The neuroecology of insect-plant interactions: The importance of physiological state and sensory integration

Abstract

Natural behaviorally important stimuli are combinations of cues that are integrated by the nervous system to elicit behavior. Nonetheless, these cues dynamically change in time and space. In turn, the animal’s internal state can cause changes in the encoding and representation of these stimuli. Despite abundant behavioral examples, links between the neural bases of sensory integration and the internal state-dependency of these responses remains an active study area. Recent studies in different insect models have provided new insights into how plasticity and the insect’s internal state may influence odor representation. These studies show that complex stimuli are represented in unique percepts that are different from their sensory channels and that the representations may be modulated by physiological state.

Graphical abstract

Introduction

An animal’s ecology and evolution will have sculpted its nervous system architecture and function. A mechanistic understanding of these ecological relationships provides opportunities to determine neural circuit function and behavior. Mutualism is a critical feature that shapes ecological communities and which many ecosystems depend on [1,2]. Plant-pollinator interactions are one example of an important mutualism that influences ecological community structure and provides an excellent framework to determine how ecologically-relevant cues are processed by the ‘insect’s sensory systems [3]. A fundamental component mediating plant-pollinator interaction can be broken into the floral displays that operate as attractive “signals” to pollinators and the neural and sensory bases of the ‘pollinators’ behaviors [4,5]. The flower’s display, involving scent, color, pattern, morphology, touch, and taste all serve to stimulate the ‘pollinator’s sensory system. For example, specific color preferences can be a dominant floral trait mediating visitation in many diurnal insects, including the bombyliid fly, Megapalpus capensis, an important pollinator for specific daisy flowers [6], whereas, for nocturnal insects, the floral odor acts as an essential long-range cue [7]. These innate responses to flowers involve specific preferences towards certain traits, including ‘flowers’ color or scent, that maintains the specialized associations.

Nonetheless, plant-pollinator associations are not static, but instead, they vary in time. Flowering plants have a peak flowering time throughout a season that can correspond with the abundances of its primary pollinators (Fig. 1A) [8]. A plant may alter its floral traits (color or amount of nectar) over the season, and for an individual flower, the scent and visual display can differ during its flowering time (Fig. 1B) [9–11]. Whereas observational studies of flower visitation and fewer studies of the effects of changing flower traits on pollinator foraging have been conducted, the sensory mechanisms controlling floral visitation under these dynamic conditions remain unclear.

Figure 1. Relationships between mutualistic partners provide realistic stimuli and model systems to uncover the neural circuits in sensory- and state-dependent behaviors.

A. Flowering plants (top) and their pollinators (bottom) can be coupled in time through their phenology. Grey bar denotes timeframe in C and D.

B. Throughout a flower life, the floral signals can change in color or scent. An exemplar for this strategy is lungwort flowers (Pulmonaria spp.), with flowers that change their color as they age. Similarly, the pollinator’s internal state can also impact its sensory-mediated behaviors and interactions with the flowering plant. Image courtesy of Rachelle Louisseize / EyeEm, Getty Images.

C. At shorter timescales, state-dependent processes like circadian rhythm can influence floral traits, such as scent emission and pollinator activity.

D. Responses of the flower and pollinator can be correlated with control of the floral scent pathways and sensitivity of the pollinator’s antennae (expression of Cryptochrome [Cry] and Odorant binding proteins [OBPs]).

The floral traits’ changes are also reflected in the insect pollinator’s local abundances and physiological changes in the ‘pollinator’s sensory systems (Fig. 1). Insect pollinators exhibit plasticity in neural and behavioral responses due to processes like age, mating-status, and diet [12]. It remains unclear how these sensory changes are linked with the dynamic changes in floral advertisements. This review details the recent progress in the links between differences in floral traits with the plasticity of pollinator neural responses. Over different timescales, the dynamics of floral traits are described to illustrate how biologically important stimuli vary in time. In parallel, the pollinator’s sensory systems are also plastic, which may correspond with these ecological changes. These two components were chosen to illustrate: (1) that natural stimuli provide outstanding opportunities to determine how neural circuits process the sensory information to drive behavior; and (2) that the context of the sensory environment – indirectly via the insect physiology and directly via changes to the ecologically-relevant floral stimuli – is dynamic.

Changes in floral stimuli

Flower traits, like color, scent, nectar abundance, can change throughout a season and over the lifespan of a single flower [9–11]. Seasonal changes in visual and olfactory cues can be due to abiotic conditions such as nutrients or water availability [13] or biological factors, such as pollination or herbivory [*14]. These biotic and abiotic conditions can elicit seemingly minor changes in scent or morphology with strong behavioral effects on the pollinators. Even quantitative changes in scent emissions or other floral traits could influence plant-pollinator interactions. For instance, most pollinators respond to the presence of certain scent compounds and the concentration and relative ratios of compounds in the scent, which can even shift from an attractant to a repellant [*15]. Water availability in Ipomopsis and other floral species can negatively correlate with monoterpene emissions, such as a-pinene and beta-ocimene [16,17], that are relevant volatile compounds that are involved in pollinator attraction [18]. Biotic factors such as herbivory or pollination can also strongly and quickly impact floral traits on the order of hours to days. For instance, in diverse angiosperm families, flowers may shift their colors, movement, and decrease in scent emissions and nectar abundances. Flowers of Pulmeneria spp. are exemplars of this phenomenon and change their displays according to floral age or pollination status, which impacts its butterfly pollinator [19]. Herbivory also impacts floral scents and morphologies. For example, flowers may change the emission of monoterpenes (α-pinene and D-limonene) in their scent profile while decreasing their inflorescence size [20]. Because of climatic and biotic impacts, insect pollinators’ floral signals will dynamically change, leading to differences in their interactions with the insect assemblage.

Changes in the neural and sensory bases of behavior

Insect pollinator responses to floral stimuli will vary due to prior experience, circadian rhythms, and internal state-dependent processes, like age, nutrient availability, and mating status. As an insect pollinator ages, developmental programming and senescence in sensory neurons and neuromodulatory systems in the brain may provide slight alterations to the behavior and resulting plant-pollinator interactions. For instance, behavioral responses to visual stimuli by bees, moths, and flies can remain robust after eclosion, although older individuals at the maximum of their lifespan can show rapid senescence in their behavioral responses [21]. This accelerated senescence is reflected in the transcription of olfactory, gustatory, and visual proteins (e.g., odorant receptors, odorant-binding proteins, rhodopsins), which can show the strongest expression immediately before eclosion, but relatively constant expression throughout the ‘pollinator’s lifespan [22,23]. For example, in honeybees, the ultraviolet, blue, and green rhodopsin receptors, gustatory receptors, and odorant-binding proteins (e.g., OBP8 and OBP14) show near-constant expression post-eclosion, and interestingly, differences in gene expression are related to caste rather than age [22,24]. In male Drosophila melanogaster – pollinators of Arum flowers [25,26] – age can impact the specific odorant receptor neurons via expression of PPK25, a member of the degenerin/epithelial sodium channel family (DEG/ENaC) [**27]. Age-dependent upregulation of PPK25 increases the Ca²⁺ influx in the odorant receptor neurons, amplifying the olfactory responses. It is unclear whether this mechanism also occurs in odorant receptor neurons involved in locating food resources.

At shorter timescales (minutes to days), state-dependent processes like circadian rhythm, hunger, and mating status can modulate the sensory systems at both the peripheral and central levels. There has been abundant work on the circadian control of behavior and insects’ neural responses [28]. Most plant-pollinator interactions occur during specific periods during the day. Many flowers are known to display their attractive qualities, such as scent emission and petal opening, in a daily rhythmic fashion [29,30]. Insect pollinators often show time-dependent increases in behavior during those periods, and this can be reflected in enhanced sensory responses and 24-hour rhythmic activity. For example, the moth Manduca sexta shows strong circadian regulation of behavioral attraction to the Petunia axillaris flowers, an important nectar source, and this behavior correlates with the rhythm of two genes involved in circadian regulation; period (per) and timeless (tim), both of which are expressed in the antennae [31]. In Drosophila, the expression of these clock genes in the olfactory receptor neurons is critical for maintaining rhythmicity [32].

Hunger is another process that strongly modulates neural responses to floral stimuli. Starvation increases the attractiveness to food-related stimuli and especially sensitizes olfactory responses. This general process occurs in insect pollinators, but also biting insects (mosquitoes, biting flies), and predatory wasps [33]. Octopamine, a critical neuromodulator in insects, acts in an analogous manner to adrenaline to mammals in its synthesis and function, especially during periods of low glucose. In starved bees, octopamine is involved in locomotor activity, hemolymph glucose concentrations, and may regulate the release of insulin-like neuropeptides [34]. Similar effects occur in starved Drosophila, where the presence of neuropeptide F (NPF) and its receptor increases the gain of the ORNs projecting to the DM1, DM2 and DM4 glomeruli [35,36]. The presence of insulin directly down-regulates the expression of the NPF receptors and thereby modulates the ORN responses and hunger-induced arousal. Octopamine and NPF and its receptor are conserved across diverse insect species (members of Diptera, Lepidoptera, Hymenoptera, and Orthoptera), suggesting that these signaling systems may have similar olfactory functions in diverse pollinating insects [37].

Hunger-induced modulation of neuronal responses also occurs in the central nervous system. Neuromodulators like octopamine and dopamine are integrally linked to the processing of sensory information in the antennal (olfactory) and optic (visual) lobes, and in the downstream memory centers, like the Mushroom Bodies. These neuromodulators are also critical for hunger-induced changes in activity and locating foods. For instance, octopaminergic neurons express the invertebrate analog of glucagon (adipokinetic hormone [ADH]), which is required for starvation-induced hyperactivity [38]. Octopaminergic neurons are expressed in the suboesophageal ganglion but also innervate the mushroom bodies. In Drosophila, knocking out the ADH receptors or the octopaminergic neurons eliminates the starvation-induced hyperactivity [38]. The Mushroom Bodies have strong octopaminergic and dopaminergic input and express insulin-like and ADH receptors, as well as integrating internal state and modulating innate behavioral responses to visual and olfactory stimuli. The Mushroom body Kenyon cells receive input from the antennal and optic lobes and synapse on to Mushroom Body output neurons (MBONs), and specific dopaminergic neurons show strong hunger-dependent modulation of these neurons. For instance, NPF and hunger state promote sugar memory expression by suppressing dopaminergic neurons’ activation, particularly the PPL1 neuron, that project to specific MBONs involved in hunger-induced attraction to food odors [39]. How these dopaminergic and MBONs respond to visual stimuli in underfed or starved conditions remains unclear.

Mating and reproductive status can also shape insect responses to floral stimuli, both behaviorally and neurobiologically. In different moth species, after mating, females show strong behavioral and neurophysiological changes [40]. For instance, they are less attracted to floral stimuli and become more attracted to hostplant odors for egg-laying. What drives these changes? Seminal proteins and peptides have shown to play influential roles in regulating female behaviors. For instance, in honeybee queens, injection of seminal fluid causes decreased sexual receptivity, increased attractiveness of queens to workers, as well as the expression pattern of genes typically expressed during mating [41]. The transfer of sex peptides during mating causes these changes in diverse insect species, including moths, flies, mosquitoes, crickets, ants, and bees [42]. In Drosophila, receptors for the sex peptides are expressed in sensory neurons in the uterus and oviduct that are silenced once activated by the peptide. The sensory neurons make synaptic connections with ascending abdominal ganglion neurons (SAG) that, in turn, project to the PC1 region of the brain [43]. PC1 is involved in aggression, courtship, and egg-laying, and once activated via sex peptide input disinhibits the neural circuits involved in egg-laying [*44]. Presumably, there are links between PC1 and the olfactory and visual brain loci that are critical for behaviors involved in locating suitable egg-laying sites.

Sensory integration of cues

As the insect navigates to a source of food, such as flowers, it integrates multiple streams of information to decide whether to feed. Nonetheless, if the flower is continually responding to environmental stressors that elicit dynamic changes in its different cues (e.g., visual, olfactory), how does the pollinator weigh the various signals to make an appropriate decision? Different sensory modalities’ relative contribution in regulating these decisions can be dictated both in time and space. Like other pollinating insects, mosquitoes respond to floral odors to elicit a strong odor plume tracking response. But the odor also sensitizes the attraction to visual objects [**45,46]. If the objects have other food-related cues (heat, water vapor, gustatory cues), then the mosquitoes will land and feed. In the mosquito, and other insects species like moths and bees, the olfactory input modulates the visual system, but the visual system may not modulate the olfactory responses [47–49]. This non-linear input may reflect the spatial scales these cues are operating: odors provide long-distance information (≫10 m), visual information intermediate (<5 m), and thermal, water vapor, and gustatory cues operating at the source of the odor. When sensory cues come into conflict or are uncoupled, for instance, when the ‘flower’s visual traits do not match the odor or mechanosensory traits, then a ‘pollinator’s response can be altered. For example, in M. sexta moths, when the flower’s visual motion does not match the mechanosensory input, then the ‘moth’s ability to track the flower is attenuated [**50]. Both mechanosensory and visual inputs are linearly summed to enable accurate feeding from the flower. Nonetheless, these behaviors can be modified depending on the physiological state. For instance, in both the example with the mosquitoes and moths, when the insect is satiated, then the behavioral attraction to food resources is suppressed, whereas when the insect is mated, different sensory modalities and inputs may be reweighted as it transitions from locating a flower to finding a site for egg-laying.

Where in the brain are these sensory inputs integrated to elicit robust responses? Several loci in the insect brain may process multimodal information, such as the Mushroom Bodies and Central Complex. As mentioned before, the Mushroom Body receives information from many different modalities, including temperature, odor, visual, mechanosensory, and gustatory inputs. In the Mushroom Body calyces, outputs from the antennal and optic lobes are anatomically separated in the lip and collar regions, respectively, although there is overlap in the basal ring area of the calyces, suggesting the Kenyon cells in this loci might integrate these sensory inputs [51,52]. Nonetheless, the downstream MBONs will also integrate the sensory inputs, and with the strong modulation of activity by neuroamines (e.g., octopamine and dopamine) and neuropeptides, these neural ensembles provide a potential substrate to examine the effects of integration and dynamical sensory inputs [39]. The Central Complex – a region in the midbrain comprised of four neuropils: the protocerebral bridge (PB), the fan-shaped body (FB), the ellipsoid body (EB), and the noduli – also integrates visual, mechanosensory, and olfactory inputs to influence navigation-related behaviors [53,54]. For instance, in the EB, neurons are critical for body position and landmark cues, and coordinated activity between Central Complex areas, via recurrent excitatory connectivity between PB and EB neurons, allowing navigation and speed tuning [55,56]. The Central Complex is also highly innervated by neuromodulatory neurons, such as aminergic and peptidergic neurons, that influence sensory behaviors [57]. In Drosophila, dopamine modulates EB neurons’ activity involved in turning and route decision-making [58]. Although we know very little about how physiological states modulate Central Complex neurons, a recent study has shown that responses to food odors are represented by certain FB neurons modulated by dopaminergic and neuropeptidergic neurons. The upstream modulation of the FB neurons allows the transmission of physiological state (fed, starved), thus enabling appropriate feeding decisions [59]. However, in both the Central Complex and Mushroom Body, it remains an open question for how different sensory modalities are dynamically weighted at other physiological states.

Another critical point is that studies examining sensory integration often rely on isolating specific sensory inputs, commonly via by ablating the antennae or painting the eyes, inhibiting the visual pathway, or eliminating visual cues. Unfortunately, these approaches may fail to reveal encoding processes for sensory integration since sensory systems will adapt and compensate for the loss of a specific modal input. Instead, conflicting and dynamically changing the sensory stimuli to dissociate different modalities while keeping the sensory channels intact – similar to human studies of reaching [60] – may contribute to our understanding of how downstream circuits in the Mushroom Body and Central Complex modulate their responses, and under different states.

Conclusions

The study of sensory integration under different physiological states can benefit from sensory ecology studies to identify how neural circuits dynamically process behaviorally important stimuli. This review’s central goal was to highlight how one ecologically-relevant mutualism, plant-pollinator interactions, provides a tractable system to understand processes involved in state-dependency and integration. Over the last decade, integrative research in these topics – coupling genetic and molecular tools and behavioral assays, combined with new methods for studying neural circuits – has led to insights into the neural bases of navigation, sensory integration, and state-dependency. Future work that incorporates sensory ecology with neurobiology and physiology should lead to an improved understanding of the coding and circuit properties that drive natural behaviors.

Highlights.

• Natural sensory stimuli dynamically change in time.

• An insect’s sensory system may reweight the inputs depending on physiological state and the strength of the sensory input.

• Internal state will modify the sensory representations and allow flexibility in behavioral decision making.

• New methods in sensory integration allows determination of the importance of different cues.