VOL. 52, 1964 CHEMISTRY: KOLIN AND COX 19

1 Woolley, D. W.. these PROCEEDINGS, 44, 1202 (1958).2Woolley, D. W., and N. K. Campbell, Biochim. et Biophys. Acta, 40, 543 (1960).8 Dombro, R. S., L. S. Bradham, N. K. Campbell, and D. W. Woolley, Biochim. et Biophys.

Acta, 54, 516 (1961).4 Woolley, D. W., and N. K. Campbell, Biochim. et Biophys. Acta, 57, 384 (1962).5 Woolley, D. W., in Transfer of Calcium and Strontium Across Biological Membranes, ed. R. H.

Wasserman (New York: Academic Press, 1963).6 Woolley, D. W., Biochemical Bases of Psychoses or The Serotonin Hypothesis About Mental

Diseases (New York: J. Wiley and Sons, 1962).7 Woolley, D. W., and E. N. Shaw, these PROCEEDINGS, 40, 228 (1954).8 Woolley, D. W., and E. N. Shaw, Brit. Med. J., ii, 122 (1954).9 Weil-Malherbe, J., J. Mental Sci., 101, 733 (1955).10 Pare, C. M. B., M. Sandler, and R. S. Stacey, Lancet, 1, 551 (1957).11 Woolley, D. W., and T. van der Hoeven, Science, 144, 883 (1964).12 Kalckar, H. M., and E. S. Maxwell, Physiol. Rev., 38, 77 (1958).13 Slack, J., K. Simpson, and D. Y. Y. Hsia, Pediatric Clinics of North America, 7, 627 (1960).14 Sterling, R. E., and P. L. Day, Proc. Soc. Exptl. Biol. Med., 78, 431 (1951).15 Vane, J. R., Brit. J. Pharmacol., 12, 344 (1957).16 Erspamer, V., Nature, 170, 281 (1952).17 Woolley, D. W., these PROCEEDINGS, 44, 197 (1958).

CONTINUOUS-FLOW ELECTROPHORESIS IN SERPENTINE LIQUIDCOLUMNS STABILIZED AGAINST THERMAL CONVECTION*

BY ALEXANDER KOLIN AND PAUL COXDEPARTMENT OF BIOPHYSICS AND NUCLEAR MEDICINE, UNIVERSITY OF CALIFORNIA (LOS ANGELES)

Communicated by Martin D. Kamen, March 31, 1964

Continuous electrophoresis in a flowing liquid of uniform density without stabi-lization by a porous medium through which it flows (such as paper curtain electro-phoresis' or electrophoresis in a liquid stream traversing a bed of glass powder2' 3)has been attempted by numerous authors. The reason for this trend of develop-ments is the desire to eliminate undesirable effects like electro-osmosis, adsorption,and "fanning" which tend to reduce the resolving power and impede the passage ofsome types of components through the electrophoretic column, especially suspendedparticles. Dobry and Finn4 used a vertical column of fluid, 4 ft in length, about2 X 3 inches in cross section streaming upward at a typical rate of 3.4 mm/sec. Apotential gradient of 6 volt/cm was established transversely to the flow. Themixture to be fractionated was injected at the bottom through a slit so as to form aribbon 2 mm thick spreading over the width of the cell. Their approach can beconsidered as an intermediate step toward a completely unsupported liquid column,since they found it necessary for stabilization to raise the viscosity of the buffer toabout 10 cp by addition of Methocel. The separated fractions were collectedthrough seven separate exit slits at the top of the column. The small number ofcollection exits limits the resolving power of this instrument.

Separations in a streaming liquid without the use of additives to increase theviscosity were obtained by Barrolier, Watzke, and Gibian5 and by Hannig6 resortingto a nonvertical fluid sheath (about 50 X 50 cm) of capillary width (about 0.5 mm)for stabilization against thermal convection. The narrow lumen of the fluid sheath

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

20 CHEMISTRY: KOLIN AND COX PROC. N. A. S.

sets a limit to the scale on which separations can be carried out, as well as to thesharpness of the flowing zones carrying the separated components, and makesseparation of particles which tend to sediment especially difficult.A different approach was used by one of the authors7 to obtain a fluid column stabi-

lized against thermal convection. This approach is based on the realization that ther-mal convection depends on the presence of a unidirectional gravitational field andthat it can be eliminated by slowly rotating a fluid volume about a horizontal axis toobtain conditions equivalent to a rotating gravitational field.8 A detailed con-sideration of the mechanism of suppression of thermal convection will be foundin a previous paper.9

B

INB B

Be/FIG. 1.-N: north pole of a bar magnet} /2iX ///j/S5X% \ surrounded by an annular sheath of electro-t/g'//W;/,,'Sm \ lyte solution. B: vectors symbolizing a

/- L Gil Bradial magnetic field. IN: injector intro-B y// ,, ' /B ducing a mixture of charged particles to be////zf,' 3,S 8 separated. The electric current, which is\/v9z;SH X {/ perpendicular to the page, causes the electro-

lyte to circulate as shown by curved arrow.

B B

B

Figure 1 shows a cross section of the electrophoretic column which has the shapeof a ring surrounding a soft iron cylinder N which is sandwiched between the northpoles of two bar magnets (not shown) so as to generate a substantially radial mag-netic field near its surface which is symbolized by the arrows B in Figure 1. Thesemagnets are coaxial with the cylinder N, the common axis being perpendicular tothe page. The electric current flows toward the reader through the annular elec-trophoretic column. The electromagnetic forces experienced by the fluid in theradial magnetic field set it in rotation as indicated by the curved arrow pointingcounterclockwise. The ions move thus on spiral paths resulting from the elec-trophoretic migration at right angles to the page combined with the circular mo-tion. The mixture to be separated is injected as a fine streak at the center of theannulus through the injector IN. It breaks up into n fine spiral streaks of differentpitch for components of different electrophoretic mobilities. These componentscan be removed and collected simultaneously in separate containers.The objective of the present paper is to describe an alternative method of stabili-

zation against thermal convection which is based on the same principle as the systemdescribed above but which does not use fluid motion in a closed circular path anddoes not employ a magnetic field. It possesses the advantage that it can be morereadily adapted to large-scale separation and that its resolving power can be in-creased by simply augmenting the length of the electrophoretic path without the

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

VOL. 52, 1964 CHEMISTRY: KOLIN AND COX 21

danger inherent in the spiral electrophoresis system that the nth spiral of com-ponent A may coincide with the mth spiral of component B for a long migrationpath.

The Serpentine Electrophoretic Column.-We shall refer for discussion of thestabilization principle to a previous paper9 and confine our attention to the electro-phoretic separation process. Figure 2 shows a serpentine channel oriented so thatthe buffer solution is transported opposite to the direction of the gravitational fieldvector g. Buffer solution is flowing from the Mariotte bottle MB through tubing Tvia the manifold M into the separation cell whose electrophoretic column has theindicated serpentine shape. The buffer enters the separation column at the bottomin the direction of the arrow F and meanders toward the top to reach the escapemanifold (collector) C where the buffer and the separated fractions leave the cellthrough small tubes t shown in perspective in Figure 3. A particle entering the

MB

C

0 - ---...X....FIG.2.-Scheme of the serpentine electro-H 1 Xophoretic separation cell. MB: Mariotte-H- - - ~ : U bottle. T: tube conveying buffer solutions

ITT from MB to the cell. IN: injector introduc-U ing the mixture to be fractionated. F:<-~=~- - - arrow indicating the direction of buffer flow.

M: manifold entrance. C: collector, which,' J/R collects the separated fractions and dis-

g tributes them via outlet tubes t among thetest tubes TT. g: gravitational field vector.

IN- *---- -----

F IN -- x

T

buffer stream at the exit of the injector IN and differing in density from the buffersolution will oscillate about the midline of the flow channel due to sedimentationas it moves in the meandering stream. Precipitation of such particles can beavoided even if their density differs considerably from the density of the buffer ifthe buffer flow is rapid enough. The particles follow then an oscillating path in-dicated by the dashed serpentine line in Figure 2. As a rule the density of the buffercan be adjusted by dissolving sucrose in it, and the stream can be made rapid enoughto make the deviations of the dashed particle path from the center line of the flowchannel inappreciable. In the present treatment, it will be assumed that thedensity and velocity of the fluid are adjusted so as to obtain this condition. Theeffects of deviations from it will be considered in a subsequent communication.

The Apparatus.-The electric current is passed through the buffer at right anglesto the page (Fig. 2) by means of electrode plates which are to be imagined below andabove the page parallel to it. The buffer enters the cell through the tube T and leaves

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

22 CHEMISTRY: KOLIN AND COX PROC. N. A. S.

through the collector tube t at the rate of about 1 drop in 10 seconds for each tube tobe collected in test tubes (TT). Electrophoretic migration occurs thus at right anglesto the buffer flow as well as to the plane of the page. To suppress contaminationof the buffer in the electrophoretic column by electrolysis products, the electrodesare separated from the serpentine fractionation channel by cellulose-dialyzing mem-branes, and the electrode compartments are perfused by buffer as shown in perspec-tive in Figure 3. The buffer enters the electrode compartments through the tubesT1 and T2 and leaves them at the rate of about two drops per second through thedrain tubes D1, D2. Tube Tm conveys the buffer to the serpentine channel via themanifold M and the buffer escapes through the collector tubes t,t. . . into the testtubes T,T.... Some of the tubes t,t... carry the separated components of themixture introduced from reservoir R into the buffer stream via injector IN. Themembranes are compressed between two rubber gaskets (not shown in the figure)which parallel the serpentine channel. The membranes form thus the sidewallsof this channel. The planes of the membranes are labeled PM in Figure 3. Theinjector IN is a gauge #24 or #22 hypodermic needle inserted through a polyethylenescrew into the manifold M so as to project into the serpentine channel as shown inFigure 2. The needle is bent slightly so that rotation of the screw moves its exithole up or down, thus allowing the centering of the origin of the streak in the ser-pentine channel.The electrodes E1 and E2 do not come down to the bottom of the electrode com-

partments EC. This leaves the first turn of the flow channel unobstructed to ob-serve the accuracy of centering of the streak containing the components to be

FIG. 3.-Perspective view of the Tb Dserpentine separation cell. T,, T2: tubesconveying buffer solution to the electrodecompartments EC. Ta, Tb: standpipes in :-which the buffer filling the electrode corm- Tpartments comes to a terminal level. D1,D2: drip tubes through which the bufferperfusing the electrode compartmentsescapes steadily. E1: one of the twoelectrodes (stainless steel plate). Bp:_--electrode binding post. +, re T --connecting the electrodes to the power Tsupply (Electronic Meas. Co., Model C636). PM: "seams" at which the elec-trode compartments can be separated from Ethe central cell body to introduce the Rcellulose membranes which separate theelectrode compartments from the cellbody. M: manifold through which thebuffer supplied by tube Tm enters theserpentine channel. IN: injector. :T-Splastic screw allowing rotation of injector ECand its removal. R: reservoir containingthe mixture to be fractionated. To: tubelinking R to the injector. C: collectorcomprising 60 collecting tubes tt. \\:IM iTT... test tubes. When in operation, T,\K j PMreservoir R is raised (as shown in Fig. 4a) A- T,to obtain sufficient hydrostatic pressure forinjection of its contents into the separationcell.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

VOL. 52, 1964 CHEMISTRY: KOLIN AND COX 23

separated, which are injected through IN from reservoir R via tubing To. Avoid-ance of an electrical current in this space has the advantage of eliminating elec-trolysis at the metal tube IN making it thus unnecessary to insulate it by a coat oflacquer.

Figure 4a shows a photograph of the separation cell with a 15-cm scale attachedto indicate dimensions. Only the thickness of the buffer sheath (2 mm) filling the7-cm-wide serpentine channel is critical. The main parts are labeled to corre-spond to Figure 3.

Figure 4b shows the complete apparatus including the Mariotte bottle and thecollecting stand CS. The test tube stand TS can be swung out from under thecircle C holding the collector tubes tt, the left stand column acting as a pivot. Thecollector tubes are arranged in a spiral pattern and are numbered. Each test tubebears a number corresponding to the collector tube which feeds it. The plastictray holding the ends of the collector tubes tt ... can be moved up and down on itsstand (together with the test tube stand) to adjust the rate of buffer flow through thecell.

Figure 4a shows a plug P next to the injector plug IN. A second injector canbe introduced through this plug so as to conduct two simultaneous independentseparations. If the cell is wide enough, many more injectors can be added. If thesame mixture is fed from all of the n injectors, the instrument carries out preparativeseparation on an n-fold scale. The separated fractions reach the collector C whichconsists of 60 0.5-mm channels (tubes ttt...) which guide them to the test tubesTT1 ....

FIG. 4.-(a) Photograph of serpentine separation cell. The lettering corresponds to Fig. 3. P:plug closing an orifice through which a second injector can be introduced. (b) View of completeapparatus. MB: Mariotte bottle. CS: collecting stand. TS: test tube stand. C: collectorterminal plate-. tt: collector tubes.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

24 CHEMISTRY: KOLIN AND COX PROC. N. A. S.

Fractionation of Ions and ofSuspended Particles.-Figure 5a shows a photograph ofa fractionation of a mixture of 3 dyes, Evans blue, rose Bengal, and "Brush" re-cording ink (green). Two simultaneous separations are carried out by means oftwo injectors. There is no hazard of remixing of the separated components in thefractionation column.

( )

:10 20 30

FIG. 5.-(a) Simultaneous separation of two mixtures injected through two separate injectors.The mixtures do not have to be identical as in this case. The streaks in both separation patternsare, from left to right, Evans blue, rose Bengal, "Brush" recording ink (green). (b) Collectionpattern for a double separation pattern of same composition as in (a). The sequence of colors isthe same as in (a). Evans blue appears in test tubes 2, 3, 16, 17; rose Bengal in 6, 7, 18, 20, 21;and the green dye in tubes 9, 10, 11, 23, 24, 25. Tubes 8 and 22 contain a dilute mixture ofrose Bengal with the green dye. The collected fractions are very dilute in tubes 6, 8, 11, 18, 25.The test tubes 5 and 19 are empty.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

VOL. 52, 1964 CHEMISTRY: KOLIN AND COX 25

Figure 5b shows a photograph of 30 test tubes removed from the spiral rack andplaced linearly in the same sequence in which they were arranged in the collectingrack. There are two collection patterns corresponding to the two separation pat-terns of Figure 5a displaying the same sequence of colors. The absence of fluidfrom tubes 4 and 19 and the occasional nonuniformities in the fluid level in the testtubes illustrate partial or complete stoppage of flow through a small fraction of thecollector tubes, which, however, does not impair the effectiveness of the collectionoperation.

FIG. 6.-Separation of a suspension of two types of particles into its constituent com-ponents. From left to right: red blood cells, yeast cells (Saccharomyces cerevisiae).The separation cell is photographed in two sections for better lighting. There is anarea of overlap between the two white arrows which mark the same point of the cell inthe upper and lower photograph.

Figure 6 shows separation of two different species of particles: Saccharomycescerevisiae and red blood cells.' The latter are the faster component. To avoidexcessive precipitation of the particles, 17 gm of sucrose were added per 100 ml ofthe final solution prepared from a pH 10 "Hydrion" buffer (1 tablet per 100 ml)diluted 1/30. The particle streaks were photographed by dark-field illumination

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020

26 CHEMISTRY: KOLIN AND COX PROC. N. A. S.

using Polaroid film. A wide beam of parallel light was passed through the celltoward the camera deviating from the line of sight sufficiently to miss the cameraobjective. The particle tracks are depicted by scattered light as white streaks ona dark background.

Cooling of the solution is not necessary. As can be seen from Figure 5a, theseparation pattern is undisturbed by thermal convection, even in the absence ofcooling provisions in a cell made entirely of thick Lucite blocks. It is, however,possible to cool the serpentine fractionation column by making the back half ofthe cell of hard-anodized aluminum painted with 5 layers of boat enamel for elec-trical insulation. The aluminum block can be cooled through channels drilledinto it by perfusion with cold water. Without cooling by perfusion, the tempera-ture of the buffer at the collector entry is about 20C above the temperature at theentrance manifold at a buffer flow rate of 5 cc/sec and a current of 40 mA at avoltage of 90 volts across the cell. By perfusion with a coolant the temperaturecan be easily lowered to a desired value.

Although Figure 3 shows the separation cell in the upright position, i.e., so thatbuffer is transported upward, it can also be operated in a horizontal position, i.e.,rotated so that the front plate of the cell facing the reader in Figure 3 is horizontal.In this case, the point of insertion of the tubes Ta and Tb must be changed by pro-viding new openings for their insertion into the electrode chambers so that theywould still be mounted vertically. The original openings must, of course, beplugged. The cell separates equally well in both positions, provided the mixturedelivered by the injector IN is not excessively dense.

Construction details of the apparatus and an analysis of its performance will bepublished elsewhere.'0

* This work has been supported by a grant from the Office of Naval Research.l Durrum, E. L., J. Am. Chem. Soc., 73, 4875 (1951).2 Grassmann, W., and K. Hannig, Naturwiss., 37, 93 (1950).3Svensson, H., and I. Brattsten, Arkiv. Kemi, 1, 4 (1930).4Dobry, R., and R. K. Finn, Science, 127, 697 (1958).6 Barrolier, J., E. Watzke, and H. Gibian, Z. Naturforsch., 136, 754 (1958).6 Hannig, K., Z. Anal. Chem., 181, 244 (1961).7Kolin, A., these PROCEEDINGS, 46, 509 (1960).8 Kolin, A., J. Appl. Phys., 25, 1442 (1954).9 Kolin, A., these PROCEEDINGS, 51, 1110 (1964).10 Kolin, A. and P. Cox, Anal. Chemn., to be published.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

020