Phenotypic plasticity in Periplaneta americana photoreceptors

Abstract

Plasticity is a crucial aspect of neuronal physiology essential for proper development and continuous functional optimization of neurons and neural circuits. Despite extensive studies of different visual systems, little is known about plasticity in mature microvillar photoreceptors. Here we investigate changes in electrophysiological properties and gene expression in photoreceptors of the adult cockroach, Periplaneta americana, after exposure to constant light (CL) or constant dark (CD) for several months. After CL, we observed a decrease in mean whole-cell capacitance, a proxy for cell membrane area, from 362 ± 160 to 157 ± 58 pF, and a decrease in absolute sensitivity. However, after CD, we observed an increase in capacitance to 561 ± 155 pF and an increase in absolute sensitivity. Small changes in the expression of light-sensitive channels and signaling molecules were detected in CD retinas, together with a substantial increase in the expression of the primary green-sensitive opsin (GO1). Accordingly, light-induced currents became larger in CD photoreceptors. Even though normal levels of GO1 expression were retained in CL photoreceptors, light-induced currents became much smaller, suggesting that factors other than opsin are involved. Latency of phototransduction also decreased significantly in CL photoreceptors. Sustained voltage-activated K⁺ conductance was not significantly different between the experimental groups. The reduced capacitance of CL photoreceptors expanded their bandwidth, increasing the light-driven voltage signal at high frequencies. However, voltage noise was also amplified, probably because of unaltered expression of TRPL channels. Consequently, information transfer rates were lower in CL than in control or CD photoreceptors. These changes in whole-cell capacitance and electrophysiological parameters suggest that structural modifications can occur in the photoreceptors to adapt their function to altered environmental conditions. The opposing patterns of modifications in CL and CD photoreceptors differ profoundly from previous findings in Drosophila melanogaster photoreceptors.

Figure 1.

Changes in photoreceptor capacitance and absolute sensitivity. (A) Mean capacitance values of photoreceptors in CL, control, and CD; here and elsewhere, n indicates the number of cells, and error bars denote SD. Mean C_m was 157 ± 58 pF in CL (n = 21; P < 10⁻¹², unpaired t test, comparison with control), 374 ± 180 pF in control (n = 83), and 560 ± 149 pF in CD photoreceptors (n = 26; P < 10⁻⁵, unpaired t test, comparison with control). **, P < 0.01. (B) Distributions of C_m values. (C) Correlations between C_m and absolute sensitivity; sensitivity values were determined by counting bump rates in response to continuous stimulation at light intensities evoking <10 bumps/s; the rates were recalculated for the common light intensity 5 × 10⁻⁶ as in Fig. 4 B. The median values were 1.2 × 10⁻³ (0.3 × 10⁻³:8.0 × 10⁻³) in CL (n = 15), 0.75 (0.19:1.85) in control (n = 59), and 5.3 (3.5:8.0) bumps/s in CD (n = 9) photoreceptors. The differences were highly significant, with P < 10⁻⁴ (MWUT) for comparisons of both CL and CD groups to control.

Figure 2.

Photoreceptor latency. (A–C) Typical responses of CL, control, and CD photoreceptors to 1-ms flashes of light; light intensity was adjusted to evoke quantum bumps with a probability of <70%; stimulus was given at 0 ms; dashed lines indicate bump latency medians for these cells; bump latency was determined as the interval between the onset of light and the time that the quantum bump amplitude reached 10% of its maximum value. (D) Normalized distributions of bump latencies; to obtain the distributions, 50 latency values from each cell were combined into a common pool, and frequencies were normalized. (E) Mean latency values were obtained by averaging mean latencies from each photoreceptor. Mean bump latency was significantly smaller in CL than in control and CD photoreceptors: the values were 54.4 ± 7.8 ms in CL (n = 12), 63.2 ± 13.7 ms in control (n = 34; P = 0.014, unpaired t test, comparison to CL), and 66.1 ± 9.2 ms in CD photoreceptors (n = 8; P = 0.007, unpaired t test, comparison to CL). *, P < 0.05; **, P < 0.01; error bars indicate SD.

Figure 3.

Elementary current and voltage responses. (A–C) Typical quantum bump responses to low-intensity continuous light stimulation in each experimental group; voltage and current recordings were obtained from the same photoreceptors using the same stimulation protocol; as shown, stimulus intensities were the same for each recording pair but different between the groups. (D) Average voltage bumps. Because of strong dependence of the voltage bump amplitude on resting membrane potential, the voltage bumps were obtained in the following way: first, mean voltage bumps were obtained for each photoreceptor by aligning the rising parts of the bumps; second, a subgroup of cells was selected so that the average resting potentials were approximately the same for all experimental groups. In the subgroups, the mean resting potentials were −56.8 ± 5.3 mV (n = 9) for CL, −56.6 ± 5.6 mV (n = 13) for control, and −56.6 ± 5.5 mV (n = 7) for CD photoreceptors. The corresponding mean C_m values were 156 ± 62, 345 ± 164, and 469 ± 169 pF. (E) Normalized voltage bumps from D.

Figure 4.

Macroscopic light-induced currents. (A) Typical light-induced currents recorded from CL, control, and CD photoreceptors; 4-s light stimuli were used at six (in control and CD photoreceptors) or seven (in CL photoreceptors) intensities in 10-fold increments; stimulus duration is shown as a horizontal bar. (B) Dependence of sustained LIC on light intensity; the values were obtained as averages of the final 3 s of current responses; although not all data samples passed the normality test, the data here for presentation purposes are shown as mean ± SD; error bars are shown in the negative direction. At all intensities, LIC in CL photoreceptors was significantly smaller than LIC in control and CD photoreceptors. For example, at light intensity 5 × 10⁻¹, sustained LIC amplitudes were −165 (−429:−46) pA in CL (n = 8), −555 (−787:−348) pA in control (n = 34; P = 0.0014, MWUT, comparison with CL), and −887 (−1,339:−540) pA in CD (n = 8; P = 0.014, MWUT, comparison with CL). LIC values in control and CD photoreceptors were not significantly different from each other at intensities 5 × 10⁻¹ and 5 × 10⁻² (P = 0.11 and 0.09, respectively, MWUT). However, in relatively dim light, LIC in CD photoreceptors exceeded that in control: at intensity 5 × 10⁻³, the sustained LIC amplitudes were −260 (−458:−111) pA in control (n = 34) versus −517 (−795:−395) pA in CD (n = 8; P = 0.019, MWUT); at 5 × 10⁻⁴, the sustained LIC amplitudes were −103 (−285:−25) pA in control (n = 33) versus −373 (−545:−135) pA in CD (n = 8; P = 0.015, MWUT). *, P < 0.05. (C) Correlations between the amplitudes of sustained LIC and C_m values at intensity 5 × 10⁻¹.

Figure 5.

Potassium currents. (A) A typical Kv current recorded from a photoreceptor after CL exposure; the recording protocol is shown to the right; each testing step was preceded by a 1-s prepulse to −102 mV to fully recover the transient IA; the first 3 ms of the current traces containing capacitive transients were removed. (B) Current–voltage relationships for sustained Kv conductance in CL, control, and CD photoreceptors; values are averages of final 200 ms for each trace; data points were fitted with a sigmoidal function; standard deviation bars are shown in different directions for presentation purposes. G_max was slightly smaller in CL than in control and CD photoreceptors: 29.1 ± 14.9 nS (n = 16), 37.3 ± 13.0 nS (n = 40; P = 0.066, t test, comparison with CL), and 40.5 ± 17.8 nS (n = 21; P = 0.041, t test, comparison with CL), respectively.

Figure 6.

Responses to Gaussian white noise–modulated light stimuli. (A) First 20 s of representative voltage responses to a 60-s GWN stimulus at light intensities that elicited IR_max responses in photoreceptors kept in CL, control, and CD; the stimulus is shown above. (B) Mean sustained membrane depolarizations during responses to GWN at different light intensities; values were obtained by averaging the entire duration of the response except for the first second and then subtracting the resting potential; error bars in B, C, and E denote SD. (C) Dependencies of mean IR on light background; in each photoreceptor, IR values associated with saturated responses in relatively bright light, which were smaller than IR_max, were excluded; the number of data points varied from 2 (for CL in two dimmest levels) to 21. At all light levels, control and CD photoreceptors transferred significantly more information than CL photoreceptors (P < 10⁻³ for all comparisons, values not shown). *, P < 0.05. (D) Distributions of IR_max values depending on light level. (E) Average maximal information rates. The IR_max values were 7.3 ± 4.3 bits/s in CL (n = 9), 12.9 ± 5.6 bits/s in control (n = 22; P = 0.008 for comparison with CL, unpaired t test), and 13.7 ± 7.3 bits/s in CD photoreceptors (n = 10; P = 0.039 for comparison with CL, unpaired t test). *, P < 0.05.

Figure 7.

Information processing for responses associated with IR_max. (A) Median signal gain functions obtained from IR_max responses to GWN in CL, control, and CD photoreceptors; in A and E, error bars represent m.a.d. and are shown in different directions for presentation purposes. Dependencies of IR_max responses on light level are shown in Fig. 6 D. (B) Mean membrane corner frequencies obtained by fitting signal gain functions of IR_max responses with a first-order Lorentzian equation. The values of f_3dB were 7.6 ± 2.0 Hz in CL (n = 9), 3.8 ± 0.9 Hz in control (n = 22; P < 10⁻⁴ for comparison with CL, unpaired t test), and 3.9 ± 1.0 Hz in CD photoreceptors (n = 10; P < 0.001 for comparison with CL, unpaired t test); error bars denote SD. The legend under B also refers to D. **, P < 0.01. (C) Median signal and noise power spectra; the legend is under D; error bars are omitted for presentation purposes. (D) Box plots compare total signal and noise power spectra integrals in the 1–50-Hz range. The total noise power was significantly higher in CL than in control and CD photoreceptors: in the range from 1 to 50 Hz, it was 0.87 (0.23:1.89) mV² in CL (n = 9; P = 0.008 and 0.001, MWUT, for comparison with control and CD, respectively), 0.13 (0.09:0.18) mV² in control (n = 22), and 0.12 (0.10:0.16) mV² in CD photoreceptors (n = 10). **, P < 0.01. (E) Median SNR functions.

Figure 8.

qPCR results for gene expression. Relative expression of GO1, UVO, GO2, Gq, Arr, PLC, TRP, and TRPL mRNA from CL, control, and CD retinas; data were obtained by qPCR and normalized to reference genes for actin and GAPDH. Values represent means ± SEM of three technical replicates of testing the same experimental samples as described in Materials and methods.