I. Phototransduction
In this section, I assume the reader has a basic familiarity with action potentials, the electrical mechanism by which neurons communicate.
The explanations and descriptions I've offered so far have been quite simplified, to give you an overview of what tools the retina has to work with. In this section, I really get into the weeds of molecular biology, which isn't everyone's favorite way to marinade a pickle. Nonetheless, I've decided to put it in because I think that the process it describes is worth understanding minutely and will give you a firm foundation for understanding the basis of vision.
The first act the visual system undertakes in constructing vision is the transformation of individual photons of light into the type of electrical signals used by the nervous system to communicate. This transformative process is a transduction; because it deals with light, it's phototransduction. The cells specialized to respond to incoming photons are the photoreceptors. As we saw in the previous section, there are two kinds in humans: rods and cones. It might be helpful to refer back to their anatomies before we continue.
Both rods and cones interact with photons in fundamentally the same way. Because it was the first to be studied, I'll use rods to analyze the process. As I mentioned above, we'll have to consider the interactions and chemistries of proteins at a microscopic level. Because of this, don't expect to understand all the information about to fly at you here at once.
I. The Dark Current
Even apart from their weird shape, which has specialized through evolutionary time to aid in their function, photoreceptors are pretty atypical neural cells. Firstly, they don’t act through action potentials, the classical all-or-nothing principle employed by many neurons. They undergo graded potentials in response to stimulation instead. Graded potentials are incremental changes in the membrane potential of a cell. Each incremental change is associated with a different rate of neurotransmitter release. These graded potentials are also localized to specific regions of the cell membrane - they don't run down the entire dendrite of the cell like an action potential does. A key to understand with this type of electric potential is that through it, some indication of the level of excitation within a neuron is continually being transmitted.
A second unusual property of photoreceptors is that they hyperpolarize in response to their stimulus, light. This means that when a photoreceptor processes light, it slows down its release of neurotransmitters. It also implies that when there is no stimulus - when there is darkness - these neurons are constantly being depolarized. This phenomenon is called the dark current. It is caused by the inflow of Na+ cations down their electrochemical gradient and into the cell.
The image above shows the membrane of a rod in light and in darkness. It shows Na+/Ca2+/K+ (NCKX) exchanger channels, which each rod expresses throughout its membrane. These channels use active transport - an energy-consuming mechanism - to eject positively charged ions (cations) out of the cell. The sodium ion, Na+, is especially targeted by this process, and a good deal of it ends up on the outside. An electrochemical gradient is thus established. The electrical part of this gradient refers to difference in charge created when an excess of positive charge is stored in the extracellular matrix outside the cell, making the inside of the cell relatively negative in charge. The chemical part is a result of diffusion - when many molecules of the same type are grouped in unequal numbers, they'll tend to want to mix until equality of numbers is reestablished. If they're prevented in doing so, like they are by the photoreceptor cell membrane, a state of potential energy will be created.
With this arrangement, Na+ ions have two mechanistic reasons for a natural, passive tendency to travel into the cell. When they do so the voltage inside of it will shoot upward. In the dark, they are permitted to do so, causing the depolarization which characterizes the dark current. When a reading of a rod cell kept in darkness is taken using a microelectrode connected to an amplifier, the cell gives steady readings of what look like valleys with baselines of negative voltage and peaks of less negative voltage:
Because cell membranes are impermeable, ions must use special channels to move to and fro. I mentioned that 'exchanger' NCKX channels force the Na+ ions out. In photoreceptors, these sodium ions can flow in again through cyclic-nucleotide-gated (CNG) channels. This type of channel interacts with a molecule called cyclic guanosine monophosphate (cGMP). The cooperative binding of at least three cGMP molecules directly on the cytoplasmic surface (which faces the inside of the cell) of the channel keeps it forced open.
The skeletal structure of a cGMP molecule
A very symbolic cyclic-nucleotide-gated (CNG) channel
The cGMP molecule is synthesized from guanosine triphosphate (GTP), which is very similar to the guanine molecule used in the construction of RNA and DNA. This conversion is catalyzed by guanylyl cyclase, and it occurs 5-10 times faster in cones than in rods; giving cones the ability to replenish their cGMP stores at a relatively quicker pace. This is a note that will become interesting when we consider cone 'light adaptation'.
II. The Absorption of Photons
A photon of light which falls on the retina is absorbed by vast numbers of photopigments present in the membranous discs of the outer segment of the photoreceptor. In rods, these photosensitive structures are called rhodopsin; in cones the job is done by related pigment molecules, photopsins.
Rhodopsin is composed of two parts. The first is the opsin, which is the structural protein actually embedded throughout the disc membranes of a rod. Tightly bound within the opsin is the second component, retinal, which is the chromophore - the molecule that absorbs the photon. It’s a derivative of vitamin A, which you get through your diet (it’s most highly concentrated in fish oils). The retinal itself is maximally sensitive to UV light. It is its interaction with the opsin to which it is bound that allows it to absorb light of different wavelengths.
Retinal can change its chemical shape - what's known in the jargon of chemistry as an ability to 'assume different isomeric forms'. When it absorbs a photon, it changes between two such forms, straightening from a ‘bent’ form known as 11-cis-retinal to the ‘straight’ all-trans-retinal, in the manner depicted below:
Because of its central position within the opsin protein, this change initiates a series of relaxations in the structure of the opsin. The structure the protein ultimately takes is a semi-stable form called metarhodopsin II. In the simplified ribbon diagrams below, note how the top right alpha-helix of the rhodopsin shifts as it transitions into metarhodopsin II. The area is marked by an asterix.
The area marked by an asterix is especially important because it is the locale which this semi-stable form of the rhodopsin protein uses to activates a G protein. This is the next step in our phototransduction cascade. One rhodopsin molecule changed into one molecule of metarhodopsin II can stimulate many hundreds of G proteins as it diffuses through the disc membrane. This ability of one molecule to set many others into motion is known as the first amplification of the light signal.
To avoid too much stimulation, metarhodopsin II molecules are quickly deactivated by a two-step process. The first step is phosphorylation - the addition of phosphate groups - to the molecule, mediated by a protein specialized for the job, rhodopsin kinase. Phosphorylation primes it for the binding of arrestin, a molecule which blocks coupling to the G protein. The metarhodopsin II molecule then decays, resulting in the separation of the all-trans-retinal and the opsin components it contains.
Just before the all-trans retinal molecule is let go by the opsin, another step occurs that I left out above. In it, an enzyme called retinol dehydrogenase turns the all-trans retinal molecule into all-trans retinol. Once a replacement retinal molecule in the correct cis configuration arrives, the retinol is transported and then recycled - turned back into 11-cis-retinal - by a group of proteins called photoisomerases. This recycling occurs in the retinal pigment epithelium, which is on the far outer edge of the retina. Once the cis version of retinal is regenerated, it can be moved back to the rod cell to be recombined with an opsin, allowing rhodopsin to reform and the cycle to occur again.
When many rhodopsin molecules are broken down in such a way, and lie waiting for replacement chromophores, the photoreceptor cell as a whole is said to be bleached; it is non-responsive to further light stimulation. The use of this word arose because rods become transparent if they’re exposed to light for long enough to lose function. Reversing this bleached state (when you return to a dark place) in rods takes about 20 minutes, while in cones the reversal is extremely fast. This ability of cones to recover rapidly partially accounts for the capacity of our eyes to adapt to varying environmental light intensity.
As a fun aside, this bleaching process once gave people the idea that it could be used to identify murderers. The reasoning went so: if the murderer's face was the last thing a victim saw, its features might be bleached onto the retina. If only the retina could be quickly enough isolated, chemically treated, and perused, then Arthur Conan Doyle could be removed from the English canon. Unfortunately, it was found that the images produced, if they could be deciphered at all, were little more than blobs.
An actual rod-bleaching experiment. A rabbit was made to stare at a window before its retina was removed and chemically prepared. The window-pattern of bleached rods, depicted on the right, appeared transparent.
But to return to a point made a few paragraphs ago: how does the altered rhodopsin molecule - metarhodopsin II - function? As I briefly stated, the semi-stable metarhodopsin II stimulates a G protein. G proteins are fully called guanine nucleotide-binding proteins. They are a protein family that functions in various biological signaling cascade systems - systems such as the one I'm describing now, which have a lot of steps that follow and cause each other. A G protein is characterized by its ability to hydrolyze GTP into GDP. This means it is able to cleave and release a phosphorus atom from a GTP molecule.
While bound to the GTP molecule, a G protein is 'on' and contributes to a signaling cascade. When it chops a phosphate group off, it deactivates itself, turning 'off':
There are many different types of G proteins, each of which are specialized to serve in some set of signaling cascades. The particular type of G protein used in phototransduction is called transducin. Proteins are composed of different amino acid chains (each of which is called a 'subunit') twined around each other. The transducin G protein is composed of three subunits: α , β and γ. The Tα subunit is normally bound to GDP, which serves to keep this subunit inactive. Metarhodopsin II acts by exchanging this GDP with GTP present in the cytoplasm. This causes the transducin molecule to break into two pieces, with the Tα subunit - the 'active' component of the G protein - dissociating and activating cGMP phosphodiesterase (PDE).
A severely fictionalized still life of transducin getting halved and this turning it on.
Each active transducin Tα subunit activates one PDE. Because it possesses inherent GTPase activity, the active subunit eventually hydrolyzes GTP into GDP and so inactivates itself, ensuring that the light signal is not relayed continuously.
PDE is the final component of the phototransduction cascade, acting directly on cGMP molecules to achieve the ultimate goal of this cascade, which is to halt the inflow of Na+ ions into the photoreceptor cell. The PDE molecule is an enzyme (a type of protein) which is specialized for breaking a class of chemical bond called phosphodiester bonds. In photoreceptors, PDEs have an inhibitory gamma subunit which keeps them dormant in the cytoplasm. The newly freed-up Tα subunit of the G-protein transducin in turn frees up a PDE by removing this inhibitory subunit.
This prompts the phosphodiesterases into action. It starts breaking cGMP down into 5c-GMP at an incredible rate - a PDE can hydrolyze over 10^3 cGMP molecules per second. This one-to-many ratio effects a second amplification mechanism.
An even more severely fictionalized lateral view of PDE, disguised as a washing machine.
With the hydrolysis of cGMP molecules, the cGMP-gated ion channels in a photoreceptor's cell membrane close down and Na+ ions no longer flow into the cell. The voltage plummets downward, because the exchanger channels still pump Na+ out. This is the net effect of light - the hyperpolarization, or decreased voltage, of a receptor cell. When hyperpolarized, the receptor releases relatively less neurotransmitter. This is the fundamental information coding process underlying our brains' ability to conjure up its interpretation of the external world that surrounds us.
As I briefly stated in the photoreceptor page, the neurotransmitter these cells release is called glutamate. Its continuous release is achieved by specialized structures in the presynaptic terminal called synaptic ribbons. The ribbon is a long electron-dense organelle, made up largely of proteins. To it are tethered many vesicles containing glutamate inside of them.
There can be hundreds of these vesicles at one such ribbon, and their release is regulated by a special kind of ion channel, the L-type calcium channel. This channel is characterized by a rapid gating mechanism; by relatively little inactivation; and by an activation range that spans the resting potential of the cell membrane, which accounts for the continuous vesicle release the photoreceptor exhibits.
Interestingly, the mechanism which maintains a steady supply of glutamate-filled vesicles to the ribbon synapse is currently unknown.