Abstract
Streptomyces clavuligerus ATCC 27064 is unable to use glucose but has genes for a glucose permease (glcP) and a glucose kinase (glkA). Transformation of S. clavuligerus 27064 with the Streptomyces coelicolor glcP1 gene with its own promoter results in a strain able to grow on glucose. The glcP gene of S. clavuligerus encodes a 475 amino acid glucose permease with 12 transmembrane segments. GlcP is a functional protein when expressed from the S. coelicolor glcP1 promoter and complements two different glucose transport-negative Escherichia coli mutants. Transcription studies indicate that the glcP promoter is very weak and does not allow growth on glucose. These results suggest that S. clavuligerus initially contained a functional glucose permease gene, like most other Streptomyces species, and lost the expression of this gene by adaptation to glucose-poor habitats.
-
The GenBank/EMBL/DDBJ accession numbers for the sequences of the regions upstream and downstream of S. clavuligerus glkA and glcP are FN377746 and FN377747.
Edited by: M. Paget
INTRODUCTION
A total of 53 putative carbohydrate uptake genes have been detected in the genome of the model organism Streptomyces coelicolor. ATP-binding cassette (ABC) transport systems account for the transport of ribose, lactose, maltose, xylose and sugar alcohols (Bertram et al., 2004). Glucose is transported by an H+–glucose symporter (van Wezel et al., 2005), and comparative analysis of gene sequences suggests that galactose is transported by an Na+–galactose symporter of the SSS (sodium solute symporters) family, similar to that found in Vibrio parahaemolyticus (Bertram et al., 2004). Fructose is incorporated by the phosphotransferase (PTS) system, as shown by the disruption of the ptsH gene (Nothaft et al., 2003b). A protein, GlpF, of the major intrinsic protein family appears to be responsible for glycerol-facilitated diffusion (Bertram et al., 2004; Hindle & Smith, 1994). Recently, it has been shown that N-acetyl-d-glucosamine (NAG) plays an important role in differentiation and antibiotic production in S. coelicolor (Rigali et al., 2006, 2008), in which this carbon source is transported by the sugar PTS system (Nothaft et al., 2003a).
The behaviour of Streptomyces clavuligerus, the producer of clavulanic acid and cephamycin C, in relation to carbon sources utilization is peculiar. This species is unable to grow on glucose as the sole carbon source; moreover, it does not use hexoses such as mannose, galactose or fructose, pentoses such as xylose, sugar alcohols such as mannitol or sorbitol, or disaccharides such as lactose or sucrose. Only glycerol, maltose and starch are efficiently used as carbon sources by S. clavuligerus (García-Domínguez et al., 1989).
If the systems for carbohydrate transport are conserved in different Streptomyces species, S. clavuligerus would be defective in: (i) ABC systems for lactose, xylose or mannitol uptake; (ii) the PTS system for fructose utilization; and (iii) the symporters required for glucose and galactose incorporation. This is an enigmatic situation for a soil-dwelling actinomycete.
One gene for glucose transport (araE, SAV2657) is present in the Streptomyces avermitilis genome, and two genes, glcP1 (SCO5578) and glcP2 (SCO7153), with 99.9 % nucleotide identity to each other, separated by 1.8 Mb in the chromosome, occur in the S. coelicolor genome. They encode 100 % identical GlcP1 and GlcP2 proteins. Both genes are expressed in glucose-grown cells of S. coelicolor, as shown by RT-PCR experiments; glcP1 is inducible by glucose while glcP2 transcripts are almost undetectable (van Wezel et al., 2005). In this work we report the effect of S. coelicolor glcP1 (for glucose transport) and glkA (encoding the glucose kinase) genes on glucose utilization by S. clavuligerus. Using a combination of the S. coelicolor and S. clavuligerus glucose permease promoters coupled to either the promoterless glcP gene of S. clavuligerus or the promoterless glcP1 gene of S. coelicolor we conclude that expression of glcP from the native S. clavuligerus promoter is limiting for glucose utilization.
METHODS
Bacterial strains, plasmids and culture conditions.
The bacterial strains and plasmids used in this study are listed in Table 1⇓. Cultures of plasmid-bearing cells were supplemented with ampicillin (50 μg ml−1), chloramphenicol (25 μg ml−1), kanamycin (25 μg ml−1) or apramycin (50 μg ml−1), as appropriate.
Plasmids and strains used in this work
Glucose transport-deficient Escherichia coli mutants and transformants carrying the glcP gene were grown on MacConkey or in Voges–Proskauer medium (Difco).
To inoculate cultures of S. coelicolor, spores were germinated in 2× tryptone yeast extract (TY) medium (Kieser et al., 2000). S. clavuligerus cultures were inoculated using culture aliquots [grown in tryptone soya broth (TSB) medium] kept frozen in 20 % (v/v) glycerol. Streptomyces strains were grown in 500 ml baffled flasks at 28 °C and 220 r.p.m.
To check the growth on different carbon sources, strains were inoculated in 50 ml of inoculum medium (IM), which contains basal medium (BM) supplemented with 0.2 % l-asparagine, 1 % glycerol, 0.1 % NH4Cl and 0.1 % yeast extract (Aharonowitz & Demain, 1978). After 24 h, the mycelium was washed and different volumes, as appropriate, were transferred to 50 ml BM containing 0.1 % NH4Cl and 1 % of the carbon source tested. Glucose solutions were treated with α-glucosidase to eliminate traces of maltose present in commercial glucose supplies. For the growth of S. coelicolor, BM was supplemented with the trace elements described by Chatterjee & Vining (1981).
Glucose in the culture supernatants was quantified with the Glucose Assay kit (Sigma).
Growth was determined by measuring total DNA by the diphenylamine method (Burton, 1968).
DNA manipulations.
Plasmid DNA isolation, restriction endonuclease digestions, ligations and transformation of E. coli were performed using standard techniques (Sambrook et al., 1989). Genomic DNA preparations from Streptomyces were carried out according to Kieser et al. (2000). Non-radioactive hybridizations were performed following the protocol given in the DIG System kit (Roche), and detection was done using 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium.
DNA sequencing.
DNA sequencing of the PCR-amplified fragments was performed using a MegaBACE 500 DNA Analysis System (Amersham Biosciences) sequencer. Subclones were sequenced from universal primers, checking the information on both strands. The sequences of the regions upstream and downstream of S. clavuligerus glkA and glcP were provided by DSM (Delft, The Netherlands) and have been deposited in the EMBL database with the accession numbers FN377746 and FN377747.
RNA isolation.
RNAprotect Bacteria reagent (Qiagen) was used for stabilization of collected samples for RNA extraction. Mycelium was treated with lysozyme (30 mg ml−1) and RNeasy Mini Spin Columns (Qiagen) were used for RNA isolation according to the manufacturer's instructions, including a phenol extraction step before transferring the lysate to the column. RNA preparations were incubated with Turbo DNase (Ambion) to eliminate any chromosomal DNA contamination. RNA quantification was done with a NanoDrop ND-1000 UV–vis spectrophotometer.
PCR and non-quantitative RT-PCR analysis.
Oligonucleotide primers used in this work are shown in Table 2⇓. All PCRs were performed in a TGradient (Biometra) thermocycler using Platinum Pfx DNA polymerase (Invitrogen). The dNTP mix was prepared from individual nucleotides using a ratio of 15 A : 15 T : 35 G : 35 C to improve amplification of high-G+C DNA content. Total DNA from S. coelicolor was used to amplify the glcP1 (SCO5578) promoter with primers CRP1/CRP2 and the glcP1-coding region with CRP3/CRP4, whereas S. clavuligerus total DNA was used to amplify the glcP promoter using the oligonucleotide pair CRP5/CRP6 and the glcP-coding region with CRP7/CRP8. PCR products were subcloned into pBluescript II SK(+) and sequenced for confirmation before assembling the promoter–gene constructs for integration in Streptomyces.
Oligonucleotides used in this work
Gene expression analysis by RT-PCR was done with the SuperScript One-Step system (Invitrogen) using 150–200 ng total RNA as template. Primers were designed to generate PCR products of approximately 500–600 bp. RT-PCR conditions were as follows: first strand cDNA synthesis, 50 °C for 30 min followed by 94 °C for 2 min; PCR amplification was performed in two phases: three cycles of 94 °C for 30 s, 68–62 °C (depending on the set of primers used) for 30 s, 72 °C for 40 s and 20–35 cycles of 94 °C for 30 s, 66–58 °C (depending on the set of primers) for 30 s and 72 °C for 40 s. Systematically, to compare non-saturated amplification products, a decreasing number of PCR cycles were performed. Negative controls were carried out with each set of primers and Platinum Taq DNA polymerase (Invitrogen) to verify the absence of contaminating DNA in the RNA preparations.
Glucose kinase assays.
S. clavuligerus was grown in TSB medium [Trypticase (30 g l−1) in soy broth] and S. coelicolor in a mixture of TSB and yeast extract/malt extract (YEME) media (40 : 60 %) (Kieser et al., 2000) for 24 h, and the mycelium was centrifuged and transferred to starch and asparagine (SA) medium (Aidoo et al., 1994) supplemented with glucose (1 %) under vigorous shaking for 10 h. Cells were harvested and washed twice, and 1 g mycelium was suspended in 50 mM triethanolamine, pH 7.4, with Complete Protease Inhibitor Cocktail (Roche). Mechanical disruption was performed with a FastPrep system (Thermo Scientific), cell debris was removed by centrifugation and the supernatant was filtered through a PD-10 desalting column (GE Healthcare).
Glucose kinase activity was measured spectrophotometrically by monitoring the formation of NADPH in a glucose-6-phosphate dehydrogenase coupled reaction, as described by García-Domínguez et al. (1989). Enzyme activity is expressed as units, where one unit is the amount of enzyme that produces 1 nmol NADPH min−1.
Construction of plasmids.
pRA is an EcoRV-digested pLUXAR recircularized. pAMB-1 is pIJ773 carrying a 4.9 kb HindIII fragment from pIJ699.
The S. clavuligerus putative glkA activator and the ORF upstream were subcloned in a 2096 bp NsiI–BclI fragment in the single-copy integrative pRA plasmid digested with EcoRV, to give pRA-2. From pRA-2 a 926 bp XmnI–EcoRI fragment containing only the putative glkA activator was rescued and subcloned in EcoRV/EcoRI-digested pRA to obtain pRA-1. The same 2096 and 926 bp DNA fragments were subcloned in the multicopy plasmid pAMB1 linearized with EcoRV to get pAMB1-2 and pAMB1-1.
The plasmids harbouring PCR fragments with S. clavuligerus or S. coelicolor glucose permease promoters or coding regions were used to assemble different arrangements of promoter and gene by means of the restriction sites included in the oligonucleotides designed. These fragments were then subcloned in EcoRI/BamHI-digested pRA to give: (i) pRA-AB (A corresponds to the promoter and B to the ORF in S. coelicolor), in which in a 1617 bp fragment the S. coelicolor glcP1 promoter controls the S. coelicolor glcP1 gene; (ii) pRA-AD, in which in a 1685 bp fragment the S. coelicolor glcP1 promoter (A) controls the S. clavuligerus glcP gene (D); (iii) pRA-CB, in which in a 1633 bp fragment the S. clavuligerus glcP promoter (C) controls the S. coelicolor glcP1 gene (B); and (iv) pRA-CD, in which in a 1701 bp fragment the S. clavuligerus glcP promoter (C) controls the S. clavuligerus glcP gene (D) (see Fig. 5⇓).
To obtain pRA-AB-K, a 1 kb DNA fragment isolated from pFT256 carrying the S. coelicolor glkA gene expressed from its own promoter, was ligated into the EcoRV site of pRA-AB.
Plasmid DNA was introduced into Streptomyces by conjugation using E. coli ET12567[pUZ8002] as the donor strain. Transconjugants were isolated on mannitol–soya medium (Kieser et al., 2000) and selected with apramycin. The integration of pRA-derived plasmids or the presence of the autonomous pAMB1-derived plasmids was checked by total DNA hybridization or by plasmid recovery.
Plasmid pglcP was constructed to check the functionality of S. clavuligerus glucose permease. A 1.5 kb SacI–XbaI fragment containing the glcP-coding region and 25 nt upstream was obtained from pKS-CD and ligated downstream of the lacZ promoter in SacI/HindIII-digested pFT76. The control plasmid pBRel was constructed by deletion of a 1.2 kb internal fragment of glcP2 in BamHI-digested pFT76.
RESULTS
S. clavuligerus contains a glkA gene and has normal glucose kinase activity
To understand the lack of glucose utilization by S. clavuligerus we searched for a glucose kinase gene (glkA) in this organism. An S. clavuligerus glkA gene was detected by hybridization of digested total DNA with a 1 kb probe containing the S. coelicolor glkA gene (Angell et al., 1992). The sequence of a hybridization-positive 4.4 kb BstEII DNA fragment revealed a gene organization similar to that of S. coelicolor, S. avermitilis and Streptomyces griseus (Fig. 1a⇓). The glkA gene encodes a 313 amino acid protein (Mr 32 273, deduced pI 5.8) of the ROK family, 81 % identical in amino acid sequence to the S. coelicolor homologue (Fig. 1b⇓).
Characteristics of the glucose kinase gene of S. clavuligerus. (a) Organization of a BstEII DNA fragment containing the incomplete orf1 (encoding a putative cyclase-dehydratase), orf2 (encoding an ion transport ATPase), orf3 (encoding a putative glkA activator), glkA (orf4) and an incomplete orf5 (encoding a protein of unknown function). The glkA gene is named SCCG_05538.1 in the S. clavuligerus Broad Institute Project Sequence. Bs, BstEII; S, SalI; Ns, NsiI; X, XmnI; B, BclI. Horizontal bars indicate the DNA fragments used in the construction of pRA-1, pRA-2, pAMB1-1 and pAMB1-2. (b) Comparison of S. clavuligerus and S. coelicolor GlkA. The characteristic set of cysteines required for activity (Mesak et al., 2004) is indicated with a horizontal line. Identical amino acids are indicated with asterisks. (c) Expression of glkA, detected by RT-PCR, in strains carrying different fragments containing orf3 in single-copy integrated form (S. clavuligerus : : pRA-1, S. clavuligerus : : pRA-2), in a multicopy plasmid (S. clavuligerus[pAMB1-1], S. clavuligerus[pAMB1-2]) and in negative controls (S. clavuligerus : : pRA, S. clavuligerus[pAMB1]). RNA was obtained in SA medium under control conditions (0) and 3 h (G) after glucose supplementation.
The glkA gene is located downstream of a gene (orf2) encoding an ion transporter ATPase, and of a gene (orf3) for a putative glucose kinase activator (with 57.6 % identity to SCO2127) (Guzmán et al., 2005). Downstream of glkA and separated by 679 nt is a gene (orf5) encoding a product of unknown function.
The in vitro glucose kinase activity of S. clavuligerus was determined in three different experiments and compared with that of S. coelicolor. Both strains were grown in SA medium supplemented with 1 % glucose. A specific activity of 23.2±3.2 mU (mg protein) −1 was found in S. coelicolor cell extracts, while the mean specific glucose kinase activity in S. clavuligerus was 17.25±2.3 mU (mg protein)−1.
Transformant S. clavuligerus [pFT256] carrying the S. coelicolor glkA gene expressed from its own promoter was also not able to use glucose, although the glucose kinase activity in the transformant was 19.7±1.6 mU (mg protein)−1.
The SCO2127 gene of S. coelicolor has a positive effect on glucose uptake and glucose kinase activities (Guzmán et al., 2005), possibly due to transcription activation. Therefore, S. clavuligerus strains carrying orf3, the SCO2127 homologous gene, in a mono-copy integrative plasmid (named pRA-1 and pRA-2), or in a multi-copy plasmid (named pAMB1-1 and pAMB1-2) along with their controls (pRA and pAMB1) were constructed. When the RNA of these strains, grown in SA medium or SA supplemented with glucose, was tested by RT-PCR, they showed the same amplification intensity for the glkA gene (Fig. 1c⇑). This indicates that S. clavuligerus orf3 is not a glkA activator and that glucose does not stimulate glkA transcription.
The S. clavuligerus glcP gene encodes a glucose permease functional in E. coli
Labelled glucose is barely incorporated in S. clavuligerus (García-Domínguez et al., 1989). To test whether a glucose permease gene exists in S. clavuligerus, total DNA was hybridized with a 1.4 kb DNA fragment containing the whole S. coelicolor glcP2 gene. A clear signal of 6 kb was found in BstEII-digested DNA; therefore, the genome of S. clavuligerus, available at DSM (Delft, The Netherlands), was searched for a gene homologous to S. coelicolor glcP1 or glcP2. One single glucose permease gene located downstream of a chromosome segregation gene (orf6) was found, while in the opposite orientation, genes for a putative acetyltransferase (orf8) and a putative docking protein (ftsY) were present (Fig. 2a⇓).
Characteristics of the glucose permease gene (glcP) of S. clavuligerus. (a) Organization of the genome in the region around glcP: incomplete orf6 (encoding a chromosome segregation protein), glcP (orf7), orf8 (encoding an acetyltransferase) and incomplete ftsY. The glcP gene is named SCCG_07234.1 in the S. clavuligerus Broad Institute Project Sequence. S, SalI; B, BclI; N, NcoI; Bs, BstEII; E, EcoRI. (b) Comparison of the proteins encoded by S. clavuligerus glcP and S. coelicolor glcP1. The 12 transmembrane motifs predicted for GlcP are indicated with horizontal lines. (c) Comparison of the promoter-containing intergenic regions upstream of araE from S. avermitilis, glcP1 from S. coelicolor and glcP from S. clavuligerus. The −10 and −35 regions and the transcription start point of the S. coelicolor glcP1 promoter (van Wezel et al., 2005) and the putative homologous regions in the S. avermitilis araE promoter are underlined.
The S. clavuligerus glucose permease gene was named glcP to differentiate it from the S. coelicolor glcP1 and glcP2 genes. It encodes a protein with 80–81 % identity to S. coelicolor GlcP1, S. avermitilis AraE and S. griseus SGR1900 (Fig. 2b⇑). The protein (Mr 50 539, pI 8.4), contains 12 membrane-spanning segments as found in other H+–glucose symporters.
Two glucose transport-negative mutants, E. coli LM1 and E. coli LR2-175 (Aulkemeyer et al., 1991; Lengeler et al., 1981), were transformed with plasmid pglcP, in which S. clavuligerus glcP is expressed from the lacZ promoter, and with plasmids pFT76 (containing S. coelicolor glcP1) and pBRel, as positive and negative controls, respectively. The strains were streaked on MacConkey agar and on MacConkey agar supplemented with 50 mM glucose. Transformants E. coli LM1[pglcP] and E. coli LR2-175[pglcP] were able to use glucose, as indicated by the yellow halo formed around the patches (Fig. 3a⇓), whereas the parental strains and transformants E. coli LM1[pBRel] and E. coli LR2-175[pBRel] did not produce yellow zones. In addition, the strains were grown for 48 h in Voges–Proskauer medium and supplemented with methyl red. The culture broth of E. coli LR2-175[pFT76] and E. coli LR2-175[pglcP] turned from slightly yellow to intense red (Fig. 3b⇓), indicating acidification due to glucose fermentation, while the parental mutant and control negative strain retained the pale yellow colour, confirming that they are not able to use glucose. This indicates that S. clavuligerus glucose permease is functional when expressed in E. coli.
Complementation of glucose transport-negative E. coli strains by S. clavuligerus glcP. (a) MacConkey agar plates without (left) and supplemented with 50 mM glucose (right). (b) Liquid Voges–Proskauer with methyl red. E. coli strains: (1) DH5-α, (2) LM1, (3) LM1[pFT76], (4) LM1[pBRel], (5) LM1[pglcP], (6) LR2-175, (7) LR2-175[pFT76], (8) LR2-175[pBRel], (9) LR2-175[pglcP].
glcP is weakly expressed in S. clavuligerus
In S. coelicolor the essential glucose permease is glcP1, while the duplicate glcP2 gene can be deleted without a significant effect on glucose utilization. The two genes have substantial differences in their 5′-upstream regions. When compared with the homologous regions in S. avermitilis araE and S. clavuligerus glcP, an overall similarity is found between glcP1 and araE upstream regions. The −10 (TAGTCT) and −35 (TTGACT) sequences of glcP1 (van Wezel et al., 2005) are perfectly conserved in araE. However, the homology is very low with the promoter regions of S. coelicolor glcP2 (results not shown) or S. clavuligerus glcP, in which the −10 and −35 regions are difficult to recognize (see Fig. 2c⇑).
The glcP gene in S. clavuligerus is 161 nt downstream of orf6 and 55 nt upstream, and in the opposite sense to orf8 (Fig. 2a⇑). RT-PCR analysis was done using RNA isolated from S. clavuligerus : : pRA (control strain carrying the integrated pRA empty plasmid) grown in: (i) BM plus glycerol, (ii) BM plus glucose (Fig. 4⇓, upper panel) or (iii) SA, and (iv) SA plus glucose (Fig. 4⇓, lower panel). To discriminate strand-specific transcription, the experiment was performed in two steps: (i) mRNA retrotranscription to obtain cDNA was done using primers GLCP5 or GLCP6 separately, and (ii) the second oligonucleotide (either GLCP6 or GLCP5, respectively) was added later for the PCR process. When the cDNA was obtained using GLCP6, complementary to the glcP mRNA (Fig. 4⇓, lanes 3 and 7), the yield of amplified product was barely detectable and clearly lower than when the retrotranscription was done with the GLCP5 primer (Fig. 4⇓, lanes 2 and 6). These results indicate that a very weak transcription of glcP occurs but that most of the signal in that region is due to an antisense transcript, probably originating in orf8. The residual expression of glcP is not affected by the presence of glucose in the cultures (Fig. 4⇓, lanes 3 and 7).
Transcription of S. clavuligerus glcP. Amplification of the S. clavuligerus : : pRA glcP gene from RNA obtained in BM medium supplemented with either glycerol (BM+Gly) or glucose (BM+G) (upper panel), or in SA medium and SA supplemented with glucose (SA+G) (lower panel). Lanes: 1 and 5, one-step RT-PCR using oligonucleotides GLCP5 and GLCP6; 4 and 8, PCR negative control with the same oligonucleotides; 2 and 6, PCR performed using cDNA obtained with GLCP5 in the reverse-transcription step (see scheme at the right side); 3 and 7, PCR performed using cDNA obtained with GLCP6 in the reverse-transcription step.
Construction of S. clavuligerus recombinant strains able to use glucose
The promoter regions of glcP1 from S. coelicolor and glcP from S. clavuligerus, as well as S. coelicolor glcP1- and S. clavuligerus glcP-coding sequences, were amplified by PCR. The four PCR products were assembled in the four combinations (two promoters × two ORFs) indicated in Fig. 5a⇓ and subcloned in the integrative single-copy plasmid pRA in such a way that each gene was expressed either from its own promoter (constructions pRA-AB and pRA-CD, where the last two letters refer to the DNA fragments combined) or from the promoter of the homologous gene (constructions pRA-AD and pRA-CB). S. clavuligerus exconjugants carrying the integrated plasmids were grown in BM medium with glucose or glycerol as sole carbon source. All the strains grew on glycerol but only S. clavuligerus : : pRA-AB and S. clavuligerus : : pRA-AD, expressing S. coelicolor glcP1 and S. clavuligerus glcP, respectively, from the S. coelicolor glcP1 promoter, grew on glucose (Fig. 5b⇓). No growth was observed of S. clavuligerus : : pRA-CD or of S. clavuligerus : : pRA-CB, in which the promoter of the native S. clavuligerus glcP gene was used. This lack of growth confirms that the promoter of glcP is weak or non-functional.
Assembly of S. clavuligerus and S. coelicolor promoters and glucose permease genes. (a) The glcP1 gene of S. coelicolor and the glcP gene of S. clavuligerus with their promoters are shown above. The different promoters and genes assembled in plasmid pRA give the constructs pRA-AB, pRA-AD, pRA-CD and pRA-CB, shown below. (b) Growth of S. clavuligerus : : pRA (▵), S. clavuligerus : : pRA-AB (•), S. clavuligerus : : pRA-AD (○), S. clavuligerus : : pRA-CD (□) and S. clavuligerus : : pRA-CB (▪) on BM medium supplemented with either 1 % (v/v) glycerol (left panel) or glucose (right panel).
The growth on glucose of S. clavuligerus : : pRA-AB and S. clavuligerus : : pRA-AD was delayed as compared with that on glycerol and the cellular DNA content was lower, suggesting a poor flow through the early steps of glucose metabolism in S. clavuligerus.
To improve the growth on glucose, the S. coelicolor glkA gene, expressed from its own promoter, was subcloned into plasmid pRA-AB to give plasmid pRA-AB-K, which therefore carries both glcP1 and glkA from S. coelicolor (Fig. 6a⇓). The good growth of S. clavuligerus : : pRA-AB-K on glucose shows that the combination of glcP1 and glkA increases the growth of the culture (Fig. 6b⇓, left panel).
Effect of S. coelicolor glcP1 and glkA genes on the growth of S. clavuligerus. (a) Plasmid pRA-AB-K used to introduce glkA-glcP1 in S. clavuligerus. (b) Growth of S. clavuligerus : : pRA (▴), S. clavuligerus : : pRA-AB-K (□) and S. clavuligerus : : pRA-AB (○) in glucose-supplemented BM medium (left panel), and glucose utilization by the same S. clavuligerus exconjugants (right panel). (c) Heterologous expression of glcP1 controlled by either the S. coelicolor glcP1 or the S. clavuligerus glcP promoter, tested by RT-PCR. RNA obtained in SA medium at 0 and 9 h (G) after glucose supplementation. (1) S. clavuligerus : : pRA, (2) S. coelicolor, (3) S. clavuligerus : : pRA-AB, (4) S. clavuligerus : : pRA-CB.
Studies of glucose utilization by the exconjugants (Fig. 6b⇑, right panel) indicate that S. clavuligerus : : pRA (control strain) did not consume glucose, but S. coelicolor (not shown) and S. clavuligerus exconjugants carrying one additional copy of the glucose permease genes consumed glucose to different degrees. Glucose depletion was faster in S. clavuligerus : : pRA-AB-K than in S. clavuligerus : : pRA-AB (Fig. 6b⇑, right panel), reflecting the cooperative effect of the glcP1 and glkA genes, while S. clavuligerus : : pRA-AD, expressing the native glcP gene from the glcP1 promoter, showed only 30 % glucose utilization at 96 h (not shown). The glucose consumption rates correlate well with the previously observed growth rates for the different strains (Fig. 6b⇑).
It is interesting that S. clavuligerus : : pRA-AB-K shows good growth in glucose but is not able to use lactose, sucrose, mannitol, galactose or fructose (1 %) as sole carbon source, confirming that GlcP1 is a glucose-specific transporter (see Discussion).
The expression of the heterologous glcP1 gene from its own promoter (in S. clavuligerus : : pRA-AB) and from the glcP promoter (in S. clavuligerus : : pRA-CB, which does not grow on glucose) was compared using the glcP1-specific oligonucleotides GLCP1/GLCP2. As observed in Fig. 6(c)⇑, no amplification of the glcP1 transcript occurred in the negative control S. clavuligerus : : pRA (lane 1) due to the specificity of the primers. In S. clavuligerus : : pRA-AB (lanes 3), a clear signal was observed for glcP1 that was slightly less intense than that observed for S. coelicolor M145 (lane 2). Amplification of the glcP1 transcript in S. clavuligerus : : pRA-CB (lanes 4) was very weak when compared with that of S. clavuligerus : : pRA-AB (lanes 3), confirming the low efficiency of the glcP promoter, but was still perceptible when compared with the negative control S. clavuligerus : : pRA (lane 1). In this figure it can be seen that glucose does not stimulate transcription of the glcP1 gene.
In summary, the lack of glucose utilization in S. clavuligerus can be complemented using the glcP1 gene of S. coelicolor or, simply, the glcP1 promoter coupled to the S. clavuligerus promoterless glcP gene.
DISCUSSION
Glucose utilization in Streptomyces species depends on two key proteins, a glucose permease and a glucose kinase. S. coelicolor glucose kinase formation is constitutive, but its activation is mediated by a glucose-dependent post-translational modification (van Wezel et al., 2007). S. clavuligerus contains a constitutively expressed glkA gene, and shows glucose kinase activity 25 % lower than the activity found in S. coelicolor. Higher levels of glucose kinase activity, obtained by expression of the S. coelicolor glkA gene in S. clavuligerus, did not confer the ability to grow on glucose upon the S. clavuligerus transformants. The same results were obtained by increasing the copy number of orf3, the homologous gene to S. coelicolor SCO2127. This gene has been reported to stimulate transcription of the glucokinase gene in Streptomyces peucetius (Guzmán et al., 2005).
S. coelicolor has two genes, glcP1 and glcP2 (probably originating by gene duplication), to transport glucose (van Wezel et al., 2005). Expression of glcP1 is essential for growth on glucose, while glcP2 can be deleted without appreciable effects (van Wezel et al., 2005). S. clavuligerus, like several other Streptomyces species, contains a single glucose permease gene (glcP), encoding a protein similar to those of S. coelicolor (80 % identical). The glcP2 promoter region, as well as the S. clavuligerus glcP promoter, differ significantly from those of glcP1 of S. coelicolor and the homologous S. avermitilis araE gene.
In this work we demonstrate that the native S. clavuligerus glcP promoter has very low expression. The residual transcription level, originating from the glcP promoter, is not sufficient for glucose utilization. A similar situation occurs with the S. coelicolor glcP2 promoter. This gene (glcP2) is able to confer glucose utilization upon S. clavuligerus when expressed from the tetracycline-induced tcp830 promoter (results not shown), indicating that the lack of function of glcP2 in S. coelicolor is, again, due to poor expression from an inefficient promoter. The similarity with the lack of expression of the second glucose permease gene of S. coelicolor further supports this hypothesis. It is likely that the redundancy of identical glcP1 and glcP2 genes in S. coelicolor has favoured the decay of expression of glcP2 in this actinomycete.
S. clavuligerus glucose permease is functional, as shown by the growth on glucose of S. clavuligerus : : pRA-AD. This was confirmed by complementation of E. coli glucose transport-negative mutants with glcP expressed from the lacZ promoter.
Carbon catabolite regulation in S. coelicolor, in contrast to the regulation mechanism in enterobacteria (Görke & Stülke, 2008), appears to be exerted through the constitutive glucose kinase (GlkA) (Angell et al., 1992, 1994). Proteins GlkA and GlcP have been proposed to form an active complex, GlkA–GlcP, to internalize and phosphorylate glucose (van Wezel et al., 2007). Our results support this proposal. The best growth of S. clavuligerus on glucose was obtained when the homologous glcP1 and glkA genes, both from S. coelicolor, were introduced into S. clavuligerus; however, when one glucose permease of S. coelicolor was combined with the glucose kinase of S. clavuligerus the growth and glucose utilization rate were lower.
S. clavuligerus appears to have adapted to grow on glycerol, maltose or even amino acids as carbon sources, losing the ability to utilize other carbon sources. Introduction of the S. coelicolor glcP1 gene results in utilization of glucose but not of other hexoses. This provides conclusive evidence that glcP, the direct orthologue of glcP1 of S. coelicolor, indeed encodes the glucose-specific transporter of S. clavuligerus. The lack of utilization of several other sugars in S. clavuligerus suggests that this actinomycete is defective in several sugar transporters. Even if the genes encoding those transporters are expressed, lack of interaction of the modified proteins with other proteins in the membrane or in the cytoplasm may explain the lack of growth on those carbon sources.
Acknowledgments
This work was supported by grants from the Spanish Ministry of Science and Technology (BIO2003-03274, BIO2006-14853) and the European Community (Actinogen LSHMCT-2004-005224). DNA sequences were provided via Dr Wilbert Heijne (DSM, The Netherlands). Plasmids pFT76 and pFT256 were provided by Dr Fritz Titgemeyer (Friedrich-Alexander-Universität, Erlangen, Germany), and the glucose transport-deficient mutants E. coli LM1 and E. coli LR2-175 by Dr Knut Jahreis (Osnabrück University, Germany).