Proc. Natl. Acad. Sci. USAVol. 76, No. 9, pp. 4335-4339, September 1979Biochemistry

Trace polypeptides in cellular extracts and human body fluidsdetected by two-dimensional electrophoresis and ahighly sensitive silver stain

(genetic screening/fetal monitoring/mutation detection/human body fluids)

CARL R. MERRIL*, ROBERT C. SWITZERt, AND MARGARET L. VAN KEURENt*Laboratory of General and Comparative Biochemistry; tLaboratory of Brain Evolution and Behavior; and fLaboratory of Clinical Sciences, National Institute ofMental Health, Bethesda, Maryland 20205

Communicated by James V. Neel, June 11, 1979

ABSTRACT Development of a highly sensitive silver stainpermits the characterization of trace cellular and body fluidproteins separated by the two-dimensional electrophoresistechnique of O'Farrell. Many of the proteins detected by thesilver stain in urine, spinal fluid, amniotic fluid, and cells wereundetected with the widely used Coomassie blue stain. Tracepolypeptides observed in Escherichia coli cell lysates with thissilver stain could be detected previously only by growth in ra-dioactive precursors followed by lengthy autoradiography. Insituations that do not permit the use of radioactive labeling, asin human clinical studies, the enhanced ability to detect proteinsachieved by the silver stain will facilitate metabolic studies andthe screening for protein abnormalities in mutational studiesand in genetic diseases.

Detection and characterization of proteins is of fundamentalimportance to many areas of biology and clinical medicine. Insome endeavors, such as genetic screening for mutational eventsand genetic diseases, the efficiency of the search is directlyproportional to the number of proteins that can be detected andcharacterized in cellular extracts or body fluids. Polypeptidesthat are present in trace amounts are often of great importance.Separation techniques, such as two-dimensional electrophoresis(1), are capable of resolving thousands of proteins in complexmixtures. However, the most commonly used nonradioactivedetection method, Coomassie blue staining, lacks the sensitivityto detect many proteins that are present in low concentrations.Unfortunately, concentration of samples is often counterpro-ductive because more abundant proteins may cause overloadingand distortion of the gel patterns. When autoradiography is usedwith radioactively labeled proteins, less abundant or traceproteins (10-4-10-5% of the total) can be detected (1). How-ever, whole animals cannot generally be studied by such ra-dioactive methods because prohibitively large amounts of ex-pensive radioactive precursors would be required. Also, ra-diation hazards would make this type of radioactive labelingin human clinical studies unacceptable.

This inability to detect rare proteins in human and otheranimal studies has limited the application of two-dimensionalelectrophoresis, particularly in endocrinology, mammalianmetabolism, developmental biology, and immunology.We have developed a highly sensitive silver staining method

that is capable of detecting rare proteins. This silver stain is atleast 100 times more sensitive than Coomassie blue stain (2) andis much less expensive and more rapid than autoradiography.When this stain is used on samples of complex biological mix-tures separated by two-dimensional polyacrylamide electro-phoresis, it is specific for polypeptides.

The application of this silver stain to two-dimensional elec-trophoretograms of biological and clinical interest is demon-strated in this report. A number of polypeptides were observedin unconcentrated spinal fluid, amniotic fluid and cells, urine,and plasma that could not be visualized in these unconcentratedsamples by Coomassie blue staining. The silver stain used incombination with autoradiography of radioactive proteins frompulse-labeled cells in culture may be helpful in screening forproteins that are undergoing altered turnover rates.

MATERIAL AND METHODSSamples. Escherichia coli [strain, SA500 (X imm21), a gift

from Max Gottesman] was grown in a minimal medium (M56)supplemented with 0.5 ml of 1% histidine, 0.47 ml of 1% valine,0.47 ml of 1% isoleucine, 0.05 ml of 1% biotin, and 0.5 ml of 2%glucose per 100 ml of medium. Cells grown to midlogarithmicphase (at 320C) were transferred to a 420C water bath for 15min. Mixed '4C-labeled amino acids, 25 ,uCi (specific activitiesof amino acids ranged from 175 to 516 Ci/mol, NEN-445, NewEngland Nuclear; 1 Ci = 3.7 X 1010 becquerels), were addedto 2 ml of bacteria and growth was continued at 42'C for 5 min.Growth was terminated by rapidly chilling the samples in anice bath followed by rapid addition of 200 ,ul of 10% (wt/vol)Casamino Acids and 20 ,ul of 2 M sodium azide. The bacteriawere centrifuged at 12,000 X g for 2 min and the pellet wasrinsed with phosphate-buffered saline. The bacterial pellet wasstored at -70°C. Specific radioactivity of the protein in thepellet was determined by measuring total protein by the Lowrymethod (3), whereas protein radioactivity was determined byprecipitation of an aliquot in 10% trichloroacetic acid. Theprecipitate was collected on a Millipore filter, dried, and placedin a scintillation cocktail for determination of radioactivity.

Plasma. Plasma was obtained by collecting blood in sterileheparinized tubes followed by centrifugation at 8000 X g for10 min at room temperature to remove cells. Urine was col-lected and centrifuged at 8000 X g for 10 min to remove anycells. Amniotic fluid was centrifuged at 8000 X g for 10 minto separate amniotic cells from the fluid. All samples, includingspinal fluid were stored at -70°C.Treatment of Samples. Samples were denatured in sodium

dodecyl sulfate (NaDodSO4) and mercaptoethanol (4). The E.coli cell pellet and the amniotic fluid cell pellet were denaturedby addition of 100,l of 2% NaDodSO4, 5% mercaptoethanol,20% glycerol, 2% 3/10 Biolyte (Bio-Rad), and 2% Nonidet P-40(BRL, Bethesda, MD) in a microfuge tube, followed by heatingat 95°C for 5 min in a heating block. Spinal fluid, amnioticfluid, urine, and plasma were denatured in a similar mannerby mixing with an equal volume of the NaDodSO4/mercap-

Abbreviation: NaDodSO4, sodium dodecyl sulfate.4335

The publication costs of this article were defrayed in part by pagecharge payment. This article must therefore be hereby marked "ad-vertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

Dow

nloa

ded

by g

uest

on

July

31,

202

1

Proc. Natl. Acad. Sci. USA 76 (1979)

toethanol solution followed by a 5-min heat treatment at 950C.Samples were rapidly cooled in an ice bath after the heattreatment and then brought to room temperature prior to ad-dition of urea (17 mg of urea per 30 A.l of prepared samples).Sample proteins were determined by the method of Lowry etal. (3).

Protease K Digestion. Protease K (Merck, Darmstadt, WestGermany), 50,g, was added to a 30-Al aliquot of each dena-tured sample in the NaDodSO4/mercaptoethanol solution.Digestion was performed at room temperature for 4 hr.Two-Dimensional Electrophoresis. Electrophoresis was

carried out according to O'Farrell (1), with 3/10 Biolyte (Bio-Rad) in the first dimension and a 10% acrylamide uniform gelin the second dimension. Isoelectric focusing was at 500 V for20 hr; slab gels were run at 20 mA/gel.

Staining. Coomassie brilliant blue R250 was used (0.1% in50% methanol/12% acetic acid) to stain the gel overnight withgentle shaking. Gels were destained with several changes of a10%/ethanol/5% acetic acid solution. Photographs were takenthrough a 5500-A narrow band interference filter (BairdAtomic, Bedford, MA).To stain the gels with silver (2), we first fixed them in 50%

methanol/12% acetic acid for 30 min (gels can be stored over-night in this solution). The gels tend to shrink in the 50%methanol solution. To expand them prior to staining we thenplaced them in 10% ethanol/5% acetic acid for 2 hr, followedby three washes with 10% ethanol (5 min each). The gels werethen soaked in 4% (wt/vol) paraformaldehyde/1.43% (wt/vol)sodium cacodylate (adjusted to pH 7.3 with HCI) for 30 min,followed by three 5-min washes with 10% ethanol. The gelswere then agitated gently for 30 min in a cupric nitrate/silvernitrate solution (made by dissolving 3.5 g of silver nitrate in 100ml of water followed by addition of 1.5 ml of 0.5% cupric ni-trate solution and then the simultaneous addition of 4 ml ofpyridine and 8 ml of ethanol). We have found it necessary touse reagent-grade absolute ethanol in this step. Next, the gelswere placed in fresh diammine solution (made within 5 minof use), prepared by mixing together 30 ml of a 19.4% (wt/vol)silver nitrate solution and 22.2 ml of a solution of sodium andammonium hydroxide [stock solution contains 100 ml of 0.36%NaOH, 45 ml of fresh concentrated NH4OH, and 55 ml of 20%(vol/vol) ethanol]. Diammine solution remaining after theprocedure must be discarded because an explosive complex mayform upon storage. After the diammine rinse, the gels werewashed for 1 min in a reducing solution containing 2.5 ml of10% formaldehyde (10 ml of commercial formaldehyde solu-tion in 100 ml of water), 6 ml of 1% citric acid, and 100 ml ofethanol in 1 liter of water. This wash was repeated, again for1 min, followed by several rinses in a second reducer, containing5 ml of 10% formaldehyde, 5 ml of 1% citric acid, and 100 mlof ethanol in 1 liter of water. The proteins may begin to appearas brown or black spots at any time in the reducing solutions.Staining can be stopped by washing the gel in successivechanges of deionized water. Image formation in the diamminestep may occur if reagent-grade absolute ethanol and fresh stocksolutions are not used. Surface contamination can be minimizedby washing the glass slab plates thoroughly, immediately aftereach electrophoresis run, and using well-washed surgical gloveswhen handling the gels. The gels are fragile after staining andshould be photographed for a permanent record.

Gels that are overdeveloped may be lightened with a pho-tographic reducer such as the copper reducer of Smith (5). Afresh mixture of equal volumes of two stock solutions ("A" and"B") are required. Solution A contains 37 g of cupric sulfate and37 g of NaCl dissolved in 1 liter of water with the addition ofsufficient concentrated ammonium hydroxide to dissolve any

precipitate. A deep blue solution should be formed. SolutionB contains 458 g of sodium thiosulfate in 1 liter of water. Usuallya 3:1 dilution of water to fresh reducer is used to lighten gels.The reduction is stopped by washing the gel in water.

Stained gels may be kept in water. Gels that are to be driedfor storage or autoradiography should be first soaked in 30%(wt/vol) sodium thiosulfate for I5 min followed by four 15-minwater rinses. The gels should then be soaked for 5 min in apreserving solution [methanol/H20/glycerol, 70:27:3 (vol/vol)](6), followed by drying on 3 MM filter paper (Whatman) at500C under reduced pressure for 2 hr.

Autoradiography. Gels that were to be autoradiographedwere dried as described above and then placed in x-ray filmcassette holders (Kodak X-omatic, with regular intensifyingscreens). Exposure was at -70'C with Kodak XR-2 x-ray filmfor appropriate timed intervals. Films were developed at 250Cfor 5 min in developer (Kodak KLX developer) followed by a5-min fix (Kodak X-omatic fixer).

RESULTS AND DISCUSSIONSilver Stain for Electrophoretic Gels. In an effort to develop

a more sensitive protein detection method than Coomassie bluestaining, a number of stains were tried, including fluorometricreagents [fluorescamine and the related reagent 2-methoxy-2,4-diphenyl-3-(2H)-furanone, gifts of Hoffmann-La Roche].No major increase over the sensitivity achieved with Coomassieblue staining was observed (unpublished observations). Ex-periments with histological staining methods led to an adaptionof a silver staining procedure described by de Olmos (7, 8), foruse in visualizing degenerating structures in the central nervoussystem. Although the basic mechanism of the stain is unknown,the presence of copper sulfate is reminiscent of the biuret re-action. Specificity of the silver stain for polypeptides wasdemonstrated by incubating E. coli cell lysates, body fluids,amniotic fluid, and amniotic fluid cell preparations with pro-tease K. This treatment resulted in the disappearance of alldiscrete spots on two-dimensional gels prepared with thesesamples that hadbeen treated with protease K and stained withthe silver method. Linear proportionality between the densityof silver stain and concentration of protein has been observedwith purified phosphorylase B and carbonic anhydrase (2).The staining procedures as described in this study depends

totally on manual manipulations and they are subject to thefollowing four major variables: reagent quality, concentration,time in each reagent, and the temperature of the reagents. Thefinal intensity of the stain is also under the investigator's controlby the decision as to when to stop the reduction reaction. Giventhe above variables, identical gels stained at the same time givesimilar images. However, comparisons of gels stained at dif-ferent times require some internal standardization, particularlyif any quantitation is to be attempted. With nonradioactiveclinical samples, it is possible to add a known quantity of ahighly purified [14C]methylated protein to the sample prior toelectrophoresis. Densitometry of the spot representing thisprotein after staining followed by radioactive quantitationserves both as a control for the staining procedure and as ameans of monitoring protein loss during the electrophoreticprocedures.

Cells Grown in Culture. A typical E. coli two-dimensionalprotein pattern is shown in Fig. 1. Only nine of the mostabundant polypeptides can be seen with Coomassie blue stainwith a sample of 30 ,ug of protein (Fig. 1A). A survey of E. coliproteins on two-dimensional electrophoresis by O'Farrell (1)demonstrated that over 80% of the proteins observed are presentas minor spots (each of these minor spots contains 0.1% or lessof the protein observed on the gel). Minor or trace polypeptides

4336 Biochemistry: Merril et al.

Dow

nloa

ded

by g

uest

on

July

31,

202

1

Proc. Natl. Acad. Sci. USA 76 (1979) 4337

I-.. ..... ........

A\ B ijC ~~~~~D

i ~ ~ 4

_..

FIG. 1. E. coli polypeptide patterns as visualized by: (A) Coomassie blue stain; (B) autoradiography, 56-hr exposure; (C) autoradiography,168-hr exposure; (D) silver stain. Total protein applied to gel was 30 jig with specific 14C activity of 1.2 ,uCi/mg. The E. coli gels are shown withthe acid isoelectric focusing range on the right. (Sample origin was at the upper left.)

can be visualized with autoradiography or the silver stain.Autoradiography after 1 week of exposure (Fig. 1C) revealed54 polypeptides, whereas 126 could be observed with the silverstain. It is possible, with cells grown in culture, to achievegreater sensitivity with autoradiography than was obtained inthe experiment illustrated in Fig. 1 by labeling for longer pe-riods of time or by the use of a larger amount of radioactiveprotein precursors or of protein precursors of higher specificradioactivity in the growth medium. However, this type ofeffort is limited to some extent by the expense of the radioactiveprecursors and possible alterations in cellular metabolism causedby radiation damage. Some increase in sensitivity could also beobtained by using fluorography or increasing the exposure timeof the autoradiography. Both of these techniques, however,generally result in some loss of resolution. This is due in thelatter case to "autoradiographic spreading" (a 10-fold increasein exposure generally results in a doubling of the spot size) (1).None of the above difficulties is inherent in the silver stainmethod.

Use of both autoradiography and the silver stain on samplesthat have been radioactively pulse labeled as in Fig. 1 may allowobservations of relative changes in individual protein turnoverrates. This possibility is illustrated by the dense spot observedby autoradiography in the upper right quadrant of Fig. 1C. Amuch less dense spot is observed with the silver stain in the sameposition in Fig. 1D. In comparison with its neighbors, the in-creased density of this spot on autoradiography may representa higher specific radioactivity, indicating either a high rate ofturnover or synthesis during the pulse labeling. This spotprobably represents synthesis of a protein caused by the tem-perature shift prior to the pulse labeling. Use of these techniquesto study relative changes in protein synthesis and degradationwill require controls for differential protein staining and carefulquantitation.Human Body Fluid Proteins. Some body fluid protein

patterns are shown in Fig. 2. Most of these proteins could notbe visualized by previously existing techniques without priorconcentration of the fluids, with the exception of the plasmaproteins (4). Unfortunately, concentration of the fluids oftenexacerbates the massive overloading problems of the more

abundant proteins. Even as little as 0.5 jA of plasma presentsan overloading of albumin, causing distortions of the nearby

protein pattern and vertical streaking. However, some over-loading of the more abundant proteins in plasma is probablynecessary to be able to visualize the minor proteins. Despitethese problems, Anderson and Anderson (4) have already begunto use two-dimensional electrophoresis to analyze humanplasma for genetic variants.

Electrophoretic analysis of urine with commonly used stainsrequires 100-fold to 1000-fold concentration to detect the traceamounts of protein that are normally present (2-4 mg/100 ml)(9). One-dimensional electrophoretic methods have proven tobe a useful tool in screening for the Bence-Jones protein foundin multiple myeloma and as an aid in characterizing types ofproteinurias. The use of two-dimensional electrophoresis byAnderson and coworkers (10) on 1000-fold concentrated urine,stained with Coomassie blue, has revealed over 230 polypep-tides, which they have categorized into eight subgroups. Un-concentrated urine stained with Coomassie blue revealed nopolypeptide spots (Fig. 2E), whereas silver staining of the samegel revealed over 24 spots (Fig. 2F). Urine concentrated 20-foldby dialysis against 10% polyethylene glycol (Mr 6000) givestwo-dimensional gel patterns, after silver staining, that arecomparable to the patterns obtained by Coomassie blue stainingof urine concentrated 1000-fold (unpublished observation). Thesilver staining method reduces the efforts required to analyzelarge numbers of urine samples by replacing the need for anexhaustive, 1000-fold concentration with a relatively simple10- to 20-fold concentration method. This should facilitate thedevelopment of electrophoresis of urine as a diagnostic tool.

Cerebrospinal fluid also requires concentration, at least10-fold for protein detection with commonly used stains. Un-concentrated spinal fluid revealed only a single polypeptide onstaining with Coomassie blue (Fig. 2C), whereas the samesamples stained with silver revealed more than 50 polypeptides(Fig. 2D). Horizontally arranged spot series, usually higher onthe acidic side of the gel, are noted in the center of the spinalfluid pattern. These patterns have been noted previously (4, 10);they are often the results of post-translational modifications,such as glycosylation. These patterns occur with cell surfaceproteins and many proteins in body fluids. The additionalresolution of silver-stained two-dimensional cerebrospinal fluidpatterns should extend clinical application of spinal fluidelectrophoresis. Electrophoretic systems have already proven

Biochemistry: Merril et aL

Dow

nloa

ded

by g

uest

on

July

31,

202

1

4338Biochemistry:Merriletal.~~~~Proc.Nati. Acad. Sci. USA 76(1979)

A

...1.3

C E

.......-

FIG. 2. Two-dimensional gels of denatured body fluid proteins: human plasma protein (0.5jul of fresh heparinized plasma), human cere-brospinal fluid proteins (15 MI of spinal fluid), and human urine (10 gl). The gels are shown with the acid isoelectric focusing range on the leftside of each gel. (Sample origin was at the upper right.) Plasma: (A) Coomassie blue stain, (B) silver stain. Cerebrospinal fluid: (C) CoomassieIblue stain, (D) silver stain. Urine: (E) Coomassie blue stain, (F) silver stain.

to be useful in detecting abnormal protein patterns in someneurological disorders (11-15).

Amniotic fluid proteins have been monitored in amnioticfluid by one-dimensional electrophoretic systems, specific en-zyme assays, and immunological techniques (16-18).

A....

i-"

B

Twenty-one protein bands have been reported, with the mostabundant protein identified as albumin (60% of the total pro-tein) (18). Some correlations between specific protein concen-trations and occurrence of birth defects and gestational age hasbeen reported (16, 19, 20). Two-dimensional electrophoretic

C1

DI....... ...- -- ...... ..

4*

FIG. 3. Two-dimensional gels of denatured human amniotic fluid (10 Al) and amniotic cell proteins (20 p~g). The gels are shown with theacid isoelectric focusing range on the left. (Sample origin was at the upper right.) Amniotic fluid: (A) Coomassie blue stain, (B) silver stain. Amnioticfluid cells: (C) Coomassie blue stain, (D) silver stain.

4338 Biochemistry: Merril et al.

Aw

Aw

4

AL wo

#. +". 0.,* , '' V:

.: ....r. ;; .*

Dow

nloa

ded

by g

uest

on

July

31,

202

1

Proc. Natl. Acad. Sci. USA 76 (1979) 4339

gels stained with silver permits the detection of over 60 poly-peptides compared to 3 with Coomassie blue in the uncon-centrated amniotic fluid (Fig. 3 A and B). The amniotic fluidcell pattern is marred by the vertical streaking that has beenobserved on numerous occasions with these samples. Even so,more than 170 polypeptides can be observed (Fig. 3 C and D).This ability to observe these cellular proteins and the numerousamniotic fluid proteins may facilitate studies such as those de-scribed above and may also prove useful in searching for en-vironmental and drug effects on the developing fetus.

CONCLUSIONTrace or less abundant polypeptides can now be characterizedby the use of two-dimensional electrophoresis and silver stain.Previously, less abundant proteins could be observed only withsystems in which high specific radioactive labeling could beaccomplished. This reliance on radioactive labeling and auto-radiography for detection of trace cellular polypeptides es-sentially limited the use of high-resolution two-dimensionalelectrophoresis to the study of cells that could be grown inculture with radioactive polypeptide precursors. The silver stainremoves this requirement and will facilitate investigations ofpolypeptide patterns in cells and extracellular fluids as theyoccur in situ. This will make it possible, for example, to studyamniotic fluids and cellular protein patterns for diagnosticevidence of birth defects or comparison of protein patterns insubpopulations of lymphocytes without the need for stimulationwith a mitogen, as is now required to obtain sufficient levels ofradioactive labeling. The silver stain will also permit theachievement of new levels of protein purification by definingcontaminants at lower levels (at least 100-fold) than was possiblewith Coomassie blue stain. Detection of proteins present in lowconcentrations, as in urine and spinal fluid, will no longer re-quire extensive concentration.The development of this silver stain in conjunction with

high-resolution protein mapping methods will facilitate theevaluation of metabolic events, pathophysiological processes,and detection of mutations. Neel (21, §), in discussions of thefeasibility of monitoring for changing mutation rates throughthe use of the protein markers to be found in blood constituents,has calculated that to detect a 50% increase in the frequencyof mutation would require two successive samples each of ap-proximately 6,000,000 bits of information. If a series of new-

borns were examined for 30 different proteins (products of 60alleles) by one-dimensional electrophoresis, each sample wouldrequire blood specimens from 100,000 children. The successfulapplication to this problem of two-dimensional polyacrylamidegel electrophoresis plus the technical developments describedin this paper could result in an order-of-magnitude reductionin the number of blood specimens required, with a corre-sponding increase in efficiency.

We thank Drs. R. Polinsky, M. Ebert, and M. Geier for assistance inobtaining samples of spinal fluid, blood, urine, and amniotic fluid. Inaddition, we thank Drs. J. V. Neel and A. Das for comments on themanuscript and Igor Prissovsky for excellent technical assistance.

1. O'Farrell, P. H. (1975) J. Biol. Chem. 250, 4007-4021.2. Switzer, R. C., Merril, C. R. & Shifrin, S. (1979) Anal. Biochem.,

in press.3. Lowry, 0. H., Rosebrough, N. G., Farr, A. L. & Randall, R. J.

(1951) J. Biol. Chem. 193,265-275.4. Anderson, L. & Anderson, N. G. (1977) Proc. Natl. Acad. Sci.

USA 74, 5421-5425.5. Walls, E. J. & Jordon, F. I. (1976) in Photographic Facts and

Formulas, revised and enlarged by Carrol, J. S., ed. (Prentice-Hall, Englewood Cliffs, NJ), pp. 179-180.

6. Mayer, J. W. (1976) Anal. Biochem. 76, 369-373.7. de Olmos, J. (1969) Brain Behav. Evol. 2, 213-237.8. de Olmos, J. & Ingram, W. R. (1971) Brain Res. 33,523-529.9. Tidstrom, B. (1963) Scand. J. Clin. Lab. Invest. 15, 167-173.

10. Anderson, N. G., Anderson, N. L. & Tollaksen, S. L. (1979) J. Clin.Chem., in press.

11. Beeley, J. A. (1977) in Biological and Biomedical Applicationsof Isoelectric Focusing, eds. Catsimpoolas, N. & Drysdale, J.,(Plenum, New York), pp. 16-19.

12. Kjellin, K. G. & Siden, A. (1977) Eur. Neurol. 15, 40-50.13. Kjellin, K. G. & Stibler, H. (1976) J. Neurol. Sci. 27, 45-57.14. Kjellin, J. G. & Stibler, H. (1976) J. Neurol. Sci. 25, 65-74.15. Kjellin, K. G., Stibler, H. & Siden, A. (1977) Eur. Neurol. 16,

79-89.16. Ogita, S., Noma, H., Kawamura, T., Shimura, K., Ohnishi, M.,

Ishiko, 0. & Sugawa, T. (1978) Am. J. Obstet. Gynecol. 131,63-68.

17. Mohide, P. T. & Hill, R. F. (1978) Am. J. Obstet. Gynecol. 132,221.

18. Fishbacher, P. H. & Quinlivan, L. G. (1970) Am. J. Obstet. Gy-necol. 108, 1051-1055.

19. Liebman, W. M. & Samloff, I. M. (1978) Biol. Neonate 33,174-176.

20. Butterworth, J., Broadhead, D. M., Sutherland, G. R. & Bain, A.O. (1974) Am J. Obstet. Gynecol. 119, 821.

21. Neel, J. V. (1971) Perspect. Biol. Med 14,522-537.§ Neel, J. V. (1978) Proceedings, XIV International Congress of Ge-netics, Moscow, in press.

Biochemistry: Merril et al.

Dow

nloa

ded

by g

uest

on

July

31,

202

1