Circadian clocks of both plants and pollinators influence flower seeking behavior of the pollinator hawkmoth Manduca sexta

Abstract

Most plant-pollinator interactions occur during specific periods during the day. To facilitate these interactions, many flowers are known to display their attractive qualities, such as scent emission and petal opening, in a daily rhythmic fashion. However, less is known about how the internal timing mechanisms (the circadian clocks) of plants and animals influence their daily interactions. We examine the role of the circadian clock in modulating the interaction between Petunia and one of its pollinators, the hawkmoth Manduca sexta. We find that desynchronization of the Petunia circadian clock affects moth visitation preference for Petunia flowers. Similarly, moths with circadian time aligned to plants show stronger flower-foraging activities than moths that lack this alignment. Moth locomotor activity is circadian clock-regulated, although it is also strongly repressed by light. Moths show a time-dependent burst increase in flight activity during subjective night. In addition, moth antennal responsiveness to the floral scent compounds exhibits a 24-hour rhythm in both continuous light and dark conditions. This study highlights the importance of the circadian clocks in both plants and animals as a crucial factor in initiating specialized plant-pollinator relationships.

Figure 1

Synchronization of plant and pollinator clocks is important for floral visitation. (A) Diagram of experimental setup: an adult male M. sexta moth at Circadian time 12 (CT12) is given 5 minutes to make a first choice between two stimuli: the control plant (also at CT12) and an experimental plant entrained to another internal time (with a comparative control for CT12). Plants were randomly assorted. (B) Raw data from the choice experiment. Each pair indicates the circadian time of the two plants in the choice assay, a control at CT12 (shown in white), and an experimental plant at another time (shown in black). The number of moths shown is the sum of of moths recorded from independent replicate choice experiments. (C) Preference index of the choice experiment data. Each bar represents the preference for 30+ individuals (tested separately) at one timepoint comparison of the 24-hour experiment. Error bars represent the standard error of the binary distribution. Asterisks denote choices which are significantly different from random in a binomial exact test (*p < 0.5, ***p < 0.001). See Supplementary Fig. 1 for a full schematic of the wind tunnel apparatus.

Figure 2

Transgenic Petunia hybrida with disrupted clocks receive reduced preference from M. sexta vs wild-type P. hybrida. Moths entrained to 12L:12D conditions were given a choice between 2 plants, a wild-type (WT) P. hybrida plant and a transgenic plant with altered clock rhythm: (35 S:PhLHY #37: arrhythmic, 35 S:PhLHY #46: early phase shift, 35 S:PhLHY #47: early phase shift). (A) Emission profiles of the plant lines for each experiment is shown above the corresponding bar (emission profiles synthesized from previously published data. All organisms were entrained to 12L:12D conditions, and the choice experiment carried out at ZT16 (denoted as the white dotted line in the emission profile). (B) The raw data of each choice experiment between clock-altered line and WT P. hybrida. (C) Preference index of the choice experiment data. Each bar represents the preference for 39+ individuals (tested separately) for each choice experiment. Error bars represent the standard error of the binary distribution. Asterisks denote choices which are significantly different from random in a binomial exact test (*p < 0.5, ***p < 0.001). See Supplementary Fig. 1 for a full schematic of the wind tunnel apparatus.

Figure 3

Circadian time of moths affects foraging activity on P. hybrida flowers. (A) Two groups of male M. sexta moths (Moth 1 and Moth 2 groups) entrained to different 12L:12D conditions offset by 12 hours; flowering P. axillaris (Plant) was also entrained to 12L:12D conditions. An arrow indicates that moths were transferred from separate environmental chambers to the experimental chamber 30 minutes prior to the experiment. Actual infrared recording occurred between ZT16 and ZT17 (indicated by the dark gray box). The number of moths visiting flowers was counted in each 10-minute window. (B) Diagram of experimental setup in wind tunnel chamber. 30 minutes prior to the 1-hour experiment, groups of CT4 and CT16 moths were simultaneously introduced to the wind tunnel, and the CT16 plant was introduced at the beginning of the experiment. (C) The 1-hour experiment was divided into six 10-minute segments, and in each segment the percentage of CT4 and CT16 moths that visited flowers is reported (total 10 moths each/experiment). The experiments were biologically repeated three times with independent samples. Asterisks denote the significance value assigned by t-test comparison of CT4 vs. CT16 numbers at each timepoint (**p-value < 0.01, *p-value < 0.05).

Figure 4

M. sexta activity is clock regulated but light repressed. Actograms of the number of male M. sexta moths flying in each 10-minute window of 24 hours or 72 hours. Experiments were repeated independently with 15 male moths three times. The results show the accumulated counts of flying moths in each time window. (A) 12L:12D. (B) Continuous dark for 3 days (3D DD). (C) Continuous light for 3 days (3D LL). (D) 12-hour T-cycle.

Figure 5

The circadian clock and light modulate behavioral response to floral scent. (A) The wind tunnel used for scent pulse experiments with diagrams of the floral and vegetative scent pulse setups administered to the intake of the wind tunnel. All floral scent pulses were administered by adding 15 cut flowers entrained to a night timepoint, ZT16. All vegetative scent pulses consisted of 8 two-month-old plants with no flowers. (B–H) Actograms of flight activity with subjective day/night stimulation pulses of floral scent or vegetative scent emissions. Each experiment consisted of 15 male moths entrained to 12L:12D and was repeated independently three times. The dotted lines indicate the time windows when either night or day scent pulses were given. (B) DD actogram (C) DD actogram with ZT40-41 “floral night pulse” (highlighted in pink). (D) DD actogram with ZT28-29 “floral day pulse”. (E) DD actogram with ZT40-41 “vegetative scent pulse” (highlighted in green). (F) DD actogram with ZT28-29 “vegetative scent pulse”. (G) LL actogram (H) LL actogram with ZT16-17 “floral scent pulse”. (I) LL actogram with ZT4-5 “floral scent pulse”. (J) LL actogram with ZT16-17 “vegetative scent pulse”. (K) LL actogram with ZT4-5 “vegetative scent pulse”.

Figure 6

M. sexta exhibits a circadian rhythm in antennal responsiveness to the floral volatile benzaldehyde, regardless of light conditions. (A) Illustration of electroantennogram (EAG) setup (B) Benzaldehyde EAG in 12L:12D. (C) Benzaldehyde EAG in DD. (D) Benzaldehyde EAG in LL. The results are means ± s.e.m. (n = 12 for each timepoint). Fitted curves in the background are derived from cosinor analysis in CircWave. CircWave analysis for LD curved fit: F-stat = 12.61, p-value = 0.000024, R² = 0.2795. CircWave analysis for DD curved fit: F-stat = 9.42, p-value = 0.000004, R² = 0.3598. CircWave analysis for LL curved fit: F-stat = 12, p-value = 0.000034, R² = 0.2581.

Figure 7

Circadian gene expression analysis in Manduca sexta antenna. qPCR data of (A) per, (B) tim, (C) Orco, (D) GOBP1, and (E) GOBP2 gene expression profiles in continuous dark conditions. The results of each timepoint are relative values to the expression levels of control gene rps13. The results show means ± s.e.m. derived from three biologically independent experiments. Fitted curves in the background of A and B derived from cosinor analysis in CircWave. CircWave analysis for per curved fit: F-stat = 12.81, p-value = 0.000569, R² = 0.6307. CircWave analysis for tim curved fit: F-stat = 17.8, p-value = 0.000109, R² = 0.7036. CircWave was not able to find sine waves for Orco, GOBP1, and GOBP2.