INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In Escherichia coli complex regulatory mechanisms control the synthesis of anaerobic respiratory enzymes. The master switch (anaerobic induction) is mediated by transcriptional regulator protein Fnr (reviewed in reference 15). In most cases the Fnr protein binds to a site (consensus, TTGAT-N₄-ATCAA) centered approximately 41.5 bp upstream from the transcription start point. This is true for the operons encoding many anaerobic enzymes including nitrate reductase-A (narGHJI), fumarate reductase (frdABCD), formate dehydrogenase-N (fdnGHI), and cytoplasmic (nirBDC) and periplasmic (nrfABCDEFG) nitrite reductases (8, 15).

The Fnr protein is homologous to the well-characterized transcriptional regulator Crp (cyclic AMP receptor protein). For naturally occurring Crp-activated promoters the Crp-binding site is located at various distances upstream of the promoter. Furthermore, it has been demonstrated that the Crp protein activates the transcription of synthetic promoters when the Crp-binding site is located at various distances upstream of the transcription start site (14). Studies indicate that the location of the Crp-binding site determines the mechanism of transcription activation with respect to interactions between Crp and RNA polymerase (reviewed in reference 3). Consequently, Crp-activated promoters have been divided into different classes determined by the location of the Crp-binding site.

Despite the similarity between Fnr and Crp, early observations that all naturally occurring Fnr-dependent promoters of E. coli had the Fnr-binding site close to position 41.5 led to speculation that Fnr was limited to only one mechanism of transcription activation. However, deletion and mutational analyses of the Fnr-activated napF (aeg-46.5; encoding periplasmic nitrate reductase) operon control region indicated that the Fnr-binding site is at position 64.5. By analogy with Crp-dependent promoters this suggested that it is a naturally occurring example of a second class of Fnr-dependent promoters (9). This possibility is supported by the demonstration that an engineered Fnr protein recognizes the Crp-binding site of the lac promoter at position 61.5 to activate lac expression (24). Furthermore, Fnr activates the transcription of synthetic promoters with the Fnr-binding site at position 61.5 (or further upstream) and the mechanism of activation is distinct from that when the Fnr-binding site is at position 41.5 (1, 29-31). However, it should be noted that the precise interactions between Fnr and RNA polymerase at the different classes of promoters are not identical to those between Crp and RNA polymerase (2, 29).

Respiratory gene expression is also regulated in response to nitrate and nitrite, the preferred anaerobic electron acceptors. Nitrate and nitrite control is mediated by homologous DNA-binding response regulators (NarL and NarP), which communicate with homologous sensor proteins (NarX and NarQ) (reviewed in references 8 and 26). The NarL and NarP proteins recognize heptamer binding sites that resemble the consensus TACYYMT (where Y = C or T and M = A or C) (6, 12, 20, 27). Both NarL and NarP bind to heptamers organized as inverted repeats with 2-bp spacing. In addition NarL, but not NarP, can also bind to heptamers in other arrangements (10). There are several distinct patterns of operon expression known, including induction by nitrate (e.g., for the narG and fdnG operons), repression by nitrate (e.g., for the frdA operon), induction by nitrate or nitrite (e.g., for the nirB operon), and induction by nitrite and inhibition by nitrate (e.g., for the nrfA operon). Despite all of this complexity some generalizations can be made with regard to operon control region architecture. Activation by NarL or NarP occurs when one or both of these proteins bind upstream of an Fnr-binding site centered at 41.5. In other cases, NarL-dependent repression is mediated by the binding of NarL to sites downstream of the Fnr-binding site (reviewed in references 8 and 26). The napF operon control region is the one exception to these generalizations about NarL- and NarP-dependent regulation.

napF operon expression is induced by nitrate or nitrite (5, 9, 22). This activation is solely dependent on the NarP protein, unlike all other operons studied, for which activation is dependent on NarL only or on either NarL or NarP (8, 26). The NarP-binding site of the napF control region is centered at position 44.5 (downstream of the Fnr-binding site), a more promoter-proximal location than those of other NarP- or NarL-activated promoters. The NarL protein is also able to bind to this 44.5 site but does not activate transcription. However, by competing with NarP for the 44.5 binding site, NarL antagonizes NarP-dependent activation. The inability of NarL to activate napF operon expression led us to hypothesize that NarL is deficient in the mechanism by which NarP activates transcription from the 44.5 binding site (9). By contrast, both NarL and NarP are competent to activate transcription activation from the more upstream binding sites of other operon control regions.

In the present study we sought to confirm the location of the Fnr-binding site in the napF operon control region and to demonstrate that Fnr may activate transcription by an atypical mechanism. We also reproduced the regulation of the napF promoter by Fnr, NarP, and NarL in vitro and confirmed and extended our previous conclusions about napF operon regulation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and plasmids. For routine manipulations, plasmids were propagated in strain DH5 [80d (lacZ)M15 (argF-lac)U169 endA1 recA1 hsdR17 deoR thi-1]. Strain VJS5577 (Table 1) is a derivative of VJS676 [(argF-lac)U169] with a single-copy (napF-lacZ) operon fusion (includes the napF control region from positions 146 to +305 [9]) and fnr-271::Tn10-, narL249::-Sp-, and narP253::Tn10d(Cm)-null alleles that were introduced by P1 kc-mediated transduction. Plasmids containing fnr⁺ (and the mutant derivatives) were a gift from Steve Busby (University of Birmingham, Birmingham, United Kingdom) and have been described previously (30).

Culture media and conditions. Defined, complex, and indicator media for routine genetic manipulations were used as described previously (11). When necessary media were routinely supplemented with ampicillin (200 µg/ml), tetracycline (25 µg/ml), chloramphenicol (20 µg/ml), or spectinomycin (30 µg/ml). Cultures for -galactosidase assay (Table 1) were grown in 3-(N-morpholino)-propanesulfonic acid (MOPS)-buffered minimal medium (pH 8.0) with glucose as the sole carbon source (25). This medium was supplemented with tetracycline (25 µg/ml) and ampicillin (60 µg/ml); 10% (vol/vol) Luria-Bertani broth was added to stimulate growth. Culture densities were measured with a Klett-Summerson photoelectric colorimeter (Klett Manufacturing Co., New York, N.Y.) with a no. 66 (red) filter.

-Galactosidase assay. -Galactosidase activities were determined at room temperature (approximately 21°C) in permeabilized cells exactly as described previously (19). Activities are expressed in arbitrary units (21). Each culture was assayed in duplicate, and reported values were averaged from three independent cultures, the standard errors of which were not more than 15%.

Protein purification. Fnr (D154A) protein (mutant Fnr protein with an aspartate-to-alanine substitution at position 154) was purified as described previously (18, 32) except that a heparin agarose chromatography step followed the Q-Sepharose anion exchange chromatography to remove a contaminating nuclease activity. The purified Fnr (D154A) protein was 10% active in site-specific DNA binding, as determined previously (32). The MBP-NarL and MBP-NarP proteins were purified exactly as described previously (9). All stated protein concentrations refer to the monomeric concentration.

DNase I footprinting. DNase I footprinting assays of the napF control region with Fnr (D154A) were done as described previously (10) except that acetyl phosphate was not included in the reaction mixtures. Reaction mixtures were incubated at 37°C for 30 min to reach equilibrium prior to the addition of DNase I. The napF control region fragment was generated from plasmid pVJS1523 by PCR and labeled with ³²P on the bottom strand as described previously (9).

Construction of a napF control region template plasmid for in vitro transcription assays. The napF control region from 202 to +91 (with respect to the transcription start site) was amplified from plasmid pVJS1515 (9) by PCR. The primers were 5'-CCTGCAAGCTTAGTGTTAAATTCTAATGAGAGAG-3' and 5'-CCGAGGATCCGCATCAATCTTCACATTGACCTTC-3'. These primers had unannealed tails (boldface and italics) to generate HindIII and BamHI sites. The product fragment was digested with HindIII and BamHI and cloned into plasmid pUC19-spf' (13) to generate template plasmid pVJS2111. This places the napF promoter approximately 200 bp upstream of the spf transcription terminator.

In vitro transcription assays. Supercoiled pVJS2111 plasmid DNA was purified with a plasmid purification kit (Qiagen Inc., Chatsworth, Calif.). Phosphorylation of the MBP-NarL and MBP-NarP proteins was essential for activity in the transcription reactions (data not shown). Therefore, for all of the experiments in this study, the MBP-NarL and MBP-NarP proteins were phosphorylated by incubating them in 50 mM Tris-Cl (pH 7.6)-10 mM MgCl₂-50 mM acetyl phosphate for 60 min at 37°C. The phosphorylation reaction mixture was then spun through a Sephadex G-25 column to remove acetyl phosphate, which interfered with the transcription reactions. The transcription reaction buffer contained 40 mM Tris-Cl (pH 7.9), 10 mM MgCl₂, 50 mM KCl, 0.1 mM dithiothreitol, 0.1 mg of bovine serum albumin per ml, and nucleotide triphosphates (0.5 mM ATP, 0.5 mM CTP, 0.5 mM GTP, 0.05 mM UTP, and 5 µCi of [-³²P]UTP at 3,000 Ci/mmol). A 20 nM concentration of supercoiled pVJS2111 plasmid DNA was preincubated in reaction buffer for 20 min with Fnr (D154A), phosphorylated MBP-NarP, or phosphorylated MBP-NarL proteins in a total volume of 20 µl. Reactions were initiated with the addition of 50 nM E. coli RNA polymerase holoenzyme (E⁷⁰; Epicentre Technologies, Madison, Wis.). After a further 5-min incubation at 37°C the reactions were terminated by the addition of 10 µl of formamide loading dye. Then, 3 to 6 µl of each reaction mixture was loaded onto a 6% (wt/vol) denaturing polyacrylamide gel. Reactions products were separated by electrophoresis, and the gels were dried and exposed to X-ray film for 4 to 16 h.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The Fnr protein binds to the 64.5 site of the napF control region. Deletion and mutational analyses had suggested that the Fnr-binding site of the napF control region is centered at position 64.5 (Fig. 1) (9), significantly further upstream than the location of that for other Fnr-dependent promoters (approximately 41.5). Therefore, we wished to confirm the location of the Fnr-binding site by DNase I footprint analysis.

View larger version (7K): [in this window] [in a new window]   FIG. 1.   Architecture of the napF operon control region. The in vivo anaerobic transcription initiation site (+1) (5) (data not shown) is indicated by the arrow, and the 10 hexamer is underlined. The Fnr-binding site is indicated by inverted hatched arrows, with residues identical to those of the consensus (TTGAT-N4-ATCAA) in boldface. The NarP- and NarL-binding site (inverted heptamer sequences) is indicated by numbered arrows. Residues identical to those of the consensus heptamer sequence (5'-TACYYMT-3') are in boldface. Each heptamer is denoted by the position of the central nucleotide with respect to the transcription initiation site.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   DNase I footprint analysis of the napF operon control region. The control region fragment was labeled on the bottom (template) strand. Each lane is labeled with the concentration of Fnr (D154A) used in the reaction mixture (in micromolar monomers). The number of base pairs from the transcription start site is indicated on the left (a G+A sequencing reaction of the napF control region fragment was used as a size marker [not shown]). The location of the Fnr-binding site is indicated by the inverted arrows. Asterisks mark the positions of sites hypersensitive to DNase I cleavage in the reactions with Fnr (D154A).

The napF operon has an atypical Fnr-dependent promoter. By analogy with Crp-dependent promoters (3) and synthetic Fnr-dependent promoters (29, 30), we hypothesized that the mechanism by which napF expression is activated by Fnr is distinct from the mechanism involved in the activation of most other naturally occurring Fnr-dependent promoters. Positive control mutations have identified amino acid side chains of the Fnr protein likely to be involved in transcription activation but not in DNA binding (1, 28-30). One of these substitutions (G85A [Gly at position 85 changed to Ala]) interferes with transcription activation when Fnr is bound at position 41.5 but enhances activation when the Fnr-binding site is at position 61.5 (or further upstream). Another Fnr substitution (S73F) strongly reduces activation by Fnr from 61.5 binding sites (or further upstream) but has a more subtle effect when the Fnr-binding site is at position 41.5 (1, 30). The different effects of these positive-control mutations suggest different mechanisms of Fnr-dependent transcription activation for the two classes of promoter. To determine the class of Fnr-dependent promoters to which the napF promoter belongs, we investigated the effect of these Fnr-positive control mutations on (napF-lacZ) expression in vivo (Table 1).

View this table: [in this window] [in a new window]   TABLE 1.   Effects of Fnr-positive control mutations on (napF-lacZ) expression

In vitro transcription from the napF promoter. Deletion, mutational, and footprint analyses had indicated that napF operon expression in vivo is induced weakly by either Fnr or NarP alone and strongly by Fnr and NarP together. In contrast, the NarL protein antagonizes napF operon expression by competing with NarP for a common DNA-binding site (9). We attempted in vitro transcription assays to further investigate this complex pattern of in vivo regulation.

View larger version (87K): [in this window] [in a new window]   FIG. 3.   In vitro transcription from the napF promoter. Each multiple-round transcription assay mixture contained 20 nM supercoiled pVJS2111, 50 nM RNA polymerase holoenzyme (E70), and either no further additions (), 0.5 µM Fnr (D154A) monomers (F), 6 µM phosphorylated MBP-NarP monomers (P), or 12 µM phosphorylated MBP-NarL monomers (L). The napF and constitutive plasmid ori transcripts are labeled.

Both MBP-NarP and Fnr (D154A) are required for maximum napF transcription in vitro. To investigate the regulation of napF transcription in more detail, we did a series of titration experiments with each of the regulatory proteins. The Fnr (D154A) protein alone (at any concentration) was unable to activate napF transcription above the level observed with RNA polymerase alone, but instead caused a slight repression (data not shown). In a control experiment, the activity of the Fnr (D154A) protein was confirmed by its ability to activate transcription of the Fnr-dependent dmsA promoter in vitro (data not shown). In the presence of MBP-NarP (7.5 µM), the addition of 0.5 µM Fnr (D154A) increased the amount of napF transcript by approximately two- to threefold. Fnr (D154A) was less stimulatory at higher or lower concentrations (data not shown). Therefore, we used 0.5 µM Fnr (D154A) for all of the subsequent reactions.

View larger version (73K): [in this window] [in a new window]   View larger version (76K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   FIG. 4.   Both MBP-NarP and Fnr (D154A) are required for maximum napF transcription in vitro. (A) Effect of phosphorylated MBP-NarP concentration on napF transcription in vitro. Reaction mixtures contained either no Fnr (D154A) (Fnr) or 0.5 µM Fnr (D154A) monomers (+Fnr). The concentration of phosphorylated MBP-NarP either was zero (lanes 1 and 10) or ranged from approximately 0.11 to 14 µM monomers in twofold increments (lanes 2 to 9 and 11 to 18). (B) Effect of phosphorylated MBP-NarL concentration on napF transcription in vitro. Reaction mixtures contained either no Fnr (D154A) (Fnr) or 0.5 µM Fnr (D154A) monomers (+Fnr). The concentration of phosphorylated MBP-NarL either was zero (lanes 1 and 10) or ranged from 0.1 to 12.8 µM monomers in twofold increments (lanes 2 to 9 and 11 to 18). (C) Representation of the data from the experiments presented in panels A and B, which were analyzed as described in Materials and Methods. For each titration, the amount of napF transcript with E70 alone [titrations in the absence of Fnr (D154A)] or with E70 plus Fnr (D154A) [titrations in the presence of Fnr (D154A)] was arbitrarily assigned the value of 1.0. , MBP-NarL without Fnr (D154A); , MBP-NarL plus Fnr (D154A); , MBP-NarP without Fnr (D154A); , MBP-NarP plus Fnr (D154A).

MBP-NarL antagonizes activation of the napF promoter by Fnr (D154A) and MBP-NarP in vitro. Mutational and DNase I footprint analysis had indicated that the NarP and NarL proteins are each capable of binding to the 44.5 site of the napF control region (Fig. 1) (9). This competitive binding presumably leads to antagonization of NarP-dependent activation by the NarL protein, since only NarP and Fnr together cause maximum napF expression. This prediction was tested in vitro by investigating the effect of increasing amounts of MBP-NarL on napF transcription in the presence of Fnr (D154A) and MBP-NarP.

View larger version (8K): [in this window] [in a new window]   FIG. 5.   Antagonization of MBP-NarP- and Fnr (D154A)-dependent napF transcription by MBP-NarL. Each in vitro transcription reaction mixture contained 50 nM RNA polymerase, 0.5 µM Fnr (D154A), and approximately 5 µM phosphorylated MBP-NarP. Increasing concentrations of phosphorylated MBP-NarL were added, and the change in napF transcription was measured as described in Materials and Methods. The amount of napF transcript with RNA polymerase, Fnr (D154A), and phosphorylated MBP-NarP only (no phosphorylated MBP-NarL) was arbitrarily assigned the value of 1.0.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The napF promoter is one of several regulated by Fnr in response to anaerobiosis and by NarL or NarP in response to nitrate and nitrite. Many of these promoters have similarities in their control region architecture. The Fnr-binding site is typically centered near position 41.5, and activation by nitrate or nitrite is dependent on NarP- and/or NarL-binding sites further upstream (reviewed in reference 8). However, the napF control region has a very different architecture, with an Fnr-binding site centered at position 64.5 and activation by NarP being mediated by a binding site downstream, at position 44.5 (Fig. 1). In this study we confirmed the location of the Fnr-binding site at position 64.5 and supported the idea that the mechanism of Fnr-dependent activation of napF expression is distinct from that of promoters with the Fnr-binding site at position 41.5. The regulation of the napF promoter was also studied in vitro, wherein we reproduced and extended previous in vivo observations. The results support our hypothesis (9) that the Fnr and NarP proteins act together to cause maximum napF operon expression whereas the Fnr and NarL proteins do not.

Location of the Fnr-binding site in the napF control region. The homologous Fnr and Crp proteins are each able to activate synthetic promoters with the Crp- or Fnr-binding sites either near position 41.5 or near 61.5 (or further upstream) (14, 30). Activation of these different classes of promoters occurs via different contacts between RNA polymerase and the regulatory protein, although the precise contacts that Fnr and Crp make with RNA polymerase are not identical (2, 29). There are numerous examples of naturally occurring Crp-dependent promoters of each class (e.g., the lac and gal promoters). However, the napF promoter is a rare example of a naturally occurring Fnr-dependent promoter that is in a class distinct from those of the more common promoters with the Fnr-binding site at position 41.5. The fact that most Fnr-dependent promoters have the latter architecture may be due to the fact that many of them are also activated by the NarL protein (e.g., the narG, fdnG, nrfA, and nirB promoters). It is possible that NarL is competent for transcription activation only when bound upstream of the 41.5 region. The napF operon provides the only known example of a promoter activated by NarP but not NarL. In this control region, the common NarP- and NarL-binding site is at position 44.5, in the region normally occupied by Fnr. In this context, only NarP is competent to activate transcription from this 44.5 binding site. The positioning of this NarP- and NarL-binding site results in the Fnr-binding site being further upstream, making this an unusual Fnr-dependent promoter.

Do Fnr and NarP activate napF transcription synergistically? One possible explanation for the observation that both Fnr and NarP are required for maximum napF transcription is that the Fnr and NarP proteins may activate transcription synergistically. By analogy with other studies of synergistic activation, Fnr and NarP may make independent contacts with RNA polymerase (4, 16, 17). Each of these contacts may have a stimulatory effect on transcription, resulting in greater transcriptional activation by Fnr and NarP together than the sum of their individual effects. An alternative hypothesis is that the binding of NarP simply allows Fnr to be an efficient activator from the 64.5 binding site, as discussed above, perhaps by bending the DNA. In this case the NarP protein may not make a stimulatory contact with RNA polymerase.

Antagonism of NarP-dependent activation by NarL. The NarL protein does not significantly activate napF operon expression in the presence of Fnr, whereas the NarP protein does. The explanation for this is dependent on the mechanism by which NarP activates napF transcription, as discussed above. NarL may be unable to make the required contact with RNA polymerase when bound to the common NarP- and NarL-binding site centered at position 44.5. Alternatively, the binding of NarL to the 44.5 site may not have the required effect on the DNA structure to enable Fnr to be an efficient activator from the 64.5 binding site. Support for the latter idea comes from the observation that DNase I footprints of the napF promoter with MBP-NarL or MBP-NarP are dissimilar. Both proteins protect the 44.5 region from DNase I attack, but they cause different sites to become hypersensitive to cleavage (9). One interpretation of this is that the two proteins may bend the DNA in different ways (9). Regardless of the mechanism, the NarL protein is unable to activate the napF promoter when bound to the 44.5 site and can be thought of as a natural positive-control mutant (relative to NarP function) in this context.

Significance of the napF operon transcription start site. It appeared that the in vitro napF transcription start site in the presence of either RNA polymerase alone or Fnr (D154A) was approximately three nucleotides upstream relative to the site in the presence of MBP-NarL or MBP-NarP (Fig. 3 and 4). We were unable to confirm this observation in vivo since a primer extension product for the napF transcript from a narL narP double-null strain was undetectable. However, we did confirm the start site of the Fnr-plus-NarP-induced promoter (+1 in Fig. 1; data not shown). In a previous study two major transcription start sites, separated by 3 or 4 nucleotides, were detected for the napF promoter in vivo (5). The upstream start site (3 in Fig. 1) was detected from aerobically grown cells, and the downstream start site (+1 in Fig. 1) was detected from anaerobically grown cells. This led to the suggestion that the upstream start site was Fnr independent whereas the downstream start site was Fnr dependent. However, it is possible that the shift in the transcript start site was caused by NarL or NarP rather than by Fnr, since only the anaerobic culture may have contained nitrate or since the level of expression of the narL and narP genes during aerobic growth with nitrate is significantly lower than that during anaerobic growth with nitrate (7). It is interesting to note that the upstream transcription start site (position 3 in Fig. 1) would place the Fnr-binding site at the permissive 61.5 position (30). Furthermore, there are alternative 10 sequences at the napF promoter, overlapping by 3 bp (TAATATCTT; Fig. 1). Understanding napF operon transcription start site selection awaits further experimental tests.