Update

TRENDS in Biochemical Sciences

Vol.29 No.4 April 2004

| Research Focus

New inhibitors targeting bacterial RNA polymerase

Seth A. Darst
The Rockefeller University, Box 224, 1230 York Avenue, New York, NY 10021, USA

The bacterial RNA polymerase is a proven target for

antimicrobials because it is inhibited by rifampicin, a
potent antibiotic that is a key component of tuberculosis therapy. The bacterial RNA polymerase is also a molecular machine that presents numerous targets for
small-molecule inhibitors that have not yet been
exploited. Recent work has uncovered new classes of
inhibitors. Further applications of high-throughput
screening, medicinal chemistry and high-resolution
structural studies hold great promise for the development of antimicrobials.
DNA-dependent RNA polymerase (RNAP) the central
enzyme of bacterial gene expression is a complex
molecular machine [1,2]. The catalytically competent
core RNAP has subunit composition a2bb0 v and a total
molecular weight near 400 kDa. In contrast to the
ribosome, surprisingly few small-molecule inhibitors
that are specific for bacterial RNAPs have been described,
despite their potential use as tools to investigate RNAP
mechanism and for antibacterial drug targets. Only one
inhibitor rifampicin (Rif; Figure 1) and its derivatives
has reached medical use [3,4] as a key component of antituberculosis therapy. Thus, recent work describing new
classes of bacterial RNAP inhibitors has generated widespread interest [5 7].
RNAP structure
Structural studies of bacterial and eukaryotic RNAPs,
combined with biochemical, biophysical and genetic data,
have provided a basic picture of the functional elongation
complex that contains RNAP, DNA template and RNA
transcript, and have begun to tease out details of the
structure function relationship for the enzyme [8 11]. In
the bacterial enzyme, the large b and b0 subunits form the
pincers of the crab-claw-shaped molecule (Figure 2a). A
large channel between the pincers holds the RNA 30 -OH
within the active site, an 8 9-base-pair RNA DNA hybrid
at the growing end of the transcript, at least ten base pairs
of duplex DNA downstream of the hybrid and , 6
nucleotides of single-stranded RNA upstream of the
hybrid. A smaller, secondary channel leads from the
surface of the enzyme directly to the active site and
might serve as a passageway for incoming nucleotide
triphosphate substrates (NTPs) because access to the
active site from the main channel is blocked by the nucleic
acids (Figure 2). Steps in the cycle of NTP entry, catalysis
and translocation are thought to require movements of
structures surrounding the active site. Much attention has
Corresponding author: Seth A. Darst ( ([email protected]).
www.sciencedirect.com

been focused on a long a helix that spans the main channel

near the active site, which is called the bridge helix [9,12]
because it bridges the two pincers of the crab claw
(Figure 2). It is easy to imagine small-molecule inhibitors
interfering with the workings of this molecular machine in
a variety of ways.
Established bacterial RNAP inhibitors
The inhibitor Rif (Figure 1) is one of the most potent and
broad-spectrum antibiotics against bacterial pathogens
and is a key component of anti-tuberculosis therapy [3,4].
It binds in a pocket of the RNAP b subunit deep within the
away from the active site
active-site channel, but . 12 A
[13] (Figure 2). The Rif-binding pocket is highly conserved
among bacterial RNAPs, but not between bacterial and
eukaryotic RNAPs. Rif acts by directly blocking the path
of the elongating RNA when the transcript becomes twoto-three nucleotides in length [13,14].
Streptolydigin (Stl) [15] (Figure 1) is effective against a
wide range of bacterial RNAPs in vitro but is not medically
important because (i) the Ki for Stl is in the 10 100 mM
range, unlike the nM range for Rif, and (ii) bacterial
membranes are generally not permeable to Stl. Stl binds
near the active site, interacting directly with the bridging
helix and nearby structures [16,17]. Stl might act by
interfering with crucial conformational changes of the
bridge helix during the nucleotide addition cycle (noncompetitive allosteric inhibition), as a-amanitin is thought
to inhibit eukaryotic RNAP II [18].
New bacterial RNAP inhibitors
Microcin J25
Microcin J25 (MccJ25) is a 21-residue, ribosomally
synthesized peptide [19,20] that is produced by strains of
Escherichia coli harboring a plasmid-borne synthesis,
maturation and export system [21]. MccJ25 exhibits
bacteriocidal activity against a range of Gram-negative
bacterial species [20]. Recently, three groups simultaneously reported that the mature peptide has an
extraordinary structural fold, with a cyclic segment
(owing to an isopeptide bond between the side chain of
Glu8 and the N terminus) and a 13-residue linear tail that
loops back and threads through the eight-residue ring
[22 24]. Two aromatic residues near the C terminus of the
tail straddle each side of the ring, sterically trapping the
tail in a noncovalent interaction, resulting in a stable,
relatively rigid fold (Figure 1). Genetic results have
indicated that the in vivo target of MccJ25 is the RNAP
[6], and this was recently confirmed in vitro [7]. Amino acid
substitutions that confer resistance to MccJ25 have been
isolated in both the RNAP b and b0 subunits [6,7]. In the

Update

160

(a)

CH3

TRENDS in Biochemical Sciences

inhibition [7]. Other inhibition mechanisms are not ruled

out, however, and it is interesting to note that regions
harboring substitutions in the RNAP b0 subunit conferring
MccJ25 resistance overlap those conferring resistance to
Stl [7]. This raises the possibility that MccJ25 acts through
a similar mechanism by interfering with conformational
changes near the RNAP active site, perhaps in the
bridge helix.

CH3

CH3

OOH3C
OH3C

OH
CH3

OH
OH

O
OH

CH3

CH3

NH

CH3

(b)

N
N
(+)

OH

CH3

CH3
O

O
OH

CH3

CH3

CH3
O

CHCONHCH3
CH3

H3C
OH

(c)

OH
N
CF3
NH

(d)

Ti BS

Figure 1. Structures of bacterial RNA polymerase inhibitors (a) rifampicin (Rif)

[3,4], (b) streptolydigin (Stl) [15,26], (c) CBR703 [5] and (d) microcin J25 (MccJ25)
[6,7,22 24].

RNAP structure, these cluster in the inner surface of the

secondary channel (Figure 2). Binding of MccJ25 within
this channel could block nucleotide substrates from
entering the enzyme active site, consistent with modeling
studies and with biochemical results on the mechanism of
www.sciencedirect.com

Vol.29 No.4 April 2004

CBR703
Most recently, a novel class of bacterial RNAP inhibitors
was identified by screening a large library of compounds
for inhibition of E. coli RNAP transcription [5]. The
progenitor compound (N-hydroxy-N0 -phenyl-3-trifluoromethyl-benzamidine; designated CBR703; Figure 1) inhibited E. coli RNAP transcription in an in vitro assay, with a
half-maximal effect (IC50) at 10 mM. Nevertheless, antibacterial assays revealed little or no activity against
Gram-negative or Gram-positive species. However, E. coli
tolC mutant strains were found to be sensitive, suggesting
that the lack of antibacterial activity against wild-type
E. coli strains was due to specific efflux from the cells
(TolC plays a major part in efflux of toxic chemicals from
Gram-negative bacteria). Variants of CBR703 were synthesized that exhibited increased potency [IC50 and
maximal inhibitory concentration (MIC) against E. coli
tolC mutant strains , 1 mM]. The most potent compounds
displayed antibacterial activity even against wild-type
E. coli strains.
The in vivo target of the CBR703 class of compounds
was confirmed by isolating mutant E. coli RNAPs with
single amino acid substitutions conferring resistance. The
positions where amino acid substitutions confer
resistance to CBR703 cluster on the surface of a groove
at the junction of the b0 bridge helix and the b subunit,
defining the putative binding site on the RNAP (Figure 2).
The location of the binding site that is adjacent to the
bridge helix, combined with a series of biochemical studies,
led Artsimovitch et al. [5] to suggest an inhibition
mechanism in which CBR703 interferes with conformational changes of the bridge helix required during the
nucleotide addition cycle
Concluding remarks
The bacterial RNAP contains multiple channels and
grooves for holding downstream double-stranded DNA,
the non-template DNA strand in the transcription bubble,
the RNA DNA hybrid and the upstream single-stranded
RNA (Figure 2). Yet another channel enables entry of
nucleotide substrates into the active site. Rif acts by
blocking the path of the RNA in the RNA DNA hybrid,
and MccJ25 might act by blocking the nucleotide entry
channel. Small molecules blocking any of the other
channels would probably interfere with RNAP function
as well.
The RNAP is thought to undergo conformational
changes at each step of the transcription cycle, such as
during open-complex formation, during the nucleotideaddition and -translocation cycle, and during termination.
Small-molecule inhibitors could interfere with these
steps allosterically, as Stl, CBR703 and possibly MccJ25

Update

TRENDS in Biochemical Sciences

Vol.29 No.4 April 2004

161

Figure 2. Model of the bacterial RNA polymerase (RNAP) elongation complex and location of inhibitor binding sites, generated using DINO (http://cobra.mih.unibas.ch/
dino/). Part (a) illustrates the inhibitor-binding sites within the context of the entire RNAP elongation complex, whereas part (b) focuses on the protein regions involved in
inhibitor binding. (a) Model of the bacterial RNAP elongation complex [9,11,27]. The RNAP is shown as a molecular surface, color coding as follows: b0 subunit, in pink; b
subunit, blue; a subunits; v, grey. Nucleic acids are shown as phosphate-backbone worms, with the non-template DNA strand in light green; template strand DNA, dark
green; RNA transcript (barely visible), red. The locations of a-carbon backbone worms and Rif [13], as highlighted in (b), are visible within the semi-transparent RNAP. The
path of the secondary channel, which leads into the active site Mg2, is indicated by the large grey arrow. (b) Stereo view (/2 38) of RNAP active-site and inhibitor-binding
regions. The viewing angle is the same as in (a). Portions of the RNAP b (cyan) and b0 (pink) subunits relevant to the binding of inhibitors (Figure 1) are shown. The b0 bridge
helix is highlighted in magenta, the active site Mg2 ion is shown as a magenta sphere. The RNA DNA hybrid present in the transcribing RNAP [9,11] is shown as a phosphate backbone worm (template DNA, green; RNA, red). Rif (orange) is shown in its binding pocket [13], sterically clashing with the RNA. Colored spheres indicate the positions of amino acid substitutions that define binding sites for inhibitors: Stl, red [16,17]; CBR703 (N-hydroxy-N0 -phenyl-3-trifluoromethyl-benzamidine); green [5]; MccJ25
(Microcin J25), blue [6,7]. The path of the secondary channel (sc) into the active site is indicated by the grey arrow.

are thought to interfere with the nucleotide-addition or

-translocation cycle.
Artsimovitch et al. [5] applied modern high-throughputscreening methods to isolate a heretofore undiscovered
class of bacterial RNAP inhibitors, and systematic
variation of the CBR703 structure improved the potency
more than tenfold. Continued variations of the CBR703
structure informed by high-resolution structural studies
holds great promise for the rational development of
antimicrobials, as does application of these methods to
find other new inhibitors (that are surely out there)
specific for the many other targets on the RNAP.
In the clinical setting, the effectiveness of otherwise
ideal compounds, such as Rif, is compromised by the
frequent occurrence of single amino acid substitutions in
the RNAP that confer resistance to the drug [25]. Single
amino acid substitutions also confer resistance to Stl,
CBR703 and MccJ25 (Figure 2). In this regard, the studies
of new inhibitors, combined with high-resolution structural studies, could potentially lead to the design of
chimeric compounds targeting more than one site on
RNAP simultaneously.
Acknowledgements
S.A.D. is supported by NIH RO1 grants GM53759 and GM61898.

References
1 Darst, S.A. (2001) Bacterial RNA polymerase. Curr. Opin. Struct. Biol.
11, 155 162
2 Murakami, K. and Darst, S.A. (2003) Bacterial RNA polymerases: the
wholo story. Curr. Opin. Struct. Biol. 13, 31 39
3 Sensi, P. et al. (1960) Isolation and properties of rifomycin B and
rifomycin complex. Antibiotics Ann. 1, 262 270
4 Sensi, P. (1983) History of the development of rifampin. Rev. Infect. Dis.
5 (Suppl. 3), 402 406
www.sciencedirect.com

5 Artsimovich, I. et al. (2003) A new class of bacterial RNA polymerase

inhibitor affects nucleotide addition. Science 302, 650 654
6 Delgado, M.A. et al. (2001) Escherichia coli RNA polymerase is the
target of the cyclopeptide antibiotic microcin J25. J. Bacteriol. 183,
4543 4550
7 Yuzenkova, J. et al. (2002) Mutations of bacterial RNA polymerase
leading to resistance to microcin J25. J. Biol. Chem. 277,
50867 50875
8 Zhang, G. et al. (1999) Crystal structure of Thermus aquaticus core
resolution. Cell 98, 811 824
RNA polymerase at 3.3 A
9 Gnatt, A.L. et al. (2001) Structural basis of transcription: an RNA
resolution. Science 292,
polymerase II elongation complex at 3.3 A
1876 1882
10 Cramer, P. et al. (2001) Structural basis of transcription: RNA
resolution. Science 292, 1863 1876
polymerase II at 2.8 A
11 Korzheva, N. et al. (2000) A structural model of transcription
elongation. Science 289, 619 625
12 Epshtein, V. et al. (2002) Swing-gate model of nucleotide entry into the
RNA polymerase active center. Mol. Cell 10, 623 634
13 Campbell, E.A. et al. (2001) Structural mechanism for rifampicin
inhibition of bacterial RNA polymerase. Cell 104, 901 912
14 McClure, W.R. and Cech, C.L. (1978) On the mechanism of rifampicin
inhibition of RNA synthesis. J. Biol. Chem. 253, 8949 8956
15 Cassani, G. et al. (1971) Inhibition of RNA polymerase by streptolydigin. Nat. New Biol. 230, 197 200
16 Heisler, L.M. et al. (1993) Four contiguous amino acids define the
target for streptolydigin resistance in the b subunit of Escherichia coli
RNA polymerase. J. Biol. Chem. 268, 25369 25375
17 Severinov, K. et al. (1995) Streptolydigin-resistant mutants in an
evolutionarily conserved region of the b0 subunit of Escherichia coli
RNA polymerase. J. Biol. Chem. 270, 23926 23929
18 Bushnell, D.A. et al. (2002) Structural basis of transcription:
resolution. Proc.
a-amanitin-RNA polymerase II cocrystal at 2.8 A
Natl. Acad. Sci. U. S. A. 99, 1218 1222
19 Blond, A. et al. (1999) The cyclic structure of microcin J25, a 21-residue
peptide antibiotic from Escherichia coli. Eur. J. Biochem. 259,
747 755
20 Salomon, R.A. and Farias, R.N. (1992) Microcin-25, a novel antimicrobial peptide produced by Escherichia coli. J. Bacteriol. 174,
7428 7435
21 Solbiati, J.O. et al. (1999) Sequence analysis of the four plasmid genes

Update

162

TRENDS in Biochemical Sciences

required to produce the circular peptide antibiotic microcin J25.

J. Bacteriol. 181, 2659 2662
22 Wilson, K-A. et al. (2003) The structure of microcin J25, a peptide
inhibitor of bacterial RNA polymerase, is a lassoed tail. J. Am. Chem.
Soc. 125, 12475 12483
23 Rosengren, K.J. et al. (2003) Microcin J25 has a threaded sidechain-tobackbone ring structure and not a head-to-tail cyclized backbone.
J. Am. Chem. Soc. 125, 1246412474
24 Bayro, M.J. et al. (2003) Structure of the antibacterial peptide microcin
J25: A 21-residue lariat protoknot. J. Am. Chem. Soc. 125,
12382 12383

Vol.29 No.4 April 2004

25 Ramaswamy, S. and Musser, J.M. (1998) Molecular genetic basis of

antimicrobial agent resistance in Mycobacterium tuberculosis: 1998
update. Tuber. Lung Dis. 79, 3 29
26 Rinehart, K.L.J. et al. (1963) Structure of streptolydigin. Antimicrob.
Agents Chemother. 161, 346 348
27 Opalka, N. et al. (2003) Structure and function of the transcription
elongation factor GreB bound to bacterial RNA polymerase. Cell 114,
335 345
0968-0004/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tibs.2004.02.005

Prions: so many fibers, so little infectivity

Barnaby C.H. May1,2, Cedric Govaerts3, Stanley B. Prusiner1,2,4 and Fred E. Cohen1,3,4
1

Institute for Neurodegenerative Diseases, University of California, 513 Parnassus Avenue, HSE-774, San Francisco,
CA 94143-0518, USA
2
Department of Neurology, University of California, San Francisco, CA 94143, USA
3
Department of Cellular and Molecular Pharmacology, University of California, San Francisco, CA 94143, USA
4
Department of Biochemistry and Biophysics, University of California, 600 16th Street, San Francisco, CA 94143, USA

A detailed mechanistic or structural understanding of

the misfolding events that render the normal cellular
prion protein, PrPC, amyloidogenic and infectious has
been elusive. Many groups have observed that, under
certain conditions, PrPC has a tendency to form b-strandrich fibers. A recently published protocol for in vitro
conversion of PrPC shows that the formation of intermolecular disulfide bonds between monomers of
recombinant PrP might play a role in the folding and
fibrilization of this protein in vitro.
Prion diseases are a group of fatal neurodegenerative
disorders characterized by the accumulation of an aberrantly folded isoform of the prion protein (PrP), designated
PrPSc. Biochemical and transgenic studies of the transmission of prion diseases have led to the conclusion that
PrPSc is the essential and, probably, the sole component of
the infectious particle [1]. Animal- and cell-based models of
prion replication are established research tools for the
study of conversion from the a-helical cellular PrP (PrPC)
to b-strand-rich PrPSc. These in vivo systems have the
relevant cofactors to mimic accurately the biology of prion
replication, as evidenced by the accumulation of infectious
PrPSc. In the pursuit of simplified models of prion replication, several groups have attempted to develop in vitro
systems for prion conversion [2]. Typically, recombinant
PrP (recPrP), or PrPC derived from healthy animal brain,
is converted to the alternative b-strand-rich fold, often
seeded by the addition of PrPSc. Hence, different conversion systems might be grouped according to whether the
input PrP substrate is brain-derived or recombinant, and
also whether the conversion is seeded with PrPSc or
unseeded. To different extents, the conversion efficiency of
these approaches is dependent on denaturing conditions,
sonication cycles [3] and even the presence of RNA [4,5].
Corresponding author: Barnaby C.H. May ([email protected]).
www.sciencedirect.com

The products of these in vitro conversions are characterized by partial resistance to digestion with proteinase K
(PK) and high b-strand content two well-known features
of PrPSc. In vitro conversion assays have sought to mimic
important features of prion biology including the species
barrier and strain differences. Although some of these
features can be mimicked in vitro, significant controversy
remains. For example, in vitro conversion assays suggest
that a particular strain of sheep carrying arginine at the
polymorphic codon 171 should be resistant to infection
with scrapie prions [6]. However, Houston et al. [7] claim
that these ARR sheep can be infected with BSE prions
in vivo. Disagreement remains as these experiments
employed different prion strains. The efficiency of in vitro
conversion is primarily judged based on the production of
PK-resistant material, because it is assumed that PK
resistance is a marker for structural similarity to PrPSc.
However, an alternative interpretation for the observed
PK resistance, at least in the case of recPrP, would be that
the PK-resistant seed, often PrPSc, protects recPrP from
PK hydrolysis in the aggregated state. It has been shown
that PrPSc interacts directly with PrPC [8], possibly
protecting its PK cleavage site. Nucleic acids might play
a similar role. In this way, the observed PK resistance
might be the result of a protection assay and not a
conformational conversion assay. This might account for
the very modest rate of PrP conversion observed in in vitro
assays. Another shortcoming of in vitro conversion assays
has been that the protease-resistant core of in-vitro-refolded
molecules are not necessarily equivalent to PrPSc, because
they typically lack the a-helical structure observed by
Fourier transform infrared spectroscopy (FTIR) studies [9].
Furthermore, it is noteworthy that it has not been possible to
demonstrate infectivity of in-vitro-converted wild-type
recPrP [10]. These factors have led many to question the
biological relevance of these conversion products.