430 POLYMERASE ACCESSORY FUNCTIONS [32]

interpreted as a demonstration of a biologically significant protein-protein interaction. The two proteins must be shown to be involved in the same process in vivo; ideally the interaction should be demonstrated by a func- tional assay or by any of several genetic techniques that can detect protein- protein interactions. Nevertheless, protein affinity chromatography is a useful technique, both as an initial method for detecting the potential components of protein machines and for the purification of such compo- nents. Protein affinity chromatography provides a unique opportunity to exploit the forces that cause protein machines to assemble to allow explora- tion of the composition and regulation of these complexes that act in so many cellular processes.

[32] R a d i o l a b e l i n g o f P r o t e i n s fo r B i o c h e m i c a l S t u d i e s

By Zvi KELMAN, VYTAUTAS NAKTINIS, and MIKZ O'DONNELL

Introduction

Several processes in nucleic acid metabolism such as replication, tran- scription, and translation require the coordinated action of numerous pro- teins. This coordination is manifest through protein-protein and protein- nucleic acid interactions ranging from stabile complexes to transient contacts. Important questions in these multicomponent systems include the following: How much of each protein is present in protein complexes or on the nucleic acid? Is the protein interacting with a specific nucleic acid structure or sequence? How tightly associated are the subunits of a complex, and how are these interactions influenced by other proteins or by ATP? Which surfaces on a protein interact with other proteins? Answers to questions such as these require a convergence of several experimental techniques. Some techniques that address these subjects utilize radioac- tive proteins.

This chapter describes two radioactive labeling methods and illustrates a few applications of labeled proteins in biochemical studies. One method is reductive methylation, 1-3 which was described in this series using [14C]formaldehyde.2 We present a modified protocol using NaB3H4 that

I G. E. Means, this series, Vol. 47, p. 469. 2 G. E. Means and R. E. Feeney, Biochemistry 7, 2192 (1968). 3 R. H. Rice and G. E. Means, J. Biol. Chem. 246, 831 (1971).

Copyright © 1995 by Academic Press, Inc. METHODS IN ENZYMOLOGY, VOL. 262 All rights of reproduction in any form reserved.

[32] RADIOLABELING OF PROTEINS 431

introduces only 1 or 2 [3H]methyl groups per protein for a specific activity of 20 to 40 Ci/mmol (0.5 to 1.0 × 106 cpm//zg of 50-kDa protein). The second method uses a specific protein kinase to 32p-end-label a protein into which a 5-7 amino acid residue kinase recognition motif has been engineered onto the N or C terminus. 4,5 This procedure results in a specific activity up to 3000 Ci/mmol (>60 × 106 cpm/tzg of 50-kDa protein). We have no experience in performing protein labeling using isotopes of iodine and for this the reader is referred to an excellent treatise in this series. ~

A drawback to chemical labeling is that the modified protein may lose one or more of its activities. Hence, the labeled protein must be tested for activity relative to unlabeled protein. Reductive methylation is the least invasive technique because it introduces only one or two methyl groups. In our experience, 12 of the 13 different proteins that were labeled retained 85 to 100% activity. End-labeling with 32p introduces a larger modification and some proteins may be expected to lose function, but the three proteins we have 32p-end-labeled retain their activity.

Proteins that are radiolabeled metabolically (in vivo) with either [35S]methionine o r 3H- or 14C-labeled amino acids should retain full activity. However, the drawback to metabolic labeling is the need to purify the radioactive protein, and during the prolonged use of centrifuges and cold room equipment it is difficult to contain the radioactivity. The chemical labeling methods described here also carry a health hazard, but the proce- dures can be performed in a fume hood in a matter of a few hours. It is important that the operations be performed using all available precautions including gloves, laboratory coat, and film badge, and that proper guidelines are followed to dispose of the radioactive waste.

3H-Labeling by Reductive Methylation

The procedure described here modifies only a few lysine residues per protein molecule and is adapted from methods used in Arthur Kornberg's laboratory] The reaction, I'2 outlined in Eq. (1), involves the addition of formaldehyde, which forms a Schiff base with lysine followed by reduction using NaB3H4 and results in net replacement of a proton for a [3H]methyl group on the primary amine of lysine. Once lysine is methylated, it is more reactive toward a second round of methylation. 1 The charge of the methylated lysine is conserved; the pK of monomethylated lysine is in-

4 B.-L. Li, J. A. Langer, B. Schwartz, and S. Pestka, Proc. Natl. Acad. Sci. USA 86, 558 (1989). 5 M. A. Blanar and W. J. Rutter, Science 256, 1014 (1992). ~' C. W. Parker, this series, Vol. 182, p. 721. 7 K.-i. Arai. S.-i. Yasuda, and A. Kornberg, J. Biol. Chem. 256, 5247 (1981).

432 POLYMERASE ACCESSORY FUNCTIONS [32]

creased by about 0.3 pH unit and the pK of dialkylated lysine is lower than lysine by about 0.6 pH units. 2 When only a few methyl groups are incorporated into a protein, the monoalkylated lysine species predomi- nates. 2

H H I I

Protein-NH 2 + C ~--- O ~ Protein-.N. ~ C I I

H H20 H

H H NaB3H4 I I

~. > Protein_N-C-3H . . I

H

(1)

The extent and velocity of the reaction increases with pH, very little reaction occurs at pH 7.0 and maximal labeling is achieved at pH 9.5. 2 The elevated pH is needed to form the Schiff base between lysine and formaldehyde (a readily reversible reaction), and to preserve the sodium borohydride, which is rapidly decomposed as the pH is lowered. To prevent extensive labeling the reaction is carried out at pH 8.5. The reaction can be performed at pH 7.0 using cyanoborohydride, 8 although the temperature must be elevated and the reaction time may be longer. Reductive methyla- tion occurs only with lysine and with the amino terminus, no reaction occurs with side chains other than lysine, and disulfide bonds are not reduced) Most of the radioactivity is not incorporated into protein but is consumed by side reactions such as reduction of formaldehyde to methanol and break- down of borohydride by solvent. Buffers containing primary and secondary amines must be avoided because they form the Schiff base and consume the reactants. We have used sodium borate although buffers with tertiary amines or phosphate buffer can be used. Besides formaldehyde, other aldehydes and ketones can be used (e.g., acetone and acetaldehyde), but the resulting alkyl group on the lysine is then bulkier (isopropyl and ethyl, respectively). 2 Hence we use formaldehyde to produce the smallest modifi- cation possible. Note that formaldehyde itself can modify proteins by form- ing inter- and intraprotein cross-links; however, the destructive reactions of formaldehyde with proteins during reductive alkylation has been examined with none of these products being observed under conditions that are much more extensive in time and temperature than those used here. 8

Materials for Reductive Methylation

We usually purchase 0.5 Ci of NaB3H4 of 50 to 75 Ci/mmol (Du Pont- New England Nuclear, NET-023X). 0.5 Ci of 75 Ci/mmol NaB3H4 is 6.7 /zmol, which is sufficient to label 3.3 ml of protein. We usually label 1 ml

8 N. Jentoft and D. G. Dearborn, J. Biol. Chem. 254, 4359 (1979).

[321 RADIOLABELING OF PROTEXNS 433

of 3 to 5 different proteins on the same occasion. Unfortunately, NaB3H4 of this specific activity is not available in smaller quantities. Lower specific activity material can be purchased such as 100 mCi NaB3H4 at 5 to 15 Ci/mmol (NET-023H). Because most of the radioactivity is lost in compet- ing side reactions, the higher the concentration of the protein, the more labeled protein is produced per mole of NaB~H4 (i.e., the resulting specific activity of the protein is unaffected whether a 1 or a 5 mg/ml solution of protein is used). Protein can also be labeled using [14C]formaldehyde,3 but the specific activity is 1000-fold lower than NaB3H4.

The other necessary materials include a small container for radioactive solid waste, a triangular file, 1 to 200-/.d Pipetman with an extended tip (explained below), ice bucket with each dialyzed protein in a separate open tube, 10 mM NaOH on ice, 2 M formaldehyde on ice, 1 M lysine, and an empty Eppendorf tube (for the NaB3H4 after it is dissolved in 10 mM NaOH). To separate protein from reagents after the reaction, one fraction collector is needed for each protein, 10 ml columns of Sephadex G-25 packed in disposable 10-ml plastic pipettes with a glass wool plug, and gel filtration column buffer. Our typical column buffer is 20 mM Tris-HC1, pH 7.5, 0.5 mM EDTA, 2 mM dithiothreitol (DTT), and 20% glycerol.

Procedure for Reductive Methylation

The night before the labeling reaction, dialyze each protein at a concen- tration of 1 to 5 mg/ml into 50 mM sodium borate, pH 8.5, 0.5 mM EDTA, 10% glycerol (DTT can be included). The next day the reagents and hard- ware must first be arranged in the fume hood. Then the sealed ampule containing the dry NaB3H4 is gently tapped to place all the powder at the bottom of the vial and the file is used to score the neck of the ampule. The ampule will have 3Hz gas under pressure and it is important when snapping the neck of the ampule to position it near the hood exhaust port. Dissolve the NaB3H4 powder in 10 mM NaOH for a final concentration of 100 mM NaB3H4 (this will be in the range of 67 to 100/~1 depending on the specific activity of the NaB3H4). Due to the length of the ampule and the narrow opening at the neck, the Eppendorf tip must be modified so that it reaches to the bottom of the ampule. A simple modification is to fasten a 2-in. length of polyethylene tubing onto the tip. The 10 mM NaOH is drawn into the Pipetman and then transferred into the ampule to dissolve the NaB3H4, then withdrawn and placed into the empty Eppendorf tube on ice. At this point one must work quickly, but steadily, because the reagent is decomposing (as evidenced by the slow appearance of bubbles).

The labeling reaction is initiated upon adding formaldehyde to the protein(s) to a final concentration of 20 mM and then NaB3H4 to a final

434 POLYMERASE ACCESSORY FUNCTIONS [321

concentration of 2 mM. After a further incubation of 15 min on ice, stop the reaction by adding lysine to a final concentration of 50 raM, and separate the protein from the reagents by passing the reaction mixture over a 10- ml column of Sephadex G-25. The column is disposed of as radioactive waste at the end of the procedure. The column is initially prepared in the cold room and the buffer is kept on ice, but just before use it is brought into the fume hood along with a fraction collector. Fractions of 400/xl are collected and placed on ice. A small amount (3 txl) of each fraction is counted to locate the protein peak, these are pooled, the protein concentra- tion is determined, and then the [3H]protein is aliquoted and stored frozen at -70 °. The labeling procedure should take no more than 15 to 20 min, the gel filtration about 30 min, and the counting and pooling of column fractions, protein determination, and aliquoting may take about 1 hr.

Variables in this reaction include time, pH, temperature, and concentra- tions of formaldehyde and NaB3H4. We find that under the conditions described above, the rate of methylation is nearly linear for 10 rain and levels off by 15 rain. Also, lowering the formaldehyde from 20 to 10 mM yields half the level of methylation. Finally, use of 2 mM NaB3H4 is sufficient for the low level of labeling desired; 4 mM NaB3H4 increased the extent of labeling by only 20%, and thus is put to better use at 2 mM to label more protein. We have always performed these procedures on ice and have not experimented with different temperatures. We have examined the effect of increasing the pH to 9.0 and find the extent of methylation increases at least 1.5-fold, but pH 8.5 is less harsh on the protein and is sufficient for incorporation of one to two methyl groups.

32p-End-Labeling of Proteins

An efficient substrate for the cAMP-dependent protein kinase has been identified as a heptapeptide Leu-Arg-Arg-Ala-Ser-Leu (or Val)-Gly (or Ala). 9'1° The Km of this sequence for phosphate transfer to the Ser by the kinase is 10 to 20/xM. 9'1° Use of only the five inner residues results in an eightfold increase in Km 10 and substitution of any of the five inner residues increases the Km by 10-400 fold, resulting in low to negligible rates of phosphorylation. 9 Radioactive proteins of high specific activity (->3000 Ci/ mmol) have been produced upon cloning five to seven residues of this sequence onto the C terminus of human interferon o~ 4 and onto the N terminus of segments of c-Fos)

We have engineered the kinase recognition sequence onto either the

9B. E. Kemp, D. J. Graves, E. Benjamini, and E. G. Krebs, J. Biol. Chem. 252, 4888 (1977). ~o (). Zetterqvist, U. Ragnarsson, E. Humble, L. Berglund, and L. Engstr6m, Biochem. Biophys.

Res. Commun. 70, 696 (1976).

132] RADIOLABELING OF PROTEINS 435

A) N-terminus wild type g e n e

IM I. R R A S V { 5' - CTGCGGCATATGCTTCGAAGAGCTTCTGTT-[15 matching nucleotides] - 3'

I I I I Ndel site BstBl site

B) C-terminus wild type gene

4 ] L R R A S V (, stop 3' - [15 matching nucleotides]-GAAGCTTCTCGAAGACAACCAATTCCTAGGTGGTC - 5'

I I I I BstBl site Bamlll site

Fie. 1. Oligonucleotide sequences for PCR amplification of either N- or C-terminal PK protein. The oligonucleotides shown are designed for use with a second oligonucleotide that matches the gene of interest to produce a PCR product encoding either (A) an N-terminal PK protein or (B) a C-terminal PK protein. The encoded amino acids of the kinase recognition motif are shown above the nucleotide sequence. In each case, the 15 nucleotides at the 3' terminus match the gene of interest. The NdeI and BamHI cloning sites and the BstBI screening site are marked. The 5-6 nucleotides at the 5' terminus of these oligonucleotides ensure efficient cleavage of the PCR product.

N or C terminus of three proteins, expressed them in E. coli, and purified them to homogeneity. We have mainly studied the/3 subunit of E. coli DNA polymerase III holoenzyme (Pol III) in which a six-residue site, Leu-Arg- Arg-Ala-Ser-Val (followed by Pro), was engineered onto the C terminus. In addition, we have placed the seven-residue site, Leu-Arg-Arg-Ala-Ser- Val-Gly, onto the C terminus of the A cro repressor and of EBNA1, the latent origin binding protein of the Epstein-Barr virus. We have also placed the seven-residue sequence Met-Leu-Arg-Arg-Ala-Ser-Val onto the N ter- minus of/3 and cro. These protein kinase (PK) proteins were engineered by PCR using an oligonucleotide encoding the kinase motif (Fig. 1) and expressed in the pET-3 system, it The kinase recognition motif can also be introduced by site-directed mutagenesis. 4 Since the pET-3 system was used for expression, the primer sequences contained an NdeI site for N-terminal PK proteins, and a BamHI site for C-terminal PK proteins. The BstBI site is not present in pET-3 and thus aids in screening clones.

Procedure for 32p-End-Labeling

Unlike reductive methylation, this technique can be performed with small amounts of protein and the procedure is compatible with a variety

~ W. F. Studier, A. H. Rosenberg, J. J. Dunn, and S. W. Dubendorff, this series, Vol. 185, p. 60.

436 POLYMERASE ACCESSORY FUNCTIONS [321

of buffers. We typically use a volume of 30 to 120 tzl containing 0.2 to 6 nmol PK protein in 20 mM Tris-HC1, pH 7.5, 1 to 60/xM [y-32p]ATP (specific activity discussed below), 2 mM DTT, 100 mM NaC1, 12 mM MgC12, and 10 mM NaF. Labeling is initiated upon adding cAMP-depen- dent protein kinase (0.008 U kinase/nmol PK protein) and shifted to 37 °. After 10 rain at 37 °, the reaction is stopped on addition of 30 mM EDTA (final concentration). The extent of 32p incorporated into the PK protein can be determined by acid precipitation or by analysis in a SDS- polyacrylamide gel (for the latter, addition of 5 mM unlabeled ATP to the sample buffer lowers background radiation). Free [y-32p]ATP can be removed either by ultrafiltration, spin dialysis, or gel filtration. The catalytic subunit of cAMP-dependent protein kinase from bovine heart can be pur- chased from Sigma and was used in previous s tudiesy however, we have used the murine version expressed and purified from E. coli 12 (gift of Dr. Susan S. Taylor, University of California at San Diego).

Under these conditions the stoichiometry of Pi incorporation is typically 0.5 to 0.8 tool Pi/mole protein (except N-terminal labeled/3 PK, discussed below). Before embarking on these studies, it is important to determine whether the wild-type protein is phosphorylated by protein kinase. Wild- type/3 was not phosphorylated to a significant extent (<0.9 mmol Pi/mol /3PK) and sequence analysis of the radioactive chymotryptic peptide of [32p]/3PK confirmed the label was in the kinase motif at the C terminus.

High specific activity protein can be achieved using straight [y-32p]ATP (3000 Ci/mmol) at approximately 1/xM, or the relative proportion of unla- beled ATP and radioactive ATP can be adjusted to achieve the desired specific activity. This can be very important in experimental designs using both 3H-labeled protein and 3ZP-labeled protein in the same experiment because the specific activity of 3H-labeled proteins is only 20 to 40 Ci/mmol and therefore one must take care to prepare the 32p-labeled protein with a comparable specific activity so as not to encounter problems of bleed over between windows while counting both isotopes.

Important Considerations

One advantage to the 32p-end-labeling method is that the isotope is available in small quantities and the manipulations are similar to end- labeling DNA, a common laboratory technique. PK proteins can also be labeled using [T-35S]ATP. a3

How often do labeled PK proteins retain activity? The modification is

12 L. W. Slice and S. S. Tylor, J. Biol. Chem. 264, 20,940 (1989). 13 V. Naktinis and M. O'Donnell, unpublished (1993).

[32] RADIOLABELING OF PROTEINS 437

su~tantial; not only are several amino acids added to the protein, but a large phosphate group is also present. In our experience the activity of both N- and C-labeled [32p]fleR is unchanged in several assays including assembly of [32p]~PK o n t o primed DNA by Pol 1II, and ability of [32p]flPK

to confer highly processive DNA synthesis onto the polymerase.14 Likewise, the [32p]EBNA1PK and both N- and C-labeled [32p]croPK retained their site- specific DNA binding activities. It is important to note that if the amount of ATP used in the labeling reaction is substoichiometric to the protein (often the case in DNA end-labeling reactions), some types of activity assays will not truly evaluate whether the phosphorylated species retains activity.

In theory one should be able to place the kinase motif at any exposed region of a molecule and label it. However, placement of this kinase motif at other positions in the/3 subunit has revealed limitations to this technique. For example, the N-terminal/3Pr: is labeled approximately 4% at best. 13 The problem is likely due to the inability of the kinase to gain access to the recognition site as addition of 0.5 M urea increased the labeling to 16%.~3 Further, we have placed the seven-residue kinase recognition sequence into two different internal positions of/3, which are located on the surface according to the crystal structure, but the kinase did not phosphorylate these proteins/5

Examples of Applications

Interaction of Proteins with DNA

To identify a particular subunit on DNA in a complex mixture of several proteins, the reaction is analyzed by gel filtration using beads with large pores such that free proteins elute in the included fractions and resolve from proteins bound to the large DNA (e.g., plasmid or M13 ssDNA), which elute in the excluded fractions. For these studies we have used agarose beads Bio-Gel A-5m and A-15m (Bio-Rad) as well as Sepharose 4B (Pharmacia-LKB), which exclude plasmids and ssDNA phage genomes, yet include proteins of up to 1 MDa.

Examples of this technique abound in studies from Arthur Kornberg's laboratory. For example, 3H-labeled proteins were used to identify primoso- mal proteins that remained on the ~bX174 ssDNA with the finding that the DnaC protein, while essential to assemble the primosome on DNA, does not remain on the D N A . 16 The E. coli Pol III contains a five-subunit

14 p. T. Stukenberg, Ph.D. thesis (1993). 15 j. Turner and M. O'Donnell, unpublished (1993). in j. A. Kobori and A. Kornberg, Z Biol. Chem. 257, 13,770 (1982).

4 3 8 POLYMERASE ACCESSORY FUNCTIONS [321

A25o

200

lSO

"6 lOO

50

C 100

90

80

70 Q

6o

50

I 40

30

20

10

E 525,

450 ! i

375

~ 3 0 0

~ 225

~_..150

75

. ~ l e x

"To 20 30 fraction

B16OO 1400

1200

" ~ ' 1 0 0 0

~ 800

, 600 DZ

400

200

0 0 i0

, • Sm~I Q 0

Sma I ear

ular

i t = 20 30 4 0 fraction

50

, . , ~ i 3 ~ j - ~ . - m T x _ " ,

I g' a

l h r

15 25 35 45 fraction

epitope . . . . .

epitope epitope epitope-132 [3H]- ~ 45Oc

37°C

• i . ~ ; - i - incubation time (hrs)

10

D 80

70

-~ 60 o e~

50

40

30

2( 1'0

F

O

8

~ NA Kd= 17rim

SSB [SH]-~

2'o ~o 40 fraction

ldnase ~ ~32p

0.E " ~K :OOH [y32p] ATP 0

0.5 1 1.5 2 4 8 30 rain

0 10 20 time (min)

0.~

0.~

FiG. 2. Applications of radiolabeled proteins to study the molecular dynamics and structural details of multicomponent complexes, (A) The [3H]y complex matchmaker assembles the [32p]/3PK dimer onto SSB coated primed M13mp18 ssDNA. Gel filtration on Bio-Gel A-5m resolves the [32p]/~eK bound to DNA (fractions 10-15) from the [3HIT complex that dissociates from DNA and elutes in the included fractions. The experiment was performed essentially as described in Fig. 2.3 of Stukenberg) 4 (B) During gel filtration, [3H]~ clamps coelute with

30

[32] RADIOLABELING OF PROTEINS 439

subassembly called the y complex, which acts in a similar fashion as DnaC protein; 3' complex places the/3 subunit of Pol III onto a primed template but then departs from the D N A . 17 Figure 2A shows the use of two labeled subunits in this latter reaction. The 3H-labeled y complex was used to place [32p]/3PK onto a primed template and the reaction was analyzed by gel filtration. Analysis of the column fractions shows a stoichiometric amount of [32p]/3PK dimers comigrating with the DNA, but the [3H]T complex mi- grates as free protein complex in the included fractions.

Topological Binding of Protein to DNA

The/3 subunit of Pol III is a ring-shaped dimer that completely encircles DNA. ~8 Once the/3 ring is assembled onto D N A (by the y complex and ATP) it tethers the rest of the Pol III machinery to the template,and slides along with it for highly processive DNA synthesis. ~9 This simple solution

~7 p. T. Stukenberg, P. S. Studwell-Vaughan, and M. O'Donnell, J. Biol. Chem. 266, 11,328 (1991).

is X.-P. Kong, R. Onrust, M. O'Donnell, and J. Kuriyan, Cell 69, 425 (1992). ~9 M. O'Donnell, J. Kuriyan, X.-P. Kong, P. T. Stukenberg, and R. Onrust, Mol. Biol. Cell 3,

953 (1992).

(bind to) a circular plasmid with a single nick, but upon linearization of DNA the [3H]/~ clamps fall off and elute in the included fractions. (Adapted from Fig. 3 in Stukenberg et al. t7

Copyright The American Society for Biochemistry & Molecular Biology.) (C) Dissociation of [3H]6' from the y complex was observed by incubating 14 pmol of y complex (reconstituted with [3H]8') with 108 pmol &' in 60/.d of buffer 20 mM Tris-HCl, pH 7,5, 100 mM NaCI, and 10% glycerol at 22 ° . At the indicated times, the samples were analyzed by filtration on a 24-ml Superose-12 column. (D) Equilibrium gel filtration analysis of the interaction between [3H]x and the SSB-DNA complex was performed on a 5-ml column of P-30 (Bio-Rad) using 500 nM [3H]x in the column buffer and injecting on the column 1.7 nmol SSB in complex with M13mpl8 (10/xg ssDNA) in column buffer. Fractions were collected and quantitated by scintillation counting. The Kd of 17 nM was calculated from the amount of [3H]x bound to the SSB-DNA in the peak. (E) Protomer interchange among/3 dimers using a mixture of 2 pmol [3H]fl 2 and 2 pmol hemagglutinin epitope tagged432 in 50/zl of 20 mM HEPES, pH 7.5, 150 mM NaCl, 0.1% Triton X-100, 10% glycerol followed by immunoprecipitation at the indicated times using monoclonal antibody to the epitope (12CA5, BAbCO) as described by Kolodziej and Young. 23 The [3H]fl that coimmunopreeipitated with the epitope tagged-fl indicates the extent of subunit exchange to form the heterodimer between the [3H]/3 and epitope tagged-/3. The experiment was repeated at the indicated temperatures. (F) In the kinase protection assay, 180 pmol /3 PK was preincubated for 30 rain at 15 ° either with or without 1 nmol 88' complex, followed by treatment with protein kinase in a 30-/.d reaction mixture as described in the text. At the indicated times 3-/zl aliquots were analyzed by SDS polyacrylamide gel and autoradiography (inset). The kinase motif is located on the C termini, which extrude out from the same face of the J3 dimer as indicated in the scheme.

440 POLYI~.ERASE ACCESSORY FUNCTIONS, [32]

to high processivity may generalize to the gene 45 protein processivity factor of the phage T4 replicase~ a~d the PCNA processi~,ity factor of the yeast and human potymerase ~19 Protein surroundir~gDNA seems so basic that it may even generalize to other DNA metabolic machineries. For example, replicative helicases such as those encoded by phage T4 and phage T7~ E. coli DnaB, and simian vies. 40 (SV40) T antigen are all hexamers and may surround duplex DNA for their action (i.e., l~e'the sixfold pseudo- symmetric/3 dimer).

raitial experiments using [3H]~ revealed this "topological binding" mode of/3 to DNA and predicted its ring shape, 17 thus motivating the crystal structure analysis. A particularly telling experiment is shown in Fig. 2B ~,. Here, [3H]/3 was placed onto a singly nicked plasmid DNA (by -/ complex and ATP) and then gel filtered to reveal that several/3 dimers coelute with the DNA in the excluded fractions. 17 However, linearization o£ the DNA results in dissociation of/3, implying that it slides along DNA and falls off the end. If/3 were to bind DNA by direct chemical interaction, as other DNA binding proteins do, then it would have stayed associated with the linear DNA through gel filtration as well as with circular DNA. Hence,/3 must be bound physically to DNA by virtue of its topology. This type of experiment is rather simple and variations on this theme may help identify the topological binding proteins of other systems.

Subunit Exchange in Multiprotein Complexes

Radiolabeled protein can be used to measure the rate of dissociation of a subunit from within a multiprotein complex. As an example, in Fig. 2C the five-subunit 3' complex (y66'XqJ) was labeled by reconstituting it using [3H]6'. To this was added an eighffold molar excess of unlabeled 6' and the mixture was gel filtered at various times. Whenever the [3Hit3'

subunit dissociates from the complex, an unlabeled 6' takes its place. Hence, over time, the column fractions containing y complex decrease in radioactiv- ity and column fractions containing free 6' increase in radioactivity. This technique can be extended to determine how the rate of subunit exchange is influenced by DNA, nucleotides and other proteins.

Weak Protein-Protein Interactions

Only strong interactions among proteins can be detected by gel filtration because it is not an equilibrium technique. However, weak interactions can be quantitated using the equilibrium gel filtration technique? ° This technique is normally used to determine the Ks between a protein and a

20 j. p. Hummell and W. J. Dreyer, Biochim. Biophys. Acta 63, 530 (1962).

[32] RADIOLABELING OF PROTEINS 441

small ligand molecule (e.g., a nucleoside triphosphate) where the ligand is radiolabeled and is present throughout the column buffer. The protein binds the radiolabeled ligand and then elutes ahead of the ligand, resulting in a peak of radioactivity that emerges above the baseline level and is followed by a trough where the unbound ligand would have eluted. The Kd can be calculated from this information. This technique has been applied in Arthur Kornberg's laboratory using [3H]/3 in the column buffer to define the Kd of a weak protein-protein interaction between the/3 subunit and the Pol III* assembly (Pol III lacking 13). 21 Figure 2D shows an analysis of a weak interaction between the X subunit of Pol III and an SSB-DNA complex using [3H]x in the column buffer. 22

Exchange Rate Among Subunits of a Dimeric Protein

How rapidly do dimers of identical subunits fall apart and come back together? For example, the protomers of the/3 dimer form a closed ring, which must open to assemble around DNA. Do the monomer units ot a /3 dimer rapidly come apart and then reassociate, thereby trapping DNA inside? The rate of exchange of monomer units of a dimer cannot be measured by gel filtration as in Fig. 2C since it is always in the dimeric s~te. An assay to measure stability of the/3 dimer, shown m Fig. 2E, uses a [3H]/3 dimer and a/3 dimer that is genetically tagged with the hemaggtminin epitope (nine amino acids) at its C-terminus. 23 Upon mixiag the t '~o/3 dimers, samples of the reaction were immunoprecipitated at time imervals. As the "heterodimer" of one [3H]/3 and one epitope .tagged-/3 ~as formed, the amount of [3H]/3 in the coimmunoprecipita~ increased. The results indicate that monorneric units of the/3 dimer reqaire :a ~ong ~ e *o ~ssoci- ate even at 37 °. This assay can be exploited to study the m~laence of other proteins and of DNA and ATP on the stability of the/3 dimer.

Kinase Protection Assay

We originally end-labeled proteins with [32p] to develop a "protein footprinting" assay for mapping the interactive surface of one protein with another (or with DNA), but this work is still in progress. However, we have developed a "kinase protection" assay. An example of this assay is shown in Fig. 2F in which/3 eK is incubated with or without the ~ ' complex, a subassembly of the y complex that binds/3. 24 Then kinase is added and

21 R. S. Lasken and A. Kornberg, .L Biol. Chem. 262, 1720 (1987). 22 Z. Kelman and M. O'Donnel l , unpubl ished (1993). 23 p. A. Kolodziej and R. A. Young, this series, Vol. 194, p. 508. 24 R. Onrust , Ph.D. thesis (1993).

~ 2 P O L Y M E R A S E A C C E S S O R Y F U N C T I O N S [ 3 3 1

the time course of phosphorylation is monitored by SDS-PAGE followed by autoradiography. The result shows that phosphorylation of/3Pi~ is almost completely btocked by the ~fi' complex. The two C termini of the/3 dimer, where the' ki,'nase motif is located, extrude from the same face of the/3 dimer (see scheme in Fig. 2F). Hence, the ability of 8fi' complex to block these sites from phosphoryl~tion suggests that &~' may interact with the "C-termiwai" face of the/3 ~ dimer.

Acknowledgments

We are grateful to Dr. Susan S. Taylor for the catalytic subunit of cAMP-dependent protein kinase. Supported by Public Health Service grant GM38839.

[33] C y c l i n g o f Escherichia coli D N A P o l y m e r a s e III f r o m O n e S l i d i n g C l a m p to A n o t h e r : M o d e l fo r L a g g i n g S t r a n d

By JENNIFER TURNER a n d MIKE O 'DONNELL

Introduction

The multiprotein replicase of the Escherichia coli chromosome, DNA polymerase III holoenzyme (Pol III), achieves a tight ATP-activated grip on DNA through its ring-shaped clamp protein, the/3 dimer ( / 3 2 ) . 1-4 The /32 ring encircles DNA, thereby acting as a sliding clamp and continuously holds Pol III to the template for remarkably high speed and processivity. 5 The model of a circular protein clamp riding along in back of the polymerase fits nicely with the continuous synthesis of the leading strand, but conceptu- ally it would hinder the discontinuous mode on the lagging strand where Pol III must rapidly dissociate from the end of one Okazaki fragment to start another fragment at the next RNA primer (i.e., Fig. 1C).

Studies of Pol III showed that it was indeed very slow to replicate more than a stoichiometric number of primed templates. 6 After Pol III completely replicated a circular single-stranded DNA (ssDNA) template, several

1 A. Kornberg and T. A. Baker, "DNA Replication." Freeman, New York, 1991. 2 X.-P. Kong, R. Onrust, M. O'Donnell, and J. Kuriyan, Cell 69, 425 (1992). 3 M. O'Donnell, J. Kuriyan, X.-P. Kong, P. T. Stukenberg, and R. Onrust, Mot. Biol. Cell 3,

953 (1992). 4 j. Kuriyan and M. O'Donnell, J. Mot. Biol. 234, 915 (1993). 5 p. T. Stukenberg, P. S. Studwell-Vaughan, and M. O'Donnell, Z Biol. Chem. 266, 11,328

(1991). 6 p. M. J. Burgers and A. Kornberg, J. Biol. Chem. 257, 11,474 (1982).

Copyright © 1995 by Academic Press, Inc. METHODS IN ENZYMOLOGY, VOL. 262 All rights of reproduction in any form reserved.