Kingsolver and Huey 2008). Many documents suggested that higher rearing temperatures produce smaller body size(French et al. 1998; Angilletta et al. 2004; Gardner et al.2011). However, we still have not a satisfied mechanical explanation for how fine differences of heat dose in early stage mediate organisms’ body size, although gut damage after heat stress was speculated to deprive nutrients and water of larvae and consequently affect body size (Krebs and Feder 1997, 1998). In addition, thermal stress could cause heat shock response which is maximally activated only at certain level of stress dose (Lindquist 1986; Huang and Kang 2007), while lower or stronger dose could not trigger enough amount of such protective proteins.

Higher dose of heat stress not only causes severe body injury (Denlinger and Yocum 1998; Rohmer et al. 2004),but also can increase the proportion of tolerant individuals in a population (Bliss 1935; Prentice 1976). When a population encounters stress, the sensitive individuals might suffer irreversible injury more easily (Denlinger and Yocum 1998; Rohmer et al. 2004) and lower the mean performance of the population. As stress dose increases to a lethal level for the sensitive individuals, the proportion of the heat tolerant individuals increases in the remained population. Tolerant individuals may restore all functions during subsequent suitable conditions by foraging more(Klok and Chown 2001; Benoit et al. 2007; Van Dievel et al.2017) or by reconstruction function in modular life cycles of insect (Potter et al. 2011). Such stress dose would result in an increase in average performance of the remained population. However, if heat dose accumulates continuously to the critical threshold for the tolerant individuals, the irreversible thermal injury would even cause detrimental impacts on the relatively resistant individuals and thus reduce population’s average performance again. Similarly,another increase in subsequent performance may occur likewise within the ‘sub-group’ consist of more resistant individuals. Continuously increasing heat dose causes thermal injury to these relatively resistant individuals and reduce average population performance. As a result, the organisms displayed the zigzag fluctuation pattern in body size and reproduction.

Many scientists have not detected the zigzag pattern because they only focused on limited ranges of exposure durations, during which the life history performances gradually reduced as heat exposures increased (Ma et al. 2004; Zhang et al. 2015). However, when the heat exposures continuously increase, individuals with low heat tolerance could be eliminated, and the individuals with higher tolerance remain in the tested population. Importantly, if the tolerant individuals have higher inherent life history performances such as fecundity and body size, the mean performance (e.g., fecundity and pupal weight) in the longer heat exposure treatment could be higher than that in shorter heat exposure treatments. If the population is consisted of different groups of individuals that have different lethal exposure thresholds, the above decrease-increase phase could be repeated during whole exposure spectrum and formed a pattern of zigzag fluctuation. Indeed, recent investigation indicated that life stages with larger body size in the same species have stronger heat tolerance(Klockmann et al. 2017a). The body size dependent heat tolerance seems to be agreed with our zigzag pattern of body size. In addition, it is widely proved that larger body size leads to a larger fecundity in insects (Honěk 1993;Kingsolver and Huey 2008). Therefore, the heat exposure dependent zigzag pattern might often exist in nature. So far, limited information can be found on the relationship between stress doses and carry-over effects, although a few recent studies had documented the carry-over effects of high temperatures in early stages. In these studies, a single dose (Piyaphongkul et al. 2012; Klockmann et al. 2017b;Zhao et al. 2017) or narrow range of heat stress (Ma et al.2004; Zhang et al. 2015) were commonly applied in the experiments, consequently the body size or reproduction simply declined. Therefore, these studies could not detect the zigzag pattern.

The sample size is important when we make inferences about a population from a sample. In our model system,the diamondback moth has large variance between different individuals in pupal weight (3.8-7.1 mg, unpublished data)and fecundity (40-240 eggs female-1, see Zhang et al.2015). We worried that few individuals would not produce reliable data of individual fecundity and therefore used larger initial populations for longer exposures to get enough survivors. We acknowledged that GLM would not eliminate the effects of sample size because we manipulated the sample size artificially. Indeed, we could not get exactly the same number of survivors although we tried our best to do so. We used GLM to differentiate the subsequent performances because it is commonly used to analyze the unbalanced design (although usually do not applied for screened samples) (Christensen 2015). Anyway, we increased the sample size to get more screened survivors with higher heat tolerance under longer exposures, which could affect the probability to reach the significant level in statistics.

Global climate change will not only lead to a substantial increase in average temperature but also in the duration of heat waves (IPCC 2012). Biodiversity conservation,bio-resources management and pest control all need an accurate prediction of population dynamics under heat waves (Harrington et al. 2001; Segan et al. 2016). Predicting how heat events influence population abundance or subsequent population growth after extreme climate events is not only relied on the number of heat-stress survivors, but also survivors’ subsequent reproduction (Weinig and Delph 2001; Harrison et al. 2011; Zhang et al. 2015). We found survivors performance is not straightforward decreasing but generates a general decline trend with zigzag fluctuating episodes as increasing heat dose in early life stages.

Our findings have broad biological and ecological implications for the studies regarding the consequences of heat stress. Given that thermal effects depend not only on the temperature intensity but also on the exposure time(Denlinger and Yocum 1998), our studies highlighted the importance for differentiating the biological effects and consequences of small changes in thermal dose at the fine scale when describing the relationship between organism performance and thermal dose. Importantly, not only the intensity and duration (in days) of heat waves, but also daily high temperature exposure time in hours should be incorporated into population dynamic models to accurately predict the effects of heat waves.

5. Conclusion

Predicting the ecological impacts of hot events needs to consider both immediate and potential carry-over effects.Our present result suggested that higher intensity in early larval stage reduced immediate survivals, and importantly produced a general decline with significant zigzag fluctuating patterns in subsequent performance as heat intensity increased. The finding highlights the importance for differentiating the biological effects and consequences of changes in heat intensity at fine scale; not only the frequency and duration of heat waves, but also daily exposure duration should be considered to predict population dynamic under climate change. Given that heat waves impose one of the greatest threats of climate change for natural and agricultural ecosystems, we expect that it will inspire new research on immediate and carry-over effects of heat waves.

Acknowledgements

This research was mainly financially supported by the National Natural Science Foundation of China (31501630 and 31471764), the China Postdoctoral Science Foundation(2015M580156), and the earmarked fund of China Agriculture Research System (CARS-29-bc-4). We appreciate the help from Assistant Researcher Zhang Shuai from Institute of Plant Protection, Chinese Academy of Agricultural Sciences,during experiments and manuscript revision.

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm