Voltage-activated potassium channels (Kv channels) in the microvillar photoreceptors of arthropods are responsible for repolarization and regulation of photoreceptor signaling bandwidth. On the basis of analyzing Kv channels in dipteran flies, it was suggested that diurnal, rapidly flying insects predominantly express sustained K+ conductances, whereas crepuscular and nocturnally active animals exhibit strongly inactivating Kv conductances. The latter was suggested to function for minimizing cellular energy consumption. In this study we further explore the evolutionary adaptations of the photoreceptor channelome to visual ecology and behavior by comparing K+ conductances in 15 phylogenetically diverse insects, using patch-clamp recordings from dissociated ommatidia. We show that rapid diurnal flyers such as the blowfly (Calliphora vicina) and the honeybee (Apis mellifera) express relatively large noninactivating Kv conductances, conforming to the earlier hypothesis in Diptera. Nocturnal and/or slow-moving species do not in general exhibit stronger Kv conductance inactivation in the physiological membrane voltage range, but the photoreceptors in species that are known to rely more on vision behaviorally had higher densities of sustained Kv conductances than photoreceptors of less visually guided species. No statistically significant trends related to visual performance could be identified for the rapidly inactivating Kv conductances. Counterintuitively, strong negative correlations were observed between photoreceptor capacitance and specific membrane conductance for both sustained and inactivating fractions of Kv conductance, suggesting insignificant evolutionary pressure to offset negative effects of high capacitance on membrane filtering with increased conductance.
- insect photoreceptor
- potassium channels
- compound eye
- visual ecology
early findings in Limulus (Fain and Lisman 1981; Pepose and Lisman 1978) and mollusk (Alkon et al. 1985; Nasi 1991) photoreceptors demonstrated the presence of voltage-gated potassium channels (Kv channels) that could potentially modify the responses to light. However, it was only with the characterization of Kv channels in fly photoreceptors and their physiological investigation (Hardie et al. 1991; Laughlin and Weckström 1993; Weckström et al. 1991; Weckström and Laughlin 1995) that their importance in counteracting the depolarizing light-induced current (LIC) became apparent. Because the equilibrium potential of K+ in insect photoreceptors is near or below the dark resting potential, the increased Kv conductance, as a consequence of activation of the Kv channels, tends to counteract the depolarizing light response (Weckström et al. 1991). Concurrently, due to the lowered membrane resistance, photoreceptor signaling speed increases. Since the temporal properties of the membrane are dictated by the product of membrane resistance and capacitance, the Kv conductance could be seen as a compensation mechanism for large capacitance arising from the highly convoluted microvillar membrane of photoreceptors.
In the previous comparative study, the Kv conductances of the photoreceptors of 20 species of Diptera were characterized (Laughlin and Weckström 1993), showing that in day-active fast-flying insects the photoreceptors strongly expressed a slowly or noninactivating Kv conductance, whereas those of dark-active or crepuscular dipterans had large inactivating currents. In the fruit fly (Drosophila melanogaster), the fast-inactivating current is based on the expression of the Shaker gene (Niven et al. 2003), whereas the slowly inactivating delayed rectifier current is based on Shab gene (Vähäsöyrinki et al. 2006). The R1–R6 photoreceptors of Drosophila express, in addition, a very small totally noninactivating current with unknown molecular identity. The slow, central R7/R8 photoreceptors in both Drosophila and the blowfly (Calliphora vicina) express a fast-activating but slowly inactivating current not based on Shaker (Anderson and Hardie 1996).
Activation of Kv channels causes a leaky membrane, and the resulting ionic currents in turn drive ion pumps. The energy consumption required for ion pumping can be very large compared with the total metabolic cost of brain activity, and this has been proposed to be one reason guiding the selection of different channel types in evolution (Laughlin et al. 1998; Niven et al. 2003).
Recent findings revealed a large diversity of Kv channels, which does not seem to conform well to the originally published comparative view. Especially, the studies on the dim-active American cockroach (Periplaneta americana) are in stark contrast with the notion derived from the studies on dipterans (Immonen et al. 2014a; Salmela et al. 2012). Although the functional variability among cockroach photoreceptors is very large (Heimonen et al. 2006), the majority of photoreceptors express a dominating delayed rectifier current, akin to that in day-active flies. Its molecular basis is not known, although preliminary characterization and pharmacology suggest that it may not belong to the Shaker-related superfamily of Kv channels, which includes the Shaker, Shab, Shaw, and Shal channel families. This raises several questions that can be approached with a wider comparative study. Possibly the functional properties of the expressed Kv channels are related to the photoreceptor size and capacitance, or they depend on the rhabdom type, being either closed or open. Also, maybe the behavioral demands as dictated by slow or fast movements, walking or flying, or some other such characteristic are crucial in selecting the expressed Kv channel types in photoreceptors, guided by the balance between information processing and energy consumption (Niven et al. 2007; Niven and Laughlin 2008).
In this work we have used 15 species of various orders of insects characterized by highly variable lifestyles: the flies D. melanogaster and C. vicina, the butterfly Papilio xuthus, the honeybee Apis mellifera, the cockroaches P. americana, Gromphadorhina portentosa, and Ectobius lapponicus, the field crickets Gryllus bimaculatus and G. integer, the bush cricket Metrioptera roeselii, the stick insects Carausius morosus and Peruphasma schultei, the water boatman Notonecta glauca, the lesser water boatman Corixa punctata, and the water strider Gerris lacustris. The Kv conductance of the photoreceptors of these species was characterized in whole cell patch-clamp experiments on isolated ommatidia, and the species were classified according to the dominance of vision in their lifestyle. The combination of electrophysiology and visual ecology suggests a novel paradigm for Kv channel selection in insects.
MATERIALS AND METHODS
Part of the data presented in this study were obtained from previously published work on C. morosus (Frolov et al. 2012), C. punctata (Frolov 2015), D. melanogaster (Krause et al. 2008), G. lacustris (Frolov and Weckström 2014), G. bimaculatus (Frolov et al. 2014), N. glauca (Immonen et al. 2014a), and P. americana (Immonen et al. 2014b). Additional experiments were performed on P. xuthus, generously provided by Prof. Kentaro Arikawa (Sokendai, Hayama, Japan), G. integer, generously provided by Petri Niemelä (Munich, Germany), C. vicina, P. schultei, and G. portentosa, purchased from Blades Biological (Edenbridge, UK), and A. mellifera, E. lapponicus, and M. roeselii, which were caught locally in Oulu (Finland). The work was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki).
Ommatidia were dissociated, and whole cell recordings were performed as described previously (Frolov et al. 2012). In brief, Sensapex micromanipulators (Oulu, Finland), an Axopatch 1-D patch-clamp amplifier, and pClamp 9.2 software (both Axon Instruments/Molecular Devices, CA) were used for data acquisition and analysis. Patch electrodes were fabricated from thin-walled borosilicate glass (World Precision Instruments, Sarasota, FL). Electrodes used had resistance between 5 and 15 MΩ. Bath solution contained (in mM) 120 NaCl, 5 KCl, 4 MgCl2, 1.5 CaCl2, 10 N-Tris-(hydroxymethyl)-methyl-2-aminoethanesulfonic acid (TES), 25 proline, and 5 alanine, pH 7.15. Patch pipette solution contained (in mM) 140 KCl (K-gluconate for D. melanogaster and C. vicina recordings), 10 TES, 2 MgCl2, 4 Mg-ATP, 0.4 Na-GTP, and 1 NAD, pH 7.15. All chemicals were purchased from Sigma Aldrich (St. Louis, MO). The liquid junction potential (LJP) between bath and intracellular solution was −4 mV (−16 mV for D. melanogaster and C. vicina recordings). Voltage values cited in the text were not corrected for the LJP, except in the case of D. melanogaster and C. vicina, for which all current-voltage relationships were shifted by −10 mV to compensate for different recording conditions. The series resistance was compensated by 80% or more. After compensation, the residual series resistance was typically lower than 10 MΩ. Kv currents were elicited from a holding potential of −70 mV, although in some experiments a holding potential of −80 or −60 mV was used. The testing voltage pulses were preceded by a prepulse to −100 or −110 mV to recover the inactivating fraction of the Kv conductance. Recordings were performed at room temperature (20–22°C).
Membrane capacitance was calculated from the total charge flowing during capacitive transients using voltage steps from −100 to −80/−70 mV in recordings without capacitance compensation. In the case of D. melanogaster, only compensated recordings were available, and since the photoreceptor capacitance was almost invariably <100 pF, it could not be determined reliably. For D. melanogaster we therefore used an average capacitance value of 60 pF (Juusola and Hardie 2001; Vähäsöyrinki et al. 2006). In 6 of the 9 recorded photoreceptors of C. vicina, the average capacitance was 108 pF. The capacitance could not be determined due to the lack of uncompensated recordings in the 3 other cells, to which we assigned a capacitance value of 100 pF (maximum capacitance compensation range of Axopatch 1-D patch-clamp amplifier). In the case of C. punctata, whose photoreceptors are characterized by a very high average capacitance, 455 ± 220 pF (Frolov 2015), voltage-clamp time constants were quite large despite capacitance compensation. This could conceivably result in an underestimation of the inactivating fraction of the current. To minimize the effects of relatively slow clamping on the estimates of inactivating current, we selected 9 cells with the fastest membrane charging out of the total sample of 21. The average capacitance of this subsample was 417 ± 234 pF. This subsample was used in the analysis of the inactivating conductance, while the complete sample was used in the analysis of the sustained conductance.
The Spearman's rank correlation coefficient (ρ) was calculated as described previously (Myers and Well 2003). The Spearman's ρ was considered significantly different from zero when the P value was <0.05.
Methodology for ranking of visual ecological and behavioral traits.
To assess the visual ecological and behavioral traits of the insects involved in the study, we used a qualitative method that yielded a rank order for the list of species. In addition to the lifestyle and mobility, three visually guided behavior aspects were assessed: general behavioral reliance on vision in everyday tasks (“general behavior”), an ability to rapidly change the course of movement/flight in response to sudden appearance of obstacles or visual stimuli (“maneuvers”), and robustness of visual escape reactions (“escaping”). Point grading was used to evaluate the lifestyle and visually demanding traits (Table 1). Lifestyle and behaviors requiring fast information processing, superior spatial and temporal resolution, and visual precision are highlighted in italics. Each of these traits received a “+1” grade. Nocturnal animals and those showing an obvious lack of a specific visually demanding trait or behavior are highlighted in bold italics and received a “−1” grade for every such trait. If a behavior was of uncertain visual loading or not described, it was graded 0 and not highlighted. The grades were totaled to yield the “visual score.”
Properties of Kv currents in photoreceptors.
To evaluate properties of Kv conductances in the context of visual ecology and behavior of the species, it is necessary 1) to meaningfully compare Kv currents and 2) to conduct an analysis of visual ecology and behavior, which would permit appropriate ranking of the species according to their visual performance. Below we first address the issue of Kv current comparison and then present the results of quantification and comparison of visual ecology and behavior. Importantly, in this work no attempt was made to separate the transient (IA) and delayed rectifier (IDR) Kv currents according to their molecular origins, since in different species they are likely to be mediated by ion channels with dissimilar kinetics. Similarly, we ignored the inward rectifier (Salmela et al. 2012) and leak currents. Moreover, the molecular identities of channels producing the currents were considered largely irrelevant in this work for the functional modifications of the voltage responses, although the identity may be important in diurnal or adaptation related modulation of the channel properties (Cuttle et al. 1995).
Kv currents were recorded from the photoreceptors of 15 different insect species from −80 to 30 mV in 10-mV increments. The averaged currents in Fig. 1 demonstrate dramatic differences in the character of Kv currents between the species, specifically, large variations in voltage dependencies of kinetics and amplitude as well as in the timing of current decay. However, most of the apparent differences appear in a voltage range (roughly above 0 mV) that is highly susceptible to series resistance errors and not very meaningful with respect to the normal functional range of photoreceptors. Therefore, we analyzed Kv current properties within a voltage range from −60 mV (resting potential) to −20 mV, which roughly represents the range where photoreceptors usually respond to varying intensities and contrasts of light. Thus this voltage range will be referred to as the physiological voltage range (PVR). The PVR is discussed in comparative context in discussion, Ranking of visual ecological and behavioral traits: the rationale. Figure 2 shows Kv currents in PVR within 400 ms of beginning of voltage pulses, with emphasis on the first and the last hundred milliseconds to better visualize the differences between the species. Although the main quantifiable parameters of the total Kv current are the amplitude, half-activation potential, and rates of activation and inactivation, the only parameters that matter for the function of Kv channels are the total momentary conductance and the speed of its change with voltage and time. Both the maximal conductance and half-activation potential have little utility for our present purpose, because the former is usually reached at positive membrane voltages, far beyond the PVR of the photoreceptor, whereas the latter is dependent on the maximal conductance. Therefore, our comparative analysis of Kv conductances was limited to evaluation and ranking of the inactivating and sustained fractions of the total K+ conductance within the PVR.
Sustained Kv conductance.
Kv currents were measured at the end of 400 ms depolarizing voltage pulses to obtain sustained Kv conductance values. Figure 3A shows profound differences in the voltage dependencies of sustained Kv conductances; note especially the high variation in the steepness of conductance-voltage relation, meaning that each species had different proportion of the sustained conductance activated at each voltage. For instance, the sustained Kv current in the fast-flying and highly visual butterfly P. xuthus was characterized by the average half-activation potential of −40 mV, resulting in the highest current at −60 and −50 mV. However, its maximum conductance was relatively low and at −20 mV the butterfly's sustained Kv current was ranked only the sixth. On the other hand, in the cockroach P. americana, sustained Kv current was among the lowest at −60 mV but became the second largest at −30 mV.
Because of the high variation in voltage dependencies, evaluation of Kv conductances requires something more than just taking into account the half-activation voltage or the maximum conductance value. To quantify and rank conductances within the PVR, we have designed the following algorithm. First, the average conductance or conductance density (conductance divided by capacitance) values at a given membrane potential were ordered from highest to smallest and the rank number assigned accordingly. Thus the largest conductance was numbered 1 and the smallest was numbered 15. This was repeated at each voltage level between −60 and −20 mV. In this way we obtained five rank values for each species. The rank values were then averaged over the voltage range, yielding the “rank score,” according to which conductances (Fig. 3B) and conductance densities (Fig. 3C) were ordered once more; see the actual rank score values in Table 1 and above the composite bars in Fig. 3, B and C. As a result, equal weights were given to each membrane potential, avoiding the bias arising from having disproportionally large values at one or several levels, which would have skewed the rating if actual conductances/conductance densities had been used for ranking. For instance, it follows from Fig. 3B that if actual conductances were first averaged over the PVR and then ranked, the sustained Kv current in P. xuthus would be in the sixth position instead of the third position now assigned.
Importantly, there were large differences in the photoreceptor capacitance, from ∼60 pF in D. melanogaster, E. lapponicus, and G. lacustris to 455 pF in C. punctata, implying proportional differences in photoreceptor membrane area. Because ion channels are expressed in the membrane, to account for dissimilarities in specific membrane performance, which can be evaluated as the membrane response rate, or speed, in the dark, proportional to the inverse of the membrane time constant, it was necessary to compare not only conductances per se but also their densities, which essentially represent the corresponding specific membrane conductances. When the average membrane capacitance was correlated to the average sustained conductance/conductance density in the PVR, which was obtained by averaging conductance/conductance density values at 5 voltage levels from −60 to −20 mV, a strong negative correlation emerged for conductance density (ρ = −0.74, P = 0.001) but not for conductance (ρ = −0.12, P = 0.5; Fig. 3, D and E). These results indicate that species with relatively small photoreceptors tend to have higher specific membrane conductances than species with larger photoreceptors.
Inactivating Kv conductance.
In most insect photoreceptors the total Kv current consists of a sustained or relatively slow inactivating fraction and a transient component. To measure the inactivating Kv current, currents at the end of 400-ms pulses were subtracted from the peak currents. Whereas in photoreceptors with relatively low capacitance values (within the compensable 100-pF range of the amplifier) membrane potential was usually settled within 1 or 2 ms after the onset of command pulse and amplitudes of maximal currents could be obtained in a straightforward manner, relatively slow membrane charging in larger cells often required digital subtraction of the capacitive transient from the recorded current traces to reveal peak currents. An example of such a procedure is shown in Fig. 4A for the average Kv current in C. punctata photoreceptors, which are characterized by the largest average capacitance among the species studied. The average time constant of membrane charging in C. punctata, which was determined by fitting the decay of capacitive transient with a single exponential function at −70 mV, was 3 ms, indicating that membrane potential was adequately clamped by 10 ms (Fig. 4A, dark gray dashed line). Although slow voltage clamp inevitably results in underestimation of fast peak currents, the actual error depends on the interplay between the speed of voltage clamp, access resistance, and channel activation rate. This error is expected to be quite small when Kv currents are evaluated in the PVR, where currents are relatively small, activate slowly, and have prolonged peaks (Fig. 4A).
Inactivating current conductances and conductance densities were ranked in the same way as in the case of sustained current (Table 1) except that due to difficulties with reliable determination of currents at −60 mV, inactivating conductances were ranked in the range from −50 to −20 mV. A strong negative correlation was found between the average photoreceptor capacitance and the average inactivating conductance density within the PVR (ρ = −0.78, P < 0.001) and a weaker one between capacitance and conductance per se (ρ = −0.51, P = 0.052; Fig. 4, D and E), indicating that species with smaller photoreceptors tend to have higher inactivating conductance per se and specific inactivating conductance than species with larger photoreceptors.
Ranking of visual ecological and behavioral traits.
The methodology and rationale for ranking of visual ecological and behavioral traits are described in materials and methods and discussion, respectively. The data in Table 1 show that A. mellifera, C. vicina, and P. xuthus are highly visual, strictly diurnal, fast-flying insects with large apposition eyes, small interommatidial angles with less than 1° in zones of highest acuity (Land 1997), and prominent visually guided behavior in general and with regard to rapid in-air maneuvering and escape reactions. In contrast, D. melanogaster, though displaying impressive aerobatics (Dubnau 2014; Fry et al. 2003), is a relatively slow flier, which received a zero score for the “flight” category. It should be noted that although small insects usually cannot fly fast, D. melanogaster seems to fly slower than hematophagic insects of similar or even smaller size, such as black flies (e.g., Simuliidae) and biting midges (e.g., Ceratopogonidae).
Visually guided maneuvering during movement was judged to be limited to the four above-mentioned species only because other fast-moving insects (G. lacustris, N. glauca, C. punctata, and P. americana) appear unable to change direction of the movement once it is initiated. This is most obvious for the surface dwellers G. lacustris and N. glauca, where a trajectory of movement normally consists of a number of interconnected straight inertial runs. Whether such species can regulate the length of each run during its implementation in response to environmental signals remains to be investigated.
The American cockroach was assigned a −1 grade for visually guided escaping, since its extremely rapid escape reactions are mediated not by vision but through the antennal mechanosensory pathway. This represents a rather special case, because the animal relies more on tactile and olfactory cues in behavior while having large eyes and sensitive vision in the dark (Honkanen et al. 2014; Mote 1990; Riemay 1984).
Conductance and conductance density rank scores for the sustained and inactivated Kv currents were correlated to the visual score. The Spearman's rank correlations were statistically significant for the sustained conductance (ρ = −0.62, P = 0.013; Fig. 5A) and its density (ρ = −0.76, P < 0.001; Fig. 5B) but not for the inactivating conductance and its density (ρ = −0.31, P = 0.26 and ρ = −0.35, P = 0.19, respectively; correlations not shown). However, due to the nonquantitative origin of the score data sets, these correlations should be treated with caution.
In this work we compared Kv currents from 15 insect species belonging to 7 different insect orders and characterized by diverse morphology, lifestyles, and behaviors. Laughlin and Weckström (1993), using the single-electrode intracellular recording technique, found that day-active, fast-flying flies strongly express sustained Kv currents, whereas nocturnal, crepuscular, and relatively slow-flying flies display large inactivating currents (Laughlin and Weckström 1993). That study was limited to dipterans, thus likely reducing biophysical variability potentially associated with more phylogenetically diverse samples. Later, supporting findings were reported in the locust, with clear circadian changes in the character of Kv currents (Cuttle et al. 1995). Patch-clamp recordings from D. melanogaster, which is a crepuscular insect (Bahn et al. 2009), showing a Kv current with a prominent IA component, have also corroborated the above-stated conclusions.
However, further patch-clamp studies have challenged this paradigm. Specifically, the crepuscular/nocturnal cockroach P. americana demonstrated a mainly noninactivated Kv current (only a minority of photoreceptors seem to express Kv current with a pronounced IA component) (Salmela et al. 2012), whereas recordings from Drosophila virilis, which is a diurnal insect (Bahn et al. 2009), surprisingly revealed a Kv current very similar in kinetics to that in D. melanogaster photoreceptors, albeit of higher magnitude (unpublished observations). The present study has now explored the topic further, expanding it to species with very dissimilar evolutionary histories and across seven insect orders. Our results indicate that specific membrane conductances as approximated by conductance densities (both the sustained and inactivated ones) tend to be larger in species with smaller photoreceptors than in species with larger photoreceptors, and that the density of sustained Kv conductance in the physiological voltage range is a better predictor of the visual prowess of the animal than the size of sustained Kv conductance per se as previously thought. Properties of IA are not usable for such a prediction.
Physiological voltage range and visual ecology.
Physiological voltage range is the range of membrane potentials used by the photoreceptor to process normal stimuli. Although the photoreceptor membrane may momentarily get hyperpolarized under physiological conditions, it is usually more useful for comparative studies to consider that the lower boundary of PVR is set by the resting potential. However, determining the upper boundary of PVR is less straightforward. In continuous light, voltage responses usually take form of a relatively sustained depolarization (“plateau”), but if the continuous light is preceded by a considerably dimmer background light or darkness, a rapid depolarizing transient will occur before the plateau, lasting from tens to hundreds of milliseconds. If the contrast between light intensities is even higher, the light-induced transient approaches values close to the reversal potential for light-activated channels (approximately from 0 to 10 mV), and this sets the absolute upper limit of the PVR. However, although the transient may play a role in signaling and adaptation, information transmission-wise the functional upper boundary of the PVR depends on the potential of sustained depolarization during photoreceptor responses to the brightest stimuli ordinarily encountered by the animal. This in turn is determined by the interplay between the self-shunting depolarizing light-activated and repolarizing Kv conductances. Moreover, under naturalistic conditions (Heimonen et al. 2012; Song et al. 2012; van Hateren 1997) the photoreceptors are more or less light-adapted most of the time, making very large transients rare. In addition, when nocturnal or crepuscular species are concerned, photoreceptors often have to operate with minimal, sometime even single-photon, signals. Taken together, when analyzing the functional significance of Kv currents, it is clearly worthwhile to consider very carefully what is the most relevant PVR.
What is the variation in the functional upper boundary among species with different visual behavior and ecologies? Conceivably, physiological variation in the upper PVR boundary arises from differences in the composition and properties of ion channels constituting the species-specific photoreceptor channelome. Moreover, the upper boundary seems to depend on the experimental approach, because voltage-light intensity relations (V-log I curves) obtained using intracellular recordings (i.e., in vivo experiments on the intact eye) differ from those procured with the patch-clamp method (in vitro recording from dissociated ommatidia), with generally lower plateau depolarization potentials observed in intracellular experiments (Immonen et al. 2014a), possibly due to increased leak introduced by the impalement. The accumulated evidence indicates that photoreceptors usually do not reach plateau levels of much beyond −30 mV on average when stimulated with prolonged, tens of seconds-long, steady or white noise-modulated pulses of progressively brighter light. This can be seen in intracellular recordings from D. melanogaster (Juusola and Hardie 2001), horseshoe crab (Limulus; Millecchia and Mauro 1969), and C. vicina (Juusola et al. 1994). This is also true for P. americana, according to results of intracellular and patch-clamp experiments, despite large intrinsic variability among cockroach photoreceptors (Heimonen et al. 2006, 2012). Similarly, average plateau depolarization has not exceeded the −30 mV mark in patch-clamp experiments in G. bimaculatus (Frolov et al. 2014), G. lacustris (Frolov and Weckström 2014), N. glauca (Immonen et al. 2014a), and C. punctata (Frolov 2015). As one exception, plateau depolarization in C. morosus surpassed −20 mV at two brightest light levels corresponding to bright daylight (Frolov et al. 2012). However, it can be argued that because the stick insect is a strictly nocturnal animal and photoreceptors in these experiments were exposed to light bypassing the ommatidial pigment screening machinery, the brightest stimuli are nonphysiological. Also, it is important to note that natural contrasts are not as saturating as steady or white-noise stimuli (Song et al. 2012). Therefore, −20 mV appears to be a reasonable conservative estimate of the general upper boundary of the PVR in insect photoreceptors.
Ranking of visual ecological and behavioral traits: the rationale.
To quantify and rank visual ecological and behavioral traits of a species in the context of biophysical properties of photoreceptors, it was first necessary to determine what is required from the visual system to maintain a particular lifestyle (diurnal, nocturnal, crepuscular, etc.) and to mediate a certain visually guided behavior (flight, maneuvering, escape reactions, etc.). The following provides the background for the visual scoring system used in this study.
The lifestyle in the context of illumination range of normal animal activity is associated with profound morphological and physiological adaptations in the visual system, which are often mutually incompatible for visual systems operating under dissimilar conditions (for review, see Cronin et al. 2014). In other words, visual systems cannot be evolutionary optimized to excel in solving all visual problems of survival and propagation posed by the environment, so trade-offs are inevitable. For instance, diurnal fliers may need high visual acuity and the ability to process rapid and large contrasts with high temporal resolution. This is generally achieved by evolving apposition-type eyes with small interommatidial and acceptance angles, conveying superior spatial resolution. Diurnal insects do not need high absolute sensitivity; in fact, they use a number of means to regulate photon flux and prevent photoreceptor saturation in bright light, including pigment screening and rhabdomere withdrawal mechanisms (Cronin et al. 2014; Immonen et al. 2014a). At the level of photoreceptors, diurnal species demonstrate low transduction gain (response amplitude per unit of stimulus intensity), high contrast gain (response amplitude per unit of stimulus contrast), short bump latency, and high corner frequency (Faivre and Juusola 2008; Juusola and Hardie 2001; Juusola et al. 1994; Niven and Laughlin 2008; Niven et al. 2003).
On the other hand, nocturnal insects need to capture and use information carried by the scarce photons that impinge on the eye in a dim environment. This is accomplished by adaptations both at the visual system periphery by optical (e.g., by superposition-type eyes) and/or physiological means (temporal integration in photoreceptors; Warrant and Dacke 2011) and at the level of optic lobes (e.g., high degree of spatial summation in the lamina; Greiner 2006; Stöckl et al. 2015). Photoreceptors of nocturnal insects usually demonstrate high gain of phototransduction, with a concomitant increase in bump latency and decrease in corner frequency (Berry et al. 2011; Fain et al. 2010; Laughlin 1981). Thus, due to evolutionary design principles, visual systems of crepuscular/nocturnal insects generally facilitate neither high temporal resolution nor visual acuity.
Although visually guided behaviors can be detected more or less in all species with eyes, the performance-related demands that different behaviors impose on the visual system can vary dramatically. For instance, 11 of 15 insects evaluated in this work possess wings and can fly (although most of them fly on very rare occasions), but only the flies, honeybee, and butterfly are known to exercise rapid visually guided maneuvers and complex aerobatics during flight. Species such as N. glauca, G. lacustris, and C. punctata fly rather awkwardly, relying for orientation on perception of sources of polarized light (Horváth and Varjú 2004), which is mediated by distinct photoreceptors characterized (in the case of the cricket, G. bimaculatus) by relatively poor contrast and temporal resolution (Frolov et al. 2014). However, the aquatic species display robust visually triggered escape reactions in their normal water milieu. As another example, visually guided behavior in crickets seems to be limited to locomotion using polarized light cues and to contrast resolution at close range (Böhm et al. 1991).
Limitations of the method.
Although this is the most encompassing comparative electrophysiological study to date of photoreceptors with regard to diversity of insect orders involved, it was limited in several technical aspects. First, only insects with apposition-type compound eyes were included, because our attempts to dissociate ommatidia and patch photoreceptors from species with superposition eyes were unsuccessful. Ommatidia in superposition eyes tend to be much smaller than in apposition eyes, due to the different optics. This may lead to their comparative biophysical properties being inconsistent with the correlations reported presently. Second, a leak conductance, which is prominently displayed in photoreceptors of the rapid diurnal flyers, was not taken into account in the present methods, because it cannot be separated from the instrumental seal leak. Although a possible physiological role of the leak channels was thus disregarded, and total sustained conductance underestimated, the error is conservative and cannot strongly affect the conclusions because the three fast-flying species, the honeybee, the blowfly, and the butterfly, already have the highest conductance and visual scores. Third, we evaluated Kv current inactivation at 400 ms after the onset of the voltage pulses. However, it is clear from Fig. 2 that in some species inactivation of the total current is not fully settled by 400 ms, which implies that conductance/conductance density scores would be slightly different if longer pulses were used. This would not have radically changed the results because the final outcome was based on ranking and not on absolute or relative numerical values. The main methodological limitation here was the necessarily qualitative nature of the visual ranking. Many parameters conventionally related to visual performance, including the interommatidial angle, absolute and relative eye sizes, and facet number, had to be omitted due to ambiguity of their relevance or difficulties of incorporating them into the visual score, even if they were known for some of the species.
Enigma of the inactivating conductances.
Expression of the rapidly inactivating IA is clearly insufficient for proper photoreceptor signaling in fast-flying insects, which instead express a rapidly activating sustained Kv conductance, as demonstrated in Diptera (Laughlin and Weckström 1993). This helps to increase the membrane bandwidth, and thereby the maximum speed of signal conduction, but also raises metabolic costs of signaling, serving as a deterrent against evolutionarily unwarranted improvements in performance (Niven et al. 2007). However, it is not clear why the pattern of IA variability as it has been observed in flies, with strong expression in nocturnal/crepuscular animals, was absent in the present broader sample. Furthermore, the role of IA in insect photoreceptors remains poorly understood, despite the advances made in understanding the biophysical, physiological, and molecular mechanisms in fruit fly photoreceptors. In D. melanogaster, IA is thought to selectively amplify weak voltage signals in dim light and reduce the metabolic costs of signaling (Niven et al. 2003).
However, IA in the fruit fly is characterized by a much more negative half-activation potential than that of IDR, thus effectively increasing the operating range of the total Kv conductance: IA conductance is activated at the level of the resting potential, when IDR channels are not conducting, and its inactivation, a voltage-dependent process, is relatively slow. This is in contrast to IA properties in other, non-dipteran species, where, as a rule, these two currents activate roughly within the same voltage range, with very similar half-activation potentials (Frolov et al. 2012, 2014; Frolov and Weckström 2014; Frolov 2015; Immonen et al. 2014a; Salmela et al. 2012). Experiments involving pharmacological inhibition and mathematical modeling have suggested that IA in the cockroach does not participate in modulation of sustained voltage responses, at least at the level of the photoreceptor soma (Salmela et al. 2012). However, the apparent lack of involvement of non-dipteran IA in contrast-response modulation does not exclude a possibility of its participation in other important processes, such as limiting durations of large voltage-response transients (that can last up to several hundred milliseconds, effectively rendering the animal blind during that period) in response to large positive contrasts.
Photoreceptor size and Kv currents.
Previously, substantial positive correlations between photoreceptor capacitance and delayed rectifier conductance were found in several species used in this study (Frolov et al. 2012; Frolov and Weckström 2014; Frolov 2015; Salmela et al. 2012). In contrast, either no correlation or a relatively small one has been found in the above-cited cases between the photoreceptor capacitances and IA amplitudes. These findings conform to the notion that variation in photoreceptor membrane area, which can be conveniently estimated by the cellular capacitance because this value per area is fairly constant, about 1 μF/μm2, is mainly due to variation in the size of the light-sensitive rhabdom, which dominates the membrane area because of its highly convoluted topology and less because of changes in the area of light-insensitive membrane. In Drosophila, the IDR channels, found presently to correlate with cellular size, are thought to be expressed in the close proximity of the rhabdom (Hardie and Raghu 2001; Krause et al. 2008), whereas IA channels, with little correlation with cell size, are very likely to be concentrated in the non-microvillar membrane (Hardie 1991; Rogero et al. 1997). However, when correlations between capacitance and Kv conductance densities (specific membrane conductances) were examined in the stick insect and the water strider, a more controversial picture emerged: whereas IA correlated strongly negatively with cellular capacitance, the IDR conductance density correlated positively with capacitance in the water strider but not in the stick insect. One possible explanation of these differences can be discerned on the basis of the capacitance-current correlations for the depolarizing LIC: whereas the LIC positively correlated with capacitance in both species, a positive capacitance to LIC density correlation trend was found only in the water strider, suggesting that the increase in the Kv conductance density with photoreceptor size might be necessary to counter the increase in the LIC conductance.
Results presented in Figs. 3, D and E, and 4, D and E, for interspecies variability in Kv conductance properties show that 1) there was no substantial correlation between capacitance and conductances when average values over the PVR were used, and 2) correlations between capacitance and conductance densities were strongly negative. These patterns of interspecies variability in Kv conductance properties appear to be quite consistent with the intraspecies findings, although it is still not understood why certain trends, especially the general decrease in specific conductances with increasing photoreceptor size, are such as they are. If specific conductance remains constant, then an increase in cell capacitance entailing an increase in membrane time constant is offset by the proportional decrease in input resistance. However, our results inexplicably paint a different picture and suggest that there is no significant evolutionary pressure to compensate for slowing of membrane responses in larger cells with a proportionally increased conductance.
This work was supported by grants from the Academy of Finland (to R. Frolov and M. Weckström) and the Finnish Cultural Foundation and the Finnish Graduate School of Neuroscience (to E.-V. Immonen).
No conflicts of interest, financial or otherwise, are declared by the authors.
R.F. and E.-V.I. conception and design of research; R.F. and E.-V.I. performed experiments; R.F. and E.-V.I. analyzed data; R.F. and E.-V.I. interpreted results of experiments; R.F. prepared figures; R.F. drafted manuscript; R.F., E.-V.I., and M.W. edited and revised manuscript; R.F., E.-V.I., and M.W. approved final version of manuscript.
Present address of E.-V. Immonen: Vision Group, Department of Biology, Lund University, Sölvegatan 35, 22362 Lund, Sweden.
- Copyright © 2016 the American Physiological Society