Apart from promoting lysogeny by lambda, CIII protects Escherichia coli r32, the heat shock-specific sigma factor, which is also a substrate of HflB [13,14].. We tried to address this issu
Trang 1inhibition of the HflB (FtsH)-mediated proteolysis
Sabyasachi Halder1, Subhamoy Banerjee1,2and Pradeep Parrack1
1 Department of Biochemistry, Bose Institute, Kolkata, India
2 Madhav Institute of Technology & Science, Gwalior, India
Development of the temperate coliphage k depends on
a set of phage-specified regulatory proteins that
inter-act with host target proteins [1–4] Among several
effects on host protein synthesis, k provokes the
over-production of some bacterial proteins induced in the
heat shock response This effect depends on early gene
expression encoded by the leftward (pL) transcription
unit of k [5,6] CIII is a k-specific regulatory protein
from the pL operon with a potential for host
inter-action It is a small 54-residue protein that favors the
lysogenic response to infection by stabilizing kCII, the
transcriptional factor that favors lysogeny and is
responsible for primary control of the lambda
develop-mental decision for lysis or lysogeny [4,7–9] In the
absence of CIII, CII is rapidly degraded by the
ATP-dependent host metalloprotease HflB (FtsH) The
molecular mechanism for CIII-mediated inhibition of
the proteolysis of CII by HflB involves CIII–HflB
interaction [10,11] CIII itself is a substrate of HflB
[12] It has also been reported that CIIIC, the central helical domain of CIII, is resistant to HflB proteolysis, and is a more effective inhibitor than the full-length protein [11]
Apart from promoting lysogeny by lambda, CIII protects Escherichia coli r32, the heat shock-specific sigma factor, which is also a substrate of HflB [13,14] This effect generates a heat-shock response in the cell [15,16] Maximal production of CIII prolongs the heat-induced synthesis of E coli heat shock proteins even at low temperature [17] The half-life of r32 ( 2 min) is increased approximately fourfold upon overproduction of CIII, resulting in an overproduction
of heat shock proteins and rapid inhibition of cell growth [15] The molecular mechanism of the CIII-mediated inhibition of proteolysis of r32 by HflB is unknown
We tried to address this issue by studying the effects of CIII and CIIIC on the proteolysis of
Keywords
antiproteolytic activity; heat shock; lysogeny;
kCII; r 32
Correspondence
P Parrack, Department of Biochemistry,
Bose Institute, P-1 ⁄ 12, C.I.T Scheme VIIM,
Kolkata 700 054, India
Fax: +91 33 2355 3886
Tel: +91 33 2569 3227
E-mail: pradeep@bic.boseinst.ernet.in
(Received 28 April 2008, revised 14 July
2008, accepted 28 July 2008)
doi:10.1111/j.1742-4658.2008.06610.x
The CIII protein of bacteriophage lambda exhibits antiproteolytic activity against the ubiquitous metalloprotease HflB (FtsH) of Escherichia coli, thereby stabilizing the kCII protein and promoting lysogenic development
of the phage CIII also protects E coli r32, another substrate of HflB We have recently shown that the protection of CII from HflB by CIII involves direct CIII–HflB binding, without any interaction between CII and CIII [Halder S, Datta AB & Parrack P (2007) J Bacteriol 189, 8130–8138] Such
a mode of action for kCIII would be independent of the HflB substrate In this study, we tested the ability of CIII to protect r32from HflB digestion The inhibition of HflB-mediated proteolysis of r32by CIII is very similar
to that of kCII, characterized by an enhanced protection by the core CIII peptide CIIIC (amino acids 14–41 of kCIII) and a lack of interaction between r32and CIII
Abbreviation
GST, glutathione S-transferase.
Trang 2r32 by HflB in vitro Our results show that both CIII
and CIIIC inhibit the HflB-mediated proteolysis
of r32 in vitro, and that CIIIC is a more effective
inhibitor We also found that there is no interaction
between CIII and r32 From these results, we suggest
that the inhibition of r32 by CIII is due to a direct
CIII–HflB interaction
Results
In vitro proteolysis of r32by HflB requires ATP
The proteolysis of r32 by HflB is very rapid and
requires the DnaK–DnaJ–GrpE chaperone machine
in vivo [18–20] However, this proteolysis is much
slower in vitro, because of a lack of chaperone
machin-ery The in vitro proteolysis requires ATP [14] We
examined the proteolysis of purified r32-C-his (1.2 lm)
by glutathione S-transferase (GST)–HflB (0.8 lm) in
the absence or presence of ATP (5 mm) Digestion
after specified time intervals was assayed by 11%
SDS⁄ PAGE (Fig 1) It is clear that the HflB-mediated
degradation of r32could not proceed without ATP In
this respect, r32resembles kCII
Inhibition of HflB-mediated proteolysis of r32by CIII and by CIIIC
The effect of CIII and CIIIC on the HflB-mediated proteolysis of r32 was checked by treating r32-C-his (1.2 lm) with GST–HflB (0.8 lm) in the presence of CIII (100 lm) for varying time intervals (Fig 2A) It was found that r32was partially protected in the pres-ence of CIII, and 65% of r32 remained undigested after 80 min, compared with 40% for the control (r32alone, Fig 2C)
The inhibitory action of CIIIC on the proteolysis of
r32by HflB was also assayed in the presence of CIIIC (60 lm) instead of CIII (Fig 2B) It is clear that in this
Fig 1 ATP-dependence of proteolysis of r32 by HflB (Upper)
Band corresponding to r 32 on 11% SDS ⁄ PAGE, in the presence or
absence of 5 m M ATP Numbers on the top of each lane indicate
the time of digestion in min (Lower) Densitometric scan showing
the amount of r32-C-his (1.2 l M) remaining after proteolysis with
GST–HflB (0.8 l M) Each data point represents the mean value from
three identical experiments.
A
B
C
Fig 2 Inhibition of HflB-mediated proteolysis of r 32 by CIII or CIIIC The bands corresponding to r32-C-his (1.2 l M) remaining after proteolysis with GST–HflB (0.8 l M) in the absence or presence of (A) His 6 –CIII (100 l M) or (B) CIIIC (60 l M) are shown Numbers on the top of each lane indicate the time of digestion in min Samples were run on (A) a 15% SDS ⁄ PAGE, or (B) a 17.5% SDS ⁄ PAGE (C) A densitometric scan of the above bands showing the amount of r32remaining after proteolysis, for r32alone ( ) or
in the presence of CIII ( ) or CIIIC (d) Each data point represents the mean value from three identical experiments.
Trang 3case the proteolysis, with 85% of r32 remaining
undigested after 80 min (Fig 2C), was inhibited more
effectively than with intact CIII
The above experiments on the inhibition of the
pro-teolysis of r32 by HflB were also carried out in the
presence of varying amounts (up to 200 lm) of CIII or
CIIIC (Fig 3) In this case, proteolysis was terminated
after 40 min Stronger inhibition by CIIIC is also
evi-dent from these experiments, with 95% of r32
remaining undigested in the presence of 40 lm CIIIC
(Fig 3, lower)
r32interacts with HflB but does not interact
with CIII
The interaction between r32and HflB was tested in an
in vitro GST pull-down assay GST-tagged HflB was
bound to glutathione-Sepharose beads followed by the
addition of r32 The proteins were analyzed on an
11% SDS⁄ PAGE and visualized by western blotting
with anti-his as the primary antibody The same
exper-iment was repeated with GST protein as a negative
control It was observed that r32co-eluted with GST– HflB but not with GST protein (Fig 4A), implying that r32interacts specifically with GST–HflB
The interaction between r32 and CIII was assayed
in an in vitro Ni2+-nitrilotriacetic acid pull-down assay The Ni2+-nitrilotriacetic acid bound His6–CIII was mixed with r32(with the 6· histidine tag removed) and incubated at 4C for 4 h The Ni2+ -nitrilotriace-tic acid beads were washed with 1· NaCl ⁄ Piand eluted
by boiling with 1· sample buffer Proteins were ana-lyzed on 17.5% SDS⁄ PAGE (Fig 4B) It was observed that r32 did not co-elute with CIII, implying that r32 does not interact with CIII
Discussion
Phage protein CIII works as an antiprotease against
E coli HflB and protects kCII by directly binding to the protease [10,11], without any detectable interaction with the substrate kCII [11] Thus, the protection of
Fig 3 Inhibition of HflB-mediated proteolysis of r 32 by different
concentrations of CIII or CIIIC The bands corresponding to r 32
-C-his (1.2 l M) remaining after proteolysis with GST–HflB (0.8 l M) for
40 min in the presence of His6–CIII or CIIIC (up to 200 l M) are
shown in the upper panels Numbers on the top of each lane
indi-cate the concentration of His6-CIII or CIIIC (in l M) Lane C indicates
the control lane showing undigested r 32 The samples were run on
a 15% (upper) or 17.5% (middle) SDS ⁄ PAGE (Lower)
Densitomet-ric scan of the above bands depicting the amount of r32remaining
after proteolysis in the presence of CIII ( ) or CIIIC (d) Each data
point represents the mean value from three identical experiments.
A
B
Fig 4 In vitro binding of HflB-r 32 and CIII-r 32 (A) Interaction between GST–HflB and r 32 -C-his was tested by GST pull-down fol-lowed by 11% SDS ⁄ PAGE, and immunoblotting with anti-His Ig Lane 1, r32-C-his (control); lane 2, fraction pulled down with GST– HflB; lane 3, fraction pulled down with GST (B) Absence of interac-tion between His 6 –CIII and r 32 (without the histidine tag) as obtained from Ni-nitrilotriacetic acid pull-down, followed by 17.5% SDS ⁄ PAGE and Coomassie Brilliant Blue staining Lane 1, His 6 -CIII alone; lane 2, r 32 alone; lane 3, fraction pulled down.
Trang 4CII by CIII works at the protease level, rather than at
the substrate level Nevertheless, competition between
CII and CIII for interaction with HflB also influences
the proteolysis of CII [11], because CIII itself is a
sub-strate of HflB [12] CIIIC, the central region of CIII, is
not a substrate of HflB, and acts as a better inhibitor
for digestion of CII by HflB, than CIII [11] Is this
mode of antiproteolytic action of CIII a general mode
of CIII activity, or does it apply only for CII? We
examined the mode of the inhibitory action of CIII on
proteolysis of the heat shock sigma factor r32, another
substrate of HflB Like proteolysis of CII, r32
proteol-ysis requires ATP Both CIII and CIIIC inhibit this
proteolysis, with CIIIC exhibiting stronger inhibition
both as a function of time (Fig 2) or as a function of
concentration (Fig 3) Under the conditions of our
experiment, near-total inhibition by CIIIC could be
observed, whereas in the presence of CIII, only partial
inhibition was achieved As in the case of CII [10,11],
CIII appears to work as an inhibitor for the
proteoly-sis of r32through direct interaction with the protease,
characterized by a lack of interaction between r32and
CIII (Fig 4) As for CII, competition between r32and
CIII for binding to HflB would also decide the extent
and efficiency of protection of r32 by CIII, because
both are HflB substrates The relative binding affinity
for r32–HflB interaction and CIII–HflB interaction
would play an important role in the inhibition of
proteolysis of r32by HflB
Interestingly, both CII and r32 are cytosolic
sub-strates for HflB In addition, HflB also acts on several
membrane-associated substrates, for which the
mecha-nism of proteolysis is probably somewhat different
[21] Whether kCIII would act as an inhibitor for such
substrates (e.g SecY, YccA, Foa) remains to be seen
However, the biological connection between kCIII and
such substrates is poor, and it is unlikely that CIII
would have any role in the proteolysis of substrates
like SecY or YccA by HflB We think that the primary
role of CIII is associated with k lysogeny In this
respect, the antiproteolytic activity of CIII needs to be
short-lived, made possible by the fact that CIII is also
an HflB substrate CIII, however, acts via direct
inter-action with HflB, which may be enabled by the
bind-ing of its central helical region to the substrate-entry
cleft of HflB [22], as pointed out previously [11] This
probably makes CIII a general inhibitor for the
cyto-solic substrates of HflB, accounting for its protection
of r32
The intriguing question that follows is why would a
lambda protein stabilize the E coli heat shock sigma
factor? r32 is an unstable protein [23,24] with a rapid
turnover during normal growth and a transient
stabilization during heat shock, followed by rapid deg-radation [18,24] The level of r32 in E coli cells is tightly controlled, through the interactions of the DnaK–DnaJ–GrpE machinery and by HflB-mediated degradation Interestingly, these latter proteins that promote the degradation of r32 are themselves pro-duced as a result of a stress response [13,25], being transcribed from heat shock promoters involving r32 They may be part of mechanisms that allow the bacte-rium to respond quickly to changing nutritional and environmental conditions When a temperate virus like lambda takes up the lysogenic pathway in response to stressed conditions of the host, the phage functions must closely follow host conditions so that a correct developmental decision can be taken During the estab-lishment of lysogeny, stabilisation of r32by kCIII may lead to elevated production of the bacterial protease lon that is transcribed from heat shock promoters [26] and degrades the phage protein N [27], indirectly helping the lysogenic response It is known that k lysogeny is reduced in lon mutants [27,28] Alternatively, inhibition
of HflB by kCIII would lead to increased levels of r32, causing elevated production of the heat shock proteins DnaJ, DnaK and GrpE which promote the replication
of k [29,30] Such an event would serve to keep the lysogenized cell prepared for stress-induced induction while degradation by HflB is compromised Various pathways, sometimes even antagonistic [31], are known
to regulate the concentration of r32in E coli Lambda could be taking advantage of this regulatory network and act at multiple levels of host–virus interactions
Experimental procedures
Materials Various fine chemicals, reagents and enzymes were obtained from Sigma-Aldrich (New Delhi, India), USB (Cleveland,
OH, USA), Merck Limited (Mumbai, India) and Sisco Research Laboratory (Mumbai, India) Resins, primers and columns were used as described in Halder et al [11]
Purification of proteins and peptides Purification of r32was carried out according to Chattopad-hyay and Roy [32] and Sambrook et al [33] The NUT-21 strain containing pUHE 211-1 was grown at 30C in 1 L
of 2XYT medium [34] with 100 lgÆmL–1 ampicillin and
50 lgÆmL–1 kanamycin At A600 1, isopropyl thio-b-d-galactopyranoside was added to a final concentration of 0.5 mm Cells were grown for a further 20 min and poured into cold tubes All subsequent steps were performed at
4C After centrifugation at 1900 g for 10 min, the cell
Trang 5pellet was resuspended in 18 mL of ice-cold buffer L
(50 mm phosphate buffer, pH 7.9, 300 mm KCl, 50 mm
iso-leucine, 50 mm phenylalanine) with 20 lgÆmL)1 of phenyl
methanesulfonyl fluoride and disrupted by sonication The
cell lysate was centrifuged for 45 min at 14 500 g The
supernatant was loaded onto a 3-mL Ni2+-nitrilotriacetic
acid-agarose column pre-equilibrated with buffer L at a
rate 0.4 mLÆmin–1 The column was subsequently washed
with 40 mL of buffer L followed by 10 mL of buffer L plus
15 mm imidazole Nickel-bound proteins were eluted with
30 mL of 15–150 mm imidazole gradient in buffer L Pure
fractions of r32proteins were dialyzed against two changes
of 1 L of 50 mm phosphate buffer, pH 7.9, containing
300 mm KCl and 50% glycerol
HflB was overexpressed as a GST fusion protein from
plasmid pAD101 containing the hflB gene cloned in
vec-tor pGEX4T and was purified with a
glutathione-Sepha-rose column (GSTrap FF; Amersham Biosciences,
Sweden) His6–CIII was obtained by overexpressing
recombinant plasmid pAB905 containing the cIII gene
and purified with Ni-nitrilotriacetic acid column as
described previously [11] A 28-residue peptide, CIIIC,
was chemically synthesized and purified by reverse-phase
C18 column (hypersil) by water–acetonitrile gradient as
stated previously [11]
Measurements of inhibition of proteolysis
by HflB
The activity of CIII (or CIIIC) was measured by its ability
to inhibit HflB-mediated proteolysis of r32 [14] Reaction
mixtures (50 lL) were prepared by taking 60 pmol of r32
in buffer P (50 mm Tris⁄ acetate, pH 7.2, 100 mm NaCl,
5 mm MgCl2, 25 lm Zn-acetate, 1.4 mm
b-mercaptoetha-nol) containing 5 mm ATP CIII or CIIIC (up to 200 lm)
was also added Proteolysis was carried out by adding HflB
(0.8 lm) followed by incubation at 37C for specified time
intervals Reactions were stopped by the addition of 5 mm
EDTA and SDS⁄ PAGE loading buffer, followed by heating
in a boiling water-bath for 5 min The amount of r32that
remained after proteolysis was analyzed after SDS⁄ PAGE
of the samples followed by Coomassie Brilliant Blue
stain-ing and quantitation in a gel documentation system
(Bio-Rad Gel Doc 1000)
Removal of the histidine tag
To get a native r32protein without the C-terminal 6·
histi-dine tag, r32( 100 lg) was mixed with thrombin (1 unit)
and incubated overnight at 22C The protein mixture was
then passed through a 1 mL benzamidine column
(Amer-sham) for removal of thrombin Finally, the native protein
was separated from the mixture by treatment with Ni2+–
nitrilotriacetic acid beads The flow-through contained only
native proteins, devoid of the histidine tag
In vitro binding assay The interaction between HflB and r32was studied by in vitro GST pull-down assay A 50 lL aliquot of bound GST–HflB (25 lg) was mixed with 30 lg of r32in buffer P The final volume was made up to 500 lL and incubated overnight on
a rotating machine at 4C Before binding, 5% of the input solution was kept aside separately and was used as a control for comparison The proteins were analyzed in an 11% SDS⁄ PAGE and by western blotting using anti-his Ig As a negative control, the same experiment was also performed with GST protein replacing GST–HflB
The interaction of histidine-tagged proteins with other proteins was examined by in vitro Ni2+-nitrilotriacetic acid pull-down assay Purified His6–CIII was immobilized at
4C for 1 h on Ni2+-nitrilotriacetic acid beads in buffer A containing 10 mm imidazole and washed with buffer P Fifty microliters of His6–CIII bound bead ( 25 lg) was mixed with 30 lg r32 The final volume was made up to
500 lL and incubated on a rotating machine at 4C for
4 h The beads were then washed three times with buffer P and resuspended in SDS gel loading buffer for elution of proteins from the beads The eluted proteins were separated
by 17.5% SDS⁄ PAGE and visualized using Coomassie Brilliant Blue staining
Acknowledgements
The authors would like to thank Professor Siddhartha Roy, Indian Institute of Chemical Biology, Kolkata, for the gift of the E coli strain NUT-21 containing pUHE211-1 that was used to purify the r32 protein S.H was supported by fellowships from CSIR, India, and by Bose Institute Subhamoy Banerjee is a summer student at the Bose Institute, from Madhav Institute
of Technology & Science, Gwalior, India
References
1 Echols H (1980) Bacteriophage k development In The Molecular Genetics of Development(Leighton TJ & Loomis
WF, eds), pp 1–16 Academic Press, New York, NY
2 Herskowitz I & Hagen D (1980) The lysis–lysogeny decision of phage k: explicit programming and respon-siveness Annu Rev Genet 14, 399–445
3 Friedman DI, Olson ER, Georgopoulos C, Tilly K, Herskowitz I & Banuett F (1984) Interactions of bacte-riophages and host macromolecules in the growth of bacteriophage k Microbiol Rev 48, 299–325
4 Echols H (1986) Bacteriophage k development: tempo-ral switches and the choice of lysis or lysogeny Trends Genet 2, 26–30
5 Drahos DJ & Hendrix RW (1982) Effect of bacterio-phage lambda infection on synthesis of groE protein
Trang 6and other Escherichia coli proteins J Bacteriol 149,
1050–1063
6 Kochan J & Murialdo H (1982) Stimulation of groE
synthesis in Escherichia coli by bacteriophage lambda
infection J Bacteriol 149, 1166–1170
7 Hoyt MA, Knight DM, Das A, Miller HI & Echols H
(1982) Control of phage lambda development by
stabil-ity and synthesis of cII protein: role of the viral cIII
and host hflA, himA and himD genes Cell 31, 565–573
8 Rattray A, Altuvia S, Mahajna G, Oppenheim AB &
Gottesman M (1984) Control of phage lambda cII
activ-ity by phage and host functions J Bacteriol 159, 238–242
9 Banuett F, Hoyt MA, Mcfarlane L, Echols H &
Her-skowitz I (1986) HflB, a new Escherichia coli locus
regu-lating lysogeny and the level of bacteriophage lambda
CII protein J Mol Biol 187, 213–224
10 Kobiler O, Rokney A & Oppenheim AB (2007) Phage
lambda CIII: a protease inhibitor regulating the lysis–
lysogeny decision PLoS ONE 2, e363
11 Halder S, Datta AB & Parrack P (2007) Probing the
antiprotease activity of kCIII, an inhibitor of the
Esc-herichia colimetalloprotease HflB (FtsH) J Bacteriol
189, 8130–8138
12 Herman C, Thevenet D, D’Ari R & Bouloc P (1997)
The HflB protease of Escherichia coli degrades its
inhib-itor lambda CIII J Bacteriol 179, 358–363
13 Herman C, Thevenet D, D’Ari R & Bouloc P (1995)
Degradation of r32, the heat shock regulator in
Escheri-chia coli, is governed by HflB Proc Natl Acad Sci USA
92, 3516–3520
14 Tomoyasu T, Gamer J, Bukau B, Kanemory M, Mori
H, Rutman AJ, Oppenheim AB, Yura T, Yamanaka K,
Niki H et al (1995) Escherichia coli FtsH is a
mem-brane bound, ATP-dependent protease which degrades
the heat-shock transcription factor r32 EMBO J 14,
2551–2560
15 Bahl H, Echols H, Straus DB, Court D, Crowl R &
Georgopoulos CP (1987) Induction of the heat shock
response of E coli through stabilization of r32by the
phage kCIII protein Genes Dev 1, 57–64
16 Kornitzer D, Altuvia S & Oppenheim AB (1991) The
activity of CIII regulator of lambdoid bacteriophages
resides within a 24-amino-acid protein domain Proc
Natl Acad Sci USA 88, 5217–5221
17 Ang D, Chandrasekhar GN, Zylicz M & Georgopoulos
CP (1986) Escherichia coli grpE gene codes for heat
shock protein B25.3, essential for both lambda DNA
replication at all temperatures and host growth at high
temperature J Bacteriol 167, 25–29
18 Straus D, Walter W & Gross CA (1990) DnaK, DnaJ,
and GrpE heat shock proteins negatively regulate heat
shock gene expression by controlling the synthesis and
stability of r32 Genes Dev 4, 2202–2209
19 Tatsuta T, Joo DM, Calendar Y, Akiyama Y & Ogura
T (2000) Evidence for an active role of the DnaK
chap-erone system in the degradation of r32 FEBS Lett 478, 271–275
20 Tomoyasu T, Ogura T, Tatsuta T & Bukau B (1998) Levels of DnaK and DnaJ provide tight control of heat shock gene expression and protein repair in E coli Mol Microbiol 30, 567–581
21 Ito K & Akiyama A (2005) Cellular functions, mecha-nisms of action, and regulation of FtsH protease Annu Rev Microbiol 59, 211–231
22 Bieniossek C, Schalch T, Bumann M, Meister M, Meier
R & Baumann U (2006) The molecular architecture of the metalloprotease FtsH Proc Natl Acad Sci USA 103, 3066–3071
23 Grossman AD, Straus DB, Walter WA & Gross CA (1987) Sigma 32 synthesis can regulate the synthesis of heat shock proteins in Escherichia coli Genes Dev 1, 179–184
24 Tilly K, Spence J & Georgopoulos C (1989) Modulation
of stability of the Escherichia coli heat shock regulatory factor r32 J Bacteriol 171, 1585–1589
25 Yura T, Nagai H & Mori H (1993) Regulation of the heat-shock response in bacteria Annu Rev Microbiol 47, 321–350
26 Gottesman S (1984) Bacterial regulation: global regula-tory networks Annu Rev Genet 18, 415–441
27 Maurizi MR (1987) Degradation in vitro of bacterio-phage lambda N protein by Lon protease from Escheri-chia coli J Biol Chem 262, 2696–2703
28 Walker JR, Usser CL & Allen JS (1973) Bacterial cell division regulation: lysogenization of conditional cell division lon)mutants of Escherichia coli by Bacterio-phage Lambda J Bacteriol 111, 1326–1332
29 Liberek K, Georgopoulos C & Zylicz M (1988) The role
of the Escherichia coli DnaK and DnaJ heat shock pro-teins in the initiation of bacteriophage k DNA replica-tion Proc Natl Acad Sci USA 85, 6632–6636
30 Johnson C, Chandrasekhar GN & Georgopoulos C (1989) Escherichia coli DnaK and GrpE heat shock pro-teins interact both in vivo and in vitro J Bacteriol 171, 1590–1596
31 Morita MT, Kanemori M, Yanagi H & Yura T (2000) Dynamic interplay between antagonistic pathways con-trolling the r32level in Escherichia coli Proc Natl Acad Sci USA 97, 5860–5865
32 Joo DM, Ng N & Calendar R (1997) A r32mutant with a single amino acid change in the highly conserved region 2.2 exhibits reduced core RNA polymerase affin-ity Proc Natl Acad Sci USA 94, 4907–4912
33 Chattopadhyay R & Roy S (2002) DnaK–sigma 32 interaction is temperature dependent: implication for the mechanism of heat shock response J Biol Chem
277, 33641–33647
34 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular Cloning, A Laboratory Manual, Vol 3 (2nd edn) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY