Single-Photon Sensitivity of Lamprey Rods with Cone-like Outer Segments

Most vertebrates have a duplex retina containing rods for dim light vision and cones for bright lights and color detection. Photoreceptors like cones are present in many invertebrate phyla as well as in chordata, and rods evolved from cones [1, 2]; but the sequence of events is not well understood. Since duplex retinas are present in both agnatha and gnathostomata, which diverged more than 400 million years ago, some properties of ancestral rods may be inferred from a comparison of cells in these two groups. Lamprey have two kinds of photoreceptors, called “short” and “long” [3, 4, 5-9], which seem to be rods and cones; but the outer segments of both have an identical cone-like morphology of stacks of lamellae without a continuous surrounding plasma membrane [3, 4, 6, 7]. This observation and other aspects of the cellular and molecular biology of the photoreceptors have convinced several investigators [2, 10-12] that “the features of ‘true’ rod transduction in jawed vertebrates, which permit the reliable detection of single photons of light, evolved after the separation of gnathostomes from lampreys [12].” To test this hypothesis, we recorded from photoreceptors of the sea lamprey Petromyzon marinus and show that their rods have a single-photon sensitivity similar to that of rods in other vertebrates. Thus photoreceptors with many of the features of rods emerged before the split between agnatha and gnathostomata, and a rod-like outer segment with cytosolic disks surrounded by a plasma membrane is not necessary for high-sensitivity visual detection.

Results

Recordings were made at room temperature (22°C) from eyes dark-adapted from 3-5 hours and dissected under infrared illumination. The retina was removed and cut into slices, and single outer segments were pulled up into suction electrodes to record photocurrents as for other species [for example 13]. In some experiments, the retinas were exposed for 3 min to 0.5 mg/ml collagenase and 0.33 mg/ml hyaluronidase to prevent clogging of pipettes by vitreous and extracellular matrix. To record from long photoreceptors (cones), we selected cells extending further from the retinal slice. The responses of short photoreceptors (rods) were recorded by moving the electrode between cones or by searching for parts of the slice where the cones had been displaced during preparation.

In Figure 1 we show mean current responses from the two kinds of photoreceptors to brief flashes of light. The short photoreceptors (A) had responses resembling amphibian rods [for example 14]. The spectral sensitivity of the response (not shown) was estimated by recording small-amplitude responses to stimuli at selected wavelengths and dividing response amplitude by flash intensity. The sensitivity measurements were then fitted to template curves [15] to estimate the peak of pigment absorbance, which for the short photoreceptors was at about 520 nm [6, 16, 17]. We therefore identify these cells as lamprey rods. The long photoreceptors (B) had responses rising and decaying much more rapidly (note difference in time scale), typical of amphibian [18] and mouse cones [19], with a spectral sensitivity peaking at about 570 nm [6, 16]. We identify these cells as lamprey cones. In our limited sample, all of the cones had the same spectral sensitivity indicating a single spectral class of cone in this species of lamprey, as previously reported [16].

Figure 1.

Current responses of lamprey rod and cone photoreceptors to brief light stimuli. For both photoreceptor types, response amplitude and duration increase with increasing stimulus intensity. (A) Mean responses of 11 rods to 20 ms 500 nm flashes given at t=0 at the following intensities (in photons μm^-2): 5, 24, 60, 222, 642 and 1576. (B) Mean responses of 8 cones to 20 ms 600 nm flashes given at t=0 at the following intensities (in photons μm^-2): 735, 2120, 5210, 1. 98 × 10⁴ , 7. 70 × 10⁴, 2.28 × 10⁵, 6.96 × 10⁵and 2.03 × 10⁶.

In Figure 2, we have plotted the mean response amplitude (with SEM) as a function of flash intensity for the cells of Figure 1. The rods at 500 nm are of the order of 1.8 log units or 65 - 70 times more sensitive than the cones at 600 nm. We have fitted the responses of each photoreceptor type to exponential saturation equations of the form r = r_max [1 − exp(-kI)] [as in 20], where r is response amplitude, r_max is the maximum value of r, I is the flash intensity, and k is a constant. The best-fitting values of r_max and k in Figure 2 were 10.1 pA and 1.5 × 10^-2 photons^-1 μm² for rods and 10.4 pA and 2×10⁻⁴ photons^-1 μm² for cones. When values of r_max and k were estimated by fitting the response-intensity curves cell by cell, we obtained 10.2 ± 0.8 pA and 1.3 ± 0.2 × 10^-2 photons^-1 μm² for rods (SE, n = 11) and 11.9 ± 2.4 pA and 1.6 ± 0.3 × 10^-4 photons^-1 μm² for cones (n = 8). These measurements were made at wavelengths of stimulation that were somewhat different from our estimates of the wavelengths of maximal sensitivity of the two kinds of photoreceptors. For this reason, we adjusted the sensitivities by a factor of 1.1 for rods and 1.24 for cones based on template curves for the photopigments [15] in order to estimate sensitivities at the λ_max of the pigments. This gave mean values of k of 1.4 × 10^-2 photons^-1 μm² for rods and 2.0 × 10^-4 photons^-1 μm² for cones, giving mean values of the intensity required to give a half-maximal response (I_½) of 50 photons μm^-2 for rods and 3500 photons μm^-2 for cones. The ratio of the values of k at the λ_max of the pigments was therefore approximately 70. We conclude that rods are of the order of 70 times more sensitive than cones, within the range of values recorded for other vertebrate species [for example 19, 21].

Figure 2.

Sensitivity of rod and cone photoreceptors in lamprey. Current response amplitudes were plotted against their corresponding flash intensities for 11 rods (closed squares) and 8 cones (open squares). Same cells as in Figure 1. The data for both cell types were fitted [as in 20] with the equation r = r_max [1 − exp(-kI)]. The best-fitting values of r_max and k were 10.1 pA and 1.52×10⁻² photons⁻¹ μm² for rods and 10.4pA and 2×10⁻⁴ photons^-1 μm² for cones.

In Figure 3, we compare the normalized waveforms of the responses of rods and cones to flashes that, for each photoreceptor type, produced a response of about half-maximal amplitude. Because the cones were less sensitive than the rods, the light intensity required to produce the cone response was of the order of 90 times brighter than the one used to stimulate the rods. The mean cone response had a much more rapid rate of activation and time to peak, at least in part the result of the brighter stimulus intensity. The decay of the cone response in Figure 3 was fitted to a single exponential decay function with a best fitting time constant of decay τ_REC of 71 ms (red curve). Values of τ_REC were also obtained by fitting the decay phases of small-amplitude responses cone by cone and gave a mean value of 54 ± 4 ms (SEM, n=8). A similar fit to the rod response in Figure 3 gave a value for τ_REC of 793 ms (red curve), and cell-by-cell fits to small-amplitude responses gave a mean value of τ_REC of 841 ± 59 ms (n=12).

Figure 3.

Comparison of response waveforms of rod and cone photoreceptors in lamprey. Normalized mean light response of 11 rods and 8 cones to 20 ms flashes that produced an approximately half-maximal response. Flash intensities were 60 photons μm^-2 for rods and 5210 photons μm^-2 for cones. The decay phases of both responses were fitted with a single-exponential decay function with a time constant of recovery (τ_REC) for rods of 793 ms and for cones of 71 ms. Same cells as in Figures 1 and 2.

To see if lamprey rods respond to single photons of light, we gave a series of dim flashes and calculated the mean single-photon response from the squared mean and variance as previously described [22]. The waveform of the response was divided cell by cell by the maximum current response to a bright flash, in order to give the fraction of channels closed as a function of time. These results are shown for lamprey rods in the upper half of Figure 4. In comparison, we give in the lower half of Figure 4 similar data for mouse rods [13]. In both cell types, 4-5% of the channels of the outer segment are closed by a single photon. Although the peak response in lamprey is somewhat smaller than in mouse, it is larger than in salamander [for example 23] and is well within the range of single-photon response amplitudes of other vertebrate rods. Lamprey single-photon responses rise to a peak and decay more slowly than mouse responses, but the slower kinetics are probably mostly the result of the difference in the temperature of the recording (22°C versus 39°C).

Figure 4.

Single-photon responses of lamprey rods (upper) and mouse rods (lower). Responses were calculated from the squared mean and variance [as in 22] for 10 lamprey rods and 41 mouse rods, normalized rod by rod to circulating current and averaged to give the mean fractional closure of channels as a function of time.

Discussion

Our results show that the lamprey retina has two physiological classes of photoreceptors closely resembling the rods and cones of other vertebrates. The “short” photoreceptors respond to brief flashes of light with high sensitivity, and their responses decay slowly as for other vertebrate rods; whereas the “long” photoreceptors have a sensitivity and time course of response nearly indistinguishable from the cones of other vertebrates [see 24]. Our experiments confirm previous evidence of two distinct lamprey photoreceptor classes from differences in the expression of membrane proteins [8] and different classes of transducin [25] and phosphodiesterase gamma [26]. Moreover extracellular recordings from lamprey retina previously indicated that there are two kinds of photoreceptors with differing sensitivities [6].

Our results suggest that, before the divergence of agnatha from gnathostomata, an ancestral rod had essentially all of the transduction machinery necessary to produce high-sensitivity photon detection. The evolution of rod transduction must therefore have occurred considerably earlier. Since cone pigments are older than rod pigments [27], and there is no evidence for high-sensitivity ciliary photoreceptors before the agnathans, we presume that the ciliary photoreceptors in primitive chordates were initially cone-like as in other invertebrates. The genome of primitive chordates must then have undergone gene duplication of the principal transduction proteins including visual pigments, transducin, phosphodiesterase, and the cyclic nucleotide-gated channels, so that in addition to cone-like ciliary photoreceptors the primitive chordates developed rod photoreceptors with most of the properties of the rods of vertebrates. This transition is likely to have occurred rather early and sometime between the emergence of the chordates during the Cambrian and the split of agnathans with gnathostomata over 400 million years ago.

Although ancestral rods before the agnatha/gnathostomata split seem to have been able to respond to light with high sensitivity, they appear to have lacked an outer segment composed of cytosolic disks surrounded by a continuous plasma membrane. Several morphological investigations [3, 4, 6, 7] have shown that lamprey rod outer segments have lamellae resembling disks, but these lamellae are not uniformly cytosolic and enclosed within a surrounding plasma membrane like the disks of other vertebrate rods. They are instead interrupted by numerous infoldings of the plasma membrane itself, much as in a cone [for example 28, 29]. Moreover, lamprey rods exposed to tritiated amino acids show random labeling over the whole of the outer segment like cones [3] rather than the banded labeling characteristic of amphibian and mammalian rods [29, 30]. This finding is further evidence that the membranes of the lamellae of lamprey rods are in direct contact and continuous with the plasma membrane. In our experiments we have observed that when we pull a lamprey rod outer segment up into our suction electrode, on occasion the outer segment does not behave as a single morphological unit but seems to fall apart into connected segments, which never happens when we record from amphibian or mouse rod outer segments. These rods nevertheless give vigorous responses to light. All of these observations taken together indicate that the structure of the lamprey rod is different from that of other vertebrate rods and more closely resembles the structure of vertebrate cones.

It is therefore likely that the evolution of a true rod morphology occurred after the separation of the agnatha from the rest of the vertebrates. We are accustomed to thinking that the peculiar anatomy of the rod outer segment with its placement of photopigment on cytosolic disks surrounded by a plasma membrane is essential for the single-photon sensitivity of rod vision, although this supposition is based on no evidence apart from the clear differences between rods and cones in sensitivity and morphology among the gnathostomata. Our recordings from lamprey rods show to the contrary that cytosolic disks are not necessary for high-sensitivity visual detection. Because the placement of the pigment on disks appears to make only a limited contribution (if any) to the sensitivity of the rod, we suggest that disks may have evolved for some other function perhaps related to the transport of pigment or renewal of the outer segment. Rhodopsin in the rods of amphibians and mammals is synthesized in packets of disk membrane added diurnally at the base of the outer segment, which are then shed at the outer segment tip and phagocytized by the retinal pigment epithelium [29, 30]. The organized generation and renewal of outer segment membrane has the advantage of keeping together pigment that has been synthesized during a fixed time period, so that only the oldest pigment is removed during disk shedding. This feature may have been sufficiently important to the cellular biology of the photoreceptor to have driven the evolution of a true rod morphology.