PROTEIN LABELLING WIT 3H-NSHP (iV-SUCCINIMIDYL-[2,3- · groups in proteins to form amide bonds (Bolton, 1977).3 ./V-succinimidyl-[2,3-H]-propionate (SH-NSP) is supposed to react in

J. Cell Sci. 43, 319-328 (1980) 219Printed in Great Britain © Company of Biologists Limited igSo

PROTEIN LABELLING WITH 3H-NSP

(iV-SUCCINIMIDYL-[2,3-3H]PROPIONATE)

G. H. MtlLLERFachrichtung Genetik, Universitat des Suarlandes, D-6600 Saarbrilcken

SUMMARY

A tritium compound with low molecular weight and diameter, iV-succinimidyl-[2,3-3H]-propionate (3H-NSP), was used to label histones and nonhistone proteins from calf thymus,and nuclear and total salivary gland proteins from larvae of the midge Chironomus thummi.Labelled proteins were visualized after electrophoresis in 10 % polyacrylamide slab gels byftuorography. The method is much more sensitive than staining procedures, allowing todiscern approx. o-i fig protein.

INTRODUCTION

Protein analysis research of cells and cellular compartments requires reduction ofthe quantity of biological material used, as a consequence of the rise of microtech-niques. This also implies improving conventional (e.g. staining) methods for proteindetection in samples gained by sophisticated micrurgical techniques (Kroeger, 1966;Lezzi & Robert, 1972; Neuhoff, 1973; Grossbach & Kasch, 1977) which by now aimat processing of defined eukaryotic chromosomal segments (Bisseling, Berendes &Lubsen, 1976).

Research has profited from the introduction of ways to label polypeptides to highspecific activities by means of radionuclides of iodine (Bolton & Hunter, 1973;Rudinger & Ruegg, 1973), a method which was also applied to the study of manuallyisolated insect polytene chromosomes and nuclear envelopes (Plagens, 1978).

The use of gamma-emitting radioisotopes of iodine, however, requires specialequipment, laboratory facilities, and high security precautions; it also bears the risk ofgenerating considerable alterations in protein structure and, mainly due to exposure toan oxidizing agent such as chloramine T, in protein function (Bolton, 1977). Of course,coupling of a protein to the molecule described below will to a lesser extent (or evennegligibly) also result in an alteration of protein structure.

Alternatively, there is the possibility of labelling proteins with beta-emitting tritiumcompounds which, due to low energy and short range of beta-particles, are efficientonly after enhancing the sensitivity of tritium detection (Randerath, 1970) by addinga scintillator that converts the energy of 3H-particles to light which eventually producean image on photographic material (originally proposed by Wilson in 1958).

Boyd, Leach & Milligan (1972) and Leach & Boyd (1973) prepared and studiediV-acylsuccinimides as reagents for the selective acylation of reactive sidechains inproteins. In short, they found that at pH 6-7 and with molar excesses of acylating

320 G. H. Midler

agent of 1*35—6'75 most of the lysine sidechains could be acylated whereas tyrosinesreacted only negligibly. The use of iV-propionylsuccinimide (pH 7) at molar excessesof 1-35, 2-7, 6-7 and 13-5 over protein amino groups resulted in 84, 100, 100 and 82%lysine modification with only 5, 9, 14 and 52% tyrosine modification (Leach & Boyd,

1973)-At about the same time the 'Bolton & Hunter' reagent, JV-succinimidyl-3-(4-

hydroxy, 5-P26I]iodophenyl)propionate was introduced which yielded proteinlabellings to high specific activities by reacting with the terminal or sidechain aminogroups in proteins to form amide bonds (Bolton, 1977). ./V-succinimidyl-[2,3-3H]-propionate (SH-NSP) is supposed to react in a similar manner and can be used aroundor at neutral pH (6-8) as compared to the required alkaline conditions for the ' Bolton& Hunter' reagent.

This paper will describe the application of 3H-NSP to label different proteins forsubsequent detection by fluorography.

MATERIALS AND METHODS

Chemicals

jV-succinimidyl-[2,3-'H]propionate, sp. act. 46 Ci/mM ~ i7TBq/mM (batch 8) waspurchased from Amersham-Buchler, Braunschweig, Germany. NSP has a molecular weight of170 (unlabelled) and its structure is:

oO II11 / —I

CH3* — CH*—C — ON

oIt is supplied in toluene solution and storage is recommended at room temperature where thedecomposition is not expected to exceed 1 % per month.

Substances for polyacrylamide gel electrophoresis (10 % acrylamide/o-27 % 6«-acrylamide)were as described by Laemmli (1970) and for fluorography instructions of Bonner & Laskey(1974) ar'd Laskey & Mills (1975) were followed.

Calibration proteins were from Boehringer Mannheim GmbH: combithekK, calibrationproteins size I.

Animals

Adult fourth instar larvae of the midge Chironomus thummi were used.

Chromatin

Calf thymus was prepared and kindly provided by Dr N. U. Bosshard; for details ofpreparations see Bosshard (1979).

Histories

Histones were extracted from calf thymus chromatin; one additional histone (H4) was fromBoehringer, Mannheim, also from calf thymus.

Protein labelling with 3H-NSP 321

Salivary glands

Glands were dissected from adult fourth instar larvae (eL4; staging according to Kroeger,I973)- One to 20 glands were transferred to a piece of parafilm ' M ' (American Can Company,Dixie/Marathon, Greenwhich, Ct.) and brought into conical centrifuge tubes filled with50-100/il denaturing sample buffer (0-0625 M Tris-HCl, pH 6-8; 2% Na-dodecylsulphate(SDS); 10% glycerol; o-ooi % bromophenol blue) (Laemmli, 1970). 2-Mercaptoethanol wasfound unnecessary; it is said to gradually lose some of its potency in solution (LeStourgeon &Beyer, 1977) resulting in smeared bands after electrophoresis. The tubes were sealed withparafilm and heated for 5 min in a boiling waterbath.

Nuclei

Nuclei were isolated from salivary glands according to a method developed by Robert (1975)which, briefly, consists of passing Ringer-detergent treated salivary glands repeatedly throughnarrow capillary pipettes. This finally results in setting free nuclei which were also dissolvedin denaturing buffer.

Chromatin and histones

Lyophilized chromatin or histone samples (mostly 1 mg/ml) were dissolved in denaturingbuffer and diluted to appropriate concentrations.

Labelling

The glass vial containing 'H-NSP (1 mCi/500 fi\) was placed into ice-water to reduceevaporation of the solvent; samples of 3H-NSP solution were put into conical centrifuge glasstubes and stoppered with silicone plugs. Inserting 2 syringe needles, the solvent was evaporatedin a gentle stream of nitrogen. According to manufacturer's instructions evaporation of 1 mltoluene takes 20 min at 20 °C. Denatured protein samples were then transferred to the glasstubes, stirred vigorously and immersed for reaction in a 37 °C waterbath for 30 min afterwhich they were ready for electrophoresis.

Electrophoresis

Buffers and gels (1-5 mm) were prepared according to Laemmli (1970) with a minor modifi-cation concerning TEMED which was used in a final concentration of C25 %. The gels wererun in a Desaga (Heidelberg, Germany) screening electrophoresis separating unit (systemHavanna); running buffer was cooled to 15 °C and gels were run for 60—90 min at 50-75 mA/150—220 V. The gels were stained overnight with 0-025 % Coomassie Brilliant Blue G-250(Serva, Heidelberg) in 50 % methanol/7 % acetic acid and destained in 5 % methanol/7 %acetic acid. They were then ready for densitometry (Kipp & Zonen densitometer DD 2 andflatbed recorded BD 7) and photography.

Fluorography

The instructions of Bonner & Laskey (1974) and Laskey & Mills (1975) were followed: thegels were soaked in DMSO for 2x30 min, then for 3 h in a 22 % solution of PPO (2,5-diphenyl-oxazole) in DMSO (w/v). PPO was precipitated in situ by immersing the gels in deionizedwater (60 min) and dried under vacuum with a slab gel drier GSD-4 (Pharmacia, Sweden).They were then brought in tight contact with a preflashed (Laskey & Mills, 1975) KodakX-Omat (X-ray) film, and stored in light-tight boxes at — 70 °C for the periods indicated.After developing (Kodak D 19), fixing and drying they were ready for densitometry andphotography.

The use of DMSO and PPO can be avoided by (i) a new product from NEN: autoradio-graphic enhancer EN'HANCE™, or by (ii) the application of a water-soluble fluor, sodiumsalicylate (Chamberlain, 1979); the latter compound is inexpensive and the method lesstime-consuming.

322 G. H. Muller

RESULTS

Calibration proteins (ioo/tg/ml each) were labelled with 20/tCi 3H-NSP and afterelectrophoresis yielded fluorograms which could be scanned densitometrically (notshown).

Chromatin. Fifty microlitres of chromatin (1 mg/ml) in denaturing buffer werelabelled with 10, 20, 40 and 60 /iCi 3H-NSP, electrophoresed, stained with Coomassieblue and further processed for fluorography. Fig. 1 shows in A the pattern afterCoomassie blue staining: only histones are to be seen. Histone Hi is represented by2 bands, H3 and H2B are not separated in this 10% gel and comigrate in 1 band,H2A and H4 give 1 band each, B and c are both fluorograms after 48 and 140 hexposure, respectively. It is quite striking that many more bands (mostly nonhistoneproteins, but also 1 additional Hi band) are resolved; this in turn depends on concen-tration of label, length of exposure, and quantity applied to the gel: it seems that inC the second run from the right represents an 'optimal' resolution as far as small andfaint nonhistone protein bands are concerned.

Salivary glands dissolve completely in buffer containing 2% SDS and, whenlabelled with, e.g. 20 /iCi 3H-NSP, show distinct banding patterns (Fig. 2) which alsoexceed the details of conventional Coomassie blue staining. In the beginning usually10 glands were used but I found that one gland will give enough resolution. Grossbach(1977), with a micromodification of the Lowry technique, determined the total proteincontent of salivary glands of the related species C. tentans to be around 15 /zg/gland(mean calculated from Grossbach's data); since the salivary gland of C. thummi issmaller with about 30 cells (40 in C. tentans) I assume a protein content of approx.io/tg/gland. Hence, application of 10/il of 1 gland dissolved in 50 [A buffer equals0-2 /tg of salivary gland protein which is visibly resolved.

Nuclei also give a clear banding pattern after 3H-NSP labelling and fluorography(Fig. 3). Here appears the advantage of the method: due to the rather small amount ofprotein present in nuclei (see below) Coomassie blue staining of the gels reveals onlysome extremely faint bands (the most prominent histones) in cases where as much as1000 nuclei had been used. I found that after labelling the amount of protein equi-valent to 50-100 nuclei gives reasonable resolution after fluorography. Assuming aratio of 1:2 for DNA:protein and a DNA content of 3 x io~9 g/nucleus (Daneholt &Edstrom, 1967; Serfling et al. 1975, have reported 1-25 x io~9 g/nucleus) then50-100 nuclei represent 0-3-0-6 (or 0-125-0-25)/^ protein which is still above thedetection limit of this method. In fact, depending on the 3H-NSP concentrationapplied, less material may be needed (in preparation).

Fig. 1. Calf thymus chromatin separated in 10% polyacrylamide. A, Coomassie bluestain; B, fluorogram, 48 h exposure; c, fluorogram, 140 h exposure. 1-4, chromatinlabelled with io, 20, 40 or 60/fCi 'H-NSP and 10/tl ( ^ 10 fig chromatin) appliedto each slot of the gel. i'-4': Ditto, but 5 /tl applied.

Protein labelling with *H-NSP 323

= H1

' 2'1 2 3 4 V 2

_ H3/2B= ^ - H2A

--H43' 4' ^

m 0

324 G. H. Miiller

Histories. Total histones from calf thymus (i fig in io/il) were labelled with 5, 10or 20/iCi 3H-NSP (molar ratios 3H-NSP: protein 1*15, 2-3 and 4-6, assuming a meanmol. wt of 15 000 for histones and using a 3H-NSP batch with a sp. act. of 66 Ci/mM);0-5, 0-3 and o-i fig were run in the gel. Peak areas of H3/H2B + H2B bands weredetermined after densitometry of the fluorogram (Fig. 4): as little as 0-3 fig totalhistone can be detected after labelling with only 5 fid 8H-NSP (= 2-5 fi\). HistonesHi (I fig/50 fil) and H4 (1 fig/50 fi\ or 10 fig/ 50 fd) were separately labelled with 5,10 or 20/tCi 3H-NSP (molar ratios for Hi : 2-28, 4-57 or 9-13; for H4 (1 fig/50 fi\):

G

Fig. 2. Chironomtis thummi salivary gland proteins labelled with 20 /tCi 'H-NSP(exposure rime 125 days). Numerals denote identifiable bands with approximatemol. wt: 1, 104000; 2, 95000; 3, 88000; 4, 80000; 5, 65000; 6, 58000; 7, 50000;8, 43000; 9, 39000; 10, 34000; 11, 30000; 12, 27000; 13, 25000; 14, 22000;15, 19000; 16, 18000; ly, 12000; 18, 11000. The fluorogram represents proteinsfrom 3 salivary glands.

1-23, 2-45 or 4"O,i), 1-5 fig or 0-1-0-5 fig applied to the gel and the fluorogramsevaluated after various exposure times (2-28 days) for visibility of bands. It could beshown for histone Hi that labelling with 10/iCi 3H-NSP and an exposure time of28 days equals labelling with 20 fid 3H-NSP and an exposure time of 7 days, i.e.doubling the label reduces the necessary exposure time 4-fold. However, the ' optimal'procedure for 0-1-0-5 f-S Hi seems to be an amount of 20/iCi 3H-NSP and 14 daysexposure. Histone H4 (o-i fig) labelled with 10 fid 3H-NSP is not visible even after28 days exposure, otherwise conditions are comparable to those for histone Hidespite the extremely different lysine/arginine ratios in the 2 materials. Reducing themolar ratio 10-fold but applying 10 times more protein to the gels gives good resultswith 10/tCi after 14 days and acceptable results with 20 fid after 48 h.

Protein labelling with 3H-NSP

DISCUSSION

It is the purpose of this paper to report of the application of a new, low-molecular-weight substance which permits careful labelling of proteins in submicrogrammequantities. As a tritium compound it is intentionally thought to be an alternative toradionuclides of iodine to the extent that facilities for work with these substances are

IH1

I L

H3, 2B, 2A, 4Fig. 3. Proteins in the sediment of salt-treated (300 mM NaCl) isolated salivarygland nuclei labelled with 30/iCi 'H-NSP (exposure time 5 days), (i) ca. 100 nuclei;(ii) ca. 200 nuclei (densitogram from this run). (Courtesy J. Bastian).

not available or that the procedure with chloramine T risks interference with bio-logical, immunological and receptor-binding activity. The latter difficulty can beovercome by the use of lactoperoxidase iodination (Parsons & Kowal, 1979) or electro-lytic procedures (Teare & Rosenberg, 1978; Sammon, Stansbury & Stahr, 1979) withsample sizes ^ 0-5 /ig protein.

It could be shown that nonhistone proteins from calf thymus chrornatin can bedepicted to a certain extent after 3H-NSP labelling and subsequent fluorographywhich until now was unsuccessful with Coomassie blue staining. It will, however,

326 G. H. Miiller

remain a qualitative procedure which can be outdone by quantitative extraction andseparation methods.

Concerning the salivary gland proteins and proteins of isolated nuclei, the mainresult is that the amount necessary for electrophoretic separation could be loweredconsiderably; it is intended to separate nuclear proteins of one single gland. Plagens(1978) claimed to be able to analyse the protein complement of 15 polytene genomeswithin 2 days after iodination with the chloramine T method; it should be interestingto label nuclei in parallel experiments with iodine and NSP and to compare theresulting patterns. The 8H-NSP method established here was, however, not intended

300 -

1:15 2-3 4-6Molar ratio, NSP/protein

Fig. 4. Total calf thymus histone (1 fig) labelled with 5, 10 or 20/tCi 'H-NSP:0-5 ( • ) , 0-3 (O), and 01 (D) M? were electrophoresed and peak areas determinedof H3/H2B + H2A bands after densitometry of the fluorogram (exposure time 3 days).

for optimization in micro-scales nor for speed in obtaining results, but rather for lessexigent routine work.

A future goal, comparing proteins of single nuclei, can only be achieved with the aidof micromethods (Neuhoff, 1973); it seems to have been approached by demonstrationof the separation of proteins from as little as 10 salivary gland nuclei by means ofmicromanipulator-assisted microelectrophoresis with gels 50-100/tm in diameter(Grossbach & Kasch, 1977).

Another tritium-labelled compound, 3H-dansyl chloride (DC1), originally and stillused in amino acid analysis, was recently proposed for the quantitation of proteins(Schultz & Wassarman, 1977). Unfortunately, the sensitivity of the 3H-DC1 assay isreduced 100-fold in the presence of nucleophilic compounds (e.g. dithiothreitol ormercaptoethanol) so in cases where these compounds are indispensable 3H-DC1 willbe of no use.

Protein labelling with 3H-NSP 327

Recently, a different and promising method for tritium labelling of submicro-gramme quantities of protein (down to around 10 ng) was proposed: proteins arelabelled in vitro by reductive methylation of amino groups with formaldehyde and highspecific 3H-potassium borohydride (8H-KBH4) (Kumarasamy & Symons, 1979).After SDS-polyacrylamide gel electrophoresis and fluorography protein bands can beevaluated in a manner similar to the one described here.

Comparable results should also be achieved by labelling proteins prior to electro-phoresis with fluorescent reagents: e.g. fluorescamine (Fluram) (Castell, Pestana,Castro & Marco, 1978; Ragland et al. 1978; Sung, Bozzola & Richards, 1978) orMDPF (2-methoxy-2,4-diphenyl-3(2H)-furanone) (Goldberg & Fuller, 1978;Douglass, LaMarca & Mets, 1978). A detection limit of 1 ng is claimed for the use ofMDPF requiring, however, instrumentation for fluorescence gel scanning andextreme precautions against fluorescent contaminants of glassware.

The use of mercaptoethanol in the 3H-NSP assay does not interfere with thereaction (not shown) but further tests will have to be carried out to determine thespecificity of the reaction and possible interferences with cellular macromolecules.

I am indebted to M. Robert for initiation of 'H-NSP experiments, to U. Plagens for intro-ducing me to iodination technique, and to N. U. Bosshard for the gift of chromatin and histonesamples.

REFERENCESBISSELING, T., BERENDES, H. D. & LUBSEN, N. H. (1976). RNA synthesis in puff 2-48BC after

experimental induction in Drosophila kydei. Cell 8, 299-304.BOLTON, A. E. (1977). Radioiodination Techniques. Amersham: The Radiochemical Centre.BOLTON, A. E. & HUNTER, W. M. (1973). The labelling of proteins to high specific radio-

activities by conjugation to a 1S4I-containing acylating agent. Biochem. J. 133, 529—539.BONNER, W. M. & LASKEY, R. A. (1974). A film detection method for tritium-labelled proteins

and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46, 83-88.BOSSHARD, N. U. (1979). Einfiuss von Ionen und NichtMston-Proteinen auf den strukturellen

Zustand von isoliertem Kalbsthymus-Chromatin. Doctoral Thesis, University of Saarbriicken.BOYD, H., LEACH, S. J. & MILLIGAN, B. (1972). iV-acylsuccinimides as acylating agents for

proteins: the selective acylation of lysine residues. Int. J. Peptide Protein Res. 4, 117-122.CASTELL, J. V., PESTANA, A., CASTRO, R. & MARCO, R. (1978). Fluorometric assays in the study

of nucleic acid-protein interactions. II. The use of fluorescamine as a reagent for proteins.Analyt. Biochem. 90, 551-560.

CHAMBERLAIN, J. P. (1979). Fluorographic detection of radioactivity in polyacrylamide gelswith the water-soluble fluor, sodium salicylate. Analyt. Biochem. 98, 132-135.

DANEHOLT, B. & EDSTROM, J.-E. (1967). The content of deoxyribonucleic acid in individualpolytene chromosomes of Chironomus tentans. Cytogenetics 6, 350-356.

DOUGLASS, S. A., LAMARCA, M. E. & METS, L. J. (1978). Methods and instrumentation forfluorescence quantitation of proteins and DNA's in electrophoresis gels at the 1 ng level. InElectrophoresis '78 (ed. N. Catsimpoolas), pp. 155-165. New York-Amsterdam-Oxford:Elsevier, North-Holland.

ELGIN, S. C. R. & WEINTRAUB, H. (1975). Chromosomal proteins and chromatin structure.A. Rev. Biochem. 44, 725-774.

GOLDBERG, R. L. & FULLER, G. C. (1978). Quantitation of collagen in polyacrylamide gels byfluorescent scanning of MDPF-labelled proteins: an improvement over densitometricscanning of gels stained by Coomassie Blue. Analyt. Biochem. 90, 69-80.

328 G. H. Miiller

GROSSBACH, U. (1977). The salivary gland of Chironomus (Diptera): a model system for thestudy of cell differentiation. In Biochemical Differentiation in Insect Glands (ed. W. Beermann),pp. 147-196. Berlin-Heidelberg-New York: Springer.

GROSSBACH, U. & KASCH, E. (1977). Microelectrophoresis of proteins at the cellular level. InTechniques of Biochemical and Biophysical Morphology, vol. 3 (ed. D. Glick & R. M. Rosen-baum), pp. 81-101. New York: Wiley-Interscience.

KROEGER, H. (1966). Micrurgy on cells with polytene chromosomes. Meth. Cell Physiol. 2,61-92.

KROEGER, H. (1973). Zur Normalentwicklung von Chironomus ikummi Kieffer. Z. Morph. Tiere74, 65-88.

KUMARASAMY, R. & SYMONS, R. H. (1979). The tritium labeling of small amounts of protein foranalysis by electrophoresis on sodium dodecyl sulfate-polyacrylamide slab gels. Analyt.Biochem. 95, 359-363.

LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head ofbacteriophage T 4. Nature, Lond. 327, 680-685.

LASKEY, R. A. & MILLS, A. D. (1975). Quantitative film detection of 3H and "C in polyacryl-amide gels by fluorography. Eur. J. Biochem. 56, 335-341.

LEACH, S. J. & BOYD, H. (1973). Thermal transitions of acylated ribonuclease. Int. J. PeptideProtein Res. 5, 239-250.

LESTOURGEON, W. M. & BEYER, A. L. (1977). The rapid isolation, high-resolution electro-phoretic characterization, and purification of nuclear proteins. Meth. Cell Biol. 14, 387-406.

LEZZI, M. & ROBERT, M. (1972). Chromosomes isolated from unfixed salivary glands ofChironomus. In Developmental Studies on Giant Chromosomes (ed. W. Beermann), pp. 35-57.Berlin-Heidelberg-New York: Springer.

NEUHOFF, V. (ed.) (1973). Micromethods in Molecular Biology. Berlin-Heidelberg-New York:Springer.

PARSONS, R. G. & KOWAL, R. (1979). Solid-phase lactoperoxidase radioiodination of micro-gram quantities of protein mixtures for electrophoretic analysis. Analyt. Biochem. 95,568-574.

PLAGENS, U. (1978). Effect of salt-treatment on manually isolated polytene chromosomes fromChironomus tentans. Ckromosoma 68, 1-19.

RAGLAND, W. L., BENTON, T. L., PACE, J. L., BEACH, F. G. & WADE, A. E. (1978). Analysis ofproteins in polyacrylamide gels by scanning for fluorescence. In Electrophoresis '78 (ed.N. Catsimpoolas), pp. 217-230. New York-Amsterdam-Oxford: Elsevier, North-Holland.

RANDERATH, K. (1970). An evaluation of film detection methods for weak /^-emitters, particu-larly tritium. Analyt. Biochem. 34, 188-205.

ROBERT, M. (1975). Isolation and manipulation of salivary gland nuclei and chromosomes.Meth. Cell Biol. 9, 377-390.

RUDINGER, J. & RUEGG, U. (1973). Preparation of iV-succinimidyl 3-(4-hydroxyphenyl)-propionate. Biochem. J. 133, 538-539.

SAMMON, P. J., STANSBURY, M. & STAHR, G. (1979). Micro-electrolytic iodination of poly-peptide hormones. Int. J. appl. Radiat. Isotopes 30, 359—364.

SCHULTZ, R. M. & WASSARMAN, P. M. (1977). (3H)Dansyl chloride: a useful reagent for thequantitation and molecular weight determination of nanogram amounts of protein. Analyt.Biochem. 77, 25-32.

SERFLING, E., MAJOROV, V. I., MIKICHUR, N. I., POPOVA, T. G. & SANDAKCHIEV, L. S. (1975).DNA and RNA content of Chironomus thummi polytene chromosomes determined by micro-column gel filtration. Cell Diff. 3, 361-370.

SUNG, M. T., BOZZOLA, J. J. & RICHARDS, J. C. (1978). In situ antigen localization withnuorescamine-conjugated antibodies: a new method. Analyt. Biochem. 84, 225-230.

TEARE, F. W. & ROSENBERG, R. A. (1978). Micro-electrolytic radioiodination of polypeptidehormones to high specific activity. Int. J. appl. Radiat. Isotopes 29, 567-569.

WILSON, A. T. (1958). Tritium and paper chromatography. Nature, Lond. 182, 524.

(Received 5 October 1979)