Tải bản đầy đủ - 0 (trang)
4 The nrf Operon Promoter: Variation on a Theme

4 The nrf Operon Promoter: Variation on a Theme

Tải bản đầy đủ - 0trang

18  Transcriptional Regulation by Nucleoid-Associated Proteins



427



a Anaerobic (Minimal medium).



-ve

IHF III



IHF II



−127



−100



IHF I



−74.5



−54



FNR



−41.5



FIS I



−15



pnrf

+1



+ve

b Anaerobic (Rich medium).



-ve

IHF III



IHF II



−127



−100



IHF I



−74.5



−54



FNR



−41.5



FIS I



−15



pnrf

+1



+ve

c Anaerobic plus NO2− (Minimal medium).

IHF I

IHF III



IHF II



−127



−100



NarL/P



−74.5



−54



-ve

FNR



FIS I



−41.5



−15



pnrf

+1



+ve

Fig.  18.3  Transcription regulation at the E. coli nrf operon promoter. (a) Minimal medium.

Transcription initiation is dependent on FNR. The binding of IHF to IHF I and FIS to FIS I inhibits FNR-activated transcription (−ve), whilst occupancy of the upstream IHF III site stimulates

transcription (+ve). In minimal medium FIS protein levels are low and so FIS I is not fully occupied (shown as dotted). (b) Rich medium. Higher levels of FIS protein lead to increased occupancy of FIS I, resulting in further repression of nrf operon expression. (c) Minimal medium with

nitrite. The binding of NarL (or its homologue, NarP) displaces IHF from IHF I, counteracting

IHF-mediated repression and enabling maximal FNR-dependent transcription



Early studies had shown that the nrf operon promoter is optimally active when

cells are grown in anaerobic conditions in the presence of nitrite or nitrate ions, and

is repressed in nutrient rich conditions (Page et al. 1990). As with the nir promoter,



428



D.F. Browning et al.



the nrf promoter is totally dependent on FNR, which binds to a target centered near

position −41 and functions as a simple activator (Tyson et al. 1994). Activation in

response to nitrate and nitrite ions is dependent on NarL or NarP binding to a site at

position −74.5 (in contrast to position −69.5 at the nir operon promoter). Studies

have shown that FNR alone is sufficient to activate the nrf promoter fully, but that

FNR-dependent activation is repressed by IHF binding to a DNA site located at position −54 (IHF site I, see Fig. 18.3). NarL or NarP binding at position −74.5 displaces

IHF from site I, thus permitting FNR-dependent activation (Browning et al. 2002b,

2006). Thus the primary role of IHF binding at site I is to confer codependence of

expression on two activators, and hence two physiological signals. This mechanism

is very similar to the one that operates at the nir promoter, save that, at the nrf promoter,

IHF binds downstream of the DNA site for NarL/NarP, and does not require support

from FIS. Interestingly, as with the nir promoter, there are other upstream sites for

IHF and the binding of IHF to one of them (IHF III) stimulates FNR-dependent

activation of the nrf promoter (Browning et al. 2006). Hence activation of the E. coli

nir and nrf promoters shares a common mechanism yet uses different arrangements

of the ‘actors’ and binding sites (compare Figs. 18.2 with 18.3).

The situation at the nrf promoter is complicated by the divergent acs promoter

that is located upstream and drives transcription from a startpoint located 280 base

pairs upstream of the nrf promoter startpoint. Expression of the acs promoter is

dependent on CRP and this activation is moderated by IHF binding at site II and

site III (Beatty et al. 2003; Browning et al. 2004b; Sclavi et al. 2007). Thus, the

primary roles of IHF bound at site II and site III appear to be concerned with the

acs promoter, and the stimulation of the nrf promoter may be serendipitous.

As mentioned above, the most striking difference between the E. coli nir and

nrf promoters is that nrf promoter expression is repressed in rich medium and that

FIS has been relieved of its duty as a ‘helper’ for IHF and given the role of ‘master’ repressor. To do this, the DNA site for FIS at the nrf promoter is located at

position -15, corresponding exactly to the location of the DNA site for FruR at the

nir promoter. Some of the supporting evidence for the effects of FIS at the nrf

promoter is shown in Fig. 18.4. Full repression in vivo requires the fis gene and



Fig. 18.4  (continued) concentrations of FIS and subjected to DNase I footprinting. The concentrations

of FIS used in each incubation were: lane 1, zero; lane 2, 0.06 mM; lane 3, 0.11 mM; lane 4, 0.22

mM; lane 5, 0.44 mM; lane 6, 0.88 mM. Gels were calibrated using Maxam-Gilbert ‘G + A’

sequencing reactions and relevant positions are indicated. The location of FIS I is shown by a box.

(c) In vitro transcription assays using plasmid pSR/ pnrf97, carrying the nrf promoter upstream of

a strong transcription terminator, as template. The RNA I transcript, encoded by sequences within

the pSR vector, serves as an internal control for transcript formation. The nrf and RNA I transcripts are indicated by arrows. The concentrations of FNR used in incubations were: lane 1, zero;

lanes 2–6, 1 mM. The concentrations of FIS were: lanes 1 and 2, zero; lane 3, 0.11 mM; lane 4,

0.22 mM; lane 5, 0.44 mM; lane 6, 0.88 mM



18  Transcriptional Regulation by Nucleoid-Associated Proteins



a



429



β-Galactosidase Units.



5000

4000

Minimal medium

3000



Rich medium



2000

(3.8)

1000



(5.4)



0



JCB3884



b



[FIS]



-



GA 1



JCB38841(fis )



2 3



4



5



6



−71



−30

−23



FIS I

−8

+1



c



[FIS]

[FNR]



- 1



2



3



4



5



6



RNA I



nrf

Fig. 18.4  FIS represses FNR-dependent transcription at the E. coli nrf promoter. (a) The panel

shows measured b-galactosidase activities of JCB3884 (fis+) and JCB38841 (fis) cells carrying

pRW50 (a lac expression vector plasmid), containing the pnrf53/D87 promoter fragment

(Browning et  al. 2005). Cells were grown anaerobically in either minimal or rich media and

b-galactosidase activities are expressed as nmol of ONPG hydrolyzed min−1 mg−1 dry cell mass.

The fold catabolite repression for each strain is indicated in brackets. (b) DNase I footprint analysis

of the nrf promoter. End-labelled pnrf53 AatII-HindIII fragment was incubated with increasing



430



D.F. Browning et al.



an intact DNA site for FIS. Binding of FIS to this site can be demonstrated in vitro

and this binding represses FNR-dependent activation of transcription in a simple

assay system.



18.5 FIS Protein as a Sensor of Nutrient Abundance in E. coli

The intracellular level of FIS rises dramatically in response to nutrient availability

and rapid growth (Ball et al. 1992). This is exploited at the nrf operon promoter to

shut off the expression of a nitrite reductase when its role as a scavenger for nitrite

in nutrient-poor conditions is not required (Wang and Gunsalus 2000). The location

of the DNA site for FIS, overlapping the −10 element, suggests that FIS acts as a

simple repressor, similar to many repressors, functioning by preventing RNA

polymerase binding (Rojo 2001). The E. coli genome contains many genes whose

function is coping with nutrient poor conditions and transcriptional repression in

response to nutrient excess is a common feature in the regulation of the promoters

of these genes. Whilst most of these effects may be due to CRP, we can expect that

FIS will be involved in repression in many cases. In addition to the nrf promoter,

three further examples, the ogt, acs and dps promoters have recently been reported.

These cases are particularly interesting since the mechanism of repression by FIS

at these promoters is more complex.

Figure 18.5 shows a sketch of the E. coli ogt regulatory region which contains a

single promoter that is activated by NarL alone without the participation of FNR (in

contrast to the nir and nrf promoters). NarL binds to two targets at position −45.5

and position −78.5. Results illustrated in Fig.  18.6 show that ogt expression is

induced by nitrate ions, that this induction is dependent on NarL, and that expression is repressed in rich medium. This repression is dependent on FIS that binds to

a single site located at position −82, which overlaps the upstream DNA site for

NarL. The footprinting experiment illustrated in Fig. 18.6c shows that FIS binding

displaces NarL, which is unsurprising since the two operator sites overlap. The

simplest explanation is that repression by FIS, which is triggered by the rise in FIS

levels that occur in rich medium, is due to the displacement of NarL. Thus, at the

ogt promoter, repression by FIS is due to the displacement of an activator, rather

than blocking RNA polymerase binding (as at the nrf promoter).

The ogt promoter is especially interesting as it controls the expression of an

enzyme that repairs methylated DNA (Potter et  al. 1987; Samson 1992). Most

NarL-dependent promoters are co-dependent on FNR and encode proteins involved

in nitrate and nitrite metabolism, and, to date, very few promoters have been found

to be activated by NarL without FNR (Constantinidou et al. 2006; Lin et al. 2007).

The fact that nitrate and nitrite metabolism generates harmful reactive nitrogen species (Taverna and Sedgwick 1996; Weiss 2006) prompts a simple explanation for

our findings. Some reactive nitrogen species interact with amino acid side chains

and that promotes DNA methylation. We suggest that the NarL-dependent induction



18  Transcriptional Regulation by Nucleoid-Associated Proteins



431



a Minimal Medium plus NO3−.

NarL I



NarL II



−78.5



−45.5



+ve

pogt

+1



b Rich Medium plus NO3−.

NarL I



FIS I



NarL II



+ve ?

pogt



−82



−45.5



+1



Fig. 18.5  Transcription regulation at the E. coli ogt promoter. (a) Minimal medium with added

nitrate. The binding of NarL at the NarL I and NarL II sites activates transcription (+ve) at the ogt

promoter. (b) Rich medium with added nitrate. FIS, binding to FIS I, displaces NarL, bound at

NarL I, repressing transcription. It is not yet known if NarL bound to NarL II is able to activate

transcription independently of NarL at NarL I



of ogt protects the cell’s DNA against the mutational consequences of such

reactions, and that such protection is deemed unnecessary in nutrient rich conditions, where DNA damage arising from the side products of high metabolic flux

poses a far more substantial threat.

A second example of FIS-dependent repression due to displacement of an essential activator is found at the E. coli acs promoter, which controls expression of

acetyl-coenzyme A synthetase. The acs promoter is divergent from the nrf promoter

and the intergenic region contains a complicated array of binding sites for FIS and

IHF that affect both promoters (Fig.  18.7). Expression from the acs promoter is

dependent on activation by tandem-bound CRP at position −69 and −122 (Beatty

et  al. 2003) and is repressed by both FIS and IHF, which function by different

mechanisms. FIS binding at position −59 (FIS III) and position −98 (FIS II)

represses the acs promoter by displacing CRP from both sites. In contrast, IHF,

which binds to three upstream sites, at positions −153, −180 and −226 represses the

acs promoter without displacing CRP (Browning et al. 2004b; Sclavi et al. 2007).

The combined effects of FIS and IHF ensure that acs expression peaks at the transition from exponential to stationary phase. Recall that acetyl-coenzyme A synthetase enables the cell to use acetate and is transiently expressed as cells enter



432



D.F. Browning et al.



β-Galactosidase Units.



b



β-Galactosidase Units.



a



600

500



- O2

- O2 + NO3- O2 + NO2-



400

300

200

100

0



1200



JCB387



JCB3883 (narL)



Minimal medium



Rich medium



1000



- O2

- O2 + NO3-



800

600

400

200

0



JCB3871 (fis)



JCB387



c

[NarL]



- - - - -



[FIS]



-



GA 1 2



3 4 5 6



JCB387



-



7



JCB3871 (fis)



8 9 10



−90



NarL I



FIS I



−102



−80

−69



NarL II



−59



−39

−30



Fig. 18.6  FIS represses the E. coli ogt promoter in rich growth media. (a) The panel shows measured

b-galactosidase activities of JCB387(narL+) and JCB3883 (narL) cells carrying pRW50, containing

the pogt100 promoter fragment. Cells were grown anaerobically in minimal medium and nitrate and

nitrite were added to a final concentration of 20 and 2.5 mM, respectively, where indicated.



433



-ve

FIS II



a

pnrf



FIS I



IHF I



−265



−226



b



IHF II



−180



IHF III



−153



CRP II



-ve



−122.5 −98



FIS III

CRP I



-ve



−69.5 −59



pacs

+1



Zone of competition.

[Active CRP]



[FIS]



[acs transcription]



Transcription



Growth



OD



Time

Fig. 18.7  Transcription regulation at the E. coli acs promoter. (a) Regulation of pacs. The binding

of CRP to the CRP I and CRP II sites activates transcription at the acs promoter. In exponential

phase when cellular FIS concentrations are high, FIS binding to FIS II and FIS III displaces CRP

and represses transcription (−ve). The binding of IHF to three upstream sites also represses CRPdependent transcription (−ve). (b) Modulation of CRP-dependent acs transcription by FIS. The

figure shows how acs transcription alters due to bacterial growth (OD). The relative concentrations of active CRP and FIS are shown



Fig. 18.6  (continued) (b) The panel shows measured b-galactosidase activities of JCB387(fis+) and

JCB3871(fis) cells carrying pRW50, containing the pogt100 promoter fragment. Cells were grown

anaerobically in either minimal or rich media and nitrate was added to a final concentration of 20 mM

where indicated. b-galactosidase activities are expressed as nmol of ONPG hydrolyzed min−1 mg−1 dry

cell mass and each activity is the average of three independent determinations. (c) DNase I footprint

analysis of the ogt promoter. End-labelled pogt100 AatII-HindIII fragment was incubated with

increasing concentrations of FIS in combination with NarL and subjected to DNase I footprinting. The

concentrations of FIS used in incubations were: lanes 1 and 6, zero; lanes 2 and 7, 0.45 mM; lanes 3

and 8, 0.89 mM; lanes 4 and 9, 1.8 mM; lanes 5 and 10, 3.8 mM. The concentrations of NarL were: lanes

1–5, zero; lanes 6–10, 3.2 mM. Gels were calibrated using Maxam-Gilbert ‘G + A’ sequencing reactions

and relevant positions are indicated. The locations of NarL and FIS binding sites are indicated by

vertical boxes



434



D.F. Browning et al.



stationary phase and take up and metabolise acetate that had been excreted during

exponential growth (Wolfe 2005).



18.6 Regulation of the Escherichia coli dps Promoter

Dps is a nucleoid-associated protein that is absent in rapidly growing E. coli but

accumulates as growth slows and cells enter stationary phase (Almiron et  al.

1992). In non-growing cells, Dps becomes the most abundant nucleoid-associated

protein and this is thought to be a key factor in maintaining the stationary phase

folded chromosome. The expression of Dps depends on a single promoter located

just upstream of the dps gene and accumulation of Dps requires the stationary

phase s factor, s38. The observation that the dps promoter can be served by RNA

polymerase containing either s38 or the major s factor, s70, raises the puzzle of

what prevents dps from being expressed in rapidly growing cells (Altuvia et al.

1994). Recent studies have shown that, just as at the nrf and ogt promoters, FIS

is the key factor in repression, but that it acts by an unusual mechanism in which

FIS jams RNA polymerase containing s70 but not s38 at the dps promoter

(Grainger et al. 2008). Thus, in rapidly growing cells, the dps promoter is silenced

by a ternary repression complex containing RNA polymerase with s70, FIS and

promoter DNA. Remarkably, FIS has little or no effect on the activity of RNA

polymerase containing s38, and hence FIS can discriminate between two different

forms of RNA polymerase. This provides an efficient switch for ensuring that the

dps promoter is silent, when FIS levels are high, but activated as FIS levels fall

and s38 levels rise.

As well as being repressed by FIS, the dps promoter is also regulated by

H-NS, IHF and OxyR (Fig. 18.8). Like FIS, H-NS acts as a repressor that discriminates between RNA polymerase containing s70 and s38 (Grainger et  al.

2008). H-NS displaces RNA polymerase containing s70 from the dps promoter,

whilst not interfering with RNA polymerase containing s38. Thus, together with

FIS, H-NS confers s factor dependence on dps expression. In contrast, a third

nucleoid-associated protein, IHF, binds upstream of the core dps promoter elements and functions as an activator during s38-dependent transcription in stationary phase (Altuvia et  al. 1994; Ohniwa et  al. 2006). Finally a second

activator, OxyR, which is triggered by oxidative stress, also binds upstream, and

is responsible for transient induction of dps during oxidative stress in rapidly

growing cells (Altuvia et al. 1994; Ohniwa et al. 2006). In these circumstances,

the repression by FIS and H-NS must be overcome, but the mechanism for this

is unclear at present. Panels a–c of Fig. 18.8 illustrate the molecular complexes

responsible for silencing the dps promoter in rapidly growing cells, and activating it in response to oxidative stress in growing cells, or as cells progress to

stationary phase (Schnetz 2008).



18  Transcriptional Regulation by Nucleoid-Associated Proteins



a



435



-ve



Es38



Es70

+1

−98.5



−51



IHF



b



OxyR



−26



+2.5



FIS H-NS



+ve



Es70



Es38

+1

−98.5



−51



IHF



OxyR



c



−26 +2.5



FIS H-NS



+ve

Oxidative stress

Es70

+1

−98.5



IHF



−51



OxyR



−26 +2.5



FIS H-NS



Fig. 18.8  Selective regulation of the E. coli dps promoter. (a) Selective repression by FIS. During

rapid growth, transcription from the dps promoter is repressed by FIS, which binds to the promoter

in unison with RNA polymerase containing s70 and shuts down the promoter, blocking access by

RNA polymerase containing s38. (b) Selective repression by H-NS. Binding of H-NS to the dps

promoter blocks the binding of RNA polymerase with s70 but permits binding of RNA polymerase

with s38. Transcription by RNA polymerase with s38 (but not with s70) can be stimulated by IHF.

(c) Activation by OxyR. In response to oxidative stress, transcription from the dps promoter by

RNA polymerase containing s70 is enhanced by OxyR, which overcomes the negative effects of

FIS and H-NS by a yet unknown mechanism (Schnetz 2008)



436



D.F. Browning et al.



18.7 Genome-Wide Effects of FIS and IHF

The above examples underscore the versatility of FIS and IHF in moderating regulation at bacterial promoters. To gain insight into the global roles of these factors,

chromatin immunoprecipitation has been exploited to find their binding locations

across the whole E. coli K-12 chromosome (Grainger and Busby 2008). To do this,

the sequence composition of DNA fragments, which had been immunoprecipitated

with antisera directed against either FIS or IHF, was analysed, using high density

microarrays (Grainger et al. 2006). Figure 18.9 shows a typical set of results illustrating the distribution of FIS and IHF. Each scan shows the enrichment (y-axis) for

DNA sequences at particular loci (x-axis) in the immunoprecipitated DNA samples.

As expected, both proteins bind at many targets. For FIS and IHF, 224 and 135

targets respectively were identified, and these include most of the previously identified targets (63 and 55 targets, respectively, listed in EcoCyc: Karp et  al. 2007).

Surprisingly, ~60% of the targets for both FIS and IHF are in intergenic regulatory

regions. Since these regions cover less than 10% of the total genome, it is clear that

FIS and IHF binding must be highly focused. This is unlike the situation with

eucaryotic histones that bind at equal densities to both coding and non-coding targets. It is clear that, if FIS and IHF are involved in chromosome compaction, they

must orchestrate this primarily by binding at intergenic regulatory regions. Analysis

of the target locations revealed 54 regulatory regions where FIS and IHF both interact, including the E. coli nir, nrf-acs and dps operon regulatory regions.

Many authors consider the E. coli nucleoid-associated proteins to be different

from transcription factors. However, the above analysis with FIS and IHF questions

this distinction, since their binding profile resembles that of some transcription factors. Figure 18.10 shows the binding profiles of two of the best characterized E. coli

transcription factors, CRP and FNR. As for FIS and IHF, these profiles were

derived from analyzing immunoprecipitated DNA using antisera directed against

purified CRP (Grainger et al. 2005) or against a FLAG tag that was attached to FNR

(Grainger et al. 2007). For both transcription factors, 60–70 clear targets were identified, which correspond to both previously identified and previously unidentified

targets. At most of these targets, factor binding can either up-regulate or downregulate transcription initiation. However, at some targets, it was impossible to

measure any detectable consequence on the activity of promoter activity, suggesting that there may be bona fide targets for transcription factor binding that serve no

purpose, at least directly, in the modulation of gene expression (Grainger et  al.

2007; Hollands et al. 2007). Note that a similar conclusion was found with the E.

coli RutR factor (Shimada et  al. 2008). Strikingly, the binding profile for CRP

shows a strong background that appears to be due to its binding to many thousands

of weak sites across the E. coli chromosome (Grainger et al. 2005). Interestingly,

these sites were predicted by bioinformatic studies (Robison et al. 1998) and are

consistent with the observed ‘non-specific’ binding of CRP observed in electromobility shift assays (Kolb et al. 1983). Clearly, it is unlikely that transcription is regulated from these sites, and, since CRP is known to bend DNA sharply upon binding



18  Transcriptional Regulation by Nucleoid-Associated Proteins



437



(Schultz et al. 1991), we suggest that at many sites, it behaves more like a nucleoidassociated protein than a transcription factor.

Taken together, the available data argue that there is no intrinsic difference

between nucleoid-associated proteins and transcription factors. As for all biological

measurements, they are best considered in the light of evolution. Thus, it is possible

that the need for bacteria to compact their genomic DNA came before the need to

regulate transcription and that nucleoid proteins evolved as one of the first DNA

binding proteins. As time passed, it is easy to believe that the genes encoding these

proteins were duplicated, that sequence specificity evolved, and that cells that could

regulate the transcription of certain loci were advantaged. Presumably, regulatory

modules were grafted onto some of these factors, many of which then lost their

function as nucleoid organizers. Prompted by data, such as that in Figs. 18.9 and

18.10, we suggest that E. coli, and probably many other bacteria, contain DNA

binding proteins with a continuum of binding specificities and functions, and that

the distinction between nucleoid-associated proteins and transcription factors is

artificial.



18.8 Perspectives

E. coli is found in many places, and most of these, such as the guts of animals and

aquatic environments, are subject to rapid and frequent fluctuations. As for most

bacteria, survival depends on the selective expression of gene products to cope with

the environment, and thus, it is no surprise that E. coli has evolved sophisticated

systems to control transcription. This is most apparent in the high proportion of its

gene products that are dedicated to regulating transcription initiation and in the

complexity of even the simplest promoter. Thus, the nir operon promoter is regulated by four transcription factors: by FNR, by NarL (and its homologue, NarP) and

by FruR and their activity is modulated by three nucleoid-associated proteins, IHF,

FIS and H-NS. Although we can assume that different combinations of these factors are used in different conditions, most studies have been performed in ‘simple’

laboratory conditions, and the relative importance of the different factors in ‘real’

environments is still poorly understood. A quick glance at the Ecocyc database will

convince anyone that the simplistic models for promoter regulation that appear in

the textbooks are misleading. These models are mostly based on a small number of

paradigm promoters (such as the lac promoter) and were established early in the

history of this subject area. We now know that many, if not most, promoters are

very complicated, with multiple factors interacting and other factors such as small

ligands, the local chromosome landscape and DNA topology intervening. The challenge now is for us to put all the facts together, to produce integrated models, and,

most important, to understand how systems are evolving.

Perhaps the most striking feature of transcriptional regulation in E. coli is its

complexity, which, surely, has arisen from its evolution. Thus a simple DNA binding protein such as FIS, which has only 98 amino acids plays a myriad of roles



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

4 The nrf Operon Promoter: Variation on a Theme

Tải bản đầy đủ ngay(0 tr)

×