Tải bản đầy đủ - 0 (trang)
3 H-NS: Histone-Like Nucleoid Structuring Protein

3 H-NS: Histone-Like Nucleoid Structuring Protein

Tải bản đầy đủ - 0trang

8  Nucleoid-Associated Proteins: Structural Properties



153



Fig.  8.2  3-dimensional structures of the N- and C-terminal domains of H-NS. (a) The NMRstructure of the dimerization domain (residues 1–46) of H-NS is shown (PDB ID: 1NI8; (Bloch

et al. 2003)). The a-helices for both monomers are indicated (a1, a2 or a3). (b) An alternative

NMR structure determined for the H-NS N-terminal domain (residues 1–57) from S. typhimurium

is shown (PDB ID: 1LR1; (Esposito et al. 2002)). (c) The three-dimensional NMR structure of the

C-terminal domain of H-NS (residues 91–137) is shown (PDB ID: 1HNR; (Shindo et al. 1995)).

The conserved residues important for DNA binding are indicated in red. At the bottom an amino

acid sequence alignment and accession numbers of the C-terminal residues from several members

of the H-NS family is shown. The amino acid residues within loop 2 are indicated in yellow as in

the three-dimensional structure above. Identical amino acids are coloured in red. (d) 3D structure

model of an H-NS dimer modeled from separate structural domains as in (b) and (c)



154



Ü. Pul and R. Wagner



higher oligomers (Bloch et  al. 2003). The structure of the N-terminal H-NS1–46

obtained from 2D NMR analysis is shown in Fig.  8.2a. Dimerization occurs by

antiparallel coiled-coil association of the a3-helices through hydrophobic residues. The a1 and a2 helices are also involved and further stabilize the dimeric

structure (Bloch et  al. 2003). Interestingly, according to the three-dimensional

structure of the H-NS1–57 dimer from Salmonella typhimurium determined in a

separate study (Esposito et al. 2002) the a3-helices are oriented in parallel, even

though the overall structures of the monomers are almost identical in both studies

(Fig. 8.2b). In both structures the two short helices a1 and a2 are folded back to

the a3-helices, but in the antiparallel orientation the a2 helix of one H-NS monomer is associated with the a3-helix of the other protomer and not with that of the

same molecule, as in the case of the parallel orientation. The observed differences

might be due to the additional amino acid residues in the H-NS1–57 variant compared to the H-NS1–46 with the antiparallel orientation of the long helices.

Although H-NS can clearly be divided into two structural domains, the assignment

of independent functions for the individual domains (with respect to dimerization/

oligomerization and DNA binding) has turned out to be difficult. Generally, mutations

in the N-terminal domain can affect DNA binding, and mutations in the C-terminal

domain affect the dimerization/oligomerization of H-NS. Since dimerization/oligomerization is an important prerequisite for DNA-binding, interpretations from

binding studies with H-NS mutants are notoriously difficult.

Amino acid sequences potentially involved in oligomerisation of H-NS in vivo

were analysed in a study using chimeric constructs in which wild-type or mutant

H-NS fragments had been fused to a l phage repressor (Stella et  al. 2005). The

analyses revealed that deletion of a proline residue in the C-terminal domain

(DP115) or substitution of the same amino acid by alanine (P115A) impeded

dimerization, and to a greater extent, tetramerization of the chimeric H-NS constructs in vivo (see Fig. 8.2c). In a separate investigation it was shown by gel retardation and footprint analyses that the mutant H-NS variant DP115 also displays

reduced affinity for curved DNA fragments (Spurio et al. 1997). It should be noted

that binding to curved DNA constitutes actually the decisive specificity for H-NSDNA recognition and that it can also discriminate between short synthetic bent

DNA fragments (Yamada et al. 1990). The H-NS DP115 mutant has lost this preferential binding, suggesting that oligomerization and DNA binding specificity may

be directly related. Similar results were also reported for amino acid substitution

mutants, where residues within the C-terminal domain, including the proline residue 115, had been replaced (Badaut et  al. 2002). Again, these mutants can bind

non-specifically to DNA but they have lost the preference for curved DNA.

Together the data indicate that the C-terminal domain contains the DNA motif

responsible for specific binding, but to some extent is also involved in proteinprotein interaction. Moreover, it has been reported that the flexible linker separating

the N- and C-terminal domains may also participate in oligomerization of H-NS.

This conclusion was derived from analyses, which show that the interactions

leading to the formation of H-NS tetramers are thermodynamically different compared to those leading to H-NS dimers (Ceschini et  al. 2000). Thus, it has been



8  Nucleoid-Associated Proteins: Structural Properties



155



suggested that after dimerization of H-NS through hydrophobic interactions of the

N-terminal domain, further oligomerization of the H-NS dimers involves amino

acids within the linker region and also partly amino acids from the C-terminal

domain. Note that there may be a difference between the solution state and the

DNA-bound state; related to that the current evidence suggests that H-NS is bound

in dimeric form when it bridges DNA (Dame et al. 2006). Oligomerization of H-NS

is important for DNA binding and the DNA bending capacity, while dimerization

is a prerequisite for the formation of tetramers or higher oligomers. Therefore it is

difficult to assign oligomerization or DNA binding to one structural domain, even

though the necessary individual functions may be catalyzed by individual domains

of the protein (Chapter 13).

The part of the H-NS structure responsible for DNA interaction does not correspond to any conserved amino acid folds known for DNA interaction. It is clearly

folded, however, in a separate domain, which has a rather large spatial distance to

the N terminus of roughly 4.5 nm as revealed by FRET analysis between Trp 108

in the centre of the C-terminal binding domain and the N-terminus of the protein

(Schröder et al. 2001). Details of the structure comprising the last 47 C-terminal

amino residues have been solved by NMR and are presented in Fig. 8.2c (Shindo

et  al. 1995, 1999). According to this analysis the C-terminal domain consists of

four small loop structures separated by two short b-strands and two helical segments. A series of unstructured amino acids, forming loop 2 harbours the amino

acid residues, which, according to mutagenesis studies, are most likely involved in

direct DNA interaction (see below). This loop is flanked by a small a-helix and a

310-helix close to the C-terminus. As also shown in Fig. 8.2c the C-terminal DNA

binding domain is highly conserved in all H-NS-like proteins, suggesting similar

structures of the DNA binding regions. Within that structure four amino acid residues (W108, G110, G112 and P115) are notably important. They are conserved in

nearly all H-NS-like proteins currently collected in the public databases. These

highly conserved amino acid residues are all located within loop 2. Based on those

sequence alignments the following consensus motif for an H-NS-type DNA binding domain has been proposed: TWTG-GR-P (Dorman et al. 1999). Previous analyses using amino acid substitutions within the C-terminal domain of H-NS support

the direct involvement of loop 2 in DNA binding (Ueguchi et al. 1996). A G112D

mutant of H-NS, for instance, is strongly impeded in DNA binding and also does

not show a comparable effect on DNA topology as the wild-type protein (Ueguchi

et  al. 1996; Pul et  al. 2005, 2007). Despite this information the exact molecular

details of H-NS-DNA interaction are still unknown. Such an analysis is primarily

hampered by the fact that for long no clearly defined DNA consensus binding

sequence was known. Even the question of whether H-NS binds to the major or

minor groove of DNA has not been answered with confidence. A recent study

(Sette et al., 2009) reports evidence for binding of the C-terminal fragment of H-NS

to the minor groove of a conserved DNA element. The only fact which is undisputed about H-NS binding, concerns its ability to interact selectively with curved

or flexible DNA segments. Such DNA properties generally correlate with a high

content of AT base pairs, which, when clustered in helical phase, result in curved DNA.



156



Ü. Pul and R. Wagner



It is not surprising, therefore, that primary H-NS binding sites have been mapped

in AT-rich sequence regions. A recent proposal of such a primary ‘high affinity’

H-NS binding site (tCGATAAATT) in principle reflects this high AT content

(Bouffartigues et al. 2007; Lang et al. 2007). Often the initial binding step involves

such AT-rich sites and further interactions are followed by cooperative polymerization of H-NS molecules along the DNA, which ultimately results in protein coating

of rather large DNA regions. This protein-coating either hinders binding of RNA

polymerase to promoter DNA or it may interfere with the interaction of other regulatory proteins. The coating mechanism as a consequence of cooperative H-NS

binding is supported by the fact that generally large regions of the DNA are protected in DNase I footprint experiments (Pul et al. 2005; Bouffartigues et al. 2007).

The fact that the variation of the distance and angular orientation of a synthetic

curved H-NS binding module does not show significant effects on transcription

inhibition at different promoters is also consistent with the coating mechanism

(Pul et al. 2008).

The capacity of H-NS for lateral oligomerization along the DNA and the presence of two DNA binding domains per H-NS dimer allows for the simultaneous

interaction with two DNA strands, which gives rise to DNA bridging (Fig. 8.2d).

Such a bridging activity of H-NS and other H-NS-like proteins has been visualized in AFM studies (Dame et  al. 2000, 2005). DNA bridging may certainly

contribute to DNA compaction of the bacterial nucleoid. More importantly, however, bridging of two adjacent DNA strands also brings about transcription regulation. As shown for the rrnB P1 promoter, H-NS-mediated bridging leads to

RNA polymerase trapping in an open promoter complex, which is inadequate for

transcription (Schröder and Wagner 2000; Dame et al. 2002). A trapping mechanism through DNA bridging has also been described for the hdeAB promoter. In

this special case sigma-factor selectivity is regulated because H-NS-mediated

bridging occurs only with Es70 but not with Es38 (Shin et al. 2005). Interestingly,

sigma-factor selectivity is also conferred by other NAPs, such as IHF and LRP at

the osmY promoter (Colland et  al. 2000) or FIS at the dps promoter (Grainger

et al. 2008).

Like most of the other NAP members H-NS often acts in concert either with

other members of the NAP family or with other DNA binding proteins, thereby

displaying synergistic or antagonistic effects. This general property of concerted

function significantly adds to the regulatory variability of the NAP family of regulators. Well studied examples are the seven ribosomal RNA operons in E. coli, where

FIS-dependent activation of transcription is counteracted by H-NS, while transcription inhibition by LRP is supported (Afflerbach et al. 1999; Pul et al. 2005). Other

examples of antagonism include the virulence gene transcriptional regulator VirB

in Shigella flexneri (Turner and Dorman 2007), transcription of the cspA mRNA,

which is under the antagonistic control of FIS and H-NS (Brandi et al. 1999) or hns

expression itself, which is under H-NS autoregulation antagonized by FIS (Falconi

et al. 1996; Stoebel et al. 2008). In line with the view of a concerted function of

different NAP members are the binding profiles of FIS, H-NS and IHF, which have

been analyzed on a genome-wide scale. This analysis reveals considerable overlap



8  Nucleoid-Associated Proteins: Structural Properties



157



in binding of the three proteins and also shows that sites of FIS and H-NS binding

are often close to sites of RNA polymerase association (Grainger et al. 2006).

H-NS does not only make contact with itself, forming homodimers or higher

oligomers but is also known to undergo heteromeric contacts with a variety of other

proteins. Notable examples are the heterodimeric forms with StpA or other H-NS

homologs, which share the dimerization domain. However, heteromeric contacts

are also formed with more distantly related proteins, such as FliG, Hfq, Hha or the

phage T7 gene product 5.5 (Kajitani and Ishihama 1991; Liu and Richardson 1993;

Williams et al. 1996; Donato and Kawula 1998; Nieto et al. 2002). Such combinations of regulatory proteins may furnish the cell with a battery of new regulatory

tools, providing novel specificity and tuning mechanisms. This may be of special

importance for adaptation reactions under rapidly changing environmental conditions and is consistent with the known involvement of H-NS in modulation of

bacterial gene expression in response to temperature and osmolarity (Falconi et al.

1998; Nieto et al. 2002; Dorman 2004).

Ever since their early identification NAPs have been recognized as important

elements for the transfer of genetic elements. This is partly reflected in their names,

which in case of FIS and IHF are acronyms of their functions characterized in the

first place. FIS (factor for inversion stimulation) is involved in phage recombination

reactions and IHF (integration host factor) also participates in phage l integration.

For H-NS it was recently shown that it binds to genes that were acquired by horizontal gene transfer, where it functions as gene silencer. Binding to the foreign

genes depends on a higher AT content relative to the resident genome, consistent

with the preference of H-NS to bind to AT-rich DNA sequences. This has led to

suggest another interesting new function for H-NS, namely transcriptional silencing

of AT-rich foreign DNA obtained through horizontal gene transfer (Lucchini et al.

2006; Dorman 2007; Navarre et al. 2007).

In summary, the modular composition and structural properties of H-NS, unrelated to conventional transcription factors, provide this protein with a variety of

unusual functions and a wide range of distinct properties, making it a highly versatile bacterial regulator to efficiently adapt the bacterial gene expression pattern to

altered conditions.



8.4 LRP: Leucine Responsive Regulatory Protein

Leucine responsive regulatory proteins constitute a family of widespread transcriptional regulators with global functions in cellular adaptation to environmental changes. E. coli LRP is a basic protein with a pI of 9.3 and exists in solution

as homodimer composed of two identical subunits with a molecular weight of

18.8 kDa each. As a transcriptional activator or repressor LRP is mainly involved

in the regulation of amino acid metabolism-related genes. Hence, LRP has been

assigned to the family of so-called ‘feast and famine’ regulatory proteins

(FFRPs) found in eubacteria and archaea, which contribute to the adaptation of



158



Ü. Pul and R. Wagner



the bacterial gene expression in response to the nutritional quality (Yokoyama

et al. 2006).

The cellular concentration of LRP strongly depends on the availability of nutrients and is inversely proportional to the bacterial growth rate (Landgraf et al. 1996).

DNA microarray analysis has shown that at least 10% of E. coli genes are affected

by LRP; most of these genes are involved in the adaptation of E. coli to nutrient

limitations (Tani et al. 2002).

The oligomerization state of many FFRPs can be regulated through allosteric

effector molecules. In the case of E. coli LRP the amino acid leucine is the allosteric effector. This is a unique feature among the NAP family members for which

otherwise no allosteric effector molecules are known. Through allosteric modulation of the quarternary structure leucine influences the association state of E. coli

LRP, which results in altered DNA binding affinity (Chen and Calvo 2002).The

resulting regulatory effects for individual promoters differ, however. Hence,

among the LRP-regulated genes leucine can either enhance, reduce or have no

effect on binding of LRP (Calvo and Matthews 1994). The association reactions

of LRP are complex and it is assumed that at nanomolar concentrations LRP dimers are the predominant species. Determination of the LRP-DNA stoichiometry

indicated that LRP can bind as a dimer but higher aggregates are frequently found

in DNA complexes (Chen et al. 2005). At micromolar concentrations LRP exists

as octamers and hexadecamers, which in the presence of leucine dissociate preferentially to octamers (Chen et al. 2001; Chen and Calvo 2002). The exact mechanism behind the differential effect of leucine on LRP binding is not fully

understood, however.

The three-dimensional structure of E. coli LRP has been solved by X-ray crystallography (de los Rios and Perona 2007). Figure 8.3a shows the modular structure of an LRP monomer, which is characterized by two structural domains also

reflecting the different functions of LRP. The N-terminal domain contains the

helix-turn-helix DNA binding motif, consisting of three a-helices. According to

mutational analysis the a3-helix of this motif was found to be responsible for

binding to the major groove of the DNA (Platko and Calvo 1993). The C-terminal

domain of LRP contains two a-helices flanking four antiparallel b-strands, forming a ab-sandwich. The resulting babbab-structure constitutes a structural motif

found in many LRP-like regulatory proteins or enzymes involved in amino acid

metabolism and has been termed RAM domain (regulation of amino acid metabolism) (Ettema et al. 2002). Within the RAM domain the b3–b4 loop contains the

binding site for the effector amino acid leucine. The N- and C-terminal domains

are linked by a flexible linker, containing a b-strand (depicted as ‘crossover

b-strand’ in Fig. 8.3a according to de los Rios and Perona (2007)), which upon

dimerization forms a two-stranded antiparallel b-sheet with the corresponding

b-strand of the second monomer at the homodimer centre (Fig. 8.3b). In addition,

dimerization is attributed to hydrophobic interactions between the two b-sheets of

the RAM domains (Chapter 15).

As shown in Fig. 8.3c four LRP dimers associate to form an octamer, arranged

as a disc, in a similar way as has been described for LrpA from Pyrococcus furi-



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

3 H-NS: Histone-Like Nucleoid Structuring Protein

Tải bản đầy đủ ngay(0 tr)

×