Proceeding~j of the National Academy of SciencesVol. 67, No. 2, pp. 620-628, October 1970

Interferon: Evidence for Subunit Structure*

William A. CarterDEPARTMENTS OF MEDICINE AND MICROBIOLOGY, JOHNS HOPKINS UNIVERSITY SCHOOL OF

MEDICINE, BALTIMORE, MARYLAND 21205

Communicaed by W. Barry Wood, May 25, 1970

Abstract. During the purification of mouse and human interferons, multipleactive components have been detected. Mouse interferon was purified over500-fold by differential precipitation, centrifugation, gel chromatography, andisoelectric focusing. On electrofocusing, two molecular forms (A and B) werenoted. Form B (isoelectric point 7.35) had a molecular weight of about 38,000and Form A (isoelectric point 7.15), which was equally active, a molecularweight of 19,000. Purified Form B was dissociable into Form A, but the re-verse reaction occurred to a much lesser extent. Human interferon, purifiedabout 1500-fold, is also composed of multiple molecular forms. Form B (pI5.60) had a molecular weight of about 24,000 and Form A (pl 5.35), which maycontain up to 85% of the total activity, a molecular weight of 12,000. Bothforms appear to be equally active. The dissociation of both human and mouseinterferons into subunits appears to take place during dialysis versus low salt(0.01 At Tris pH 7.4). The data are consistent with the idea that the nativemolecule exists as a dimer of similar or identical subunits. Dimer formation,which probably occurs within the cells, does not seem to lead to a measurablecooperative effect between the subunits.

Little is known of the chemical structure of interferon proteins. Chick andmouse interferons have been reported to have molecular weights of about25,000-35,000 and isoelectric points near neutrality.' 2 Human interferon isreported to have a molecular weight of about 25,000 and an isoelectric point alsonear neutrality.3 There has been little evidence regarding the possible subunitstructure of any of these molecules. Chick interferon is apparently inactivatedby the SH-containing compound, j3-mercaptoethanol, and the process is en-hanced by simultaneous exposure to 8 M urea.' Sensitivity to -SH reagentshas been explained by either a reduction of disulfide bonds, or a dependence onmetal ions for function. However, interferon inactivated by f3-mercaptoethanolhas not been reactivated by either reoxidation or the addition of metal ions.1As an initial step in the further characterization of interferons, we have ex-

tensively purified mouse and human interferons by means of the new techniqueof electrofocusing in polyacrylamide gels. We observed that both interferons,although initially homogeneous on gel chromatography, display two componentson electrophoresis in pH gradients. Molecular weight measurements of the twoforms are consistent with the hypothesis that native interferon is a dimer ofsimilar or identical subunits.

620

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

VOL. 67, 1970 SUBUNITS OF INTERFERON 621

Materials and Methods. All of the experiments reported here were carried outwith virus-induced interferon from mouse or human cells.4 Minimal Eagle's medium,MEM (less amino acids) contained 0.15 M NaCl, 0.01 M sodium phosphate (or Trishydrochloride, as specified) pH 7.4, and 0.001 M MgCl2. Mouse L cells and human fetalfibroblasts were grown in monolayers under standard conditions, described more fullyelsewhere.4 The growth and assay of vesicular stomatitis virus (VSV) and Newcastledisease virus (NDV) are described elsewhere.4 Interferon was measured' in the cen-trifuged culture fluid 48 hr after infection with NDV (multiplicity of infection usually 10).During purification, aliquots of electrophoretic or chromatographic fractions were

stored at -70'C to permit repeated assays.Purification of mouse and human fetal interferons: These steps were used in the

purification of interferon proteins from tissue-culture supernatants: acidification topH 2 for 48 hr followed by centrifugation (8000 X g for 20 min at 40C), gel chro-matography, isoelectric focusing, and a final gel chromatography step. The proteinsfulfilled the criteria generally applied to interferon.6 Mouse interferon after Step 1 wasstable at 40C for over 14 months, while human interferon was stable at 4VC for at least2 months. Both proteins were relatively unstable after further purification.Gel chromatography: A column of Sephadex, G-200 superfine, 20 X 1.6 cm, was

exhaustively washed with 0.05 M NaCl-0.01 M Tris pH 7.4 at 40C. The voidvolume was determined using dextran blue, and the column was calibrated withknown molecular weight markers.7 Protein samples (0.35-0.40 ml) were eluted at ahydrostatic pressure of 11 cm of water. Isoelectric focusing8 was carried out within 48 hr.Mouse and human interferons, purified by gel filtration, rechromatographed identically

in 0.05 M NaCl-0.01 M Tris pH 7.4. Purified fractions obtained by electrofocusing weremixed with MEM with and without bovine serum albumin (BSA, 100lg/ml). Dif-ferent molecular forms chromatographed identically in both solutions. Any form of theprotein is relatively stable in the standard chromatographic buffer in respect to activityand molecular size.

Isoelectric focusing: For isoelectric focusing, carrier ampholytes (LKB, Stock-holm) were used to create pH gradients in thin layers of acrylamide gel.8 Gels wereprepared by the photopolymerization (riboflavin 0.273 mg/ml) of 5% (w/v) acrylamidecontaining 2% (w/v) ampholytes.9 Samples were dialyzed versus at least 200 volumesof 0.01 M Tris pH 7.4 for 4-6 hr at 40C before application. 20-50 ,ug could be appliedwithout influencing band width or contour. Above 50 ,g, the pH gradient deviatedslightly from linearity, and the bands became more curvilinear.

After electrophoresis at 400 V for 18 hr, slices from the edges of the gels were placedin 1.0 ml of water and the pH determined. From the inner portion, 7-mm segments werecut for assay and MEM containing BSA (100 lg/ml) was added to each fraction. Thehemoglobin standards eluted from the slices into the medium within 4 hr, and 24 hr wasallowed for elution of the interferon proteins. Duplicate gels for protein determinationswere precipitated and exhaustively washed with 5% TCA.Chemical determinations: When possible, protein was measured by the method

of Lowry et al.'0 using bovine serum albumin as a standard. To estimate protein contentin isolated gel slices, in which protein content was as low as 0.2 ,ug/ml, peptide bond ab-sorbance at 194 nm was determined using a Beckman DBG spectrophotometer with adeuterium lamp." Samples were exhaustively dialyzed against deionized water beforespectrophotometry since many inorganic salts absorb in the 190-195 nm range.

Results. Mouse interferon was consistently recovered in yields of 60-100%,and human fetal interferon was usually 50-75% recoverable, with a range of12-80%.

Isoelectric focusing gave the major purification. It is difficult to measurespecific activity accurately because of the low protein content, but a lower limitof specific activity could be determined. By these estimates, the cumulativepurification of mouse interferon was 593-fold to yield a specific activity (either

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

622 MICROBIOLOGY: W. A. CARTER Puoc. N. A. S.

form) of 4.5 million units/mg protein. The cumulative purification of humaninterferon was 1500-fold (Form A) yielding a specific activity of 2.1 million units/mg protein, and 280-fold (Form B) with a specific activity of 386,000 units/mgprotein. 12Gel chromatography of unpurified interferons: Control observations indicated

that interferon activity and molecular size were stable in 0.01 M\I Tris pH 7.4-0.05 lv NaCl. llouse interferon behaved as a single molecular species with amolecular weight of 38,000 (Fig. 3, top panel). Viral-induced human interferon-was eluted predominantly in. a position corresponding to a molecular weight of24,000 (Fig. 4, top panel). Both interferons thus exist largely as single molecularspecies at this stage of purification.

Isoelectric focusing of mouse and human interferon: Isoelectric focusing wasperformed on the interferons obtained from gel chromatography (Fig. 1). Mouse

pl 7 35 7 IS

HEMOGLOBIN WD _ 12

8.0 - 10

7.0 8;\N

I~~~~~~~~~~~~~~~~~~ox6.0-6 E

5.0-0

4.0-2

3.00 10 20 30 40

FRACTION NUMBER

FIG. 1. Isoelectric focusing of mouse interferon. Mouse interferoi (sp act 100,000 units/mg)was focused on pH 3-10 ampholyte gradients. After electrophoresis, 7-mm gel segments werecut out for assay. 2400 units (24 ug protein) were applied and all the activity was recoveredin two equally active fractions. 0, pH gradient; X, interferon.

interferon activity was found in two peaks (A and B) containing equal activity,with isoelectric points of 7.15 and 7.35 respectively. Essentially 100% of theinput activity was recoverable and the specific activity of the two peaks wasthe same. In other experiments, the relative proportions of Forms A and Bchanged. With increasing concentration of protein, peak A contained moreactivity than peak B (Fig. 5, bottom panel). The observed isoelectric pointswere constant over the range of protein used.Human interferon focused with more acidic isoelectric points (Fig. 2). Three

peaks were recovered with isoelectric points of 5.35, 5.60, and 5.70. Thesefractions, unlike those obtained with mouse interferon, contained differentpercentages of the total activity. The major fractions, Forms A (pI 5.35) and

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

VOL. 67, 1970 SUBUNITS OF INTERFERON 623

9.0- A60'Dqt

8.0 - p1 6 50q pl 5.75 5.605.35Xx

7.0 40-,

56.0 30z0La_

5.0 2020

I--z

4.0- 10

3.0Clo 20 30 40

FRACTION NUMBER

FIG. 2. Isoelectric focusing of human fetal interferon. Human fetal interferon (sp act20,000 units/mg protein) was focused on ampholyte gradients as for Fig. 1. 800 units (40 pg)was applied; recovery, 65%. In duplicate gel slices, protein content was estimated by peptidebond absorbance. Human fetal interferon was usually 50-75% recoverable in the focusingstep, but yields as low as 12% were obtained. 0, pH gradient; 0, human interferon.

B (pI 5.60), contained 99% of the activity. A minor fraction (Form C) (pI5.70) with 1% of the total activity was not further characterized. The relativeactivity existing as Form A or B varied in different electrophoretic runs, butFormA usually contained 50-85% of the total activity.Gel chromatography after electrofocusing: Molecular weight measurements

were made on interferon peaks obtained from isoelectric focusing. MouseForm B yielded 86% of the original 38,000 mol wt species and 14% of a pre-viously undetected 19,000 mol wt species (Fig. 3, bottom panel). Form A yielded90% of the 19,000 mol wt species and 10% of the original 38,000 mol wt form.Under similar conditions, crude mouse interferon was an exclusively 38,000 molwt species. Thus, electrofocusing has separated two forms of the molecule.

Similar experiments were performed with human interferon Form A whichyielded 90% of a 12,000 mol wt species, previously detected in trace amounts,and 10% of the original 24,000 mol wt form (Fig. 4, bottom panel). It was notpossible to study human Form B accurately by the techniques previously appliedto the two forms of mouse interferon because of the small amount of activity.We obtained indirect evidence that Form B (pI 5.60) is the undissociated 24,000species by devising conditions in which the native molecule (mol wt 24,000)could be electrophoresed without dissociation. Under these conditions onlyone species (pI 5.60) is formed, which has a molecular weight of 24,000.Both mouse and human interferons are separable into two active forms, one

of which is twice the mol wt of the other. The simplest explanation is that thelarger mol wt form is a dimer and the lower mol wt form the monomeric unit.While it is possible that the monomeric units are different, it seems more likelythat the dimers are made up of identical subunits. Only one monomeric species

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

624 MICROBIOLOGY: W. A. CARTER PROC. N. A. S.

MOLECULAR WEIGHT 38,000 19,000 0

-120 X

° 80

60 Z0

-40 ":.20 w

2 80 alk. Phosx 303

bovine serum albumin X0~~~~~~~~~~~~~~~~~~60 N,

feron(pat600nt/grtiwa rmtgpeo ehd-0a,420 C

Btp G40 c chymotrypsin

Form A pC.15) an zp .5,ioae yeetohrssi h xeietdsrbdi

0 ''w2 20- ceuoplosmin A

Fig.1,weehrmaogapedonSehaexG-00 Btponc. RNase h

30 35 40 45 50 55V0 FRACTION NUMBER

FIG. 3. Top panel: Gel chromatography of mouse interferon. Unpurified mouse inter-feron (spe act.16,000 units/migprotein) was chromatographedson Sephadex G-200 at 40C.Eluent, 0.05 M NaCi-0.01 M Tris pH 7.4; flow rate, 2.5 mi/hr. 0.5 ml fractions were col-lected and assayed in mouse L cells by the colorimetric method. About 95% of the 2800 unitsapplied was recovered.Bottom panel: Gel chromatography of peaks of mouse interferon isolated b electrosfocusing.

Forms A p1 7.15) and B (pl 7.35), isolated by electrophoresis in the experiment described inFig. 1, were chromatogaphedOd Sephadex G-200. Both forms had been previously elutedfrom gel slices into MEM containing BSA, 100tg/mi. Theaiamples were applied directly,without freezing, to the column. Both forms, as isolated from gel slices, were equally activewith specific activities of 4.5 million units/mg protein. (These activities showed a gradualdecay over 10r-14 days. By the end of two weeks at rC,activity in both forms had fallen10- to 15-fold.) 300 units of each form were chromatographedwith recoveries of 85% (FormB) and 55% (Form A). 0, molecular weight standards and interferon (Form A); 0, interferon(Form B).

is seen, and it has exactly one half the molecular weight of the larger species.The dimer to- monomewr interconversion may be symbolically written as I2 ~~21.

Interconversions of Forms B and A: Mouse interferon Form B was isolatedby isoelectric focusing, dialyzed versus 0.01 M Tris pH 7.4, and refocused at itsisoelectric point. About a quarter of the activity was now present as Form A(Fig. 5). Form B can thus continue to dissociate into A, the presumptive mono-meric species. It Was ndt technically possible to carry out similar studies withhuman interferon Form B because of the much lower activity appearing in thispeak.

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

VOL. 67, 1970 SUBUNITS OF INTERFERON 625

MOLECULAR WEIGHT 24,000 12,000L~~~~~~~ 4

40

30

20

010_

l00~~~~~~~~~~~z0

so80 - flk. phos. 40 a

I bovine serum albumin

I

60 3 30

FI.4 o anel20

hoaorpyo ua ealitreo Uprfe ua

40 2

w chyrmotrypsin-LJi 20 ceruloplasmin 10

I ~~~~~~~~~~~~~~ponc..RNose

0 0 - 1030 35 40 45 50 5V0 FRACTION NUMBER

FIG. 4. Top panel: Gel chromatography of human fetal interferon. Unpurified humanfetal interferon (sp act 2000 units/mg protein) was chromatographed under conditions de-scribed in Fig. 3. About 85% of the 500 units applied was recovered. Interferon was assayedin primary cells of human foreskin.

Bottom panel: Gel chromatography of human fetal interferon (Form A) isolated by electro-focusing. Form A (pI 5.35), isolated by electrophoresis in the experiment described in Fig. 2,was chromatographed on Sephadex G-200. 400 units of Form A (sp act 2 X 106 units/mgprotein) was applied and all of the activity was recovered. 0, molecular weight standards;0, purified human interferon Form A.

Something associated with the focusing procedure causes an apparent dis-sociation of interferon. Could it be the low ionic strength which prevailsthroughout the electrofocusing procedure? The idea of a sait dependency forstability of the dimeric form was supported by attempts to focus in higher saltconcentrations (0.05 M NaCl-0.01 M Tris or phosphate, pH 7.4). When focus-ing was carried out under these conditions, no evidence for dissociation wasobtained. Mouse interferon yielded only Form B (pI 7.35) and human interferonyielded Form B (pI 5.60) as the principal species. In other similar experiments,human Form A has been present from trace amounts to 15% of the recoverableactivity.These observations suggest that low salt concentrations prevailing during

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

626 MICROBIQLOGY: W. A. CARTER PROC. N. A. S.

9.0 FORM B A ISOELECTRIC FORM B REFOCUSED 15

_12!

8.0 CO0

6 XE

7.03go x-x-x-x-x.X-X-X-X- B 0 oo60 R < CONC. OF -X- IS 0.75 CONC.-6- 18 a

6. 10 20304

aIs z0

XC505 12il

CC

1pI7.3)9 Wz

3

3.0 -a---I00 10 20 30 40

FRACTION NUMBER

FIG. 5. Refocusing of isolated mouse interferon Form B. Tcp panel: Isolated Form B(pI 7.35) was dialyzed versus a low salt concentration (0.01 M Tris pH 7.4) and refocused atits isoelectric point. 1200 units applied; recovery 90%. 0, pH gradient; X, activity.

Bottom panel: Two other gels were simultaneously electrophoresed with different amountsof mouse interferon purified by gel chromatography, to detect possible effects of concentrationon the observed pI or the percentage activity in either isoelectric peak. 25 pg of protein (2400units) and 35 jug of protein (3360 units) were applied; recoveries were 90 and 95%, respectively.The results suggest that over this relatively narrow range the isoelectric points are unchanged,although the relative amount of activity in Form A increases with more total protein present.X, 25 pg applied; A, 35 pg applied.

isoelectric focusing allow the detection of subunits. If this were so, dissociationshould be demonstrable by gel chromatography. Mouse interferon Form Bwas dialyzed versus low salt (0.01 Tris pH 7.4) for 16 hr and then chromato-graphed (Fig. 6, top panel). The dialysis step alone was sufficient to dissociatethe molecule as evidenced by the displacement of interferon to a 19,000 mol wtpeak. The 38,000 mol wt region contained little activity. There is no concomi-tant loss of activity, which indicates that the smaller molecular form must beequally active. It provides additional support for subunits of similar structure.The reverse reaction, the conversion of mouse Form A to Form B was studied

by chromatography of Form A which had been dialyzed against MEM (Fig.6, bottom panel). This buffer was selected since interferon is present as Form Bin tissue culture fluid. About 5% of the activity was transformed into FormB, which indicates that the monomeric form is not quantitatively transformedquickly into the oligomer by simply raising the salt concentration. This in-efficient transformation of A into B may be due to the low concentration of themonomeric species. 13Mixture experiments performed with the A, B, and C forms of human inter-

feron revealed that the activities were additive. There does not seem to be ameasurable cooperative effect between the subunits in the dimer form. Similarstudies in progress with polymer-induced human interferon give analogous re-sults. 14

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

VOL. 67, 1970 SUBUNITS OF INTERFERON 627

FIG. 6. Gel chromatog- 30raphy of mouse interferon MOLECULAR WEIGHT 38,000 19,0Form B dialyzed versus low A

salt concentration. Top pan- A _20 T0

el: Mouse interferon was lXdialyzed versus high (0.05 M T|NaCl-0.01 M Tris pH 7.4 Al0.001 M MgCl2) or low (0.01 aX IL oM TrispH 7.4) salt for 16hr la

before chromatography on 80 _t alk ph*$.

Sephadex G-200 (equilibrated \0-Uwith 0.05 M NaCl-0.01 M xbovine Swum albu8n1in

Tris pH 7.4). In each in- i 60 \0stance, crude interferon (8p ZZact 16,000 units/mg protein) 340 \hymotr~n. 20was used; over 95% of the \2800 units applied were recov- /ered. No detectable loss of U 20 10activity occurred on dialysis 0 c Rose.versus low salt. 0, dialyzedversus high salt; A, dialyzed 30 35 40 45 50 55versus low salt concentration. VO FRACTION NUMBER

Bottom panel: To studythe proposed reverse reaction,Form A B, Form A was isolated in the experiment described in the top panel and dialyzedversus higher salt (0.05 M NaCl-0.01 M phosphate pH 7.4-0.001 M MgC12) with and with-out fetal calf serum (6% v/v) for 20 hr. The proteins were rechromatographed on SephadexG-200. The results were unchanged by the presence of the calf serum. 0, molecular weightstandards; 0, mouse interferon Form A after dialysis.

Discussion. Mouse and human interferons, viral-induced, initially behavedas single molecular species. At one step during purification (isoelectric focusing),two forms of each protein were detected. These two forms were then charac-terized as to molecular weight. Form A (mouse) has a molecular weight of19,000 and Form A (human) a molecular weight of 12,000. Form B (mouse)has a molecular weight of 38,000 and Form B (human) a molecular weight of24,000. These data strongly suggest that interferon exists as a dimer of twosimilar, perhaps identical, subunits. In the case of mouse interferon, the mono-meric unit mass is 19,000 daltons and with human interferon 12,000 daltons.The different isoelectric points for the dimer and monomer suggest that certainsurface charges must rearrange during oligomer formation, but without influenc-ing activity, since there does not seem to be a measurable cooperative effect be-tween the subunits in the dimer form.The conversion of Form B into A is promoted by the exposure to low salt.

In the presence of 0.05 M NaCG, little evidence for two molecular forms exists.The mechanism of the conversion, which is currently under study, could relate toeither an environment of low ionic strength or the removal of a necessary metalion by dialysis. The latter possibility seems more likely since high ionic strengthgenerally promotes dissociation because of the electrostatic forces provided byhigh salt.

Recently, several highly purified glycolytic enzymes, homogeneous by gel elec-trophoresis, have been found on isoelectric focusingto contain a number of molecu-lar species.'5 The authors suggested that these complex profiles may be related to

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0

628 MICROBIOLOGY: W. A. CARTER PROC. N. A. S.

combinations of closely related subunits which dissociate and rejoin randomlyduring electrofocusing. Our current studies with molecular weight measure-ments indicate that isoelectric focusing does in fact separate different oligomericforms, and that the prevalence of low salt concentrations during electrofocusingmay promote the conversions between different molecular species.Recent preliminary reports suggest many isoelectric components of interferons

from different animal tissues'6 (Stancek, Gressnerova, and Paucker, 1970; per-sonal communication). These studies have been carried out largely by electro-focusing in sucrose gradients. It is not clear how to reconcile these results withthe findings reported in this paper since these results appear more complex.Further study should serve to clarify the important relationships between sub-unit structure and the biological properties of interferon.

The author thanks Drs. Hamilton 0. Smith, Bernard Weiss, and Thomas J. Kelly, Jr., fortheir helpful suggestions during the course of this work.

Abbreviations: BSA, bovine serum albumin; MEM, minimal Eagle's medium; TCA, tri-chloroacetic acid; pl, isoelectric point.

* This work was supported by USPHS Research Career Development Award A142565 andresearch grant CA06973; and grants from the Council for Tobacco Research, USA, no. 694,and American Cancer Society, Maryland Division, no. 69-01. A preliminary report of thiswork was presented at the meeting of the Federation of American Societies for ExperimentalBiology, April 15, 1970, and appeared in Fed. Proc. 29, 635 (1970).

1 Merigan, T. C., C. A. Winget, and C. B. Dixon, J. Mol. Biol., 13, 679 (1965).2 Lampson, G. P., A. A. Tytell, M. M. Nemes, and M. R. Hilleman, Proc. Soc. Exp. Biol.

Med., 112, 468 (1963).3 Merigan, T. C., D. F. Gregory, and J. K. Petralli, Virology, 29, 512 (1966).4Carter, W. A., K. R. Hande, B. Essien, and M. M. Kabach, in preparation.Finter, N. B., J. Gen. Virol., 5, 419 (1969).

6 Lockhart, R. Z., Jr., in Interferons, ed. by N. B. Finter (Amsterdam: North HollandPubl. Co., 1966), pp. 1-20.

7Determann, H., in Gel Chromatography (New York: Springer-Verlag, 1968).8 Awdeh, Z. L., A. R. Williamson, and B. A. Askonas, Nature, 219, 66 (1968).9 The reaction was carried out between two glass plates held apart by 1 mm thick glass

slides. A rotary motion of the glass slide lifted one plate off the gel which was then transferredto an electrophoresis apparatus at 40C with an atmosphere previously equilibrated with water.Several hours at room temperature may cause the outer boundary of the gel to dry out, lead-ing to a curvilinear appearance of the protein bands.

10 Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall, J. Biol. Chem., 193, 263(1951).

11 Mayer, M. M., and J. A. Miller, Anal. Biochem. 36, 91 (1970).12 There are several possible explanations for the nearly complete recovery and the ap-

parently lower specific activity of one band as compared to the other. The most likely oneis the presence of small amounts of inorganic salts which absorb at 194 nm. There was nomeasurable loss of protein during TCA precipitation and washing.

's If Form B consists of two nonidentical subunits, complete reassociation of the two sub-units into the dimer form can only occur with equal quantities of the two, whereas if the sub-units are identical, any admixture should be convertible to the dimeric form. We cannot testthese alternatives at present because of the slow conversion of Form A into B.

'14Carter, W. A., and P. M. Pitha, in Biological Effects of Polynucleotides, eds. R. F. Beers,Jr. and W. Braun (New York: Springer-Verlag), in press.

16 Susor, W. A., M. Kochman, and W. J. Rutter, Science, 165, 1260 (1969).16 Fantes, K. H., Science, 163, 1198 (1969).

Dow

nloa

ded

by g

uest

on

Aug

ust 3

, 202

0