T I B T E C H - APRIL 1990 [Vol. 8] 105 Microbial isopenicillin N synthase genes: structure, function, diversity and evolution Gerald Cohen, Dov Shiffman, Moshe Mevarech and YairAharonowitz Clinically and economically, penicillins and cephalosporins are the most important class of the ~-lactam antibiotics. They are produced by a wide variety of microorganisms including numerous species of Streptomyces, some unicellular bacteria and several filamentous fungi. A key step common to their biosynthetic pathways is the conversion of a linear, cysteine-containing tripeptide to a bicyclic [~-lactam antibiotic by isopenicillin N synthase. Recent successes in the cloning and expression of isopenicillin N synthase genes now permit production of a plentiful supply of this enzyme, which may be used for structural and mechanistic studies, or for biotechnological applications in the creation of novel ~-lactam compounds from peptide analogues. New ideas concerning the evolution and prevalence of the penicillin and cephalosporin biosynthetic genes have emerged from studies of isopenicillin N synthase genes. Many microorganisms have the capacity to synthesize [~-lactam antibiotics 1. Penicillins and cephalosporins are among the most valuable therapeutic agents of this kind and were originally discovered in a few filamentous fungi, simple eukaryoti c organisms. Subsequently they were found to be produced by numerous species of mycelium-forming Grampositive actinomycetes, especially within the subgroups Streptomyces and Nocardia. The salient features of penicillin and cephalosporin biosynthesis in these organisms are well established (Fig. 1). During the past few years, [3-1actam antibiotics have also been isolated from a variety of unicellular, Gram-negative bacteria including such diverse species as Pseudomonas, Agrobacterium, Serratia, Gluconobacter, Flavobacterium and Xanthomonas (Table 1). Prokaryotes produce not only the classical bicyclic sulfur-containing penicillins and cephalosporins, but G. Cohen, D. Shiffman, M. Mevarech and Y. Aharonowitz are at the George S. Wise Faculty of Life Sciences, Department of Microbiology, Tel Aviv University, Ramat Aviv 69978, Israel. cleic acid techniques to detect antibiotic biosynthetic genes. Recent advances in the cloning and expression of genes involved in penicillin and cephalosporin synthesis have given some partial answers to these questions and needs, tn this review, we focus on one of these genes, that coding for the enzyme isopenicillin N synthase (IPNS), and summarize present knowledge of its structure, mode of action, distribution in nature, and evolution. The IPNS genes IPNS catalyses a key reaction in the biosynthesis of penicillins and cephalosporins (Fig. 1). It mediates the oxidative conversion of the linear fripeptide, 6-(L-ec-aminoadipyl)I~-cysteinyl-I>valine (ACV), to isopenicillin N. Eight microbial IPNS genes have so far been cloned and sequenced, and most have been expressed in Escherichia coli. Three are derived from fungal strains 3-5, four from Gram-positive streptomycetes (Refs 6-8 and J. F. Martin, pers. commun.) and one has very recently been isolated and characterized from a Gram-negative unicellular prokaryote 29. also a wide range of structurally Table 2 summarizes some properrelated ~-lactam antibiotics. Among ties of IPNS structural genes. The these are the carbapenems and the primary amino acid sequences of the clavams, which have carbon and protein-coding regions of these genes oxygen atoms substituted for the (deduced from the DNA sequence) sulfur atom, respectively, and the are aligned for maximum homology monobactams which possess just a (Fig. 2), revealing the high degree of single ring. The overriding interest in conservation between the IPNS prothese compounds and in finding new teins from different sources. Com[~-lactams stems from the severe puter analysis of the sequence data clinical problems that still exist in indicates that the IPNS proteins the treatment of infectious disease, possess considerable similarity in and because of the unique mode of their secondary structure and hyaction of many of these antibiotics as dropathy features. There is no signifiinhibitors of bacterial cell-wall bio- cant homology between IPNS prisynthesis 2. mary amino acid sequences and The widespread occurrence of [~- those of other enzymes, implying that lactam-producing microorganisms in the IPNS proteins are unrelated to nature raises several intriguing any previously characterized class of questions: h o w closely related are the enzymes. genes and enzymes responsible for ~The overall sequence similarity of lactam synthesis; and what are the IPNS genes from different micromechanisms that account for the organisms is very high; in all seven existence of ~-lactam antibiotics in reported sequences, more than 60% certain groups of microorganisms but of the nucleotide bases and 50% of not others? More practical aspects of the deduced amino acid residues are ~-lactam production also need to be identical or have conservative readdressed, including the use of placements. Moreover, regions of enzyme-based processes for syn- extensive sequence similarity are thesis of novel antibiotics and the scattered, with no obvious pattern, development of sensitive microbial • throughout the genes and proteins. It screening procedures based on nu- is therefore not possible at present to ~) 1990, Elsevier Science Publishers Ltd (UK) 0167-9430/90/$2.00 106 TIBTECH- APRIL1990[Vol.8] --Fig. 1 define those regions of the IPNS proteins that are necessary for substrate binding and catalytic activity. The research group at Eli Lilly (Indianapolis, USA) has used sitedirected mutagenesis to examine the role of the two cysteine residues in IPNS from Cephalosporium acremonium 9. These cysteines are contained in a stretch of eight amino acid residues that are identical or conserved in all IPNS proteins (Fig. 2). Studies by Baldwin and coworkers 1° had established earlier that IPNS can exist in two interconvertible forms, presumed to be the cysteine disulfide and dithiol forms. Substitution of a serine for Cysl04 in the IPNS polypeptide chain reduced its activity by about 95%; replacement of Cys25 by a serine had a much less dramatic effect on activity. Thus, Cysl04 and Cys251 are important, but not essential, for activity. A further observation pointing to the importance for enzyme activity of the region surrounding Cysl04 is the finding that the penicillin ringexpansion enzymes of C. acremonium and Streptomyces clavuligerus (deacetoxycephalosporin C synthetase), which convert penicillin N to deacetoxycephalosporin C (Fig. 1), both possess a short cysteinecontaining decapeptide that shares 50% homology with that IPNS region 11. In addition, the Streptomyces and Flavobacterium IPNS enzymes contain one or two additional non-conserved cysteine residues located at positions 37 and 142; however, their importance for enzyme activity is not known. Another way to map functionally important IPNS amino acid residues is to use E. cob strains that express a cloned IPNS gene efficiently. If ACV is taken up by such cells and converted to isopenicillin N, they will be killed; some of the cells that survive should therefore possess a defective IPNS gene. The value of this approach is that it involves direct selection for IPNS mutants. A technique that may prove useful in identifying protein domains involved in substrate binding is the use of antibodies directed against short synthetic peptides corresponding to highly conserved regions of IPNS. Specific inhibition of IPNS activity by such antibodies could narrow the choice of possible substrate-recognition or -binding domains (but L-0c-AAA-I- LCys-t- LVal H /SH N4kL N H 3_O2 N +c ~ ~ v " ~ ~ I Oo _ ~,~cH3 LLD-ACV _ -DY/ , , , OH3 CO2H H ~ Isopenicillin N synthase H N ~ , ~ "I S~,CH3 H 3 N + ~ ~ '~N u -O2C Penicillin G O IsopenicillinN D ' ~ '%'Gas H#'''CO2H ••'•imerase Penicillin N l deacetoxycephalosporin C synthetase DeacetoxycephalosporinC DeacetylcephalosporinC CephalosporinC Cephamycins Penicillin and cephalosporin biosynthesis in fungi and bacteria, illustrating the role of isopenicillin N synthase in the formation of /J-lactam antibiotics. L-O~AAA, aminoadipic acid; ACV, 6-(L-~-aminoadipyl)-L-cysteinyI-D-valine. might, misleadingly, imply the involvement of domains essential for the three-dimensional structure of the enzyme in direct substrate binding). Alternatively, by designing lowmolecular-weight oligopeptides that resemble the conserved regions of the IPNS molecule and that compete for ACV in cell-free IPNS reactions, it may be possible to locate the substrate-binding site. A recently described photoaffinity method for labelling IPNS promises to be valuable for investigating the binding of substrates to IPNS 12. Covalent labelling of IPNS was carried out by laserflash photolysis in the presence of a diazirinyl-containing substrate analogue of ACV. Molecular graphic techniques based on the structure of ACV and its analogues, which aim to explore the architecture of the binding site, may help to identify the substrate-binding site. Since we do not yet have an X-ray crystal structure of IPNS, studies of this kind will be invaluable for undertaking a rational genetic analysis of enzyme structure-function relationships. IPNS catalysis Until recently, only small amounts of IPNS proteins could be prepared from the soluble fraction of peni- TIBTECH - APRIL 1990 [Vol. 8] 107 ~Table I Distribution of [3-1actam antibiotic-producing microorganisms Bacteria Class o f 13-1actam Fungi Pen a rn Aspergillus Penicillium Epidermophyton Trichoph yton Polypaecillum Malbranchea Pleurophomopsis Cep h e m Cephalosporium Spiroidium Scopulariopsis Diheterospora Paecflomyces Gram-positive Gram-negative Streptomyces Nocardia Flavobacterium Xanthomonas Lysobacter Clavam Streptom yces Carbapenem Streptomyces Serratia Erwinia Monobactam Nocardia Pseudomonas Gluconobacter Chromobacterium Agrobacterium Acetobacter cillin- and cephalosporin-producing cells. In Flavobacterium, an inactive membrane-bound form has been detected 13, and in C. acremonium and S. clavuligerus there are reports of latent membrane-bound IPNS activity. Since most of the cloned IPNS genes can now be expressed efficiently in E. coli, substantial amounts of enzyme should become available for structural and mechanistic studies. The cloned proteins appear to be indistinguishable in their gross biochemical properties from those obtained from the parent microorganisms, but in the case of C. acremonium, the cloned protein lacks the N-terminal methionine and glycine residues predicted from the DNA sequence ~4. All the enzymes have an absolute requirement for ferrous ion and for an electron donor, generally provided as ascorbate, and use molecular oxygen as cosubstrate. However, careful comparison of the biochemical properties of the bacterial and fungal enzymes with respect to different peptide substrates has not been carried out. Studies of the conversion of isotopically labeled ACV to isopenicillin N showed that four hydrogen atoms are removed from the tripeptide with consumption of one molecule of oxygen. The two new bonds, C-N and C-S, of the ~-lactam and thiazolidine rings, respectively, appear to be made in a stepwise fashion with retention of configuration and with intermediates attached to the enzyme ~5. Although the detailed mechanism of formation of the ~-lactam ring is still uncertain, Baldwin has proposed that the cysteine thiol in ACV is probably used initially to form a bridge between the 3-cysteinyl C-H bond and the dioxygen molecule via an iron atom 15, and that oxidation of the C-H bond is coupled with reduction of the oxygen. Once the oxo-iron species is formed at the active site, the reaction is driven forward by the intense energy of the chemical intermediates. IPNS recognizes a broad range of tripeptide substrates. A large variety of novel isopenicillin-like molecules have been synthesized in vitro starting from ACV analogues 15,1~. Some of the most interesting of these, from the point of view of the [~-lactam products and for the insight they provide into the reaction mechanism, have a substitution of the D-valine residue with other D-amino acids. Unusual bicyclic ~-lactams made in this w a y include molecules containing six-, seven- and eight-membered rings fused to the ~-lactam ring. It is possible, in a cell-free system, to couple IPNS (a cyclase) with two successive biosynthetic enzymes, the epimerase and the ring-expansion enzyme (deacetoxycephalosporin C synthetase) (Fig. 1), to create cephalosporins from tripeptide analogues 16. Some of these unnatural [3lactams are useful antibiotics, acting on both Gram-positive and Gramnegative bacteria. Clearly, such antibiotics cannot be obtained by conventional fermentation methods since the peptide analogues are not cellular metabolites. Practical in-vitro synthesis of [3-1actams will depend, therefore, on generating substantial quantities of IPNS enzyme, which should be possible through expression of cloned IPNS genes. Considerations of this sort will intensify efforts to create recombinant IPNS proteins with altered substrate specificity and enzyme activity. One potentially useful w a y to achieve this end is to make use of conserved restriction ~Table 2 Properties of isopenicillin N synthase genes Microorganism Genome %GC IPNS ORF a IPNS % GC Sequence identity b Cysteine codons Flavobacterium S. lipmanii S. jumonjinensis S. clavuligerus C. acremonium P. chrysogenum A. nidulans 70 71 71 71 55 52 326 331 329 329 338 331 331 64 66 64 66 63 56 52 100 69 69 68 65 64 61 3 3 3 4 2 2 2 a Number of amino acids in open reading frame; bcomparison of nucleotide sequence with Flavobacterium IPNS gene. 108 T I B T E C H - A P R I L 1990 [Vol. 8] --Fig. 2, S. clavuligerus S. jumonjinensis S, lipmanii Flavobacterium A. nidulans C. acremonium P. chrysogenum S. clavuligerus S. jumonjinensis S. lipmanii Fla vobacterium A. nidulans C. acremonium P. chrysogenum mmmm::mmm • LIMIPISIA E V PITII DIIIS P L PL MN R H A D V P V I D~S G L M G S V S - - K ArNlv PFII DIVIS P L M G S V P V P V AINIV PIRII DIVlS P L b.~v LIRIPISIA D V PITII DHS rim • mn I mm:~D • Elm FY,T..,VOVOOL N,,HOA [] S G D D A K A K Q B V A ~ E I N K A A RIGIS G F F Y A S N H G V D V Q L LIQ D VVIN E F H R N ~ F F G T D P D A K A H V A Q I N E A C RIGIS G F F Y A S H H G I D V R R LIQ~LE.~N E F H R T~4[r S G N D M D V K K D T A A R Irma A C RIGIS G F F Y AA N H GV D L A A L ~ K FIT1T DW H M A. ~ F G D D Q A A K M R V A Q Q lID A-IAr~BrDIT G F F Y AF~N H G I N V Q R LISIQ KITIK E F H MIISlIT F G D D K E K K L E V A B A I p AIAISIRIDIT G F F Y ALVIN H G V D L P W LISIR EITIN K F H M~S I F MAST PKAINIVPKL~ID~JSPL FGDNMEEKMKVARAI~) AJAS[SJRIDITGFFYAVLV~NHGVDVKRLS~SJNKTT [ JREFHFS ~ 67 67 64 66 68 66 il~l!~!~ii~ i~li?;~!~iPGRKTVESWCYLNPSFGEDHP r~ ~ m m r~mIKAGTPMHEVNVWPDEELE]R ,M~ ~ - l ~ F 134 KGKKAVESFCYLNPSFSDDHP IKSETPMHEVNLWPDEIE 134 PGRKAVESFCYLNPDFGEDHP i~ii~iiii~i)!i~i~')ii~iiiCi IAAGT 134 EG GK KK KA AV NE ES S FF C CY Y LL N NP PN SF FT D~ADDHHPARTIIQ KA AK GTL P PT SH HE V N V W PEDVENTI K WPDEAR P PGKKAVESFCYLNPSFSMDHPRIKEPTPMHEVNVWPDEAK PEKKAVESFCYLNPNFK~JDHPLTQSKTPTHEVNVWPDEKK 11 33 14 136 134 mmmm~ [] m r-mm m 7 m m • • • rim II HPRFRPFCEGYY~QML~LST~LMRGLALALGRPEHFFDAALAEQDSLSSV~LIRYPYLEEYPP--VKT 200 HPRFRPFCEDYY~QLL~LST~IMRGYALALGRREDFFDEALAEADTLSSV~LIRYPYLEEYPP--VKT 200 HPDFRSFGEQYYR~VF~LSK~LLRGFALALGKPEEFFENEVTEEDTLSAV~MIRYPYLDPYPEAAIKT 202 HP~MRRFYEAYF~DIVFDVAA|VlILRGFAIALGR~ESFFE~'-~FSMDDTLSAV~LIRYPFLENYPP--LKL 197 HP~FQDF~EQYY~VF~LS~-~LLKGYALALGK~ENFFA~HIF~PDDTLASV~LIRYPYLDPYPEAAIKT 202 HP~FRAF~EKYY VF~LSISAIVLRGYALALGR~EDFFT~HIS~RDTTLSSV~LIRYPYLDPYPEPAIKT 204 HP~FREF~EQYY VF~LSISAILLRGYALALGK~EDFFS~HIF~KEDALSSV~LIRYPYLNPIPPAAIKT 202 m S. clavuligerus S. jumonjinensis S. lipmanii Flavobacterium A. nidulans C. acremonium P. chrysogenum m S. clavuligerus S. jumonjinensis S. lipmanfi Fla vobacterium A. nidulans C. acremonium P. chrysogenum S. clavuligerus S. jumonjinensis • lipmanii Fla vobacterium A. nidulans C. acremonium P. chrysogenum m m m _ m : i J D m mm7 [] r-m-m-, rim mm~ m LFQTQVQNLQVETF~GWRDI~T~ENDFLVNCG~YMAHVTNDYFPAP IiDGQLLSFE~H~DVSNIT ;DGTKLSFE~DVS~ITVLYQTEVQNLQVETIVD[GWQDI~R[S~EDFLVNCGIT~MGHITHDYFPA~ DGTRLSFE~JH~DVS~ITVLFQTEVQNLQVETIVDIGWQSL~TIS~ENFLINCG[TIYLGYLTNDYFPA DGEKLSFEHHQDVSHITVLYQTAIPNLQVETAEGYLDIIP[VIS~EHFLVNCGITffMAHITNGYYPAPV DGTKLSFE~H~DVS~ITVLYQSNVQNLQVETAAGYQDIE--T~TGYLINCG~YMAHLTNNYYKAP~ DGTKLSFE~IH~DVS~ITVLYQSDVQNLQVKTPQGWQD!QIAD~TGFLINCGIS[YMAHITDDYYPA;U DGTKLSFE~J~DVS~ITVLYQSDVANLQVEMPQGYLDIEJADIDNAYLVNCG~YMAHITNNYYPA --- EPFVP KPFHP 267 267 269 264 269 271 269 _EGASEEVR~_EAL~I S Y G D Y L Q H G L R ~ L I V K N G Q329 T -EGAAGTVK~-PTTSYGEYLQHGLR~LIVKNGQT 329 m m m -EDTGDRKL~-PAVTYGEYLQEGFH~LIAKNVQT i i i i ~ i l I ~ I l i i ~ i i ! i i ! D P F A P P P Y A PDPFDP--REPN P ~ - - - G ~ - P T V SS---~REPLSYGDYLQNGLVSLI~KNGQT YGDYLQHGLLDLIRANGQT QPWDPATAKDGA~DAAK~KPAISYGEYLQGGLRGLI~KNGQT QPWDP--SKEDG~T---~QRPISYGDYLQNGLVSLI~KNGQT 331 326 331 33B 331 A l i g n m e n t o f the derived p r i m a r y amino acid sequences of the enzyme-encoding regions o f seven IPNS genes. Filled s q u a r e s above sequences indicate positions of amino acid residues that are present at the same site in all seven sequences; open squares indicate positions o f similar but not identical residues that are present at the same site in all s e v e n sequences. Rectangular open boxes are drawn around amino acid residues that are conserved in streptomycete proteins; shaded boxes s h o w residues that are conserved in fungal proteins. Arrowheads indicate the two cysteine residues present in all seven sequences. sites in IPNS genes, or introduce new ones, to create hybrid IPNS genes 17. Diversity of IPNS genes in nature One of the most exciting recent developments in the field of ~-lactam antibiotics has been the discovery of n e w types of [3-1actam antibiotics and of new microbial species that synthesize them (see Table 1). H o w widespread are the ~-lactam producers in the microbial world? The common practice of screening for antibiotic-producing organisms is based on detecting their biological activity or chemical characteristics. Biological assays employing agar plates have been the primary criterion for distinguishing nonproducers from producers. Even with the use of highly sensitive indicator strains, these methods may fail to detect very low levels of antibiotics, or antibiotics with weak activity. Moreover, many of the actinomycete ~-lactam-producing species also make other classes of antibiotic which may interfere with the screening of microorganisms for [3-1actams; for example, Streptomyces griseus excretes a bewildering variety of different antibiotics. Equally frustrating is the fact that penicillinand cephalosporinproducing strains may possess potent enzymes that degrade or modify these antibiotics. The ~-lactamases are the best-known example of enzymes that destroy penicillins and cephalosporins, and it has been demonstrated that a cephalosporinproducing species of Flavobacterium also makes a [3-1actamase (H. von Dohren et al., pers. commun.). [A fascinating digression on this theme is the finding that potent inhibitors of ~-lactamase exist, such as the clavams and carbapenems (thienamycins) discovered in S. clavuligerus and Streptomyces cattleya, which are themselves [3-1actam antibiotics. Furthermore, these same streptomycetes also synthesize penicillins and cephalosporins, suggesting a complex regulatory interrelationship between [~-lactam antibiotic, putative ~-lactamase and [3-1actamase inhibitor.] Ingenious procedures for detecting low levels of fi-lactams and other inhibitors of cell-wall biosynthesis have been developed 18. Many of these depend on the use of enzymebased screening procedures in which the antibiotic is assayed by its inhibitory effect on a selected en- TIBTECH- APRIL 1990 [Vol. 8] 109 --Fig. 3 A 1 2 B 3 4 1 2 C 3 4 1 2 3 4 : ! !"i " .i~"~I~:i :, ~" ~i ~:4=:~i~ ~ = = =-:~ :=/== i:: =i::'!~ i= Hybridization analysis of genomic DNA from different streptomycetes with three IPNS DNA probes. Total cellular DNA of two isolates of S. g riseus (one of which is a producer) and S. lividans and S. venezuelae (non-producer strains) was digested with a restriction endonuclease and the DNA fragments separated by agarose gel electrophoresis and transferred to nitrocellulose filters for hybridization with radiolabeled IPNS probes from S. clavuligerus (A), S. lipmanii (B), and S. jumonjinensis (C), followed by autoradiography. Lanes 1, S. griseus (fi-lactam producer); 2, S., griseus (non-producer); 3, S. lividans; 4, S. venezuelae. zyme, such as the cell-wall o-alanine carboxypeptidase. A variation of this approach employs 6-1actamase as the target enzyme. Used in conjunction with a suitable penicillin or cephalosporin substrate containing a chromogenic or fluorogenic reporter group, this approach has identified 6-1actamase inhibitors, many of which have turned out to be 6-1actams. While some of these enzyme screening techniques are extremely sensitive, in particular that based on the induction of Bacillus licheniformis 6-1actamase TM,it is desirable to develop alternative techniques to detect antibiotic-biosynthesis genes in microorganisms. Such techniques would reflect the genetic capacity for antibiotic production in a given strain rather than its actual production. Hybridizatio n methods have been used recently to screen non-penicillin-producing species of Streptomyces for the presence of IPNS genes 7. Several DNA probes containing known streptomycete IPNS genes are hybridized with genomic DNA from non-producing strains - the approach is based on that used by Hopwood and co-workers to detect polyketide genes in Streptomyces2°. The results were unexpected (Fig. 3). Some non-producing species responded to one or more of the IPNS probes: for example, two independent isolates of S. griseus that were characterized as penicillin and cephalosporin non-producers hybridized strongly to two of the IPNS probes. A third S. griseus antibioticproducing isolate hybridized to all three probes. A non-producer strain of Streptomyces lividans hybridized weakly to one of the three IPNS probes, whereas a non-producer strain of Streptomyces venezuelae did not hybridize at all. This suggests that IPNS-like genes may occur in microorganisms that have, until now, been considered non-producers. It raises the possibility that the inability of certain Streptomyces strains to make penicillins and cephalosporins may be a consequence of their containing defective or silent IPNS genes (i.e. genes that are, for unknown reasons, poorly expressed). For example, some wild-type strains of Aspergillus nidulans appear to be unable to make ACV despite having an otherwise complete penicillin-biosynthesis pathway 21. The idea of silent or cryptic antibiotic synthetase genes is not n e w - it was first described for the gene encoding phenoxazinone synthetase, a key enzyme in the biosynthesis of actinomycin, in a species not previously known to produce that antibiotic 22. An additional reason that ~-lactams might not be produced in certain streptomycete strains is the presence of defects in structural genes other than IPNS or in a regulatory element needed for antibiotic synthesis. DNA hybridization is currently being used in our laboratory to screen a collection of unicellular Gramnegative microorganisms for the presence of IPNS-like genes. The choice of the IPNS gene for these studies is dictated, in large part, by the fact that IPNS occupies an early and key role in the biosynthetic pathway of all the penicillin and cephalosporin antibiotics. IPNS probes are therefore expected to be reliable sensors of production potential of such antibiotics, even when the product may turn out to be a new O-lactam. To improve the chance of identifying distantly related sequences, several IPNS DNA probes, including those from Streptomyces and from the recently isolated IPNS gene of Flavobacterium, are used. With this approach, we have demonstrated the presence of an IPNS-like gene in a species of Xanthomonas (unpublished). Detection may be further improved by amplification of the signal using the polymerase chain reaction (PCR) technique to screen chromosomal DNA. We plan to extend our analysis to cover a much broader group of prokaryotic and eukaryotic microorganisms. If IPNS and other antibiotic synthetase genes are more commonly distributed in the microbial world than has been previously thought, their role, still very much a debated issue, may need to be re-examined. The products of the antibiotic syntheta.se genes may possess functions as yet unrecognized. Evolutionary relationships in IPNS genes Perhaps the most remarkable feature of the microbial IPNS genes is the extremely high degree of sequence homology they exhibit. The predicted primary amino acid sequences of fungal and bacterial IPNS proteins share 55-60% identity. 110 (These values take into account only identical matches of amino acid residues and ignore conservative replacements.) Such high values for sequence comparisons between eukaryotic and prokaryotic proteins are rare, and are generally attributed to relatively slow changes occurring in those genes during evolution, presumably as a consequence of strict functional requirements for many of the constituent amino acids and because of the importance of the product to the organism. This view seems, however, unlikely to be correct, since the nucleotide sequence differences within the fungal and streptomycete IPNS genes appear to be consistent with normal rates of evolutionary change and because IPNS genes do not appear to be essential for cell growth. A quite different explanation proposed by the group at Eli Lilly8 on the basis of a comparison of fungal and Streptomyces IPNS genes - is that the identity is a consequence of a lateral [horizontal] gene transfer between prokaryotes and eukaryotes that took place some 370 million years ago, well after they diverged about 2 billion years ago 23. According to the hypothesis, this ,was probably the result of a single transfer event from Streptomyces to the fungi because the former have a more elaborate biosynthetic pathway for ~-lactam antibiotics. Furthermore, in one of the fungal strains, C. acremonium, the ~-lactam biosynthesis genes are located on separate chromosomes. We recently proposed a modified version of this transfer hypothesis, taking into account new sequence data for the IPNS gene of a penicillinproducing species of Flavobacterium that belongs to a third major group of microorganisms, the Gram-negative unicellular bacteria. The phylogenetic tree for IPNS genes (Fig. 4) is based on our analysis of sequence data using a parsimony method 24. The gene tree has three major features: first, the IPNS genes of the fungi, Streptomyces and Flavobacterium fall into three welldefined groups that correspond to the known species tree; second, the IPNS genes of all the species are extremely similar despite the fact that they belong to widely different organisms; and third, the similarity of IPNS genes creates a marked distortion in TIBTECH7-APRIL1990[Vol.8] ~Fig. 4. P. chrysogenum Eukaryotes S. jumonjinensis L[~Gram-p°sitive S. clavuligerus Prokaryotes S. lipmanfi t Lateral transfer I -negative I 2 1-1.5 Flavobacterium I 0.3 (Billionyears) Phylogenetic tree of IPNS genes derived by the method of m a x i m u m parsimony, and using the neighbor-joining method 28. Branch length is proportional to the number of informative substitutions per site. The root of the tree was placed at the middle point of the longest branch connecting two species, assuming a constant rate of evolutionary change. Distances between IPNS genes are measured by the s u m of the horizontal branches connecting them. Estimated times of known events in evolution, in billion years, splitting of fungal species, splitting of Gram-positive and Gram-negative bacteria and spfitting of prokaryotes and eukaryotes are based on 5S and 16S RNA studies 23 the evolutionary time-scale separating the splits between the eukaryotes and prokaryotes on the one hand and the Gram-positive and Gramnegative bacteria on the other. The transfer scheme depicted in Fig. 4 accounts satisfactorily for the properties of the IPNS gene tree, provided that a single transfer event occurred close to the divergence between the Gram-positive and Gram-negative bacteria, estimated as 1.0-1.5 billion years ago 23'25. If, as originally thought, the transfer event took place only about 370 million years ago (an estimate based on assumptions of evolutionary rates), the Streptomyces IPNS genes would be expected to resemble the fungal IPNS genes more closely than the Flavobacterium IPNS gene. The IPNS genes of the Streptomyces and fungi would have evolved separately over only 370 million years, whereas the evolution of the Streptomyces and Flavobacterium IPNS genes from a common ancestral gene would have occurred over a much longer period, corresponding to the split between the two groups of eubacteria. The IPNS gene tree apparently rules out this possibility. Otherwise, it is necessary to postulate that multiple transfer events occurred at about the same time between Gram-positive, Gram-negative and eukaryotic microorganisms. A quite different hypothesis, which cannot be entirely ruled out, is that the rate of evolution of IPNS genes is not constant in all lineages. For example, if the rate were much slower during the initial stages following the eukaryotic/prokaryotic divergence than at later times, this would result in high sequence similarity and, consequently, a distortion of the IPNS gene tree. It would be possible to test the transfer hypothesis if an IPNS gene (or, more plausibly, a related gene that shares a common ancestor, provided their divergence occurred before the eukaryotic-prokaryotic split) were to be identified in the archaebacteria. In this case the o u t g r o u p sequence could be used to locate the root of the phylogenetic tree and the conclusion would then be based on a topological argument and be independent of rate or molecular clock assumptions. An important prediction of the transfer hypothesis is that certain other ~-lactam biosynthetic genes TIBTECH - APRIL 1990 [Vol. 8] s h o u l d also display the same high degree of s e q u e n c e identity in w i d e l y different microorganisms as is f o u n d for the IPNS genes. Indeed, it has r e c e n t l y b e e n s h o w n that the C. a c r e m o n i u m and S t r e p t o m y c e s lipm a n i i genes encoding penicillin r i n g - e x p a n s i o n e n z y m e s (deacetoxyc e p h a l o s p o r i n C synthetase) and their p r e d i c t e d proteins possess almost the same percentage sequence i d e n t i t y as do t h e c o r r e s p o n d i n g IPNS genes and proteins from these species 26. More extensive studies of IPNS genes, especially those belonging to p r e v i o u s l y u n c h a r a c t e r i z e d species, and genes encoding other ~lactam biosynthesis e n z y m e s m a y therefore resolve the q u e s t i o n of w h e t h e r a transfer m e c h a n i s m - a controversial issue in other areas of e v o l u t i o n - best accounts for the s p r e a d of IPNS genes in nature, and if so w h e n exactly this h a p p e n e d . 111 higher species, we may expect to find answers to some f u n d a m e n t a l questions c o n c e r n i n g the evolution of these extraordinary biosynthetic systems and what precisely is their biological role. Acknowledgements The research carried out at Tel Aviv U n i v e r s i t y was s u p p o r t e d by a joint grant from the National Council for Research and Development, Israel and G. S. F. Mfinchen, FRG. We w o u l d like to thank Prof. A. L. Demain and Dr D. Grauer for fruitful c o m m e n t s and discussion and Mrs H. Koltai for technical assistance. References 1 Elander, R. P. (1983) in Antibiotics Containing the Beta-Lactam Structure 2 Conclusions M a n y of the ideas s u m m a r i z e d here dealing w i t h IPNS genes m a y soon be, or are already being, i m p l e m e n t e d with other penicillin and cephalosporin biosynthetic genes. An obvious c a n d i d a t e for s t u d y is ACV synthetase (see Fig. 1) w h i c h is a large, c o m p l e x multie n z y m e system 27. By analogy with IPNS it m a y be possible to exploit this e n z y m e , w h e n it becomes available in sufficient amounts, to generate n o v e l p e p t i d e substrates for penicillin and c e p h a l o s p o r i n synthesis in a m o r e e c o n o m i c a l fashion than traditional chemical synthesis. As IPNS and ACV synthetase can be coupled, enzymatic methods may p r o v e to be a powerful ally to f e r m e n t a t i o n practice in p r o d u c i n g u n u s u a l penicillins. Similar considerations a p p l y to biosynthetic reactions i n v o l v e d in e x p a n d i n g and m o d i f y i n g the penicillin nucleus. In each case a major objective will be to use r e c o m b i n a n t DNA t e c h n i q u e s to alter in a d e f i n e d w a y the properties of the antibiotic biosynthetic enzymes. It seems likely that this kind of a p p r o a c h for e x p a n d i n g the range of ~-lactam antibiotics will be paralleled by m o r e sophisticated ways to search for n e w e n z y m a t i c activities and ~-lactam c o m p o u n d s in nature. F u r t h e r m o r e , as w e delve d e e p e r into the distribution of ~-lactams among the microbes, and potentially in 3 4 5 6 7 8 9 I (Demain, A. L. and Solomon, N. A., eds), pp. 97-146, Springer-Verlag Zimmerman, S. B. and Stapley, O. S. (1983) in Antibiotics Containing the Beta-Lactam Structure I (Demain, A. L. and Solomon, N. A., eds), pp. 285-300, Springer-Verlag Samson, S. M. et al. (1985) Nature 318, 191-194 Carr, L. G., Skatrud, P. L., Scheetz, M. E., Queener, S. W. and Ingolia, T. D. (1987) Gene 48, 257-266 Ramon, D., Carramolino, L., Patino, C., Sanchez, F. and Penalva, M. A. (1987) Gene 48, 171-181 Leskiw, B. K. et al. (1988) Gene 62, 187-196 Shiffman, D., Mevarech, M., Jensen, S. E., Cohen, G. and Aharonowitz, Y. (1988) Mol. Gen. Genet. 214,562-569 Weige], B. J. et al. (1988) J. Bacteriol. 170, 3817-3826 Samson, S. M., Carr, L. G. and Ingolia, T. D. (1987) Proc. Nati Acad. Sci. USA :: i::-::: :: : Next 84, 5705-5709 10 Baldwin, J. E., Cagnon, J. and Ting, H-H. (1985) FEBS Lett. 188, 253-256 11 Samson, S. M. et al. (1987) Bio/ Technology 5, 1207-1214 12 Baldwin, J. E., Coates, J. B., Halpern, J. B., Moloney, M. M. and Pratt, A. J. (1989) Biochem. J. 261,197-204 13 Palissa, H. et al. (1990) J. Bacterio]. 171, 5720-5728 14 Baldwin, J. E. et al. (1987) J. Antibiotics 40, 652-659 15 Baldwin, J. E. and Abraham, E. P. (1988) Natural Prod. Rep. 5,129-145 16 Wolfe, S., Demain, A. L., Jensen, S. E. and Westlake, D. W. S. (1984) Science 226, 1386-1392 17 Floss, H. G. (1987) TrendsBiotechnol. 5, 111-115 18 Nisbet, J. L. (1982) J. Chem. Tech. Biotechnol. 32, 251-270 19 Sykes, R. B. et a]. (1981) Nature 291, 489-491 20 Malpartida, F. et a]. (1987) Nature 325,818-821 21 Cole, D. S., Holt, G. and Macdonald, K. D. (1983) J. Gen. MicrobioL 96, 423-426 22 Jones, G. H. and Hopwood, D. A. J. Biol. Chem. 259, (1984) 14158-14164 23 Hori, H. and Osawa, S. (1979). Proc. Natl Acad. Sci. USA 76, 381-385 24 Landan, G. et al. Mol. Biol. Evol. (in press) 25 Ochman, H. and Wilson, A. C. (1987) J. Mol. Evol. 26, 74-86 26 Kovacevic, S., Weigel, B. J., Tobin, M. B., Ingolia, T. D. and Miller, J. R. (1989) J. Bacteriol. 171, 754-760 27 Banko, G., Demain, A. L. and Wolfe, S. (1987) J. A m . Chem. Soc. 109, 2858-2860 28 Nei, M. and Gojobiri, T. (1986) Mol. Biol. Evo]. 3, 418-426 29 Shiffman, D. et al. Nucleic Acids Res. (in press) TO Is: of ¼he Trade : :::Tools:Of ~he Trade i~ :aft o:~casio~i:fea~ure fnTrendsinBiotechnoIogy - its aim is : ::to present'ihformation of:interest to the bench biotechnolegist. As Tools of the : :Trade isa service to readers, commercial itemswit| be considered for publication :::only if the produdts are genuinely new::6~ if ~heypr0vide some cost benefit over : existingtechnblogy~::i: : :: :::::: ::: :::::::::::: - : : . /;i : : Itern:~ on: :~ew teehnoJogy~f ~:no~;commercia~:nat~i-ewiii:ai~o b~:considered for pBblicatiom* -i-he:benchbioteChnoiogisiimptovises~ you may have adapted a p~e~iOUs!9~eScribedteChfiiquefor::ane~ppii~ation ~ to improve performance,or :~,to::reducethd tim6::oFmshey reqdired.SuCh iin~provigationsCan be of use to: cNieagU~s:[n :other l a b ~ : iet:them kno~:y:0Q thOught:o{ it f i B t ] : :: C0nt/i6U~i~ns (iess t,an:500 ~ord~ :~ieaSe~houid b~ ~ent . ::: :.: ToSi~oft~e:T~ade,::Trendsin:B~otecfi~ology :>:::- : Elsevie-¢Trends Journals;68: Hills R0a~: :: :::: CamBridge CB2: 1LA,~UK.:: i::::::::::: :. : : : :: : :: : /:: : : :::::::::::::: ::: : :::: : _ :: : :: !t i First:aiscio~Ure$ i a { p u 6 i : ~ c : c l i ~:i~iead#~s~: 0 s G ~~:eremi b f : :h~ae8 ew::{~hniq~es~ maY limit ::~;the future: comrnerciN Opportunities.Adequate conSultati0n::with colleagues and institute :11:administrator# is advised before Submission if:y6u have any doubts: : :