ŌThe discovery of how light is changed into a nerve signal in our eyes.Ķ 


The following is a transfer to the web of my slides and lecture notes for a brief talk given at noon on May 14, 2013, for the McPherson Eye Research Institute seminar series.  (It is a brief and biased personal account that references only a fraction of the workers in my own and other laboratories who contributed to solving the issue.)



When I was asked to do one of these brief talks, I decided  to give a participants account of how we figured out how light hitting our eyes is changed into a nerve signal, I messed with this problem for about 40 years. 


Speculations on how our vision works started in antiquity... and heated up in the  17th century with Descartes and others.




Our modern account starts in the 19th century...  Wilhelm Kuhne looked at the Visual Purple described by Fran Christian Boll in 1876, finding that adding a solution of bile salts to retinas dissected in the dark would bring the colored molecules that were bleached by light into solution.




Kuhne influenced Otto Warburg, who had among his many projects an interest in the Vitamin A recently discovered by Paul Karrer.  He assigned different post docs to look for vitamin A in different tissues. 



George Wald,  got lucky when he was assigned by Warburg to look for it in the eye.




The right slide is Wald at the time of his Nobel prize in 1967.


At any given time the questions  permitted are limited by the tools at hand,  at this point there was some good organic chemistry going on, sorting out and crystallizing polyene isomers .   Purifying and crystalizing the different isomers of retinal that could be formed by light led to a furious competition in the late 1940Õs to find which one combined with the opsin protein to make rhodopsin. 



George Wald won the race,  discovering that it was the neo-b, or 11-cis isomer of retinal that complexed with opsin, to then be released on illumination as all trans-retinal. [prezi]  Synthetic 11-cis retinal could be used to couple with the opsin protein to resynthesize rhodopsin, either in retinas, or solubilized by the cardiac glycoside digitonin. 



In the 1950Õs WaldÕs lab, along with others,  began detailed studies of the visual cycle,  the reduction of retinal free by illumination to retinol with transport then to and from the pigment epithelium.



He also described the brief intermediates formed after illumination, that could be stabilized at lower temperatures.  (The figure shows the chicken cone pigment, then called iodopsin.)





and he also characterized the spectra of cone pigments  [prezi]



By now we know in exquisite detail how visual pigments can come in many

different colors.  This actually is a drawing of RBP, retinal binding protein, in which you can tweak the wavelength absorption maximum with different amino acid substitutions.




A serendipitous random walk got me into the field of vision in 1960,




As a bright eyed Harvard freshman I was taken on in a new freshman seminar program by George Wald. (Edwin Land, who invented the Polaroid camera, had given the money to Harvard for the program.)


I had been told to meet Wald by his recent Ph.D. student, Austen Riggs, a new Assistant professor at the University of Texas professor in Austin Texas where I grew up. 


Riggs had sent a letter to my high school biology teaching wanting a dish washer, and hired me.  I washed dishes and learned how to make buffers and do molar calculations. I then lost a bet with a high school homeroom buddy that if I applied to Harvard I wouldnÕt get in.


 I timidly walked in on Wald in my freshman year, he was in an expansive mood, and set me to learning how to do organic extractions of carotenoids from frog retinas.


A lot of what we did in the early 1960s seems really quaint now,  I doubt that many current Ph.D. candidates are required to demonstrate proficiency in glass blowing, hand lettering graphs of data, not to mention reading two foreign languages. 


Wald required that I learn his graphic hand lettering style, that he claimed had passed down through four scientific generations.  I used an ink pen to letter figures in my first Nature paper in 1965 (top right of figure), but later got lazy and did leroy lettering.



Journals at that point were still doing it just like Gutenberg,  figures printed from acid etched lead plate on wood  blocks, IÕve kept this block mailed to me by the Journal of General Physiology on my office book shelf since the late 1960s. 



In my junior and senior years at Harvard I was taking biochemistry courses, and learned about Fisher and Krebs work on sodium borohydride reduction of the schiff base binding pyridoxal phosphate to proteins. 


I used this reagent to irreversibly fasten retinal to opsin during its bleaching, and this let me from 1963-66 do the first modern protein chemistry on visual pigments by irreversibly attaching retinal to opsin and then then identifying the lysine to which it was attached along with adjacent amino acids.



After a two year post doc stint in Neurobiology at Harvard Med, I came to Wisconsin in late 1968 with the first of what turned in to 30 years of EY-00463 ready to support my work on visual transduction in Frog ROS.


The central focus that persisted throughout this project was trying to integrate the physiology and biochemistry of transduction, figuring frogs would be the best image of man, using these large frog rod outer segments, shown here in a scanning e.m. I made with Stan Carlson in about 1969.



 These large structures shear away and reseal when a retina is gently shaken, cones mostly remain on the retina.


So.  how do we get from light hitting rhodopsin to a nerve to closing channels in these plasma membranes to cause a hyperpolarization of the cell.... flail around looking for some new chemistry.



People were starting to find regulatory protein phosphorylations almost everywhere they looked in biochemical pathways, and so in 1970 I went to a Gordon conference in this new field, came back to the lab, made some gamma P-32 labeled ATP,  and finally made an addition to the photo-isomerization story of the late 1940s by finding that the visual pigment rhodopsin became phosphorylated upon illumination, a reaction we now known is characteristic of many other G-protein membrane receptors. 




Properly seeing this reaction required that we realize that seeing effects of light on chemistry must start in complete darkness. This was made possible by the recent availability of infrared image converters developed for the Vietnam war.



A number of  laboratories failed to get our data because they tried the experiment using as ÔdarkÕ the dim red lights we were all accustomed to, that bleached less than 1% of the rhodopsin present, but still quite enough to turn on all the interesting light sensitive chemistry.  


It turned out that at the same time, in  late 1971 Kuhn in DreyerÕs laboratory at Cal Tech was able to see more phosphorylation in brightly illuminated membrane than ones kept in the dim red light, this complemented our work, which was able additionally to do the stoichiometry, finding that many phosphate groups could be attached to each bleached rhodopsin. 


In a series of papers we proceeded to do the sort of things biochemists do, removing components from the system, like rhodopsin kinase,  and then putting them back to make it work again.  (1974-75). 



HereÕs a cartoon slide from Vadim Arshavsky  showing how phosphorylation is necessary to shut off excited rhodopsin by causing binding to the capping arrestin protein, this arrestin function being common to other G protein coupled receptors.  He, along with Peter Calvert purified and reconstituted these components in the early 1990s, including recoverin, a calcium binding regulator of rhodopsin kinase. 


Another step in hunting around for the second messenger needed to communicate between rhodopsin bleaching in the disc membrane and sodium channels in the surrounding plasma membrance  happened in the early 1970s.




Bill Hagins had been inserting electrode alongside rods in the retina to measure the dark current from inner to outer segments that was suppressed by light,




and he found that exposure of rods to high calcium levels caused a similar response to the light induced signal, and so in 1971 the calcium hypothesis came on the scene,  it was modeled on muscles, suggesting that calcium released by disc membranes after rhodopsin excitation closed sodium channels in the plasma membrane. 


This was also a period of intense interest the internal messengers cyclic AMP and cyclic GMP being found anywhere you wanted to look...we found that drugs that cause an increase in cyclic GMP levels by inhibiting the phosphodiesterase that breaks it down increase the permeability of the plasma membrane of suspended rods.


Then,   in the 1974-76 period, we, along other groups, found that cyclic GMP phosphodiesterase was activated by light, by steps unknown at the time, to convert cyclic GMP to GMP




We used the osmotic assay plasma membrane permeability in suspensions of ROS to show that cGMP levels, phosphodiesterase activity, and membrane permeability, changed together over the low levels of illumination that saturate the rod photoresponse.  (1975-76) 


The conventional wisdom was that the cyclic GMP decrease caused by PDE activation would be much too slow to be involved in visual excitation, so that cyclic GMP was most likely involved in rod light adaptation.


We decided to measure how fast the cyclic GMP disappeared, designed a Rube Goldberg mechanical device that the Molecular Biology shop built, to rapidly quench the system at short times after illumination, and in 1976 Mike Woodruff found that cyclic GMP decrease occurred less than 100 msec after illumination, fast enough for cyclic GMP, as well as calcium, to be a candidate for the internal transmitter.



Michael Woodruff, and then Ric Cote, designed fancier rapid quenching techniques to look at the light sensitivity and kinetics of the rapid cGMP decrease and slower recovery. The figure below shows that the cGMP decrease  after a bright and half saturating flash of light, is as rapid as the current change, but recovers more slowly.



We continued to look  for further light sensitive protein phosphorylations and graduate student Arthur Polans  found that two small proteins were dephosphorylated on illumination.



Meanwhile,  in the late 70s  we and other labs were finding light induced changes in GTP levels in rod suspensions. 


Then another clue came from Cassel and Selinger studying beta-adrenergic receptor mediated adenylate cyclase activation in turkey erythrocytes. They found that GTP displaced bound GDP when catecholamines were added, suggesting that adenylate cyclase activation involved binding at a regulatory site.


Fung, Hurley, and Stryer immediately tried this with rod membranes and reported in 1981 that photolyzed rhodopsin catalyzes the exchange of GTP for bound GDP on about 500 molecules of a three subunit protein they named transducin, the alpha subunit of transducin bound to GTP could activate phosphodiesterase.  (Jim Hurley was at ARVO last week an at the celebratory dinner for Vadim.)




Here is a summary cartoon slide sent  by my former postdoc Heidi Hamm,   whose work in my lab with Patricia Witt in the early 1980s was to generate monoclonal antibodies that block light-mediated G Protein activation.  (It only took 27 years to find out the epitopes involved!)


Fast forwarding to the present, Heidi, who is now Earl Sutherland Professor and chair of  the department of Pharmacology at Vanderbilt , was good enough to send this movie from collaborative work showing that for G protein activation and GDP release, a domain of Ga has to open up. 



It animates rhodopsin getting activated, then recruiting the Galpha C terminal, leading to a conformational change that causes separation between the GTPase and helical domains of Galpha, and GDP release.  


So... we had at this point, in the early 1980s a prevailing calcium hypothesis for rapid light shutting of channels in the rod plasma membrane, with the presumption that this cGMP and phosphodiesterase pathway most likely regulated response inactivation or adaptation. 


Against this, as I just mentioned, we were reporting increasingly detailed kinetic data showing the cGMP decrease to be rapid enough for it to qualify as an internal transmitter. 


Also,  we were finding that transferring suspensions of intact outer segments from low to high calcium was also lowering cyclic GMP levels. 


When I suggested in 1980 that photoreceptors in essence worked ŌbackwardsĶ  , being in the dark like the excited states of muscle or other hormone-sensitive cells, high calcium and high cyclic nucleotide levels,  with both going down on,  rather than up, in excitation, the outrage was immediate and loudly voiced during discussions after my talks at several meetings.   I had to grow a very thick skin at this point. 



I was one of the organizers of a Berlin Dahlem conference held in 1983, at which a dramatic climax in this field occurred.  A nature editor who was attending revealed that Nature had received a manuscript from an unknown (to us) young Russian russian researcher named Fesenko (sent to King Y Yau, who was also at the meeting, for review). Fesenko the new the patch electrode recording technique to find channels in the rod plasma membrane that were gated by cGMP. (Five or six labs in the US had been trying, but hadnÕt been able to get it...they were scooped).   This was the missing piece we needed, published in 1985.   The excitation pathway was coming into focus. 




In the vertebrate rod we have found it much easier to discover the pathway between photons and channel regulation than in the invertebrate drosophila folks, in spite of all their elegant genetics. 


During the period of 1984-87 we were obtaining a definitive disproof of the calcium hypothesis, made possible using a new technique we had developed to make purified suspensions of frog rod out segments still attached to their mitochondria rich inner segments,  these structures generated the same dark currents observed in rods in the living retina, which we could measure with the suction electrode technique developed by Baylor, Hodgkin, Lamb and their collaborators.






Work by Grant Nicoll, Benjamin Kaupp, and Paul Schnetkamp in the lab showed that clamping calcium levels to almost zero with calcium chelators and ionophores had no effect on the light induced conductance.  Trevor Lamb and collaborators in Cambridge obtained data leading to the same conclusion.  We described sodium calcium exchange mechanisms in the rod membrane, further helped cement the identify of cGMP as internal messenger by showing in 1986 that an Amiloride derivative blocked both the light and cyclic GMP sensitive conductances.


Our development of this in vitro preparation of purified living truncated rod outer segments in the mid-1980 permitted simultaneous assay of chemical and electrophysiological changes, and made it possible for us to do a new generation of quantitative rapid kinetic experiments on the roles of calcium, ATP, GTP, cGMP, and protein phosphorylation in transduction.  


Once general acceptance of cyclic GMP as the primary internal messenger settled in,  we had the next problem.... a single photon response is quite rapid,  how do we turn off the phosphodiesterase, so that guanylate cyclase can generate cGMP to open them again.  The recovery chemistry to date was too slow. This slide from Vadim Arshavsky shows the problem...


The beginning of figuring this out began when I first met Vadim Arshavsky in the summer of 1986 at an international conference organized by Ovchinnikov, a big wheel in the Russian Academy.  They flew us first to Moscow, where we stayed in the gargantuan Russia hotel,  being privileged guests we were offered all you can eat caviar and cucumbers,  and that was pretty much it for plant and animal protein. 


Then they flew us via Novosibirsk to Irkutsk, where we stayed, and the conference was held in a building constructed for a visit by Dwight Eisenhower in the 1950's that was aborted by the U-2 incident.


The high point of the meeting for me was a cruise and picnic on Lake Baikal, and wandering through primeval Russian forest for the first timeÉ magical, like you expected to see Baba Yaga go swishing past in her mortar and pestle. 


Anyway, on the cruise boats deck, a guy named  Phillipov comes up to me and says he wants to introduce a creature standing deferentially behind him and bobbing up and down - this is his graduate student, Vadim Arshavsky, who he is hoping I might be able to sponsor as a post doc in my laboratory.  We more of less sealed the deal then and there (Vadim was PhillipovÕs propery, he had no say in the matter), and after quite a hassle we got him over to Wisconsin in 1989. 


What we got into when Vadim arrived was trying to develop an observation that cGMP suppressed the GTPase activity of an amount of transducin equal to the PDE. We blew this up into the gamma as GAP story (GAP being a family of small protein regulators of GTPase), got a great press, Nature review etc., did a whole rash of papers on noncatalytic CG sites on the PDE alpha and beta subunit influencing interactions with the PDE gamma subunit, did stuff with recombinant PDE gammas.



We generated these elaborate model drawing of cGMP binding sites doing negative feedback regulation which we would like to forget now, but you have to stumble around before you get it right, which Vadim Arshavky (who went on to Harvard and now is at Duke) and Ted Wenzel finally did. This is why they got the proctor award at ARVO last week.


At the same time  we were dissociating and reconstituting the rhodopsin phosphorylation activation and inactivation process, looking at recoverin, the calcium sensitive regulator of rhodopsin kinase.   Getting complicated, lots of players, lots of people working on them. 


A collaborative effort between our lab and several others was able to provide a complete and quantitative model of the gain stages in the transduction pathway.  Trevor Lamb and Ed Pugh were the main players in this effort and were awarded the Proctor Prize at ARVO a few years ago.



Here is the summary figure from 2000 neuron paper,  which by now has been rather expanded by work of Vadim and others...


IÕm afraid the first person narrative I can give you of what was happening in the transduction field decays by the end of the 1990s,  after the high of excitement in the 1980s of nailing steps in the transduction pathway.


I  figured it was getting to be a good time to quit while youÕre at the top of your game,  and so over a period of years in the mid-1990s I dialed down the transduction research factory and got jobs for my people and dialed up the second line of interest I had been developing for some time, starting a new course on structures, function, and evolution of the human mind, with the Biology of Mind Book appearing in 1999. 


Since I withdrew from the fray, numerous new components have been discovered to tweak and regulate the system. 


Here are some players added to the Transducin PDE story since then



Transducin is inactivated by complex with an RGS protein and several other components involved.  PDE gamma turns out to be an affinity adaptor




Here, finally is how you get the excitation chemistry to shut down.  In R9AP knockout mutants made by Marie Burns and her collaborators, if the GTPase activating complex doesnÕt form the flash response becomes much longer.



Here is the current grand summary slide that Vadim uses.  



OK,  IÕll stop there, I hope IÕve given you some taste of the good old days when we were bumbling around getting the basic transduction pathway figured out.