Inhibition of nuclear pre-mRNA splicing by antibiotics
in vitro
Maren Hertweck, Reinhard Hiller and Manfred W. Mueller
Vienna BioCenter, Institute of Microbiology and Genetics, Vienna, Austria
A number of antibiotics have been reported to disturb the
decoding process in prokaryotic translation and to inhibit
the function of various natural ribozymes. We i nvestigated
the eect of several antibiotics on in vitro splicing of a
eukaryotic nuclear pre-mRNA (b-globin). Of the eight
antibiotics studied, erythromycin, Cl-tetracycline and
streptomycin were identi®ed as splicing inhibitors in nuclear
HeLa cell extract. The K
i
values were 160, 180 and 230 l
M
,
respectively. Cl-tetracycline-mediated and streptomycin-
mediated splicing inhibition were in the molar inhibition
range for hammerhead and human hepatitis d elta virus
ribozyme self-cleavage (tetracycline), of group-I intron self-
splicing (streptomycin) a nd inhibition of RNase P cleavage
by some am inoglycosides. Cl-tetracycline a nd the amino-
cyclitol glycoside streptomycin were found to have an
indirect eect on splicing by unspeci®c binding to the
pre-mRNA, s uggesting that t he inhibition is the r esult of
disturbance of t he correct foldin g o f the pre-mRNA into the
splicing-compatible tertiary structure by the charged groups
of these antibiotics. The macrolide, erythromycin, the
strongest inhibitor, had only a slight eect on formation o f
the presplicing complexes A and B, but almost completely
inhibited formation of the splicing-active C c omplex by

factors required for catalysis have not yet occurred (e.g. U2/
U6 base pairing). Complex C represents the catalytically
active stage in which intron r emoval from pre-mRNA by a
two-step transesteri®cation reaction occurs.
The ®rst step of nuclear p re-mRNA splicing is g overned
by a nucleophilic attack of the 2¢ OH of the branch
adenosine at the 5¢ splice site. The 5¢ exon is released and the
splicing intermediate, t he lariat)3¢ exon, is for med. The
second step involves a nucleophilic attack of the terminal
3¢ OH of the 5¢ exon at the 3¢ splice site leading to the
formation of ligated exons and the excised intron RNA in
lariat form [8±10].
It has been demonstrated that several a ntibiotics act as
inhibitors of various biological RNA-catalyzed key pro-
cesses. The binding of antibiotics t o different functional
RNAs may result from r ecognition o f e lectrostatic comple -
mentary and evolutionarily conserved tertiary structure
motifs, which can lead to impairment or l oss of R NA
function [11,12]. Recently, two t ypes of aminoglycoside
antibiotic-binding sites on RNA have been proposed.
Type-I binding sites c onsist of a symmetric internal loops
as in the r ibosome-decoding site, a t which the a minoglyco-
sides induce slight distortion of the RNA structure and
interfere with the binding o f the functional sub strate.
Type-II binding sites are the central metal ion-binding pock-
ets i n the catalytic cores of ribozymes. In type-II binding
sites the aminoglycosides act by displacing es sential bivalent
metal ions via their positively c harged groups [13,14].
The best known example of an inhibitory actio n by
antibiotics in vivo is the inhibition of prokaryotic protein

cleaving hammerhead a nd h airpin r ibozymes, t he human
hepatitis delta virus and HIV-1 ribozymes, and tRNA
processing RNase P RNAs in vitro [11,31±35]
(M. Hertweck, R. Hiller & M.W. Mueller, unpublished
work)
2
. However, in vitro inhibition of nuclear pre-mRNA
splicing by any antibiotic has not been described s o far.
Here, we report on the inhibitory actions of erythromy-
cin, Cl-tetracycline a nd streptomycin on nuclear splicing of
the second b-globin intron in nuclear HeLa cell extract.
Denaturing and native g el a nalyses provided information
about the nature of the splicing inhibition, especially on
which component o f the nuclear splicing reaction, the pre-
mRNA or the nuclear extract containing the snRNAs and
associated splicing factors, the antibiotics exert t heir action.
Of the antibiotics tested, only the macrolide erythromycin
was fou nd to have a direct effect on splic ing. Formation of
the spliceosomal complexes A and B was slightly reduced
and formation of complex C was completely inhibited by
erythromycin, resulting in speci®c inhibition of the second
step of pre-mRNA splicing. This ®nding, together w ith
recent data on macrolide-mediated inh ibition of eukaryotic
pre-mRNA transcription factor activation in vivo,suggest
that erythromycin interferes with at least two major stages
of eukaryotic gene expression, transcription i nitiation and
pre-mRNA splicing.
MATERIALS AND METHODS
Plasmid and oligonucleotides
The pre-mRNA b-globin, derived from the second b-globin

NaCl, 30 m
M
dithiothreitol and
250 U T3 RNA polymerase. The RNA was internally
labeled b y t he addition of [a-
32
P]UTP to t he in v itro
transcription a ssay. Freshly transcribed pre-mRNA was
puri®ed by electrophoretic separation on a denaturing 5%
polyacrylamide (30 : 1) 8
M
urea gel
3
and eluted from the gel
after UV-shadowing.
In vitro
splicing assays
Standard splicing of the b-globin pre-mRNA in HeLa
nuclear extract was per formed at 30 °C for 60 min. Splicing
reactions were prepared in a 10-lL volume u sing the
following standard conditions: 40% (v/v) H eLa nuclear
Fig. 1. Struc tural drawings of the antibiotics
investigated in nuclear splicing i nhibition. (A)
The macrolide e rythromycin containing a
14-membered m acrolide rin g with a t ypical
lactone group (framed) and two sugars,
D
-desosamine (top) and
L
-cladinose (bottom).

were analyzed on a Molecular D ynamics P hosphorImager.
The sum of the four fractions, lariat)3¢ exon, excised lariat,
ligated exons and pre-mRNA, was de®ned as 100% educt
plus product.
For a nalysis of spliceosomal complexes, splicing reaction
mixtures were loaded directly on to a native gel consisting of
3% polyacrylamide (100 : 1)
4
/0.5% a garose, 0.5 ´ Tris/
borate/EDTA (pH 8.5), and run in 0.5 ´ Tris/borate/
EDTA at 4W for 16h at 4°C. The complexes were
analyzed on a Molecular Dynamics PhosphorImager. The
gel w as then exposed to ®lm at )80 °C w ith a n intensifying
screen.
Band-shift assay
Sense or antisense b-globin pre-mRNA (50 fmol) was
incubated with 500 l
M
antibiotic in a ®nal v olume of 10 lL
under the following conditions: 5 m
M
Hepes (pH 7.9),
480 l
M
MgCl
2
,20m
M
creatine ph osphate, 400 l
M

AB
/k
con
vs. antibiotic
concentration were ®tted to a binding equation, giving an
inhibition constant, K
i
, for splicing inhibition as described
previously [31].
RESULTS
Erythromycin, Cl-tetracycline and streptomycin inhibit
nuclear splicing
in vitro
The effects of eight different antibiotics (the aminoglycoside,
kanamycin, the a minocyclitol glycoside, streptomycin, t he
penicillins, ampicillin and penicillin G, the aminocyclitol,
spectinomycin, tetracycline and Cl-tetracycline, and the
macrolide, erythromycin) on n uclear b-globin splicing in
HeLa nuclear extract were investigated ( Fig. 2A). Antibiotic
was added to the splicing reaction mixture at a concentra-
tion of 450 l
M
. Kanamycin and penicillin G ( Fig. 2A, lanes
4 and 8) showed no effect on pre-mRNA splicing, whereas
ampicillin and spectinomycin (lanes 6±7) had a s light
enhancing effect. Some 60% of the b-globin pre-mRNA
was converted into excised lariat RNA in the presence of the
latter two antibiotics, compared with 47% excised lariat
formed in the control without antibiotic. Tetracycline
showed a slight inhibitory effect on pre-mRNA splicing,

)3¢ exon being almost as abundant as the excised lariat
RNA (Fig. 2D, as shown in Fig. 2A, lane 5 and Fig. 2B,
lanes 8 ±10). This i ndicates that e rythromycin may ha ve a
direct effect on catalysis by speci®cally inhibiting comp o-
nent(s) of the nuclear splicing apparatus involved particu-
larly in the second transesteri®cation step of splicing.
Like ribozymes, the nuclear splicing apparatus also
includes bivalent metal ions, especially Mg
2+
ions, for
stabilization of the spliceosome structure and catalysis,
despite the accompanying protein factors [2 ,36,37]. To
investigate whether splicing i nhibition is based on compe-
tition of the antibiotic s with Mg
2+
ions, Mg
2+
competition
experiments u sing 500 l
M
antibiotic and 3±10 m
M
Mg
2+
(3 m
M
Mg
2+
employed for standard splicing, see Materials
and methods and [38]) were performed (Fig. 3). The

160  12 l
M
, t hat for Cl-tetracycline was 180  9 l
M
,
and that for streptomycin was 230  32 l
M
.
Ó FEBS 2002 Nuclear splicing inhibition by antibiotics (Eur. J. Biochem. 269) 177
The c ourse of inhibition by Cl-tetr acycline was sim ilar t o
that by erythromycin (the course refers to titration of
antibiotic incubated for 60 min; Fig. 2C). A distinct course
of inhibition was exhibited b y streptomycin. An extended
delay phase up to a concentration of 150 l
M
was followed
by a burst phase leading to complete inhibition at 500 l
M
streptomycin (K
i
230 l
M
) (Fig. 2C). The course of inhibi-
tion is in line with the suggestion that streptomycin
competes with Mg
2+
ions for binding to the nuclear
splicing apparatus.
Cl-tetracycline and streptomycin, in contrast
with erythromycin, have different targets

con®rmed the possible binding of Cl-tetracycline a nd
streptomycin to the pre-mRNA. With Cl-tetracycline
and s treptomycin, the preincubation resulted in a 65%
and 46% increase in splicing i nhibition, compared with
the r eaction without preincubation (Fig. 4B). In c ontrast,
preincubation with erythromycin resulted in a 20%
decrease in splicing inhibition (Fig. 4B), con®rming that
erythromycin does n ot inhibit nuclear splicing by binding
to the pre-mRNA.
Next we looked at the potential interaction of erythro-
mycin w ith factor(s) of the nuclear extract in m ore detail
(Fig. 4C). Erythromycin (0, 300 l
M
or 500 l
M
)wasincu-
Fig. 2. Sp licing inhibition of the b-globin i ntron by several an tibiotics. (A) Eect of eight dierent antibiotics on nuclear splicing. b-globin splicing
employing 50 fmol of internally a-
32
P-labeled pre-mRNA was performed under st andard splicing cond itions at 30 °C for 60 min. Lane 1, no
addition of an tibiotic (±); lanes 2±9, in the presence of 450 l
M
antibiotic. CTC, C l-tetracycline; Tet, tet racycline; Kan, kanamycin; Ery, erythro-
mycin; Amp, ampicillin; Spec, spectinomycin; Pen, penicillin G; Strep, streptomycin; L-3¢Ex, lariat)3¢ exon ; L, lariat; Prc, mRNA precursor. (B)
Inhibition of b-globin splicing by Cl-tetracycline, tetracycline, erythromycin and streptomycin, shown in the presence of 0 l
M
(±), 100, 250 or
500 l
M
antibiotic. Reactions were performed as described in (A). Bands are labeled as in (A). (C) Results from (B) plotted as a graph. Th e eciency

antisense b-globin pre-mRNA was analyzed. The gel-
puri®ed sense a nd antisense RNAs showed main confor-
mation(s) without antibiotic as well as several other less
prominent native bands in the middle or lower area of the
gel (Fig. 5, lanes 1 and 2). The addition of Cl-tetracycline
and streptomycin to both RNAs r esulted in a change in the
native band pattern. Detectable upward band shifts (lanes 3,
4 and 6) as well as the appearance of a very s trong and
compact main conformation (lane 5 ) w ere observed . The
middle and lower bands disappeared in the presence of
Cl-tetracycline a nd strepto mycin. The observation that
changes in the band pattern occurred with both sense and
antisense R NAs, wh ich have similar secondary structures
but complementary primary sequences and different tertiary
structures, implied that C l-tetracycline and streptomycin
interact unspeci®cally with b-globin RNA.
Incubation of sense a nd antisense RNA with erythromy-
cin did not result in any change in the electrophoretic
mobility in n ative gel systems (Fig. 5, lanes 7 and 8 ). This is
Fig. 4. Cl-te tracycline and streptomycin have a dierent target for
inhibitory action from that observed for erythromycin. (A) Preincuba-
tion assay. Eect on splicing i nhibiton by preincubation of H eLa
nuclear extract with 500 l
M
Cl-tetracycline ( CTC), streptomycin
(Strep)orerythromycin(Ery),comparedwithacontrolreaction(±)
without antibiot ic. ( B) Inverse preincubation assay. Eect o f splicing
inhibition by preincubation of the b-globin pre-mRN A with 500 l
M
CTC, Strep o r Ery, compared with a control r eactio n (±) without

nuclear extract as suggested above.
Erythromycin completely inhibits spliceosomal
complex C formation
Next we investigated which spliceosomal complex is nega-
tively affected by the action o f erythromycin (Fig. 6).
Formation of the four different spliceosomal complexes E,
A, B and C was identi®ed in a time course splicing a ssay
using native gel systems (Fig. 6, lanes 1±4). Complex C,
representing the c atalytically active splicing c omplex, w as
formed after 20 min (lanes 3 and 4) . The ident ity of the
respective complexes was veri®ed independently by gel
extraction followed b y denaturing g el electrophoresis wh ere
the corresponding b-globin splicing educt and products were
detected (not shown).
Adding erythromycin in increasing concentrations (300,
600 and 900 l
M
) to t he 30 min s plicing r eaction resulted in
inhibition of spliceosome assembly. Formation of the
presplicing complexes A and B was slightly reduced by
increasing concentrations of erythromycin (lanes 5±7).
Interestingly, formation of complex C, in which the catalytic
steps o f s plicing o ccur, was strongly affected (lanes 5±7). In
detail, 300 l
M
erythromycin resulte d in formation of lo w
levels of complex C (lane 5). When the c oncentration o f
erythromycin was increased to 600 and 900 l
M
, formation

band-shift assay. Each antibio tic (500 l
M
)was
incubated with the RNA [sense (S) or an ti-
sense (A)] at 30 °C for 15 min. The samples
were separated on a native 5% polyacrylam ide
(50 : 1)/50 m
M
Tris/glycine gel. The native
patterns obtained were compared with those
without antibiotic (±) (lanes 1and 2). b and-
shift, area of upward band-shift; main, m ain
conformation(s); middle, bands in the m iddle
area of the g el; lower, bands in th e lower area
of the g el .
180 M. Hertweck et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Cl-tetracycline and streptomycin have an indirect
effect on splicing
With Cl-tetracycline and streptomycin, a general reduction
in b-globin p re-mRNA t urnover was observed (Fig. 2).
Results obtained in preincubation assays (Fig. 4) indicated
that these antibiotics affect splicing by binding to the pre-
mRNA. Their interaction with factor(s) of the HeLa nuclear
extract could b e excluded ( Fig. 4B). A band-shift assay
using sense and antisense b-globin pre-mRNA (Fig. 5)
suggested an unspeci®c interaction of these antibiotics with
b-globin R NA.
The data i mply that Cl-tetracycline and streptomycin
affect b-globin intron splicing by unspeci®c binding to
RNA, resulting i n a general reduction in pre-mRNA

2+
with the i nhibitory
action of Cl-tetracycline was observed.
Erythromycin has a direct effect on the second
catalytic step of splicing
In contrast with the a bove a ntibiotics, e rythromycin s pecif-
ically inhibited t he second transesteri®cation step of nuclear
pre-mRNA splicing. In the presence of increasing concen-
trations o f the antibiotic, the splicing i ntermediate l ari-
at)3¢ exon accumulated by a f actor 2±3, being as abundant
as the r educed excised lariat, whereas a relatively high
concentration of pre-mRNA was retained (Fig. 2). The
greatly d ec reased pre-mRNA turnover accompanied by the
change in molecular ratio of lariat )3¢ exon and e xcised
lariat compared with the splicing reaction without antibiotic
was a clear indication that erythromycin affects both steps
of splicing but speci®cally blocks the second step. The assays
using preincubation and dilution implied speci®c interaction
of erythromycin with one or several components of the
HeLa nuclear extract (Fig. 4). Thus, erythromycin showed
no af®nity for t he pre-mRNA (Fig. 4), which was con®rmed
by a band-shift assay ( Fig. 5). Furthermore, an erythromy-
cin-mediated total inhibition o f catalytically active spliceos-
omal complex C formation w as shown b y native g el
analysis. A ssembly of t he presplicing complexes A a nd B
were slightly affected by e rythromycin (Fig. 6).
These ®ndings together suggest erythromycin-mediated
splicing inhibition in which t he antibiotic has a dir ect effect
on the second catalytic step of splicing; e rythromycin is
thought to interact with one or more RNA and/or protein

omal catalytic center [49±53].
So far, information about the m echanism of i nhibitory
interaction between erythromycin and its target(s) within the
spliceosome is not known . Er ythromycin contains differently
charged groups. Prominent negative charges, such as the
hydroxy and carbonyl groups, the dimethylated amino
group of the
D
-desosamine and many uncharged methyl
groups characterize the e rythromycin m olecule (Fig. 1A).
Binding to the spliceosome could possibly result from
hydrophobic interactions or van der Waals forces. B inding
via the negative charges of erythromycin to essential bivalent
metal ions positioned i n binding pockets within the snRNAs
(without displacing them) is possible as w ell [12]. A s there is
no competition b etwe en erythromycin and Mg
2+
for b inding
to the s pliceosome (Fig. 3), and as t here are n o p rotonated
amino and other positively charged groups within erythro-
mycin, splicing inhibition by displacement of essential Mg
2+
ions as proposed for streptomycin can be excluded.
Interestingly, recently p ublished data [54±56] c learly
demonstrate that the macrolide antibiotics act, irrespective
of their antimicrobial activity, in the eukaryotic cell and
have inhibitory effects on m RNA expression a t t herapeutic
concentrations in vivo. Macrolides, e specially erythromycin,
are evidently capable of downregulating the expression of
various genes o f i n¯ammatory mediators such as cytokines

409.
5. Pikielny, C.W., Rymond, B.C. & Rosbash, M. (1986) Electro-
phoresis of ribonucleoproteins reveals an o rdered assembly path-
way of y east splicing complex es. Nature (Lon don) 324, 3 41±345.
6. Lamond, A.I., Konars ka, M.M., Grabowski, P.J. & Sharp, P.A.
(1988) Spliceosome assembly involves the b inding and release of
U4 small nu clear ribon ucleoprotein. Proc.NatlAcad.Sci.USA85,
411±415.
7. Konarska, M.M. (1990) Analysis of splicing complexes and small
nuclear ribonucleoprotein particles by native gel electrophoresis.
Methods Enzymol. 18 0, 442±453.
8. Padgett, R.A., Konarska, M., Grabowski, P.J., Hardy, S.F. &
Sharp, P.A. (1984) Lariat RNAs as intermediates and products in
the splicing o f messenger RNA p recursors. Science 225, 898±903.
9. Konarska, M.M., Grabowski, P .J., P adgett, R.A. & Sharp, P.A.
(1985) Characterization of the branch site in lariat RNAs pro-
duced by splicing of mRNA precursors. Nature (London) 313,
552±557.
10. Sharp, P.A. (1987) Splicing of mRNA precursors. Science 235,
766±771.
11. von Ahse n, U. & Schroeder, R. (1991) Strep tomycin inhib its
splicing of group I introns by competition with the guanosine
substrate. Nucleic Acids Res. 19 , 2261±2265.
12. Hermann, T. & Westhof, E. (1999) Docking of cationic antibiotics
to negatively charged pockets in RNA folds. J. Med. Chem. 42,
1250±1261.
13. Hermann, T. & Westhof, E. (1998) Aminoglycoside binding to the
hammerhead ribozyme: a general model for the interaction of
cationic antibiotics with R NA. J. Mol. Biol. 276, 9 03±912.
14. Sch roeder, R., Waldsich, C. & Wank , H. (2000) Modula tion of

translation are equivalent targets for erythromycin inhibition in
Staphylococcus aureus. Antimicrobial Agents Chemother. 40, 1301±
1303.
182 M. Hertweck et al.(Eur. J. Biochem. 269) Ó FEBS 2002
25. Leclerc, D., Melancon, P. & Brakier-Gingras, L. (1991) Mutation
in the 915 re gion of Escherichia c o li 16 S r ibosomal R NA r educe
the binding o f strept omycin t o th e ribosom e. Nucleic Acids Res.
19, 3973±3977.
26. Pinard, R., Payant, C., Melancon, P. & Brakier-Gingras, L. (1993)
The 5¢ proximal helix of 16S rRNA is involved i n the binding of
streptomycin to the rib osome. FASEB J. 7, 173± 176.
27. Epe, B., Woolley, P. & Hornig, H. (1987) Competition between
tetracycline a nd tRNA at both P and A sites of the rib osome of
Escherichia coli. FEBS Lett. 213, 443± 447.
28. Oehler, R., Polacek, N., Steiner, G. & Barta, A. (1997) Interaction
of tetracycline with RNA: photoincorporation into ribosomal
RNA of Escherichia c o li. Nucleic Acids Res. 25 , 1219±1224.
29. Brodersen, D.E., Clemons, W.M., Jr,, Carter, A.P., Morgan-
Warren, R.J., Wimberly, B.T. & Ramakrishnan, V. (2000) The
structural basis f or th e action of the antibiotics tetra cycline,
pactamycin, and hygromycin B on the 30S ribosomal subunit. Cell
103, 1143±1154.
30. Pioletti, M ., S chlu
È
nzen, F., Harms, J., Zarivach, R., Glu
È
hmann,
M.,Avila,H.,Bashan,A.,Bartels,H.,Auerbach,T.,Jacobi,C.,
Hartsch,T.,Yonath,A.&Franceschi,F.(2001)Crystalstructures
of the com plexes of the small ribosomal subunit with tetracycline,

merhead inhibitor. J. Biochem. 317, 855±860.
40. Senapathy, P., Shapiro, M .B. & Harris, N.L. (1990) Splice ju nc-
tions, branch point sites and exons: sequence statistics, ide nti®-
cation, a nd applications t o genome project. Methods E nzymol.
183, 252±278.
41. Goodall, G .J. & Filipowicz, W . (1991) D ierent eects o f i ntron
nucleotide co mposition and secondary struct ure on pre-mRNA
splicing in monocot and dicot plants. EMBO J. 9, 2635±2644.
42. Mount,S.M.,Burks,C.,Hertz,G.,Stormo,G.D.,White,O.&
Fields, C. (1992) Splicing signals in drosophila: intron size, infor-
mation co nte nt, and consensu s sequences. Nucleic A cids Res. 20 ,
4255±4262.
43. Mitscher, L.A. (1978) The Chemistry of Tetracycline Antibiotics.
Marcel Dekker, New Y o rk.
44. Newman, A. & Norman, C. (1991) Mutations in y east U5 snRNA
alter the speci®city of 5¢ splice-site cl eavage. Cell 65, 115±123.
45. Madhani, H.D. & Guthrie, C. ( 1992) A novel base-p airing inter-
action between U2 a nd U6 snRNA suggests a mechanism for the
catalytic activation of t he spliceosome. Cell 71, 803±817.
46. Chang, J.S. & McPheeters, D.S. (2000) Identi®cation of a U2/U6
helix Ia mutant that in¯uences 3¢ splice site selection during
nuclear pre-mRNA splicing. RNA 6, 1 120±1130.
47. Reed, R. ( 2000) Mechanism o f ®delity i n pre-mRNA s plicing.
Curr. Opin. C ell Biol. 12 , 340±345.
48. Staley, J.P. & Guthrie, C. (1998) Mechanical devices of the
spliceosome: motors, clocks, springs, and things. Cell 92, 315±326.
49. Krainer, A.R. & Maniatis, T. (1985) Multiple factors including the
small nuclear ribonucleoproteins U1 and U2 are necessary for pre-
mRNA splicing in vitro. Cell 42, 725 ±736.
50. Sawa, H. & Shimura, Y. (1991) Alterations of RNase H sensitivity