Tải bản đầy đủ - 0 (trang)
9 GalR Dimer–Dimer Interface Involved in DNA Looping

9 GalR Dimer–Dimer Interface Involved in DNA Looping

Tải bản đầy đủ - 0trang

405



17  Role of HU in Regulation of gal Promoters



a



OE



b



Bender



OI



OE



c



Adaptor



OI



Combination



OE



OI



Fig. 17.7  Model of HU role in gal DNA looping. (a) HU acting as a bender to bring about GalR-GalR

tetramerization, resulting in looping repression of gal promoters. (b) HU acting as an adapter to achieve

GalR tetramerization. (c) HU acting both as a bender and adapter during loop formation. The symbols

are described in the legend of Fig. 17.1



dimers is too low to overcome the energy required for DNA bending and possibly

twisting needed for DNA loop formation. Although DNA supercoiling overcomes

part of the energetic barrier, HU can further help by binding at an architecturally

critical position in the DNA to bend the DNA and sustain the loop. HU is known to

be a DNA bending protein (Drlica and Rouvière-Yaniv 1987; Pettijohn 1988;

Bonnefoy and Rouvière-Yaniv 1991). In (b), GalR dimers cannot form tetramers by

themselves thus necessitating the need for a mediator to bring them together. HU

may play that role. However, the demonstration of GalR tetramer formation by

ultracentrifugation studies (Roy et al. 2005), the ability of mutant GalR dimers to

interact with each other (Geanacopoulos et  al. 1999), and the demonstration of a

specific HU-DNA interaction in the GalR loop rule out a simple adapter model

(Aki et al. 1996; Aki and Adhya 1997). In (c), HU plays the role of both DNA bender

and GalR adapter. Given the limited size of the DNA segment (113 bp), a loop

containing three proteins bound to DNA at spatially separated sites and simultaneously

contacting each other would be energetically and/or sterically prohibitive (Majumdar

et al. 1987). We will distinguish between models (a) and (c) below.



17.12 Assembly of the gal Loop: Role of HU

Models (a) and (c) above are distinguished by the prediction that in the latter, there

would an interaction between GalR and HU. In fact, studies of HU binding to DNA

revealed that there is a tripartite co-operativity between GalR and HU in binding to

gal DNA; binding of HU to gal DNA is absolutely dependent upon binding of GalR

to both gal operators, and HU binding in turn results in increasing the strength of

GalR binding (Aki and Adhya 1997). The binding of GalR to the operators in the

absence of HU is non-cooperative (Brenowitz et al. 1990). The basis of a synergistic binding of GalR and HU could be an interaction between GalR and HU.

Co-immuno-precipitation of GalR and HU, indeed, shows a specific interaction

between the two proteins both in crude extracts and with purified proteins that was



406



D.E.A. Lewis et al.



Fig. 17.8  Model of HUaa showing the locations of Serine17, Lysine18, and Threonine19 (Kar

and Adhya 2001)



not evident between an HU homolog, IHF (Oberto et al. 1994), and GalR (Kar and

Adhya 2001). The amino acid residues in HU that define the contact(s) between HU

and GalR have been identified by characterizing mutants of HUa (S17P, K18A,

and T19D) that show inefficient co-immunoprecipitation with GalR and do not help

gal repression efficiently both in  vitro and in  vivo (Fig.  17.8) (Kar and Adhya

2001). Electrophoretic mobility-shift assays to test the DNA-binding properties of

the HU mutants show that the mutants are proficient in DNA-binding.

The crystal structure of the B. stearothermophilus HU homodimer (HBs),

which shares a 59% sequence homology with E. coli HU heterodimer (Drlica and

Rouvière-Yaniv 1987), was used to model the E. coli structure (Kar and Adhya

2001). The amino acid residues S17, K18, and T19 in E. coli HUa are located

contiguously in a small turn between the first and second alpha helices (Fig. 17.7).

This region lies on the opposite face of the DNA-binding surface, in a prominently

accessible portion of HU. All three amino acids have solvent-exposed side chains

and are likely candidates to contact GalR. No information is available related to

the corresponding residues in GalR.

Given the facts discussed above, a pathway for HU-mediated DNA looping in

gal emerges (Fig.  17.9) (Kar and Adhya 2001). (a) GalR dimers recruit HU.

(b) GalR−HU complexes bind to the operators. (Step b could precede step a.) (c) The



407



17  Role of HU in Regulation of gal Promoters

Fig. 17.9  A pathway of GalR and HU to DNA looping

in the presence of supercoiled DNA. Steps (a)–(e) are

described in the text. The symbols are described in the

legend of Fig. 17.1



Pathway to DNA looping

a



b



+



DNA



c



d



e



operator-bound GalR−HU complexes transiently interact to form tetramers via GalR

dimer–dimer interaction and generate a DNA loop that has a distortion around

the apical region (hbs). (d) Spontaneous dissociation of HU from GalR increases the

local concentration of HU. (e) Because HU has a stronger affinity to distorted DNA

(~10 nM) than to GalR (~10 mM), the dissociated HU preferentially binds to the

architecturally critical but transiently bent/distorted DNA, stabilizing the loop. By this

scenario, HU binds and stabilizes the loop by binding to an architecturally critical

position in DNA. This pathway is consistent with the energetics of the various interactions

reported above. The concentration of HU in the cell is not high enough to interact

with a rare and transient target in gal DNA. The recruitment of HU by GalR as a

piggyback in gal looping to increase local concentration for DNA binding may be



408



D.E.A. Lewis et al.



DNA loop



Fig.  17.10  Electron microscopy image of DNA looping with LacI and lac operators on a linear

DNA (reproduced with permission from Cold Spring Harbor Laboratory Press (Mandal et al. 1990))



an example of a common theme in the formation of similar higher order structures

involving both architectural and sequence-specific DNA binding proteins in other

DNA transaction reactions (Charlier et al. 1980; Kabsch et al. 1982; Borowiec et al.

1987; Lewis et al. 1996; Muller et al. 1996; Dickerson 1998).



17.13 Electron Microscopic Evidence of a DNA Loop

When electron microscopy is used to study loop formation on a gal DNA fragment

containing either the OEG – OIL, OEL – OIG or OEL – OIL genotype in the presence of

GalR (in the absence of HU), LacIadi or LacI, DNA loops of the expected size are

observed only with the OEL – OIL DNA and wild type LacI tetramer protein but not

with LacIadi, a mutant dimer protein (Fig. 17.10) (Mandal et al. 1990). Looping is

never observed with the OEL – OIL DNA in the presence of GalR. This result was

expected because LacI tetramer represses transcription in vitro but LacIadi and GalR

do not (Choy and Adhya 1992; Choy et al. 1995a).



17.14 Evidence for an Anti-parallel gal DNA Loop: Modeling

We mentioned above that two GalR dimers form a V-shaped tetramer stabilized by

stacking interactions. DNA loop formation by GalR tetramer as predicted by

genetic data mentioned above can align the operators in four possible arrangements

(Fig. 17.11a and b), two antiparallel loops (A1 and A2) and two parallel loops (P1

and P2) (Geanacopoulos et  al. 2001; Virnik et  al. 2003; Semsey et  al. 2004; Lia

et al. 2008). Antiparallel loops mean that the 5¢ end of one operator is arranged in

the opposite direction of the 5¢ end of the other operator (A1 and A2). In parallel

loops, the 5¢ end of one operator is arranged in the same direction as the 5¢ end of

the other operator (P1 and P2).



409



17  Role of HU in Regulation of gal Promoters



a



+1

P1



5’

OE



60.5



A1



b



hbs

+6.5



P2



5



3’

OI



+53.5



A2



hbs



P1



hbs



P2



hbs



3’



3’

OE

OI



OE



5’



3’



c



Hetero-GalR P2

P1



OI



5’



3’







+



RNA1



1



OE



2



5’

5’



hbs







+



+



P2

P1



P2

P1



P2

P1



RNA1



RNA1



RNA1



3



4



OI



OI



OE



5



6







+



7



8



Fig. 17.11  Models of DNA loops formed by GalR. (a) Regulatory region of the gal regulon as

described above in Fig. 17.2. The monomer of GalR is red and the other monomer is white. The

red and black arrows indicate the direction from 5¢ to 3¢ of the operators, OE and OI, respectively.

(b) Possible loops of GalR-DNA interactions. A1 and A2 represent two forms of “antiparallel”

loops and P1 and P2 represent two forms of “parallel” loops. (c) RNAs made from galP1 and P2

with GalR heterodimer and different operators to generate the four possible loops: A1 (lanes 1–2),

A2 (lanes 3–4), P1 (lanes 5–6) and P2 (lanes 7–8) (reproduced with permission from Cold Spring

Harbor Laboratory Press (Semsey et al. 2004))



These four possibilities were analyzed by stereochemical model building using an

empirical elastic model of DNA (Geanacopoulos et al. 2001). First, the GalR dimer

structure was modeled after the x-ray structure of a GalR homolog, PurR (Schumacher

et al. 1994). The model of the HU-DNA complex was based on the crystal structure

of the IHF-DNA complex (Rice et al. 1996). GalR tetramer configurations were constructed using the sequence-dependent structural parameters of the interoperator

DNA and conformation changes caused by GalR and asymmetric HU binding.

Evaluation of their DNA elastic energies gives unambiguous preference to a loop

structure in which the two gal operators adopt an antiparallel orientation. The antiparallel modes (A1 and A2) of DNA looping proved to be more favorable energetically

than the parallel modes (P1 and P2) (Table 17.2). From an energetic point of view,

formation of the optimized parallel loop, P2, results in a five-fold increase in the

bending energy compared to the optimized antiparallel loop, A1.

The A1 loop, unlike the parallel loops, would unwind the inter-operator DNA and

would be stabilized by negative supercoiling, an absolute requirement for repressosome

formation (Aki and Adhya 1997). The A1 structure is consistent with the asymmetric

(closer to OE than to OI) binding of HU when the 5¢ end of the loop is longer than



410



D.E.A. Lewis et al.

Table  17.2  Energy of 113 bp DNA looping for the

o­ ptimized GalR loops (Geanacopoulos et al. 2001)

Loop1



−DGtotal2



A1

9.4

A2

25.5

P1

92.3

P2

25.5

1

The four loops are schematically presented in Fig. 11.

2

Energy values are in kcal mol–1. −DGtotal = −DGbend −DGtwist.



the 3¢ end. The sufficiency of negative supercoiling to stimulate binding of an HU

heterodimer to a specific site (Kobryn et al. 1999) suggests that the unwinding of the

interoperator DNA in the A1 loop is important for HU binding to its site, hbs.



17.15 In Vivo Evidence for an Anti-parallel Loop

To distinguish which loop arrangement is the correct structure in the repressosome,

heterodimers of GalR were constructed and the operators designed to orient GalR

in one fixed direction and thus generate the four possible loop structures (Semsey

et  al. 2004). The results showed that simultaneous repression of gal P1and P2

operators by the GalR heterodimer is achieved only when the DNA trajectory was

in A1 form (Fig. 17.11c, lanes 1–2) and not in the A2 form (lanes 3–4) or in the

two parallel forms (P1 and P2) (lanes 5–8), confirming the energetic prediction that

A1 loop formation requires least energy. Moreover, HU binds to its site located at

+6.5 in the A1 model. In the A2 model, HU would bind to a site that is not located

at the apex of the loop and would not be able to stabilize the loop in the A2 case.

This result shows that the antiparallel form is the correct trajectory of GalR in DNA

looping. However, as discussed below, the energetic calculations for loop formation

by single DNA molecule analysis indicates that both A1 and A2 loops form with

similar energies. The discrepancy is discussed later.



17.16 Evidence of Anti-parallel Loops: AFM Studies

The distinction between the two trajectories that are most feasible from energetic considerations (A1 and P2) was studied by AFM which confirmed the anti-parallel nature

of the gal loop (Fig. 17.12) (Virnik et al. 2003). For convenience of comparison of theoretical prediction and experimental results, the two models (a) and (c) are presented as

parts incorporated into imaginary DNA minicircles depicted with broken lines. Arrows

are used to indicate the 5-prime to 3-prime directions for the operators, which concur

with the directions of transcription from the gal promoters. It is clear from the AFM

images that all the loops observed in the presence of GalR and HU conform to a trajectory of an anti-parallel A1 loop (b); loops as expected for a parallel trajectory were

never observed, giving credence to the idea that anti-parallel loops are more stable.



17  Role of HU in Regulation of gal Promoters



411



Fig. 17.12  Atomic force microscopy of GalR-HU loop with minicircle. (a) Proposed model of

“antiparallel” loop. (b) Experimental AFM image of “antiparallel” loop. (c) Proposed model of

“parallel” loop. The arrows indicate the direction of the operators from 5¢ to 3¢. The blue and red

arrows represent OI and OE, respectively (Virnik et al. 2003)



17.17 Stability of the gal Loop: Single DNA Molecule Analysis

By using single-molecule micromanipulation to generate and finely tune tension in

DNA molecules, the kinetics, thermodynamics, and supercoiling dependence of

GalR/HU-mediated DNA looping have been characterized. The factors required for

gal DNA looping in single-molecule experiments (HU, GalR and DNA supercoiling) correspond exactly to those essential for gal repression in  vitro and in  vivo.

Magnetic tweezers were used to determine the thermodynamics of loop formation

on DNA tethered at one end (Fig.  17.13) (Lia et  al. 2003, 2008). The technique

involves detecting the transition between unlooped (L) and looped (L-1) structures,

where L represents the length of the tethered particle and L-1, represents the change

in the length of the particle due to looping. By using an optimal force (F) of 0.88 pN

(to keep the DNA somewhat stretched), and superhelical density (s) of −0.03 (to

keep the DNA slightly untwisted), looping and unlooping is observed (Lia et  al.

2003). DNA molecules that are negatively supercoiled by at least 3% (s = −0.03)

and stretched with a force (F) of 0.88 pN intermittently switched between two conformations in the presence of both GalR and HU. No length-changes are observed

in the absence of GalR and/or HU or in the presence of d(+)-galactose. Similarly,

GalR and HU do not generate looping in molecules containing only one operator, OE

or OI. The extrapolated mean lifetimes of looped and unlooped conformation in the

presence of GalR and HU were <1 ms and ~21 sec, respectively (Lia et al. 2003).

Looping and unlooping are not observed when the DNA was relaxed (s = 0) or

contained positive supercoiled (s = +0.03). The associated free energy change

involved in the GalR/HU-mediated loop is ~−12 kBT or −7.1 kcal/mol (Geanacopoulos



412



D.E.A. Lewis et al.



Fig.  17.13  Experimental layout to measure DNA looping by magnetic tweezers. A gal DNA

molecule is attached to a cover slide and a magnetic force (green arrow) is applied to it. The length

of the tethered molecule (unloop) is indicated by (L) and that for the loop molecule by (L−1). The

transitions from the unloop state (broken blue arrow) to the loop state (broken red arrow) are

shown in the middle of the figure



et al. 2001; Lia et al. 2003), which compares to −9.4 Kcal/mole estimated from the

modeling studies described above (Geanacopoulos et al. 2001).



17.18 Energetics of A1 and A2 Loops Estimated by Single

Molecule Studies

The single DNA molecule assay employing magnetic tweezers when applied to

distinguish between locked A1 and A2 arrangements indicate that the A1 and A2

loops formed with similar energies in DNA. The two loops show nearly equivalent

probabilities of formation and the change in free energy for loop formation increases

with increasing force i.e., tension destabilizes the loops. Therefore, it appears that

there is no thermodynamic reason accounting for the observed functional difference

in transcriptional repression between A1 and A2 loops. However, the loop lifetimes



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

9 GalR Dimer–Dimer Interface Involved in DNA Looping

Tải bản đầy đủ ngay(0 tr)

×