Elevated temperatures diminish the effects of a highly resistant rice variety on the brown planthopper

This study compares the effects of temperature (constant at 15, 20, 25, 30 and 35 °C) on adult longevity, oviposition, and nymph development of the brown planthopper, Nilaparvata lugens, on susceptible and resistant rice varieties. The resistant variety contained the BPH32 gene. In our experiments, nymphs failed to develop to adults at 15, 20 and 35 °C on either variety. Host resistance had its greatest effect in reducing adult survival at 20–25 °C and its greatest effect in reducing nymph weight gain at 25 °C. This corresponded with optimal temperatures for adult survival (20–25 °C) and nymph development (25–30 °C). At 25 and 30 °C, adult females achieved up to three oviposition cycles on the susceptible variety, but only one cycle on the resistant variety. Maximum egg-laying occurred at 30 °C due to larger numbers of egg batches produced during the first oviposition cycle on both the susceptible and resistant varieties, and larger batches during the second and third oviposition cycles on the susceptible variety; however, resistance had its greatest effect in reducing fecundity at 25 °C. This revealed a mismatch between the optimal temperatures for resistance and for egg production in immigrating females. Increasing global temperatures could reduce the effectiveness of anti-herbivore resistance in rice and other crops where such mismatches occur.


Results
Adult survival and oviposition. We examined survival and egg-laying in adult females on the susceptible rice variety IR22, and on the resistant variety IR62 at 15, 20, 25, 30 and 35 °C. In particular, we examined cyclic oviposition responses to the varieties and temperatures. Adult survival declined over the course of the experiment with significant two and three-way interactions due to differences in the rates of decline at different temperatures and for the two varieties (Table 1). Longevity was lower on IR62 and was greatest at 15 °C, but not significantly different for temperatures ≥ 25 °C (Fig. 1a-e). A significant three-way interaction was due to similar patterns in longevity on both hosts at 15 °C, but reduced longevity on the resistant host at 20-30 °C. Whereas total female longevity (time to 100% mortality) peaked at 15 °C, the time to 50% mortality peaked at 20 °C (Fig. 2a). Furthermore, total longevity on IR62 showed a clear increase at 30 °C due to a comparatively extended longevity among late survivors on the variety.
The number of egg batches per female, the size of the egg batches, and the total numbers of batches and of eggs laid varied throughout the course of the experiment (Table 1; Fig. 1). Egg-laying on both IR22 and IR62 was highest at 30 °C ( Fig. 2; Table 2). Individual females on IR22 displayed clear oviposition cycles in the numbers of batches produced and the size of batches. At 25 and 30 °C, peaks were observed at about 1, 9 and 14-15 days after the initiation of experiments, representing 1st, 2nd and 3rd oviposition cycles, respectively ( Fig. 1m-n). At 30 °C, batch sizes increased over successive cycles on IR22 (Fig. 1s). On the resistant variety IR62, adults laid eggs during only a single, initial cycle ( Fig. 1k-o). The larger batches (Fig. 1p-t) and higher numbers of batches ( Fig. 1u- www.nature.com/scientificreports/ vival was relatively constant between 15 and 30 °C, but declined rapidly over the course of 15 days at 35 °C (Table 3; Fig. 3a-e). This produced significant time and temperature effects and a significant [time × temperature] interaction (Table 3). Host variety had no significant effect on nymph survival over the course of the experiment (Table 3); however, fewer nymphs had survived on IR62 by the end of the experiment (Temperature: F 4,36 = 91.632, P < 0.001; Variety: F 1,36 = 7.634, P = 0.009; Interaction: F 4,36 = 0.928, P = 0.459; Fig. 4a). Nymphs gained biomass over the course of the experiment (Table 3; Fig. 3f-j). This was affected by temperature and host variety with the greatest biomass achieved at 25 and 30 °C (Table 3). All two-way and three-way interactions were significant because of similar weight gains at 15 and 35 °C (irrespective of variety) and similar biomass during early parts of the experiment on both varieties, but a later divergence to produce smaller nymphs on IR62 (Table 3). Biomass at the end of the experiment was greatest for planthoppers reared at 25 °C and on IR22 (Temperature: F 4,36 = 193.612, P < 0.001; Variety: F 1,36 = 57.724, P < 0.001; Interaction: F 4,36 = 12.754, P < 0.001; Fig. 4b). We assessed YLS densities in planthopper nymphs exposed to 25 °C and 35 °C for 3 days and maintained at 27 °C for a further 7 days. Nymphs feeding on IR62 had higher YLS densities (F 1,16 = 5.874, P = 0.028) under both temperature regimes. After exposure to 35 °C, YLS densities declined by ca 56% and 84% in planthoppers feeding on IR62 and IR22, respectively (F 1,16 = 67.296, P < 0.001). Despite differences in the proportional losses in symbionts on the two varieties, nymphs on IR62 showed a greater loss in body weight (70%) compared to nymphs on IR22 (43%) (F 1,16 = 10.289, P = 0.005) and at 35 °C (F 1,16 = 4.559, P = 0.049) (Fig. 5).
The times for nymphs to develop to second (N1), third (N2) and fourth (N3) instars were affected by temperature and host variety (Table 3). Development was slower at 15 and 20 °C, but rates were similar at higher temperatures ( Fig. 3k-o). Development to fourth instar was between 1 day (> 20 °C) and 3 days (15 °C) slower on IR62 (Table 3; Fig. 4c). There was a significant [temperature × variety] interaction for development time to second (N2) instar because of similar rates on both varieties at above 30 °C, but slower rates on the resistant variety at between 15 and 25 °C (Table 3; Fig. 3k-o). Interactions were not significant for the other nymph stages. Adults emerged only at 25 and 30 °C (Table 3; Fig. 3h,i). Survivors were monitored at 15 °C until 30 days and at 20 °C until 23 days without any adults emerging. The proportion of surviving nymphs emerging as adults was similar between 25 and 30 °C (F 1,12 = 3.175, P = 0.100), but fewer adults emerged on IR62 (F 1,12 = 38.485, P < 0.01; Fig. 3h,i).

Stability of resistance.
We estimated the separate effects of temperature and host resistance on planthopper fitness by measuring fitness reductions at sub-optimal temperatures in IR22 (compared to optimal) and fitness reductions on IR62 compared to IR22 at each exposure temperature. The optimal temperature for adult longevity was 20 °C (Fig. 6a). Resistance had its greatest effects on longevity at 20 and 25 °C (significantly higher than at 15, 30 or 35 °C). The combined effects of both temperature and resistance gene(s) resulted in highest mortalities at 25 °C (Table 4; Fig. 6a). Egg-laying was highest at 30 °C with significantly lower numbers of eggs laid at 15, 20 and 35 °C. However, resistance was most effective at 25 °C. This mismatch between optimal temperatures for egg laying and host resistance to oviposition produced a significant decline in the combined reducing effects of temperature and resistance at 30 °C (Table 4; Fig. 6b).
Nymph mortality was highest at 35 °C and significantly higher than at 20-30 °C. Temperature had no effect on the functioning of resistance against nymph survival and the overall effects of the resistance gene(s) on survival were low such that combined temperature and resistance effects were greatest at 15 and 35 °C (Table 4; Fig. 6c). 25 °C was the optimal temperature at which nymphs gained biomass (Fig. 6d). Biomass declined at higher (30-35 °C) and lower (15-20 °C) temperatures. Resistance was most effective at 25 °C and moderately effective Table 1. F-values from repeated measures GLM for adult female longevity and oviposition parameters (see Fig. 1). 1 Time = time in days as experiment progresses; run = temporal replicate that includes observations for all temperatures conducted across different climate chambers. 2 Degrees of freedom in parentheses are for batches, eggs and batch size per female (i.e., last three columns). 3 ns = P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. www.nature.com/scientificreports/ at 20 and 30 °C. Nevertheless, the lowest reductions in weight gain were observed at 25 °C because of the large effect of optimal temperature on nymph development and weight gain (Table 4; Fig. 6d).

Discussion
Recently, Horgan et al. 37     www.nature.com/scientificreports/ rice varieties, nymph weights also declined, and development was delayed with no adults emerging. One of the most significant effects of high temperatures (e.g., ≈ 35 °C) is to alter the binding of enzymes (i.e., shifting the Michaelis constant 42 ) affecting interactions between enzymes and their substrates. This will affect the functioning of important metabolites including digestive proteases and juvenile hormones 43,44 , thereby retarding development-as observed in our experiments with planthopper nymphs. Many of the effects of high temperatures can also be attributed to the loss of YLS: we monitored YLS in planthoppers exposed during 3 days to 35 °C and found that densities declined by 56-84%, with greater losses in the susceptible variety. Because of the essential role of YLS in planthopper nutrition, aposymbiotic nymphs fail to gain weight or develop to adults 21,[24][25][26] . Furthermore, YLS are abundant in gravid females and are passed to the eggs prior to oviposition 24 . The removal of YLS from females does not affect their normal survival and development; however, it does affect reproduction and egg development 25 . We did not examine hatchability in the present study, but because YLS densities were significantly reduced at 35 °C, we suggest that any eggs produced at that temperature were unlikely to develop. The effects of temperature on hatchability of brown planthopper eggs have been reviewed by Horgan et al. 37 -hatchability declines to < 40% at temperatures of ≥ 35 °C. Low temperatures directly affect insect herbivores by increasing lags in neural signal transmission and by reducing the insect's ability to generate action potentials in muscle 42,45 . Juvenile hormones can also fail at certain low temperatures 43 . However, the threshold temperatures for such effects are species-specific 45 . Based on the failure of nymphs to gain weight in our experiments, we suggest that neural and muscle functions were already reduced at 15 °C; this caused a decline in food intake, and the nymphs appeared sluggish and were less responsive to mechanical stimuli. At both 15 and 20 °C, nymphs had delayed development and failed to emerge as adults. Such effects can also be due to changes in rice plant physiology at low temperatures including a reduction in concentrations of soluble sugars (i.e., < 20°C 34 ). For example, similar survival, growth rates and development rates of nymphs on IR22 and IR62 at 20 °C in our experiments, indicate that although nymphs consumed sufficient phloem to gain body weight, the phloem sap was of relatively poor quality at that temperature, even on IR22. Unlike nymphs, adult planthoppers survived for longer at low temperatures; however, despite this longevity, the adults displayed only weak oviposition cycles (20 °C) or displayed only an initial, first cycle of egg laying (15 °C). These patterns depict a lower success in ovariole development at temperatures of ≤ 20 °C, even on IR22. Nymphs on IR22 gained the greatest biomass at 25 °C. The nymphs also had high survival and rapid development at 25 °C, albeit with marginally faster development at 30 °C. Meanwhile, adults deposited larger egg batches (on both varieties) at 30 °C. Changes in planthopper fitness at these temperatures (25-30 °C) are important for understanding the consequences of global warming for herbivore pests. As temperatures increase between the lowest and highest tolerable temperatures for normal planthopper development, the insects are predicted to increase feeding activity and may gain increased nutrients from their rice host 46 . For example, concentrations of soluble sugars increase in rice and concentrations of several defensive allelochemicals decline as temperatures rise (including in susceptible varieties 33,47 ). Furthermore, the activity of planthopper feeding effectors such as endo-β-1,4-glucanase increases as temperatures rise (i.e., from 27 to 37 °C) 48 . We suggest that changes such as these increase egg-laying and nymph development rates at 30 °C.
Our results indicate that IR62 maintains strong resistance against the Laguna (Philippines) population of the brown planthopper. In our study, IR62 exhibited clear antibiosis effects on planthopper nymphs. Rice responds to planthopper feeding by activating both the Jasmonic Acid (JA) and Salicylic Acid (SA) pathways. This induces a range of responses in both susceptible and resistant varieties, including augmentation in the concentrations Table 3. Results (F-values) of repeated measures GLMs for nymph survival and biomass over 15 days with results (F-values) of multivariate GLM for nymph development stages (nymphs 2 to 4)(see Fig. 3). 1 Time = time in days as experiment progresses; run = temporal replicate that includes observations for all temperatures conducted across different climate chambers. 2 ns = P > 0.05, * = P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. www.nature.com/scientificreports/ of a wide variety of defensive secondary chemicals; however, compared to susceptible varieties the rate and magnitude of such responses is higher in varieties with the BPH32 gene 29 . Rice lines with the BPH32 gene also produce a range of constitutively expressed defense chemicals. For example, Stevenson et al. 28 have shown that high schaftoside concentrations in varieties derived from Rathu Heenati were directly associated with planthopper mortality 28 . Furthermore, in a study by Saxena and Okech 30 , volatiles from Rathu Heenati and PTB33 (both of which share resistance genes with IR62) were shown to increase adult and nymph mortality and reduce adult feeding. The volatiles also reduced female settling 30 . Low and high temperatures can affect the production of constitutively expressed defensive secondary metabolites in rice 33,34 and may alter the efficiency of the JA and SA pathways 49 . However, the apparent loss of host resistance at 35 °C in our study was mainly due to severe direct effects of very high temperatures on the planthopper and its YLS that obscured any effects of resistance on nymphs, and much of the effects on adults. However, resistance still functioned to reduce fitness (by ≈ 20%) at the higher temperature. Similarly, low feeding rates at 15 °C, likely reduced nymph exposure to rice defenses and obscured any effects of the resistance gene(s) at low temperatures. IR62 reduces egg-laying of planthoppers in two ways: firstly, antixenosis deters females from settling and ovipositing on the plant; secondly, antibiosis reduces nutrient intake by gravid females and the conversion of nutrients to eggs. In choice studies, avirulent planthoppers tend to lay fewer eggs on IR62 than on susceptible varieties (e.g., TN1 and IR22) -although the effect can be weak 21 . In our experiments, adults were confined to their host plants. Reduced egg-laying in such non-choice experiments is mainly due to a lower production of eggs while feeding on the adult host plant. Our results clearly indicate that adults (≈ 20%) on IR22 exhibited three oviposition cycles (at 25 and 30 °C, and to some extent at 20 °C). Similar cycles have been observed in the green rice leafhopper, Nephotettix cincticeps 50 , but to our knowledge, they have not been observed previously in   www.nature.com/scientificreports/ the brown planthopper. Based on the feeding history of the insects in our experiments, these oviposition cycles were due to resources attained during nymph development on the natal host (in our case on TN1) with further ovariole development on the adult host (1st cycle), as well as acquisition of resources from the adult host during the 2nd and 3rd cycles. Planthoppers on IR22 produce large numbers of eggs relative to planthoppers on TN1 and other susceptible hosts 23 , supporting the idea that the large numbers of eggs produced during the 2nd and 3rd cycles were due to resources from the adult host, IR22. On IR62, adults failed to produce a 2nd or 3rd cycle -even where they survived for up to 10 (25 °C) or 15 (30 °C) days. This indicates that the eggs deposited on IR62 were mainly derived from ovarioles produced during the pre-oviposition stage when planthoppers were still on the natal host TN1. Defenses in IR62 reduced the number of eggs per batch at 20, 25 and 35 °C, but had no apparent effect at 15 °C. Reductions in the size of batches on IR62 (compared to the 1st cycle on IR22) suggest that the planthoppers failed to acquire extra nutrients to support ovariole maturation-supporting the idea that resistance is related to antifeedants in the host phloem. This effect was greatest at 25 °C (see also Lu et al. 51 ). However, at 30 °C, egg numbers were relatively high on both IR22 and IR62 indicating that the planthoppers successfully acquired further resources from the plants to increase egg production at that temperature. Because IR62 was effective against nymphs at 30 °C, but less effective against adults at the same temperature, we suggest that the www.nature.com/scientificreports/ defenses of IR62 were compromised by the increased activity of adults at the higher temperature together with a greater feeding capacity (as suggested by trends in egg production). In a similar case, Havko et al. 35 found that at relatively high temperatures (i.e., 29 °C versus 22 °C), high feeding rates of the cabbage looper (Trichoplusia ni) on Arabidopsis overwhelmed JA-mediated defenses; but the high temperature did not affect the expression of JA-responsive genes or the production of glucosinolates. One important mechanism by which phloem feeders neutralize host defenses is by consuming large amounts of xylem to dilute phloem-based toxins. High production of xylem-based honeydew has been reported for planthoppers feeding on IR62, even after several generations of adaptation 23 . Detailed studies on the production and actions of defense metabolites across gradients of tolerable temperatures are recommended to further elucidate the mechanisms leading to a lower efficiency of antibiosis defenses against adult planthoppers in rice at elevated temperatures. Currently, a range of near-isogenic rice lines with planthopper resistant genes are available to support further research in this area 17,32 . IR62 has been widely planted in Cambodia and in Mindanao (Philippines); although current rates of adoption are not known 19 . Temperatures in these regions have increased in recent decades 52 Figure 6. Reductions in the fitness of Nilaparvata lugens on IR62 over a temperature gradient. Graphs indicate the total reductions (violet circles) in adult longevity (a), the number of eggs laid (b), nymph survival (c) and nymph biomass (d). In each case, this is composed of a fitness reductions due to temperature (blue circlesestimated based on reductions relative to optimal temperatures on a highly susceptible variety) and reductions due to the resistance gene(s) (red circles -estimated as the total reduction minus the temperature-related reduction in fitness). Standard errors are shown (N = 4 for a and b, N = 5 for c and d). Lowercase letters indicate homogenous temperature groups (Tukey, P < 0.05). Table 4. F-values from univariate GLMs for factors producing a decline in fitness of Nilaparvata lugens on resistant rice over a range of temperatures (see also Fig. 5). 1 ns = P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. 2 Degrees of freedom = 4,12; no significant effect of run. 3 Degrees of freedom = 4,16; no significant effect of run. www.nature.com/scientificreports/ same period, the annual temperature anomaly increased by about 0.23 °C each decade in Cambodia 52 . During 2019, temperatures of 30 + °C were recorded in Malaybalay (Mindanao) on 148 days and in Kampong-Chhnang (Cambodia) on 349 days, with minimum temperatures above 27-28 °C during extended periods at the latter site between February and May 54 . At such temperatures, BPH32-derived resistance against immigrating females could be comprised for much of the time. Potential climate-related reductions in the efficiency of host plant resistance, as depicted in the present study, indicate that rice producers must broaden their pest management actions to increase the resilience of future crops. This is further highlighted by observations that several agrochemicals will increase the tolerance of brown planthoppers to adverse high temperatures 55,56 . Rice production systems that incorporate host plant resistance as a component of landscape approaches to promote the diversity of natural enemies 13,15,16 will enhance the resilience (including the durability) of novel resistance genes and prevent losses from insect pests as global temperatures continue to increase.

Materials and methods
Brown planthopper. We 37 . We also conducted experiments at 40 °C; however, nymphs and adults failed to survive beyond 2 days and the adults did not lay eggs. Further information on the responses by brown planthopper adults and nymphs to temperatures ranging from 15 to 40 °C have been presented by Horgan et al. 37 . The bioassays were conducted as follows: Oviposition experiments: Plants of each variety were individually covered with acetate rearing cages (50 × 10 cm: H × D). The cages had a mesh top to allow air circulation. A single mated gravid female was introduced to each cage at 20 DAS using a suction aspirator. Temperatures were 15, 20, 25, 30 and 35 °C. Temperatures were replicated across the chambers (i.e., N = 4), with chambers assigned randomly to each temperature. Each replicate consisted of continual observations from one day to 20 days after caging the females. The plants under each acetate cage were changed daily and the condition of the adults noted (i.e., surviving or dead). Plants that were exposed to females were dissected to count the numbers of egg clusters and the numbers of eggs per cluster. Replicates for each complete set of temperatures (henceforth a 'run') took ≈ 60 days to complete.
Nymph survival and development: Sufficient rice seedlings were prepared to be able to assess daily nymph survival and development at each temperature through destructive sampling (i.e., 15 days × 5 subsamples × 2 varieties = 150 seedlings per temperature replicate [up to 300 seedlings for three of the replicates at 15 °C]). Each temperature was replicated five times (N = 5) as described above. Ten newly emerged nymphs were placed on plants (one plant per cage) of each variety at 20 DAS under each temperature treatment. The plants were covered with acetate rearing cages (50 × 10 cm: H × D) with mesh widows for ventilation. Nymphs were allowed to feed and develop for 15 days with groups of ten plants (susceptible and resistant) randomly selected for sampling per temperature, per day. The number of survivors and their developmental stages were recorded and the insects were dried in an oven for 5 days and weighed to estimate total nymph biomass per plant. Each run usually took ≈ 60 days to complete.
Yeast like symbiont densities after optimal and high temperatures. Ten  www.nature.com/scientificreports/ for 3 days after which the plants were placed in an insectary at 27 °C until neonates were 10 days old. Nymphs were allowed to develop for a total of 10 days to improve estimates of symbiont densities. The nymphs were then weighed. Yeast-like symbiont densities were estimated using the method described by Ferreter et al. 23 . The nymphs were homogenized in 500 μl physiological saline solution (0.9% NaCL). For each sample, an aliquot of 10 μl was transferred to a hemocytometer and the YLS counted under a compound microscope (× 40 magnification) 23 .

Data analyses.
Results from the oviposition and nymph survival experiments were analyzed using repeated measures general linear models (GLM) with days after first exposure as the repeated measure and temperature, variety and their interaction as main factors. Because experimental runs took up to 60 days to complete, we included 'run' as a blocking factor in our analyses to control for possible changes in the planthopper colony during the time that the research was conducted (e.g., short-term temperature assimilation, etc.). Survival, the numbers of egg batches, and the numbers of eggs laid were analyzed only for the first 6 days of the experiment because of low survival after that time at some temperatures. Similarly, because egg batches were produced by planthoppers across all replicates and treatments for only 4 days, batch numbers per female, eggs per female, and the size of egg batches were each analyzed only for the first 4 days -representing the peak of the first oviposition cycle. We analyzed nymph survival and nymph biomass during 15 days using repeated measures GLMs. Prior to analyses, survival was arcsine-transformed; nymph biomass, the number of egg batches, the numbers of eggs, and batch size per female were log-transformed; and the total number of batches and eggs were ranked. Nymph biomass and survival at the end of 15 days, and the total number of batches and eggs laid, adult longevity, batches and eggs per females, and batch size at the end of each experiment (including all days) were further analyzed using univariate GLM. For these analyses, longevity was measured as the time to 50% and 100% mortality. Batch number, total egg number, the number of eggs per planthopper, and batch size were logtransformed before analyses. We analyzed nymph development based on the time for 50% of nymphs to develop to second (N1), third (N2) and fourth instars (N3). We used a multiple GLM to analyze nymph development times (N1, N2, N3) and univariate GLM to analyze the proportion of nymphs developing to adults at 25 and 30 °C with 'run' included as a blocking factor (see above) 37 .
We estimated the separate effects of temperature and host plant resistance on planthopper fitness. To estimate the effects of temperature we calculated percentage reductions in fitness measured on IR22 at each temperature (T) compared to optimal temperatures for each parameter (i.e., the reduction in fitness due to sub-optimal temperature at T a = 1-(fitnessT a /fitnessT optimal ), where a = 15 °C, 20 °C, etc.). To estimate the effects of host resistance, we calculated percent reductions in fitness at each temperature by comparing fitness measures on IR62 and IR22 (i.e., the reduction in fitness due to resistance at T a = 1-(fitness IR62 /fitness IR22 ) Ta , where a = 15 °C, 20 °C, etc.). Total reductions in fitness were then determined as the sum of both fitness reductions at each temperature. Reductions calculated as such for adult longevity (until 50% mortality), total number of eggs laid, nymph survival at 15 days, and nymph weight at 15 days, were analyzed using univariate GLMs.
Post-hoc Tukey tests were performed for the factor 'temperature' . Residuals were plotted following parametric analyses to test for normality and homogeneity. www.nature.com/scientificreports/