MicrObiology (1997), 143,617-624
Printed in Great Britain
Natural kirromycin resistance of elongation factor Tu from the kirrothricin producer Streptomyces cinnamoneus Carmela Cappellano, Federica Monti, Margherita Sosio, Stefano Donadio and Edoardo Sarubbi Author for correspondence: Stefano Donadio. Tel: +39 2 96474 243. Fax: +39 2 96474 400. e-mail :
[email protected]
Lepetit Research Center, via R. Lepetit 34, 2 1040 Gerenzano, Italy
The antibiotic kirromycin (Kr) inhibits bacterial protein synthesis by binding to elongation factor T u (EF-Tu). Streptomyces cinnamoneus and Nocadia lactamdurans, producers of antibiotics of the Kr class, are known to possess an EF-Tu resistant to Kr. Both micro-organisms appear to possess a single tuf gene and we have characterized the one from 5. cinnamoneus, which belongs to the tuff family. To assess the molecular determinants of Kr resistance, the S. cinnamoneus tuf gene was expressed in Escherichia coli as a translational fusion to ma/€, which enabled the recovery by affinity chromatography of the recombinant protein uncontaminated by the host factor. The recombinant EFT u was able to catalyse polyU-directed polyphe synthesis in two heterologous cell-free systems, even as an uncleaved fusion. When tested for antibiotic sensitivity it behaved like the natural S. cinnamoneus protein, showing equivalent resistance to Kr but sensitivity to the antibiotic GE2270, indicating that all determinants for Kr resistance are intrinsic to the EF-Tu sequence. Multiple sequence analysis of EF-Tu proteins, together with knowledge of mutations conferring Kr resistance, allowed the identification of key residues as likely candidates for the natural Kr resistance of the S. cinnamoneus EF-Tu. One of these, Tht.378, was mutated to the consensus Ala and the resulting mutant protein was sensitive to Kr. Interestingly, it retained some activity (30% of the control) even a t high Kr concentrations. Keywords : Streptomyces cinnamoneus, antibiotic, expression system, protein synthesis, tuf gene
INTRODUCTION
Elongation factor (EF) T u is an essential component of bacterial protein synthesis (Weijland et al., 1992). It forms a ternary complex with GTP and aa-tRNA and interacts with the elongating ribosome to place aatRNA in the A-site. After codon-anticodon interaction, GTP is hydrolysed, causing the dissociation of EFTu*GDP from aa-tRNA and from the ribosome. EFTu*GDP is subsequently recycled to an active form via a nucleotide exchange reaction promoted by EF-Ts. The polyenic antibiotics of the kirromycin (Kr) class are Abbreviations: EF, elongation factor; Kr, kirromycin; MBP, maltose binding protein. The EMBL accession number for the nucleotide sequence reported in this paper is X98831. 0002-0996 0 1997 SG M
known to interact with both EF-Tu*GDP and EFTu"GTP, affecting protein synthesis by inhibiting the release of EF-Tu"GDP from the ribosome, thus preventing elongation (Parmeggiani & Swart, 1985). GE2270 (also known as MDL 62879), a structurally unrelated antibiotic, also binds to EF-Tu, but it inhibits protein synthesis through a different mechanism, i.e. by preventing the formation of the aa-tRNA*EF-Tu*GTP complex (Anborgh & Parmeggiani, 1991;Landini et al., 1992). The protein synthesis machinery of actinomycetes producing antibiotics of the Kr class, although all resistant to these molecules, exhibit different properties (Glockner & Wolf, 1984). Some producers, such as Streptomyces collinus and Streptomyces ramocissimus, possess a KrS EF-Tu, with an IC,, of 0.2-0-6pM. Others, such as Streptomyces cinnamoneus and 617
C. CAPPELLANO and OTHERS
Nocardia lactamdurans (formerly Streptomyces lactamdurans), possess a Kr' factor (IC5,, > 500 pM). Different possibilities can be conceived to explain high level resistance in uiuo. For instance, a KrSEF-Tu might be replaced with a resistant factor at the onset of antibiotic production. A mechanism of this type, analogous to that found in producers of novobiocin (Thiara & Cundliffe, 1989) and pentanenolactone (Frohlich et al., 1989),would be consistent with the recent finding of multiple tuf genes in Streptomyces (Vijgenboom et a!., 1994; van Wezel et al., 1994).Alternatively, an otherwise sensitive target might be made resistant by enzymic modification, a protective mechanism commonly present in producers of protein synthesis inhibitors (Cundliffe, 1989). Finally, the EF-Tu might be intrinsically Kr' because of its amino acid sequence. In the latter case, the high conservation of EF-Tu sequences from different organisms and the known mutations conferring Kr' (Abdulkarim et al., 1994; Mesters et al., 1994) might provide clues to specific residues implicated in resistance. S. cinnamoneus produces kirrothricin, an antibiotic of the Kr class presenting only minor structural differences from the latter and presumed to act by a similar mechanism (Thein-Schrammer et al., 1982).In this study we present evidence that S. cinnamoneus possesses a single tuf gene which encodes all the information necessary for a Kr' EF-Tu. Remarkably, a similar resistance mechanism also operates in the actinomycete Planobispora rosea, producer of the EF-Tu inhibitor GE2270 (Sosio et al., 1996).The deduced sequence of the naturally resistant S . cinnamoneus EF-Tu provides clues to the evolution of a resistant target in a producing organism. A preliminary account of this work has been presented (Alderson et al., 1994). METHODS Bacterial strains and plasmids. Escherichia cofi DH5a was the routine host for cloning, whereas E. cofi JM109 was used for gene expression. S. cinnamoneus Tu89 (Thein-Schrammer et al., 1982), Streptomyces coelicofor M145 and Streptomyces fividans 1326 (Hopwood et af., 1985), Streptomyces gfaucescens ETH 22794 and N. factamdurans ATCC 27382 were grown in standard media (Hopwood et af., 1995). E. cofi DH5a and Bacillus subtifis ATCC 6633 were used for protein synthesis assays. Plasmids of the pUC series (Vieira & Messing, 1982) and pUCBM21 (Boehringer Mannheim) were used for cloning. pMAL-c2 was from New England Biolabs. Cloning of the S. cinnamoneus tuf gene. For cloning the tuf gene, chromosomal DNA from S. cinnamoneus was digested with BamHI and a 3-5-45 kb fraction was recovered from an agarose gel and ligated to BamHI-digested pUCBM21. Plasmid DNA from about 800 of the resulting recombinant colonies was prepared as 100 pools of eight colonies each and analysed by Southern hybridization with the S . coeficofor tufprobe. A single positive pool was identified, from which the single hybridizing plasmid pGEl5O was identified. DNA manipulations. Genomic DNA from actinomycetes was prepared as described by Hopwood et af. (1985). The S. coeficofor tufsegment was amplified using the same conditions used for the P . rosea tufsegment (Sosio et af., 1996), but with
618
an annealing temperature of 50 "C. DNA sequences were determined with an ALF automated sequencer (Pharmacia), following the manufacturer's instructions. Sequence analyses were performed using GCG programs (Devereux et af., 1984). Southern hybridization of genomic DNAs with the S. coeficofortuf segment was carried out with a hybridization stringency set at 6 x SSC, 65 "C, with final washes at 2 x SSC, 65 "C (1x SSC is 0.15 M sodium chloride, 0.015 M sodium citrate) . With the S. cinnamoneus tuf probe, the hybridization stringency was set at 6 x SSC, 58 "C, with final washes at 2 x SSC at this same temperature. Plasmid constructions. A 90 bp fragment was PCR-amplified from pGE150, introducing a SmaI site at the tuf5' end and changing the third position in the first few codons. The primers used were 5' TCCCCCGGGATGGCAAAAGCAAAATTCGAGCGGAC 3' (bases differing from the S . cinnamoneus sequence are in italics) and 5' GGTGATCGCCGCGGTAAGAGT 3'. This fragment, after digestion with SmaI and SacII, and the 1.4 kb SacII-PstI fragment from pGE1.50 (Fig. 1; the PstI site originates from the polylinker) were simultaneously ligated into pMAL-c2, previously digested with XmnI and PstI. The resulting plasmid was named pMAL-TUF1. To construct plasmid pMAL-TUF2, the 319 bp NruI-BamHI fragment from pMAL-TUF1 was replaced with the equivalent segment obtained from pGEl.50 after PCR with primers 5' AGCCTGTCGCGATGGAGGAAGGCCTTCGCTTCGCCATCC 3' and 5' GACTCTAGAGGATCCACAGGTCGT 3'. The first primer changed the ACC codon for Thr3'* into GCC and introduced a silent XmnI site for following the mutant allele. The fidelity of PCR synthesis was verified by DNA sequencing. Production and purification of recombinant EF-Tu. E. cofi JMl09 cells harbouring pMAL-TUF1 or pMAL-TUF2 were grown at 30 "C in 1 1 LB medium containing 50 mg ampicillin ml-'. At an OD,oo of 0 . 4 4 6 , IPTG to 0.2 mM and MgCl, to 10 mM were added. After a further 2 h of growth, cells were harvested and the resulting bacterial pellet was resuspended in 40 ml buffer A (20 mM Tris/HCl, pH 7.5,0.2 M NaCl, 10 mM MgCl,, 1 mM D T T and 5 pM GDP) containing 1 mg lysozyme m1-l. Cells were incubated for 30 min in ice, lysed by sonication and then centrifuged to remove cellular debris. The supernatant was diluted fivefold with buffer A and loaded on a 30 ml amylose column (New England Biolabs) preequilibrated with the same buffer. Fusion proteins were eluted with buffer A containing 10 mM maltose, concentrated in a speed-vac, dialysed against 20 mM Tris/HCl, pH 76, 5 mM MgCl,, 2 mM D T T and 2 pM GDP, and stored at -80 "C. Typically 20-25 mg pure fusion proteins were obtained from a 11 culture. Fusion proteins were digested with Factor Xa (New England Biolabs) in a 1:100 enzyme/substrate stoichiometric ratio, in the presence of 2 mM CaCl, for 4 h at 4 "C. Proteolytic reactions were stopped by adding PMSF to 0 2 m M and soybean trypsin inhibitor in a twofold molar excess over protease. SDS-PAGE analysis was performed using the Phast-System (Pharmacia), which was also used for immunoblots with anti-MBP (maltose binding protein) antibody (New England Biolabs) and anti-EF-Tu antibody from E. cofi (kindly provided by M. L. Nolli, this laboratory). Protein concentrations were determined with the dye-binding Bio-Rad Protein Assay kit, using BSA (Sigma) as reference protein. Protein synthesis assays. S30 extracts were prepared from cultures of E. cofi DH5a and Bacillus subtifis ATCC 6633 as described by Landini et a f . (1993). PolyU-directed polyphe synthesis was carried out essentially according to Traub et af. (1971). Briefly, assays were performed in 100 pl 30 mM
Streptomyces cinnamoneus EF-Tu
Tris/HCl, pH 7.7, 10 mM MgCl,, 80 mM NH4Cl, 3 mM DTT, 1 mM GTP, 0.8 mM ATP, containing 80 mg polyU, 0 6 mg E. coli tRNAPhe, 17 pmol L-Phe, 3.2 pmol ~ - [ ~ H l P h e (2.2 x lo6 Bq mol-', Amersham), a calibrated amount of S30 extract (to give 10000-20000 d.p.m.) and exogenous EF-Tu when appropriate. Reactions were carried out for 30 min at 30 "C, then stopped by the addition of trichloroacetic acid to 5% (w/v). After heating for 10 min at 80 "C, the precipitate was collected on glass fibre filters using a cell harvester (LKB) and the filter-associated radioactivity measured in a Beta-Plate counter (Pharmacia). T o block the endogenous EF-Tu, the S30 extract was pre-incubated for 5 min at 0 "C with various amounts of GE2270. The lowest antibiotic concentrations inhibiting protein synthesis activity over 95 o/' were experimentally determined as 015 and 0 4 mg ml-l for the E. coli and the B. subtilis extracts, respectively. Kr and GE2270 factor A, produced at Lepetit Research Center, were dissolved in dimethylsulfoxide and diluted in water immediately before use. Preparation of the S. cinnamoneus 5100 fraction. S. cinnarnoneus mycelium, grown in 3 0 g Tryptone Soya 1-1 (Oxoid) for 48 h at 30 "C, was washed with 10-3YO (w/v) sucrose and resuspended in lysis solution (Hopwood et al., 1985) containing 2 mg lysozyme ml-'. After 1 h at 30 "C, the suspension was sonicated on ice, centrifuged for 30 min at 30000 g and then for 4 h at 100000g (both spins were at 4 "C). The resulting SlOO fraction was mixed with NH4C1-washed E. coli ribosomes, prepared as described by Landini et al. (1993), and assayed for polyU-directed polyphe synthesis, as described above.
RESULTS Characterizationof the S. cinnamoneus tuf gene
A size-enriched library of S. cinnamoneus BamHI fragments (see Methods) was screened by sib selection, probing plasmid DNA isolated from pools of transformants with a tuf fragment. With this procedure we isolated plasmid pGE150, carrying a 3.8 kb insert, with the tuf gene located within the 1-8kb NcoI-BamHI fragment (Fig. 1). The nucleotide sequence of this segment revealed an ORF of 1194 nt, encoding a 396 aa polypeptide (after removal of the initial Met) with a calculated M , of 43760 and PI of 4.96. The translated
Fig. 7. Organization of the 5. cinnamoneus tuf region. A restriction map of the insert from pGE1SO is shown (only the Sacll, Smal and Nrul sites relevant to the text are indicated). The Pstl site shown in bold originates from the vector polylinker. The partial fus sequence is denoted by the grey broken arrow.
sequence was closely related (identity scores ranging from 91.7 to 92.2%) to the tufl gene products from S. ramocissimus (Vijgenboom et al., 1994), S. coelicolor (van Wezel et al., 1994) and S. collinus (Mikulik & Zhulanova, 1995); lower identity scores (59-9 and 61.0%)were observed with the tuf3 products from the two former species, and with the S. ramocissimus tuf2 product (84.4'YO identity). Southern hybridizations employing different digests of each genomic DNA were carried out to establish the number of tuf genes in S. cinnamoneus. When a first experiment was performed, the sequence of the S. coelicolor tufgenes had not been reported. We amplified a tuf segment from S. coelicolor DNA using consensus oligonucleotides (Sosio et al., 1996) and used this as the probe. A single band was invariably observed with S. cinnamoneus and N . lactamdurans DNA, while two bands of approximately equal intensity were observed with S. coelicolor DNA (data not shown). [The size and intensity of these bands were subsequently found to correspond to S. coelicolor tufl and tuf3 (van Wezel et al., 1994).]We then used the SmaI-NruI fragment from the S . cinnamoneus gene (Fig. 1) as the probe. Under moderate stringency, we could detect a strong and a weak band in each digest of S. coelicolor DNA, corresponding in size to tufl and tuf3, respectively. A similar hybridization profile (a strong and a weak band) was also observed with S. glaucescens and S. lividans DNA. However, only a single strong band was detected with S. cinnamoneus and N . lactamdurans DNA (data not shown). The results of these hybridizations strongly suggest that the latter two species possess a single tuf gene. A GC content of 63.6 YO for the S. cinnamoneus tuf gene is a relatively low but not unprecedented value for Streptomyces genes (Wright & Bibb, 1992).The frequent use of some T-ending codons (e.g. GGT and CGT account for 44 and 55% of the Gly and Arg codons, respectively) observed in tuf is believed to be characteristic of highly expressed genes in high GC microorganisms (Ohama et al., 1990; Wright & Bibb, 1992). Upstream to tuf we found the last 62 codons of an ORF highly related to fus, encoding EF-G. Downstream, 39 bp after the stop codon TAA, there is a putative transcription terminator (a 15 bp inverted repeat followed by six Ts), with a calculated AG of -47 kcal mol-l (Tinoco et al., 1973) for the corresponding RNA. The fus-tuforganization typical of tufl loci (Sosio et al., 1996; van Wezel et al., 1994; Vijgenboom et al., 1994) and the high relatedness to known tufl genes indicate that the single S. cinnamoneus tuf gene belongs to the tufl class. Interestingly, the segment immediately preceding the tuf GTG start codon was found to be highly conserved among the four Streptomyces tufi sequences available (Fig. 2). With the accommodation of a few gaps, there are 52/64 invariant positions among the four species. The extent of conservation declines considerably further upstream from this conserved segment. Conversely, no significant similarity among the four species was 619
C. CAPPELLANO and OTHERS
cGtTCaCaCgctttaGGC cGtTCtCaCgccgtaGGC CGtTCtCaCgctttaGGC aGcTCaCgCctttagGGC
Scol Sram Scoe Scin
-GTC-C-C------m
-60 AAAGATCACCTGGCGCCGa. AAAGATCACCTGGCGCCGat AAAGATCACCTGGCGCCGat AAAGATCACCTGGCGCCG.t
Scol Sram Scoe Scin
-140 -120 -100 -80 TTgACtCCGgagccT.cgtG GGgCAaaCaaCcgcaaAcaC GggtGtttgcCcCgGgttCC ggG.Cttt cCAGC TTgACtCCGgagtcT.cgcG GGcCAttCatCctcatAtgC GggtGaatggCcCgGggcCC ggGaCtta aCAGC TTgACcCCGgagccTgcatG GGgCAttCcgCcgtgaAccC GgtgGaatgcCcCcGgcaCC cgGgCttt. cCAGC TTtACaCCGactgtT..cgG GGgCAtaCcgCaggcaAtcC GaagGatgccCgCgGatgCC ccGgCggacggcacccCAGC
....... ....... ......
TT-AC-CCG----T----G
=CA--C--C-----A--C
G--G----C-C-G--CC
-4 0
-20
t1
tgaAgtAAGGCGTACAGAAC CACTCCacAGGAGGAcccCA QTQ tgaAccAAGGCGTACAGAAC CACTCCacAGGAGGAcccCA QPQ ..gAgtAAGGCGTACAGAAC CACTCCacAGGAGGAcccCA QPQ tgaAgtAAGGCGTACAGAAC CACTCCgtAGGAGGA..aCA
AAAGATCACCTGGCGCCG- ---A--AAGGCGTACAGAAC
GCG AAG GCG AAG GCG AAG GCG AAG CACTCC--AGGAGGA---CA QTQ GCG AAG f M A K
--GC-----------
CAGC
GCG GCG GCG GCG GCG A
.................................................................................................................................................. .......................................................................................................... Fig. 2. Alignment of the region upstream to tufl in S. collinus (Scol), 5. ramocissimus (Sram), 5. coelicolor (Scoe) and 5. cinnamoneus (Scin). The unmarked bottom line indicates the deduced consensus sequence. The translational start site is taken as position + 1.
SO-
SCOl
sram scin
SCOe SCOl
sram scin
21
90 DEGKlTLTM ITKVLEDAYP DIaEAslvDQ IDKAPKKRQR CITISIAWB YQTMRHYAE VDCPGEADYI F L T L TP 19 I L PFTP B S
dhgkttltm it-vl-----
1s1
fd- id-apee--r
giti---live
I -
y-t-rhyah
210
BYBFPGDDVP WKVSALXAL BGD-HSV LBuaAVDBA IPBPBRDVDK L R Q n o H L R MTQ D S D = 0 BKLC H Q
ve-evrell- ---f-gd--p
w--gal-a1
egd--w---i
--h-avd--
ip-p-rd-dk
..... 121
150
VLLARQvcVp YIWAtloluD MVDDBBILBL
x
w
S
v l l m q ~ g v p- i ~ - l n k 4-vdde--d
vd-pghady-
271 . . CQkGElW6LL LRCImDVB
310
RGQVIIKPGS VTPETEFBAQ
C
--agdu-g-1
C
T
lrgi-redv- rgqv--kpg- i-ph--f---
T
311 396 AYILSKDBGG RETPFIlolwR PQFYFRlTDV TGW'MAPW;T B(VIIPGDUTB HKVBLIQPVA MEBGWAIR B G G R T V W VTKINK Scol Q sram B R V SCW
scin
K
-yil-kdegg rhtpff--yr pqfyfrttdv tg-v-l----
A
T I
emvmpgd--- - - v - l i - p a
t
R T
-e-gl-fair
t
V
eggrtvgag- v--i--
......................................................................................................................................................................................................................
............................................................................
Fig. 3. Alignment of selected portions of Streptomyces EF-Tu sequences. Only positions different from the S. coelicolor EF-TU (Scoe) are indicated. The unmarked bottom line indicates the consensus sequence (18 out of 22 positions) calculated using the programs PILEUP and PRETTY, using a set of aligned EF-TU proteins (see Sosio et a/., 1996, for details). Major deviations from the consensus in the 5. cinnamoneus sequence are marked by arrows. Abbreviations are defined in the legend to Fig. 2.
Analysis of the deduced EF-Tu sequence
responding to 125 in E . coli) with a Ser in place of the invariant Val, and at position 378 (375 in E. coli) where a Thr substitutes the consensus Ala. This latter position is particularly intriguing, considering that the A375T mutation confers Krr to E. coli and Salmonella typhimurium (Parmeggiani & Swart, 1985;Abdulkarim et af., 1994; Mesters et al., 1994). Additional deviations from consensus are at positions 294 (Cys in place of the invariant Val) and 359 (Ala instead of Val/Ile/Thr).
Several single amino acid substitutions conferring Krr have been mapped on the EF-Tu sequence (Abdulkarim et af.,1994; Mesters et af.,1994). We thus checked the S . cinnamoneus EF-Tu for substitutions equivalent to those found in resistant mutants, as well as for deviations from the consensus in highly conserved positions. Four positions were identified in this way (Fig. 3). The most striking substitutions are found at position 127 (cor-
The availability of three highly related sequences from KrS Streptomyces EF-Tu proteins enabled a further inspection of amino acid residues peculiar to the S . cinnamoneus protein. The four sequences contain 327 identical residues out of 397. Only at 27 positions is one sequence different from the other three: the S . cinnamoneus EF-Tu has 18 unique residues, whereas the other three proteins differ at 1 , 3 or 5 positions (Fig. 3).
detected 3' to tuf. A putative ribosome binding site can be recognized in this conserved segment. However, no sequence clearly resembling any from a compilation of Streptomyces promoters (Strohl, 1992) could be detected. These observations suggest an important, although still unknown, function for the region preceding tuf2 in Streptomyces.
620
Streptomyces cinnamoneus EF-Tu
Interestingly, half of these unique residues are found in the 291-397 segment of the S. cinnamoneus EF-Tu.
> 1000
1000
Expression in E. coli
If the above-described deviations from consensus are sufficient per se to confer Krr, then expression of the S. cinnamoneus tufgene in a heterologous host such as E. coli should produce a resistant EF-Tu. However, any contamination of the recombinant protein by the endogenous KrS EF-Tu must be avoided, since this would interfere in protein synthesis assays, due to the dominance of KrS (Parmeggiani & Swart, 1985). T o achieve this separation easily, it was necessary to add an affinity tag to the recombinant protein. We tried the Cterminal hexaHis tag expression system (Boon et al., 1992); however, under the same conditions in which some soluble E. coli EF-Tu was obtained, the His-tagged S. cinnamoneus protein was completely insoluble ( S . Donadio, unpublished results). As an alternative method, we fused the tufgene to the 3' end of malE. The resulting fusion protein could be purified easily by affinity chromatography and subsequently released from the fusion partner by exploiting a Factor Xa cleavage site located in the linker region between the two moieties. The final EF-Tu product would contain, at its N terminus, one (Gly) or two (if Met is cleaved off in the natural protein) additional residues. SDS-PAGE analysis showed an IPTG-inducible, amylose-binding protein of the expected 85 kDa size which, after cleavage by Factor Xa, yielded two proteins of 47 and 40 kDa. These were identified as EF-Tu and MBP, respectively, by Western blotting with anti-EF-Tu and anti-MBP antibodies (data not shown). The intact fusion protein or the EF-Tu obtained upon cleavage contained a functional GDPbinding domain, as indicated by their ability to bind GDP-Sepharose. The recombinant EF-Tu was tested for its ability to catalyse protein synthesis in cell-free systems. We found that addition of the S. cinnamoneus protein to S30 extracts from E . coli or B . subtilis resulted, in both cases, in a marked increase in polyU-directed polyphe synthesis. A similar effect was shown by the uncleaved fusion and by E. coli EF-Tu, used as a positive control. This finding indicated that in both S30 extracts the amount of active endogenous EF-Tu present was limiting and that the S. cinnamoneus protein was functional. T o confirm this finding and to obtain quantitative data on the functionality of the recombinant protein, the endogenous EF-Tu in both S30 extracts was specifically inhibited by adding a calibrated amount of GE2270, which inactivates EF-Tu, but not ribosome function (Anborgh & Parmeggiani, 1991; Landini et al., 1992). After inhibition of the endogenous factor, the ability of exogenously added EF-Tu to restore protein synthesis was evaluated by comparing the S. cinnamoneus protein and the similarly produced P. rosea EF-Tu (Sosio et al., 1996) with the natural E . coli EF-Tu. The S. cinnamoneus protein restored full protein synthesis
100
=
k O
10
v
$
l i
1
W
g
1000
I=
a,
s
P,
W
100 10 1
E. coli
F! rosea
5. cinn.
T378A
Fig, 4. Activity of recombinant EF-TU proteins in protein synthesis. The bars denote the amount of protein required t o restore 50 % of the original polyU-directed polyphe synthesis activity t o GE22704nactivated 530 extracts from (a) E. coli and (b) B. subtilis. The respective activities before GE2270 addition were 13800 and 12600 c.p.m. Factor Xa-cleaved and uncleaved fusion proteins are indicated by white and grey bars, respectively. Natural E. coli EF-TU is indicated by the black bars. T378A denotes the 5. cinnamoneus mutant EF-Tu.
activity in both systems, both as an uncleaved fusion or as a cleaved product. Fig. 4 shows that similar amounts of S. cinnamoneus EF-Tu were required to restore 50% of the original activity in either system; however, the E . coli EF-Tu was active at much lower concentrations. The recombinant P. rosea EF-Tu behaved essentially as the S. cinnamoneus protein, although it was consistently less active as an uncleaved fusion (Fig. 4). The lower activity exhibited by the recombinant EF-Tu in comparison to the natural E. coli factor could be due to the additional residue(s) at the N terminus of the cleaved product. Protein preparations active in cell-free protein synthesis were obtained only after growing the E. coli host cells at 30 "C; higher temperatures resulted in totally inactive products. Nonetheless, our results show that N-terminal MBP fusions can represent a useful method to produce an EF-Tu easily resolvable from the endogenous factor and sufficiently active for characterization of its antibiotic resistance. Resistance to Kr
Before evaluating the behaviour of the recombinant EFT u towards Kr, its natural counterpart was assayed. To this end, an SlOO fraction (containing EF-Tu and all other soluble factors) from S. cinnamoneus was combined with E. coli ribosomes and tested in polyUdirected polyphe synthesis. This system was highly resistant to Kr, with only 40% inhibition observed at concentrations as high as 100 pg ml-l, in agreement with previous reports (Glockner & Wolf, 1984).As expected, 621
C. CAPPELLANO a n d OTHERS
-0
0.1
I
I
1
10
I
100
Kirromycin(pg ml-l) Fig. 5. Effect of Kr on EF-Tu. About 500 pmol of recombinant 5. cinnamoneus wild-type EF-TU ( 0 )and T378A EF-TU (D), and l00pmol of natural E. coli EF-TU (+) were added t o (a) untreated or (b) GE2270-inactivated8. subtilis 530 extracts. The dashed line in (a) indicates the effect of Kr with no exogenous EF-Tu added (about 15000c.p.m.), which was taken as 100%. In (b), the restored activity in the absence of Kr is taken as 100% for each EF-Tu. The data shown are representative of experiments performed at least two times.
complete loss of activity was observed at 1 pg GE2270 ml-'. The recombinant S. cinnamoneus EF-Tu was checked for Kr' employing the B. subtilis S30 extract, which contains an endogenous Kr' EF-Tu (Landini et al., 1993) and thus overcomes the problem of KrSdominance. The addition of S. cinnamoneus EF-Tu to this system resulted in high protein synthesis activity, well above control even at high Kr concentrations (Fig. Sa). Conversely, upon addition of the E. coli EF-Tu, the system was completely inhibited by 1 pg Kr ml-', as expected from the simultaneous presence of a KrSand a Kr' factor. The same results were observed employing a B. subtilis system where the endogenous EF-Tu had been inactivated previously by GE2270. Also in this case the activity restored upon addition of the S. cinnamoneus EF-Tu was only marginally affected by Kr (even at 100 pg ml-l), while the E. coli EF-Tu was highly sensitive (Fig. Sb). As expected, both proteins were instead completely inhibited by further addition of GE2270 (not shown). In conclusion, these results indicate that the E. coliderived S . cinnamoneus EF-Tu shows the same behaviour as the natural protein with respect to antibiotic sensitivity. This indicates that the S. cinnamoneus tuf gene, per se, is sufficient to produce a Kr' factor. Characterization of the T378A mutant EF-Tu
As described above, Thr378in the Kr' S. cinnamoneus EF-Tu corresponds to a Kr'-conferring mutation mapped in other EF-Tu proteins (Parmeggiani & Swart, 622
1985; Abdulkarim et al., 1994; Mesters et al., 1994). T o assess the contribution of this residue to Kr', the tuf gene was modified by site-directed mutagenesis (see Methods). The resulting T378A mutant protein was produced as an MBP fusion and purified as its wild-type counterpart. When assayed for protein synthesis activity, it behaved essentially like the parental EF-Tu, both in the E . coli and B. subtilis systems (Fig. 4), indicating that the introduced mutation had not significantly impaired activity. The response to Kr, in contrast, was very different between the two proteins : the T378A EF-Tu behaved essentially as a KrS protein (Fig. Sa), with most of the activity being lost at less than 1 pg Kr ml-'. However, considerable residual activity (30OO/ of the control) was retained even at 100 pg Kr ml-', in contrast with the KrS E. coli EF-Tu. When the experiment was performed with the GE2270-inhibited extract, the T378A protein again showed an apparently biphasic inhibition by Kr, with over 30% residual activity at 100 pg Kr ml-l (Fig. Sb). This result makes it unlikely that the observed residual activity was due to the endogenous Kr' B. subtilis EF-Tu. In conclusion, these data confirm the role of Thr378as an important determinant of Krr in the naturally resistant S. cinnamoneus protein. They also suggest that other residues in the EF-Tu primary structure may contribute to resistance. DISCUSSION
By expressing the tufgene in E. coli and characterizing the resulting product, we have found that all the information required for the synthesis of a Krr EF-Tu is contained within the single S. cinnamoneus tuf gene. The importance of Thr378in conferring Krr to the S. cinnamoneus EF-Tu, suggested by the gene sequence, was verified experimentally by characterizing the T378A mutant, which behaved essentially as a KrS factor. However, this mutant retained some residual protein synthesis activity at high Kr concentrations. This may suggest either a low affinity for the antibiotic, or alternatively that the mutant EF-Tu*Kr complex is somehow able to catalyse protein synthesis, although at much lower efficiency. The fact that the mutant activity decreased by 20% at 0-1pg Kr ml-l and that it did not change significantly between 10 and 100 pg Kr ml-l (Fig. Sa), argues against the former hypothesis. In any case, the residual activity of the T378A mutant strongly suggests that some other amino acid residue is involved in conferring protection from the antibiotic. Most of the mutations conferring Krr to the E . coli and S . typhimurium factors, when superimposed on the three-dimensional structure of the Thermus tbermophilus EF-Tu complexed with a GTP analogue (Berchtold et al., 1993), map at the interface region between domains I and 111. It has been postulated that Kr binds to this region, possibly at the interface itself (Mesters et al., 1994). Consistent with this hypothesis is the finding that many of the amino acid residues unique
Streptomyces cinnamoneus EF-Tu to the S. cinnamoneus sequence, when compared to the other Streptornyces EF-Tu proteins, occur within domain 111, particularly in the C-terminal 50 aa segment (Fig. 3). Among the deviations from consensus seen in the S. cinnamoneus EF-Tu, Ser12' (in place of the invariant Val) is particularly interesting. To our knowledge, this position has never been implicated in Kr'. Nevertheless, it is located at the domain 1-111 interface, adjacent to Gln124 ( E . coli numbering), a residue that when mutated to Lys confers high level Kr' (Zeef 8c Bosch, 1993).In addition, Ser127is spatially very close to Thr378 in the GTP-bound structure of the T . thermophilus EF-Tu (Berchtold et al., 1993). Most Streptomyces species possess multiple tuf genes, and a tuf3-like gene appears to be present in each case (van Wezel, 1994). For example, S. ramocissimus, a Kr producer with a KrS EF-Tu, has three such genes (Vijgenboom et al., 1994). We failed to detect a tuf3 homologue in S. cinnamoneus, however, although we cannot exclude the presence of a very divergent tuf3. In E. coli, due to the dominance of KrS, resistance can be usually selected in the presence of a null mutation in tufB (Parmeggiani & Swart, 1985).It is tempting to speculate that acquiring a Kr' EF-Tu in Streptornyces might require loss of the extra tufgenes, although tuf2 and tuf3 expression in S. ramocissimus has yet to be observed (Vijgenboom et al., 1994). [The only indication of tuf3 expression, restricted to a very narrow growth range, comes from S. coelicolor (van Wezel et al., 1995).] The existence of some Kr producers with a sensitive EFTu (Glockner & Wolf, 1984) suggests the existence of additional resistance determinant (s), likely to be linked to the Kr biosynthesis genes, as found in other antibiotic producing actinomycetes (Martin & Liras, 1989).These additional protective mechanisms may allow the evolution of a resistant target which is as functional as the sensitive one. This has probably occurred through multiple mutations in the tufgene, as suggested by the observation that the Kr' S. cinnamoneus EF-Tu has diverged somewhat from the other streptomycete tufl gene products, particularly in domain 111. This hypothesis is consistent with the finding that most mutants selected for Kr' appear to have an EF-Tu less active than its wild-type counterpart (Abdulkarim et al., 1994). The need for multiple substitutions to reconcile antibiotic resistance with full functionality has been recently reported for the rpsL gene involved in conferring streptomycin resistance (Schrag & Perrot, 1996). In conclusion, a resistant EF-Tu from an antibiotic producer can offer valuable insights not only into the molecular determinants of antibiotic resistance, but also into specific interactions important for normal functioning of this essential protein synthesis factor. ACKNOWLEDGEMENTS We are grateful to Khalid Islam, Anna Maria Puglia and Adolfo Soffientini for helpful discussions and suggestions, t o Elena Bossi for DNA sequencing and Wolfgang Guba for a molecular model of EF-Tu. T h e EF-Tu co-ordinates were
kindly provided by Prof. Rolf Hilgenfeld. We thank Khalid Islam for critical reading of the manuscript and are also indebted t o Maurizio Denaro for supporting this work.
REFERENCES Abdulkarim, F., Liljas, L. & Hughes, D. (1994). Mutations to
kirromycin resistance occur in the interface of domains I and 111 of EF-Tu.GTP. FEBS Lett 352, 118-122.
Alderson, G., Richtie, D. A., Cappellano, C., Cool, R. H., Ivanova, N. M., Huddleston, A. S., Flaxman, C. S., Kristufek, V. & Lounes, A. (1994). Physiology and genetics of antibiotic production and
resistance. Res Microbiol 144, 665-672.
Anborgh, P. H. & Parmeggiani, A. (1991). New antibiotic that acts specifically on the GTP-bound form of elongation factor Tu. EMBO ] 10,779-784. Berchtold, H., Reshetnikova, L., Reiser, C. 0. A., Schirmer, N. K., Sprinzl, M. & Hilgenfeld, R. (1993). Crystal structure of active
elongation factor Tu reveals major domain rearrangements. Nature 365, 126-132.
Boon, K., Vijgenboom, E., Madsen, L. V., Talems, A., Kraal, B. & Bosch, L. (1992). Isolation and functional analysis of histidine-
tagged elongation factor Tu. Eur J Biochem 210, 177-183. Cundliffe, E. (1989). How antibiotic-producing organisms avoid suicide. Annu Rev Microbiol43, 207-233. Devereux, J., Haeberli, P. & Smithies, 0. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res 12,387-395. Frohlich, K.-U., Wiedmann, M., Lottspeich, F. & Mecke, D. (1989).
Substitution of a pentanenolactone-sensitive glyceraldehyde-3phosphate dehydrogenase by a genetically distinct resistant isoform accompanies pentanenolactone production in Streptomyces arenae. ] Bacteriol 171, 66964702. Gltickner, C. & Wolf, H. (1984). Mechanism of natural resistance to kirromycin-type antibiotics in actinomycetes. FEMS Microbiol Lett 25, 121-124. Hopwood, D. A,, Bibb, M. J., Chater, K. F., Kieser, T., Bruton, C. J., Kieser, H. M., Lydiate, D. J., Smith, C. P., Ward, J. M. & Schrempf, H. (1985). Genetic Manipulation of Streptomyces. A
Laboratory Manual. Norwich : The John Innes Foundation.
Landini, P., Bandera, M., Goldstein, B. P., Ripamonti, F., Soffientini, A., Islam, K. & Denaro, M. (1992). Inhibition of
bacterial protein synthesis by elongation-factor-Tu-binding antibiotics MDL 62,879 and efrotomycin. Biochem J 283, 649-652. Landini, P., Bandera, M., Soffientini, A. & Goldstein, B. P. (1993).
Sensitivity of elongation-factor-Tu (EF-Tu) from different bacterial species to the antibiotics efrotomycin, pulvomycin and MDL 62879. ] Gen Microbiol 139,769-774. Martln, J. F. & Liras, P. (1989). Organization and expression of genes involved in the biosynthesis of antibiotics and other secondary metabolites. Annu Rev Microbiol43, 173-206. Mesters, J. R., Zeef, L. A. H., Hilgenfeld, R., de Graaf, J. M., Kraal, B. & Bosch, L. (1994). The structural and functional basis for the
kirromycin resistance of mutant EF-Tu species in Escherichia coli. EMBO ] 13,4877-4885. Mikulik, K. & Zhulanova, E. (1995). Sequencing of the tufl gene and the phosphorylation pattern of EF-Tul during development and differentiation in Streptomyces collinus producing kirromycin. Biochem Biophys Res Commun 213,454-461. Ohama, T., Muto, A. & Osawa, 5. (1990). Role of GC-biased mutation pressure on synonymous codon choice in Micrococcus luteus, a bacterium with a high genomic GC-content. Nucleic Acids Res 18, 1565-1569.
623
C. C A P P E L L A N O and O T H E R S
Parmeggiani, A. & Swart, G. W. M. (1985). Mechanism of action of kirromycin-like antibiotics. Annu Rev Microbiol39,557-577. Schrag, S. J. & Perrot, V. (1996). Reducing antibiotic resistance. Nature 381, 120-121.
Vijgenboom, E., Woudt, L. P., Heinstra, P. W. H., Rietveld, K., van Haarlem, J,. van Wezel, G. P., Shochat, S. & Bosch, L. (1994). Three
Sosio, M., Amati, G., Cappellano, C., Sarubbi, E., Monti, F. & Donadio, S. (1996). An elongation factor Tu resistant to the EF-
Weijland, A,, Harmark, K., Cool, R. H., Anborgh, P. & Parmeggiani, A. (1992). Elongation factor Tu: a molecular switch in protein
Tu inhibitor GE2270 in the producing organism Planobispora rosea. Mol Microbiol22,43-51. Strohl, W. R. (1992). Compilation and analysis of DNA sequences associated with apparent streptomycete promoters. Nucleic Acids Res 20,961-974. Thein-Schrammer, I., Zahner, H., Hopper, H.-U., Hummel, 1. & Zeeck, A. (1982). Metabolic products of microorganisms. Kirrothricin, a new member of the kirromycin-group. J Antibiot 35,948-956. Thiara, A. S. & Cundliffe, E. (1989). Interplay of novobiocin resistant and sensitive DNA gyrase activities in self-protection of the novobiocin producer, Streptomyces sphaeroides. Gene 81, 65-72. Tinoco, I., Jr, Borer, P. N., Dengler, B., Levine, M. D., Uhlenbeck, 0. C., Crothers, D. M. & Gralla, J. (1973). Improved estimation of secondary structure in ribonucleic acids. Nature 246, 40-41. Traub, P., Mizushima, S., Lowry, C. V. & Nomura, M. (1971).
Reconstitution of ribosomes from subribosomal components. Methods Enzymol20, 391407. Vieira, J. & Messing, J. (1982). The pUC plasmids, an M13mp7derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19, 259-268.
624
tuf genes in the kirromycin producer Streptomyces ramocissimus. Microbiology 140, 983-998.
biosynthesis. Mol Microbiol6, 683-688. van Wezel, G. P. (1994). Transcriptional regulation of trans-
lational genes in Streptomyces coelicolor A3(2). PhD Thesis, University of Leiden, The Netherlands. van Wezel, G. P., Woudt, L P., Vervenne, R., Verdurmen, M. L. A., Vijgenboom, E. & Bosch, L. (1994). Cloning and sequencing of the
tufgenes of Streptomyces coelicolor A3(2).Biochim Biophys Acta 1219,543-547. van Wezel, G. P., Takano, E., Vijgenboom, E., Bosch, L. & Bibb, M. J. (1995). The tuf3 gene of Streptomyces coelicolor A3(2) encodes
an inessential elongation factor Tu that is apparently subject to positive stringent control. Microbiology 141, 2519-2528. Wright, F. & Bibb, M. 1. (1992). Codon usage in the G+C-rich Streptomyces genome. Gene 113,55-65. Zeef, L. A. H. & Bosch, L. (1993). A technique for targeted mutagenesis of the EF-Tu chromosomal gene by M13-mediated gene replacement. Mol Gen Genet 238,252-260.
Received 29 May 1996; revised 27 August 1996; accepted 12 September 1996.
