Báo cáo khoa học: Interaction of the general transcription factor TnrA with the PII-like protein GlnK and glutamine synthetase in Bacillus subtilis potx

Interaction of the general transcription factor TnrA with
the PII-like protein GlnK and glutamine synthetase in
Bacillus subtilis
Airat Kayumov
1,2
, Annette Heinrich
3
, Kseniya Fedorova
2
, Olga Ilinskaya
2
and Karl Forchhammer
3
1 Kazan State University of Architecture and Engineering, Russia
2 Kazan Federal University, Department of Microbiology, Russia
3 Interfaculty Institute of Microbiology and Infection Medicine, Eberhard-Karls-Universita
¨
tTu
¨
bingen, Germany
Keywords
Bacillus subtilis; GlnK; glutamine synthetase;
nitrogen regulation; PII protein; transcription
factor TnrA
Correspondence
K. Forchhammer, Interfaculty Institute of
Microbiology and Infection Medicine,
Eberhard-Karls-Universita
¨
tTu
¨

bingen, Auf der
Morgenstelle 28, D-72076 Tu
¨
bingen,
Germany
Fax: +49 7071295843
Tel: +49 70712972096
E-mail:
(Received 26 January 2011, revised
10 March 2011, accepted 14 March 2011)
doi:10.1111/j.1742-4658.2011.08102.x
TnrA is a master transcription factor regulating nitrogen metabolism in
Bacillus subtilis under conditions of nitrogen limitation. When the preferred
nitrogen source is in excess, feedback-inhibited glutamine synthetase (GS)
has been shown to bind TnrA and disable its activity. In cells grown with
an energetically unfavorable nitrogen source such as nitrate, TnrA is fully
membrane-bound via a complex of AmtB and GlnK, which are the trans-
membrane ammonium transporter and its cognate regulator, respectively,
originally termed NrgA and NrgB. The complete removal of nitrate from
the medium leads to rapid degradation of TnrA in wild-type cells. In con-
trast, in AmtB-deficient or GlnK-deficient strains, TnrA is neither mem-
brane-bound nor degraded in response to nitrate depletion. Here, we show
that TnrA forms either a stable soluble complex with GlnK in the absence
of AmtB, or constitutively binds to GS in the absence of GlnK. In vitro,
the TnrA C-terminus is responsible for interactions with either GS or
GlnK, and this region appears also to mediate proteolysis, suggesting that
binding of GlnK or GS protects TnrA from degradation. Surface plasmon
resonance detection assays have demonstrated that GS binds to TnrA not
only in its feedback-inhibited form, but also in its non-feedback-inhibited
form, although less efficiently. TnrA binding to GlnK or GS responds dif-

ferentially to adenylate nucleotide levels, with ATP weakening interactions
with both partners.
Structured digital abstract
l
tnrA binds to glnK by surface plasmon resonance (View interaction)
l
GS binds to tnrA by pull down (View interaction)
l
tnrA binds to glnK by pull down (View interaction)
l
tnrA binds to GS by pull down (View interaction)
l
GS physically interacts with tnrA by anti bait coimmunoprecipitation (View interaction)
l
glnK binds to tnrA by pull down (View interaction)
l
glnK physically interacts with tnrA by anti bait coimmunoprecipitation (View interaction)
l
tnrA physically interacts with GS by anti bait coimmunoprecipitation (View interaction)
l
tnrA physically interacts with glnK by anti bait coimmunoprecipitation (View interaction)
l
tnrA binds to tnrA by cross-linking study (View interaction)
l
tnrA binds to GS by surface plasmon resonance (View interaction)
Abbreviations
FC, flow cell; GlnK-ST, Strep-tag II-tagged variant of GlnK; GS, glutamine synthetase; GS-ST, Strep-tag II-tagged variant of glutamine
synthetase; ITC, isothermal titration calorimetry; NAGK, N-acetyl-
L-glutamate kinase; SPR, surface plasmon resonance.
FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS 1779

Introduction
Spore-forming bacteria of the genus Bacillus have a
variety of regulatory responses to changes in the
environment. TnrA, a major transcription factor
in Bacillus subtilis under nitrogen-limited conditions
(conditions in which the nitrogen source becomes
growth-limiting), controls gene expression in response
to nitrogen availability. During nitrogen-limited
growth, TnrA serves either as an activator or a repres-
sor of genes involved in nitrogen assimilation. TnrA
activates its own gene [1], the nitrate and nitrite utiliza-
tion genes [2], the nrgAB (amtBglnK) operon (ammo-
nium transport) [3], and some other target promoters
[1,4,5]. TnrA is a negative regulator of glnA and gltAB ,
encoding the ammonium assimilatory enzymes gluta-
mine synthetase (GS) and glutamate synthase, respec-
tively [6–8]. TnrA belongs to the MerR family of
transcription factors, and is present as a homodimer of
two 12-kDa subunits [1]. The signal for its activation
remains unclear [1,6,8,9]. Several lines of evidence indi-
cate that GS acts as a sensor of nitrogen availability in
B. subtilis [1,9]. The feedback-inhibited form of GS
binds tightly to TnrA, preventing its binding to DNA,
with the most effective feedback inhibitors of GS being
glutamine and AMP [9]. Mutations in TnrA that result
in constitutive expression of the TnrA-activated amtB
promoter all lie within the C-terminal region of TnrA,
and impair the interaction between GS and TnrA
[9,10].
Another mechanism for controlling TnrA activity

was recently found. When B. subtilis cells were grown
with a poor nitrogen source such as nitrate, TnrA was
found to be almost completely associated with the cell
membrane via the ammonium uptake proteins AmtB
and GlnK, originally termed NrgA and NrgB, respec-
tively [11,12]. AmtB is a homotrimeric transmembrane
ammonium transporter that is active under nitrogen-
limited conditions [13]. GlnK consists of three 12-kDa
monomers, and is a small regulatory protein that
belongs to the PII protein family. GlnK homologs
bind to AmtB and regulate its activity, depending on
the cellular nitrogen status [14]. Like other GlnK pro-
teins, B. subtilis GlnK was shown to bind to the mem-
brane in an AmtB-dependent manner [11,12].
Furthermore, B. subtilis GlnK exhibits the unique fea-
ture of lacking a response to 2-oxoglutarate, but seem-
ing to primarily respond to ATP. Depending on the
ATP levels, B. subtilis GlnK was shown in vitro to be
soluble or membrane-bound: 4 mm ATP caused almost
full solubilization of GlnK [12]. In wild-type B. subtilis,
TnrA was shown to bind specifically to the membrane-
bound AmtB–GlnK complex, but not to soluble, ATP-
saturated GlnK. TnrA-dependent expression of the
nrgAB (amtBglnK) promoter was shown to be reduced
in a GlnK-deficient strain under conditions of ammo-
nium-limited growth [11], indicating that GlnK could
be involved in fine-tuning TnrA-dependent gene
expression. Furthermore, the cellular levels of TnrA
are modulated by proteolysis [15]. After shifting of
nitrate-grown cells to a medium containing no usable

nitrogen source, TnrA is released from the membrane
and, concomitantly, it is degraded within 15 min by
proteolysis. By contrast, no degradation of TnrA was
observed during this kind of shift experiment in B. sub-
tilis AmtB-deficient and GlnK-deficient strains, despite
TnrA being soluble in these cells [12,15]. To gain dee-
per insights in the involvement of proteolysis in modu-
lating TnrA-dependent gene expression, we aimed to
elucidate why TnrA is resistant to proteolysis in the
GlnK-deficient and AmtB-deficient strains.
Results
Immunoprecipitation of TnrA with GlnK or
GS in B. subtilis
In contrast to what is seen in wild-type cells, in amtB
or glnK knockout mutants TnrA was only detectable
in the soluble fraction of cell-free extracts, but was
never membrane-bound and no proteolytic degrada-
tion occurred after nitrate depletion [12,15]. To explain
the mechanism of TnrA protection from proteolysis,
we investigated which proteins TnrA is bound to in
these mutants, considering GlnK and GS, in particu-
lar, as potential partner proteins of TnrA. To this end,
immunoprecipitation assays were performed with cell-
free extracts from both mutant and wild-type nitrate-
grown cells, and from cells shifted to nitrogen-depleted
medium. Cell-free extracts were incubated with TnrA-
specific, GlnK-specific or GS-specific antibodies cou-
pled to Protein A Sepharose, in the presence of non-
ionic detergent. These antigen–antibody complexes
were collected, and after rigorous washing in nonionic

detergent-containing buffer and elution of antibody-
bound protein, the samples were separated by
SDS ⁄ PAGE and analyzed by immunoblotting.
In agreement with earlier data [12], GlnK was copre-
cipitated with TnrA from crude extracts of nitrate-
grown wild-type cells, when antibodies against TnrA
were used for immunoprecipitation (Fig. 1A). When
the cells were shifted to nitrate-deprived medium prior
to extraction of the proteins, much less TnrA was
immunoprecipitated, and, in consequence, less GlnK
Interaction of TnrA with GlnK and GS A. Kayumov et al.
1780 FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS
was detected, as TnrA is degraded by proteolysis fol-
lowing the shift to nitrate-deprived medium [15]. By
contrast, in the AmtB-deficient mutant strain, the same
amount of TnrA was immunoprecipitated and the
same amount of GlnK was copurified with TnrA in
both nitrogen regimes (Fig. 1A, lanes I and II). It
should be noted that GlnK is present only as soluble
protein in the AmtB-deficient mutant, whereas, in
wild-type cells, it is predominantly bound to the trans-
membrane AmtB channel, and only AmtB-bound
GlnK was able to interact with TnrA [11,12,15]. In the
TnrA immunoprecipitate of GlnK-deficient cells, again
no effect was observed on the recovery of TnrA fol-
lowing nitrate deprivation, in agreement with the lack
of TnrA degradation in this strain. GS was copurified
with TnrA and the recovery of GS was independent of
whether the cells were nitrate-grown or shifted to
nitrate-deprived medium. By contrast, no GS was co-

purified with TnrA in the wild-type or AmtB-deficient
mutant under either condition (Fig. 1A). These obser-
vations were confirmed by reversing the immunopre-
cipitation experiments, using antibodies against GlnK
or GS. TnrA was copurified with immunoprecipitated
GlnK from extracts of both wild-type and AmtB-defi-
cient cells (Fig. 1B), and was recovered by GS immu-
noprecipitation only in the GlnK-deficient mutant
(Fig. 1C). Taken together, these data demonstrate, that
TnrA binds constitutively to GlnK in AmtB-deficient
mutants, and to GS in GlnK-deficient mutants. This
constitutive binding in the mutant strains probably
protects TnrA from proteolytic degradation.
Surface plasmon resonance analysis (SPR) of the
GlnK–TnrA interaction
As a next step, the interaction of TnrA with GlnK was
investigated in vitro by BIAcore SPR detection. For
this analysis, a Strep-tag II-tagged variant of GlnK
(GlnK-ST) and a His-tagged recombinant TnrA were
overproduced in Escherichia coli BL21 and purified to
apparent electrophoretic homogeneity [12]. His
6
-tagged
TnrA was immobilized on flow cell (FC) 2 of a chelat-
ing nitrilotriacetic acid sensor chip, and GlnK-ST was
used as an analyte. His
6
-N-acetyl-l-glutamate kinase
(NAGK) from Synechococcus elongatus [16] was bound
to the reference cell (FC 1) as a control for nonspecific

interactions.
Figure 2A shows a response difference sensorgram
(FC2 – FC1) of interactions of GlnK with immobilized
TnrA. For this analysis, an analyte concentration of
40 nm GlnK (trimer) was used. Binding of GlnK
was not observed when another His-tagged protein
(His
6
-NtcA from S. elongatus) was immobilized on the
sensor chip (not shown), revealing that the observed
binding was specific for TnrA. The GlnK–TnrA com-
plex appeared to be quite stable, as revealed by the
very slow dissociation of the complex following the
injection phase (Fig. 2A). In the course of the mea-
surements, we found that 2 mm ATP (in the absence
of Mg
2+
) led to rapid dissociation of the GlnK–TnrA
complex (see below), which was subsequently used to
regenerate the TnrA-coated chip surface. The dissociat-
ing effect of 2 mm ATP on the GlnK–TnrA complex is
shown in Fig. 2A. Immediately after application of
2mm ATP to the preformed GlnK–TnrA complex,
rapid dissociation was observed, reaching the basal
levels of resonance units (GlnK free surface) within
seconds. To test the effects of various molecules on the
interaction of TnrA with GlnK, 40 nm GlnK (trimer)
Fig. 1. Coimmunoprecipitation of TnrA, GlnK and GS. Immunopre-
cipitation experiments were performed with either TnrA-specific
(A), GlnK-specific (B) or GS-specific (C) antibodies. Cells were

grown under nitrogen-limited conditions in SMM supplemented
with 20 m
M NaNO
3
(I). At late exponential growth phase, cells
were washed and shifted to combined nitrogen-free medium, incu-
bated at 37 °C with shaking for 10 min, and then harvested (II).
The crude cell extracts were used for immunoprecipitation as
described in Experimental procedures. The washed immunoprecipi-
tates were analyzed by immunoblotting with antibodies against
TnrA, GlnK, or GS, as indicated on the right.
A. Kayumov et al. Interaction of TnrA with GlnK and GS
FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS 1781
was incubated with various effector molecules, and the
mixture was used as an analyte in SPR analysis. ATP
and 2-oxoglutarate are known to be the primary effec-
tors involved in PII signaling, and they strongly affect
interactions of many GlnK proteins with their recep-
tors [17]. The divalent cations Mg
2+
or Mn
2+
were
previously shown to negatively affect the binding of
ATP to GlnK [12]. Therefore, we investigated the
binding of TnrA to GlnK in the presence of different
mixtures of Mg
2+
or Mn
2+

with the effector molecules
ATP and 2-oxoglutarate. As shown in Fig. 2B, MgCl
2
or MnCl
2
alone did not affect TnrA binding to GlnK.
However, Mg
2+
and Mn
2+
gradually relieved the
inhibitory effect of ATP on the GlnK–TnrA inter-
action, so that, in the presence of 1 mm Mg
2+
or
Mn
2+
, ATP at 2 mm was not fully inhibitory, and
2mm Mg
2+
or Mn
2+
restored more than 50% of the
GlnK–TnrA interaction in the presence of 2 mm ATP.
On the other hand, 2-oxoglutarate did not influence
the GlnK–TnrA interaction, either alone, in the
absence of divalent metals, or in combination with
ATP and Mg
2+
or Mn

2+
. To resolve the inhibitory
effect of ATP on the GlnK–TnrA interaction in the
absence of divalent cations more clearly, ATP was
titrated to the binding assays in the absence or pres-
ence of 2-oxolguatarate. As shown in Fig. 3A, 0.2 mm
ATP was sufficient to inhibit 50% of the GlnK–TnrA
interaction. The inhibitory effect of ATP was not
further enhanced by 2-oxoglutarate, in agreement with
earlier studies showing that B. subtilits GlnK does
not respond to 2-oxoglutarate [12]. Other nucleotides,
such as ADP, AMP, and GTP, at concentrations of
Fig. 3. Influence of various effector molecules on the interaction of
GlnK with the His
6
-TnrA surface. GlnK was preincubated with effec-
tor molecules at the concentrations indicated, and injected onto the
His
6
-TnrA surface. GlnK incubated in pure HBS buffer served as a
control (set as 100% binding). (A) Effect of increasing ATP concen-
trations (as indicated), with or without 1 m
M 2-oxoglutarate (2-OG)
present. (B) Effects of various nucleotides (ATP, ADP, AMP, and
GTP) and 2-OG at different concentrations on GlnK binding to TnrA.
Fig. 2. BIAcore analysis of GlnK–TnrA complex formation and ATP
effect on dissociation of the GlnK–TnrA complex. The analyte GlnK
was injected in a volume of 30 lL at a flow rate of 15 lLÆmin
)1
.

The graph shows the response difference between FC 2 (His
6
-
TnrA) and FC 1 (His
6
-NAGK). (A) ATP effect on dissociation of the
GlnK–TnrA complex. First, GlnK (40 n
M trimers) was injected onto
the His
6
-TnrA surface. After 50 s of washing with HBS buffer,
25 lL of 2 mK ATP was injected (indicated by the arrow), which
removed the GlnK bound to the His
6
-TnrA surface within a few sec-
onds. (B) Binding of GlnK to TnrA in the presence of different Mg
2+
or Mn
2+
concentrations with or without 2 mM ATP and 1 mM 2-oxo-
glutarate (2-OG) present, as indicated. GlnK in pure HBS buffer
served as a control (set as 100% binding).
Interaction of TnrA with GlnK and GS A. Kayumov et al.
1782 FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS
1–3 mm had only a small effect on the GlnK–TnrA
interaction, except for ADP, which was moderately
inhibitory, although less so than ATP (Fig. 3B). Taken
together, these measurements, although performed
under rather artificial conditions, indicate that the
GlnK–TnrA complex could be stable in vivo in the

presence of divalent cations and that the complex is
negatively affected most efficiently by ATP.
Isothermal titration calorimetry (ITC)
The affinity of binding of GlnK to nucleotides, which
affected the GlnK–TnrA interaction as revealed by
SPR analysis, was quantified by ITC. Previously, bind-
ing of different combinations of ATP and 2-oxogluta-
rate to GlnK was measured by this method [12]. Under
optimal binding conditions, strong binding of ATP was
observed (Fig. 4), which could be perfectly fitted with a
three sequential binding sites model. Data analysis
resolved two high-affinity binding sites (dissociation
constant for the first two sites: K
d1
=12±4lm and
K
d2
=77±15lm) and one low-affinity site (site 3)
(K
d3
= 4 ± 0.05 mm). No binding of other nucleotides
(ADP, AMP, and GTP) was detectable (Fig. S1), con-
firming their weak effect on the GlnK–TnrA interac-
tion (Fig. 3). These data support the idea that, in
B. subtilis, GlnK senses the intracellular ATP level at
site 3, as the binding affinity of this site is in the milli-
molar range, which is considered to be physiologically
relevant, and that this signal is then transmitted to
TnrA.
BIAcore analysis of the GS–TnrA interaction

The results of the immunoprecipitation experiments
revealed a constitutively present GS–TnrA complex in
the GlnK-deficient cells transferred to nitrate-deprived
conditions, as well as in nitrate-grown cells. Under
these conditions, GS is supposed to be in an active
state, whereas only feedback-inhibited GS was previ-
ously reported to bind TnrA [9,18]. To test whether,
indeed, non-feedback-inhibited GS can also bind
TnrA, a Strep II-tagged variant of GS (GS-ST) was
overproduced in E. coli BL21, purified to apparent
electrophoretic homogeneity [12], and used in BIAcore
analysis on immobilized His
6
-tagged TnrA immobilized
on a chelating nitrilotriacetic acid sensor chip. NAGK
from S. elongatus was bound to the reference cell as a
control for nonspecific interactions.
The response difference sensogram (FC2 – FC1) in
Fig. 5A shows the binding of non-feedback-inhibited
GS to immobilized TnrA. The GS–TnrA complex was
quite stable under the conditions used: almost no com-
plex dissociation appeared after the injection phase. In
contrast to what was found for GlnK, no efficient
effector molecule was found to remove GS from the
His
6
-TnrA surface. ATp at 10 mm caused only partial
release of GS from TnrA (Fig. 5A).
The effects of various metabolites on the GS–TnrA
interaction were also investigated. GS at 40 nm was

incubated with effector molecules for 5 min in ice, and
the mixture was used as an analyte in SPR analysis.
AMP and glutamine are known to be the most effec-
tive inhibitors of GS [9]. Figure 5B shows the effect of
the feedback inhibitors AMP and glutamine on GS
binding to a TnrA-coated sensor chip. The presence of
either AMP or glutamine led to an approximately two-
fold signal increase in comparison with non-feedback-
inhibited GS. ATP, at a physiological concentration,
negatively affected GS binding to the TnrA sensor
surface. At a concentration of 2.5 mm, it decreased
complex formation by approximately 60%, and at a
concentration of 5 mm, by 80% (Fig. 5C). ATP at
10 mm was required to completely abolish complex
formation; however, once the complex was formed,
Fig. 4. ITC of ATP binding to GlnK. The raw data were fitted with a
three-site binding model for a PII trimer. The upper panel shows
the raw data in the form of the heat effect during the titration of
25 l
M GlnK solution (trimer concentration) with ATP (titration from
2.1 to 73.5 l
M). The lower panels show the binding isotherm and
the best-fit curve according to the three sequential binding sites
model.
A. Kayumov et al. Interaction of TnrA with GlnK and GS
FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS 1783
this concentration could not efficiently dissociate the
complex (see above).
The C-terminus of TnrA is required for interaction
with both GlnK and GS, as well as for

intracellular proteolysis
Previously, it had been reported that the DNA-binding
domain of TnrA is located on its N-terminus, whereas
the C-terminus is responsible for GS binding [10]. Six
amino acids required for this interaction on the C-termi-
nus were identified (Met96, Leu97, Gln100, Leu101,
Ala103, and Phe105) (Fig. S2). A previous study
showed [19] that the TnrA-dependent nrg and nasB pro-
moters were constitutively expressed when seven or 20
amino acids were deleted from the C-terminus of TnrA,
whereas deletion of 34 amino acids from the C-terminus
resulted in a TnrA null mutation phenotype. This
implied that the TnrA signal transduction domain is
most likely located at the C-terminus. In nitrate-grown
cells, TnrA is almost completely membrane-bound via
GlnK [12]. We have speculated that GlnK may also
interact with the C-terminus of TnrA, and may play a
role in the regulation of TnrA activity and its proteo-
lysis [15]. To test this assumption, various truncations
of TnrA (lacking six, 20 and 35 amino acids from the
C-terminus) were constructed and overproduced in
E. coli (Fig. S2). Glutaraldehyde crosslinking assays
revealed that all proteins were in a dimeric state, con-
firming that the C-terminus is not required for dimeriza-
tion (Fig. S3) [10,19]. Interactions of the truncated
TnrA proteins with GlnK and with GS were determined
by pulldown and SPR analysis, as described above
(Fig. 6). As expected, the C-terminus of TnrA was abso-
lutely required for GS binding: deletion of even six
amino acids abolished this interaction (Fig. 6A,B).

Truncated forms of TnrA with the C-terminus lacking
six or 20 amino acids still bound to GlnK; however,
removal of 35 amino acids completely abolished binding
of GlnK (Fig. 6A,C). This result implies that a region in
TnrA located between 20 and 35 amino acids from the
C-terminus is required for GlnK interaction, whereas
the ultimate C-terminal amino acids of TnrA are appar-
ently needed for GS binding.
In addition, the in vitro proteolysis of truncated
TnrA variants was investigated, as described previously
for full-length TnrA [15]. B. subtilis 168 (wild-type) cells
were grown in SMM medium supplemented with
sodium nitrate until the late exponential growth phase,
and the cells were then washed and resuspended in
SMM without nitrate, and finally incubated for a
further 20 min. Samples were taken before and after
the shift, and soluble cell-free extract was prepared by
ultracentrifugation. The soluble extract (containing
20 lg of total protein) was supplemented with 50 ng of
TnrA, and the mixture was incubated at 37 °C for
60 min; TnrA incubated in buffer served as a control.
The fate of TnrA in the samples was then analyzed by
immunoblotting with TnrA-specific antibodies (Fig. 7).
During incubation in the soluble cytoplasmic extract,
TnrA6, TnrA20 and wild-type TnrA were almost com-
pletely degraded, but not TnrA35. This indicates that a
Fig. 5. BIAcore analysis of GS–TnrA complex formation. (A) GS–
TnrA interaction. First, non-feedback-inhibited GS was injected onto
the His
6

-TnrA surface. After 180 s of washing with HBS buffer,
25 lLof10m
M ATP was injected (indicated by the arrow), which
partially removed the GS bound to the His
6
-TnrA surface. (B, C)
Effects of AMP, glutamine and ATP on GS binding to TnrA. GS
was preincubated with effector molecules at the concentrations
indicated, and injected onto the His
6
-TnrA surface. GS incubated in
pure HBS buffer served as a control.
Interaction of TnrA with GlnK and GS A. Kayumov et al.
1784 FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS
region located between 20 and 35 amino acids from
the C-terminus of TnrA is required for protease recog-
nition and, at the same time, overlaps with the GlnK
recognition site (see above). This finding agrees with the
assumption that binding of GS or GlnK protects TnrA
from proteolytic degradation [15], as these proteins
would shed the recognition site for proteolytic degrada-
tion. As soon as GlnK or GS dissociate from TnrA, the
C-terminus becomes accessible to proteolysis.
Discussion
TnrA, a major transcription factor in B. subtilis for the
control of nitrogen assimilation, is active under nitro-
gen-limited conditions and is membrane-bound via the
AmtB–GlnK complex [6,12]. Its activity was shown to
be regulated by complex formation with feedback-
inhibited GS, and in the absence of a nitrogen source

TnrA is eliminated from the cells by proteolysis
[9,10,15]. The findings in the present study strongly
imply that, in vivo, TnrA is stable only in a complex
with a partner protein. In nitrate-grown wild-type cells,
TnrA is active and bound to the GlnK–AmtB complex
[12] (Fig. 1A,B). After shifting of the cells to nitrate-free
medium, this complex dissociates and TnrA becomes
degraded, whereas no degradation occurs in AmtB-defi-
cient or GlnK-deficient cells [12]. Our data show that,
in these mutant strains, TnrA interacts constitutively
with either soluble GlnK or GS, respectively, and in
consequence is protected from proteolysis (Fig. 1). The
reason for this protection, according to the present
study, is that binding of GS or GlnK to the C-terminus
Fig. 7. In vitro proteolysis of truncated TnrA proteins. Soluble cell-
free extracts were prepared from (I) cells that had been shifted into
nitrogen-free medium and incubated for 20 min, and (II) nonshifted
cells. Purified TnrA protein variants, full-length or different C-termi-
nal truncations (each 50 ng of protein), were incubated with these
extracts for 30 min (as described in [15]). TnrA incubated in assay
buffer served as a control (C). Subsequently, proteolytic removal of
TnrA was analyzed by western blotting.
Fig. 6. The interaction of truncated TnrA proteins with GlnK and
GS. (A) BIAcore analysis of GlnK and GS binding to wild-type TnrA
(TnrAwt), TnrA6, TnrA20, and TnrA35. The analyte (40 n
M GlnK or
GS oligomers) was injected in a volume of 30 lL onto the TnrA sur-
face at a flow rate of 15 lLÆmin
)1
. His

6
-NAGK served as a control
in FC 1. (B) Pulldown analysis of GS binding to TnrAwt, TnrA6,
TnrA20, and TnrA35 (see Experimental procedures for details). (C)
Pulldown analysis of GlnK binding to TnrAwt, TnrA6, TnrA20, and
TnrA35. TnrA (dimer) at 10 n
M was premixed with 10 nM GS (12-
mer) or 10 n
M GlnK (trimer), and incubated in buffer B at 20 °C for
30 min. The protein mix was loaded onto an Ni
2+
–nitrilotriacetic
acid Sepharose column to affinity-purify TnrA (I) or Strep-Tactin
Sepharose to affinity-purify GS or GlnK (II), after the columns had
been washed with buffer B. Proteins were eluted with 250 m
M
imidazole (I) or with 2.5 mM destiobiotin (II), and the eluates were
analyzed by western blot with TnrA-specific, GlnK-specific and GS-
specific antibodies, as indicated on the left.
A. Kayumov et al. Interaction of TnrA with GlnK and GS
FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS 1785
of TnrA shields the recognition site for proteolytic deg-
radation of TnrA (Figs 6 and 7).
In the AmtB-deficient strain, GlnK is located in the
cytoplasm and constitutively binds TnrA. Previously,
the AmtB-deficient strain (with constitutive GlnK–
TnrA binding) was shown to display high levels of tran-
scription from the TnrA-dependent nrgAB promoter
under ammonia-limited conditions (ammonium at low
pH) [11]. This suggests that TnrA bound to GlnK is still

able to activate gene expression. The assumption that
GlnK binding does not impair TnrA activity is consis-
tent with the observation that, in nitrate-grown wild-
type cells, TnrA is bound to the AmtB–GlnK complex
despite being transcriptionally active.
Nitrate depletion leads to dissociation of the AmtB–
GlnK–TnrA complex and subsequent TnrA degrada-
tion, whereas in AmtB mutants TnrA remains bound
to GlnK and is therefore protected from proteolysis.
This suggests a role of AmtB in dissociation of the
GlnK–TnrA complex. A possible regulatory role of
AmtB proteins has been suggested previously [20];
however, the mechanism leading to AmtB-dependent
GlnK–TnrA dissociation has remained elusive so far.
In B. subtilis wild-type cells growing on a poor
nitrogen source (nitrate), GS is active and does not
bind TnrA (Fig. 1A,C), and the latter is sequestered by
the AmtB–GlnK complex. However, in the GlnK
mutant, TnrA is constitutively bound to GS; a shift to
a nitrate-deprived medium does not lead to dissocia-
tion of the complex, and TnrA remains protected from
proteolysis. The constitutive binding of TnrA to GS
seems to contradict previous reports that only feed-
back-inhibited GS is able to bind TnrA [9]. However,
the sensitive SPR analysis has demonstrated that non-
feedback-inhibited GS is, in fact, able to bind TnrA,
although with reduced affinity as compared with feed-
back-inhibited GS (Fig. 5). The reduced affinity could
account for the fact that this interaction is not detected
by examining it indirectly through TnrA–DNA binding

assays [9,10,18]. Constitutive binding of GS to TnrA
in the GlnK-deficient strain provides an explanation
for the so-far elusive observation that TnrA-dependent
transcription from the nrgAB promoter is impaired in
a GlnK-deficeint strain growing under ammonia-
limited conditions (ammonium at low pH) [11], as GS
binding was shown to depress the transcriptional activ-
ity of TnrA [9,10,18].
Taken together, the results from this investigation
provide indications of the physiological role of the
GlnK–TnrA interaction, which has previously been
unclear. In the GlnK-bound state, TnrA is protected
from proteolysis without affecting its ability to induce
gene expression. When TnrA dissociates from the
AmtB–GlnK complex (after a shift to nitrate-deprived
conditions), it becomes rapidly degraded. Under these
conditions, GS should be in a highly active, non-
feedback-inhibited state, which has reduced affinity
for TnrA, Therefore, TnrA could be preferentially rec-
ognized by a protease as an idle protein and
degraded, as has been proposed for many proteins in
B. subtilis [21]. When, however, TnrA is complexed by
GS before nitrate downshift, as is the case in the
GlnK-deficient mutant, it remains bound and is pro-
tected from proteolysis.
Experimental procedures
Bacterial strains and growth conditions
The B. subtilis strains used in this study – strain 168 (wild
type), the AmtB-deficient strain GP 254, and the GlnK-
deficient mutant GP 253 – have been described previously

[11]. B. subtilis cells were grown in Spizizen minimal med-
ium (SMM) [22] containing glucose [0.5% (w ⁄ v)] as a car-
bon source. Sodium nitrate (20 mm) served as a nitrogen
source. l-Tryptophan was added to a final concentration of
50 mg L
)1
.
Protein preparation
TnrA from B. subtilis 168 cells and NAGK from S. elonga-
tus, carrying His
6
-tags on their N-terminal, were overpro-
duced in E. coli BL21 with the pET15b expression vector
(Novagene, San Diego, CA, USA), and purified on Ni
2+
–
nitrilotriacetic acid columns to apparent electrophoretic
homogeneity, as described previously [12,16]. GlnK-ST and
GS-ST were overproduced in E. coli BL21 with the pDG148
expression vector and purified with a Strep-Tactin column
(IBA, Go
¨
ttingen, Germany), as described in detail in Doc. S1.
Immunoblot analysis
For immunoblot analysis, the samples were separated on
15% SDS ⁄ PAGE gels. After electrophoresis, the proteins
were transferred to a nitrocellulose membrane by semi-dry
electroblotting. Antibodies were visualized with secondary
antibodies (anti-rabbit IgG–POD) (Sigma-Aldrich, Tauf-
kirchen, Germany) and the LumiLight detection system

(Roche Diagnostics, Mannheim, Germany).
Coupling antibodies to Protein A Sepharose
One hundred milligrams of Protein A Sepharose beads (GE
Healthcare, Munich, Germany) were incubated for 2 h at
24 °C in 0.5 mL of NaCl ⁄ P
i
(4.3 mm Na
2
HPO
4
, 1.8 mm
KH
2
PO
4
, 137 mm NaCl, 2.7 mm KCl, pH 8.0). The beads
were harvested by short centrifugation (11 500 g,30s,
Interaction of TnrA with GlnK and GS A. Kayumov et al.
1786 FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS
4 °C), and incubated with 0.5 mL of antiserum for 1 h at
24 °C with gentle shaking. After being washed times with
5 mL of 0.2 m Na
3
BO
3
(pH 9.0), the Sepharose beads were
resuspended in 5 mL of 0.2 m Na
3
BO
3

(pH 9.0), and
dimethyl pimelimidate dihydrochloride was added to a final
concentration of 20 mm. The incubation was continued for
30 min at 24 °C with gentle shaking. The Protein A Sepha-
rose beads were washed twice with 5 mL of 0.2 m ethanol-
amine (pH 8.0), resuspended in 5 mL of this, and incubated
for 2 h at 24 °C. To remove the unbound antibodies, the
beads were washed twice with 5 mL of NaCl ⁄ P
i
, twice with
5 mL of 100 mm glycine (pH 3.0), and twice with NaCl ⁄ P
i
,
and resuspended in 0.5 mL of NaCl ⁄ P
i
.
Immunoprecipitation
The immunoprecipitation experiments were performed as
described in [23]. Cultures of B. subtilis were grown in SMM
with 20 mm NaNO
3
to a D
600 nm
of 0.8, harvested by centri-
fugation (8500 g, 10 min, 4 °C), resuspended in buffer I
(50 mm Hepes ⁄ NaOH, pH 7.0, 50 mm KCl, 100 mm EDTA,
2mm MgCl
2
,1mm benzamidine), and broken with a
FastPrep-24 (M.P. Biomedical, Irvine, CA, USA). After cen-

trifugation (15 000 g, 10 min, 4 °C) to remove debris and
unbroken cells, the samples, containing 3 mg of total
protein, were diluted with detergent-containing buffer
[NET buffer I: 50 mm Tris ⁄ HCl, pH 7.0, 150 mm NaCl,
0.1% (v ⁄ v) nonionic detergent Nonidet P-40, 1 mm EDTA]
to a total volume of 1.5 mL, and following a 15-min incuba-
tion at 24 °C, the sample was briefly centrifuged (16 000 g,
30 s) to remove debris. To this extract, 100 lL of a suspen-
sion of Protein A Sepharose beads with coupled antibodies
was added. After a 3-h incubation at 4 °C, Sepharose beads
were harvested by centrifugation (16 000 g,30s,4°C), and
the sediment was washed twice with NET buffer I, once with
NET buffer II (NET buffer I with 500 mm NaCl), and once
with buffer IP [10 mm Tris ⁄ HCl, pH 7.5, 0.1% (v ⁄ v) Noni-
det P-40]. The bound proteins were eluted from Protein A
Sepharose by 10 consecutive additions of 50 lL each of
buffer IE (100 mm glycine, pH 2.4), and the elutions were
pooled and analyzed by immunoblot analysis with TnrA-
specific, GlnK-specific and GS-specific antibodies.
BIAcore SPR detection
SPR experiments were performed with a BIAcore X biosen-
sor system (Biacore AB, Uppsala, Sweden). To immobilize
His
6
-TnrA on the nitrilotriacetic acid biosensor surface,
Ni
2+
was first bound to the nitrilotriacetic acid surfaces of
both flow chambers through injection of 10 lLofa5mm
NiSO

4
solution. Then, His
6
-TnrA was injected into FC 2 in a
volume of 50 lL at a concentration of 2 nmol ⁄ mL in HBS
buffer (10 mm Hepes, 200 mm NaCl, 0.005% Nonidet P-40,
pH 7.5). His
6
-NAGK from S. elongatus was injected into
FC 1 in a volume of 50 lL at a concentration of 2 nmol ⁄ mL
in HBS buffer. This resulted in increases in resonance units
of 500 in FC 2 and 800 in FC 1. Experiments were per-
formed at 25 °C in HBS buffer at a flow rate of 15 lLÆmin
)1
,
with GlnK-ST or GS-ST as analyte at the concentrations
indicated. To analyze the effect of small molecules on GlnK
or GS binding to the His
6
-TnrA surface, the analyte was
preincubated for 5 min on ice with the various effector mole-
cules as indicated, and was then injected into the sensor chip.
For novel reloading of the nitrilotriacetic acid sensor chip
with fresh His
6
-TnrA, 50 lL of 0.5 m EDTA was injected to
completely remove His
6
-TnrA and Ni
2+

. Subsequently, the
chip was loaded again with Ni
2+
and His
6
-TnrA or His
6
-
NAGK as described above. This procedure was performed
when the performance of analyte binding to the His
6
-TnrA
surface started to decrease.
ITC
ITC experiments were performed on a VP-ITC microcalo-
rimeter (MicroCal, LCC, New York, USA) in 10 mm
Hepes ⁄ NaOH, 50 mm KCl and 100 mm NaCl (pH 7.4) at
20 °C [24]. For determination of ATP, ADP, AMP and
GTP binding isotherms for wild-type GlnK, 25 lm protein
(trimer concentration) was titrated with 2 mm ATP, 2 mm
ADP, 2 mm AMP, or 2 mm GTP, respectively. The ligand
(5 lL) was injected 35 times into the 1.4285-mL cell with
stirring at 350 r.p.m. The binding isotherms were calculated
from received data, and fitted to a three-site binding model
with MicroCal origin software (Northampton, MA, USA).
Construction of mutant tnrA genes
All DNA manipulations were performed by standard meth-
ods as described in [23]. Mutant tnrA genes were amplified
with pfu polymerase from chromosomal DNA of B. subtil-
is 168. Briefly, the tnrA gene coding for the protein with dele-

tion of six amino acids from C-terminus was obtained with
primers TnrAN (5¢-GCT CGA GGA TCC GAT GAC CA-
C AGA AGA TCA TTC TT-3¢) and TnrA6 (5¢-TTA
ACG GGA TCC GTA CCG TTA GTG AGC ATT AAG-
3¢). The PCR products were purified, digested with BamHI,
and ligated into the BamHI-digested pET-15b vector (Nov-
agen). This vector provides N-terminally His
6
-tagged protein
overexpression in E. coli BL21 cells. To obtain TnrA pro-
teins lacking 20 and 35 amino acids from the C-terminus,
TnrA20 (5¢-TCC AGC GGA TCC TTC CGC ACT TAC
GGA TC-3¢) and TnrA35 (5¢-TTC TTT GGA TCC CAT
ATC CTT TTA AAT CTC TGC-3¢) oligonucleotides were
used, respectively, instead of TnrA6. The sequences of all
cloned genes were confirmed by DNA sequencing.
Pulldown
For these assays, the purified His
6
-tagged TnrA proteins
(wild-type and truncated versions), GlnK-ST and GS-ST
A. Kayumov et al. Interaction of TnrA with GlnK and GS
FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS 1787
were used. Initially, the proteins (10 nm each protein) were
diluted in 300 lL of buffer B (100 mm Tris ⁄ HCl, pH 8.0,
200 mm NaCl, 2 mm MgCl
2
,1mm EDTA) and incubated
at 20 °C for 30 min. Afterwards, the protein mixture was
loaded onto Ni

2+
–nitrilotriacetic acid Sepharose (Qiagen,
Hilden, Germany) or Strep-Tactin Sepharose (IBA, Go
¨
ttin-
gen, Germany) equilibrated with 10 column volumes
(10 · 0.2 mL) of buffer B, with subsequent washing four
times with five volumes of the same buffer. Proteins were
eluted with buffer E (buffer B supplemented with 250 mm
imidazole from Ni
2+
–nitrilotriacetic acid Sepharose or
2.5 mm destiobiotin from the Strep-Tactin column). The
samples were collected and analyzed by western blot with
TnrA-specific, GlnK-specific and GS-specific antibodies.
Acknowledgements
J. Stu
¨
lke (Go
¨
ttingen) is gratefully acknowledged for
providing B. subtilis strains. This work was supported
by DFG grant Fo195, the Russian–German program
‘Michail Lomonosov’ A ⁄ 08 ⁄ 75091, and the Ministry of
Education and Science of the Russian Federation (gov-
ernment contract No. P2573 from 25 November 2009).
References
1 Fisher SH (1999) Regulation of nitrogen metabolism in
Bacillus subtilis: vive la difference! Mol Microbiol 32,
223–232.

2 Nakano M, Hoffmann T, Zhu Y & Jahn D (1998)
Nitrogen and oxygen regulation of Bacillus subtilis
nasDEF encoding NADH-dependent nitrite
reductase by TnrA and ResDE. J Bacteriol 180,
5344–5350.
3 Wray LV, Zalieckas JM, Ferson AE & Fisher SH
(1998) Mutational analysis of the TnrA-binding sites in
the Bacillus subtilis nrgAB and gabP promoter regions.
J Bacteriol 180, 2943–2949.
4 Wray LV, Atkinson MR & Fisher SH (1994) The nitro-
gen-regulated Bacillus subtilis nrgAB operon encodes a
membrane protein and a protein highly similar to the
Escherichia coli glnB-encoded PII protein. J Bacteriol
176, 108–114.
5 Wray LV, Ferson AE & Fisher SH (1997) Expression
of the Bacillus subtilis ureABC operon is controlled by
multiple regulatory factors including CodY, GlnR,
TnrA and Spo0H. J Bacteriol 179, 5494–5501.
6 Wray LV, Ferson AE, Rohrer K & Fisher SH (1996)
TnrA, a transcription factor required for global nitro-
gen regulation in Bacillus subtilis. Proc Natl Acad Sci
USA 93, 8841–8845.
7 Belitsky BR, Wray LV, Fisher SH, Bohannon DE &
Sonenshein AL (2000) Role of TnrA in nitrogen
source-dependent repression of Bacillus subtilis
glutamate synthase gene expression. J Bacteriol 182,
5939–5947.
8 Fisher SH & Debarbouille M (2002) Nitrogen source
utilization and its regulation. In Bacillus Subtilis and Its
Closest Relatives: From Genes to Cells (Sonenshein AL,

Hoch JA & Losick R eds), pp. 181–191. American
Society for Microbiology, Washington, DC.
9 Wray LV, Zalieckas JM & Fisher SH (2001)
Bacillus subtilis glutamine synthetase controls gene
expression through a protein–protein interaction with
transcription factor TnrA. Cell 107, 427–435.
10 Wray LV Jr & Fisher SH (2007) Functional analysis of
the carboxy-terminal region of Bacillus subtilis TnrA a
MerR family protein. J Bacteriol 189, 20–27.
11 Detsch C & Stulke J (2003) Ammonium utilization in
Bacillus subtilis: transport and regulatory functions of
NrgA & NrgB. Microbiology 149, 3289–3297.
12 Heinrich A, Woyda K, Brauburger K, Meiss G, Detsch
C, Stu
¨
lke J & Forchhammer K (2006) Interaction of the
membrane-bound GlnK–AmtB complex with the master
regulator of nitrogen metabolism TnrA in Bacillus sub-
tilis. J Biol Chem 281, 34909–34917.
13 Khademi S & Stroud RM (2006) The Amt
⁄ MEP ⁄ Rh
family: structure of AmtB and the mechanism of ammo-
nia gas conduction. Physiology (Bethesda) 21, 419–429.
14 Javelle A, Severi E, Thornton J & Merrick M (2004)
Ammonium sensing in Escherichia coli. Role of the
ammonium transporter AmtB and AmtB–GlnK
complex formation. J Biol Chem 279, 8530–8539.
15 Kayumov A, Heinrich A, Sharipova M, Iljinskaya O &
Forchhammer K (2008) Inactivation of the general tran-
scription factor TnrA in Bacillus subtilis by proteolysis.

Microbiology 154, 2348–2355.
16 Maheswaran M, Urbanke C & Forchhammer K (2004)
Complex formation and catalytic activation by the PII
signaling protein of N-acetyl-L-glutamate kinase from
Synechococcus elongatus strain PCC 7942. J Biol Chem
279, 55202–55210.
17 Forchhammer K (2008) P(II) signal transducers: novel
functional and structural insights. Trends Microbiol 16,
65–72.
18 Fisher SH & Wray LV (2006) Feedback-resistant muta-
tions in Bacillus subtilis glutamine synthetase are clus-
tered in the active site. J Bacteriol 188, 5966–5974.
19 Shin BS, Choi SK, Smith I & Park SH (2000) Analysis
of tnrA alleles which result in a glucose-resistant sporu-
lation phenotype in Bacillus subtilis. J Bacteriol 182,
5009–5012.
20 Tremblay PL & Hallenbeck PC (2008) Of blood brains
and bacteria the Amt ⁄ Rh transporter family: emerging
role of Amt as a unique microbial sensor. Mol Micro-
biol 71, 12–22.
21 Gerth U, Kock H, Kusters I, Michalik S, Switzer RL &
Hecker M (2008) Clp-dependent proteolysis down-regu-
Interaction of TnrA with GlnK and GS A. Kayumov et al.
1788 FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS
lates central metabolic pathways in glucose-starved
Bacillus subtilis. J Bacteriol 190, 321–331.
22 Saxild HH & Nygaard P (1987) Genetic and
physiological characterization of Bacillus subtilis
mutants resistant to purine analogs. J Bacteriol 169,
2977–2983.

23 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular
Cloning: A Laboratory Manual. Cold Spring Harbor
Laboratory Press, New York, NY.
24 Fokina O, Chellamuthu VR, Zeth K & Forchhammer
K (2010) A novel signal transduction protein P(II) vari-
ant from Synechococcus elongatus PCC 7942 indicates a
two-step process for NAGK–P(II) complex formation.
J Mol Biol 399, 410–421.
Supporting information
The following supplementary material is available:
Fig. S1. Isothermal titration calorimetry of (A) ADP,
(B) AMP and (C) GTP binding to GlnK.
Fig. S2. TnrA C-terminal truncations.
Fig. S3. Crosslinking analysis of truncated TnrA
proteins.
Doc. S1. Purification of His
6
-tagged TnrA proteins,
purification of GlnK-ST and GS-ST, and glutaralde-
hyde crosslinking assays.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
A. Kayumov et al. Interaction of TnrA with GlnK and GS

FEBS Journal 278 (2011) 1779–1789 ª 2011 The Authors Journal compilation ª 2011 FEBS 1789