|
|
|
|||||||
|
|||||||||
(Received for publication, July 19, 1994; and in revised form, November 30, 1994) From the
The two-component system sensor/response regulator pair,
FixL/FixJ, controls the expression of Rhizobium meliloti nitrogen fixation (nif and fix) genes in
response to changes in oxygen concentration. A truncated version of
FixL, FixL*, is an oxygen-binding hemoprotein kinase that
phosphorylates and dephosphorylates the nif and fix gene transcriptional activator, FixJ. Phosphorylation of FixJ is
required for optimal transcriptional activation, and anaerobic
conditions in vitro result in a substantial increase in the
level of FixJ-phosphate. In this study, site-directed mutagenesis was
carried out at histidine residues in FixL*. Mutant FixL* derivatives
were purified and analyzed in vitro for their heme/oxygen
binding properties and phosphorylation/dephosphorylation activities.
Mutation of histidine 285, the putative autophosphorylation site, to
glutamine results in the loss of FixL* phosphorylation activities.
However, this mutant protein retains a substantial level of
FixJ-phosphate dephosphorylation activity. Mutation of histidine 194 to
asparagine results in the loss of heme binding and in the failure of
FixL* to regulate its phosphorylation/dephosphorylation activities in
response to changes in oxygen concentration. The FixL*H194N mutant
protein also exhibits an increased FixJ phosphorylation activity under
aerobic conditions. This study provides further evidence for the
importance of the heme binding domain of FixL* in regulating FixJ
phosphorylation and dephosphorylation activities in response to oxygen.
The soil bacterium, Rhizobium meliloti, fixes
dinitrogen (N Our laboratory
has undertaken a biochemical approach to understanding the mechanism by
which the FixL protein senses oxygen and transduces this signal to
FixJ, ultimately resulting in the transcriptional activation of nif and fix genes (see (13) for review). Because
FixL is a transmembrane protein with four transmembrane
helices(14) , a soluble, truncated version of FixL, FixL*, was
constructed to facilitate purification and in vitro studies(15) . FixL* is an oxygen-binding hemoprotein and a
kinase capable of autophosphorylation as well as the phosphorylation of
the FixJ protein(15) . Analysis of purified deletion
derivatives established that FixL* consists of two separable,
functional domains; the N-terminal domain that functions in heme and
oxygen binding and the C-terminal domain that possesses
autophosphorylation and FixJ phosphorylation activities(16) .
The rate of FixL* autophosphorylation increases when reactions are
performed under anaerobic conditions relative to aerobic
conditions(16, 17, 18) , and this increased
activity requires the presence of the heme binding domain(16) .
The phosphorylation of FixJ occurs through direct phosphotransfer from
FixL*-phosphate to FixJ, and the rate of the phosphotransfer reaction
is not affected by oxygen concentration(17, 18) .
Additionally, FixL* possesses a FixJ-phosphate dephosphorylation
activity which decreases in response to anaerobic
conditions(17) . The FixJ dephosphorylation activity resides in
the C-terminal kinase domain. ( In this work, we analyze the
effects of mutations of several histidine residues in FixL* on oxygen
sensing and signal transduction in vitro. The histidine at
position 285 in the kinase domain of FixL* was mutated, because it is a
highly conserved histidine that is predicted to be the site of
autophosphorylation based on sequence alignments with other
two-component system sensors(5) . Consistent with this
prediction, mutation of histidine 285 results in loss of
autophosphorylation and FixJ phosphorylation activities. All three
histidine residues that lie in the previously defined heme binding
region of the protein (16) were also mutationally altered
because of the similarity of the absorption spectra of FixL* and
hemoglobin (15) and the likelihood that a histidine is involved
in coordination of the heme iron. Histidine 194 is identified as the
putative site of heme coordination. Finally, it is shown that mutations
in the heme binding region have a significant impact on C-terminal
kinase activity.
Mutagenesis was carried out by the method of
Kunkel(27) . Oligonucleotides used for the mutagenesis were
phosphorylated using T4 polynucleotide kinase purchased from U. S.
Biochemical Corp. Single-stranded DNA was prepared from E. coli strain RZ1032 bearing plasmid pEM13.8, pEM19.2, or pEM28.5 by
using the helper phage M13K07(28) . Ten pmol of each of the
phosphorylated oligonucleotides were annealed to 0.5 pmol of the
appropriate single-stranded DNA (prepared from strain RZ1032) in PE1
buffer (20 mM Tris-HCl, pH 7.5, 10 mM MgCl
Figure 1:
Schematic
structure of FixL and the truncated FixL derivative, FixL
FixL*H138Q, FixL*H285Q, and
FixL*H194N were purified to approximately 90% purity based on Coomassie
Blue staining of SDS-PAGE gels. FixL*H138Q and FixL*H285Q are
orange-red in color, like purified FixL*, indicating the presence of
heme. The absorption spectra of FixL*H138Q and FixL*H285Q are very
similar to those of FixL*, both in the presence (Fig. 2A) and absence (Fig. 2B) of
oxygen. The absorption maxima for each peak for the mutant and
wild-type proteins, both oxygenated and deoxygenated, are within 0.5 nm
of one another (data not shown). This shows that changing either
histidine 138 or 285 to glutamine has no major effect on the abilities
of FixL* to interact with heme or to bind oxygen (although changes in
oxygen binding affinity would not be addressed by this analysis).
Figure 2:
Absorption spectra and gel heme assays of
FixL* and FixL*histidine mutants. A, oxygenated spectra of
FixL* (thick solid line), FixL
The purified FixL*H194N protein is colorless and fails to show the
characteristic heme absorption spectrum of FixL*, FixL*H138Q, and
FixL*H285Q (data not shown). FixL*H194N also fails to show a
FixL*-specific heme staining band in crude extracts suggesting that the
lack of heme in FixL*H194N is not simply due to the loss of heme during
purification (Fig. 3, lane 2). The failure of
FixL*H194N to bind heme suggests that the histidine at position 194 is
involved in coordination of the heme iron. In hemoglobin, the fifth
coordination position of the heme iron is occupied by a histidine
residue, the so-called proximal histidine(33) , and mutation of
this histidine to glutamine in human
Figure 3:
Nondenaturing gel heme assay of FixL* and
FixL* mutant derivatives. Lanes 1-5 contain
approximately 200 µg of total protein of the indicated crude
extracts: lane 1, BL21 (pEMH192N); lane 2, BL21
(pEMH194N); lane 3, TG1 (pGG820); lane 4, BL21
(pUC9), lane 5, BL21 (pAF1). Lane 6 contains 3.6
µg of purified FixL*H285Q. Each strain used to make the crude
extracts also carried the nifA::lacZ reporter plasmid pCHK57.
Plasmid pAF1 overexpresses fixJ from the lac promoter
in pUC9 (A. Lois, personal communication).
Although FixL*192N was not
purified and examined further, analysis of crude extracts revealed that
FixL*192N does indeed contain heme (Fig. 3, lane 1).
This indicates that the histidine at position 192, despite its close
proximity to the important histidine at position 194, is not likely to
be involved in heme binding.
FixL*H194N was quantitated relative to the other FixL* derivatives
on the basis of Coomassie Blue staining material on SDS-PAGE gels. On
the basis of densitometric scans, we estimate that the variation in the
amounts of the FixL* derivatives used in each of the assays was no more
than 15%.
Fig. 4A shows the
autophosphorylation activity of FixL* compared with that of FixL*H138Q
in the presence and absence of oxygen. As reported previously, the
FixL* protein shows approximately a 3-fold increase in the rate of
autophosphorylation under anaerobic relative to aerobic conditions at
22 °C(16) . FixL*H138Q shows an elevated
autophosphorylation rate in the presence of oxygen, a rate that is
essentially equivalent to the rate of FixL* autophosphorylation in the
absence of oxygen. Upon deoxygenation, the rate of FixL*H138Q
autophosphorylation increases by about 2-fold; an oxygen response
similar in magnitude to that of FixL*. Thus, the autophosphorylation
activity of FixL*H138Q depends on oxygen concentration, consistent with
the fact that H138Q retains oxygen binding ability. Furthermore,
FixL*H138Q has even greater autophosphorylation activity than FixL*
both in the presence and in the absence of oxygen.
Figure 4:
Autophosphorylation of FixL* and mutants. A, autophosphorylation of FixL* (circles) and
FixL*H138Q (squares). B, autophosphorylation of FixL* (circles) and FixL*H194N (triangles). Open
symbols indicate aerobic conditions and filled symbols anaerobic conditions. Each of the FixL* derivatives (1
µM) was incubated with [
Fig. 4B shows the autophosphorylation of FixL*H194N
in comparison with FixL*. As would be expected for a heme-less FixL*
protein, the autophosphorylation activity of FixL*H194N fails respond
to oxygen. Since the absolute level of FixL*194N autophosphorylation is
higher than wild-type FixL* in the presence of oxygen, this suggests
that a major function of the oxygenated heme moiety is to repress the
autophosphorylation activity of the C-terminal portion of the protein.
Although the rate of FixL*H194N autophosphorylation was always higher
than that of oxygenated FixL*, the values measured in different
experiments were somewhat variable, ranging from 1.5 to 3 times that of
FixL* in the presence of oxygen. This is in contrast to the very
reproducible rates exhibited by FixL*H138Q and FixL*. A possible
explanation for this is that FixL*H194N may be less stable than FixL*.
It is well known that apomyoglobin and apohemoglobin denature much more
readily than do their heme containing counterparts(36) .
Unlabeled FixJ was added to each
Figure 5:
Transfer of phosphate from phosphorylated
FixL*, FixL*H138Q and FixL*H194N. A, transfer of phosphate
from FixL* (open circles) to FixJ (filled circles)
and from FixL*H138Q (open squares) to FixJ (filled
squares). B, transfer of phosphate from FixL* (open
circles) to FixJ (filled circles) and from FixL*H194N (open triangles) to FixJ (filled triangles).
Phosphorylated FixL*, FixL*H138Q, or FixL*H194N (0.6 µM)
was incubated under aerobic conditions with FixJ (1.2 µM)
at 22 °C. Proteins were analyzed by SDS-PAGE and autoradiography.
Quantitation was by laser densitometry.
The
FixJ-phosphate dephosphorylation activity of FixL* is shown in Fig. 6A. The dephosphorylation activity of FixL*
decreases under aerobic conditions. In contrast to earlier results (17) , we have found that this decrease in phosphatase activity
in response to anaerobiosis is of the same magnitude whether or not
FixL* is preincubated with ATP (data not shown). In addition, the
amount of the decrease is somewhat less than was reported previously
(see ``Discussion'').
Figure 6:
Dephosphorylation of FixJ-phosphate by
FixL* and FixL* mutant derivatives.
FixL* 138Q exhibits aerobic and
anaerobic phosphatase activities equivalent to those of FixL* (Fig. 6B), demonstrating that the mutation at position
138 has no significant effect on phosphatase activity. Since the H138Q
mutation substantially affects kinase activity (Fig. 4A); however, the data suggest that the kinase
and phosphatase activities are to some extent independently regulated
by the N-terminal heme binding domain of FixL*. FixL*H194N and FixL*
dephosphorylate FixJ-phosphate at equivalent rates under anaerobic
conditions (Fig. 6C), but FixL*H194N consistently
displays a reduced rate of FixJ-phosphate dephosphorylation under
aerobic conditions. The rate of FixJ-phosphate dephosphorylation by
FixL*H194N is essentially unaffected by oxygen concentration, i.e. the aerobic and anaerobic rates are similar. This is consistent
with the fact that FixL*H194N lacks heme and suggests that an
oxygenated heme domain could be required for full FixJ
dephosphorylation activity. The FixL*H285Q protein dephosphorylates
FixJ-phosphate, although at a significantly reduced rate relative to
that of FixL* (Fig. 6D), suggesting that the histidine
at position 285 plays a role in, but is not absolutely required for,
phosphatase activity. The FixL*H285Q phosphatase activity is still
down-regulated in response to anaerobiosis, consistent with the ability
of this protein to bind heme and oxygen.
Figure 7:
Phosphorylation of FixJ by FixL* and FixL*
mutant derivatives. FixL*, FixL*H138Q, or FixL*H194N (0.5
µM) was incubated with FixJ (1 µM) as
described under ``Materials and Methods'' under aerobic (open symbols) or anaerobic (filled symbols)
conditions. The amount of phosphorylated FixJ protein is plotted. A, phosphorylation of FixJ by FixL* (circles) and
FixL*H138Q (squares). B, phosphorylation of FixJ by
FixL* (circles) and FixL*H194N (triangles). Proteins
were analyzed by SDS-PAGE and autoradiography, followed by laser
densitometry.
Under
aerobic conditions, FixL*H194N also shows an increased level of FixJ
phosphorylation relative to that seen for FixL* (Fig. 7B). Although the results with FixL*H194N were
again more variable than for the other FixL* derivatives, the level of
FixJ phosphorylation by FixL*H194N under both aerobic and anaerobic
conditions is always 4-7-fold higher than that seen for FixL* in
the presence of oxygen. The increased level of FixJ phosphorylation by
FixL*H194N under aerobic conditions is consistent with the increased
autophosphorylation activity of FixL*H194N under aerobic conditions (Fig. 4B). The H285Q mutant does not show any FixJ
phosphorylation activity (data not shown), consistent with its lack of
autophosphorylation activity. In this work, we have examined the effects of mutations of
selected histidine residues on the heme binding properties as well as
the in vitro phosphorylation and dephosphorylation activities
of FixL* in the presence and absence of oxygen. Two mutations in the
region of the protein previously demonstrated to bind heme, H138Q and
H194N, increase FixL* autophosphorylation under aerobic conditions.
Significantly, a change of histidine 194 to glutamine abrogates heme
binding. Since this is the only histidine residue in the heme binding
domain which, when mutated, results in the loss of heme binding, this
histidine is likely to be the site of heme iron coordination. Finally,
we showed that mutation of the highly conserved histidine residue at
position 285 results in the loss of autophosphorylation and FixJ
phosphorylation activities, suggesting that histidine 285 is the site
of autophosphorylation. The H285Q mutant is also defective in FixJ
dephosphorylation activity, but a significant level of activity
remains, and this activity is regulated by oxygen. The FixL*H194N
data are consistent with a model in which the heme binding region of
FixL* exerts a negative effect on the autophosphorylation activity of
the C-terminal kinase domain, since the loss of heme in FixL*H194N
results in rates of autophosphorylation and FixJ phosphorylation under
aerobic and anaerobic conditions that are roughly equivalent to the
rates of these reactions catalyzed by FixL* under anaerobic conditions.
This increased in vitro activity of FixL*H194N in the presence
of oxygen is consistent with data showing that deletion of the heme
binding region (but not the transmembrane regions) results in elevated
transcriptional activity at the nifA and fixK promoters under aerobic conditions in vivo in E.
coli(38) . The H138Q protein shows a significantly
higher level of autophosphorylation activity than does FixL* under both
repressed (aerobic) and derepressed (anaerobic) conditions. In this
regard, it is similar to the FixL*362 mutant, an alanine-to-valine
mutation at amino acid position 362(17) . However, FixL*H138Q
does not show a reduction in phosphatase activity as does
FixL*362(17) . The H138Q mutation may affect the conformation
of the C-terminal kinase domain in such a way that it has increased
access to the ATP substrate or that the active site is better
positioned to carry out ATP hydrolysis. Alternatively, the H138Q
mutation could alter the monomer-dimer equilibrium such that a greater
proportion of dimers are formed. Gel filtration data are consistent
with FixL* existing as a dimer(15) , and there is a growing
body of evidence that two-component system sensors function as dimers
with autophosphorylation occurring through a transphosphorylation
between the subunits within a dimer(39, 40) . It
has been shown previously that anaerobiosis decreases the rate of
FixJ-phosphate dephosphorylation by FixL*(17) . The FixJ
dephosphorylation activity of FixL* was shown to be significantly
inhibited under anaerobic conditions when FixL* was incubated with ATP
(to generate FixL*-phosphate) prior to the addition of the
FixJ-phosphate substrate. Anaerobic conditions had a less dramatic
effect on reducing phosphatase activity when FixL* was in the
unphosphorylated state (not preincubated with ATP)(17) .
However, in this study we saw no significant difference in the
phosphatase activities between unphosphorylated and phosphorylated
FixL* either in the presence or absence of oxygen (data not shown).
Similarly, we observed no effect of phosphorylation on the
dephosphorylation activities of the FixL*H138Q, FixL*H194N, or
FixL*H285Q mutant proteins (data not shown). The magnitude of the
oxygen effect that we observed in this work was also less dramatic than
that seen in the earlier study. The reason for these differences is
presently unclear but might be due to differences in the FixL* and FixJ
preparations used in each study. It is interesting that FixJ-phosphate
dephosphorylation by FixL* appears to occur in at least two phases: an
initial rapid phase followed by a slower phase (Fig. 6A). For FixL* and the FixL* histidine mutants,
the decreased rate of dephosphorylation under anaerobic conditions is
evident only in the later, slower phase of the reaction. In a previous
study of phosphorylated FixL*, the effect of oxygen was evident at
earlier times in the reaction(17) . More detailed kinetic
experiments must be done in order to understand the biphasic nature of
FixL* dephosphorylation of FixJ-phosphate and the effects of oxygen in
the early, fast phase of the reaction. The phosphatase activities of
the histidine mutants demonstrate that mutations in both the heme
binding region (H194N) and the conserved kinase domain (H285Q) can
affect phosphatase function. The FixL*H194N mutant shows a reduction in
aerobic, but not anaerobic, phosphatase activity that is consistent
with the loss of oxygen sensing capability. This suggests that an
oxygenated heme domain is required for full phosphatase activity. This
situation is analogous to NtrB, whose NtrC-phosphate dephosphorylation
activity is repressed (41, 42) in the absence of an
activating signal from the PII protein. Mutation of the conserved,
putative autophosphorylation site, histidine 285, reduces, but does not
eliminate FixL* phosphatase activity. In addition, the phosphatase
activity remaining is regulated by oxygen concentration. Work with
other two-component system sensors has shown that mutations of the
conserved histidine sometimes result in the complete loss of
phosphatase activity. For example, a histidine-to-valine mutation at
the conserved position 139 in NtrB results in a mutant protein that is
devoid of phosphatase activity. However, an asparagine replacement at
the same position retains phosphatase activity (as assayed by the
ability of the mutant NtrB protein to negatively regulate transcription
from the glnAp2 promoter)(43) . Such data together with the
FixL*H285Q results reported here suggest that although the conserved
histidine is not absolutely required for phosphatase activity,
perturbations in structure around this histidine can significantly
influence this activity. Interestingly, the H138Q mutation has a
significant effect on FixL* kinase activity but no apparent effect on
phosphatase activity. This suggests that the kinase and phosphatase
activities are to some degree independent. The independence of
phosphatase and kinase activities has recently been demonstrated for
NtrB by showing that it is possible to mutate one activity without
altering the other(43) . This study has focused on the
effects of specific mutations on the in vitro activities of a
soluble truncated derivative of FixL, FixL*. The in vitro response of FixL* to its oxygen ligand will facilitate the
understanding of the molecular basis of oxygen sensing in R.
meliloti as well as the understanding of signal transduction by
two-component regulatory system proteins in general. It is important to
keep in mind, however, that in vitro studies with FixL* have
limitations. Although it has been demonstrated that the membrane
spanning regions are not absolutely required for oxygen sensing and
signal transduction in E. coli(38, 44) or R. meliloti(14) , maximal responsiveness to oxygen
requires that the transmembrane segments be
present(14, 44) . The membrane-spanning regions are
therefore likely to play a role in FixL function. The membrane region
of FixL may sense other signals related to low oxygen growth such as
the redox state of the cell. There is precedent that a single sensor
can respond to multiple input signals. VirA, which is required for
virulence of Agrobacterium tumefaciens, responds in its
periplasmic domain to sugars (45, 46) and in its
cytoplasmic domain to phenolic compounds(47) . In addition, it
has recently been suggested that the full-length FixL protein responds
to ammonia in vivo(48) . Clearly, more work needs to
be done with the full-length FixL protein before we understand the
basis of signal transduction by FixL in R. meliloti.
Volume 270,
Number 10,
Issue of March 10, 1995 pp. 5243-5250
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ROLE OF HISTIDINE RESIDUES IN HEME BINDING, PHOSPHORYLATION,
AND SIGNAL TRANSDUCTION (*)
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
) into ammonia when it is in symbiotic
association with its plant host, Medicago sativa (alfalfa).
There are at least 23 genes from R. meliloti required for
nitrogen fixation (nif and fix genes) as well as a
host of other both bacterial and plant genes that are required for the
nodulation process. The nitrogen fixation genes in R. meliloti are expressed relatively late in nodule development and the key
signal that induces the expression of most of the nif and fix genes is low oxygen concentration(1, 2) .
This is consistent with the fact that the nitrogenase enzyme is oxygen
labile, making it imperative for the bacterium to express the nitrogen
fixation machinery only when oxygen concentration is low. Two genes, fixL and fixJ, are responsible for sensing and
transducing the low oxygen signal which results in the activation of nif and fix gene expression(2, 3) .
The products of the fixL and fixJ genes are
homologous to the two-component regulatory system proteins (2) that are involved in signal transduction in
bacteria(4) . Genetic and biochemical evidence from many of
these systems supports a phosphotransfer model for signal
transduction(5, 6, 7, 8) . FixL is
an oxygen sensor that is homologous to the sensor class, and FixJ is a
transcriptional activator homologous to the response regulator class of
two-component system proteins. Genetic studies show that FixL responds
to low oxygen concentration by activating the transcriptional
activator, FixJ(9) . Activated FixJ then increases
transcription at the nifA and fixK promoters(9) , whose products are transcriptional
activators of other nif and fix genes(10, 11, 12) .
)The net result of anaerobic
conditions in vitro is an increase in the level of
phosphorylated FixJ that is available to activate transcription from
the nifA and fixK promoters. Recently the entire
signal transduction pathway starting with anaerobiosis and ending with
transcriptional activation at the nifA and fixK promoters was reconstituted in vitro(19, 20) using purified truncated FixL
derivatives, purified FixJ, and purified RNA polymerase from Escherichia coli or R. meliloti. The FixLJ system
thus presents a unique system for understanding the molecular basis of
oxygen sensing in R. meliloti.
Bacterial Strains
E. coli MC1061
(F
araD139 
(ara-leu)7696 (lac)X74 galU galK
hsdR2 (r
m
) mcrB1 rpsL (Str
) (21) or E. coli TB1 [JM83 hsdR (rm)) (22) was used for
plasmid constructions. E. coli TG1[(lac-proAB)
supE thi (F` traD36 LacIq proA
proBLacZM15(23) or BL21 (F hsdS gal) (24) was used for overexpression of FixL derivatives. E.
coli strain RZ1032 (HfrKL16 PO/45 (lysA (61-62)) dut1 ung1 thi1 relA1 Zbd-279::Tn10 supE44) (25) was
used for preparation of single-stranded DNA for site-directed
mutagenesis.Plasmids
DNA manipulations were carried out using
standard procedures(26) . To facilitate mutagenesis of specific
histidine residues, small restriction fragments encoding the residues
of interest were subcloned into pBluescript SK (Stratagene Cloning Systems). The 128-base pair HindIII-SstII fixL fragment bearing
histidine 138 was cloned into the HindIII-SstII sites
of pBluescript SK, the 220-base pair XhoI-EcoRI fragment bearing histidine 285 was
subcloned into XhoI/EcoRI-digested pBluescript
SK, and the 143-base pair ClaI-SstII
fragment bearing histidines 192 and 194 was subcloned into ClaI-SstII-digested pBluescript SK.
The plasmids were designated pEM13.8, pEM28.5, and pEM19.2,
respectively. The source of the fixL DNA fragments was the
FixL* overexpression vector pGG820(15) .Mutagenesis
Oligonucleotides were designed to
mutate a specific histidine residue as well as to create or eliminate a
restriction enzyme cleavage site to facilitate identification of mutant
clones. The mutagenic oligonucleotides were as follows: (i) His to Gln
at position 138; creation of a PvuII restriction site, 5` CGC
GAC GCT CAG CTG AGA TCC ATA C 3`; (ii) His to Asn at position 192; loss
of EcoRV restriction site, 5` GCG CTG CAG ATA GCC GTC
GTG TTC GTT GCG GTA GGG TT 3`. (iii) His to Asn at position 194; loss
of EcoRV restriction site, 5`GCG CTG CAG ATA GCC GCC
GTC GTT TTC GTG GCG GT 3`. (iv) His to Gln at position 285; creation of SacI restriction site, 5`G ACG CTT GCC CAG GAG CTC AAC CAG CCG TTG 3`. Underlined bases indicate the base changes that
alter the amino acid sequence. Bold bases are base changes that do not
alter the amino acid sequence but create or eliminate a restriction
enzyme cleavage site., 50 mM NaCl, 1 mM dithiothreitol)(26) . Second strand synthesis and ligation
were performed using 0.5 mM fast protein liquid
chromatography-purified dNTPs (Pharmacia Biotech Inc.), 1 mM ATP, 2 units of T4 DNA ligase, and 25 units of exonuclease-free
Klenow (U. S. Biochemical Corp.) in PE1 buffer without NaCl. The
reaction was incubated overnight at 15 °C. The reaction mixture was
transformed into E. coli strain MC1061, and clones with the
appropriate restriction enzyme changes were sequenced. The entire
stretch of fixL DNA was sequenced to ensure that the sequences
were as expected and that there were no additional mutations. The
restriction fragments bearing the appropriate mutations were cloned
back into the pGG820 FixL* overexpression vector(15) . The
resulting pGG820 derivatives were named according to the position
number and identity of the amino acid change. Thus, pEMH138Q, pEMH192N,
pEMH194N, and pEMH285Q are plasmids expressing mutant FixL* proteins
that have a glutamine at position 138, asparagines at positions 192 or
194, and a glutamine at position 285, respectively.Purification of FixL*, FixJ, and FixL* Mutant
Proteins
Purification of FixL* from E. coli strain TG1
bearing pGG820 and plasmid pCHK57 has been described(16) . The
purification of FixJ from TG1 bearing pCHK57 and pMW5 has been
described(29) . FixL*H138Q, FixL
H194N, and
FixL*H285Q were purified exactly as was FixL*, except that the strain
used for overexpression of these three plasmids was BL21 bearing
pCHK57. The presence of the nifA::lacZ reporter plasmid,
pCHK57, did not substantially affect the expression of the FixL*
derivatives or FixJ.Autophosphorylation and FixJ Phosphorylation
Reactions
Autophosphorylation reactions and FixJ phosphorylation
reactions were carried out at 22 °C as described(16) .Phosphotransfer Reactions
Phosphotransfer
reactions were conducted as described in (17) except that the
FixL* derivatives were labeled with [
-P]ATP
for 90 min under anaerobic conditions at 22 °C prior to G-25 column
chromatography. 0.6 µM phosphorylated FixL* and 1.2
µM FixJ were used in the transfer reactions.FixJ-Phosphate Dephosphorylation
Reactions
Purification of P-labeled FixJ-phosphate
and phosphatase reactions were carried out as described (17) using FixL* and FixL* derivatives that had been
preincubated with 0.8 mM ATP for 60 min at 22 °C prior to
assaying phosphatase activity. 0.8 mM ATP was included in the
phosphatase reaction mixtures.Gel Electrophoresis
Autophosphorylation, FixJ
phosphorylation, phosphotransfer, and phosphatase reactions were
stopped at the appropriate times by the addition of SDS-PAGE ()gel sample buffer (30) containing 10 mM EDTA. Reaction products were analyzed on 10% SDS-PAGE gels (30) followed by autoradiography and laser densitometry.Densitometry
Autoradiograms were scanned using a
laser densitometer (LKB), and the data were analyzed using software
from the Ambis Corp. (San Diego, CA). All the phosphorylation and
dephosphorylation experiments reported were repeated a minimum of four
times, and representative experiments are shown.Gel Heme Assays
Nondenaturing gel heme assays were
conducted as described previously(16) .
Construction and Expression of FixL* Mutants
To
investigate the roles of specific histidine residues in FixL* function,
we mutated selected histidines using oligonucleotide-directed
mutagenesis. In addition to mutating the histidine that is the putative
site of autophosphorylation, all 3 histidines in the heme binding
domain were altered. The substituting amino acids represent changes as
conservative as possible based on the criteria of ``safe''
substitutions defined by Bordo and Argos (31) and the secondary
structure predicted by the method of Chou and Fasman(32) .
Mutagenesis was carried out by the method of Kunkel (27) as
described under ``Materials and Methods,'' and the mutant
proteins were cloned into the FixL* overexpression vector
pGG820(15) . Amino acid substitutions are as indicated in the
legend to Fig. 1.
. Filled boxes indicate the four transmembrane (TM)
regions(14) . Stippled boxes indicate the heme and
oxygen binding region (16) , and the hatched boxes indicate the conserved histidine kinase
domain(2, 16) . The positions of the histidines that
were individually mutated, and the resulting amino acid replacements
are shown. Numbers refer to amino acid position based on the
first ATG in the open reading frame(2) . The mutant FixL*
proteins are designated according to the position and identity of the
amino acid change. Thus, FixL*H138Q is a FixL* protein with a histidine
to glutamine mutation at position 138. All of the FixL* derivatives
were produced in E. coli from the FixL* overexpression vector,
pGG820, and therefore all the FixL* derivatives have the same
N-terminal fusion to lacZ as does wild-type
FixL*(15) .
Heme and Oxygen Binding Properties of FixL*
Mutants
The mutant constructs were overexpressed in E.
coli, and the production of the mutant FixL* derivatives was
confirmed by SDS-PAGE followed by both Coomassie Blue staining and
Western blotting. Mutant FixL* proteins bearing the H138Q, H192N, or
H285Q mutations were expressed to similar levels relative to total
cellular proteins in E. coli strain BL21 (pCHK57) as was the
wild-type FixL protein in E. coli strain
TG1(pCHK57). The FixL* protein with the H194N mutation was expressed in E. coli strain BL21 (pCHK57) at approximately one-half the
level of the other FixL* derivatives.
H138Q ( ), and
FixLH285Q(- - - -). B,
deoxygenated spectra of FixL (thick solid line),
FixLH138Q ( ), and FixLH285Q(- -
- -). Spectra were measured at 0.5-nm intervals using a
Hitachi U-2000 spectrophotometer. Proteins were in 20 mM Tris-HCl, pH 7.8, 125 mM NaCl, 20% (v/v) glycerol, 10
mM 2-mercaptoethanol. Deoxygenation was achieved by the
addition of a crystal of sodium dithionite to a 0.5-ml
sample.
-hemoglobin results in the
loss of heme binding(34) .
Quantitation of FixL* and Mutant Derivatives
The
mutant FixL* proteins were tested for autophosphorylation, FixJ
phosphorylation, and FixJ dephosphorylation activities. In order to
compare activities of different mutant proteins in vitro, it
was necessary to determine accurately the concentration of each of the
purified FixL* protein derivatives. For most proteins, this can be done
by determining the extinction coefficient at 280 nm for the purified
protein(35) . However, this is not possible for hemoproteins
because heme absorbs in the ultraviolet range. The heme-containing
proteins, FixL*, FixL*H138Q, and FixL*H285Q were quantified based on
the previously determined extinction coefficient for FixL* at 417
nm(15) . Because this method assumes an equal heme content for
each protein preparation, we also quantified the proteins by running
various amounts of each protein on SDS-PAGE gels followed by staining
with Coomassie Blue and laser densitometry. The relative amounts of
purified protein determined by absorbance at 417 nm correlated with the
amount of Coomassie Blue staining material, suggesting that similar
proportions of each protein preparation contain bound heme. Autophosphorylation
The mutant proteins were
assayed for their autophosphorylation activities relative to wild-type
FixL*. The mutant FixL*H285Q protein showed no autophosphorylation
activity either in the presence or absence of oxygen (data not shown).
The lack of autophosphorylation activity is consistent with the
histidine at position 285 being the site of autophosphorylation as
expected from sequence alignment studies with two-component system
sensors(5, 7) .
-P]ATP
as described under ``Materials and Methods.'' Proteins were
analyzed by SDS-PAGE and autoradiography. Quantitation was by laser
densitometry.
Phosphotransfer from FixL* Mutants to
FixJ
FixL*H138Q-phosphate and FixL*H194N-phosphate were tested
for their abilities to phosphorylate FixJ. Autophosphorylation of
FixL*, FixL*H138Q, and FixL*H194N was carried out by incubating each of
the proteins with [-P]ATP under anaerobic
conditions. The P-labeled proteins were then separated
from ATP by G-25 column chromatography, and the amount of phosphate
incorporated was quantitated by liquid scintillation counting.
Approximately 70-75% of FixL*, FixL*H138Q, and FixL*H194N were
phosphorylated after 90 min at 22 °C (data not shown). The H285Q
mutant protein fails to autophosphorylate and therefore could not be
tested in this assay.P-labeled FixL* derivative and the rate of transfer from
FixL* to FixJ was monitored. The initial rates of phosphate loss and
FixJ-phosphate accumulation from FixL*H138Q and FixL*H194N correlate
well with each other and are equivalent to the rates observed for FixL* (Fig. 5, A and B). These results suggest that
the FixL*H138Q and the FixL*H194N mutations have no significant impact
on the initial rate of phosphotransfer from FixL* to FixJ. At later
time points, there is evidence of loss of transfer activity with a
subsequent gradual hydrolysis of phosphate from FixJ-phosphate.
FixJ-phosphate ultimately accumulates to somewhat higher levels for
FixL*H138Q (Fig. 5A) and FixL*H194N (Fig. 5B) than for FixL*, suggesting that there may
differences in the stability of FixJ-phosphate in the presence of the
different L* derivatives. The rates of phosphotransfer from the FixL*
derivatives to FixJ were unaffected by oxygen concentration (data not
shown) as is true for the wild-type FixL* protein(17) .
Dephosphorylation of FixJ-Phosphate by FixL* and FixL*
Mutant Proteins
We tested whether the FixL* histidine mutant
proteins are altered in their abilities to dephosphorylate
phosphorylated FixJ both in the presence and absence of oxygen. FixJ
was labeled with P by incubating it in the presence of
FixL*362, a mutant shown to have increased FixJ phosphorylation
activity(17) , and [-P]ATP under
anaerobic conditions. FixJ-phosphate was then separated from ATP by
G-25 column chromatography and from FixL*362 by using an anti-FixL
immunoaffinity column(17) . FixJ-phosphate has been shown to
have a half-life of approximately 4 h under standard assay
conditions(17, 37) . FixJ-phosphate dephosphorylation
assays were conducted with wild-type FixL* and its mutant derivatives
that had been preincubated with ATP as described (17) .
P-Labeled
FixJ-phosphate (0.45 µM) was incubated with FixL*,
FixL*H138Q, FixL*H194N, or FixL*H285Q (1 µM) under aerobic (open symbols) or anaerobic (filled symbols)
conditions as described under ``Materials and Methods.''
Dephosphorylation of FixJ-phosphate by FixL* (A, circles),
FixL* (B, circles) and FixL*H138Q (B, squares); FixL* (C, circles) and FixL*H194N (C, triangles); and FixL* (D, circles) and FixLH285Q (D, diamonds)
is shown. FixJ-phosphate incubated under aerobic assay conditions
without a FixL* derivative is indicated by the 
symbol. Proteins
were analyzed by SDS-PAGE, autoradiography, and laser
densitometry.
FixL and FixJ Phosphorylation Reactions
To analyze
the net effect of the competing FixJ phosphorylation and
dephosphorylation activities of the FixL* mutants, the proteins were
incubated with FixJ in the presence of
[-P]ATP. Fig. 7A shows the rates
of FixJ phosphorylation by FixL* and by FixL*H138Q. The phosphorylation
of FixJ by FixL* is enhanced approximately 8-fold under anaerobic
relative to aerobic conditions at 22 °C, consistent with previously
reported results (16, 17) . FixL*H138Q gives increased
FixJ phosphorylation relative to FixL* both in the presence and absence
of oxygen. This is consistent with the increased autophosphorylation
activity of FixL*H138Q under aerobic and anaerobic conditions.
))
We thank E. P. Geiduschek, P. Agron, and A. Lois for
helpful discussions and G. Kassavetis for advice on densitometry.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. L. Smart and C. E. Bauer Tetrapyrrole Biosynthesis in Rhodobacter capsulatus Is Transcriptionally Regulated by the Heme-Binding Regulatory Protein, HbrL J. Bacteriol., February 15, 2006; 188(4): 1567 - 1576. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Nakamura, H. Kumita, K. Imai, T. Iizuka, and Y. Shiro ADP reduces the oxygen-binding affinity of a sensory histidine kinase, FixL: The possibility of an enhanced reciprocating kinase reaction PNAS, March 2, 2004; 101(9): 2742 - 2746. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-A. Gilles-Gonzalez and G. Gonzalez Signal transduction by heme-containing PAS-domain proteins J Appl Physiol, February 1, 2004; 96(2): 774 - 783. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Harper, L. C. Neil, and K. H. Gardner Structural Basis of a Phototropin Light Switch Science, September 12, 2003; 301(5639): 1541 - 1544. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Stephenson and J. A. Hoch PAS-A domain of phosphorelay sensor kinase A: A catalytic ATP-binding domain involved in the initiation of development in Bacillus subtilis PNAS, November 29, 2001; (2001) 251408398. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. G. Puskás, M. Inui, K. Zahn, and H. Yukawa A periplasmic, {alpha}-type carbonic anhydrase from Rhodopseudomonas palustris is essential for bicarbonate uptake Microbiology, November 1, 2000; 146(11): 2957 - 2966. [Abstract] [Full Text]
|
|
|
|
 |
|
 A.-M. Garnerone, D. Cabanes, M. Foussard, P. Boistard, and J. Batut Inhibition of the FixL Sensor Kinase by the FixT Protein in Sinorhizobium meliloti J. Biol. Chem., November 5, 1999; 274(45): 32500 - 32506. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Miyatake, M. Mukai, S.-i. Adachi, H. Nakamura, K. Tamura, T. Iizuka, Y. Shiro, R. W. Strange, and S. S. Hasnain Iron Coordination Structures of Oxygen Sensor FixL Characterized by Fe K-edge Extended X-ray Absorption Fine Structure and Resonance Raman Spectroscopy J. Biol. Chem., August 13, 1999; 274(33): 23176 - 23184. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Masuda, Y. Matsumoto, K. V. P. Nagashima, K. Shimada, K. Inoue, C. E. Bauer, and K. Matsuura Structural and Functional Analyses of Photosynthetic Regulatory Genes regA and regB from Rhodovulum sulfidophilum, Roseobacter denitrificans, and Rhodobacter capsulatus J. Bacteriol., July 15, 1999; 181(14): 4205 - 4215. [Abstract] [Full Text]
|
|
|
|
 |
|
 B. L. Taylor and I. B. Zhulin PAS Domains: Internal Sensors of Oxygen, Redox Potential, and Light Microbiol. Mol. Biol. Rev., June 1, 1999; 63(2): 479 - 506. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Sartori, L. Aldegheri, G. Mazzotta, G. Lanfranchi, H. Tournu, A. J. P. Brown, and G. Carignani Characterization of a New Hemoprotein in the Yeast Saccharomyces cerevisiae J. Biol. Chem., February 19, 1999; 274(8): 5032 - 5037. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Aono, K. Ohkubo, T. Matsuo, and H. Nakajima Redox-controlled Ligand Exchange of the Heme in the CO-sensing Transcriptional Activator CooA J. Biol. Chem., October 2, 1998; 273(40): 25757 - 25764. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Tal, H. C. Wong, R. Calhoon, D. Gelfand, A. L. Fear, G. Volman, R. Mayer, P. Ross, D. Amikam, H. Weinhouse, et al. Three cdg Operons Control Cellular Turnover of Cyclic Di-GMP in Acetobacter xylinum: Genetic Organization and Occurrence of Conserved Domains in Isoenzymes J. Bacteriol., September 1, 1998; 180(17): 4416 - 4425. [Abstract] [Full Text]
|
|
|
|
|
|
H. Nakajima, Y. Honma, T. Tawara, T. Kato, S.-Y. Park, H. Miyatake, Y. Shiro, and S. Aono Redox Properties and Coordination Structure of the Heme in the CO-sensing Transcriptional Activator CooA J. Biol. Chem., March 2, 2001; 276(10): 7055 - 7061. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Stephenson and J. A. Hoch PAS-A domain of phosphorelay sensor kinase A: A catalytic ATP-binding domain involved in the initiation of development in Bacillus subtilis PNAS, December 18, 2001; 98(26): 15251 - 15256. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |