Pleiotropic effect of the SCO2127 gene on the glucose uptake, glucose kinase activity and carbon catabolite repression in Streptomyces peucetius var. caesius

Micro-organisms have developed very efficient mechanisms for sensing their nutritional environment, and they respond by adjusting their metabolic activities. This is a global regulatory process, which in the case of carbon source is referred to as carbon catabolite repression (CCR). This mechanism involves control at various levels, including transcriptional expression and protein activity (Brückner & Titgemeyer, 2002).

In various Streptomyces species, glucose exerts CCR that not only affects the utilization of various carbon sources, but also the synthesis of secondary metabolites (Demain, 1989). The mechanisms used by streptomycetes to carry out carbon regulation have been studied primarily in Streptomyces coelicolor. One of these mechanisms is the phosphoenolpyruvate (PEP)-dependent sugar phosphotransferase system (PTS) (Saier et al., 1995; Stülke & Hillen, 1999). However, this mechanism responds only to the presence of fructose and N-acetylglucosamine (Nothaft et al., 2003; Wang et al., 2002). Thus, in contrast to mechanisms found in enterobacteria, the PTS system appears to play a secondary role in CCR of streptomycetes (Titgemeyer et al., 1995). Besides the PTS system, the analysis of Streptomyces mutants insensitive to CCR has suggested the involvement of the expression products of a number of genes, such as bld and ccrA (Champness, 1988; Ingram et al., 1995). Additionally, Hodgson (1982) observed that S. coelicolor mutants isolated by their resistance to growth on the glucose analogue 2-deoxyglucose (2-DOG) present a decreased sensitivity to repression by various carbon sources, as well as reduced glucose kinase (Glk) activity. When these 2-DOG-resistant (Dog^R) mutants were complemented with the glkA gene (Ikeda et al., 1984), Glk activity was completely restored, while CCR sensitivity was only partially recovered (Angell et al., 1994), supporting a role for Glk in this regulatory process. However, a complete recovery can be achieved when the 572 bp SCO2127 gene, located just upstream, is introduced with glkA (Angell et al., 1994). The authors suggest that SCO2127 may encode a protein involved in glucose transport or its metabolism. However, there has been no evidence to support these possibilities.

Our group has worked with Dog^R mutants from Streptomyces peucetius var. caesius that show a similar phenotype (Segura et al., 1996). However, in addition to presenting CCR insensitivity and low Glk activity, our mutants also show difficulties in transporting glucose (Escalante et al., 1999), and this phenotype is partially corrected in revertants of a Dog^R mutant that have recovered 2-DOG sensitivity (Dog^S). However, although the Glk of these Dog^S mutants is increased compared with the parent strain, there is no correlation between their Glk activity levels and the degree of glucose repression (Ramos et al., 2004).

In spite of a large number of studies on the effect of glucose in the genus Streptomyces, the molecular bases for CCR have yet to be clearly established, and there is not a satisfactory regulatory model to explain this phenomenon. With this background, we began to review the effect of the S. coelicolor glkA and SCO2127 genes on Glk activity, glucose transport and CCR sensitivity in a series of mutants derived from S. peucetius var. caesius insensitive to CCR, which present low levels of Glk activity and deficient glucose transport. Our results support a stimulatory role for SCO2127 on CCR in Streptomyces species.

Bacterial strains, plasmids and growth conditions.
All bacterial strains and plasmids used in this study are described in Table 1. The strains are deposited in the UNAM-Biomedicas Culture Collection, Mexico. For anthracycline production and glucose kinase assays, we used 2·5 ml samples of a seed culture (Segura et al., 1996), washed and resuspended in sterile distilled water, to inoculate 50 ml YM medium with the desired glucose concentration. The medium was contained in 250 ml baffled Erlenmeyer flasks. YM medium contains (l¹): 4 g yeast extract, 10 g malt extract, pH 7·2. All cultures were grown at 29 °C on a rotary shaker at 180 r.p.m.

Table 1. Strains and plasmids

The glkA and SCO2127 genes from S. coelicolor were amplified using PCR from total DNA of the M145 strain. The oligonucleotides were designed based on the reported glkA and SCO2127 gene sequences (Angell et al., 1992), and they covered all the promoter and the transcriptional terminator. For glkA, the primers used were: forward (5'-CACGAGATCTGGACGAACG-3') and reverse (5'-GCGAAGCTTCCGGAGTCGG-3'). For SCO2127, the primers were: forward (5'-CGGAGATCTGGCCGCGGGG-3') and reverse (5'-GGCAAGCTTACCCGAGGC-3'). Both primers contained restriction sites for BglII and HindIII for cloning into the pIJ486 vector. This vector was digested with BamHI and HindIII in the multiple cloning site. PCR conditions for glkA and SCO2127 amplification were an initial denaturation step at 94 °C (5 min), then 94 °C (1 min), 57 °C (1 min) and 72 °C (1 min) for 20 cycles, and a final extension period at 72 °C (5 min). PCR products with 694 and 1122 bp were purified, and cloned into pIJ486, generating pSG210 and pSG200, containing SCO2127 and glkA genes, respectively. These plasmids were used to transform the S. peucetius var. caesius Dog mutants. Recombinants containing both genes were obtained by introduction of pIJ513, which contains a 2·9 kb fragment including both SCO2127 and glkA regions (Ikeda et al., 1984), into Dog mutants.

Protoplast formation and regeneration.
Protoplast formation was carried out from strains grown in 50 ml YEME medium supplemented with 5 mM MgCl₂.6H₂O and 0·5 % glycine, as described by Kieser et al. (2000). Protoplast regeneration was carried out in R6 medium (Baltz & Matsushima, 1981) for 40 h. After regeneration, 1 ml sucrose (10·3 %) and 50 µg thiostrepton were added, and incubation continued for another 48 h.

Southern blot analyses.
The glkA and SCO2127 genes from S. coelicolor were used as probes to detect their possible homologues in DNA from S. peucetius var. caesius. Using the method of Kieser et al. (2000), chromosomal DNA was extracted from 48 h cultures grown in YEME supplemented with 5 mM MgCl₂.6H₂O. Total DNA from the original S. peucetius var. caesius strain, and from S. coelicolor M145 as a positive control, was digested separately with BamHI, EcoRI, PstI and XhoI. After digestion, 10 µg of each digest was run on a 1 % agarose gel. Southern blotting and hybridization were carried out according to the manufacturer's recommendations for the DIG High-Prime DNA Labelling and Detection Kit II (Roche Molecular Biochemicals).

Anthracycline and protein determination.
Anthracyclines were extracted from harvested mycelia (96 h cultures) using acetone and 0·05 M sulfuric acid (4 : 1), according to Arcamone et al. (1969). Anthracyclines were quantified at 495 nm using a molar absorption coefficient of 220. For protein determination, samples were processed as previously reported (Segura et al., 1997), and assayed by the Lowry method, using bovine serum albumin as a standard.

Glucose kinase assays.
Glk activity from cell-free extracts prepared from 48 h mycelial cultures of the strains was measured spectrophotometrically by monitoring the reduction of NADP in a glucose-6-phosphate-dehydrogenase-coupled reaction at pH 7·2 and 25 °C, as previously reported (Imriskova et al., 2001; Ramos et al., 2004). Glk activity was determined at 25 °C and pH 7·5, and expressed as the amount of enzyme that produces 1 nmol NADPH min¹.

Glucose uptake experiments.
A 50 ml volume of the seed culture was used to inoculate a 2·8 l Fernbach flask containing 500 ml of a chemically defined medium (CD medium) with 100 mM D-glucose, as previously reported (Escalante et al., 1999). After 36 h, mycelia (250 mg wet weight) were harvested, washed with 0·85 % NaCl, and resuspended in a vial containing 4·5 ml saline. The suspension was incubated under agitation, and glucose transport was initiated by the addition of 5 µCi (185 kBq) D-[¹⁴C]glucose (38·8 MBq mmol¹) in 475 µl 10 mM cold glucose, as previously reported (Escalante et al., 1999). Radioactivity was determined by soaking the filter in vials containing 4 ml of a commercial liquid scintillation counting solution.

Isolation of RNA, and dot blot analyses.
Total RNA of S. peucetius var. caesius was extracted from mycelia of the original strain and its derived Dog^R mutant, as well as from the Dog^R mutant transformed with SCO2127. RNA was isolated from cultures grown in YM medium with 100 or 500 mM glucose for 24 h (exponential growth phase) by the method of Chomczynski & Sacchi (1987). Residual DNA was eliminated by incubating the samples with DNase I for 1 h at 37 °C. Total RNA was precipitated by addition of 0·1 vol. 3 M sodium acetate and 1 vol. ice-cold 2-propanol. RNA concentrations were determined spectrophotometrically, and RNA quality was checked by agarose-formamide electrophoresis, according to Kieser et al. (2000). RNA was stored at 70 °C for no more than 2 weeks. For dot blot analysis, the RNA concentration was adjusted to 1 µg µl¹ with GFP (1 M glyoxal, 10 mM phosphate and 70 % formamide), and then denatured at 55 °C for 15 min. The resulting RNA was kept in an ice-bath until loaded onto Hybond-N⁺ membranes (Amersham). The amounts of total RNA applied to the membranes were 2, 5, 10 and 20 µg. To detect the glk mRNA, a digoxigenin-UTP-labelled S. peucetius var. caesius glk probe (900 bp) was utilized (DIG High-Prime DNA Labelling and Detection Starter Kit II; Roche). The hybridization assay was carried out overnight at 54 °C. The glk mRNA was developed with an alkaline-phosphataseantidigoxigenin conjugate, and CSPD [disodium 3-(4-methoxyspiro{1,2-dioxetane-3,2'-(5'-chloro)tricyclo[3.3.1.1^3,7]decan}-4-yl)phenyl phosphate; Roche] as the light emission substrate, and the membrane was exposed to an autoradiography film. Total RNA from Paenibacillus amylolyticus was used as a negative control.

Reproducibility of results.
The results shown are the mean values of at least two independent experiments carried out in triplicate. The observed variations were consistently less than 10 %.

Transformation of Dog mutants
Protoplasts from Dog mutants were transformed with plasmids containing the S. coelicolor glkA and SCO2127 genes individually or together. From each Dog mutant we isolated between one and three transformants containing the different constructs. Table 1 describes each of the seven strains derived from the transformation experiments. After subculturing these transformants for various numbers of generations under selective pressure, we were still able to recover our plasmids, suggesting that recombination into the bacterial chromosome does not occur.

Presence of glkA and SCO2127 homologues in S. peucetius var. caesius
In order to corroborate the presence of the endogenous glkA and SCO2127 genes in S. peucetius var. caesius, we carried out Southern blot experiments using probes obtained from S. coelicolor. As can be seen in Fig. 1, there was good hybridization in S. peucetius var. caesius with both probes, and it was comparable with that found with the S. coelicolor DNA used as the positive control. Additionally, the hybridization profiles found using glkA and SCO2127 were identical to each other regardless of the restriction endonucleases used, showing that these genes are encoded on the same DNA fragment, and thus suggesting that the genomic organization of these genes is similar to that found in S. coelicolor, where they are found adjacent to each other, possibly in the same operon.

(49K): Fig. 1. Southern blot hybridization of total DNA from (a) S. coelicolor M145 and (b) S. peucetius var. caesius original strain. The S. coelicolor A3(2) M145 glkA and SCO2127 genes were used as probes. Total DNA (10 µg) was digested separately with BamHI, EcoRI, PstI and XhoI. Lanes M contained molecular size markers.

Sensitivity to glucose repression of the recombinants
In order to examine the effect that glkA and SCO2127 have on CCR sensitivity, we analysed the effect of two different glucose concentrations upon anthracycline biosynthesis in our Dog mutant strains and their transformants. As shown in Table 2, compared with synthesis on 100 mM glucose (non-repressive concentration), all our mutants showed resistance to CCR in 500 mM glucose. However, with the exception of the Dog^S-2 mutant transformed with glkA, all other transformed strains obtained showed CCR sensitivity at both repressive and non-repressive glucose concentrations, and sensitivity to glucose repression was higher when compared with that of the original strain. Dog mutants transformed with plasmids without inserts did not show differences in sensitivity from that of the untransformed mutants. An effect similar to that found for production of anthracyclines was observed when the synthesis of β-galactosidase was monitored in these strains (not shown). This suggests that an increase in flux through glycolysis, even at lower glucose concentrations, is able to induce CCR. According to their non-integrative characteristic (Ward et al., 1986), after subculturing these transformants in the absence of a selective pressure (without thiostrepton), the plasmids were lost, and the mutants recovered their resistance to CCR in 500 mM glucose, showing that no recombination took place between plasmids and chromosomal DNA.

Table 2. Anthracycline production by different S. peucetius var. caesius mutants and their recombinants Anthracycline production was determined after 120 h incubation in YM medium supplemented with low (100 mM) and high (500 mM) D-glucose.

Glucose kinase activity
To elucidate the mechanisms by which the mutant strains transformed with glkA and/or SCO2127 recover CCR sensitivity, we investigated whether there was a connection between this phenotype and Glk activity. As can be seen in Table 3, with relation to the Dog mutants, all of the recombinant strains showed an increase in Glk activity. In some strains, Glk activity increased to almost 320 %. Additionally, the increase in activity was observed regardless of which gene was expressed (glkA or SCO2127). In situ Glk activity analysis shows a single band with identical electrophoretic mobility (±32 kDa) in all of our strains (data not shown).

Table 3. Glucose kinase activity levels of cell extracts from the original strain and several DogR and DogS mutants of S. peucetius var. caesius and their recombinants Strains were grown for 48 h in YM medium supplemented with 100 mM D-glucose.

Glucose uptake
Considering that glucose metabolism starts with its incorporation into the cell, we explored the connection between the phenotype of our recombinants and this process by determining the capacity of these strains to incorporate D-[U-¹⁴C]glucose. Table 4 shows that all of the recombinants expressing SCO2127 alone, SDR-2 (obtained from Dog^R), SDS-2 (obtained from Dog^S-2) and SDS-12 (obtained from Dog^S-11), showed a significant increase in glucose incorporation (between 1·7 and 3 times). Similarly if the mutant strains were transformed with the 2·9 kb fragment expressing both SCO2127 and glkA (transformants SDS-3 and SDS-13), there was a strong increase in glucose transport. This effect does not seem to be due to the plasmid copy number, since the 2·9 kb fragment was contained in the low-copy-number plasmid pIJ513. Conversely, none of the transformants containing only glkA showed significant changes in transport efficiency compared to the non-transformed strains.

Table 4. D-Glucose uptake by the original strain of S. peucetius var. caesius, and several Dog mutants and their recombinants Strains were grown for 36 h in CD medium supplemented with 100 mM glucose.

Interestingly, when Dog^S-2 was transformed with SCO2127, either alone or in combination with glkA, glucose incorporation was increased three times. However, this increase barely reached 50 % of the transport capacity found in the original strain.

RNA transcription
To gain more information about the stimulatory mechanism of SCO2127, transcription of glk was monitored by performing RNA dot blot experiments. As shown in Fig. 2, when glk mRNA levels from the original strain and a CCR-resistant mutant (Dog^R) were compared with those of the Dog^R mutant transformed with SCO2127 (SDR-2), an increase in the glk mRNA levels was observed in the transformed mutant compared with those of the other strains. In contrast, lower glk mRNA levels were obtained in the Dog^R mutant. In addition, in comparison with cultures grown with 100 mM D-glucose, higher glk mRNA levels were observed when two of the strains (original and Dog^R mutant) were cultivated with 500 mM D-glucose. On the other hand, these differences were not found in cultures of the Dog^R mutant transformed with SCO2127. The glucose permease mRNA levels were not determined, since the sequence of this gene was only recently described (Bertram et al., 2004).

(66K): Fig. 2. RNA dot blot of glk mRNA from the original strain, a DogR mutant of S. peucetius var. caesius, and a recombinant derived from this mutant grown in YM medium with 100 (L) or 500 mM (H) D-glucose (SDR-2). The amounts of total RNA applied are indicated on the left. As an example of the relative signal intensities, the bar graph shows a densitometric quantification derived from the 2 µg RNA dot blot signals. The data are the means of three independent experiments. Standard deviations are indicated by error bars.

There is mounting evidence that Glk is not sufficient to elicit CCR in streptomycetes. It has been shown in S. coelicolor that resistance to 2-DOG yields mutants (Dog^R) with decreased Glk activity, and reversal of glucose repression (Hodgson, 1982). However, when these mutants are transformed with the glkA gene there is only a partial recovery of CCR sensitivity (Angell et al., 1994). Complete recovery is only attained when the region containing both glkA and the SCO2127 gene, located directly upstream of glkA, is introduced (Angell et al., 1994). Additionally, Glk is not necessary for glucose repression of the chi63 (chitinase) in S. coelicolor (Ingram & Westpheling, 1995), or for the expression of α-amylase in Streptomyces kanamyceticus (Flores et al., 1993). This would suggest the existence of other molecules important for the regulation of CCR exerted by glucose.

Mutants of S. peucetius var. caesius resistant to the inhibition of carbon source utilization by 2-DOG (Dog^R mutants) exhibit a pleiotropic phenotype, with decreased Glk activity and low glucose incorporation, which seem to cause reversal of glucose repression (Segura et al., 1996; Escalante et al., 1999). By applying an enrichment procedure to the Dog^R strain, several mutants were isolated that regained sensitivity to 2-DOG inhibition (Dog^S). Although sensitive to 2-DOG, these mutants exhibit different glucose uptake/Glk ratios, which render them totally or partially sensitive to CCR (Ramos et al., 2004). In the present work, we have shown that both Dog^R and Dog^S phenotypes can be reverted to that of the original strain by transforming the mutants with plasmid pSG210, which expresses the 620 bp SCO2127 gene of S. coelicolor. In these recombinant strains, glucose uptake and Glk activity values were restored to similar or even higher levels than the original strain, and the recombinants regained sensitivity to CCR. The effect of SCO2127 on Glk activity correlated well with the glk mRNA levels observed in a CCR-resistant mutant transformed with this region. The above-mentioned results were quite unforeseen considering that the SCO2127 region does not seem to code for either a glucose permease (Bertram et al., 2004) or a glucose kinase (Angell et al., 1994). Similar results were obtained by transformation with plasmid pIJ513 (Ikeda et al., 1984), which contains both SCO2127 and glkA. As expected, recombinants with plasmid pSG200 containing a 950 bp fragment with the glkA gene alone reverted to normal Glk activity values, but glucose uptake was unchanged.

Our hybridization experiments suggest that the SCO2127 and glkA genes have a similar genomic organization in both S. coelicolor and S. peucetius var. caesius, or are at least in close proximity within the chromosome. However, there appear to be no data published on the activity and expression of SCO2127 in either of these systems (Angell et al., 1994).

With regard to CCR sensitivity in the mutant strains transformed with either SCO2127 or glkA, anthracycline production showed increased sensitivity to the glucose effect, even at low carbohydrate concentrations (100 mM). This oversensitivity could be explained in terms of a better performance of S. coelicolor SCO2127 and Glk activities in the S. peucetius var. caesius intracellular environment, or by the high-copy-number plasmids harbouring either SCO2127 or glkA. An exception to this observation was obtained with the Dog^S-2 mutant transformed with glkA. However, this recombinant exhibited about 80 % less glucose incorporation compared to that of the original strain. Therefore, although containing a high Glk activity (threefold higher), its glucose availability is probably very low.

As expected, under glucose-repressive conditions (500 mM), the original strain displayed higher glk mRNA levels than those observed in non-repressive concentrations (100 mM), again supporting the role of Glk activity in CCR sensitivity. These differences were also observed in a mutant resistant to CCR (Dog^R). In this case, the low glk mRNA levels observed under both repressive and non-repressive conditions correlated well with its CCR-resistant phenotype (Ramos et al., 2004). On the other hand, although higher glk mRNA levels were observed in the Dog^R strain transformed with SCO2127, no differences were observed in this parameter under repressive and non-repressive conditions, explaining the oversensitivity of this recombinant to the glucose effect.

The question arises of how SCO2127 mediates its stimulating effect on glucose uptake and Glk activity. Our data from the SCO2127-transformed strains do not support the occurrence of recombination between pSG210 and the mutants' chromosomal DNA, thus eliminating the possibility of a highly transcribed SCO2127 promoter as the cause of the stimulatory effect on Glk activity and glucose uptake. In addition, although the glk gene is located next to SCO2127, and in theory may be susceptible to a stimulatory effect from the SCO2127 promoter, the glcP1 (SCO5578) region, which was recently described as encoding a non-PTS glucose permease in S. coelicolor (Bertram et al., 2004), was found to be located not in the vicinity of SCO2127, but on the opposite side of the conserved region of the chromosome. Considering the increase in glk mRNA levels observed in one of the CCR-resistant mutants transformed with SCO2127, we suggest that SCO2127 is involved in stimulating transcription of glk, and probably that of the glucose permease gene as well. In conclusion, our data suggest the participation of an integral regulatory system, which is initiated with an increase in glucose incorporation and metabolism, resulting in an increased synthesis of catabolites, which may be involved in eliciting CCR in this micro-organism. In agreement with this possibility, among several products of glucose metabolism, fructose 1,6-bisphosphate and PEP exert CCR on anthracycline formation in S. peucetius var. caesius (Ramos et al., 2004), with fructose 1,6-bisphosphate being the most effective. This effect resembles that reported for fructose 1,6-bisphosphate on the phosphorylation of Hpr kinase from Bacillus subtilis as a preliminary step for CCR (Jault et al., 2000). This would suggest an increase in glycolytic flux as a requisite for the establishment of the phenomenon of CCR in S. peucetius var. caesius.

We thank Rafael Cervantes for excellent technical assistance. We are indebted to Marco Antonio Ortiz for strain-preservation studies. This work was partially supported by the grant IN-202903 from Dirección General de Asuntos del Personal Académico, UNAM, México. Silvia Guzmán was recipient of a doctoral fellowship from Consejo Nacional de Ciencia y Tecnología, México, and from Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México.

Footnotes

^†, Ruth López †Present address: Department of Medical Genetics, 8-33 Medical Sciences Building, University of Alberta, Edmonton, Alberta, Canada T6G 2H7. ‡Present address: Facultad de Medicina, Universidad Autónoma de Campeche, Campeche, Cam. 24090, Mexico.