Streptomyces coelicolor produces a brown pigment on nutrient-limited agar medium (Tyr-PM) using l-tyrosine as the sole nitrogen and carbon source. The pigment production is associated with the second step of l-tyrosine catabolism catalysed by 4-hydroxyphenylpyruvate dioxygenase (HppD), which converts 4-hydroxyphenylpyruvate (4HPP) to 2, 5-dihydroxyphenylacetate (homogentisate) to provide the carbon and energy substrates for the growth of S. coelicolor on Tyr-PM. An hppD mutant did not produce brown pigment, and its normal growth was impaired on Tyr-PM. hpdA and hpdR, located close to hppD, were identified as activator and repressor genes for hppD transcription in the presence of tyrosine. hpdA, divergently transcribed from hppD, is negatively autoregulated in the absence of tyrosine, whereas it is repressed by both its own protein and HpdR in the presence of tyrosine. Electrophoretic mobility shift assays and footprinting experiments showed that HpdA and HpdR each bind to an overlapping region spanning the promoters of both hppD and hpdA, and that 4HPP, instead of tyrosine, is the specific ligand modulating the binding patterns and footprints of HpdA and HpdR on the hppD–hpdA promoter region. These results suggested that the transcription of hppD is subject to coarse and fine control by a complex regulatory system.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Tyrosine has characteristic anabolism and catabolism in organisms (Yang et al., 2002; Moran, 2005). The oxidative degradation of tyrosine is significant not only for human and animals (Grompe, 2001), but also for the normal growth and development of some bacteria (Milcamps and Bruijn, 1999). The main catabolic pathway of tyrosine includes five biochemical reactions and generates fumarate and acetoacetate, which are integral to the Krebs cycle (Fig. 1). As ketogenic and glucogenic products, fumarate and acetoacetate make direct energetic contributions to growth. Therefore, tyrosine catabolism could supply the nitrogen, carbon and energy substrates for organisms through the Krebs cycle. In the second step of tyrosine catabolism, 4-hydroxyphenylpyruvate dioxygenase (HppD) converts 4-hydroxyphenylpyruvate (4HPP) to 2,5-dihydroxyphenylacetate (homogentisate, HMG or HGA). This conversion involves decarboxylation, aromatic hydroxylation and substituent migration in a single catabolic step (Moran, 2005). HppD, as a member of the α-keto-acid-dependent dioxygenase subclass, exists in all organisms and requires FeII, α-ketoglutarate and dioxygen for catalysis (Neidig et al., 2005). Homogentisate, the product of HppD transformation, usually becomes oxidized to a quinoid derivative, which eventually generates melanin compounds by spontaneous polymerization, conferring a characteristic brown colour to the medium (Denoya et al., 1994; Johnson-Winters et al., 2003).
There is limited information on the regulatory elements related to the genes of the tyrosine catabolism. A tyrosine aminotransferase gene (tyrB), functioning in both the anabolism and the catabolism of tyrosine in Escherichia coli, is repressed by TyrR (Yang et al., 2002; Pittard et al., 2005). In Pseudomonas putida, the global regulator Crc (for ‘catabolic repressor control’) represses the transcription of hppD and hmgA (homogentisate dioxygenase), but does not bind to their promoters (Morales et al., 2004); and the hmgABC operon involved in the degradation of l-tyrosine is repressed by an IclR-type regulator, HmgR (Arias-Barrau et al., 2004). The transcription of hmgA in Sinorhizobium meliloti is induced by nutrient deprivation and controlled by an ArsR family regulator (Milcamps et al., 2001). However, no specific regulator of hppD transcription has been reported in bacteria so far.
Streptomycetes are Gram-positive, filamentous soil bacteria known for the production of many useful antibiotics and for their complex development (Chater, 2000), and they also display amazing versatility in their ability to use relative poor sources of carbon and nitrogen. Tyrosine is one of a wide variety of compounds that can be used as sole nitrogen and carbon source by streptomycetes in nutrient-deprivation condition (Pometto and Crawford, 1985). Genetic analysis of the tyrosine catabolic pathway in Streptomyces has largely been confined to hppD. Overexpression of hppD from Streptomyces avermitilis (Sav hppD) in E. coli resulted in brown pigment production in the culture broth (Denoya et al., 1994). The structure of HppD and the basis for ordered substrate addition for HppD conversion were also studied for Sav HppD (Johnson-Winters et al., 2003; Brownlee et al., 2004). Tyrosine and its catabolic intermediate 4HPP are also important biosynthetic precursors for some antibiotics, including saframycin A (Mikami et al., 1985), clorobiocin (Pojer et al., 2003), lincomycin A (Neusser et al., 1998), novomycin (Eustaquio et al., 2003) and calcium-dependent antibiotic (Hojati et al., 2002) in streptomycetes.
Despite its important roles in tyrosine catabolism and antibiotic biosynthesis in Streptomyces (Fig. 1), the in vivo function and transcriptional regulation of HppD are still unknown. Progress in genome sequencing allowed us to compare the sequences of hppD and its flanking genes among different streptomycetes and other related strains. We found that a gene for an AsnC-type regulator (or feast/famine regulatory protein) and another for an IclR (isocitrate lyase regulator)-type regulator constitute a potential regulon of hppD in S. coelicolor, S. avermitilis and S. ansochromogenes (Fig. 2). Both AsnC and IclR members play important regulatory roles in bacteria as DNA binding proteins (Molina-Henares et al., 2006; Yokoyama et al., 2006). AsnC regulators control the metabolism of leucine (Cui et al., 1996), asparagine (Poggio et al., 2002), l-alanine (Lodwig et al., 2004) and proline (Keuntje et al., 1995; Yokoyama et al., 2006), but no AsnC regulator has been shown to control tyrosine catabolism in bacteria up to now. IclR family members play various functions in bacteria (Molina-Henares et al., 2006). Some of them control metabolism of aromatic compounds, such as 3-(3-hydroxyphenyl) propionic acid and protocatechuate (Barnes et al., 1997; Popp et al., 2002; Torres et al., 2003). HmgR is an IclR-type regulator involved in the degradation of l-tyrosine, repressing the expression of the hmgABC operon in P. putida (Arias-Barrau et al., 2004), but whether it regulates the transcription of hppD as well is not clear.
In this study, we report on the ligand-dependent regulatory system of hppD transcription in S. coelicolor.
hppD, hpdA and hpdR play important roles in the tyrosine catabolism and normal growth of S. coelicolor
blast searches in GenBank revealed that an AsnC-type gene and an IclR-family gene are close to hppD in three Streptomyces (Fig. 2), whereas in most other bacteria, hppD is linked with only one of the potential regulatory genes (summarized by Arias-Barrau et al., 2004). To study whether the organization of these linked genes has specific biological significance in Streptomyces, and whether they all participate in tyrosine catabolism, disruption mutants of hppD (SCO2927), the asnC-like gene (SCO2928, named hpdA) and the iclR-like gene (SCO2935, named hpdR) were constructed in S. coelicolor M145 (Fig. 3A). On Tyr-PM (Tyrosine-containing and poor nutrient medium) agar containing 5 mM tyrosine as the sole nitrogen and carbon source, M145 developed normally and produced brown diffusible pigment after 48 h growth, whereas the hppDdisruption mutant (hppDDM) and hpdAdisruption mutant (hpdADM) both displayed poor growth with little aerial mycelium and reduced sporulation, and also lacked brown pigment. The hpdRdisruption mutant (hpdRDM) showed slightly enhanced growth and production of brown pigment in comparison with M145 (Fig. 3B). Wild-type pigment production and growth were restored by the introduction of the relevant wild-type gene into hppDAM, hpdADM and hpdRDM, excluding potential polar effects on adjacent genes. The biomass and pigment production of M145, hppDDM, hpdADM and hpdRDM during the life cycle were also determined on Tyr-PM (Fig. 3C and D), which showed that the normal growth and development of all the strains reached stationary phase after 48 h incubation. Thus, S. coelicolor can utilize l-tyrosine as the sole nitrogen and carbon source in nutrient-limited conditions and produce diffusible brown pigment, both hppD and hpdA play positive roles in the tyrosine catabolism, and hpdR plays a negative role.
The transcription of hppD is activated by hpdA and repressed by hpdR in response to tyrosine in nutrient-deprived conditions
The effect of hpdA and hpdR on the transcription of hppD was analysed by S1 mapping in the wild-type M145, hpdADM and hpdRDM strains grown for 48 h on minimal medium (MM) in the absence of tyrosine, or on Tyr-PM medium in the presence of tyrosine. No hppD transcript was detected in M145 grown on MM, whereas transcript was readily detected in M145 grown on Tyr-PM with tyrosine (Fig. 4), indicating that hppD transcription was tyrosine-induced in M145. In the hpdR mutant, hppD transcript was likewise undetectable in the absence of tyrosine, whereas it was about fourfold higher than that of M145 in the presence of tyrosine, implying that hpdR negatively regulates hppD transcription in the presence of tyrosine. In the hpdA mutant, hppD transcript was undetectable either in the absence or in the presence of tyrosine (Fig. 4), suggesting that hpdA is necessary for hppD transcription. Similar transcriptional results were obtained with RNA samples isolated from later time points (72, 96, 120, 144 h) (data not shown). These results indicated that the l-tyrosine-induced transcription of hppD is dependent on hpdA-mediated activation and subject to hpdR-mediated repression.
HpdA and HpdR bind specifically to Pd-a
To prove whether the regulatory effects of hpdA and hpdR on hppD transcription are mediated in a direct manner, electrophoretic mobility shift assays (EMSAs) were performed using a 435 bp DNA probe (Pd-a) spanning the hppD–hpdA intergenic region and partial coding sequences as a target. HpdA could bind to Pd-a fragment at protein concentrations ranging from 15 to 960 nM. The saturating concentration of HpdA for stable protein–DNA complex formation was 480 nM. HpdR could bind to Pd-a fragment at protein concentrations from 120 to 960 nM, and the saturating HpdR concentration for stable protein–DNA complex formation was 480 nM (Fig. 5A). Specificity of binding was examined with the addition of 50-fold excess unlabelled probe, which abolished HpdA (480 nM) or HpdR (480 nM) binding to the labelled fragment (data not shown).
HpdA and HpdR bind to an overlapping region containing the promoters of hppD and hpdA
The EMSAs and transcription results indicated that HpdA and HpdR might regulate the transcription of hppD via direct binding to the Pd-a region, potentially including the hppD and hpdA promoters. To analyse the regulation mechanism clearly, the promoter structures of hppD and hpdA and the binding sites of HpdA and HpdR on Pd-a were identified. The transcription start point of hppD was localized by high-resolution S1 nuclease mapping to G or T at position 64 or 65 nt upstream of the potential start codon (ATG) of hppD (Fig. 5B and D). Putative −10 (AGAACC) and −35 (TATTGC) regions with a spacer of 18 bp were indicated. The transcription start point of hpdA was localized to a G at position 58 nt upstream of the potential start codon (ATG) of hpdA (Fig. 5B and D). A −10 region (CAACAT) and a −35 region (GCAATA) with a spacer of 18 bp were deduced. The divergent transcription start points of hppD (Pd) and hpdA (Pa) were separated by 67 nt, and their potential −35 regions exactly overlapped (Fig. 5D). This organization is similar to that of a Lrp-like activator gene (putR) and its target gene (putA) in Rhodobacter capsulatus, where putR and putA are divergently transcribed and have overlapped −35 regions in their promoters (Keuntje et al., 1995).
To determine the binding sites of HpdA and HpdR, DNase I footprinting experiments were performed, using the Pd-a fragment end-labelled on either the top or bottom strand (Fig. 5C). On the top strand, HpdR protected one region stretching from positions −25 to −119 relative to the transcription start point of hppD, and on the bottom strand, three separate sites were protected or affected by HpdR from positions −2 to −8, −28 to −72 and −85 to −120 nt relative to the transcription start point of hppD. In the case of HpdA, the footprint regions extended from positions −11 to −109 on the top strand and from positions −17 to −120 on the bottom strand relative to the transcription start point of hppD (Fig. 5C and D). Thus, HpdA and HpdR bind to an overlapping region spanning the complete promoter region of hpdA and the −35 region of the hppD promoter (Fig. 5C and D).
hpdA is negatively autoregulated in the absence of tyrosine and repressed by hpdA and hpdR in the presence of tyrosine
As the binding sites of HpdR and HpdA spanned the entire promoter region of hpdA and part of the hppD promoter region (Fig. 5D), hpdR and hpdA might regulate the transcription of hpdA as well. Additionally, as a new member of the AsnC family, HpdA might show negative autoregulation, like other homologues reviewed by Yokoyama et al. (2006). To confirm these hypotheses, S1 mapping was performed to detect hpdA transcription in M145, hpdADM and hpdRDM strains grown on MM or Tyr-PM agar medium in the absence or presence of tyrosine, using a smaller probe corresponding to the 5′-end of its coding region and its upstream regulatory region (see Experimental procedures). The result showed a clear, albeit complex, hpdA transcription pattern. In M145 grown on MM or Tyr-PM medium, there was equal and low-level hpdA transcription (Fig. 6A). In the absence of tyrosine, hpdA transcription in the hpdA mutant was fourfold more than that of the parental strain, whereas that in the hpdR mutant was at the same level as that of the parental strain, indicating that hpdA was negatively autoregulated and was not affected by hpdR in the absence of tyrosine. In the presence of tyrosine, hpdA transcription in the hpdA mutant was about 20-fold more than that in the parental strain, whereas that in the hpdR mutant was fourfold more than that of the parental strain, indicating that hpdA is subject to autorepression and hpdR-mediated repression in the presence of tyrosine. The reasons for the different autorepression of hpdA in the absence or presence of tyrosine will be considered later (see Discussion). Taken together, hpdA has a basal transcription level whether tyrosine is present or not, is negatively autoregulated whether tyrosine is present or not, and is specifically repressed by hpdR in the presence of tyrosine.
The transcription of hpdR is independent of tyrosine
HpdR binds to the overlapped promoter regions of hppD and hpdA (Fig. 5D) and represses the transcription of hppD and hpdA only in the presence of tyrosine, which suggested that: (i) the transcription of hpdR might be tyrosine-induced, or (ii) hpdR is constitutively expressed independent of the presence of tyrosine, but represses hppD and hpdA expression only in the presence of tyrosine or tyrosine catabolic intermediates. To distinguish between these possibilities, the transcription of hpdR was determined by S1 mapping in the wild-type strain (M145) grown in the absence or presence of tyrosine. Transcription of hpdR was constitutive and independent of tyrosine (Fig. 6B). Only the second hypothesis, that tyrosine or its catabolic intermediate may function as a co-repressor with HpdR in the repression of hppD and hpdA, was compatible with these results.
4HPP modulates the binding patterns of HpdA and HpdR
To determine the effect of tyrosine or its catabolic intermediate 4HPP on the binding capacities of HpdA or HpdR, EMSAs were carried out. When l-tyrosine was added from 1 to 28 mM, the gel shift patterns of HpdA or HpdR on Pd-a were not changed, indicating that l-tyrosine is not the direct ligand for HpdA or HpdR (Fig. 7A). Considering the potential existence of substrate-dependent regulation for hppD, we tested whether 4HPP, the substrate of HppD, might affect the DNA binding affinities of HpdA and HpdR. Indeed, both proteins changed their binding patterns obviously when 4HPP was added at 3 or 6 mM (Fig. 7B and C). HpdA, in the absence of 4HPP, formed several HpdA–Pd-a complexes at different protein concentrations (Figs 5A and 7B), while in the presence of 4HPP at more than 3 mM, only a single protein–DNA complex with short mobility shift accumulated in the gel regardless of the concentrations of HpdA, indicating that 4HPP could alter the multimeric state of HpdA on the target DNA. Combining these results with the transcriptional analysis of hppD and hpdA (Figs 4 and 6) and the EMSA results for HpdA without or with 4HPP (Fig. 7B) led to the following in vivo model. HpdA(4HPP) assembly was deduced to be an effective form of HpdA, functioning both in the activation of hppD transcription and in the repression of hpdA in the presence of tyrosine, whereas the HpdA–DNA complexes formed in the absence of 4HPP (Figs 5A and 7B) might function to repress hpdA in the absence of tyrosine.
The equivalent HpdR experiment showed that HpdR could bind to Pd-a in the absence of 4HPP to form different HpdR–DNA complexes at altered protein concentrations (Fig. 5A and 7C). In the presence of 4HPP (3 or 6 mM), various HpdR(4HPP)–DNA complexes formed, depending on the molar ratio between [4HPP] and [HpdR]. When the concentration of HpdR was 480 nM, several altered HpdR(4HPP)–DNA complexes accumulated, whereas when the concentrations of HpdR were lower than 240 nM, a single HpdR(4HPP)–DNA complex of low mobility shift accumulated. Considering these data along with our observations that the repression function of HpdR in vivo specifically required the presence of tyrosine (Figs 4 and 6), HpdR was deduced to be able to sense the concentration changes of 4HPP in vivo and form various HpdR(4HPP) assemblies with different binding capacities on Pd-a, hence repressing the transcription of hppD and hpdA in a [4HPP]-dependent way.
In summary, 4HPP, instead of tyrosine, is the specific ligand for HpdA and HpdR, and it alters the functions of HpdA or HpdR in the transcriptional regulation of hppD and hpdA by triggering assembly transitions of HpdA or HpdR.
4HPP changed the footprints of HpdA and HpdR on Pd-a
To disclose the potential effect of 4HPP on HpdA or HpdR footprints, DNase I digestions were performed at the protein concentration of 240 nM, at which distinct HpdA(4HPP)–Pd-a or HpdR(4HPP)–Pd-a complexes formed in EMSAs (Fig. 7B and C). The DNase I sensitivity of 10 sites in the HpdA footprint was changed in the presence of 10 mM 4HPP (Fig. 8). However, DNase I sensitivity was unchanged in the presence of 5 mM 4HPP (Fig. 8), in contrast with the EMSA results, where distinct HpdA(4HPP)–Pd-a complex could be formed in the presence of 240 nM HpdA and 3 or 6 mM 4HPP. We suppose that this difference may reflect the use of abundant DNA probe in DNase I footprinting experiments compared with small amounts of DNA probe in EMSAs. For HpdR, the DNase I sensitivity of two sites in the footprint was changed in the presence of 5 mM 4HPP, but that of five sites was changed in the presence of 10 mM 4HPP. Thus, 4HPP changed the footprints of both HpdA and HpdR, indicating that it caused changes of DNA conformation that might directly influence the binding of other regulators or RNAP to Pd-a, and ultimately affect the transcription of hppD or hpdA.
Tyrosine catabolism is important for the normal growth of S. coelicolor under nutrient-deprivation and tyrosine-supply condition
In previous studies, it was shown that S. coelicolor responds to nitrogen deprivation by inducing pathways for alterative nitrogen sources, such as valine or nitrate (Tang and Hutchinson, 1995). Our data proved that S. coelicolor could use tyrosine as nitrogen or carbon sources in nutrient-limited conditions. The first step of tyrosine catabolism is responsible for the supply of nitrogen, and the following four steps after deamination of tyrosine are responsible for the supply of the carbon and energy substrates (Fig. 1). In this pathway, HppD is one of the key enzymes (Moran, 2005). Therefore, if hppD is knocked out or silent, tyrosine catabolism should be shut off. But our data showed that hppDDM and hpdADM could still use tyrosine as the sole nitrogen and carbon source, albeit inefficiently. We try here to explain this unexpected phenomenon. In S. avermitilis, both HmaS (4-hydroxy mandelate synthase) and HppD can utilize the same substrate, 4HPP, but they give different products: (S)-(4-hydroxy) mandelate (HMA) and homogentisate respectively (Neidig et al., 2005). An hmaS-like gene (SCO3229) is present in S. coelicolor, so the accumulated 4HPP in hppDDM and hpdADM might be partially converted by HmaS to HMA and, eventually, provide limited carbon and energy substrates for growth.
hppD transcription level is determined by HpdA, HpdR and 4HPP
Even though both HpdA and HpdA(4HPP) could bind specifically to about 100 bp of DNA sequence spanning the −10 and −35 regions and the upstream sequence of the hppD promoter, HpdA(4HPP) is considered to be the functional form in the activation of hppD transcription in vivo. The binding of AsnC/Lrp regulators to their target DNA usually modulates the DNA geometry and induces DNA bending, which can enhance the binding of RNAP and other regulators (Thaw et al., 2006). Thus, the binding of HpdA(4HPP) to the hppD promoter region might recruit RNAP to initiate the transcription of hppD.
HpdR binds to a long sequence almost overlapping that of HpdA, suggesting that HpdR might compete with HpdA for DNA binding or directly interfere with the binding of RNAP to the hppD promoter to repress hppD transcription. Study of the interaction between E. coli IclR and RNA polymerase has shown that IclR represses the transcription of aceB by two modes (Yamamoto and Ishihama, 2003). Likewise, our data showed that there are two characteristic sequences in the binding region of HpdR spanning the hppD promoter: one comprises two imperfect 5′-CCCGCACCCN(5/8)CTGGTCATC-3′ direct repeats located between positions −125 and −51, and the other is an A-T-rich region 5′-TGAAAACGCAAACTATT-3′ between positions −51 and −35. The latter shows high similarity to the E. coli IclR-recognized box II 5′-TTAAATGGAAATTGTT-3′. Therefore, HpdR may repress hppD through similar modes as those of IclR (Yamamoto and Ishihama, 2002).
Tyrosine is the primary exogenous signal to trigger the transcription of hppD (Fig. 4), but it is not the signal sensed by HpdA and HpdR. This role is fulfilled by 4HPP. In the presence of 4HPP, HpdA is modulated to form a single low-order HpdA(4HPP)–DNA complex, and HpdR forms several high- or low-order HpdR(4HPP)–DNA complexes, depending on the molar ratio between 4HPP and HpdR. The 4HPP-responsive regulatory system consisting of HpdA activator and HpdR repressor presumably enables Streptomyces to properly express HppD in response to the concentration changes of 4HPP produced by tyrosine catabolism and other possible pathways in vivo. This regulatory system may permit a wide range of regulation, from fine-tuning to complete shutdown of hppD transcription.
hpdA transcription is repressed by HpdA and HpdR, and influenced by 4HPP
Previous studies showed that negative autoregulation is a common characteristic within the AsnC/Lrp family of regulatory proteins, and in general, this repression is independent of effectors (Brinkman et al., 2003; Yokoyama et al., 2006). However, our data demonstrated that exogenous tyrosine has distinct effects on the autoregulation of hpdA. In the absence of tyrosine, hpdA transcription is about 4-fold higher in the hpdA mutant than in the parental strain, whereas in the presence of tyrosine, hpdA transcription is about 20-fold higher in the hpdA mutant than in the parent (Fig. 6A). Although both HpdA and HpdA(4HPP) could bind to Pd-a, they formed different protein–DNA complexes with changed footprints (Figs 7B and 8).These might involve different conformations of the hpdA promoter and its adjacent regulatory region, and further influence the binding of RNAP or HpdR to the Pd-a regulatory region. Additionally, 4HPP induced changes in DNase I sensitivities of 10 sites in the HpdA footprint and 5 sites in the HpdR footprint, 12 of which are located just between the hpdA translation start codon and −35 region. Therefore, HpdA(4HPP) and HpdR(4HPP) assemblies are likely to have some transient cooperative binding and influence hpdA transcription in vivo. Taken together, the autoregulation of hpdA is affected by tyrosine, actually by its catabolic intermediate 4HPP, while HpdR and other regulators may play significant roles. Therefore, the mechanism of hpdA autoregulation is different from the known ligand-independent autoregulation of other AsnC/Lrp family members, such as E. coli asnC (Yokoyama et al., 2006).
HpdR is a specific repressor in the IclR family
Many IclR-family repressors are ligand-inducible transcription factors, the ligand accelerating disassociation of the regulator from the target DNA (Molina-Henares et al., 2006). However, HpdR represses hppD and hpdA only in the presence of tyrosine, HpdR(4HPP)–DNA complexes apparently being the in vivo functional forms for repression. The feature that HpdR can form multiple complexes in the absence or presence of ligand is novel among the IclR-type regulators. We do not understand the biological significance of the HpdR–Pd-a complexes formed in the absence of 4HPP, but HpdR is believed to sense changes of 4HPP concentration in vivo to form different HpdR(4HPP)–Pd-a complexes, thereby exerting a range of repression levels.
The mechanism of hppD transcriptional regulation elucidated here extends the limited knowledge of the expression of genes involved in amino acid catabolism. The substrate of HppD, 4HPP, acts as the specific ligand which triggers the transitions between protein–Pd-a complexes and protein(4HPP)–Pd-a complexes. HpdA and HpdR proteins sense the concentration changes of 4HPP and show different responses, with HpdA acting as a potential on–off regulator to hppD transcription, while HpdR acts as a possible fine-tuning regulator in response to the in vivo 4HPP. Streptomyces are therefore emerging as a further example of the diversity of tyrosine catabolism regulation in prokaryotes.
Bacterial strains, plasmids and growth conditions
Streptomyces coelicolor strain M145 and its derivatives used in this study are listed in Table 1. Usually, these strains were grown at 30°C. Yeast extract-malt extract liquid medium (YEME), protoplast regeneration medium with yeast extract (R2YE) and agar MM for sporulation were prepared as described by Kieser et al. (2000). The agar-based medium Tyr-PM used for 4-hydroxyphenylpyruvate dioxygenase expression and brown pigment production was prepared by modification of MM: asparagine was omitted, FeSO4 was added at a higher concentration (40 mg l−1), K2HPO4 and MgSO4.7H2O were used at 0.5 g l−1 and 0.2 g l−1, and l-tyrosine was used at 5 mM as the carbon and nitrogen source. E. coli DH5α BL21(DE3)pLysS and E. coli ET12567 (Table 1) were usually grown at 37°C in Luria–Bertani (LB) medium.
Table 1. Bacterial strains described in this study.
Plasmids pBluescript KS+ (Stratagene) and pIJ2925 (Janssen and Bibb, 1993) were used for routine cloning and subcloning experiments in E. coli. The Streptomyces/E. coli shuttle plasmid vector pKC1132 containing the apramycin-resistance gene [aac(3)IV] was used to construct the gene disruption mutants via homologous recombination. pSET152, which can integrate into the Streptomyces chromosome by site-specific recombination at the bacteriophage ΦC31 attachment site (attB) (Kuhstoss and Rao, 1991; Bierman et al., 1992), was used to introduce single copies of genes into the S. coelicolor chromosome. pET23b (Pharmacia Biotech) containing an ampicillin-resistance gene was used for gene overexpression in E. coli. The cosmid SCE19A containing the hppD (SCO2927), hpdA (SCO2928) and hpdR (SCO2935) and their flanking sequences was used for cloning these genes.
For plasmid selection in E. coli, LB medium was supplemented with ampicillin (100 μg ml−1), kanamycin (100 μg ml−1) or apramycin (50 μg ml−1). In Streptomyces, apramycin or kanamycin was added at 100 μg ml−1 to R2YE, and 10 μg ml−1 to MM and YEME.
DNA manipulation and sequence analysis
Isolation of plasmid and chromosome DNA, procedures for Southern blotting, and transformation of Streptomyces protoplasts were carried out as described previously (Tan et al., 1998). All other molecular techniques, including transformation of competent E. coli cells, and blunt ending of DNA by Klenow fragment and T4 polymerase, were performed as described by Sambrook et al. (1989). The hpdR probe used in Southern blot analysis was labelled non-radioactively (Digoxigenin-11-UTP kit; Roche). For database searches, the programs tfasta (Lipman and Pearson, 1985) and blast (Altschul et al., 1997) were employed. Sequence analysis and homologue comparisons were performed with the programs of NCBI blast-X. KEGG was used to search for likely tyrosine catabolic genes in different genomes.
Primers and polymerase chain reaction (PCR) conditions
All the oligonucleotides used in this study were listed in Table 2. The PCRs were performed using Taq DNA polymerase or Pfu high-fidelity DNA polymerase (Shanghai Sangon Biological Engineering and Technology Services Corporation, China): an initial denaturation at 95°C for 4 min was followed by 30 cycles of amplification (95°C for 1 min, 65°C for 1 min, and 72°C for 2 min) and additional 7 min at 72°C. In some cases, the annealing temperature and time, as well as the elongation time, were changed to accommodate different DNA templates and oligonucleotides.
Table 2. Primers used in this work.
Restriction sites (NdeI and XhoI) are underlined. The sequences are presented from 5′-end to 3 ′-end.
Primers for the identification of gene disruptants hppdDDM and hpdADM
Primers for S1 nuclease mapping of hrdB, hppD, hpdA and hpdR
Primer used for DNA sequencing of hppD and hpdA
Primers for gene overexpression in E. coli
Primers for EMSA
Construction of disruption mutants and complementation
To construct hppD disruption mutant, two fragments flanking hppD were prepared by restriction digestion of SCE19A. In the final construction, a 552 bp BstXI–NcoI fragment in hppD was replaced by the kanamycin-resistance gene (aphII), and then the ΔhppD::aphII DNA fragment (nearly 5.5 kb) blunted with Mung bean nuclease was inserted into the EcoRV site of pKC1132 to generate pKC1132::ΔhppD::aphII, which was passed through ET12567 to generate non-methylated DNA, and subsequently introduced into M145 protoplasts. The kanamycin-resistant and apramycin-sensitive strains were selected, and a disruption mutant generated by double cross-over was confirmed by PCR using oligonucletides DP1 and DP2 as primers (Table 2). The disruption mutant was designated hppDdisruption mutant (hppDDM) (Fig. 3A).
The construction of the hpdA and hpdR disruption mutants was similar to that of hppDDM. A 500 bp SacII–HincII fragment internal to hpdA was replaced by aphII to generate ΔhpdA::aphII, and the resulting recombinant plasmid pKC1132::ΔhpdA::aphII was introduced into M145 to give an hpdA disruption mutant (hpdADM) (Fig. 3A). In this mutant strain, the hpdA promoter region and its closely downstream sequence containing 56 bp coding region is still retained, which permitted the transcription of the hpdA promoter to be studied in appropriate conditions. The mutant was further confirmed by PCR using oligonucletides AP1 and AP2 as primers (Table 2). Likewise, a 304 bp ApaI fragment containing part of the promoter region and the 5′-end coding region corresponding to the α–helix–turn–α–helix motif of hpdR was replaced by aphII to generate ΔhpdR::aphII, and the resulting recombinant plasmid pKC1132::ΔhpdR::aphII was introduced into M145. A KanRAprs-putative hpdR disruption mutant (hpdRDM) (Fig. 3A) was confirmed by Southern blot hybridization.
For complementation of hppDDM, a 1.3 kb ApaI–SacII DNA fragment carrying the promoter and coding region of hppD was isolated from cosmid SCE19A, filled in with T4 polymerase, and inserted into the EcoRV site of pSET152 to generate the recombinant plasmid, which was then introduced into hppDDM to obtain the complemented strain, hppDC. For hpdADM complementation, a 950 bp DNA fragment from cosmid SCE19A containing hpdA was digested with HincII, and subcloned into the EcoRV site of pSET152. The resulting plasmid was then introduced into hpdADM to get the complemented strain, hpdAC. For hpdRDM complementation, a 1.4 kb KpnI fragment containing the complete hpdR was purified, filled in by T4 polymerase, and subcloned into the EcoRV site of pSET152 to generate pSET152::hpdR, which was then introduced into hpdRDM to give rise to the complemented strain hpdRC.
The spores of M145, hppDDM, hpdADM and hpdRDM grown on MM were collected in sterilized water, diluted and counted by haemocytometer. Suspensions containing 109 spores were spread on cellophane on 30 ml agar Tyr-PM plates, and the mycelium was collected and weighed at 24 h intervals. All the assays were repeated three times.
Brown pigment assay
After the mycelia were collected and assayed for the biomass, the medium, containing brown pigment, was frozen at −70°C for 4 h and then thawed at room temperature. A total of 0.2 ml solution was used for the absorbance assay at 600 nm.
S1 nuclease mapping
Total RNAs were isolated from cultures on cellophane-overlaid MM or Tyr-PM plates as described by Liu et al. (2005). Quality and quantity of RNAs were examined by UV spectroscopy and checked by agarose gel electrophoresis.
Hybridization probes for S1 nuclease mapping were prepared by PCR from S. coelicolor M145 chromosomal DNA using a 5′end-labelled primer located in the open reading frame and an unlabelled upstream primer (Table 2). For each S1 nuclease reaction, 40 μg of RNA was used for hybridization with a [γ-32P] end-labelled probe in NaTCA buffer at 45°C for 15 h, following denaturation at 65°C for 15 min. S1 nuclease (Promega, USA) digestions were performed as described by Kieser et al. (2000), and the reaction products were separated on 6% sequencing gels, along with sequencing reactions prepared using the fmol DNA cycle sequencing kit (Promega, USA) and radiolabelled oligonucleotide DP4. The hpdA and hpdR probes were generated by PCR using the unlabelled oligonucleotides AP3 or RP1 and radiolabelled AP4 or RP2 respectively. To detect the hpdA transcription in hpdA mutant strain as well as other strains simultaneously, the probe was designed to correspond to the 56 bp coding region and the upstream regulatory sequence of hpdA. Because of the quite low level of hpdA transcript, 160 μg of RNA was used for the S1 nuclease protection analysis and the denaturing temperature was 70°C for 15 min. The S1 nuclease digestions were as described above.
Overexpression and purification of HpdR-His6 and HpdA-His6 proteins
The hpdR-coding region was amplified by PCR using SCE19A cosmid DNA as template and oligonucleotides RP3 and RP4 as forward and reverse primers (Table 2). Similarly, the hpdA-coding region was obtained by PCR using AP5 and AP6 as forward and reverse primers. The amplified fragments digested with NdeI and XhoI were inserted into pET-23b to generate expression plasmids pET23b::hpdR and pET23b::hpdA respectively, which were further confirmed by DNA sequencing and then introduced into E. coli BL21 (DE3) for protein overexpression. E. coli BL21(DE3) harbouring pET23b::hpdR or pET23b::hpdA was grown at 37°C in 50 ml LB with 100 μg ml−1 ampicillin to an OD600 of 0.6. IPTG was then added to a final concentration of 0.1 mM, and the cultures were incubated for an additional 2 h at 30°C. The cells were harvested by centrifugation at 7700 g, 4°C for 10 min, washed twice with binding buffer [22.5 mM sodium borate, 110 mM boric acid, 300 mM NaCl, 5 mM imidazole, 5% glycerol (pH 8.9)] and then resuspended in 5 ml of the same buffer. The cell suspension was broken by sonication on ice. After Triton X-100 was added to the cell lysate (final concentration 1%), the mixture was gently shaken for 30 min at 4°C to get more soluble protein. After centrifugation (13 000 g for 10 min at 4°C), the supernatant was recovered, and HpdR-His6 or HpdA-His6 were separated from the whole-cell lysate using Ni-NTA agarose chromatography (Qiagen). After extensive washing with buffer [22.5 mM sodium borate, 110 mM boric acid, 300 mM NaCl, 30 mM imidazole, 5% glycerol (pH 8.9)], the HpdR-His6 or HpdA-His6 proteins were specifically eluted from the resin with 4 ml elution buffer [22.5 mM sodium borate, 110 mM boric acid, 300 mM NaCl, 250 mM imidazole, 5% glycerol (pH 8.9)] and concentrated to about 1 μg μl−1 by ultrafiltration (Mili-pore membrane, 3 kDa cut-off size) according to the protocol provided by the manufacturer. Protein purity was determined by Coomassie blue staining after SDS-PAGE on a 10% polyacrylamide gel. The purified protein was stored in 10% glycerol at −70°C.
Preparation of DNA probes
In total, 20 pmol of the downstream primer DP4, corresponding to the hppD promoter, was end-labelled with [γ-32P]-ATP using 10 U of T4 polynucleotide kinase (Promega, USA). The reaction was incubated for 30 min at 37°C, and then heated for 2 min at 90°C to eliminate the activity of T4 kinase. The labelled DP4 was used with the unlabelled DP3 to generate a 435 bp intergenic fragment between hppD and hpdA by PCR, using SCE19A DNA as template. The labelled DNA probe containing the hppD–hpdA intergenic region and 5′-end partial sequences of the two genes was purified using the PCR Purification Kit Uniq10 spin columns from Shangon (Shanghai Sangon Biological Engineering and Technology and Service), giving an approximate final concentration of 20 000 cpm μl−1.
Electrophoretic mobility shift assays
The EMSAs were performed according to the method of Yamazaki et al. (2000) and Elliot et al. (2001) with some modifications. 32P-labelled DNA probe (1000 cpm) was incubated individually with varying quantities of HpdR-His6 or HpdA-His6 at 25°C for 20 min in a buffer containing 1 μg of poly-(dI–dC) (Sigma) and 20 mM HEPES (pH 7.4), 2 mM dithiothreitol (DTT), 150 mM KCl, 5 mM MgCl2, 0.5 μg μl−1 calf BSA, 5% glycerol in a total volume of 20 μl. BSA was used as a stabilizer of HpdR-His6 and HpdA-His6. After incubation, protein-bound and free DNA were separated by electrophoresis on non-denaturing 5% polyacrylamide gels (mono/bis, 80:1) with a running buffer containing 40 mM Tris-HCl (pH 7.8), 20 mM boric acid, and 1 mM EDTA at 10 V cm−1 and 4°C. Gels were dried and exposed to Biomax radiographic film (Kodak).
For competitive gel shift assays, about 50-fold of unlabelled probe (specific competitor) was preincubated with HpdR-His6 or HpdA-His6for 20 min at 25°C, followed by the addition of labelled probe and incubation for another 20 min at 25°C. The resulting DNA–protein complexes were then subjected to electrophoresis and autoradiography as described above.
l-tyrosine and 4HPP, as the potential ligands of HpdA and HpdR, were separately dissolved in dimethylsulfoxide, to prepare solutions with concentrations of 400 mM.
DNase I footprinting
In order to characterize the binding sites of HpdA and HpdR on both strands of the hppD–hpdA intergenetic region, two probes were prepared separately by labelling the 5′-ends of the sense or antisense strands. For preparation of the labelled antisense strand, unlabelled DP3 primer and 5′end-labelled DP4 primer were used in the PCR amplification. For the sense strand, 5′end-labelled DP3 primer and unlabelled DP4 primer were used. The DNase I footprinting was performed essentially according to the procedures of Elliot et al. (2001). The reaction mixture (50 μl) contained 40 kcpm 32P-labelled DNA probe (100 nM), 0–3.2 mM of HpdA or 0–6.4 mM of HpdR protein, 20 mM HEPES (pH 7.4), 2 mM DTT, 100 mM KCl, 5 mM MgCl2, 0.5 μg μl−1 calf BSA and 5% glycerol. After incubation of the mixture for 20 min at 30°C, 1 U of DNase I (Promega, USA) was added to each reaction mixture, and then it was further incubated at 37°C for 1 min and 20 s. The reaction was stopped by the addition of 50 μl stop solution (3 M ammonium acetate, 0.25 M EDTA, 0.1 mg tRNA ml−1) and 100 μl phenol-chloroform. DNA fragments in the aqueous phase were precipitated with 3 vols ethanol, washed with 70% ethanol, dried and directly suspended in 10 μl of 90% formamide-loading gel buffer (10 mM Tris-HCl, pH 8.0, 20 mM EDTA, pH 8.0, 0.05% bromophenol blue, 0.05% xylene cyanol). Samples were then denatured at 95°C for 2 min and fractionated on a 6% polyacrylamide–urea gel. The sequence ladder was made using an fmol DNA cycle sequencing kit (Promega, USA) with the labelled primer DP4 or DP3 for the top or bottom strand. After electrophoresis, the gels were dried and exposed to Kodak X-ray film. Experiments were repeated at least three times, with similar results.
We are grateful to Professor Keith Chater (John Innes Centre, Norwich, UK) for providing E. coli ET12567, pKC1132 and pSET152, and for helpful discussion during this work and critical reading of the manuscript. We also thank Dr Helen Kieser (John Innes Centre, Norwich, UK) for providing the cosmid SCE19A, and Dr Brenda Leskiw (University of Alberta, Canada) for the gift of apramycin. This work was supported by grants from the National Natural Science Foundation of China (Grant No. 30430010), the Ministry of Science and Technology of China (Grant Nos. 2003CB114205 and 2006AA02Z206), and Chinese Academy of Sciences (Grant No. KSCX2-YW-N-027).