Installed: 1/97--------------------------------Updated: 3/13/08
Figure 1. From: Disc Electrophoresis 1
In 1974, Eugene Garfield, Founder and Charman Emeritus of the , Institute for Scientific Information, which publishes Current Contents and the Science Citation Index , and publisher of The Scientist, reported in
"Selecting the All-Time Citation Classics: Here are Cited the Fifty Most Cited Papers for 1945-1988" Current Contents, Feb. 12, #7, 3-148 (1990), that
Davis, B.J. "Disc Electrophoresis. 2, Method and application to human serum proteins".Ann. New York Acad. Sci. 121:404-427 (1964),
View First page of Disc Electrophoresis. 2
Download original Davis, B.J. "Disc Electrophoresis. 2,(526K)
Ornstein, L. "Disc Electrophoresis, 1, Background and Theory". Ann New York Acad. Sci. 121:321-349 (1964),
View First page of Disc Electrophoresis. 1
Download original Ornstein, L. "Disc Electrophoresis. 1,(602K)
(if you don't already have it !)
To help explain why, I quote from the section on Electrophoresis, in A History of Analytical Chemistry, Laitinen, H.A., and Ewing, G.W., Eds. (Am. Chemical Soc., NY, 1977) pp. 324-325:
"The search for gels with high resolving power eventually led to the deveopment of a special synthetic polyacrylamide which was reported by B. J. Davis and L. Ornstein at a scientific meeting in 1959. S. Raymond and Y. J. Wang described this gel in a publication the following year. Ornstein described the use of the gel for 'disc electrophoresis' in a paper that was privately published and widely disseminated by Distillation Products Industries (an arm of Eastman Organic Chemicals Div. of Eastman Kodak) around 1961. This paper was truly revolutionary not only because it led to the immediate acceptance of acrylamide gels but because the method was consciously based upon the principles that underlie the most sophisticated methods that have been developed to this date. Polyacrylamide gels can be formed with any desired degree of crosslinking and in a wide range of concentrations; both characteristics can be adjusted to form molecular sieving gradients within a bed. Thus, molecules can be discriminated on the basis of size as well as mobility. The versatility of these gels has thus added a new dimension to zone electrophoresis. In addition to giving elegant separations, the technique provides a reliable means for measuring molecular constants, including molecular weight, Einstein-Stokes radii, and electrophoretic mobilities.
It is likely that separation of substances on the simple basis of differential rates of electromigration will soon be obsolete. In 1923 J. Kendall and E. D. Crittenden undertook the separation of isotopes by electrical means. This approach used the fact that when a solution of heterogeneous particles with like charge but different electrophoretic mobilities is subjected to an electrical field, the particles move at different speeds until they separate into contiguous zones each of which contains a single species of particle. The interfaces between the zones are extremely sharp and are self-maintaining because any disruption causes a discontinuity in the voltage gradient which acts to restore the pattern. The concentration of material within each band is determined by requirements for uniform conductivity. Once particles have been classified, all fractions migrate at the same rate because any difference in speed would break electrical continuity and interrupt migration until diffusion restored contact between separated zones. L. Ornstein and B. J. Davis, in 1964, were the first to realize the value of this principle. F. M. Everaerts exploited it successfully, in 1968, in developing a method for separating ions. H. Haglund and his group, in 1970, developed this method to its present degree of sophistication and called it isotachophoresis."
first to use polyacrylamide gels electrophoresis (PAGE), to appreciate the potential of polyacrylamide gels to selectively sieve molecules based on size and to provide detailed prescriptions for their use (the first paragraph above); and
first to appreciatie the utility of, and the rules for manipulating buffer systems to acheive what we have called controlled "steady-state-stacking" in order to concentrate and separate dilute samples, either as an automatic precursor step to further separation by PAGE, or as the sole means of accomplishing the separation (the second paragraph above).
The continuing impact of our work is easy to recognize: Perusal of practically every current issue of scientific journals in the fields of genetics, bioengineering, biochemistry, cell and molecular biology, reveals many papers which illustrate and depend upon PAGE separations of proteins, nucleic acids or polysaccharides, often using an initial steady-state-stacking step. The methods have become so widely used, that the original references are now rarely cited*.
*Current Measure of Impact of our work
As noted above, citation counts, extracted from the Science Citation Index, provide a useful measure of the impact of a scientific work on the scientific publications that are produced afterwards. The dependability of this forensic tool has diminished somewhat because of a gradual change in standards of scholarship that have developed in the past 40 years. In the first half of the 20th century, editors generally demanded that the reference section of a scientific paper contain citations to the original publications of any basic theoretical or experimental protocols used in the work. Over time, that requirement has been relaxed so that reference to any more recent version of such protocol can substitute for the original, on the questionable assumption that by chaining backwards in the literature, the original work can be found. As a result, even when the use of particular protocols persist at very high levels in the scientific literature, frequency of citations to the original publications decay with time.
With the advent of internet search engines, like Google, a very convenient, alternate, forensic tool has become available to fairly easily assess the current cultural impact of protocols and inventions.
For example, the Polymerization Chain Reaction (PCR), invented by Kary Ellis, and for which he received a 1993 Nobel Prize, can be searched with [(PCR OR "Polymerization Chain Reaction") AND Mullis)] to return hits on about 225,000 web pages. (The 'anding' of PCR with Mullis is necessary to eliminate completely unrelated uses of 'PCR'.) Although most of these hits will contain citations to his original paper, this count grossly underestimates the continuing impact of his work. Searching, instead, with [PCR AND (Mullis OR DNA OR Taq OR Polymerase)] returns over 20 million hits. This count is a far better measure of the continuing impact of his seminal contribution.
This can be considered to be a genetic scan for ancestral, lexical 'genes' uniquely- and contingently-linked to the Mullis 'mutation'.
A search with [(electrophoresis OR polyacrylamide) AND ("Davis BJ" OR Ornstein OR Disc OR Disk)], returns about one million hits; and these are mostly citations of our original papers. If the search is for [(electrophoresis AND acrylamide) OR (electrophoresis AND PAGE)], about 8 million hits are returned. A search on only [electrophoresis] returns about 16 million hits. So about half of all current literature that discusses electrophoresis 'descends' from our work. The continuing impact of those 1964 papers is still significant compared to that of the very different and more recent PCR technology.
But that's only half of the story: DNA sequencing studies almost always use polyacrylamide electrophoresis, gel-slab or capillary, for separation, and usually use PCR for prior amplification of samples. A search for [DNA AND (sequencing OR sequence)] yields about 70 million hits. A search for [DNA AND (sequencing OR sequence) NOT polyacrylamide NOT PCR] yields about 50 million hits. So the lion's share of all discussions of DNA sequences and sequencing, most all of which are descendents of both our work and Mullis', currently make no mention of their technological roots.
Similarly, Southern blots, Northern Blots and Western blots (immunoblots), all of which usually use some form of PAGE as their first step, usually fail to even mention electrophoresis. Thus, [("Southern blot" OR "Northern blot" OR "Western blot" OR immunoblot)] yields about 7.5 million hits. [("Southern blot" OR "Northern blot" OR "Western blot" OR immunoblot) NOT electrophoresis] yields about 6 million hits.
The following quotes, excerpted from "Tenuous but Contingent Connections", give bits and pieces of the contingencies that led BJ Davis and me into the electrophoretic arena.
"...Gerry Oster had been working with polyacrylamide gels and photopolymerization (e. g., [40-42]). Knowing of my interests in embedding media and microtomy , he invited me to Brooklyn to introduce me to the chemistry of acrylamide polymerization. He thought polyacrylamide might be useful as an embedding medium for cutting sections of hydrated tissues. I returned to Columbia with a kit of monomers, peroxy acids and photocatalysts, and proceeded to explore the worlds of polyacrylamide. .....
Baruch Joel Davis, 'BJ', then a Fellow in Hematology, began to work with me at the Cell Research Lab. S. J. Holt, of the Courtauld Institute of London, using indoxyl esters, had recently published the first truly high-resolution, enzyme-cytochemical photomicrographs . This convinced me and BJ to initiate an intensive effort to develop other high-resolution, enzyme-cytochemical reagents. Unexpectedly (a story of serendipity, which, however, I will skip here), we were quickly rewarded with the discovery of 'hexazotized pararosaniline'. This compound couples with l-naphthol, indoxyl and 8-hydroxyquinoline faster than any other known coupling agent, producing, particularly with l-naphthol, extremely insoluble pigments [48-50]. This generated a plethora of new methods for phosphatases and esterases. To profit from this new power, we needed better methods for preserving structure and enzymatic activity, simultaneously. With our colleagues, we perfected new freeze-substitution, paraffin-embedding protocols for use with our new staining reagents, and promptly discovered multiple discrete sites of intracellular enzyme activity in various kinds of tissues with each substrate (e. g., l-naphthyl acetate) . Were each of those the same enzyme or different enzymes with similar substrate-specificities?
In 1955, Oliver Smithies had introduced starch gel electrophoresis for high-resolution separation of soluble proteins . M. Poulik, who worked with Smithies, introduced discontinuous buffer systems and enzyme-histochemical staining of electrophoretically separated enzymes, "zymograms", in 1957 . R. Hunter and associates quickly applied Smithies and Poulik's methodology to the separation of enzymes of histochemical interest (e.g.,[54,55]). Here was a set of tools for answering our question. .....
Ralph Engle was also busy preparing to introduce Smithies' starch gel electrophoresis to clinical hematologists . BJ would often visit with Ralph after a session at Rockefeller, and came back with tales of the beauty of the separations, and the woes of the poor reproducibility of gel preparations, the fragility of the gels and their opacity.
[Not from "Tenuous but contingent connections"] Another "defect" of both starch gel and agar gel electrophoresis, which was already appreciated by their practitioners, was the fact that, due to the presence of various anionic residues, such as glucuronic acid in the purest starch fractions (and sulfate residues in agar) the measured mobilities were always biased by electroosmotic (endosmotic) flow of water in the gels; in a sense, since the gels had net negative charges, and therefore "wanted" to migrate in the electric field, but couldn't, the water moved relative to the stationary gels to just compensate. And because the concentration of glucuronic acid residues varied in starch preparations from batch to batch, reproducibility was often severly compromised. Pure acrylamide monomer is a neutral molecule, and polymers and gels produced from it are also neutral and can not induce endosmotic flow.
I naively assumed that I understood all the physical chemistry involved. 'Obviously', molecular sieving by gel pores was the important feature. I immediately concluded that transparent and relatively strong cross-linked polyacrylamide gel, which I knew well, could be easily substituted for starch gel, with the added advantage of the potential to better tailor pore-size to the separation at hand.
We did a quick experiment, preparing a 7 % acrylamide/ 1 % methylenebisacrylamide gel, initiated by a high concentration of ammonium persulfate and accelerated with beta-dimethylamino-propionitrile in a Smithies' starch gel mold, with borate electrode buffers and a sample of human serum. The results were fantastic - better than starch gel! Well, it was 'clear' that we had not optimized the system. Surely we could do better if we prepared a system which was homogeneous with respect to all electrolytes other than the sample proteins. However, the closer we approached this goal, the poorer our resolution. Our attempts to apply 'electrophoretic dogma' were not working. But when we repeated the original 'dirty' experiment, the resolution was extraordinary. Something peculiar was going on.
Like Alexander Kolin (, pp. 9 and 12), we turned to H. Abramson's text  as well as to Longsworth's chapter in M. Bier's book  to try to better understand the possibilities.*
*Incidentally, Kolin and Abramson had been at Mount Sinai. In fact, Abramson only departed a year after I arrived. But neither I nor BJ ever met him, nor was there any connection between us and his co-workers. So Mount Sinai's nurturing (?) of these three lines of electrophoretic workers at different times was purely coincidental.
We also discovered Kolin's 1954 and 1955 papers on 'isoelectric focusing' [65, 66] and thought that perhaps pH gradients had something to do with our results. We decided to permit a mixture of 'universal' anionic pH-indicator-dyes to electrophorese through the gel to look for pH gradients or pH discontinuities. The dyes were introduced into the cathodic reservoir in very low concentration and the power supply was turned on. Lo and behold! A sharp dye front immediately formed at the cathodic end of the gel, and as the exquisitely sharp front migrated into the gel, a colored band of uniformly high optical density grew in thickness behind the front! There was no indication of pH discontinuities behind the front, but the moving, concentrating, steadily thickening band of dyes had to hold a clue to what was going on. The experiment was repeated with added serum protein. By now, we were regularly working with vertical gels in glass tubes to facilitate observation during the run (see Fig. 2). The dyes behaved essentially the same, but a stained albumin band concentrated and then dropped behind the dye stack. We also observed a group of highly refractile, very sharp discs (the other serum proteins) dropping behind albumin.
Thereafter, we quickly unravelled the role of the Kohlrausch regulating function  in these phenomena, (borate had been stacking behind sulfate, with dyes in-between) and successfully designed a system, where sieving effects of polyacrylamide 'pores' plus controlled stacking and unstacking with appropriate weak acids and bases were optimally meshed, to produce even higher resolution and more reproducible separations using the 'deficiencies' of moving boundary electrophoresis . Now, we also understood the basis for the performance of Poulik's discontinuous buffer systems (which he did not understand). On March 24, 1959, BJ presented our first results at the meeting of the N.Y. Blood Society, and on April 12, 1959 I discussed them briefiy at the Histochemical Society Meeting . Independently, on May 21, 1959, S. Raymond and L. Weintraub, of the Pepper Laboratory of the University of Pennsylvania in Philadelphia, submitted their paper on the use of polyacrylamide gel for electrophoresis of proteins to Science, and it was published on September 18,1959 .
Our contribution differs from that of Raymond and Weintraub. They and we appreciated and demonstrated the valuable sieving properties of polyacrylamide. However, we also introduced the choice of use of differences in gel pore-size and/ or differences in mobilities of weak acids and bases at various pH's to program discontinuities into the system to automatically concentrate, stack and unstack proteins (see Fig. 3) from dilute solutions to provide extremely thin starting zones . Our steady-state-stacking is of course identical to, and anticipated, 'isotachophoresis' . We explored isotope separations, preparative steady-state-stacking and separations with un-gelled linear, uncross-linked polyacrylamide [72,73]. We also investigated the relationship between the percentage cross-linker and gel swelling, rehydratable gels, electrophoretic destaining and electrophoretic elution of proteins from gels...."
A decade passed, and I was by then deep in a long-term development program with Technicon Instruments Corp. (now Bayer Corp.) to develop an automated diferential white blood cell counter, (which is reviewed in "Tenuous but Contingent Connections" and on my "Hematology and Cytometry" Page).
In the interim, the use of acrylamide gels had spread world wide, and their many advantages and potentials were by then apparently widely appreciated.
Kolin had demonstrated the separation of proteins with "Isoelectric Focusing" in 1954. To oversimplify, the idea was that if you could set up a stable pH gradient in an electric field, and then introduced proteins (or other amphoteric substances), each would migrate to the region of its isolelectric point and then sit there "permanently".
Establishing stable pH gradients in solution usually required heroic manipulatioins or the establishment of a rather restrictive overlaping density gradient to inhibit convective destruction of the gradient.
But the oportunity to make such gradients within a neutral, large-pore, polyacrylamide gel, changed the picture completely. The research of a group in Sweden, led by O. Vesterberg and H. Svensen (1966-1968) and affiliated with LKB Producter (later purchased by Leica), developed a class of synthetic amphoteric small molecules with pKa's spanning a large pH range, which they called "Carrier Ampholytes". If these were electrophoretically run into a polyacrylamide gel they would distribute themselves (over a period of hours) in such a way as to produce a more or less continuous and stationary pH gradient which could be used to separate proteins by isoelectric focussing.
Other members of the same group (led by H. Haglund, 1970), appreciated that if those Carrier Ampholytes were mixed with a sample of proteins and run in a steady-state-stacking mode, the proteins and ampholytes would stack, often with ampholytes between proteins, in such a way that might facilitate both analytical and large-scale preparative separations. They chose to call this method "Isotachophoresis".
In 1972 the New York Academy of Sciences organized the International Conference on Isoelectric Focusing and Isotachophoresis.
I attended those meetings as a way of "catching up". I had assumed that since by then, "everyone" was using polyacrylamide gels, and thousands had read my "Disc Electrophoresis: 1" paper, (which includes fairly extensive Appendices on the physical chemistry of virtually all aspects of electrophoretically-relevant phenomena), most in attendance would by now know at least as much about the pertinent electro-physics as I. I was mistaken!
In a paper "Instability of pH gradients Formed by Isoelectric Focusing in Polyacrylamide Gel" by A. Chambach, P. Doerr, G.R. Finlayson, L.E.M. Milers, R. Sherins, and D. Rodbard Annals New York Acad. Sci. 209, 44-60 (1973), the authors carefully documented what, by then, was a frequently observed set of phenomena: during prolonged Isoelectric Focussing runs in gels. There was typically a continous spreading of the gradient around the neutral pH region; and not infrequently also a continuous drift of the whole gradient towards one electrode. There seemed to be great confusion at the conference as to the explanation of these "artifacts". To me, there was no confusion at all.
In the published discussion (pp 61-61) following this paper I said:
"I'm rather surprised that there really seems to be any mystery about why the gradient spreads symetrically. There are a couple of different ways of explaining it, both of which are physically equivalent although they sound like different physical devices. In fact, two of the ones that you mentioned explain it and they're really the same device. You discounted (addressed to Dr. Chrambach) electroosmosis because the pH instability effect is bidirectional. But, it isn't electroosmosis stimulated by the gel, it is electroosmosis stimulated by the ampholyte. In other words, if you had equilibrium conditions you would have stationary ampholyte throughout the gel. That is the same as if it were a charged staionary substrate. And, in fact, you have different conditions at both ends of the gel, therefore there is endosmotic flow into the gel from both ends. This is voltage and time dependent. In fact, you have a spreading gradient around the neutrality point that is faster than would occur by diffusion alone.
Another way of explaining it is: Consider, with the ampholyte gradient already established, the addition of an anion (weak acid) into the alkaline reservoir. That anion is going to pass through the gradient and as it passes through the gradient, of course it is still a weak acid, let's say carbonate, which if your'e careless gets absorbed from the air into the alkaline reservoir. As that carbonate passes through the gradient and the interaction of the pK of the carbonate and the ampholytes occurs, conditions similar to steady-state stacking are produced and cause ampholyte to migrate. If, for example, you choose an ampholyte gradient that is below pH 7 then there will be one-directional migration and the entire gradient will move out of the tube with time.
Now, the stronger the weak anion or cation in either of the two reservoirs, the faster such migration will be.
Water is, in fact, a weak buffer, with a hydrogen ion and hydroxyl ion coming from each end, you're getting the same effect---and that turns out to be electroosmosis. You can look at it either as an approach to steady-state stacking that is in fact, causing the ampholytes to spread, or you can look at it physically from the point of view of electroosmotic theory, and it is electroosmosis. You never can get real equilibrium conditions with isoelectric focusing. There will always be net migration because, in the limiting case where you exclude all other possible anions and cations, you have your hydroxyl and hydrogen ions contributing.Therefore, there is an effect that can be studied either as an approach to steady-state stacking or as electroosmosis.
In addition, during the early stages of formation of the so-called stable pH gradient there are, in fact, steady-state stacking conditions developing in both ends of the gradient. In fact, a good deal of the rapid early sharpening is due to steady-state stacking and not isoelectric focusing."
As a result of these and other published and unpublished contributions that I had made to the discussions of the presentations at the Conference, I was invited to participate in a concluding (published) extemporaneous Panel Discussion which reveals much of interest about the science-politics of the period; the surprisingly deficient physical intuition and perspective of many of the electrophoresis "experts" at the time (with exceptions, e.g., A.C., T.M.J. and D. R.); and again, just how much the progress and consolidation of scientific understanding depends on contingencies ("acidents" of who says or thinks what, when and where).