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 binding to nuclear extract component(s). This results in direct inhibition of the second step of pre-mRNA splicing. To our knowledge, this is the ® rst report on s peci®c inhibition of nuclear splicing by an a ntibiotic. The func- tional g roups involved in the interaction of erythromycin with snRNAs and/or splicing factors require further investigation. Keywords: 1 erythromycin; C l-tetracycline; nuclear splicing inhibition; spliceosomal complex C. Recognition o f nuclear mRNA precursors (pre-mRNAs) by heterogeneous nuclear RNPs in¯uences the s ubsequent assembly of the spliceosome [1]. The spliceosome is a multicomponent complex consisting of ®ve different snR- NAs and a large number of spliceosomal and nonspliceos- omal proteins which assemble on the pre-mRNA before the splicing reaction is started. Although the snRNAs are thought to be the catalytically essential participants in splicing, accompanying proteins are necessary for co-ordinating and regulating spliceosome assembly and the splicing reactions themselves [2±4]. Spliceosome assem- bly is a multistep process, and n ative gel analysis has been employed to detect the f ormation of four distinct spliceo - somal complexes termed E, A, B and C [5±7]. Complex E includes U 1 snRNP bound to the nascent pre-mRNA, while complex A contains U1 as well as U2 snRNP bound to the pre-mRNA. Complex B represents a precatalytic assembly in whic h the U4/U6 hybrid and the U5 s nRNP have joined the complex, but where interactions between the s plicing 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 synthesis [15±18]. For instance, macrolide antibiotics, espe- cially erythromycin (Fig. 1 A), are supposed to inhibit translation in bacteria by interacting with RNA and protein components o f t he 50S ribosomal subunit. An effect on the peptidyltransferase reaction a nd prevention of positioning of the peptidyl-tRNA in the peptidyl site (P site) is the consequence [19±21]. Macrolides are also known to inhibit the initial step of 50S subunit assembly in g rowing bacter ia [22±24]. The a minocyclitol glycoside s treptomycin (Fig. 1B) Correspondence to M. W. Mu eller, VBC Genomics, Bioscience Research GmbH, Rennweg 95B, A-1030 Vienna, Austria. Fax: + 43 1 7966572 21, Tel.: + 43 1 7966572 20, E-mail: mueller@vbc-genomics.com (Received 1 6 August 2001, revised 22 October 2001, accepted 29 October 2001) Eur. J. Biochem. 269, 175±183 (2002) Ó FEBS 2002 as well as tetracyclines (Fig. 1C) inhibit bacterial translation by interacting with the 30S ribosomal subunit. Streptomycin induces misreading of the genetic code. It binds to four different parts of 16S rRNA, i.e. close to the decoding site, and makes contact with protein S12, affecting codon± anticodon interaction and proofreading [ 18,25,26]. Tetracy- clines in duce c onformational changes in the 30S subunit by interfering with t he binding of tRNA to both the ribosomal aminoacyl s ite (A site) and t he P site [ 18,27]. 16S RNA and the proteins S4, S7, S9 and S17 are thought to be essential components of the strong binding sites o f tetracyclines on the 3 0S subunit [ 28±30]. Furthermore, man y aminoglyco- sides and tetracyclines are known to i nhibit the catalytic activity of self-splicing group-I and group-II introns, self- 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 intron of rabbit pre-mRNA ( 573 nucleotides), ¯anked by the corresponding 5 ¢ exon sequence ( 105 nucleotides) and 3¢ exon sequence (45 nucleotides), was used to assay pre- mRNA splicing. To construct the plasmid KS+/b-globin, the i ntron p lus ¯anking exons of genomic rabbit DNA was ampli®ed by PCR techniques. For PCR, the primers b-glb1/ P(5¢-AATCTGGTACCTCACCTGGACAACCTCAAA- 3¢)andb-glb2/M (5¢-AATAAGAATTCGCCAAAAT GATGAGACAGCA-3¢) were used. The product was cloned between the KpnIandEcoRI site of the bluescript phagemid vector KS+. The c loning step has been con- ®rmed by dideoxy sequencing with the KS+-speci®c T3 primer. In vitro transcription The b-globin pre-mRNA was prepared from EcoRI-linea- rized KS+/b-globin using T3 RNA polymerase. Template DNA (5 lg) was transcribed in a 200-lL volume containing 2.5 m M each NTP, 8 m M MgCl 2 ,2m M spermidine, 4 0 m M Tris/HCl, pH 8.0, 50 m M 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). (B) T he aminocyclitol g lycoside streptomycin with a s treptidine ring as the main building block: 1,3-diguanidinoinositol. (C) Te tra- cycline consisting of four f used rings. The hydrogen at position 7 in tetracycline is replaced by chlorine in Cl- t etracycline. 176 M. Hertweck et al.(Eur. J. Biochem. 269) Ó FEBS 2002 extract (Promega), 5 m M Hepes (pH 7.9), 3 m M MgCl 2 , 20 m M creatine p hosphate, 400 l M ATP, 0.6% polyv inyl alcohol, 18 U RNase inhibitor, 50 fmol pre-mRNA, a nd antibiotic at concentrations indicated in the ®gures. Reac- tions were stopped b y the addition of HeLa nuclear extract stop mix (Promega). The mixture was then extracted with phenol/chloroform/isoamyl alcohol (39 : 59 : 2, v/v), and the aqueous phase ethanol precipitated. After a washing step, the pellets were resuspended in f ormamide-containing gel-loading dye, a nd the p roducts separated on denaturing 5% polyacrylamide (30 : 1) 8 M urea gels. The products 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 ATP, 0.6% polyvinyl alcohol and 5% glycerol. Reactions were carried out at 30 °C for 15 min, and the mixtures put on i ce and loaded o n to a native 5% polyacrylamide (50 : 1), 50 m M Tris/glycine (pH 7.5) gel and run in 50 m M Tris/ glycine at 12 W for 3 h at room temperature. The gel was analyzed as described above. Kinetic analyses b-Globin s plicing over 0±60 min was followed in the absence and presence of various concentrations of antibiotic (0, 50, 100 , 250, 500 l M ). Reactions were carried out as described for in vitro splicing assays and gels were scanned on a PhosphorImager. The fraction inhibition at each antibiotic concentration was calculated from the rates of pre-mRNA t urnover for the antibiotic reaction (k AB(antibiotic) ) and the corresponding control reaction [kcon(control without AB)] from the relationship 1 ) k AB /k con . The values of 1 ) k 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, as indicated by a reduction in excised lariat RNA by 9% only (lane 3). T he other three antibiotics, Cl -tetracycline, streptomycin and erythromycin, are clearly inhibitors of nuclear pre-mRNA splicing, as indicated by a large reduction in excised lariat R NA and a corresponding amount of unchanged pre-mRNA, respectively (lanes 2, 9 and 5). The t hree inhibitory antibiotics and tetracycline were tested in the concentration range 0±5 00 l M (Fig. 2B,C). Conversion of pre-mRNA into excised lariat RNA was reduced by 60% in the presence of 500 l M Cl-tetracycline and by 95% in t he presence of 500 l M erythromycin and streptomycin (Fig. 2B,C). Tetracycline at 500 l M inhibited splicing by only 1 0% (Fig. 2B,C), c on®rming that it is negligible. The different r esults obtained f or Cl-tetracycline and tetracycline suggest an important role for the chlorine group in inhibition. Splicing in the presence of Cl-tetracycline a nd strepto- mycin r esulted i n a general r eduction in pre-mRNA turnover (Fig. 2A,B, lanes 1±4 and 11±13). Splicing inhibi- tion by erythromycin, the strongest in hibitor observed, led to the accumulation of the splicing intermediate l ariat )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 experiments indicated interference of Mg 2+ with the inhibitory action of streptomycin only. An increase in Mg 2+ concentration to 1 0 m M completely reversed s tre p- tomycin-mediated splicing inhibition (Fig. 3). This may suggest that streptomycin competes with Mg 2+ ions for binding to the nuclear splicing apparatus. The inhibitory effect of erythromycin and Cl-tetracycline were not allevi- ated by Mg 2+ (Fig. 3). Kinetic analyses using increasing concentrations of erythromycin, Cl-tetracycline and streptomycin up to 500 l M were performed. The concentration required to inhibit 50% of pre-mRNA splicing in HeLa nuclear extract (K i ) was determined for each a ntibiotic ( see Mate rials and methods and [31]). The K i for erythromycin was 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 for inhibitory action We investigated on which component of the nuclear splicing reaction, the b-globin pre-mRNA or t he HeLa nuclear extract containing the snRNAs and associated splicing factors, the antibiotics exhibit their inhibitory action (Fig. 4). For this, we designed preincubati on assays. In the ®rst preincubation assay (Fig. 4A), the HeLa nuclear extract was preincubated with 500 l M Cl-tetracycline, streptomycin or erythromycin for 1 5 min at 30 °Cinthe absence of pre-mRNA, followed by the addition of pre- mRNA in prewarmed buffer (30 °C) and incubation for 1 h at 30 °C. The splicing ef®ciencies were calculated from the percentage of the excised lariat fraction. With Cl-tetracyc- line, preincubatio n resulted in a 4 0% diminished inhibitory effect, compared w ith t he reaction without preincubation (Fig. 4A). For streptomycin the decrease in splicing inhibi- tion was 12% (Fig. 4A). This indicates that these antibiotics do not inhibit nuclear s plicing by b inding to factor(s) of t he nuclear extract. The same e xperiment u sing erythromycin showed the opposite effect. Splicing inhibition was increased by 90% after preincubation (F ig. 4A), suggesting that erythromycin binds to factor(s) of t he HeLa extract e ven in the absence of pre-mRNA. In the inverse preincubation experiment, the poten tial binding of Cl-tetracycline, streptomycin and erythromycin to b-globin pre-mRNA (Fig. 4B) was examined. The pre- mRNA was preincubated with the antibiotics before being mixed with the other components. The d ata 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 of lariat formation ( %) is plotted as a function of increasing concentrations of antibiotic (l M ). (D) Inhibition o f b-globin s plicing by increasing concentratio ns of e ryth romycin. The eciency of lariat)3¢ ex on and lariat formation (%) is ill ustrated as a function o f erythromycin ( l M ). The ®gure represents means  SEM from four independent experiments. 178 M. Hertweck et al.(Eur. J. Biochem. 269) Ó FEBS 2002 bated at 30 °C under standard splicing conditions for 60 min. As a control, half of each set up w as stopped (undiluted r eaction) and t he remainder adjusted to the ®nal reaction volume with HeLa nuclear e xtract and splicing buffer; the incubation was then c ontinued for 30 min (diluted r eaction). The data revealed that splicing i nhibition by erythromycin (undiluted reaction) is signi®cantly dim in- ished by a 1 : 1 dilution with nuclear extract and buffer (diluted r eaction) by a factor of 2±5, resulting in 58% lariat formation when c ompared w ith t he respective control reaction (Fig. 4C). As the r atio between pre-mRNA and erythromycin in the undiluted and diluted reaction is constant, the result of signi®cant reduction in splicing inhibition indicates t hat erythromycin does not inh ibit nuclear splicing b y interacting with the pre-mRNA. Thus, this ®nding is consistent with the idea t hat erythromycin interacts speci®cally with one or several components of the nuclear extract. The suggested inhibitory interaction between Cl-tetracyc- line or streptomycin and the pre-mRNA was further investigated in a band-shift assay ( Fig. 5), in which the af®nity of all three inhibitory antibiotics for sense a nd 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 antibiotic. (C) Signi®cant reduction in erythromycin-mediated sp licing inhibition by a 1 : 1-dilution of b-globin pre-mRNA and antibiotic (0, 300 or 500 l M Ery) with nuclear extract and buer. The ®gure represents means  SEM f rom three independent experiments. Fig. 3. In terference of Mg 2+ with the inhibitory action of streptomycin analyzed by Mg 2+ competition experiments. Reactions were performed as described in Fig. 2A using 500 l M streptomycin (Strep), Cl-tetra- cycline (CTC) or erythromycin (Ery) and increasing concentrations of MgCl 2 (3, 6 and 10 m M MgCl 2 ), compared with a control reaction (±) without antibiotic. The ®gure represents means  SEM f rom two independent experiments. Ó FEBS 2002 Nuclear splicing inhibition by antibiotics (Eur. J. Biochem. 269) 179 in agreement with the idea that erythromycin does not interact directly with the pre-mRNA. Thus, the inhibitory effect on splicing may result from interaction of this antibiotic with RNA a nd/or protein component(s) o f the 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 of complex C was almost completely ( lane 6) or completely (lane 7) inhibited. This is in line with t he large r eduction in pre-mRNA turnover with increasing concentrations of erythromycin (cf. Figure 2B). DISCUSSION In this study, three antibiotics, erythromycin, Cl-tetracycline and s treptomycin, were identi®ed in vitro as inhibitors of nuclear b-globin pre-mRNA splicing ( Fig. 2). They are members o f t hree different classes o f antibiotic: macrolides, tetracyclines and aminoglycosides. Of the antibiotics tested, erythromycin was found to be the strongest inhibitor. The K i values for erythromycin, Cl-tetracycline and streptomy- cin were f ound to be in the m icromolar range (160, 180 and 230 l M , respectively). The K i values are similar to those observed in tetracycline-mediated inhibition of the s elf- cleavage reaction in hammerhead and human hepatitis delta virus ribozymes [33,39], the streptomycin-mediated inhibition of group-I intron self-splicing [11], and some aminoglycoside class-mediated inhibitions of RNase P cleavage [35]. Fig. 5 . An i ty of Cl-tetracycline (CTC; lanes 3±4), streptomycin (Strep; lanes 5±6) and ery- thromycin (Ery; la nes 7±8) f or the s ense and antisense b-globin pre-mRNA analyzed by a 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 turnover. A possible explanation is that charged groups on these antibiotics perform ionic interactions with b-globin RNA. Thereby, the a ntibiotics may have a gen eral af®nity for several RNA sequence motifs and/or tertiary structure elements. However, the unspec i®c interaction between these antibiotics and b-globin RNA cannot be generalized to all eukaryotic pre-mRNAs because of the low sequence conservation among nuclear introns [40±42]. Focusing on the presence or a bsence of a speci®c side group, tetracycline and Cl-tetracycline h ad signi®cantly different effects o n nuclear pre-mRNA splicing. The two tetracyclines, composed of four fused rings with several substituents, only differ by a single atomic substitution of chlorine for hydrogen at position 7 (Fig. 1C). Inhibition by Cl-tetracycline w as found to be sixfold more t han by tetracycline (Fig. 2). The extent of splicing inhibition by tetracycline is n egligible (Fig. 2). The sixfold stron ger splicing inhibition by Cl-tetracycline may be explained by the a dditional s ize and hydrophobicity introduced by the substitution, as is often the case in ribozyme inhibition by tetracyclines [33,39,43]. In Mg 2+ competition experiments (Fig. 3), only splicing inhibition caused by streptomycin was suppressed. It is possible that s treptomycin c ompetes w ith e ssential M g 2+ ions for overlapping binding sites on b-globinintronRNA. Although tetrycyclines are known to be chelators of metal cations [43], no interference o f M g 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 components of t he nuclear extract t hat correspond to the splicing-active C complex, playing an important part in the central stage(s) of the second catalytic step o f s plicing. Fig. 6. Eect of erythromycin on spliceosome assembly shown by native gel analysis. Reactions were performed as described in Fig. 2 A and incubated for the time indicated (0, 10, 20 or 30 min). Time-dependent formation of the four spliceosomal complexes E, A, B and C (lanes 1±4) and eect on complex formation of increasing concentrations of erythromycin (3 00, 600 and 9 00 l M ; l anes 5±7) a re shown. The com- plexes we re separated o n a native 3% polyacrylam ide (100 : 1 )/0.5% agarose, 0.5 ´ Tris/bor ate/EDTA gel. Ó FEBS 2002 Nuclear splicing inhibition by antibiotics (Eur. J. Biochem. 269) 181 Potential interference of erythromycin with the formation of complex A and/or B c annot be excluded. We may assume that three of the ®ve spliceosomal snRNAs, U 2, U5 and U6, in pa rticular are potent ially major RNA targets of erythromycin. They are central participants of complex C and all are required in the catalytic steps of splicing [44±47]. Furthermore, interference of erythromycin with the function of the s pliceosomal and nonspliceosomal protein cofactors and DEAD/H box ATPases cannot be exclud ed. These accessory proteins trigger speci®c rearrangements of the spliceosome at every s tage of the s plicing pathway [48]. For instance, the mammalian U5 snRNP protein U5 220 is actually thought to be contributed directly to the spliceos- 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 and c hemokines in several h uman cell types by inhibiting the activation of different transcription factors, including nuclear factor-jB and activator protein-1. These data, together with our ®nding that the nuclear splicing apparatus is a speci®c target of erythromycin in vitro, suggest that the u se of erythromycin may h ave serious consequences: it may interfere with at least two major stages of eukaryotic gene expression, transcription i nitiation and pre-mRNA splicing. Thus, our data support the call for moderate use of antimicrobial-acting antibiotics, particu- larly erythromycin, in humans and animals. ACKNOWLEDGEMENT This work was supported by the Austrian Fonds zur Fo È rderung der wissenschaftlichen Forschung (FWF) . REFERENCES 1. Dreyfuss, G. (1993) HnRNP proteins and the biogenesis of mRNA. Annu.Rev.Biochem.62, 289±321. 2. Lu È hrmann, R ., Kastner, B. & B ach, M. (1990) Structure and spliceosomal s nRNPs and their role in pre-mRNA splicing. Biochim. Bio phys. Acta 1087, 265±292. 3. Moore, M.J., Query, C.C. & Sharp, P.A. (1993) Splicing o f p re- cursors to mRNA by the spliceosome. I n The RNA World (Gesteland, R.F. & Atkins, J.F., eds), pp. 303±358. Cold Spring Harbor Laboratory Press, C old Spring Harbor, New Yo rk . 4. Kramer, A. (1996) The structure and function of proteins involved in mammalian pre-mRNA splicing. Annu. Rev. Biochem. 65, 367± 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 RNA function by aminoglycoside antibiotics. EMBO J. 1, 1±9. 15. Teraoka, H. (1971) Binding of erythromycin to Escherichia coli ribosomes. J. Antibiot. 24, 302±309. 16. Vazquez, D . ( 1979) I nhibitors of protein biosynthe sis. Mol. Biol. Biochem. Biophys. 30, 8 9±95. 17. Gottesmann, M.E. (1967) Reaction of ribosome-bound peptidyl transfer ribonucleic acid with aminoacyl transfer ribonucleic acid or pu romycin. J. Biol. Chem. 242 , 5564±5571. 18. Moazed, D. & Noller, H.F. (1987) In teraction of antibiotics with functional sites in 16S ribosomal RNA. Nature (Londo n) 327, 389±394. 19. Pestka, S. (1974) Binding of [ 14 C]erythromycin to Escherichia coli ribosomes. Antimic rob. Agents C hemother. 6, 474± 478. 20. Vester, B. & G arrett, R.A. (1987) A plasm id-coded a nd site- directed mutation in Escherichia coli 23S RNA that confers resistance to er ythromycin: i mplications for the mechanism of action o f erythromycin. Bioc himie 69, 891±900. 21. Chinali, G., Ny ssen, E., diGiambattista, M. & Cocito, C. (1988) Action of er ythromyc in and virgin iamycin S on polype ptid e syn- thesis in cell-free systems. Biochim. Bioph ys. Acta 951, 42±52. 22. Chittum , H.S. & C hampney, W.S. (1995) Erythromycin inhibits the assembly of the large ribosomal subunit in growing Escherichia coli cells . Curr. Microbiol. 30, 2 73±279. 23. Champney, W.S. & Burdine, R. (1995) Macrolide antibiotics inhibit 50S ribosomal subunit assembly in Bacillus subtilis and Staphylococcus a ureus. Antimicrobial Agents Chemother. 29, 241± 2144. 24. Champney, W.S. & Burdine, R. (1996) 50S ribosomal subunit and 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, edeine and IF3. EMB O J. 20 , 1829±1839. 31. Stage, T.K., Hertel, K. & Uhlenbeck, O.C. (1995) Inhibition of the hammerhead ribozyme by neomycin. RNA 1, 95±101. 32. Earnshaw, D.J. & Gait, M.J. (1998) Hairpin ribozyme cleavage catalyzed by aminoglycoside antibiotics and the polyamine sper- mine in the absence of metal ions. Nucleic Acids Res. 26, 5551± 5561. 33. Rogers, J., Chang, A.H., von Ahsen, U., Schroeder, R. & Davies, J. (1996) Inhibition of the self-cleavage reaction of the human hepa- titis delta virus ribozyme by antibiotics. J. Mol. Biol. 259,916± 925. 34. Zapp, M.L., Stern, S. & Green, M.R. (1993) Small molecules that selectively block RNA binding of HIV-1 Rev. protein inhibit Rev. function and viral production. Cel l 74, 969 ±978. 35. Mikkelsen, N.E., Brannvall, M., Virtanen, A. & Kirsebom, L .A. (1999) Inhibition of RN ase P RNA cleavage by aminoglycosides. Proc. N atl Acad. Sc i. USA 96 , 6155±6160. 36. Sontheimer, E.J., Gordon, P.M. & Piccirilli, J.A. (1999) Metal ion catalysis during group I I intron self-splicing: parallels with the spliceosome. Genes D ev. 13, 1729 ±1741. 37. Gordon, P.M., Sontheimer, E.J. & Piccirilli, J.A. (2000) Metal ion catalysis du ring the exon-ligation step of nuclear pre-mRNA splicing: extending the parallels between the spliceosome and group II i ntrons. RNA 6, 199±205. 38. Lamond, A.I., Konarska, M.M. & Sharp, P.A. (1987) A muta- tional analysis of spliceosome assembly: evidence for splice site collaboration d uring s pliceosome formation. Genes Dev. 1, 532±543. 39. Murray, J.B. & Arnold, J.R.P. (1996) Antibiotic interactions with the hammerhead ribozyme: tetracyclines as a new class of ham- 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 of the 3¢ splice site region during the in vitro splicing reaction. Nucleic A cids Res. 19, 3 953±3958. 51. Chua, K. & Reed, R. (1999) Human step I I splicing f actor hSlu7 functions in restructuring the spliceosome between the catalytic steps o f splicing. Genes Dev. 13, 841 ±850. 52. Luo, H .R., Moreau, G .A., Levin, N . & Moore, M.J. (1999) The human Prp8 protein is a component of both U2- and U12- dependent spliceosomes. RNA 5, 8 93±908. 53. Siatecka, M., Reyes, J.L. & Konarska, M.M. (1999) Functional interactions of Prp8 with both splice sites at the spliceosomal catalytic center. Genes D ev. 13, 1983±1993. 54. Aoki, Y. & K ao, P.N. (1999) Er ythromycin inhibits transcrip- tional activation of N F- K B, b ut not NFAT, t hrough calcineurin- independent signaling in T cells. Antimicrobial Agents Chemother. 43, 2678±2684. 55. Desaki, M., Takiza wa, H., Ohtoshi, T., Kasama, T., Kobayashi, K.,Sunazuka,T.,Omura,S.,Yamamoto,K.&Ito,K.(2000) Erythromycin suppresses nuclear factor- K B and ac tivator protein-1 activation i n human bronchial epithelial cells. Biochem. Biophys. R es. Commun. 26 7 , 124±128. 56. Kohri, K., Tamaoki, J., Kondo, M., Aoshiba, K ., Tagaya, E. & Nagai, A. (2 000) Macrolide a ntibiotics inhibit nitric o xide generation by rat pulmonary alveolar macrophages. Eur. Respir. J. 15, 62±67. Ó FEBS 2002 Nuclear splicing inhibition by antibiotics (Eur. J. Biochem. 269) 183 . 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. of splicing inhibition by tetracycline is n egligible (Fig. 2). The sixfold stron ger splicing inhibition by Cl-tetracycline may be explained by the a