Transcription Termination Sequence Research Articles

High-level expression of porcine fructose-1,6-bisphosphatase in Escherichia coli: purification and characterization of the enzyme.

A cDNA encoding porcine fructose-1,6-bisphosphatase was isolated from total pig liver RNA. The enzyme's coding sequence was fused to the bacteriophage T7 gene 10 promoter and transcription terminator sequence and expressed in E. coli under control of the T7 RNA polymerase. Induced cells contain elevated levels of fructose-1,6-bisphosphatase enzymatic activity and an abundant 37,000 dalton protein. The enzyme was purified to apparent homogeneity and judged identical to wild-type porcine fructose-1,6-bisphosphatase. The kinetic parameters are similar to those reported for the pig kidney enzyme. The N-terminal amino acid sequence is identical to the predicted sequence and the kinetic parameters are consistent with freedom from proteolysis. As estimated from enzymatic activity and visual inspection of coomassie blue-stained SDS-PAGE gels, porcine fructose-1,6-bisphosphatase constitutes as much as 20% of the soluble protein from the over-expressing E. coli strain.

Transcription Factors and Termination of Transcription in Methanococcus

Summary We have recently shown that specific transcription in cell extracts of Methanococcus thermolithotrophicus depends upon a transcription factor that was separated from the RNA polymerase by phosphocellulose chromatography. Here, we provide evidence for a second transcription factor. This factor copurified with RNA polymerase during initial Chromatographic steps, but it was resolved as a distinct activity required for cell-free transcription after centrifugation in sucrose density gradients. The native molecular weight of this factor was estimated by gel filtration to be 56 000. After electrophoresis under denaturing conditions, a 28 kDa polypeptide was correlated with factor activity. Thus, the native transcription factor appeared to be a dimer composed of two polypeptides of identical molecular mass. Oligo-dT sequences at the 3′-end of a tRNA Val gene and internal sequences of this tRNA were modified by DNA deletions in order to investigate the nature of archaeal transcription terminators. Deletion of two residues of the decameric sequence 5′-TTTTAATTTT-3′ reduced termination efficiency to about 27 percent of wild-type levels. Deletion of two additional residues from the 3′-end completely abolished terminator function. Deletions in the DNA region encoding the TΨC stem and loop of tRNA also significantly reduced termination efficiency. In addition a Rho-independent terminator of Escherichia coli perfectly replaced the decameric Methanococcus terminator sequence. These findings suggested that transcription termination sequences in archaea are similar to the terminator sequences in bacteria.

Analysis of the Spirochaeta aurantia flaA gene and transcript

The flaA gene, which codes for the Spirochaeta aurantia flagellar filament outer layer polypeptide, FlaA, was cloned, sequenced and analysed. The gene appears to be transcribed into a monocistronic mRNA from a σ70-like promoter. The translational start is 31 base pairs after the start of the transcript, the open reading frame is 1011 base pairs, and a rho-independent-like transcription terminator sequence begins about 19 base pairs after the translational stop codon. The deduced amino acid sequence of the S. aurantia FlaA showed 40% identity with the Treponema pallidum FlaA, but these polypeptides did not show a significant similarity to other polypeptides for which sequence information was available.

Structure of the 87-kDa beta-1,3-glucanase gene of Bacillus circulans IAM1165 and properties of the enzyme accumulated in the periplasm of Escherichia coli carrying the gene.

The nucleotides of a gene for the extracellular 87-kDa beta-1,3-glucanase of Bacillus circulans IAM1165 and its flanking regions were sequenced. The sequence showed an open reading frame for 877 amino acids, which corresponds to a precursor of the beta-1,3-glucanase. The coding region of 2631 bp is flanked by putative promoter and transcription terminator sequences. The signal peptide was considered to be consisted of 38 amino acids. The amino acid sequence of the mature enzyme composed of 839 amino acids showed high homology to that of the enzyme from B. circulans WL-12, although these enzymes are different in their sizes. A catalytic domain of the enzyme was estimated central region of the sequence on the basis of comparison of amono acid sequences of beta-1,3- or beta-1,3:1,4-glucanases. Properties of the periplasmic enzyme produced in Escherichia coli carrying the gene were identical with those of the extracellular enzyme produced by B. circulans IAM1165.

Expression of biologically active HIV glycoproteins using a T7 RNA polymerase-based eucaryotic vector system.

Bacteriophage T7 RNA polymerase and a derivative containing a nuclear localization signal were transiently expressed in CV-1 cells and were shown to localize to the cytoplasm and nucleus, respectively. A vector was constructed containing T7 promoter and transcription terminator sequences flanking a picornaviral 5' untranslated sequence for cap-independent translation and a polyA signal. Expression of the HIV-1 envelope glycoproteins in this vector system gave high levels of specific transcripts and translation products, independent of the subcellular localization of T7 RNA polymerase. The synthesis of HIV glycoproteins was also completely independent of the coexpression of the HIV rev protein, which is normally required for the expression of HIV structural proteins. In addition, a polyA signal was not required, whereas the presence of the picornaviral 5' untranslated region was necessary for efficient expression. Different possibilities to account for these findings are discussed. The HIV glycoproteins synthesized in this system were normally processed and assembled; they could induce syncytium formation and complement an env-deletion mutant of HIV-1.

Studies of the N-terminal half of human lactoferrin produced from the cloned cDNA demonstrate that interlobe interactions modulate iron release.

The factors influencing iron binding and release by lactoferrin have been addressed by comparison of the native full length molecule (Lf) with the N-terminal half of human lactoferrin (LfN) produced from the cloned cDNA expressed in baby hamster kidney (BHK) cells. The coding sequences for LfN were inserted into the expression vector pNUT between the metallothionein promoter and the human growth hormone transcription termination sequences. Transformed BHK cells were grown in roller bottles where concentrations of LfN as high as 35 mg/liter were obtained. The pure protein, produced by the transformed BHK cells, was characterized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, protein blotting and immunodetection, N-terminal sequence analysis, UV-visible spectroscopy, electron spin resonance spectroscopy, and measurements of metal binding and release. By these criteria LfN was found to be correctly processed, glycosylated, and able to bind iron reversibly. Both UV-visible and electron spin resonance spectra of the half molecule were very similar to those of native lactoferrin and the full length lactoferrin produced in BHK cells, but there were marked differences in the pH at which iron release occurred. Iron release from LfN occurs in the pH range 6.0-4.0, compared with 4.0-2.5 for native lactoferrin and 6.2-4.0 for transferrin. These results suggest that the more facile release of iron from LfN compared with native lactoferrin results from the absence of stabilizing contacts between the N- and C-terminal halves and that the characteristic difference in pH stability between lactoferrins and transferrins is due primarily to differences in these interactions.

Conjugative transfer of Enterococcus faecalis plasmid pAD1: nucleotide sequence and transcriptional fusion analysis of a region involved in positive regulation.

The Enterococcus faecalis plasmid pAD1 undergoes conjugative transfer in response to cAD1, a peptide sex pheromone emitted by potential bacterial recipients. Regulation of pAD1 transfer involves a number of plasmid-encoded determinants:iad, which determines a peptide-competitive inhibitor iAD1; signal sensing and transducing elements; and negative and positive regulators. The key positive regulator(s) of the pheromone response is believed to be encoded within a segment designated the E region of the plasmid. In this study, we analyzed the nucleotide sequence and transcription within the E region. An open reading frame designated traE1 was identified; its inferred protein consists of 118 amino acids. Insertional mutagenesis of traE1 resulted in a complete loss in plasmid transfer capability. Analysis of Tn917-lac insertions giving rise to transcriptional lacZ fusions showed that traE1 is transcribed only under cAD1 inducing conditions. Analysis of additional lacZ fusions within the region provided some insight into the roles of potential regulatory signals within and around the nucleotide sequences reported here. A regulatory role appearing to involve read-through of certain key transcription termination sequences seemed evident.

A monoamine-regulated Klebsiella aerogenes operon containing the monoamine oxidase structural gene (maoA) and the maoC gene.

The Klebsiella aerogenes gene maoA, which is involved in the synthesis of monoamine oxidase, was induced by tyramine and the related compounds, subjected to catabolite and ammonium ion repression, and cloned. The nucleotide sequence of the region involved in monoamine oxidase synthesis was determined. Two open reading frames, the maoA gene and a hitherto unknown gene (maoC), were found. These are located between a potential promoter sequence and a transcriptional terminator sequence. A region of the Escherichia coli chromosome that was highly homologous to the Klebsiella maoA gene was found. The potential maoA gene is located at 30.9 min on the E. coli chromosome. Analysis of the amino acid sequences of the first 11 amino acids from the N terminus of the purified monoamine oxidase agrees with those deduced from the nucleotide sequence of the maoA gene. The leader peptide extends over 30 amino acids and has the characteristics of a signal sequence. Primer extension and S1 nuclease mapping of transcripts generated in vivo suggests that the tyramine-induced mRNA starts at a site 62 bases upstream from the ATG initiation codon of the maoC gene. In the putative promoter region, a high degree of similarity to the consensus sequence for the binding site of cyclic AMP receptor protein was found. Thus, the mao region is composed of two cistrons, and the mao operon is regulated by monoamine compounds, glucose, and ammonium ions.

Structure of the Clostridium thermocellum gene licB and the encoded β‐1,3‐1,4‐glucanase

The nucleotide sequence of the Clostridium thermocellum gene licB, coding for a thermoactive beta-1,3-1,4-glucanase, has been determined. The gene is located downstream, but in opposite orientation to the beta-glucosidase gene bglA. A coding region of 1002 bp is flanked by canonical promoter and transcription terminator sequences. The primary translation product of the licB gene has a predicted molecular mass of 37,896 Da. The protein sequence can be divided into several discrete segments: an N-terminal signal peptide, a catalytic region, a segment rich in Pro and Thr residues and a C-terminal reiterated domain. The catalytic region shows close similarity to lichenases of bacilli (52-58% identity) and Fibrobacter succinogenes (35% identity), but is unrelated to barley beta-1,3-1,4-glucanases. It consists of two domains, which in the case of the F. succinogenes lichenase are arranged in reversed order to that of C. thermocellum and Bacillus lichenases. The C-terminal reiterated domain of C. thermocellum lichenase is homologous to the duplicated non-catalytic domain of endo-beta-1,4-glucanases and xylanase Z from the same organism. This domain is considered a characteristic feature of clostridial cellulases organized as multienzyme complex (cellulosome). The beta-1,3-1,4-glucanase encoded by the licB gene might therefore be an additional enzyme component of the C. thermocellum cellulosome.

RNA editing in the phosphoprotein gene of the human parainfluenza virus type 3

RNA editing of the human parainfluenza virus type 3 (HPIV3) phosphoprotein (P) gene was found to occur for the accession of an alternate discontinuous cistron. Editing occurred within a purine-rich sequence (AAUUAAAAAAGGGGG) found at the mRNA nucleotides 791–805. This sequence resembles an HPIV3 consensus transcription termination sequence and is located at the 5′-end of the putative D protein coding sequences. Editing at an alternate site (AAUUGGAAAGGAAAGG), mRNA nucleotides 1121–1136, for accession of a conserved V cistron, which is present in a number of paramyxovirus P genes, was not found to occur in HPIV3. In contrast with many other paramyxoviruses, editing was indiscriminate with the insertion of 1–12 additional G residues not present in the gene template. RNA editing was found to occur in both in vivo (HPIV3 infected cells) and in vitro (purified nucleocapsid complexes) synthesized mRNAs. Further, the in vitro prepared mRNA was edited regardless of whether the nucleocapsid complexes were transcribed in the presence or absence of uninfected human lung carcinoma (HLC) cell lysates. These results support the notion that RNA editing appears to be exclusively a function of viral proteins.

Faithful in vivo transcription termination of Xenopus laevis rDNA

DNA sequencing and subsequent functional in vitro analysis of the Xenopus laevis rDNA transcription termination has led to the identification of three transcription termination sequence elements: T1, located at the 3' end of the 28S rDNA; T2, a putative processing site 235 bp downstream of T1; T3, the principal terminator positioned 215 bp upstream of the gene promoter. As demonstrated for nuclear run-off assays, T3 was found to be the main terminator for Xenopus rDNA transcription. These in vitro data are in obvious contradiction to results obtained by electron microscopic (EM) spread preparations from rapidly isolated amplified oocyte nucleoli, i.e., an rDNA chromatin probe thought to represent the in vivo situation, indicative of transcription termination at sites T1-2. However, most interestingly, T3 had--again by the EM method--been identified as the exclusive terminator for NTS spacer transcription units. In order to answer the question of whether read-through transcription of the complete rDNA spacer sequence is obligatory for 40S pre-rRNA in vivo transcription, we analyzed several hundreds of spread rRNA genes from Xenopus oocyte nucleoli in great detail, applying two different spreading procedures, e.g., dispersal of amplified oocyte nucleoli shortly in detergent-free or detergent containing low-salt media prior to the EM spreading technique. Quantitation of EM spreads resulted in the finding that read-through rDNA spacer transcription beyond T1-2 termination sites (i.e., indicative of T3 transcription termination) can be visualized for the in vivo situation at a frequency of less than 3% of rRNA genes analyzed.(ABSTRACT TRUNCATED AT 250 WORDS)

Structure of the β‐glucosidase gene bglA of Clostridium thermocellum

The nucleotide sequence of the Clostridium thermocellum gene bglA, coding for the thermostable beta-glucosidase A, has been determined. The coding region of 1344 bp was identified by comparison with the N-terminal amino acid squence of recombinant beta-glucosidase A purified from Escherichia coli. The deduced amino acid sequence corresponds to a protein of 51,482 Da. The coding region is flanked by putative promoter and transcription terminator sequences. The protein is unrelated to beta-glucosidase B of C. thermocellum, but has a high level of similarity with other bacterial beta-glucosidases and phospho-beta-glucosidases. Similarity is also observed with the beta-galactosidase of the archaebacterium Sulfolobus solfataricus. Unexpectedly, it was found that human lactase-phlorizin hydrolase contains three copies of a sequence closely related to C. thermocellum beta-glucosidase A (up to 40% sequence identity). These diverse beta-glucosidases can therefore be grouped into an enzyme family (BGA) of common structural design. Sequence comparison by hydrophobic cluster analysis revealed that all BGA enzymes share a well conserved region which is homologous to the catalytic domain of the widely distributed cellulase family A. A distinctive feature of this domain is the sequence motif His-Asn-Glu-Pro in which the catalytic residues His and Glu are separated by 35-55 amino acid residues. The cellulase family A and the beta-glucosidase family BGA might thus be considered as members of a protein super-family comprising beta-glucanases and beta-glycosidases from all three primary kingdoms of living organisms.

Host-derived 5' ends and overlapping complementary 3' ends of the two mRNAs transcribed from the ambisense S segment of Uukuniemi virus

Two mRNAs, coding for the N and NSS proteins, are transcribed from the small (S) Uukuniemi virus RNA segment by an ambisense strategy (J. F. Simons, U. Hellman, and R. F. Pettersson, J. Virol. 64:247-255, 1990). In this report, we describe the analysis of the 5' and 3' ends of the two mRNAs. Primer extension as well as cloning and sequencing of individual mRNAs showed that the 5' ends of both mRNAs contained nonviral sequences ranging from 7 to 25 residues in length (mean, 12 residues), indicating a cap-snatching mechanism similar to the one originally described for priming of influenza virus mRNA synthesis. In 35% of the cases, the first virion-specified nucleotide (an A residue) was substituted with a G residue. Between the translation termination codons of N and NSS, there is a 74-residue-long noncoding intergenic region (Simons et al., J. Virol. 64:247-255, 1990). Nuclease protection assays using both RNA and DNA hybridization probes showed that the 3' ends of the N and NSS mRNAs overlap each other by about 100 nucleotides. The 3' end of the NSS mRNA extends into the coding sequence of the N mRNA, whereas the N mRNA is terminated just prior to the stop codon of NSS. To our knowledge, this is the first example of overlapping complementary mRNAs in viruses with an ambisense coding strategy. No obvious transcription termination sequence was identified. However, because of a short palindromic sequence in the intergenic region, the 3' ends of both mRNAs (and consequently also the template RNAs) can be folded into an A/U-rich hairpin structure. It remains to be determined whether this structure plays any role in transcription termination.

Functional organization and nucleotide sequence of the Bacillus subtilis pyrimidine biosynthetic operon

A 12.5-kilobase segment of Bacillus subtilis chromosomal DNA containing the entire pyrimidine biosynthetic (pyr) gene cluster has been cloned and sequenced. The sequenced DNA has seven cistrons encoding the six enzymes of de novo pyrimidine nucleotide biosynthesis and two open reading frames of unknown function. Based on the sequence and mapping of transcripts, the genes in this cluster appear to be transcribed on one large polycistronic message in the order ORF1, pyrB, pyrC, pyrAA, pyrAB, ORF2, pyrD, pyrF, pyrE. The deduced amino acid sequences for six pyrimidine biosynthetic enzymes from B. subtilis and comparisons with the corresponding sequences from numerous other species are presented. The 3' ends of the reading frames overlap the 5' ends of the downstream open reading frames for all cistrons in the cluster except ORF1 and pyrB, which are separated by a 145-base pair intercistronic region. The start of transcription was mapped by primer extension to a G residue 158 nucleotides upstream from the translation initiation codon of ORF1. This site is preceded by a typical B. subtilis sigma A-dependent promoter. A promoter indicator plasmid was used to show that this region carried a promoter from which transcription was regulated by pyrimidines. Transcription from this promoter was not detected in primer extension experiments using mRNA prepared from B. subtilis cells grown in the presence of excess uracil. No transcripts initiating from the intercistronic space between ORF1 and pyrB were detected with S1 nuclease mapping; however, a transcription terminator was detected in this region that reduced but did not fully block transcriptional readthrough. This terminator was not regulated by pyrimidines in the growth medium under the conditions tested. The region immediately following the promoter and 5' to ORF1 has a potential transcription terminator sequence. The roles, if any, of these transcription terminators in the regulation of pyr operon expression are yet to be determined. However, deletion of the terminator immediately following the promoter abolished pyrimidine regulation of transcription in the indicator plasmid.

Improved electroporation and cloning vector system for gram-positive bacteria

A protocol for transformation of intact Enterococcus faecalis cells by electroporation was developed through a systematic examination of the effects of changes in various parameters, including (i) growth conditions; (ii) composition of the electroporation solution; (iii) electroporation conditions, such as field strength and resistance; (iv) size, concentration, and purity of DNA used for transformation; and (v) conditions used to select for transformants. Key features of this protocol include the use of exponential-phase cells grown in inhibitory concentrations of glycine and the use of an acidic sucrose electroporation solution. Frequencies of greater than 2 x 10(5) transformants per microgram of plasmid DNA were obtained for E. faecalis cells, whereas various strains of streptococci and Bacillus anthracis were transformed at frequencies of 10(3) to 10(4) transformants per microgram of plasmid DNA with the same protocol. A novel Escherichia coli-Streptococcus and Enterococcus shuttle cloning vector, pDL276, was constructed for use in conjunction with the electroporation system. This vector features a multiple cloning site region flanked by E. coli transcription termination sequences, a relatively small size (less than 7 kb), and a kanamycin resistance determinant expressed in both gram-positive and gram-negative hosts. Various enterococcal and streptococcal DNA sequences were cloned in E. coli (including sequences that could not be cloned on other vectors) and were returned to the original host by electroporation. The vector and electroporation system was also used to clone directly into E. faecalis.

Genetic analysis of the enterobactin gene cluster in Shigella flexneri

The genes for transport and synthesis of the phenolate siderophore enterobactin are present on the chromosomes of both Ent+ and Ent- clinical isolates of Shigella flexneri. To determine why Ent- S. flexneri isolates fail to express a functional enterobactin system, the structure and expression of enterobactin genes were examined. Several alterations may be responsible for the inability of S. flexneri to express enterobactin. (i) The mRNA levels produced from the entC and fepB genes were not derepressed in low-iron media. (ii) DNA sequence analysis of the entC-fepB intergenic region revealed an 83-bp noncontiguous deletion in the putative fepB leader sequence. The deleted sequences are in a region which would be capable of forming extensive stem-and-loop structures. (iii) An amber codon in the 5' portion of the entC gene was also detected. (iv) An IS1 element, previously mapped to the Ent- S. flexneri enterobactin gene cluster, was found to lie within a potential transcriptional termination sequence in the entF-fepE intergenic region. (v) A mutation responsible for the inactivation of the entF gene was mapped to the entF coding region by using entF hybrid gene fusions. (vi) A comparison of outer membrane profiles from an E. coli strain harboring the cloned fepA gene from either an Ent+ or Ent- Shigella isolate revealed that the Ent- FepA protein is present in the outer membrane but at greatly reduced levels than that of the Ent+ FepA protein. This observation, along with additional studies, suggests that the Ent- FepA may be defective in translation and/or translocation.

Reliable Transient Promoter Assay Using Fluorescein-di-β-D-galactopyranoside Substrate

The promoter region of the mouse myelin proteolipid protein (PLP) gene was cloned into a promoter testing vector, pIP111. The pIP111 vector is a promoterless derivative of pCH110 (SV40 early region promoter-lacZ) and contains the Escherichia coli lpp transcription terminator sequence at the 5' end of the cloning site. The newly constructed PLP-lacZ fusion plasmid (pWP) was transfected into PLP-nonproducing NIH-3T3 fibroblasts or PLP-producing C6 cells. When the measured beta-galactosidase activity in the pWP-transfected cells was normalized to the pCH110-transfected cells (an appropriate control if the SV40 early region promoter functions constitutively in various cell lines), the results suggested that the promoter region of the PLP gene contains the information necessary for initiation of transcription in a C6 cell-specific manner. However, the beta-galactosidase produced in viable cells was also detected by fluorescein-di-beta-D-galactopyranoside (FDG) treatment followed by image analysis using inverted fluorescent microscopy, which allowed the transfection efficiency to be calculated, and the beta-galactosidase activity obtained by the regular ONPG method was normalized with the value obtained. This procedure indicated that the promoter region of the PLP gene did not show C6-specific expression, because the SV40 early-region promoter was 10 times more active in NIH-3T3 cells than in C6 cells. Thus, the standard experiment gave misleading results. As our detection method is simple and can be used to analyze the promoter activity in a single cell, many applications should be possible.(ABSTRACT TRUNCATED AT 250 WORDS)

A sulfur- and tyramine-regulated Klebsiella aerogenes operon containing the arylsulfatase (atsA) gene and the atsB gene

The structural gene for arylsulfatase (atsA) of Klebsiella aerogenes was cloned into a pKI212 vector in Escherichia coli. Deletion analysis showed that the atsA gene with the promoter region was located within a 3.2-kilobase cloned segment. In E. coli cells which carried the plasmid, the synthesis of arylsulfatase was repressed by various sources of sulfur; the repression was relieved, in each case, by tyramine. Transfer of the plasmid into atsA or constitutive atsR mutant strains of K. aerogenes resulted in complementation of atsA but not of atsR. The nucleotide sequence of the 3.2-kilobase fragment was determined. Two open reading frames, the atsA gene and an unknown gene (atsB), were found. These are located between a potential promoter and a transcriptional terminator sequence. Deletion analysis suggests that atsB is a potential positive factor for the regulation of arylsulfatase. Analysis of the amino acid sequences of the first 13 amino acids from the N terminus of the purified secreted arysulfatase agrees with that of the nucleotide sequence of atsA. The leader peptide extends over 20 amino acids and has the characteristics of a signal sequence. Primer extension mapping of transcripts generated in vivo suggests that the synthesis of mRNA starts at a site 31 or 32 bases upstream from the ATG initiation codon of the atsB gene. By Northern (RNA) blot analysis of the transcripts induced by tyramine, we found a 2.7-kilobase transcript which is identical in size to the total sequence of the atsB and atsA genes. Thus, the ats operon is composed of two cistrons and is regulated by sulfur and tyramine.

Construction of an efficient overproducer clone of HinfI restriction endonuclease using the polymerase chain reaction

We describe the use of the polymerase chain reaction (PCR) technique to alter transcriptional and translational signals surrounding a gene so as to achieve overexpression in Escherichia coli. By changing the ribosome-binding site sequence preceding the hinfIR gene to match the consensus E. coli signal and by adding a transcription terminator sequence immediately following the gene, the yield of HinfI was increased about tenfold over that obtained from the natural Haemophilus influenzae signals. The addition of the positive retroregulator stem-loop sequence derived from the crystal protein-encoding gene of Bacillus thuringiensis downstream from the hinfIR gene further increased yields by twofold to a level of 13% of the total cellular protein.

Production of a hybrid macrolide antibiotic in Streptomyces ambofaciens and Streptomyces lividans by introduction of a cloned carbomycin biosynthetic gene from Streptomyces thermotolerans

The structurally related macrolide antibiotics carBomycin (Cb) and spiramycin (Sp) are produced by Streptomyces thermotolerans and Streptomyces ambofaciens, respectively. Both antibiotics contain 16-membered lactone rings to which deoxysugars are attached. There are three sugars m Sp (forosamine, mycaminose and mycarose) and two sugars in Cb (mycaminose and a derivative of mycarose containing an isovaleryl group at position 4). We have identified the gene from S. thermotolerans (designated carE ), which appears to encode an enzyme that acylates this mycarose sugar, and have shown that recombinant strains containing carE can use Sp as a substrate and convert it to the hybrid antibiotic, isovaleryl Sp (ivSp). Expression of carE was demonstrated in two heterologous hosts: in S. ambofaciens, where endogenously synthesized Sp was converted to ivSp, and in Streptomyces lividans where exogenously added Sp was converted to ivSp. The carE gene was isolated on a cosmid that also encodes genes required for Cb-lactone formation. These genes reside on a DNA segment of about 70 kb and are part of a Cb biosynthetic gene cluster that is flanked by two Cb-resistance genes, carA and carB. Mapping studies and nucleotide sequence analysis revealed that carE is located at one end of this gene cluster, immediately adjacent to the carB gene. Genes carB and carE are transcribed convergently and may share a common transcriptional terminator sequence.