Microbiology (2012), 158, 1656–1663
DOI 10.1099/mic.0.058446-0
Influence of the ADP/ATP ratio, 2-oxoglutarate and
divalent ions on Azospirillum brasilense PII protein
signalling
Edileusa C. M. Gerhardt,13 Luı́za M. Araújo,13 Ronny R. Ribeiro,2
Leda S. Chubatsu,1 Marcelo Scarduelli,1 Thiago E. Rodrigues,1
Rose A. Monteiro,1 Fábio O. Pedrosa,1 Emanuel M. Souza1
and Luciano F. Huergo1
Correspondence
Luciano F. Huergo
[email protected]
1
Instituto Nacional de Ciência e Tecnologia da Fixação Biológica de Nitrogênio, Departamento de
Bioquı́mica e Biologia Molecular, Setor de Ciências Biológica, Universidade Federal do Paraná,
Centro Politécnico, Jardim das Américas, Caixa Postal 19046, UFPR Curitiba, Paraná, Brazil
2
Departamento de Quı́mica, Centro Politécnico, Jardim das Américas, Caixa Postal 19081,
UFPR Curitiba, Paraná, Brazil
Received 22 February 2012
Revised
20 March 2012
Accepted 23 March 2012
Proteins belonging to the PII family coordinate cellular nitrogen metabolism by direct interaction
with a variety of enzymes, transcriptional regulators and transporters. The sensing function of PII
relies on its ability to bind the nitrogen/carbon signalling molecule 2-oxoglutarate (2-OG). In
Proteobacteria, PII is further subject to reversible uridylylation according to the intracellular levels
of glutamine, which reflect the cellular nitrogen status. A number of PII proteins have been shown
to bind ADP and ATP in a competitive manner, suggesting that PII might act as an energy sensor.
Here, we analyse the influence of the ADP/ATP ratio, 2-OG levels and divalent metal ions on in
vitro uridylylation of the Azospirillum brasilense PII proteins GlnB and GlnZ, and on interaction
with their targets AmtB, DraG and DraT. The results support the notion that the cellular
concentration of 2-OG is a key factor governing occupation of the GlnB and GlnZ nucleotide
binding sites by ATP or ADP, with high 2-OG levels favouring the occupation of PII by ATP. Both
PII uridylylation and interaction with target proteins responded to the ADP/ATP ratio within the
expected physiological range, supporting the concept that PII proteins might act as cellular energy
sensors.
INTRODUCTION
The PII family comprises a group of widely distributed
trimeric signal transduction proteins found in nearly all
Bacteria and also present in Archaea and in the chloroplasts
of plants (Forchhammer, 2008). The archetypes of this
family are the two Escherichia coli PII proteins, GlnB and
GlnK, the functions and structures of which have been
extensively characterized. PII proteins regulate the activity
of a variety of target proteins through direct protein–
protein interactions. These interactions are modulated by
the allosteric binding of ATP, ADP and 2-oxoglutarate (2OG) to PII altering its 3D structure (Forchhammer, 2008).
The major structural rearrangements in response to the
binding of these molecules occur in a flexible region of PII
3These authors contributed equally to this work.
Abbreviations: 2-OG, 2-oxoglutarate; D/T, ADP/ATP ratio.
Four supplementary figures are available with the online version of this
paper.
1656
known as the T-loop (Forchhammer, 2008), and in some
organisms PII activity is further modulated by posttranslational modification of the T-loop. In Proteobacteria, PII is subjected to reversible uridylylation catalysed
by the bifunctional uridylyltransferase/uridylyl-removing
enzyme GlnD. The uridylylation status of PII is mainly
regulated by the intracellular levels of glutamine which
binds to GlnD and controls its opposing activities (Jiang
et al., 1998).
Structural and biochemical analysis showed that ATP and
ADP bind in a competitive manner to the lateral clefts
between each monomer of the PII trimer (Conroy et al.,
2007; Jiang & Ninfa, 2009; Xu et al., 1998). Therefore, PII
proteins have the potential to act as sensors of the ADP/
ATP ratio (D/T). Recent studies demonstrated that the
regulation of NtrB and ATase by E. coli GlnB (Jiang &
Ninfa, 2009) and the regulation of NAGK and PipX by
Synechococcus elongatus GlnB (Fokina et al., 2011) were
sensitive to the D/T in vitro, suggesting that PII might be
able to sense the cellular energy.
Downloaded from www.microbiologyresearch.org by
058446 G 2012 SGM
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
Printed in Great Britain
A. brasilense PII protein signalling
METHODS
The binding mode of 2-OG to PII remained elusive until
the recent resolution of the crystallographic structures of
PII bound to this molecule. In the structures of Azospirillum brasilense GlnZ (Truan et al., 2010), S. elongatus
GlnB (Fokina et al., 2010) and Archaeoglobus fulgidus
GlnK3 (Maier et al., 2011), the 2-OG binding site was
located in the vicinity of the ATP binding site where 2-OG
contacts the Mg2+ bound to the ATP phosphates. Another
important contact is a salt bridge formed between the 5carboxyl group of 2-OG and the side chain of the highly
conserved K58 residue (Fig. 1a). These structural data
explained the previous observation that MgATP and 2-OG
bind to PII cooperatively (Truan et al., 2010). The binding
of 2-OG to S. elongatus and E. coli GlnB in the presence of
saturating MgATP exhibits strong anti-cooperativety such
that the occupation of the first 2-OG binding site occurs
within the low micromolar range, whereas the occupation
of the second and third sites occurs with higher dissociation constants (Fokina et al., 2010; Jiang & Ninfa, 2009).
Protein purification. The A. brasilense GlnD, DraT, DraG and AmtB
proteins were expressed and purified as N-terminally 66His-tagged
proteins, as described previously (Huergo et al., 2007, 2009). The A.
brasilense GlnB, GlnZ, GlnZ Q39K and GlnZ Q39E proteins were
expressed and purified as described previously (Moure et al., 2012;
Rajendran et al., 2011). The GlnZ Q39E mutant was obtained using
the Quick-Change site-directed mutagenesis kit (Agilent) using the
pMSA4 plasmid (Moure et al., 2012) as a template. The HisDraT–
GlnB complex was purified as described previously (Huergo et al.,
2009). Proteins were quantified using the Bradford reagent (Sigma).
Uridylylation reaction. Uridylylation assays were performed as
described previously (Araújo et al., 2008). The reactions contained
10 mM Tris/HCl pH 7.5, 10 mM KCl, 25 mM MgCl2, 1 mM UTP,
3 mM GlnB or GlnZ, 100 nM GlnD and the indicated concentration
of ATP, ADP and 2-OG. ATP (.99 %), ADP (.95 %) and 2-OG
(.98 %) were purchased from Sigma. In some experiments, MgCl2
was substituted with 5 mM MnCl2. Reactions were incubated at 30 uC
and stopped at the times indicated in each experiment by adding
20 mM EDTA. Samples were analysed by native-PAGE and gels were
stained with SyPro-Ruby (Invitrogen).
The first recognized function of PII proteins was to regulate
nitrogen metabolism according to the cellular nitrogen
status. In diazotrophs, both the synthesis and the activity of
nitrogenase are regulated by PII (Dixon & Kahn, 2004;
Dodsworth & Leigh, 2006; Huergo et al., 2012; Sarkar et al.,
2012). In A. brasilense, GlnB binds to and activates the
ADP-ribosyltransferase enzyme DraT promoting nitrogenase inactivation upon an ammonium shock. The second PII
protein in this organism, GlnZ, regulates the activity of the
enzyme catalysing the nitrogenase ADP-ribosyl removing
reaction, DraG. The regulation of DraG occurs through the
formation of a ternary complex between AmtB, GlnZ and
DraG on the cell membrane (Huergo et al., 2012).
Protein pull-down assay. In vitro complex formation was
performed using MagneHis-Ni2+ beads (Promega), as described
previously (Huergo et al., 2009). Interaction reactions were conducted
in buffer containing 50 mM Tris/HCl pH 8, 0.1 M NaCl, 10 %
glycerol, 20 mM imidazole and 5 mM MgCl2 in the presence or
absence of effectors (2-OG, ATP and ADP) as indicated in each
experiment. In experiments using HisDraG or HisAmtB, 0.05 % (w/v)
lauryldimethylamine-oxide (LDAO) was used to keep AmtB soluble
and reduce non-specific interaction of PII with the resin. For
experiments using HisDraT, LDAO was substituted with 0.05 % (v/
v) Tween 20. Samples were analysed by using 12.5 % SDS-PAGE gels
stained with Coomassie blue. The data reported were confirmed in at
least two independent experiments.
ATP determination. The contaminant levels of ATP in ADP
In order to further characterize the putative energy sensing
function of PII proteins, we analysed the influence of the D/
T, 2-OG levels and the metal ions Mg2+, Mn2+ and Ca2+
on in vitro uridylylation of A. brasilense GlnB and GlnZ and
on the in vitro interactions between GlnB, GlnZ and their
protein targets AmtB, DraG and DraT. The results support
the concept that the cellular concentration of 2-OG is a key
factor governing the occupation of the GlnB and GlnZ
nucleotide binding sites by ATP or ADP. The interactions of
GlnB and GlnZ with their target proteins responded to the
D/T within the physiological range, supporting the proposal
that PII proteins might function as cellular energy sensors.
solutions were quantified by using the ATP bioluminescence kit
HSII (Roche). Briefly, 50 ml samples or ATP standards diluted in the
range of 1026–10212 M were mixed with 50 ml luciferase reagent. The
luminescent signals were recorded in a 96-well plate reader (Tecan
Infinity 200) using 1 s delay and integration from 1 to 10 s.
EPR. For the EPR analysis, 100 ml purified A. brasilense GlnZ at
120 mM (trimer concentration) was incubated on ice in 50 mM Tris/
HCl pH 7.5, 100 mM KCl, 20 % glycerol (v/v), 2 mM MnCl2, with
2 mM ATP, ADP and 2-OG as indicated in each experiment. EPR
experiments were carried out on a Bruker Elexsys E500 spectrometer
equipped with a Bruker ER4122SHQE resonator. The experiments
were performed at 77 K.
(b)
(a)
Q39
2-OG
Q39
K58
K58
ADP
ATP
http://mic.sgmjournals.org
Fig. 1. Comparison between the structures of
GlnZ bound to MgATP 2-OG (PDB 3MHY)
(a) and to ADP (PDB 3O5T) (b). Key: cyan,
one GlnZ monomer; blue stick, 2-OG; black
sphere, Mg2+; dashed lines, octahedral coordination of Mg2+. In the ADP-bound form the
Q39 residue interacts with K58 (green sticks),
whereas in the MgATP plus 2-OG structure,
Q39 participates in the coordination of Mg2+.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
1657
E. C. M. Gerhardt and others
0
2-OG concentration (M)
ATP concentration (mM)
(a)
1
5
10
25
0
50 100 250 1000
GlnZ
GlnZ-UMP1
GlnZ-UMP2
GlnZ-UMP3
(b)
0
ADP concentration (mM)
1
5
10
5
10
(c)
1
2.5
5
10
(d)
25 50 100 250 1000
GlnB
GlnB-UMP3
ADP concentration (mM)
(d)
1
2.5
5
10
25
50 100 250 1000
GlnB
GlnB-UMP3
Fig. 2. Effect of ATP and ADP on the uridylylation of A. brasilense
GlnZ and GlnB. The uridylylation reaction was carried out in the
presence of 5 mM 2-OG and varying the ATP and ADP
concentrations as indicated. (a) and (b), GlnZ; (c) and (d), GlnB.
Samples were analysed by native PAGE.
RESULTS
Effects of adenine nucleotides and 2-OG levels on
in vitro uridylylation of A. brasilense PII
We showed previously that in vitro uridylylation of the A.
brasilense PII proteins by GlnD required ATP and 2-OG
(Araújo et al., 2004, 2008). Given that PII can also bind
ADP we investigated whether this molecule could promote
or affect PII uridylylation in vitro. Using a saturating concentration of 2-OG (5 mM), ATP activated the uridylylation of both GlnZ and GlnB at a concentration as low as
1 mM. Nearly full uridylylation was observed using 1 mM
ATP, confirming previous results (Fig. 2a, c). Note that
partially uridylylated GlnB trimers are not detected in
native PAGE, as reported previously (Araújo et al., 2008).
When ATP was substituted with ADP, uridylylation of
both proteins started to occur only at 10–25 mM ADP,
and uridylylation was partial even when up to 1 mM ADP
was used (Fig. 2b, d). These results are in accordance with
those observed with E. coli GlnB, Rhodospirillum rubrum
GlnJ and Herbaspirillum seropedicae GlnB; all these
proteins were uridylylated less efficiently when ADP was
the activating nucleotide in comparison with ATP (Jiang
et al., 1998; Teixeira et al., 2008). One exception is H.
seropedicae GlnK, which was efficiently uridylylated when
ADP was used together with high 2-OG levels (Bonatto
et al., in press).
1658
GlnZ
GlnZ-UMP1
GlnZ-UMP2
GlnZ-UMP3
GlnZ
GlnZ-UMP1
GlnZ-UMP2
GlnZ-UMP3
GlnB
GlnB-UMP3
ATP concentration (mM)
(c)
0
50 100 500 2000 5000
(b)
50 100 2501000
25
GlnZ
GlnZ-UMP1
GlnZ-UMP2
GlnZ-UMP3
0
1
(a)
GlnB
GlnB-UMP3
Fig. 3. Effect of 2-OG on the uridylylation of GlnZ and GlnB. The
reactions were carried out in the presence of 1 mM ATP (a and c) or
ADP (b and d) and the 2-OG concentration indicated. (a) and (b),
GlnZ; (c) and (d), GlnB. Samples were analysed by native PAGE.
We next tested the uridylylation response of PII in a fixed
concentration of 1 mM ATP or ADP and varying 2-OG
concentrations. Both GlnZ and GlnB were uridylylated at low
micromolar concentrations of 2-OG when ATP was present
(Fig. 3a, c). However, uridylylation only occurred in high
micromolar 2-OG concentrations when ADP was used as the
activating nucleotide (Fig. 3b, d). Interestingly, the ATP and
2-OG requirements for GlnB uridylylation were lower than
those for GlnZ uridylylation, suggesting that GlnB and GlnZ
have different affinities for binding these effectors.
The fact that ADP was a much less efficient effector of GlnB
and GlnZ uridylylation could be explained by an inefficient
binding of ADP to GlnB and GlnZ under the assay conditions. This hypothesis is excluded by the fact that interaction between PII, AmtB and DraG, which is dependent on
the occupation of PII by ADP, was observed in the presence
of 1 mM ADP and 2-OG up to 2 mM (see below). Reactions
in the presence of 5 mM 2-OG and AMP in concentrations
ranging from 1 mM to 1 mM did not result in detectable
uridylylation (data not shown).
The results depicted in Figs 2 and 3 suggest either that GlnB
and GlnZ are better substrates for uridylylation when
saturated with MgATP and 2-OG or that GlnD is activated
by these molecules. In order to distinguish between these
models we analysed the ability of two GlnZ variants, Q39K
and Q39E, to be uridylylated in vitro. According to the crystal
structure of GlnZ bound to MgATP and 2-OG, the side chain
of Q39 participates in the coordination of Mg2+ and is
required for the formation of the 2-OG binding site (Fig. 1a).
The addition of a positive (Q39K) or negative (Q39E) charge
in this region would affect 2-OG binding. Indeed, an E. coli
GlnB Q39E mutant is impaired in 2-OG binding (Jiang et al.,
1997) and the same substitution in R. rubrum GlnJ inhibited
GlnJ uridylylation (Teixeira et al., 2008). Both A. brasilense
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
Microbiology 158
A. brasilense PII protein signalling
Q39K and Q39E could not be uridylylated in vitro (Fig. S1,
available with the online version of this paper), supporting the
proposal that the occupation of the 2-OG binding site of PII is
required for uridylylation.
Biochemical studies with E. coli and S. elongatus PII
proteins showed that ADP has an antagonistic effect upon
2-OG binding to PII (Fokina et al., 2010, 2011; Jiang &
Ninfa, 2009), and this is supported by structural data
showing that the ADP-bound conformation of PII occludes
the 2-OG binding site (Truan et al., 2010) (Fig. 1b). It is
still not clear whether or how 2-OG binds to PII proteins
when the PII nucleotide-binding site is occupied by ADP.
Furthermore, calorimetric titration studies using S. elongatus GlnB and H. seropedicae GlnK failed to detect 2-OG
binding in the presence of ADP (Fokina et al., 2011; M. A.
S. Oliveira and others, personal communication).
Strikingly, we noted that high levels of ADP and 2-OG
combined were able to promote GlnB and GlnZ uridylylation (Figs 2 and 3), suggesting that 2-OG could interact
with PII in the presence of ADP. An alternative explanation
for these results would be that the uridylylation in the
presence of ADP is caused by traces of ATP contaminating
the ADP solutions. Measurement of ATP using luciferase
indicated 1.6 % contamination of ATP in the ADP solution
used and, indeed, the GlnZ uridylylation profile in the presence of ADP correlates well with the level of contaminating ATP. For instance, GlnZ uridylylation using 250 mM
ADP, which has about 5 mM contaminating ATP, was
similar to that obtained using 5 mM ATP (Fig. 2). Hence,
the uridylylation observed with ADP could be caused by
contaminant traces of ATP which would outcompete ADP
under high 2-OG levels because ATP and 2-OG bind to PII
in a cooperative manner.
Recent studies suggested that PII proteins respond to the D/
T in vitro (Fokina et al., 2011; Jiang & Ninfa, 2009). To
verify whether the uridylylation of A. brasilense GlnZ was
responsive to the D/T, uridylylation assays were carried out
under different D/Ts, keeping the total nucleotide concentration at 1 mM and using three different 2-OG levels
(0.01, 0.1 and 2 mM) covering the physiological range (Fig.
4). As expected from the results using ATP and ADP
separately, increasing the D/T reduced GlnZ uridylylation
in all three 2-OG conditions (Fig. 4). However the effect
was relatively small at 0.01 mM 2-OG, presumably because
the very low 2-OG limits the MgATP 2-OG form of GlnZ.
Within the physiological 2-OG range (0.1–2 mM) an
increasing D/T clearly reduced GlnZ uridylylation though
there was relatively little influence on GlnZ uridylylation by
varying the 2-OG concentration.
Effects of combinations of ATP, ADP, 2-OG and
divalent metal ions on GlnZ–DraG, GlnZ–AmtB
and GlnB–AmtB complex formation
Biochemical and structural analysis showed that A. brasilense
GlnZ interacts with DraG in vitro only when bound to ADP.
http://mic.sgmjournals.org
2-OG
0.01 mM
1
2
3 4
2-OG
0.1 mM
5
6
7
2-OG
2 mM
8 9 10 11 12 1314 15
GlnZ
GlnZ-UMP1
GlnZ-UMP2
GlnZ-UMP3
1 0.75 0.5 0.25 0
1 0.75 0.5 0.25 0
1 0.75 0.5 0.25 0 ATP (mM)
0 0.25 0.5 0.75 1
0 0.25 0.5 0.75 1
0 0.25 0.5 0.75 1 ADP (mM)
Fig. 4. Effects of the D/T on the uridylylation of GlnZ. The reactions
were carried out with 0.01, 0.1 or 2 mM using the D/Ts indicated.
Samples were analysed by native PAGE electrophoresis.
This interaction is abolished when the GlnZ binding sites are
empty or when bound only to ATP (Huergo et al., 2009;
Rajendran et al., 2011). A similar response has been observed
for the AmtB–GlnZ and AmtB–GlnB complexes (Rodrigues
et al., 2011). Therefore, the capacity of PII to interact with
DraG or with AmtB is an indirect method to analyse the
competitive binding of ATP and ADP to PII.
Complex formation between His-tagged DraG and native
GlnZ was assayed by pull-down in the presence of different
D/Ts. The final nucleotide concentration was kept constant
at 1 mM. As expected, in the presence of ADP alone, GlnZ
co-eluted with His–DraG, indicating complex formation
(Fig. 5a). There was no co-elution when using ATP only,
confirming previous results. In the presence of combinations of ATP and ADP up to a 9 : 1 ratio, the DraG–GlnZ
interaction was still detected, though the intensity of GlnZ
co-eluted corresponded to 48 % of that observed when
using ADP only (Fig. 5a). Similar results were obtained
with the AmtB–GlnZ and AmtB–GlnB complexes (Fig. S2).
These data support the proposition that GlnZ and GlnB
nucleotide binding sites are partially occupied by ADP even
when ATP is nine times more abundant than ADP.
Addition of 2-OG to the reaction containing different D/Ts
altered the results in a 2-OG-concentration-dependent
manner (Figs 5b and S2). The higher concentration of
2-OG used (2 mM, which reflects the highest physiological level reported under nitrogen limitation) (Radchenko
et al., 2010; Senior, 1975) strongly reduced the DraG–GlnZ
interaction in most of the D/Ts assayed (Fig. 5b). When
ADP was the sole nucleotide, 2 mM 2-OG reduced only
23 % of the DraG–GlnZ protein interaction (Fig. 5b,
compare lanes 5 and 15). As we still do not know whether
2-OG can bind to GlnZ when the protein is saturated with
ADP, this slight inhibition might be caused by contaminant ATP in the ADP solution as suggested by our
uridylylation assays. The use of 0.01 mM 2-OG generated
results similar to those observed without the addition of 2OG, this is not surprising since this concentration is lower
than the reported physiological range. An intermediate
concentration of 2-OG (0.1 mM, which reflects the lowest
physiological level reported under nitrogen abundance)
(Radchenko et al., 2010; Senior, 1975) resulted in an
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
1659
E. C. M. Gerhardt and others
(a)
1
2
3
4
5
His–DraG
2-OG
2 mM
(b)
2-OG
0.1 mM
2-OG
0.01 mM
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
His–DraG
GlnZ
GlnZ
ATP (mM)
1
0.9
0.75
0.5
0
ADP (mM)
0
0.1
0.25
0.5
1
ATP (mM) 1
0.9 0.75 0.5 0 1
0.9 0.75 0.5
0
1 0.9 0.75 0.5 0
ADP (mM) 0
0.1 0.25 0.5 1 0
0.1 0.25 0.5
1
0 0.1 0.25 0.5 1
Fig. 5. Effects of the D/T on the interaction between DraG and GlnZ under different 2-OG levels. Complex formation between
His–DraG and GlnZ was assessed by pull-down using the D/T ratio indicated. No 2-OG was added (a) or 2-OG was added at
the concentrations indicated (b). Proteins eluted from the Ni2+ beads were analysed by SDS-PAGE. The graphs below each
gel show the quantification of GlnZ signal that co-eluted with His–DraG as arbitrary units.
intermediate response (Fig. 5b, lanes 6–10). The effect of
varying the D/T and 2-OG levels on AmtB–GlnB and
AmtB–GlnZ complexes were, in general, similar to those
observed with DraG–GlnZ except that the complexes with
AmtB were more sensitive to 2-OG (Fig. S2).
Effects of Mg2+, Mn2+ and Ca2+ on PII effector
molecule binding
According to the structural data, high 2-OG would facilitate
ATP binding only when Mg2+ is present (Fig. 1a). To
confirm this we assayed DraG–GlnZ complex formation
using a D/T of 1 in the absence or presence of 2 mM of 2OG and in the absence or presence of 5 mM of the divalent
metal ions Mg2+, Mn2+ or Ca2+. As expected, complex
formation was sensitive to 2-OG in the presence of Mg2+
but not in its absence (Fig. 6a). Replacement of Mg2+ with
Mn2+ resulted in a similar response, as observed with
Mg2+, whilst Ca2+ had no effect.
In order to analyse whether the binding mode of Mn2+ to
GlnZ was similar to that proposed for Mg2+ in the GlnZ
structure (Fig. 1a) an EPR approach was used. We expected
that the coordination of Mn2+ would change when GlnZ is
simultaneously bound to MnATP and 2-OG and this
change could be potentially detected by EPR (Fig. 1a). The
EPR analysis indicates the appearance of a typical six line
hyperfine pattern of Mn2+ species in the region of 150–200
mT when GlnZ was in the presence of Mn2+ in
combination with both ATP and 2-OG. When ATP or 2OG was omitted, this signal was not observed (Fig. S3).
Furthermore, this signal was not observed in control
solutions containing Mn2+, ATP and 2-OG but without
the addition of GlnZ (data not shown). Similar responses
1660
were observed with A. brasilense GlnB (data not shown).
These data indicate that the coordination of Mn2+ changes
in the presence of GlnZ, ATP and 2-OG, suggesting that
Mn2+ binds to PII in a similar manner as Mg2+ (Fig. 1a).
The dissociation of the R. rubrum GlnJ–AmtB1 complex
showed different sensitivities to 2-OG whether Mg2+ or
Mn2+ was present (Teixeira et al., 2008), suggesting that,
when compared with MgATP, MnATP could increase the
affinity of GlnJ for 2-OG binding, and hence that the relative
concentrations of Mg2+ and Mn2+ could alter the PII
output signal. We did not observe any significant differences
in the apparent affinity of nucleotides and 2-OG for the in
vitro interaction assay between DraG and GlnZ when Mg2+
was substituted with Mn2+ (Fig. 6a). We also studied the
effect of different metal ions on the in vitro dissociation of
the A. brasilense DraT–GlnB complex. The DraT–GlnB
complex is stable in the presence of MgATP but is readily
dissociated by 2-OG in the presence of ATP and Mg2+
(Huergo et al., 2009). As expected, the complex was sensitive
to 2-OG in the presence of Mg2+ but not in its absence.
Mn2+ resulted in a similar response, whilst Ca2+ again
had no effect (Fig. S4a). To investigate if the dissociation
constant for 2-OG binding to GlnB could be altered by the
ATP counter-ion we assessed the stability of the DraT–GlnB
complex using a fixed concentration of ATP and variable 2OG in the presence of Mg2+ or Mn2+. The results support
the proposal that the affinity for 2-OG binding is similar
whether Mg2+ or Mn2+ is used as the ATP counter-ion
(Fig. S4b).
Even though Ca2+ can form octahedral complexes with
oxygen, its larger ionic radius and increased coordination
distances in comparison to Mg2+ and Mn2+ (Harding,
2000) could explain its inability to facilitate the binding of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
Microbiology 158
A. brasilense PII protein signalling
2-OG
0 mM
(a)
1
2
3
2-OG
2 mM
4
5
6
7
8
His–DraG
range (Dominguez, 2004). Consequently, according to the
data in Fig. 6(b), Mg2+ would outcompete Ca2+ for ATP
binding, and modulation of 2-OG binding to PII by Ca2+
is unlikely to occur in vivo under regular physiological
conditions. On the other hand, we cannot completely rule
out that transient Ca2+ upshifts could affect PII signalling
in vivo.
DISCUSSION
GlnZ
Mg2+
+
+
Mn2+
Ca2+
(b)
+
+
+
1
2
+
3
4
6
5
1
1
5
His–DraG
GlnZ
Mg2+ (mM) 1
Ca2+ (mM)
1
1
1
Fig. 6. Effect of different metals on DraG–GlnZ complex formation. (a) Complex formation between His–DraG and GlnZ was
assessed by pull-down using 0.5 mM ATP and ADP and 5 mM of
the indicated metal in the absence of 2-OG (lanes 1–4) or in the
presence of 2 mM 2-OG (lanes 5–8). (b) Complex formation
between His–DraG and GlnZ was assessed by pull-down using
0.75 mM ATP and 0.25 mM ADP, 100 mM 2-OG and different
metals at the concentrations indicated. Samples were analysed by
SDS-PAGE.
2-OG to PII in the presence of ATP (Figs 6a and S4a).
However, like Mg2+, Ca2+ relieved the inhibitory effect of
ATP on S. elongatus NAGK–PII interaction (Maheswaran
et al., 2004), suggesting that this ion can interact with
the ATP phosphates in PII. Hence, Ca2+ may have the potential to act as a modulator of the PII response to 2-OG,
because Ca2+ could compete with Mg2+ for binding to the
ATP phosphates, though only PII-MgATP and not PIICaATP would be able to engage in interaction with 2-OG.
To verify this hypothesis, we analysed complex formation
between DraG–GlnZ in the presence of a D/T of 0.33 and
0.1 mM 2-OG using different Mg2+/Ca2+ ratios. The
results indicate that Ca2+ is indeed able to abrogate the
inhibitory effect of 2-OG when this ion is five times more
abundant than Mg2+ (Fig. 6b, lane 5). There is a precedent
for a link between Ca2+ signalling and regulation of nitrogen metabolism by PII (Leganés et al., 2009); however, the
reported intracellular levels of Ca2+ are within the nanomolar range whilst those of Mg2+ are in the micromolar
http://mic.sgmjournals.org
We have analysed the effects of the D/T, 2-OG levels and
divalent metal ions on A. brasilense PII protein uridylylation and interaction with the target proteins in vitro. Our
data support the view that occupation of the PII trimer
by MgATP and 2-OG is a prerequisite for the A. brasilense
PII uridylylation. Therefore, uridylylation is an indirect
method to determine the status of PII occupation by
MgATP 2-OG. Conversely, the interaction between PII and
DraG or between PII and AmtB was used as an indirect
method to determine the levels of PII bound to ADP. The
combination of these analyses indicates that the cellular
concentration of 2-OG is a key factor governing the
occupation of the GlnB and GlnZ by ATP or ADP. High 2OG favoured ATP binding to PII, as supported by the
increasing uridylylation rates (Fig. 3), and reduction in the
DraG–GlnZ (Fig. 5), AmtB–GlnZ and AmtB–GlnB interactions (Fig. S2). These same analyses indicate that low 2OG levels favour the binding to PII of ADP over ATP.
Comparison between the structures of A. brasilense GlnZ
bound to MgATP 2-OG and that bound to ADP (Fig. 1)
provides a rational explanation for these results. ATP,
2-OG and the side chain of Q39 participate in the
coordination of Mg2+ in the GlnZ MgATP 2-OG structure
and the 5-carboxyl group of 2-OG forms a salt bridge with
the side chain of K58 (Truan et al., 2010) (Fig. 1a). When
GlnZ is bound to ADP, the side chain of Q39 interacts with
K58 and occupies the same space to which 2-OG binds
(Rajendran et al., 2011) (Fig. 1b). Therefore, 2-OG binding
would facilitate ATP binding but would perturb the ADPbound conformation of GlnZ.
In agreement with the results reported here, high 2-OG
stimulated ATP binding but either did not affect or inhibited
the binding of ADP to E. coli GlnB and GlnK in vitro. This
response was confirmed by interaction assays between AmtB
and GlnK, in which a D/T of 0.08 (ATP 136 more abundant
than ADP) led to the same response as that observed when
using ADP only (Radchenko et al., 2010). Similar results
were observed here for the A. brasilense AmtB–PII (Fig. S2).
The addition of 2 mM 2-OG under this condition
completely disrupted AmtB–PII complex formation (Fig.
S2) and E. coli AmtB–GlnK complex (Radchenko et al.,
2010). Hence, the fact that increasing concentrations of 2OG promote the exchange of the ADP bound to GlnZ for
ATP could be a universal feature in proteobacterial PII.
In prokaryotes, 2-OG is the major carbon skeleton used for
assimilation of inorganic nitrogen. Its pool decreases from
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
1661
E. C. M. Gerhardt and others
~1.4 to ~0.3 mM a few seconds after a nitrogen-deprived
culture of E. coli receives an ammonium shock (Radchenko
et al., 2010). Such 2-OG reduction would facilitate the
exchange of the ATP bound to PII for ADP, generating a 2OG-controlled conformational switch in PII which, in turn,
would regulate the interaction between PII and nitrogenrelated proteins as AmtB and DraG.
Is this 2-OG-controlled conformational switch the only
reason for the ability of PII to bind ATP and ADP
competitively or could the PII structure also respond to
variations in the available D/T in a fixed 2-OG concentration? The reported physiological D/T in bacteria varies
from 0.06 to 1.48 (Bennett et al., 2009; Paul & Ludden,
1984; Radchenko et al., 2010; Upchurch & Mortenson,
1980). Variations in D/T within this range altered the
DraG–GlnZ, AmtB–GlnZ and AmtB–GlnB interactions
significantly, especially at 0.1 mM 2-OG (Figs 5b and S2).
Hence, in principle, the A. brasilense PII proteins have the
potential to act as D/T sensors.
The best-documented functions of A. brasilense PII are to
control nitrogenase expression and activity and to regulate
the global nitrogen transcriptional factor NtrC, which
activates the expression of a plethora of genes under
nitrogen deprivation (de Zamaroczy, 1998; Huergo et al.,
2012). Both of these processes are energy demanding and it
would be rational to use PII as a sensor to downregulate all
these processes when energy is scarce. Further analysis
correlating the PII activity with intracellular fluctuations in
the D/T will be required to clarify the predicted energysensing function of PII proteins.
ACKNOWLEDGEMENTS
We are grateful to Valter Baura and Roseli Prado for their technical
support and to Marco A. S. Oliveira for help with ATP determination
(all UFPR). Mike Merrick (JIC) and Fritz Winkler (PSI) are thanked
for their comments and discussions. This work was supported by
CNPq, INCT, CAPES and Fundação Araucária.
affected by ATP, ADP and 2-oxoglutarate in vitro. Arch Microbiol
[Epub ahead of print].
Conroy, M. J., Durand, A., Lupo, D., Li, X. D., Bullough, P. A., Winkler,
F. K. & Merrick, M. (2007). The crystal structure of the Escherichia coli
AmtB-GlnK complex reveals how GlnK regulates the ammonia
channel. Proc Natl Acad Sci U S A 104, 1213–1218.
de Zamaroczy, M. (1998). Structural homologues PII and PZ of
Azospirillum brasilense provide intracellular signalling for selective
regulation of various nitrogen-dependent functions. Mol Microbiol 29,
449–463.
Dixon, R. & Kahn, D. (2004). Genetic regulation of biological nitrogen
fixation. Nat Rev Microbiol 2, 621–631.
Dodsworth, J. A. & Leigh, J. A. (2006). Regulation of nitrogenase by 2-
oxoglutarate-reversible, direct binding of a PII-like nitrogen sensor
protein to dinitrogenase. Proc Natl Acad Sci U S A 103, 9779–9784.
Dominguez, D. C. (2004). Calcium signalling in bacteria. Mol
Microbiol 54, 291–297.
Fokina, O., Chellamuthu, V. R., Forchhammer, K. & Zeth, K. (2010).
Mechanism of 2-oxoglutarate signaling by the Synechococcus elongatus
PII signal transduction protein. Proc Natl Acad Sci U S A 107, 19760–
19765.
Fokina, O., Herrmann, C. & Forchhammer, K. (2011). Signal-
transduction protein P(II) from Synechococcus elongatus PCC 7942
senses low adenylate energy charge in vitro. Biochem J 440, 147–156.
Forchhammer, K. (2008). P(II) signal transducers: novel functional
and structural insights. Trends Microbiol 16, 65–72.
Harding, M. M. (2000). The geometry of metal-ligand interactions
relevant to proteins. II. Angles at the metal atom, additional weak
metal-donor interactions. Acta Crystallogr D Biol Crystallogr 56, 857–
867.
Huergo, L. F., Merrick, M., Pedrosa, F. O., Chubatsu, L. S., Araújo,
L. M. & Souza, E. M. (2007). Ternary complex formation between
AmtB, GlnZ and the nitrogenase regulatory enzyme DraG reveals a
novel facet of nitrogen regulation in bacteria. Mol Microbiol 66, 1523–
1535.
Huergo, L. F., Merrick, M., Monteiro, R. A., Chubatsu, L. S., Steffens,
M. B., Pedrosa, F. O. & Souza, E. M. (2009). In vitro interactions
between the PII proteins and the nitrogenase regulatory enzymes
dinitrogenase reductase ADP-ribosyltransferase (DraT) and dinitrogenase reductase-activating glycohydrolase (DraG) in Azospirillum
brasilense. J Biol Chem 284, 6674–6682.
Huergo, L. F., Pedrosa, F. O., Muller-Santos, M., Chubatsu, L. S.,
Monteiro, R. A., Merrick, M. & Souza, E. M. (2012). PII signal
REFERENCES
transduction proteins: pivotal players in post-translational control of
nitrogenase activity. Microbiology 158, 176–190.
Araújo, M. S., Baura, V. A., Souza, E. M., Benelli, E. M., Rigo, L. U.,
Steffens, M. B., Pedrosa, F. O. & Chubatsu, L. S. (2004). In vitro
Jiang, P. & Ninfa, A. J. (2009). Sensation and signaling of alpha-
uridylylation of the Azospirillum brasilense N-signal transducing GlnZ
protein. Protein Expr Purif 33, 19–24.
Araújo, L. M., Huergo, L. F., Invitti, A. L., Gimenes, C. I., Bonatto, A. C.,
Monteiro, R. A., Souza, E. M., Pedrosa, F. O. & Chubatsu, L. S.
(2008). Different responses of the GlnB and GlnZ proteins upon in
vitro uridylylation by the Azospirillum brasilense GlnD protein. Braz J
Med Biol Res 41, 289–294.
Bennett, B. D., Kimball, E. H., Gao, M., Osterhout, R., Van Dien, S. J.
& Rabinowitz, J. D. (2009). Absolute metabolite concentrations and
implied enzyme active site occupancy in Escherichia coli. Nat Chem
Biol 5, 593–599.
Bonatto, A. C., Souza, E. M., Oliveira, M. A., Monteiro, R. A.,
Chubatsu, L. S., Huergo, L. F. & Pedrosa, F. O. (2012). Uridylylation
of Herbaspirillum seropedicae GlnB and GlnK proteins is differentially
1662
ketoglutarate and adenylylate energy charge by the Escherichia coli PII
signal transduction protein require cooperation of the three ligandbinding sites within the PII trimer. Biochemistry 48, 11522–11531.
Jiang, P., Zucker, P., Atkinson, M. R., Kamberov, E. S., Tirasophon,
W., Chandran, P., Schefke, B. R. & Ninfa, A. J. (1997). Structure/
function analysis of the PII signal transduction protein of Escherichia
coli: genetic separation of interactions with protein receptors.
J Bacteriol 179, 4342–4353.
Jiang, P., Peliska, J. A. & Ninfa, A. J. (1998). Enzymological
characterization of the signal-transducing uridylyltransferase/uridylyl-removing enzyme (EC 2.7.7.59) of Escherichia coli and its
interaction with the PII protein. Biochemistry 37, 12782–12794.
Leganés, F., Forchhammer, K. & Fernández-Piñas, F. (2009). Role of
calcium in acclimation of the cyanobacterium Synechococcus elongatus
PCC 7942 to nitrogen starvation. Microbiology 155, 25–34.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
Microbiology 158
A. brasilense PII protein signalling
Maheswaran, M., Urbanke, C. & Forchhammer, K. (2004). Complex
formation and catalytic activation by the PII signaling protein of Nacetyl-L-glutamate kinase from Synechococcus elongatus strain PCC
7942. J Biol Chem 279, 55202–55210.
Maier, S., Schleberger, P., Lü, W., Wacker, T., Pflüger, T., Litz, C. &
Andrade, S. L. (2011). Mechanism of disruption of the Amt-GlnK
AmtB and partially uridylylated forms of the P(II) protein GlnZ.
Biochim Biophys Acta 1814, 1203–1209.
Sarkar, A., Köhler, J., Hurek, T. & Reinhold-Hurek, B. (2012). A novel
regulatory role of the Rnf complex of Azoarcus sp. strain BH72. Mol
Microbiol 83, 408–422.
complex by P(II)-mediated sensing of 2-oxoglutarate. PLoS ONE 6,
e26327.
Senior, P. J. (1975). Regulation of nitrogen metabolism in Escherichia
coli and Klebsiella aerogenes: studies with the continuous-culture
technique. J Bacteriol 123, 407–418.
Moure, V. R., Razzera, G., Araújo, L. M., Oliveira, M. A., Gerhardt,
E. C., Müller-Santos, M., Almeida, F., Pedrosa, F. O., Valente, A. P. &
other authors (2012). Heat stability of Proteobacterial PII protein
Teixeira, P. F., Jonsson, A., Frank, M., Wang, H. & Nordlund, S.
(2008). Interaction of the signal transduction protein GlnJ with the
facilitate purification using a single chromatography step. Protein
Expr Purif 81, 83–88.
cellular targets AmtB1, GlnE and GlnD in Rhodospirillum rubrum:
dependence on manganese, 2-oxoglutarate and the ADP/ATP ratio.
Microbiology 154, 2336–2347.
Paul, T. D. & Ludden, P. W. (1984). Adenine nucleotide levels in
Rhodospirillum rubrum during switch-off of whole-cell nitrogenase
activity. Biochem J 224, 961–969.
Truan, D., Huergo, L. F., Chubatsu, L. S., Merrick, M., Li, X. D. &
Winkler, F. K. (2010). A new P(II) protein structure identifies the 2-
Radchenko, M. V., Thornton, J. & Merrick, M. (2010). Control of
AmtB-GlnK complex formation by intracellular levels of ATP, ADP,
and 2-oxoglutarate. J Biol Chem 285, 31037–31045.
Rajendran, C., Gerhardt, E. C., Bjelic, S., Gasperina, A., Scarduelli,
M., Pedrosa, F. O., Chubatsu, L. S., Merrick, M., Souza, E. M. & other
authors (2011). Crystal structure of the GlnZ-DraG complex reveals a
different form of PII-target interaction. Proc Natl Acad Sci U S A 108,
18972–18976.
Rodrigues, T. E., Souza, V. E., Monteiro, R. A., Gerhardt, E. C., Araújo,
L. M., Chubatsu, L. S., Souza, E. M., Pedrosa, F. O. & Huergo, L. F.
(2011). In vitro interaction between the ammonium transport protein
http://mic.sgmjournals.org
oxoglutarate binding site. J Mol Biol 400, 531–539.
Upchurch, R. G. & Mortenson, L. E. (1980). In vivo energetics and
control of nitrogen fixation: changes in the adenylate energy charge
and adenosine 59-diphosphate/adenosine 59-triphosphate ratio of
cells during growth on dinitrogen versus growth on ammonia.
J Bacteriol 143, 274–284.
Xu, Y., Cheah, E., Carr, P. D., van Heeswijk, W. C., Westerhoff, H. V.,
Vasudevan, S. G. & Ollis, D. L. (1998). GlnK, a PII-homologue:
structure reveals ATP binding site and indicates how the T-loops may
be involved in molecular recognition. J Mol Biol 282, 149–165.
Edited by: J. Moir
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 09:47:14
1663