Photoreceptors, carburetors, and intelligent design

OCT 07, 2010

A paper in today’s Nature caught my eye. E. J. Chichilnisky of the Salk Institute for Biological Studies in San Diego and his collaborators set out to determine how the cells in primate retinas are wired to sense color.

DOI: 10.1063/PT.5.010035

Charles Day

Now you might think, as I did before reading the paper, that the problem had already been solved. Our retinas contain three types of photoreceptive neuron that are maximally sensitive to red, green, or blue light. When, say, the blue photoreceptors fire, we see blue, right? Wrong.

It turns out that neuroscientists have known for some time that our brains don’t receive direct signals from the R, G, and B photoreceptors. That’s in part because the photoreceptors’ spectral responses overlap. A dim red light would elicit the same level of response in an R photoreceptor as a bright red light would elicit in a G photoreceptor.

Our eyes rely instead on “opponency.” The signals that run from our retinas to our brains correspond to two opposing combinations: B − (R + G) and R − G. The combinations are calculated by specialized neurons called ganglions that receive input from mixed groups of photoreceptors.

Chichilnisky and his team connected hundreds of ganglion cells in vitro to electrodes. They then recorded the cells’ response to a spectrally varying light field whose spatial resolution was fine enough to trigger responses in individual photoreceptors.

The figure shows schematically how the photoreceptors (colored dots) are grouped to feed data to individual ganglions (at the focal points of the white lines). In principle, the grouping of randomly distributed photoreceptors could account for the retina’s sensitivity to color. But Chichilnisky found that an extra ingredient is needed: The photoreceptors closest to the focal points are weighted more heavily than those farther out.

Sensing color with randomly distributed and simply connected R, G, and B photoreceptors seems elegant, but if you look under the hood at the proteins responsible, you see a baroque edifice of bizarre complexity. Here, for you to skip, skim, or scrutinize, is how the Wikipedia entry on photoreceptors describes the protein-to-protein transduction chain:

The rhodopsin or iodopsin in the outer segment absorbs a photon, changing the configuration of a retinal Schiff base cofactor inside the protein from the cis-form to the trans-form, causing the retinal to change shape.

This results in a series of unstable intermediates, the last of which binds stronger to the G protein in the membrane and activates transducin, a protein inside the cell. This is the first amplification step - each photoactivated rhodopsin triggers activation of about 100 transducins. (The shape change in the opsin activates a G protein called transducin.)

Each transducin then activates the enzyme cGMP-specific phosphodiesterase (PDE).

PDE then catalyzes the hydrolysis of cGMP. This is the second amplification step, where a single PDE hydrolyses about 1000 cGMP molecules. (The enzyme hydrolyzes the second messenger cGMP to GMP.)

With the intracellular concentration of cGMP reduced, the net result is closing of cyclic nucleotide-gated ion channels in the photoreceptor membrane because cGMP was keeping the channels open. (Because cGMP acts to keep Na⁺ion channels open, the conversion of cGMP to GMP closes the channels.)

As a result, sodium ions can no longer enter the cell, and the photoreceptor hyperpolarizes (its charge inside the membrane becomes more negative). (The closing of Na⁺channels hyperpolarizes the cell.)

This change in the cell’s membrane potential causes voltage-gated calcium channels to close. This leads to a decrease in the influx of calcium ions into the cell and thus the intracellular calcium ion concentration falls.

The lack of calcium means that less glutamate is released to the bipolar cell than before (see below). (The decreased calcium level slows the release of the neurotransmitter glutamate, which can either excite or inhibit the postsynaptic bipolar cells.)

Reduction in the release of glutamate means one population of bipolar cells will be depolarized and a separate population of bipolar cells will be hyperpolarized, depending on the nature of receptors (ionotropic or metabotropic) in the postsynaptic terminal (see receptive field).

Evolution and the limits of what can be achieved within cells with proteins are behind the rather involved transduction chain. To achieve its current performance, the primate eye has made use of a succession of incremental changes that began in the Cambrian era 540 million years ago.

Human engineers aren’t limited to making only incremental changes. My first car, a 1977 Chevrolet Malibu , had a bulky—and balky—carburetor to mix petrol and air. My second (and current) car, a 1993 Honda Civic , has a fuel injector to do the same job.

The fuel injector didn’t evolve from the carburetor, nor did the transistor evolve from the thermionic valve. Both innovations, which are simpler and more effective than their predecessors, resulted from leaps of engineering and scientific imagination—which brings me, at last, to my main point.

The devices in our bodies are intricate and complex, but they’re too fussy to be the work of an intelligent designer.