Philip Matsumura is in the Department of
Microbiology & Immunology, College of Medicine, University of Illinois at
Chicago, Chicago, Illinois 60612,
USA. matsumur@uic.edu
CheZ is a protein involved in signal transduction in bacterial
chemotaxis and accelerates the dephosphorylation of the activated response
regulator, CheY-phosphate. The crystal structure of CheZ in complex with
activated CheY provides insights into the function of CheZ.
Signal transduction in bacterial chemotaxis is a well-studied
phenomenon. With extensive biochemical and genetic information available, it is
the best-characterized two-component signal system. Partial or complete
structures have been determined for all of the signaling components from the
transmembrane receptor to the response regulator save CheZ. On
page 570 of this issue of Nature Structural
Biology, Zhao et al.1 report the co-crystal structure
of CheZ bound to its substrate, activated CheY, and provide the first
structural insight into the enigmatic protein.
Bacterial chemotaxis is an example of the two-component signal
transduction pathway that allows bacteria, lower eukaryotes and plant cells to
communicate with the environment. Bacterial such as Escherichia coli
have dozens of these systems to monitor a variety of environmental conditions.
Most of the two-component systems lead to the transcriptional regulation of
genes that code for proteins that allow the bacterial cell to cope with an
environmental stimulus. Typically, these systems contain a transmembrane
component that senses the environmental stimulus, as well as a histidine kinase
activity and phosphorylated response regulator that carries the signal to the
target.
The output of the chemotactic two-component system is the regulation of
the rotation of the flagella filament. In this system, a number of
methyl-accepting chemotaxis proteins are the transmembrane receptors of
extracellular signals; CheA is the histidine kinase; and CheY is the response
regulator. Phosphorylated CheY binds to the flagellar motor, specifically the
FliM protein in the basal body of the flagellar motor that extends into the
interior of the cell. Binding of CheY-phosphate causes the rotation of the
flagellar motor to change the sense of rotation from counterclockwise to
clockwise; dissociation of CheY-phosphate from the motor and dephosphorylation
reverts this action. Counterclockwise rotating flagella form a cohesive bundle
that propels the bacterium along a smooth path. Reversal of one or more
flagellar motors destablizes this bundle and causes the bacterium to change
directions. This regulation mechanism allows the bacterial cell to navigate
around its environment.
Unlike other chemotaxis proteins that make up the core of the
sensor-response regulator paradigm for two-component signaling systems, CheZ is
not represented in all organisms that display chemotaxis. McNamara and
Wolfe2 have shown that CheZ is found only in members of the
family Enterobacteriaceae. CheZ is almost always co-expressed with
CheAS (an N-terminally truncated form of CheA generated from an
internal translation start site at codon Met 98 of cheA), and existing
evidence indicates that binding of CheZ to CheAS leads to enhanced
dephosphorylation of CheY3. CheZ and CheAS found only
in enteric bacteria might be important for chemotaxis in this specialized
niche2, although it is not known why enteric bacteria would
require CheZ (or CheAS) to enhance an already rapid
auto-dephosphorylation of CheY-phosphate. Perhaps, in the steep gradients found
in the gut, a rapid quenching of the chemotaxis signal allows for better time
resolution in unstable environments. Aquatic bacteria are likely exposed to
shallow gradients and dilute concentrations of attractants and repellents.
CheZ can stimulate the dephosphorylation of CheY-phosphate, but because
CheZ appeared to only have this activity on CheY-phosphate, it was not clear if
CheZ was a CheY-phosphate specific phosphatase or an allosteric regulator of
the intrinsic dephosphorylation activity of CheY-phosphate. By co-crystallizing
CheZ with the BeF3- activated CheY, Zhao et
al.1 provide results that suggest that elements of both
mechanisms are present. The positioning of Gln 147 from CheZ at the active site
of the BeF3- activated CheY clearly suggested
that CheZ directly participates in the dephosphorylation reaction; however, it
does so by orienting a water molecule for nucleophilic attack that is thought
to be the mechanism for auto-dephosphorylation. The authors suggest that the
proposed mechanism of CheZ activity shares similarity with Ras GTPase.
CheZ has been long known to aggregate into multimers4, and
oligomerization of CheZ has recently been proposed as a mechanism of its
activation5. However, this property has made attempts at growing
crystals for X-ray diffraction problematic. Zhao et al.1
solved this problem by crystallizing CheZ in complex with its substrate,
activated CheY. Because CheY-phosphate is unstable, the authors used
BeF3- as a stable phosphate mimic. Unlike free
CheZ, CheZ−CheY formed well-diffracting crystals, demonstrating again
that sometimes stable complexes can be crystallized where the individual
components are more difficult. In this case, the binding of the
BeF3- activated CheY appeared to prevent
oligomerization of CheZ.
The structure reveals that two CheZ molecules form a long four-helix
bundle, and a single hairpin loop connects the two main helices in each CheZ
subunit. Two CheY molecules bind to the middle portion on opposite face of the
helical bundle. Based on previous characterization of cheZ mutants, Zhao
et al.1 mapped loss-of-function mutations to regions of
the molecule that contain the CheY binding sites and the hairpin loop
structure. The gain-of-function mutants map to the other half of the molecule,
as seen in Fig. 1c of ref. 1 Sourjik and Berg6 demonstrated that a
CheZ-YFP fusion protein localizes to the polar chemoreceptor patch in E.
coli. Cantwell and Manson (pers. comm.), using a slightly different
CheZ-GFP fusion, have extended this work to show that localization requires the
presence of CheAS. By examining a wide spectrum of the available
cheZ missense mutations, as well as through their own mutational
analysis, Cantwell and Manson found that the hairpin loop comprising residues
80−120 targets CheZ to the receptor patch (Fig. 1).
These results suggest a function for the long helix-turn-helix extension seen
in the CheZ structure.
Figure 1. Schematic view of the helix-turn-helix hairpin of CheZ.
In the CheZ dimer, the hairpins in each subunit flair away
from each other, leaving them free to interact with other proteins. The
hydrophobic residues of the two amphipathic helices face inward and polar
groups face outward, except that there are several exposed hydrophobic residues
in or adjacent to the loop. The phenotypes of mutations affecting different
regions of this structure, which includes residues 83−121, are indicated
by different shadings, as shown in the figure. All of the mutated residues that
impair chemotaxis are located at the helix−helix packing face or the
subunit interface.
The co-crystal of CheZ and CheY also provides insight into the
interactions between CheZ and CheY. Results from the Eisenbach lab suggested
that CheZ does not act on CheY-phosphate when CheY-phosphate is bound to the
motor7. Further, NMR studies have shown that an N-terminal
peptide of FliM and the C-terminal peptide from CheZ affect the resonances of
the same residues in CheY upon binding8. In the structure of
CheZ−CheY1, the C-terminal end of CheZ is bound to the same
region of activated CheY as the FliM peptide9 (Fig. 2). Zhao et al.1 also show that the
C-terminal end of CheZ is required for CheZ−CheY interactions; other
interactions between the two proteins are not sufficient by themselves to hold
the complex together. Since both the motor protein FliM and CheZ bind to CheY
in the same area, it would be difficult to have simultaneous binding of both
proteins. The structural results are therefore consistent with CheZ acting on
CheY-phosphate when it is not attached to the flagellar motor.
Figure 2. Comparison of CheY−CheZ and CheY−FliM
interactions.
Activated CheY is shown as a ribbon diagram and CheZ and
FliM peptide as surface models. Tyr 106 of CheY, in its 'in' position, is shown
as a stick model. a, Overall structure of activated
CheY−BeF3- interacting with CheZ.
b, Activated CheY−BeF3- bound to
the N-terminus of FliM9. In both cases the interaction take place
between helix 4, -strand 4 and helix 5 of CheY. Coordinates for the CheY
bound to CheZ were obtained from Zhao et al.1. Coordinates
for the CheY bound to the N-terminus of FliM were obtained from the Protein
Data Bank (accession code 1F4V).
The side chain of Tyr 106 in CheY can adopt two rotameric
conformations10. It is noteworthy that this side chain is in the
'in' or hydrophobic position in both CheZ−CheY and CheY−FliM
peptide structures. Since the side chain in the 'out' or hydrophilic position
would impose steric interference with the bound polypeptides, the movement or
the restriction of this rotamer is a prerequisite for binding to both FliM and
CheZ. Therefore, the position of Tyr 106 side chain of CheY is important to
CheZ dephosphorylation activity as well as communicating the signal from CheY
to the motor.
In summary, the structure of the CheZ dimer bound to two activated CheY
molecules has given insight into the mechanism of its activity. With the
structure, future studies can begin to answer questions of how this activity is
regulated and why enteric bacteria need this enhanced activity to cope with
their environment.
-strand 4 and helix 5 of CheY. Coordinates for the CheY
bound to CheZ were obtained from Zhao et al.
