Proc. Nat. Ad. Sci. USA
Vol. 71, No. 11, pp. 425-4428, November 1974

A Novel Form of RNA Polymerase from Escherichia coli

(M13/+X174/rifampicin)

WILLIAM WICKNER AND ARTHUR KORNBERG

Department of Biochemistry, Stanford University School of Medicine, Stanford, California 94305

Contributed by Arthur Komberg, September 6, 1974

ABSTRACT   A new form of RNA polymerase, termed
RNA polymerase 111, has been recognized as a large frac-
tion of the rifampicin-sensitive enzyme in E. coli. It is
physically separable from RNA polymerase (holoenzyme,
RNA polymerase I) by gel filtration and is distinguished
by its capacity to discriminate between M13 and 4x174
viral DNA templates in priming DNA synthesis. This tem-
plate specificity is manifested only with saturating levels
of DNA unwinding protein and characterizes the priming
of DNA synthesis on viral single strands in cell-free ex-
tracts and in vivo. RNA polymerase 111 has less than 5%
of the specific activity of RNA polymerase 1 in transcribing
duplex DNA of phages X and T4, salmon sperm DNA, and
the copolymer poly[d(A-T)). Rifampicin inactivation of
RNA polymerase 111 releases a factor, presumably a small
subunit, which can be isolated and used to confer on
RNA polymerase I the properties of 111, namely, dis-
crimination between M13 and 6x174 templates in priming
DNA synthesis, and a relative inability to transcribe duplex
DNA.

The observation that rifampicin, a specific inhibitor of the
&subunit of bacterial RNA polymerase, blocks conversion of
MI3 single-stranded circular viral DNA (SS) to the duplex
replicative form (RF) in vi00 first led us to postulate a role for
RNA priming in DNA synthesis (1). Silverstein and Billen
(2) found that conversion of dX174 SS to RF was not sensi-
tive to rifampicin under similar conditions. Soluble extracts
of gently lysed Escherichia coli, like the intact cells, were
rifampicin-sensitive in the conversion of SS to RF of M13,
but not of 6x174 (3). Further studies (4, 5, 20) have shown
that replication of 4x174 SS is primed by a multienzyme
RNA synthetic system using many of the same factors re-
quired for host chromosome replication. Recently, reconstitu-
tion of M13 SS to RF has been achieved (6) with purified
DNA polymerase 111 holoenzyme (7), DNA unwinding pro-
tein (8, 9), and RNA polymerase I (holoenzyme). These
enzymes resemble cells and crude extracts in producing M13
RF with a full-length linear complementary strand and a
unique gap (3, 6), but they have the nonphysiological prop
erty of catalyzing rifampicin-sensitive conversion of .$X SS
to RF.
A search for the enzymatic basis of M13 DNA-specific
priming has led us to purify a new form of RNA polymerase,
termed RNA polymerase 111, the subject of this communica-
tion. This enzyme is assayed by its ability, in the presence
of DNA unwinding protein, to prime DNA polymerase I11
holoenzyme replication of M13, but not 4X, DNA. RNA poly-
merase 111, purified approsimately 100-fold by ammonium
sulfate fractionation and gel filtration, is estimated to be
50% pure by sodium dodecyl sulfate-acrylamide gel analysis.

Abbreviations: SS, single-stranded circular viral DNA; RF,
double-atranded, circular viral DNA; 6X, bacteriophage 6x174.

It is separated from RNA polymerase I during gel filtration.
In addition to its selectivity in priming DNA synthesis, it is
distinguished from RNA polymerases I and I1 by the remark-
able property of being inactive in transcription of duplex
DNA under RNA polymerase I assay conditions. Upon ex-
posure to rifampicin, a small factor is released from RNA poly-
merase 111 which renders RNA polymerase I (holoenzyme)
both template-specific in priming viral DNA synthesis and in-
active in transcribing duplex DNA. These studies raise
questions about the physiological role of RNA polymerase 111,
its structure and interconversion with RNA polymerase I,
and the role of each in the RNA synthetic events of the cell.

MATERIALS AND METHODS

Materials were from previously described sources (7). E.
coli B, harvested */, through the logarithmic phase of growth
on rich medium, was purchased from Grain Processing Corp.,
Muscatine, Iowa.

Enzymes. DNA polymerase 111 holoenzyme was prepared
as described (7). DNA unwinding protein was prepared ac-
cording to Weiner et al. (9). We thank Dr. Michael Chamber-
lin for gifts of RNA polymerase I holoenzyme and core (IO)
as well as for rifampicin-resistant holoenzyme. tinless other-
wise noted, RNA polymerase I was prepared by a minor
modification of the method of 13abinet (11).

Templates. M13 and $XI74 SS were prepared as described
previously (12). Salmon sperm DNA was purchased from
Sigma. Phage T4 and X DNAs were the generous gifts of Dr.
J. Thorner and Dr. J. M. Syvanen of this department.

Buflers. Buffer I contained 20% glycerol, 0.05 M Tris . HCI
(pH 7.5), 1 mM dithiothreitol, 10 mM MgCIz, 1 mM EDTA,
and 0.15 R.1 ammonium acetate. Buffer I1 contained 30%
sucrose, 0.02 31 Tris.HCI (pH 7.5), 1 mM dithiothreitol, 2
mM MgCI?, 0.1 mM EDTA, and 0.25 M ammonium acetate.
Buffer I11 was the same as buffer I1 but with only 0.04 M
ammonium acetate.

Assays of RNA Polymerase. DN.4 synthesis on M13 or
#X SS was measured in a final volume of 25 pl containing:
5p1 of assay buffer (10% sucrose, 0.05 M TrisSHCI (pH 7.5),
20 mM clithiothreitol, 0.05 M NaCI, 0.2 mg/ml of bovine
serum albumin), 3 pl of deoxynucleoside triphosphate mix-
ture (150 pM [a-3*P]dCTP (400 cpm/pmole); 400 pM each
of dATP, dGTP, and dTTP; 40 mM MgCl?], 1 pl of ribo-
nucleoside triphosphates (25 mal ATP, 5 mM each of GTP,
CTP, and UTI'), 2.5 p1 of DN.4 (500 pmoles of nucleotide),
3.5 p1 of water, and 5 p1 of DNA unwinding protein (1.8 pg).
These components (20.0 pl) aere mixed and incubated for 2
min at 30'. Water, RNA polymerase, and DNA polymerase

4425

4426 Biochemistry: Wickner and Kornberg

Proc. Nat. Acad. Sci. USA 71 (1974)

TABLE 1.

Purification of RNA polynerase III

                    Specific
              Total activity,
         Total protein, units/mg
Fraction units mg of protein

I. Extract 90 11,Ooo 0.008
11. Ammonium sulfate I 74 595 0.12
111. Ammonium sulfate I1 30 90 0.33
IV. Gel filtration I 27 60 0.45
V. Gel filtration I1 12 17 0.71

I11 holoenzyme (0.05 unit) were then added to chilled tubes
to a final volume of 25 pl. After a 10-min incubation at 30",
acid-insoluble nucleotide was determined (13). One unit of
RNA polymerase I11 primes 1 nmole of deoxynucleotide in-
corporation into M13 RF per rnin at 30'.
RNA synthesis was measured in a volume of 50 pl con-
taining: 0.05 M Tris-HCl (pH 7.5), 0.4 mg/ml of bovine
serum albumin, 4 mM Zmercaptoethanol, 15 mM MgCl,, 1
mM ['HIATP (10' cpm/pmole), 1 mM each of GTP, CTP,
and UTP, and 25 pgjml of DNA template. Acid-insoluble
nucleotide was measured after 10 min at 30".

RESULTS

Purification of RNA Polymerase III. All operations were at

Preparation of Extract. Two hundred grams of frozen E.
coli 13 cell paste (see Materials and Methods) was broken into
small pieces and placed in a 1-liter Waring Blendor. Lysis
buffer [lo% sucrose, 0.05 M Tris.HC1 (pH 7,5), 0.2 M am-
monium sulfate, and 10 mM spermidine-HCl] was added to
the suspension to a volunie of one liter and the cells were
suspended with three full-power bursts. Lysozyme (100 mg)
was added and the suspension was immediately transferred
to five 250-ml plastic bottles placed on ice. After 30 rnin the
bottles were transferred to a 37" bath for 5 inin and centri-
fuged at 0" for 60 min at 12,000 rpm in the Sorvall GSA rotor.
The clear amber supernatant is Fraction I (755 ml).

0-4'. A summary of the purification is in Table 1.

Ammonium Suljafe I. Solid ammonium sulfate (0.3 g/ml
of Fraction 1) was added with rapid stirring. After 20 min, the
suspension was centrifuged for 10 mill at 12,000 rpm in the
GSA rotor. Pellets were successively suspended in one-fifth
the volume of Fraction I in buffer I plus: (a) 0.30 g (b) 0.24 g
(c) 0.22 g (d) 0.20 g (e) 0.18 g and (f) 0.16 g of ammonium sul-
fate per ml. Suspensions were performed with glassTeflon
homogenizers and were followed at once by centrifugation for
10 min at 12,000 rpm in the GSA rotor. The supernatant
from each centrifugation was mixed with an equal volume
of saturated, neutralized ammonium sulfate and, after 10
min, centrifuged as before. Each precipitate was dissolved
in 10 ml of buffer I1 and fractions of high specific activity
(usually d-f) were pooled: Fraction I1 (44 ml).

Amnionium Sulfate II. Fraction I1 was mixed with an equal
volume of saturated ammonium sulfate and centrifuged after
10 min. The centrifugation here and subsequently was for 5
rnin at 16,000 rpm in the Sorvall SS34 rotor. Pellets were
successively suspended in one-fifth the volume of Fraction I
in buffer I1 plus: (a) 0.24 g (6) 0.22 g (c) 0.20 g (d) 0.18 g and
(e) 0.16 g of ammonium sulfate per ml. After each centrifuga-

MI3 DNA synthesis

-

n
0

-

I 5 200

Y U

0 Y

U 3 C

Y
3

.-

-

-

-

n

-

C
I U
U

2 loo

._

.-

a

I5 20 25 50 35
         Froction

MI3 DNA synthesis

-

n
0

-

I 5 200

Y U

0 Y

U 3 C

Y
3

.-

-

-

-

n

-

C
I U
U

2 loo

._

.-

a

I5 20 25 50 35
         Froction

FIG. 1.

     Gel filtration. Fraction I11 enzyme was applied to a
Bio-Gel A-5m column (240 ml, see ResuZts) and 6-ml fractions were
collected. Fractions 25 to 33 were used for further purification of
RNA polymerase 111.

tion, the supernatants were precipitated with saturated am-
monium sulfate as before. Each precipitate was dissolved in
5 ml of buffer 111; fractions of high specific activity (usually
c and d) were pooled: Fraction 111 (12 ml).

Gel Filtration I. Fraction I11 was applied to a Bio-Gel A-5m
column (100-200 mesh, 240 ml) equilibrated with buffer 11.
Fractions that primed DNA synthesis on M13, but not 4X-
174, SS (Fig. 1) were pooled: Fraction IV (60 ml).

Gel Filtration II, Fraction IV was concentrated by addition
of 0.3 g of solid ammonium sulfate per ml. After 30 min, the
suspension was centrifuged and the precipitate was dissolved
with 1.5 ml of buffer I11 (1.5 ml) and applied to a Bio-Gel
A-5m column (100-200 mesh, 60 ml) equilibrated with buffer
111. Fractions of peak RNA polymerase I11 specific activity
were pooled: Fraction V (14 ml). This fraction was stable for
at least 2 weeks at 0".

Requirements for Specijc Priming of M1S DNA Replication.
Discrimination by RNA polymerase between M13 and 4X
DNA for priming DNA synthesis requires two factors: (i)
the novel form of RNA polymerase, designated RNA poly-
merase I11 to distinguish it from the classic holoenzyme,
called RNA polymerase I, and (ii) DNA unwinding protein
in an amount sufficient to coat the single-stranded DNA
(Table 2). The unwinding protein markedly stimulates the
RNA polymerase I-prinied reaction (6) but does so for both
M13 and 4X DNA replication; the effect on the RNA poly-
merase 111-primed reaction is to suppress specifically the
4X DNA replication.

RNA polymerase III has feeble transcriptional actirily. RNA
polymerase I11 was 15- to 70-fold less active than RNA4 poly-
merase I in transcription of standard templates, such as A,
T4, and salmon sperm DNA, and poly[d(h-T)] (Table 3).
When the two polymerases were present in the same incuba-
tion, the RNA polymerase I remained active, thus indicating
the absence of a diffusible inhibitor to account for the in-
activity of RN.4 polymerase 111.

Proc. Nat. Acad. Sei. USA 71 (1974)

RNA Polymerase 111 4427

TABLE 2.

Requirements for specijk priming of MI3

Template for
DNA synthesis
(pmoles)

Form of

RNA polymerase Unwinding protein M13 @X

I
I11
I
I11

- 13 15
- 225 107
+ 232 122
+ 188 6

DNA synthesis on M13 or @X SS was assayed with 0.01 unit
of RNA polymerase I11 or RNA polymerase I (Matmials and
Methods). DNA unwinding protein was addedi where indicated.

The properties of RNA polymerase I observed with the
preparation obtained by the nabinet procedure were also
observed with another preparation prepared by M. Cham-
berlin. RNA polymerase I holoenzyme primed 3113 and $X
replication equally well, while the core enzyme was totally
inactive (Table 4), in contrast to the behavior of RN.4 poly-
merase 111, which primed M13, but not @X, replication and
was inactive in transcription of poly [d(A-T) 1. Rifampicin at 5
pg/ml of completely inhibited RNA polymerase 111, as it is
known to do to RNA polymerase I.

Rifampicin Releases a Diffusible Faclor from RNA Poly-
merase IIZ. RNA polymerase 111 treated with rifampicin and
then precipitated with ammonium sulfate to separate pro-
teins from free, unbound rifampicin was inactive in priming
DNA, but when mixed with RNA polymerase I prevented it
from priming $X SS while still permitting it to prime M13.
This result suggested that rifampicin released a factor from
RNA polymerase 111 that could confer the M13-specificity
property on RNA polymerase I. This diffusible factor, re-
leased by rifampicin, nas isolated by Sephades G-150 gel
filtration (Fig. 2), using as an assay its inhibition of the prim-
ing of $X DNA synthesis by RNA polymerase I. Little or
none of this factor activity was seen on gel filtration in a con-
trol experiment with RNA polyinerase 111 that had not been
exposed to rifampicin. As judged by gel filtration, the fac-
tor is smaller than hemoglobin, but its size and other prop-
erties remain to be determined.
The rifampicin-induced release of a diffusible factor was
also studied with the aid of a rifampicin-resistant RNA poly-
merase I (I-rifR) (Table 5). When I-rifR was mixed with RNA
polymerase I11 in the absence of rifampicin (Esps. 1,3, and
5), RNA-primed DNA synthesis and transcription showed ap-

TABLE 3.

Transcription oj duplez DNA

RNA synthesis (pmol) on: DNAsynthesis

                       (pmol) on:
polymerase sperm T4 X [d(A-T)] M13 @X

I 62 84 146 132 56 41

I11 4 32 8 49 3
I + I11 63 50 138

Form of RNA Salmon poly-

As described in Materials and Mdhods, 0.005 unit (defined for
priming of DNA synthesis) of RNA polymerase I (1 pg) or RNA
polymerase I11 (2 fig) was added. In some assays, 0.005 unit of
each polymerase was mixed and added together to the assay.

TABLE 4.

Comparison oj jom oj RNA polymerase I and III

in priming of DNA synthmis and in transcription

DNA
synthesis
(pmol) on:

RNA polymerase M13 @X

I-holoenzyme (Chambedin*)   12 10
I-holoenzyme [Babinet (ll)]    38 36
I-core (Chamberlin')         0 0
I11                    66 5
I11 + rifampicin            0 0

Form of

RNA Bynthesis
(pmol) on
poly [d(A-T)I

13
83
10
2
0

Holoenzyme and core RNA polymerase I, RNA polymerase I
prepared by the method of Babinet (ll), and Fraction IV RNA
polymerase I11 (0.007 unit) were assayed as described in Ma-
tetials and Melhods. Where indicated, the enzyme was incubated
for 5 min at 0' with 6 a/ml of rifampicin before addition to the
assay.
* Holoenzyme RNA polymerase I has been purified by an un-
published procedure employing phosphocellulose chromatography
(Prof. M. Chamberlin and Dr. N. Gonzalez, University of Cali-
fornia, Berkeley).

proximately additive results. In contrast, when rifampicin-
treated RNA polymerase I11 was mixed with RNA poly-
merase I (Exps. 2, 4, and 6), the priming of $X replication
was depressed whereas that of hI13 was stimulated: RNA

void
volume

SAMRE A:

60 -

40-

20-

6 8 IO 12 14 16 18 2l

Fraction

FIQ. 2.

     Isolation of a factor released from RNA polymerase
I11 by rifampicin. Samples of RNA polymerase I11 (100 pg in
0.1 ml buffer 111) were incubated for 5 min at 0" (A) with rifam-
picin (2 pg) or (B) without the drug. Each sample was then
mixed with 100 fil of saturated ammonium sulfate. After 10 min
at Oo, the precipitates were collected by centrifugation (OO, 5
min, 16,000 rpm in the Sorvall SE12 rotor). Each precipitate
was suspended twice in 3 ml of a mixture of equal volumes of
buffer I and saturated ammonium sulfate and collected by
centrifugation (as above). Rifampicin-treated (sample A) and
untreated RNA polymerase I11 (sample B), dissolved in 100 pl
buffer I were each filtered over a Sephadex G-150 column (0.6 X
8 cm, equilibrated with buffer I at 4'). Aliquots (3 pl) of the
1Wp1 fractions were assayed for their ability to inhibit RNA
polymerase I-primed CpX SS -+ RF. Incorporation of 20 pmoles
of deoxynucleotide into the acid-insoluble fraction was observed
in the uninhibited reaction.

4428 Biochemistry: Wickner and Kornberg

Proc. Nat. Acad. Sci. USA 71 (2974)

TABLE 5.  Rifampicin release of a diflusibk factor from RNA

polynerase 111

Form of

~.

Experi- RNA

ment polymerase Rifampicin M13 +X T4

RNA

DNA
synthesis
(pmol) on: synthesis

(pmol ) on

1 I11           - 55     5 0
2 I11          + 0     0 0

                      11 8 11
3 I-rifR -
4 I-rifR + 8 11 10
5 I-rifR + 111 - 96 17 9
6 I-rifR + I11 + 36 2 2

Rifampicin-resistant RNA polymerase I (I-rifn, 1 pg) and
RNA polymerase I11 (2 pg, rifampicin-sensitive) were assayed
as described in M~criuls and Methods. Where indicated, the
enzymes were incubated for 5 min at 0" with 2 pg of rifampicin
per ml.

synthesis on a T4 duplex DNA template was likewise de-
presed upon mising of the two rifampicin-treated enzymes.
In sum, rifampicin, known to act on the p subunit, appears
to release a small factor from RNA polymerase 111, which can
cause RNA polymerase I to discriminate between M13 and
+X SS and to become relatively incrt in transcript.ion of
duplei DNA.

              DISCUSSION
The capacity of cell extracts to discriminate in the replication
of M13 and #X single-stranded DNA in a rifampioiii-sensitive
reactioi! was not maintained in a purified RNA polymerase-
multienzyme system. Our efforts to restore discriminatory
ability to the purified system have led us to a new form of
RNA polymerase, termed RNA polymerase 111. This enzyme
is physically separable from RNA polymerase I (Fig. 1) and
is further recognized by (i) its selective abilit,y to prime the
replication of M13, but not #X DNA, and (ii) its inability to
transcribe standard duples DNA. The close relation of RNA
polymerase 111 to RNA polymerase I is shown by its sensi-
tivity to rifampicin. Upon exposure to rifampicin, RNA poly-
merase I11 releases a factor, presumably a low-molecular-
weight subunit, that inhibits RNA polymerase I transcription
and confers the aLility to prime M1.3 hut not #X replicat.ion.
This factor may be the major difference between RNA poly-
merases I and 111.

Modification of RNA polynierase functions by addition or
substitution of subunits has been shown in phage T3 and T4
infections of E. coli (14, 15) and in SP82 infection of B. szrhtilis
(16). Chao and Speyer (17) found that E. coli RNA poly-
merase is chaiiged in chromatographic properties and teni-
plate specificity \vheii cells ent,er a stat'ionary phase. How-
ever, this form of RNA polymerase is not separated from
REA polymerase I by gel filtration and, while inactive ill
t,ranscriptioii of natural duples DNA, it, is still quite active
on poly[d(A-T)]. In contrast, RNA polymerase I11 has been
obtained from cells growing esponentially, is separable from

RNA polymerase I by gel filtration, and is unable to tran-
scribe poly [d(A-T) 1. Ishihama and coworkers (18) identified
yet another form of the enzyme, which they called RNA
polymerase 11. It contains a different u subunit with a molecu-
lar weight of 56,000 rather than 90,OOO. Yet RNA poly-
merases I and I1 were not distinguishable in their capacities
to transcribe duplex DNA.
The ability of RNA polymerase 111 to maintain the cellular
discrimination between M13 and #X and its abundance in
extracts of gently lysed cells suggest that RNA polymerase
111 may constitute a significant fraction of the cellular RNA
polymerase. The complete inability of cell estracts to prime
#X replication with rifampicin-sensitive enzymes points to a
predominance of RNA polymerase I11 over RNA polymerase
I or else to a specific repressor of the latter in priming the
replication of +X single-stranded DNA (19). These studies
raise questions that can be answered only by purification of
RNA polymerase 111 to homogeneity so that the factor
releasable by rifampicin can be characterized and its functioiis
in regulating transcriptional activity properly assessed.

This work WBS supported by grants from the National Insti-
tutes of Health and the National Science Foundation. W.W. is a
Fellow of the Mellon Foundation.

1.

2.

3.


4.

5.

6.

7.

8.


9.

10.

11.

12.

13.


14.


15.

16.


17.


18.


19.

20.

Brutlag, D., Schekman, R.. & Kornberg, A. (1971) Proc.
Nut. Acad. Sei. USA 68, 2826-2829.
Silverstein, S. & Billen, D. (1971j Bwchim. Biophys. Acta

Wickner, W., Brutlag, D., Schekman, R.. & Kornberg, A.
(1972) Proc. Nut. Acad. Sci. USA 69, 965-969.
Schekman, It., Wickner, W., Westergaard, O., Brutlag, I).,
Geider, K., Bertsch, L. L. & Kornberg, A. (1972) Proc. Nut.

Wickner, R., Wright, M., Wickner, S. & Hurwita, J. (1972)
Proc. Nat. Ad. Sca. USA 69, 3233-3237.
Geider, K. & Kornberg, A. (1974) J. Riol. Chem. 249,3999-
4005.
Wickner, W. & Kornberg, A. (ld'4) J. Bid. Chem., in press.
Sigd, N., Ilelius, H., Kornberg, T., Gefter, M. & Alberts, B.
(1972) Proc. Nat. Acad. Sci. USA 69, 3537-3541.
Weiner, J. A., Bertsch, L. L. & Kornberg, A. (1974) J. Biol.
Chem., in press.
Berg, D., Barrett, K. & Chamberlin, M. (1971) in Methods
in Enzymology, eds. Grossman, L. & Moldave, K. (Academic
Press, New York), Vol. 21, pp. 506-519.
Babinet, C. (1967) Bwchem. Biophys. Res. Commun. 26,

Wickner, W., Schekman, It., Geider, K. 6 Kornberg, A.
(1973) Proc. Nat. Acad. Sci. USA 70, 1764-1767.
Jovin, T. M., Englund, P. T. & Kornberg, A. (1969) J. Biol.
Chem. 244,2996-3008.
Ilharmgrongartama, B., Mahadik, S. P. & Srinivasan, P. It.
(1973) Proc. Nat. Acad. Sci. USA 70, 2845-2849.
Horvitz, H. R.. (1973) Nature New Diol. 244, 137-140.
Spiegelman, G. B. & Whiteley, H. R. (1974) J. Rid. Chem.
249, 1476-1482.
Chao, L. & Speyer, J. F. (1973) Hiochem. Biophys. Res.
Commun. 51, 39940.5.
Fukuda, R., Iwakura, Y. & Ishihama, A. (1974) J. Mol.
Bid. 83, 353-367.
Hurwitz, J., Wickner, S. & Wright, M. (1973) Biochem.
Biophys. Res. Comm. 51, 237-267.
Schekman, R., Weiner, A. & Kornberg, A. (1974) Science,
in press.

247, 383-390.

Aud. Sci. USA 69,2691-2695.

639-644.