होम Molecular Microbiology Characterization of the GlnK protein of Escherichia coli

Characterization of the GlnK protein of Escherichia coli

,
यह पुस्तक आपको कितनी अच्छी लगी?
फ़ाइल की गुणवत्ता क्या है?
पुस्तक की गुणवत्ता का मूल्यांकन करने के लिए यह पुस्तक डाउनलोड करें
डाउनलोड की गई फ़ाइलों की गुणवत्ता क्या है?
खंड:
32
साल:
1999
भाषा:
english
पृष्ठ:
13
DOI:
10.1046/j.1365-2958.1999.01349.x
फ़ाइल:
PDF, 477 KB
Conversion to is in progress
Conversion to is failed
0 comments
 

अपनी समीक्षा पोस्ट करने के लिए साइन इन करें या साइन अप करें
आप पुस्तक समीक्षा लिख सकते हैं और अपना अनुभव साझा कर सकते हैं. पढ़ूी हुई पुस्तकों के बारे में आपकी राय जानने में अन्य पाठकों को दिलचस्पी होगी. भले ही आपको किताब पसंद हो या न हो, अगर आप इसके बारे में ईमानदारी से और विस्तार से बताएँगे, तो लोग अपने लिए नई रुचिकर पुस्तकें खोज पाएँगे.
Molecular Microbiology (1999) 32(2), 301–313

Characterization of the GlnK protein of Escherichia coli
Mariette R. Atkinson and Alexander J. Ninfa*
Department of Biological Chemistry, University of
Michigan Medical School, Ann Arbor, MI 48109-0606,
USA.
Summary
The GlnK and PII signal transduction proteins are paralogues that play distinct roles in nitrogen regulation.
Although cells lacking GlnK appear to have normal
nitrogen regulation, in the absence of PII, the GlnK protein controls nitrogen assimilation by regulating the
activities of the PII receptors glutamine synthetase
adenylyltransferase (ATase) and the kinase/phosphatase nitrogen regulator II (NRII or NtrB), which controls
transcription from nitrogen-regulated promoters. Here,
the wild-type GlnK protein and two mutant forms of
GlnK were purified, and their activities were compared
with those of PII using purified components. GlnK and
PII were observed to have unique properties. Both PII
and GlnK were potent activators of the phosphatase
activity of NRII, although PII was slightly more active.
In contrast, PII was approximately 40-fold more potent
than GlnK in the activation of the adenylylation of glutamine synthetase by ATase. While both GlnK and PII
were readily uridylylated by the uridylyltransferase
activity of the signal-transducing uridylyltransferase/
uridylyl-removing enzyme (UTase/UR), only PII⬃UMP
was effectively deuridylylated by the UR activity of the
UTase/UR. Finally, there were subtle differences in the
regulation of GlnK activity by the small molecule effector 2-ketoglutarate compared with the regulation of PII
activity by this effector. Altogether, these results suggest that GlnK is unlikely to play a significant role in
the regulation of ATase in wild-type cells, and that
the main role of GlnK may be to contribute to the
regulation of NRII and perhaps additional, unknown
receptors in nitrogen-starved cells. Also, the slow
deuridylylation of GlnK⬃UMP by the UTase/UR suggests that rapid interconversion of GlnK between uridylylated;  and unmodified forms is not necessary for
GlnK function. One mutant form of GlnK, containing
the alteration R47W, was observed to lack specifically
the ability to activate the NRII phosphatase in vitro ; it
was able to be uridylylated by the UTase/UR and to
Received 28 October, 1998; revised 15 January, 1999; accepted 19
January, 1999. *For correspondence. E-mail aninfa@umich.edu; Tel.
(þ1) 734 763 8065; Fax (þ1) 734 763 4581.
䊚 1999 Blackwell Science Ltd

activate the adenylylation activity of ATase. Another
mutant form of GlnK, containing the Y51N alteration
at the site of uridylylation, was not uridylylated by the
UTase/UR and was defective in the activation of both
the NRII phosphatase activity and the ATase adenylylation activity.
Introduction
The PII signal transduction protein of Escherichia coli , product of glnB , plays a key role in the regulation of nitrogen
assimilation in response to intracellular signals of nitrogen
and carbon status. PII activates the phosphatase activity
of the glnL (ntrB ) product, NRII (NtrB) and, by so doing,
controls the extent of phosphorylation and activity of the
enhancer-binding transcription factor NRI (NtrC) encoded
by glnG (ntrC ) (Ninfa and Magasanik, 1986; Keener and
Kustu, 1988; Kamberov et al ., 1994). Recently, we have
observed that PII also inhibits the kinase activity of NRII
(P. Jiang and A. J. Ninfa, unpublished data). Thus, the
binding of PII converts NRII from a kinase to a phosphatase. The phosphorylated form of NRI is an activator of
transcription from nitrogen-regulated promoters that are
transcribed by ␴54-RNA polymerase (Ninfa and Magasanik,
1986). PII prevents transcription from these promoters by
bringing about the dephosphorylation of NRI⬃P by NRII.
In addition, PII activates the adenylylation activity of adenylyltransferase (ATase), product of glnE , which regulates
the activity of glutamine synthetase (GS) by reversible
adenylylation [Brown et al ., 1971; Jiang et al ., 1998a]. Glutamine synthetase is the most important enzyme for the
assimilation of the preferred nitrogen source, ammonia,
when ammonia is limiting. Adenylylation of a GS subunit
results in its inactivation.
Much previous work, including studies in which the regulatory system was reconstituted from purified components,
indicated that the key signal of nitrogen status is the concentration of glutamine, while the key signal of carbon
status is the concentration of 2-ketoglutarate (Senior,
1975; Engleman and Francis, 1978; Jiang et al ., 1998a,b).
These effectors act antagonistically to regulate PII activity
by different mechanisms (Kamberov et al ., 1995; Liu and
Magasanik, 1995; Jiang et al ., 1998a,b). Glutamine regulates the activity of PII by controlling its uridylylation state
and by controlling the interaction of PII with the ATase. PII
is reversibly uridylylated by the signal-transducing uridylyltransferase/uridylyl-removing enzyme (UTase/UR, product
of glnD ); glutamine inhibits the uridylyltransferase reaction

302 M. R. Atkinson and A. J. Ninfa
and activates the uridylyl-removing reaction (Engleman
and Francis, 1978; Jiang et al ., 1998c). Thus, under conditions of nitrogen excess, when the internal glutamine
concentration is high, PII is mainly unmodified, while in
conditions of nitrogen limitation, PII is mainly uridylylated.
In addition, glutamine activates the adenylylation reaction
of ATase synergistically with PII. When the concentration
of glutamine is high, a low concentration of PII is able to
activate the adenylylation activity of ATase, whereas at
low concentrations of glutamine, a higher concentration of
PII is required to activate ATase (P. Jiang and A. J. Ninfa,
unpublished). In the absence of PII, glutamine is a weak
activator of the ATase adenylylation activity (Brown et al .,
1971; Rhee et al ., 1989).
The uridylylation of PII prevents its interaction with NRII,
and thus prevents the dephosphorylation of NRI⬃P
(Atkinson et al ., 1994; Kamberov et al ., 1995). PII⬃UMP
is an essential activator for the deadenylylation activity of
ATase (Brown et al ., 1971; Rhee et al ., 1989), which catalyses the conversion of the inactive GS⬃AMP to active
GS.
The carbon signal, 2-ketoglutarate, regulates the activity
of unmodified PII allosterically (Liu and Magasanik, 1995;
Kamberov et al ., 1995; Jiang et al ., 1998a,b,c). The trimeric
PII contains three sites that bind 2-ketoglutarate. However,
binding of this effector to the three sites is not independent
(Kamberov et al ., 1995; Jiang et al ., 1998c). The first molecule of 2-ketoglutarate binds to PII at a very low concentration (⬇5 ␮M; Kamberov et al ., 1995; Jiang et al ., 1998a),
such that PII is probably bound all the time in vivo , where
the 2-ketoglutarate concentration is reported to range
from 0.1 to 0.9 mM (Senior, 1975). Binding of additional
molecules of 2-ketoglutarate to PII only occurs at higher
concentrations, with dissociation constants estimated from
kinetic studies to be approximately 50–250 ␮M (Jiang
et al ., 1998a,b,c). Thus, binding of the first molecule of
2-ketoglutarate to PII exerts negative co-operativity on
the binding of additional effector molecules. PII trimers
containing a single bound 2-ketoglutarate interact optimally with NRII and ATase (Kamberov et al ., 1995; Liu
and Magasanik, 1995; Jiang et al ., 1998b,c). When the
negative co-operativity is overcome at elevated concentrations of 2-ketoglutarate, the ability of PII to interact with
NRII and ATase is diminished (Kamberov et al ., 1995;
Liu and Magasanik, 1995; Jiang et al ., 1998a,b,c). Thus,
growth in the presence of a good carbon source, resulting
in a high intracellular concentration of 2-ketoglutarate, prevents the dephosphorylation (inactivation) of NRI⬃P and
the adenylylation (inactivation) of GS.
Recent work has indicated that E. coli and several other
bacteria contain multiple versions (paralogues) of the PII
protein. In E. coli , GlnK, product of glnK , is a PII paralogue
that plays a role in nitrogen regulation (Allikmets et al .,
1993; van Heeswijk et al ., 1995; 1996; Atkinson and Ninfa,

1998). Unlike PII, which is present at a constant level
under all conditions, the expression of glnK is nitrogen
regulated such that essentially no GlnK is found in nitrogen-replete cells, and a high intracellular concentration
of GlnK is found in nitrogen-starved cells (van Heeswijk
et al ., 1996; Atkinson and Ninfa, 1998). Experiments with
a lacZ transcriptional (operon) fusion to the glnK promoter
indicate that this promoter is strongly expressed in nitrogenstarved cells (Atkinson and Ninfa, 1998). Under starvation
conditions, GlnK may contribute to nitrogen regulation by
controlling both ATase and NRII (Atkinson and Ninfa,
1998). Indeed, E. coli has a severe growth defect on minimal medium in the absence of exogenous amino acids if
it lacks both GlnK and PII because, in such cells, the high
expression of the Ntr regulon apparently results in the
degradation of several amino acids (Atkinson and Ninfa,
1998). This high expression of the Ntr regulon is caused
by the unregulated activity of NRII (Atkinson and Ninfa,
1998).
Genetic analysis of glnK indicated that, in cells containing PII, the loss of GlnK had little effect on the expression
of the nitrogen-regulated glnAp2 promoter and only a modest effect on the regulation of the glnK promoter (Atkinson
and Ninfa, 1998). Also, the loss of GlnK had little effect on
the regulation of the GS adenylylation state, as long as PII
was present (Atkinson and Ninfa, 1998). Thus, PII plays the
dominant role in the regulation of these phenotypes (Bueno
et al ., 1985). However, in cells deleted for PII, GlnK was
responsible for regulation of these phenotypes, as already
noted. In cells lacking PII, the regulation of the adenylylation state of GS by ammonia was wild type and required
the presence of GlnK (Atkinson and Ninfa, 1998). In contrast, in cells lacking PII, the regulation of glnA expression
by ammonia is eliminated, and the regulation of glnK
expression by ammonia is defective. The residual regulation of glnK expression by ammonia is caused by GlnK
(Atkinson and Ninfa, 1998). These observations suggested
that GlnK could interact well with the ATase, but was not as
effective as PII in activating the NRII phosphatase activity.
Nevertheless, GlnK is required for the regulation of NRII
in the absence of PII.
The regulation of GlnK activity by the UTase/UR is
indicated by several observations. In the absence of the
UTase/UR, but not in its presence, GlnK prevents the
expression of some Ntr genes that apparently require a
high level of NRI⬃P for expression (Atkinson and Ninfa,
1998). Also, the uridylylation of GlnK by the UTase/UR
has been observed directly in vivo (van Heeswijk et al .,
1996).
Cells lacking PII and the UTase/UR cannot fully activate
the Ntr regulon, because of the presence of unmodified
GlnK, which elicits enough of the NRII phosphatase activity
to maintain the concentration of NRI⬃P below the threshold needed for activation of some Ntr genes. At least one
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

E. coli GlnK protein 303
of the genes required for the use of arginine as a nitrogen
source is not expressed in such cells, rendering the cells
unable to grow when arginine is the sole nitrogen source.
In contrast, the triply null glnB glnD glnK strain lacking
UTase/UR, PII and GlnK grows poorly on arginine as the
sole nitrogen source, because of its inability to prevent
degradation of amino acids, but this growth is considerably
better than that of its glnK þ parent. Spontaneous mutations in glnK that permitted growth on arginine were isolated in the doubly mutant glnB glnD strain (Atkinson and
Ninfa, 1998). Some of the glnK mutants obtained had
growth properties suggesting that the glnK alleles were
null, while other glnK mutants had properties that suggested that the altered GlnK was moderately defective in
the interaction with NRII, and interacted normally with the
UTase/UR and ATase.
We purified wild-type GlnK, the GlnKY51N mutant protein, which appeared to be null or nearly so in cells (Experimental procedures ), and the GlnKR47W mutant protein,
which appeared to be specifically defective in the interaction
with NRII in cells (Atkinson and Ninfa, 1998), and compared
their activities with those of PII using purified components.
Surprisingly, GlnK was a potent activator of the phosphatase activity of NRII (about half as active as PII) but was a
very poor activator of the adenylylation of glutamine synthetase by ATase (about 1/40th as active as PII). The GlnK
protein was readily uridylylated by the UTase/UR. However, the deuridylylation of GlnK⬃UMP by the uridylylremoving activity of the UTase/UR was very slow (about
1/10th the rate of PII⬃UMP deuridylylation). Thus, when
reconstituted with the UTase/UR in a monocyclic system,
the PII uridylylation state was effectively regulated by glutamine, as shown before, but GlnK uridylylation, while slowed
by glutamine, did not reach a steady state. Thus, GlnK has
unique properties that may illustrate its role in nitrogen regulation. We also observed that the GlnKR47W protein was
specifically defective in the interaction with NRII in vitro , as
predicted (Atkinson and Ninfa, 1998), while the GlnKY51N
protein was essentially non-functional in vitro .

Results

Purification of GlnK, GlnKR47W and GlnKY51N
We engineered the overexpression of the desired glnK
alleles and purified their products using the methods used
previously for PII (Kamberov et al ., 1994). A second preparation of GlnK was also used in our studies, prepared
identically except for the last chromatography step (Experimental procedures ). In each case, the purified proteins
were about 90% or more pure, as estimated by SDS–
PAGE (data not shown). As one of the purification steps
involved gel filtration chromatography, we could readily
observe that, like PII, GlnK and the mutant GlnK proteins
appeared to be trimers (data not shown). The trimeric
nature of GlnK was examined by examining the time course
of GlnK uridylylation by the purified UTase/UR using nondenaturing gel electrophoresis, as described previously
for PII (Atkinson et al ., 1994). This analysis also suggested
that GlnK was a trimer, as three distinct uridylylated species
could be observed corresponding to GlnK molecules containing one, two or three uridylyl groups (data not shown).
While our work was under way, the structure of the trimeric
GlnK was solved (Xu et al ., 1998)

Interaction of GlnK, GlnKR47W and GlnKY51N
with NRII
To assess the interaction of GlnK with NRII, we measured
the rate of NRI⬃P dephosphorylation in reaction mixtures
containing NRI, NRII and ATP, as described previously
(Fig. 1; Kamberov et al ., 1994; 1995). During an initial
phase of the reaction, NRI, present in excess, was phosphorylated by NRII and ATP. The rate of NRI⬃P dephosphorylation was then assessed after the addition of PII,
GlnK or a buffer control. Previous results have shown that
the rate of NRI⬃P dephosphorylation and the steadystate level of NRI phosphorylation in this assay is dependent on the PII concentration and on the 2-ketoglutarate
concentration (Kamberov et al ., 1994; 1995). In Fig. 1A,
Fig. 1. Effect of PII, GlnK and GlnK * proteins on
the dephosphorylation of NRI⬃P by NRII.
Reactions were incubated at 25⬚C and contained
100 mM Tris-Cl, pH 7.5, 100 mM KCl, 25 mM
MgCl 2 , 300 ␮M 2-ketoglutarate, 300 ␮g ml¹1 BSA,
600 ␮M [␥-32P]-ATP, 18 ␮M NRI and 0.3 ␮M NRII.
Reactions were incubated for 20 min to permit the
phosphorylation of NRI by NRII, after which PII,
GlnK or a GlnK * protein was added.
A. Effect of addition of PII, GlnK or GlnK *
proteins to 0.3 ␮M. Symbols are: (þ), PII; (*),
GlnK; (filled square), GlnKR47W; (X), GlnKY51N;
(filled circle), storage buffer control.
B. Effect of addition of PII, GlnK or GlnK *
proteins to 3.0 ␮M. Symbols are as in (A).

䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

304 M. R. Atkinson and A. J. Ninfa
typical results are shown for experiments with 2-ketoglutarate at the intermediate concentration of 300 ␮M and
PII or GlnK at 0.3 ␮M, a limiting concentration for PII. As
shown, PII brought about the rapid dephosphorylation of
NRI⬃P, and GlnK clearly had activity, but was less effective. When present at this concentration, GlnKR47W and
GlnKY51N did not have significant activity. In contrast, in
Fig. 1B, typical results are shown for experiments at
300 ␮M 2-ketoglutarate and 3 ␮M PII or GlnK, which is
a saturating concentration for PII. At this concentration,
both PII and GlnK brought about the very rapid dephosphorylation of NRI⬃P, GlnKR47W was clearly less effective and GlnKY51N seemed to lack activity.
As elevated 2-ketoglutarate diminishes the interaction of
PII with NRII (Liu and Magasanik, 1995; Kamberov et al .,
1995; Jiang et al ., 1998b), we examined the activity of GlnK
at various 2-ketoglutarate concentrations. A low concentration of 2-ketoglutarate stimulated the ability of GlnK to
activate the NRII phosphatase, whereas elevated 2-ketoglutarate inhibited the activity of GlnK (Fig. 2A, curves with
symbols X, filled diamonds, filled triangles and house
show the effects of 50 ␮M, 150 ␮M, 500 ␮M and 2 mM
2-ketoglutarate respectively). Under the conditions used,
the inhibition of GlnK activation of the NRII phosphatase
was more dramatic than that observed with PII (Fig. 2B).
These results suggest that, like PII, GlnK is allosterically
regulated by 2-ketoglutarate.

Activation of the adenylylation activity of ATase by
PII, GlnK, GlnKR47W and GlnKY51N
Previous results have indicated that PII can activate the
adenylylation of GS by ATase, even in the absence of

2-ketoglutarate (Brown et al ., 1971; Jiang et al ., 1997).
For PII, this activity is greatly increased at low concentrations of 2-ketoglutarate and is prevented by 2-ketoglutarate
at high concentrations (Jiang et al ., 1998c). Also, the activation of ATase by PII (and 2-ketoglutarate) is synergistic
with glutamine. At high glutamine concentrations, a low concentration of PII is able to activate the enzyme, while at low
glutamine concentrations, a high concentration of PII is
required to activate the ATase (P. Jiang and A. J. Ninfa,
unpublished).
We examined first the activation of ATase by PII, GlnK,
GlnKR47W and GlnKY51N in the absence of 2-ketoglutarate (Fig. 3). Under these conditions, PII was clearly the
most potent activator, GlnKR47W and GlnK had clearly
detectable activity and GlnKY51N appeared to lack activity
(Fig. 3). Interestingly, GlnKR47W was a better activator of
ATase then was wild-type GlnK (Fig. 3).
The ability of GlnK to activate the ATase in the presence
of glutamine (6 mM) and 2-ketoglutarate (25 ␮M) was examined and compared with that of PII. Under these conditions,
PII is a potent activator of ATase (Jiang et al ., 1998c). When
the initial rate of GS adenylylation was examined, 20 ␮M
GlnK was required to attain an initial rate of GS adenylylation similar to that obtained with 0.5 ␮M PII (data not shown).
Thus, while GlnK could activate ATase under these conditions, it was about 40-fold less potent than PII.
The ability of purified GlnK to activate the adenylylation
of GS by ATase has previously been studied indirectly by
measurement of the activity of GS after treatment with the
ATase ⫾ GlnK or PII (van Heeswijk et al ., 1996). In that
study, GlnK was also observed to be approximately 1/40th
as active as PII in activating the adenylylation of GS by
ATase.

Fig. 2. Effect of 2-ketoglutarate concentration on the activation of the NRII-catalysed dephosphorylation of NRI⬃P by PII and GlnK. Reaction
conditions were as in Fig. 2, except that 2-ketoglutarate was varied as indicated below.
A. The symbols are: (X), 1 ␮M GlnK and 50 ␮M 2-ketoglutarate; (filled diamond), 1 ␮M GlnK and 150 ␮M 2-ketoglutarate; (filled triangle), 1 ␮M
GlnK and 500 ␮M 2-ketoglutarate; (house), 1 ␮M GlnK and 2 mM 2-ketoglutarate; (filled square) 1 ␮M GlnK and 0 2-ketoglutarate; (asterisk),
300 nM PII and 50 ␮M 2-ketoglutarate; (þ) 300 nM PII and 0 2-ketoglutarate; (filled circle), storage buffer control.
B. The symbols are: (filled diamond), 300 nM PII and 550 ␮M 2-ketoglutarate; (filled square), 300 nM PII and 850 ␮M 2-ketoglutarate; (þ)
300 nM PII and 2 mM 2-ketoglutarate; (filled triangle), 300 nM GlnK and 500 ␮M 2-ketoglutarate; (X) 300 nM GlnK and 850 ␮M 2-ketoglutarate;
(asterisk), 300 nM GlnK and 2 mM 2-ketoglutarate; (filled circle), storage buffer control.
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

E. coli GlnK protein 305
Heeswijk et al ., 1996; Atkinson and Ninfa, 1998). We
examined whether the regulation of ATase by glutamine,
in synergy with GlnK, could explain this phenomenon. Reaction mixtures initially contained ATase, GS and either GlnK,
a mixture of GlnK and GlnK⬃UMP, or a buffer control. In all
cases, negligible adenylylation of GS was observed (Fig.
5). Glutamine (6 mM) was then added to each reaction
mixture; this addition resulted in increased adenylylation
of GS in all cases, and the rate of GS adenylylation was
greatly increased by GlnK and by the combination of GlnK
and GlnK⬃UMP (Fig. 5). This experiment suggests that
synergy between GlnK and glutamine in the activation of
ATase, or the activation of ATase by glutamine alone,
may explain the earlier results obtained with intact cells.

Uridylylation and deuridylylation of PII, GlnK,
GlnKR47W and GlnKY51N

Fig. 3. Activation of GS adenylylation by PII, GlnK and the GlnK *
proteins. Reactions were incubated at 30⬚C and contained 100 mM
Tris-Cl, pH 7.5, 25 mM MgCl 2 , 300 ␮g ml¹1 BSA, 5 mM glutamine,
500 ␮M [␣-32P]-ATP, 10 nM ATase and 1 ␮M GS (dodecamers).
The symbols are: (þ) 400 nM PII; (filled square) 400 nM
GlnKR47W; (asterisk) 400 nM GlnK; (X) 400 nM GlnKY51N; (filled
circle), storage buffer control.

The ability of PII to activate the adenylylation activity of
ATase is stimulated by very low concentrations of 2-ketoglutarate and is inhibited by higher concentrations of 2ketoglutarate (Jiang et al ., 1998c). Previous work has
established that all 2-ketoglutarate effects on the PII-activated ATase adenylylation reaction are caused by the
binding of 2-ketoglutarate to PII (Jiang et al ., 1998c). We
therefore examined whether GlnK was similarly allosterically regulated by 2-ketoglutarate.
The adenylylation of GS by ATase was examined in
reaction mixtures containing saturating glutamine (6 mM),
fixed GlnK (25 ␮M) and various concentrations of 2-ketoglutarate (Fig. 4). The highest rate of GS adenylylation
was obtained in reaction mixtures lacking 2-ketoglutarate
and, at elevated concentrations of 2-ketoglutarate, the adenylylation of GS was inhibited considerably (Fig. 4). Thus,
GlnK seems to be regulated allosterically by 2-ketoglutarate, like PII, but the activation of adenylylation at low concentrations of effector that was observed previously with
PII was not observed with GlnK.
Experiments with intact cells lacking PII have shown
that GlnK can mediate the regulation of the GS adenylylation state (Atkinson and Ninfa, 1998). In starved cells lacking PII but containing GlnK, the GS adenylylation state is
low when nitrogen is limiting, and is rapidly increased
if the cells are shifted to medium containing ammonia (van
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

Previous results have indicated that the uridylylation of PII
by the UTase/UR requires 2-ketoglutarate at low concentrations, corresponding to the concentration resulting in the
binding of a single molecule of this effector to PII, and was
not greatly inhibited by elevated concentrations of this effector (Kamberov et al ., 1995; Jiang et al ., 1998a). The rate of
uridylylation was examined in experiments in which PII
or GlnK was present in large excess (10 ␮M), enzyme
was limiting (15 nM) and 2-ketoglutarate was present at
300 ␮M. Under these conditions, the rate of PII uridylylation was linear for 4 min (Fig. 6, þ symbols), and glutamine
at the intermediate concentration of 125 ␮M brought about
a significant decrease in the rate of PII uridylylation (Fig. 6,
× symbols). GlnK was uridylylated nearly as well as PII
(Fig. 6, squares), and the rate of GlnK uridylylation was
also diminished in the presence of glutamine (Fig. 6, circles).

Fig. 4. Effect of 2-ketoglutarate concentration on the stimulation of
GS adenylylation by GlnK. Initial rates were determined in reactions
that contained 100 mM Tris-Cl, pH 7.5, 100 mM KCl, 1 mM DTT,
10 mM KPi, 25 mM MgCl 2 , 300 ␮g ml¹1 BSA, 6 mM glutamine,
500 ␮M [␣-32P]-ATP, 50 nM ATase, 1 ␮M GS (dodecamers), 25 ␮M
GlnK and various concentrations of 2-ketoglutarate.

306 M. R. Atkinson and A. J. Ninfa

Fig. 5. Synergistic stimulation of the adenylylation activity of ATase
by GlnK and glutamine. Reactions were performed at 30⬚C and
contained 100 mM Tris-Cl, pH 7.5, 100 mM KCl, 25 mM MgCl 2 ,
1 mM DTT, 300 ␮g ml¹1 BSA, 100 ␮M 2-ketoglutarate, 10 mM KPi,
1 ␮M GS (dodecamers) and 500 ␮M [␣-32P]-ATP. GlnK or both
GlnK and GlnK⬃UMP were present from the beginning of the
experiment, as indicated. After 10 min of incubation, glutamine was
added to the reaction mixtures to 6 mM. Symbols are: (þ) no GlnK
or GlnK⬃UMP added; (asterisk), 25 ␮M GlnK and 10 ␮M
GlnK⬃UMP added; (filled circle), 25 ␮M GlnK added.

Interestingly, in the presence of glutamine, GlnK was uridylylated slightly faster than PII. The uridylylation of
GlnKR47W (Fig. 6, triangles) was slower than that of GlnK
and was also decreased by glutamine (Fig. 6, wedges).
We could not detect the uridylylation of GlnKY51N (data

not shown), as expected, as position 51 is the site of PII
uridylylation (Jiang et al., 1997). In additional experiments,
we observed that the uridylylation of GlnK and GlnKR47W
required 2-ketoglutarate, and PII uridylylation was barely
detectable in the absence of 2-ketoglutarate (data not
shown).
The K act for 2-ketoglutarate activation of uridylylation
was examined for PII and GlnK. Previous results had indicated that, for PII, the 2-ketoglutarate K act corresponded
well to the K d for the dissociation of 2-ketoglutarate from
PII bound to a single molecule of this effector (Jiang
et al ., 1998a). The K act for PII uridylylation was about
4 ␮M, in good agreement with earlier results (Jiang et al .,
1998a), whereas that for GlnK uridylylation was about
7–15 ␮M 2-ketoglutarate (Fig. 7).
To examine the deuridylylation of GlnK⬃UMP by the
UR activity of the UTase/UR, we purified PII⬃UMP and
GlnK⬃UMP as described previously (Jiang et al .,
1998a). Under conditions of saturating glutamine (6 mM)
and 100 ␮M 2-ketoglutarate, the initial rate of deuridylylation of GlnK⬃UMP (Fig. 8, wedge) was about 10-fold
slower than the rate of PII⬃UMP deuridylylation (Fig. 8,
þ). At very high enzyme concentrations, GlnK⬃UMP
could be slowly deuridylylated (Fig. 8, filled wedge). In
additional experiments, we examined the effect of variation in the 2-ketoglutarate concentration on the rate of
GlnK⬃UMP deuridylylation. We observed that, as shown
earlier for PII⬃UMP deuridylylation (Jiang et al ., 1998a),
100 ␮M 2-ketoglutarate was nearly saturating, and the
rate of GlnK⬃UMP deuridylylation was not significantly
increased at higher 2-ketoglutarate concentrations (data
not shown). These experiments demonstrate that the
deuridylylation of GlnK⬃UMP was very slow under conditions in which PII⬃UMP was readily deuridylylated.
Fig. 6. Uridylylation of PII, GlnK and GlnKR47 W
by the UTase activity of the UTase/UR, and
inhibition of uridylylation by glutamine. Reactions
were incubated at 30⬚C and contained 100 mM
Tris-Cl, pH 7.5, 100 mM KCl, 25 mM MgCl 2 ,
1 mM DTT, 300 ␮M 2-ketoglutarate, 500 ␮M ATP,
500 ␮M [␣-32P]-UTP and 15 nM UTase/UR.
Symbols are: (þ), 10 ␮M PII; (square), 10 ␮M
GlnK; (triangle), 10 ␮M GlnKR47W; (circle),
10 ␮M GlnK þ 125 ␮M glutamine; (X), 10 ␮M
PII þ 125 ␮M glutamine; (wedge), 10 ␮M
GlnKR47W þ 125 ␮M glutamine.

䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

E. coli GlnK protein 307

Fig. 7. K act for 2-ketoglutarate activation of uridylylation of PII and
GlnK. Initial rates were determined in reactions performed at 30⬚C
containing 100 mM Tris-Cl, pH 7.5, 100 mM KCl, 500 ␮M ATP,
25 mM MgCl 2 , 1 mM DTT, 300 ␮g ml¹1 BSA, 400 ␮M [␣-32P]-UTP,
10 nM UTase/UR, either 15 ␮M PII or 15 ␮M GlnK and the
indicated 2-ketoglutarate concentrations. Symbols are: (filled circle),
GlnK; (open circle), PII.

In additional experiments, we examined whether GlnK⬃
UMP inhibited the deuridylylation of PII⬃UMP by the UR
activity of the UTase/UR. For these experiments, PII⬃UMP
was present at 2 ␮M, which corresponds approximately to
the K m for this reaction (Jiang et al ., 1998a). Under these
conditions, GlnK⬃UMP was a poor inhibitor of PII⬃UMP
deuridylylation (data not shown). We could not add a
high enough concentration of GlnK⬃UMP to reaction
mixtures to obtain 50% inhibition of the rate of PII⬃UMP
deuridylylation (data not shown), therefore K i must be
very high.

Reconstitution of the UTase/UR-GlnK monocycle
The regulation of the steady-state level of PII and GlnK uridylylation by glutamine was examined in reaction mixtures

that initially contained unmodified PII or GlnK at 1.25 ␮M,
UTase/UR at 0.1 ␮M and 2-ketoglutarate at 200 ␮M. Previous studies have shown that, under these conditions, PII
uridylylation rapidly reaches a steady-state level that is
dependent on the glutamine concentration (Jiang et al .,
1998b); as a control, these experiments were repeated
(Fig. 9A). In contrast, GlnK uridylylation state never reached
a steady state under similar conditions, unless GlnK was
completely modified (Fig. 9B). While glutamine slowed
the rate of GlnK uridylylation, eventually GlnK became
highly uridylylated even in the presence of elevated glutamine concentrations.
Discussion
Biochemical characterization of purified GlnK cannot prove
the function of GlnK in intact cells, but may serve to limit
the possibilities. The biochemical studies presented here
and the genetic studies presented previously (Atkinson
and Ninfa, 1998) indicate that PII and GlnK have distinct
properties, which probably reflect their unique roles in
regulating nitrogen assimilation. PII, which is present at a
constant, relatively low concentration, was a potent activator of NRII and ATase, and its interaction with the UTase/
UR permitted efficient regulation of its uridylylation state
by glutamine. As the interaction of PII with the NRII and
ATase receptors is regulated allosterically by 2-ketoglutarate, PII is ideally suited to integrate information on the
carbon and nitrogen status of the cell and to regulate
these receptors accordingly (Jiang et al ., 1998b,c).
GlnK, on the other hand, is not present at a constant
concentration in the cell; the activation of glnK transcription by NRI⬃P places the appearance of GlnK under the
control of PII and, in starved cells, under the control of
Fig. 8. Deuridylylation of PII⬃UMP and
GlnK⬃UMP by the UR activity of the UTase/UR.
Reactions were performed at 30⬚C and contained
100 mM Tris-Cl, pH 7.5, 100 mM KCl, 25 mM
MgCl 2 , 1 mM DTT, 100 ␮M 2-ketoglutarate,
500 ␮M ATP, 300 ␮g ml¹1 BSA and either 3.2 ␮M
[32P]-GlnK⬃UMP or [32P]-PII⬃UMP. The
symbols are: (circle), GlnK⬃UMP, 6 mM
glutamine, no enzyme; (square), PII⬃UMP, 6 mM
glutamine, no enzyme; (triangle), GlnK⬃UMP,
1 ␮M UTase/UR; (filled square), GlnK⬃UMP,
3 ␮M UTase/UR; (wedge), GlnK⬃UMP, 6 mM
glutamine, 1 ␮M UTase/UR; (filled triangle),
GlnK⬃UMP, 5 ␮M UTase/UR; (filled circle),
GlnK⬃UMP, 6 mM glutamine, 3 ␮M UTase/UR;
(X), PII⬃UMP, 1 ␮M Utase/UR; (filled wedge)
GlnK⬃UMP, 6 mM glutamine, 5 ␮M UTase/UR;
(þ), PII⬃UMP, 6 mM glutamine, 1 ␮M UTase/UR.

䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

308 M. R. Atkinson and A. J. Ninfa
Fig. 9. Regulation of uridylylation state by
glutamine in reconstituted PII-UTase/UR and
GlnK-UTase/UR monocycles. All reactions were
performed at 30⬚C in 100 mM Tris-Cl, pH 7.5,
100 mM KCl, 25 mM MgCl 2 , 500 ␮M ATP, 200 ␮M
2-ketoglutarate, 1 mM DTT, 500 ␮M [␣-32P]-UTP,
300 ␮g ml¹1 BSA and 100 nM UTase/UR.
A. 1.25 ␮M PII was used.
B. 1.25 ␮M GlnK was used.
Symbols are: (filled diamonds) 4 mM glutamine;
(X) 1.2 mM glutamine; (filled squares), 0.8 mM
glutamine; (asterisk), 0.5 mM glutamine; (þ),
0.25 mM glutamine; (filled circle) 0 glutamine.

GlnK (van Heeswijk et al ., 1996; Atkinson and Ninfa,
1998). As PII is a potent activator of the NRII phosphatase
activity, GlnK probably never accumulates in the cell until
most of the PII has been converted to PII⬃UMP. Previous
results with a glnK::lacZ transcriptional fusion suggested
that, upon nitrogen starvation, the glnK gene is strongly
transcribed (Atkinson and Ninfa, 1998). As the intracellular
concentration of glutamine is low under these conditions,
there will be little inhibition of the transferase activity
of the UTase/UR, and GlnK will become uridylylated.
However, the presence of both GlnK and GlnK⬃UMP in
nitrogen-starved cells has been observed in two separate
studies (van Heeswijk et al ., 1996; He et al ., 1998). Thus, it
is possible that, when the rate of GlnK synthesis becomes
sufficiently high, the fixed level of UTase activity cannot
keep up, and both GlnK and GlnK⬃UMP will be present.
Indeed, the ratio of GlnK to GlnK⬃UMP may increase to
a fixed maximum as cells become starved.
As GlnK is only present in nitrogen-starved cells, it must
provide some function beneficial for such cells. The presence of unmodified GlnK in starved cells would ensure
that there is always some NRII phosphatase activity. We
observed that GlnK was a potent activator of the NRII phosphatase activity (Fig. 1). One hypothesis for the role of GlnK
is that it contributes to the regulation of NRII under nitrogenlimiting conditions. Consistent with this hypothesis, previous
genetic studies have indicated that, in the absence of both
PII and GlnK, the presence of NRII resulted in unrestrained
Ntr gene expression, which can be catastrophic under
certain conditions (Atkinson and Ninfa, 1998). Also, in
the absence of PII, GlnK clearly acts through NRII to regulate its own promoter and other Ntr promoters (Atkinson
and Ninfa, 1998). However, there is a problem with this
hypothesis. Previous genetic studies failed to identify
growth conditions in which GlnK was of benefit to cells
containing PII (Atkinson and Ninfa, 1998). If PII is entirely

converted to PII⬃UMP in starved cells, GlnK should be
required to maintain control of NRII. Yet, cells containing
PII and lacking GlnK survive starvation and the shift to
nitrogen-limiting conditions as well as do isogenic cells
containing GlnK (Atkinson and Ninfa, 1998). Thus, while
not necessary for the regulation of NRII in cells containing
PII, GlnK may have a subtle role in fine-tuning NRII activity
in starved cells.
GlnK was about 40-fold less potent than PII in activating
the adenylylation of GS by ATase. This result is consistent
with the earlier observation that purified GlnK was ineffective in bringing about the inactivation of GS by ATase (van
Heeswijk et al ., 1996). While other explanations cannot be
ruled out, these observations suggest that GlnK may not
play a significant role in regulating ATase in starved cells
containing PII. As PII⬃UMP is present in such cells, a possible role for GlnK would be to facilitate the adenylylation
of GS if ammonia were encountered. We observed that,
like PII, GlnK can act synergistically with glutamine to activate the adenylylation of GS by ATase (Fig. 5). However,
previous experiments have indicated that the rate of GS
adenylylation in response to ammonia shock of starved
cells (containing PII) is the same in the presence and in
the absence of GlnK (Atkinson and Ninfa, 1998). Apparently, in cells lacking GlnK, the rapid deuridylylation of
PII⬃UMP in response to ammonia shock of starved cells
is adequate to ensure the rapid adenylylation of GS (Atkinson and Ninfa, 1998). Alternatively, the activation of ATase
by glutamine may be sufficient to account for the rapid
adenylylation of GS under these conditions (Fig. 5).
Experiments with intact cells indicated that, in the
absence of PII, GlnK was required for the regulation of the
adenylylation state of GS (Atkinson and Ninfa, 1998). In
such cells, the GlnK requirement is probably the result of
a requirement for GlnK⬃UMP for activation of the deadenylylation of GS⬃AMP by ATase. The observation that cells
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

E. coli GlnK protein 309
lacking PII could adenylylate GS rapidly in response to
ammonia shock led us to the conclusion that the GlnK
uridylylation state was rapidly regulated by the UTase/UR
(Atkinson and Ninfa, 1998). However, the current study
with purified components indicated that the uridylylation
state of GlnK was not rapidly regulated by UTase/UR;
rather, the deuridylylation of GlnK⬃UMP was very slow
(Fig. 8). Thus, the rapid deuridylylation of GlnK⬃UMP
cannot account for the ability of starved cells lacking PII
to adenylylate GS rapidly in response to ammonia shock.
In our experiments, GlnK⬃UMP was deuridylylated
slowly by the UR activity of the UTase/UR, and GlnK⬃
UMP was a poor inhibitor of PII⬃UMP deuridylylation.
This may be advantageous to the cell. If a starved wildtype cell (containing GlnK, GlnK⬃UMP and PII⬃ UMP)
suddenly encounters ammonia, the internal glutamine concentration will rise rapidly. The UTase activity will become
inhibited, and the UR activity will become activated. It may
then be advantageous for the cell to deuridylylate PII⬃
UMP preferentially, providing the highly potent PII protein
to activate the NRII phosphatase activity and ATase adenylylation activity. The difference in potency between PII and
GlnK and the fact that PII is synthesized constitutively
ensures that, once PII⬃UMP is deuridylylated, the resulting PII dominates nitrogen regulation.

Allosteric regulation of GlnK activity by 2-ketoglutarate
Earlier studies with reconstituted systems have indicated
that PII is regulated allosterically by 2-ketoglutarate, with
different consequences for each receptor interaction. A
low concentration of this effector stimulates the interaction
of PII with each of its receptors. However, at higher effector concentrations, the interaction with ATase is strongly
inhibited, the interaction with NRII is inhibited to a lesser
extent and the interaction with the UTase/UR is not significantly inhibited (Jiang et al ., 1998a,b,c). Apparently, PII
adopts distinct conformations, depending on its degree of
saturation with 2-ketoglutarate.
Our previous experiments with PII suggested that the
K act for the activation of PII uridylylation by 2-ketoglutarate
corresponded to the K d for the binding of a single effector
molecule to the PII trimer (⬇5 ␮M; Kamberov et al ., 1995;
Jiang et al ., 1998a; this study). Here, we observed that the
activation of GlnK uridylylation occurred at approximately
7–15 ␮M, suggesting that GlnK also binds an effector
molecule with high affinity.
The activation of NRII and ATase by GlnK was inhibited
by high concentrations of 2-ketoglutarate, similar to earlier
observations with PII (Liu and Magasanik, 1995; Kamberov
et al ., 1995; Jiang et al ., 1998a). Therefore, it seems that,
like PII, the binding of additional effector molecules at elevated effector concentrations results in another conformation that does not interact well with NRII and ATase.
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

However, there was a clear difference in the activation of
the ATase by PII and GlnK. In this reaction, low concentrations of 2-ketoglutarate resulted in significant stimulation of
GS adenylylation by PII (Jiang et al ., 1998c), but not by
GlnK. Thus, the binding of the first molecule of 2-ketoglutarate had different effects on the conformation of PII and
GlnK, as reported by the ATase.

An altered form of GlnK specifically defective in the
interaction with NRII
Experiments with PII have indicated that the apex of the
T-loop (Carr et al ., 1996) is essential for the interaction
of PII with its three protein receptors (UTase/UR, NRII
and ATase), and that mutations in this part of PII may
have different effects on each of these interactions (Jiang
et al ., 1997). For instance, a PII mutation, A49P, resulted
in an altered protein specifically defective in the interaction
with NRII (Jiang et al ., 1997). Our earlier genetic experiments suggested that glnK mutations resulting in the conversion of R-47 to a number of other amino acids may also
result in proteins specifically defective in the interaction
with NRII (Atkinson and Ninfa, 1998). Here, one such
GlnK protein containing the alteration R47W was observed
to be specifically defective in the interaction with NRII in
experiments with purified components. This protein was
uridylylated by the UTase/UR nearly as well as wild-type
GlnK and activated the ATase more effectively than wildtype GlnK, but was defective in the activation of NRII.
Thus, as proposed from genetic experiments (Atkinson
and Ninfa, 1998), R-47 of GlnK seems to be very important for the interaction of GlnK with NRII.
In our experiments, the GlnK-Y51N protein appeared
essentially to lack the ability to activate NRII and ATase.
This was somewhat surprising, as earlier studies with this
protein in intact cells indicated that it was able to bring
about the relief of NifL-mediated inhibition of NifA (He et
al ., 1998; see below).

In addition to its role in the regulation of NRII and
ATase, GlnK may interact with additional receptors
The synthesis of nitrogenase in Klebsiella pneumoniae
requires the transcriptional activator NifA and is regulated
by the NifL signal transduction protein. Nitrogenase synthesis is repressed at conditions of nitrogen sufficiency,
reflecting the fact that the nifLA operon is an Ntr operon
requiring a high intracellular concentration of NRI⬃P for
transcription (Wong et al ., 1987). However, recent work
in which the nifL and nifA genes from K. pneumoniae were
introduced into E. coli along with a reporter gene fusion
to a NifA-activated gene (nifH::lacZ ) indicated that nitrogen starvation and transcription of the Ntr regulon was
required for relief of NifA from NifL-mediated inactivation,

310 M. R. Atkinson and A. J. Ninfa
even when the synthesis of NifL and NifA did not depend
on NRI⬃P (He et al ., 1997). This requirement is caused
by the need for the GlnK protein, which somehow prevents
the inactivation of NifA by NifL (He et al ., 1998). Interestingly, GlnK-Y51N was also able to bring about the activation
of NifA (He et al ., 1998). While E. coli does not typically
contain NifL and NifA, the regulation of these by GlnK raises
the possibility that E. coli contains additional unknown
receptors that are regulated by GlnK.

Experimental procedures

Bacteriological techniques
The bacterial strains and plasmids used in this study are summarized in Table 1. Rich LB þ 0.2% (w/v) glutamine medium
and W-salts minimal medium were prepared as described previously (Bueno et al ., 1985). Ampicillin, when present, was at
100 ␮g ml¹1. The preparation of competent cells and transformation of competent cells with plasmid DNA was performed
as described previously (Maniatis et al ., 1982) Strains containing pJLA503-based plasmids were transformed and maintained at 30⬚C, as before (Kamberov et al ., 1994).

Plasmid constructions
The overexpression of GlnK has been reported previously (van
Heeswijk et al ., 1996). However, analysis of the reported cloning method suggested that several amino acids from the lac␣
gene of the cloning vector had been added to the N-terminus
of GlnK. When we reconstructed the GlnK overexpression
plasmid as described previously, forming plasmid pKOP3, we
observed that induction resulted in the expression of a protein
that was significantly more massive than PII (data not shown).
Yet, uninduced cultures containing this plasmid were reported
to produce native-length GlnK (van Heeswijk et al ., 1996). To

Table 1. Bacterial strains and plasmids used in
this study.

simplify matters, pJLA503-based overexpression plasmids
pKOP2 (wild-type GlnK), pKOPR47W (glnK4 ) and pKOPY51N (glnK5 ) were constructed as follows. The desired glnK
alleles were obtained by polymerase chain reaction (PCR)
amplification of chromosomal DNA from strains YMC10, OB25
and OB133 (Atkinson and Ninfa, 1998), according to the
method of Saiki et al . (1990). The PCR primers used in all
three cases were: (upstream primer) 5⬘-CGAATTCCATATGAAGCTGGTGACCGTGAQTAATC and (downstream primer)
5⬘-CGGATCCGTCGACTTCCTGTTGCTGTGTGCCAGAG.
These primers introduce a unique Nde I site overlapping the
first ATG codon of glnK and a unique Bam HI site immediately
downstream from the glnK coding sequence. The PCR products were purified using a Qiagen PCR purification kit,
digested with Nde I and Bam HI, and cloned into similarly
cleaved pJLA503. The plasmids were sequenced using the
Oncor Fidelity DNA sequencing system to ensure that additional mutations were not introduced during PCR amplification. Plasmid DNA was prepared for DNA sequencing using
the Qiagen plasmid miniprep kit.
Plasmid pKOP3 is a reconstruction of pWVH149 (van Heeswijk et al ., 1996). It was constructed by cloning the Eco RI/ Eae I
fragment containing glnK þ from pDK4 (Atkinson and Ninfa,
1998) into Eco RI- and Not I-digested pBluescript-II KSþ
(Stratagene).

Characterization of the phenotype of strain OB113
(⌬glnB glnD99::Tn 10 glnK5/ pgln2; Atkinson and
Ninfa, 1998)
To determine whether the glnK5 allele was null, we compared the growth of strain OB113 (Atkinson and Ninfa,
1998) with the wild-type and control strains that were genetically identical (⌬glnB glnD99::Tn10 /pgln2), except containing
either glnK þ or ⌬mdl-glnK (Atkinson and Ninfa, 1998). Cells
were plated for single colonies on various solid media, and
growth was observed after 72 h at 37⬚C. Media used were:
glucose–ammonia–glutamine–tryptophan–Xgal (GNglnXtrp),

Strain

Relevant genotype

Source/reference

JM109
YMC10
BD
OB25
OB113
BDK

Wild type
Wild type
⌬glnB glnD99::Tn10
⌬glnB glnD99::Tn10glnK4 /pgln62
⌬glnB glnD99::Tn10glnK5 /pgln2
⌬glnB glnD99::Tn10⌬mdl-glnK::kan

Gibco BRL
Backman et al . (1981)
Atkinson and Ninfa (1998)
Atkinson and Ninfa (1998)
Atkinson and Ninfa (1998)
Atkinson and Ninfa (1998)

Plasmid

Property/construction

Source/reference

pJLA503
pKOP2
pKOPR47W
pKOPY51N
pKOP3

Hyperexpression vector
PCR glnK þ Nde I/ Bam HI into pJLA503
PCR glnK4 Nde I/ Bam HI into pJLA503
PCR glnK5 Nde I/ Bam HI into pJLA503
Reconstruction of pWVH149 (van
Heeswijk et al ., 1996) Eco RI/ Eae I of
pDK4 into Eco RI/ Not I cleaved
pBluescript-II KSþ
Bam HI of Kohara 149 into pUC18
Hyperexpression vector
glnALG operon (Hin dIII fragment) in
pBR322

Schauder et al . (1987)
This study
This study
This study
This study

pDK4
pBluescript-II KSþ
pgln2

Atkinson and Ninfa (1998)
Stratagene
Backman et al . (1981)

䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

E. coli GlnK protein 311
glucose–glutamine–tryptophan–Xgal (GglnXtrp), glucose–
ammonia (GN) and glucose–arginine (Garg), as described
previously (Atkinson and Ninfa, 1998). All media also contained 0.004% (w/v) vitamin B1.
On GNglnXtrp, GglnXtrp and GN media, the wild-type strain
YMC10 grew well, as did strains BD (⌬glnB glnD99::Tn10 )
and BD/pgln2 (⌬glnB glnD99::Tn10 /pgln2). In contrast, strains
BDK (⌬glnB glnD99::Tn10⌬mdl-glnK ) and BDK/pgln2 (⌬glnB
glnD99::Tn10⌬mdl-glnK/pgln2) only formed pinpoint colonies.
Strain OB113 formed colonies that were intermediate in size
between those formed by BD/pgln2 and BDK/pgln2. On
Garg medium, OB113 grew as well as the wild type, whereas
its parent BD/pgln2 (⌬glnB glnD99::Tn10 /pgln2) did not grow
at all. These results with intact cells indicate that, while the
glnK5 mutation (Y51N) results in a greatly altered GlnK protein,
this mutation is not null.

GlnK hyperexpression and purification
To check protein induction using pKOP3 and to determine
whether a fusion protein was programmed by this vector, cultures were grown overnight in LB þ glutamine þ ampicillin ⫾
1 mM IPTG at 37⬚C with vigorous aeration. Cells from a 1 ml
aliquot of each culture were harvested by centrifugation,
resuspended in 100 ␮l of 4 × cracking buffer/dye solution,
heated to 95⬚C for 15 min and subjected to electrophoresis
on a 15% SDS–polyacrylamide gel as described previously
(Kamberov et al ., 1994). After electrophoresis, the gel was
stained with Coomassie brilliant blue R250.
To check induction of pKOP2 in strain BD (Atkinson and
Ninfa, 1998), a 2 ml LB þ glutamine þ ampicillin culture was
grown overnight at 30⬚C with vigorous aeration. In the morning,
a 1 ml aliquot of the overnight culture was used to inoculate
10 ml of fresh media, and this culture was grown for 1 h, after
which it was split into two portions, which were grown at
30⬚C (non-inducing) and 43⬚C (inducing) for 3 h with vigorous
aeration. Cells from 1 ml aliquots of the uninduced and induced
cultures were harvested by centrifugation and analysed by
SDS–PAGE as described above.
To grow large-scale batch cultures of cells for protein purification, strain BD containing either pKOP2, pKOPR47W or
pKOPY51N was grown overnight in 2 ml of LB þ glutamine þ
ampicillin medium at 30⬚C with vigorous aeration. This overnight culture (1 ml) was expanded to 500 ml using the same
medium and growth conditions. After overnight growth, the
500 ml culture was divided equally into eight 2 l flasks, each
containing 500 ml of fresh medium, and the culture was incubated at 30⬚C for 1 h, after which the incubation was at 44⬚C
for 4 h with vigorous aeration. Cells were harvested by centrifugation, and cell pellets were stored at ¹80⬚C.
GlnK and GlnKR47W were purified using the method described previously for PII (Kamberov et al ., 1994), with modifications as noted below. The overall scheme included three
chromatography steps: DE52, Sephadex-G75 and hydroxylapatite, as before (Kamberov et al ., 1994). Ammonium sulphate fractionation, used previously for purification of PII
(Kamberov et al ., 1994), was not used. Instead, ammonium
sulphate precipitation at 50% saturation was used only to concentrate GlnK before loading the G75 gel filtration column,
and to concentrate GlnKR47W before the G75 and hydroxylapatite columns. In both cases, the pooled peak fractions from
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

the hydroxylapatite column were dialysed against storage
buffer (Kamberov et al ., 1994) and stored in aliquots at
¹80⬚C.
A second purified preparation of GlnK and purified
GlnKY51N was provided by Dr Quan Sun of our laboratory.
These were purified by the same method, except that the
final hydroxylapatite column was replaced by a phenylSepharose chromatography step. For this, ammonium sulphate was added to the pooled peak from the G75 column
to 25% saturation. This sample was loaded directly onto a
40 ml phenyl-Sepharose column (Pharmacia) equilibrated
in 50 mM Tris-Cl, pH 7.5, 1 mM EDTA, 100 mM KCl, 25%
saturation ammonium sulphate, and the column was washed
extensively with the same buffer. Under these conditions, GlnK
binds quantitatively to the column. The column was eluted
with a 250 ml gradient of ammonium sulphate from 25% to
0% saturation in the same buffer. GlnK and GlnKY51N eluted
near the end of the gradient. The purified peak fractions from
the phenyl-Sepharose column were pooled, dialysed against
storage buffer and stored in aliquots at ¹80⬚C, as above.
The phenyl-Sepharose step described above was also
shown to work well for the purification of PII (P. Jiang and
A. Pioszak, unpublished data). This step is superior to the
hydroxylapatite step described before (Kamberov et al .,
1994), as hydroxylapatite chromatography required dialysis
of the sample to remove salt that was present during the
prior gel filtration chromatography step. Use of phenyl-Sepharose in place of hydroxylapatite eliminates the need for a slow
dialysis step, while providing similar purification.
Protein concentrations were determined according to the
method of Lowry et al . (1951).

Purified UTase/UR, PII, ATase, GS, NRI and NRII
Preparations of these proteins have been described previously
(Jiang et al ., 1997; Kamberov et al ., 1994). Protein concentrations were determined according to the method of Lowry et al .
(1951).

Preparation of PII ⬃ UMP and GlnK ⬃ UMP
To prepare these uridylylated proteins, 500 ␮l reaction mixtures
containing 100 mM Tris-Cl, pH 7.5, 100 mM KCl, 25 mM MgCl 2 ,
500 ␮M ATP, 200 ␮M 2-ketoglutarate, 1 mM dithiothreitol
(DTT), 120 ␮M GlnK or PII and 500 ␮M [␣-32P]-UTP were incubated at 37⬚C for 30 min. Glycerol and KCl were then added
to 10% (v/v) and 350 mM final concentrations respectively.
The reaction mixtures were then heated to 60⬚C for 15 min
to inactivate the UTase/UR. Reaction mixtures were subjected
to chromatography on 20 ml Sephadex-G50 columns in 50 mM
Tris-Cl, pH 7.5, 10% glycerol, 1 mM DTT, 200 mM KCl to
separate protein from unincorporated UTP. Peak fractions
containing GlnK⬃UMP or PII⬃UMP were stored directly at
¹20⬚C.

Assays for the phosphorylation of NRI, adenylylation
of GS, uridylylation of PII or GlnK and deuridylylation
of PII ⬃ UMP or GlnK ⬃ UMP
These assays were carried out essentially as described

312 M. R. Atkinson and A. J. Ninfa
previously (Atkinson et al ., 1994; Kamberov et al ., 1994; Jiang
et al ., 1997) with modifications as noted in the figure legends.
For the ATase, UTase and UR assays, initial rates were determined after fitting of progress curves using Enzfit, as described
previously (Kamberov et al ., 1995; Jiang et al ., 1998a).

Acknowledgements
This work was supported by grant GM47460 from the NIH.
We thank Dr Quan Son for providing a preparation of purified
GlnK and the preparation of GlnKY51N used here, and Dr
W. C. van Heeswijk for reviewing an earlier version of this
paper.

References
Allikmets, R., Gerrard, B., Court, D., and Dean, M. (1993)
Cloning and organization of the abc and mdl genes of
Escherichia coli : relationship to eucaryotic multidrug resistance. Gene 136: 231–236.
Atkinson, M.R., and Ninfa, A.J. (1998) Role of the GlnK
signal transduction protein in the regulation of nitrogen
assimilation in Escherichia coli . Mol Microbiol 29: 431–
447.
Atkinson, M.R., Kamberov, E.S., Weiss, R.L., and Ninfa, A.J.
(1994) Reversible uridylylation of the Escherichia coli
PII signal transduction protein regulates its ability to stimulate the dephosphorylation of the transcription factor
nitrogen regulator I (NRI or NtrC). J Biol Chem 269:
28288–28293.
Backman, K., Chen, Y.-M., and Magasanik, B. (1981) Physical
and genetic characterization of the glnA-glnG region of the
Escherichia coli chromosome. Proc Natl Acad Sci USA 78:
3743–3747.
Brown, M.S., Segal, A., and Stadtman, E.R. (1971) Modulation of glutamine synthetase adenylylation and deadenylylation is mediated by metabolic transformation of the PIIregulatory protein. Proc Natl Acad Sci USA 68: 2949–
2953.
Bueno, R., Pahel, G., and Magasanik, B. (1985) Role of
glnB and glnD gene products in the regulation of the
glnALG operon of Escherichia coli . J Bacteriol 164: 816–
822.
Carr, P.D., Cheah, E., Suffolk, P.M., Vasudevan, S.G., Dixon,
N.E., and Ollis, D.L. (1996) X-ray structure of the signal
transduction protein PII from Escherichia coli at 1.9A. Acta
Cryst D52: 93–104.
Engleman, E.G., and Francis, S.H. (1978) Cascade control
of glutamine synthetase. II. Metabolic regulation of the
enzymes in the cascade. Arch Biochem Biophys 191:
602–619.
He, L., Soupene, E., and Kustu, S. (1997) NtrC is required for
control of Klebsiella pneumoniae NifL activity. J Bacteriol
179: 7446–7455.
He, L., Soupene, E., Ninfa, A.J., and Kustu, S. (1998) A
physiological role for the GlnK protein: relief from NifL inhibition under nitrogen-limiting conditions. J Bacteriol 180:
6661–6667.
van Heeswijk, W.C., Stegeman, B., Hoving, S., Molenaar, D.,
Kahn, D., and Westerhoff, H.V. (1995) An additional

PII in Escherichia coli : a new regulatory protein in the
glutamine synthetase cascade. FEMS Microbiol Lett 132:
153–157.
van Heeswijk, W.C., Hoving, S., Molenaar, D., Stegeman, B.,
Kahn, D., and Westerhoff, H.V. (1996) An alternative PII
protein in the regulation of glutamine synthetase in Escherichia coli . Mol Microbiol 21: 133–146.
Jiang, P., Atkinson, M.R., Kamberov, E.S., Zucker, P.,
Schefke, B.R., Chandran, P., et al . (1997) Structure/function analysis of the PII signal transduction protein of Escherichia coli . Genetic separation of the interactions with
receptors. J Bacteriol 179: 4342–4353.
Jiang, P., Peliska, J.A., and Ninfa, A.J. (1998a) 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.
Jiang, P., Peliska, J.A., and Ninfa, A.J. (1998b) Reconstitution of the signal-transduction bicyclic cascade responsible
for the regulation of Ntr gene transcription in Escherichia
coli . Biochemistry 37: 12795–12801.
Jiang, P., Peliska, J.A., and Ninfa, A.J. (1998c) The regulation
of Escherichia coli glutamine synthetase revisited: role of
2-ketoglutarate in the regulation of glutamine synthetase
adenylylation state. Biochemistry 37: 12802–12810.
Kamberov, E.S., Atkinson, M.R., Feng, J., Chandran, P.,
and Ninfa, A.J. (1994) Sensory components controlling
bacterial nitrogen assimilation. Cell Mol Biol Res 40:
175–191.
Kamberov, E.S., Atkinson, M.R., and Ninfa, A.J. (1995) The
Escherichia coli PII signal transduction protein is activated
upon binding 2-ketoglutarate and ATP. J Biol Chem 270:
17797–17807.
Keener, J., and Kustu, S. (1988) Protein kinase and phosphoprotein phosphatase activities of nitrogen regulatory proteins NTRB and NTRC of enteric bacteria: roles of the
conserved amino terminal domain of NTRC. Proc Natl
Acad Sci USA 85: 4976–4980.
Liu, J., and Magasanik, B. (1995) Activation of the dephosphorylation of nitrogen regulator I-phosphate of Escherichia coli . J Bacteriol 177: 926–931.
Lowry, O.H., Rosenbrough, N.J., Farr, A.L., and Randall,
R.J. (1951) Protein measurement with Folin phenol
reagent. J Biol Chem 193: 265–275.
Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982) Molecular Cloning, a Laboratory Manual . Cold Spring Harbor,
NY: Cold Spring Harbor Laboratory Press, p. 68 and
p. 250.
Ninfa, A.J., and Magasanik, B. (1986) Covalent modification
of the glnG product, NRI , by the glnL product, NRII , regulates the transcription of the glnALG operon in Escherichia
coli . Proc Natl Acad Sci USA 83: 5909–5913.
Rhee, S.G., Chock, P.B., and Stadtman, E.R. (1989) Regulation of Escherichia coli glutamine synthetase. Adv Enzymol
(related areas Mol Biol) 62: 37–92.
Saiki, R.K. (1990) Amplification of genomic DNA. In PCR Protocols, a Guide to Methods and Applications. Innis, M.A.,
Gelfand, D.H., Sninsky, J.J., and White, T.J. (eds). San
Diego, CA: Academic Press, pp. 13–20.
Schauder, B., Blocker, H., Frank, R., and McCarthy, J.E.G.
(1987) Inducible expression vectors incorporating the
䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

E. coli GlnK protein 313
Escherichia coli atpE translation initiation region. Gene 52:
279–283.
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.
Wong, P.-K., Popham, D., Keener, J., and Kustu, S. (1987) In
vitro transcription of the nitrogen fixation regulatory operon

䊚 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 301–313

nifLA of Klebsiella pneumoniae. J Bacteriol 169: 2876–
2880.
Xu, Y., Cheah, E., Carr, P.D., van Heeswijk, W.C., Westerhoff, H.V., Suhhash, V.G., and 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.