Signal Transduction in Bacteria

Ann Stock, Ph.D.

Howard Hughes Medical Institute and The University of Medicine and Dentistry of New Jersey

Robert Wood Johnson Medical School

TWO-COMPONENT SIGNAL TRANSDUCTION

The focus of this presentation will be signal transduction in bacteria. Signal transduction is something that we commonly associate with eukaryotic systems, but nowhere is it more important than in the prokaryotic world where single-celled organisms such as bacteria are constantly faced with dramatic changes in their environment. In order to be able to compete, survive, and thrive in their environment, they need to elicit appropriate adaptive responses to changing environmental conditions.

A majority of signal transduction in bacteria is mediated by phosphotransfer signaling systems, commonly referred to as “two-component” signal transduction pathways. These systems involve two conserved protein components. The first component is a histidine protein kinase that autophosphorylates at a histidine residue, creating a high-energy phosphoryl group that is poised for transfer to the next component. The phosphorylation of the kinase is not stoichiometric; it does not regulate the kinase activity as does phosphorylation of eukaryotic kinases. Rather, in this particular case, phosphorylation serves to provide phosphoryl groups for transfer to the second component, a protein called the response regulator.

In bacteria, these systems largely replace the small GTPase signaling pathways (such as those involving members of the ras super family) that are prevalent in eukaryotes. The response regulator functions as a phosphorylation-activated switch to control the output responses. The



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 80
The Markey Scholars Conference: Proceedings Signal Transduction in Bacteria Ann Stock, Ph.D. Howard Hughes Medical Institute and The University of Medicine and Dentistry of New Jersey Robert Wood Johnson Medical School TWO-COMPONENT SIGNAL TRANSDUCTION The focus of this presentation will be signal transduction in bacteria. Signal transduction is something that we commonly associate with eukaryotic systems, but nowhere is it more important than in the prokaryotic world where single-celled organisms such as bacteria are constantly faced with dramatic changes in their environment. In order to be able to compete, survive, and thrive in their environment, they need to elicit appropriate adaptive responses to changing environmental conditions. A majority of signal transduction in bacteria is mediated by phosphotransfer signaling systems, commonly referred to as “two-component” signal transduction pathways. These systems involve two conserved protein components. The first component is a histidine protein kinase that autophosphorylates at a histidine residue, creating a high-energy phosphoryl group that is poised for transfer to the next component. The phosphorylation of the kinase is not stoichiometric; it does not regulate the kinase activity as does phosphorylation of eukaryotic kinases. Rather, in this particular case, phosphorylation serves to provide phosphoryl groups for transfer to the second component, a protein called the response regulator. In bacteria, these systems largely replace the small GTPase signaling pathways (such as those involving members of the ras super family) that are prevalent in eukaryotes. The response regulator functions as a phosphorylation-activated switch to control the output responses. The

OCR for page 80
The Markey Scholars Conference: Proceedings phosphorylated protein has a relatively short lifetime. The protein is phosphorylated at an aspartate residue, a high-energy modification that has a half-life of several hours. The chemical stability of the aspartyl phosphate is often further reduced in these systems by both intrinsic autophosphatase activity within the response regulator and by auxiliary phosphatase proteins, resulting in half-lives that range from seconds to hours in different signaling systems. Two-component systems are very widespread throughout the bacterial kingdom. In a typical genome there are about twenty to thirty of these regulatory systems that are involved in many different functions. Some of these are housekeeping functions, but in addition, these proteins are often involved in the regulation of expression of various types of toxins and virulence factors, and in mediating antibiotic resistance through a variety of different mechanisms. This regulation is important for host-pathogen interactions. These proteins are present and abundant in bacteria and are present to a limited extent in some eukaryotic cells. They have been found in yeasts, and in the slime mold Dictyostelium. They are quite prevalent in plants, but they have not been identified in animals. And this, of course, makes them promising targets for the development of antimicrobial agents. There are a very large number of these systems that have been identified. There are well over a thousand different two-component systems that have been found in a variety of bacterial genomes; not surprisingly, like components of any signal transduction system, these proteins are incredibly modular in nature. The conserved domains can be assembled in a variety of different ways into proteins, and the proteins can be assembled in a variety of ways into pathways, creating diverse and complex schemes. In a typical system, consisting of just two components, the histidine kinase is a transmembrane protein with a variable extracellular sensing domain and a conserved intracellular kinase domain. Phosphoryl transfer occurs to the conserved domain of the response regulator that in turn, controls the activity of an associated variable effector domain. The response regulator protein is often a transcription factor that regulates expression of a specific group of genes. In other systems, these components can be arrayed in a much more complex fashion with multiple kinases feeding into single response regulators, single kinases feeding into multiple-response regulators, and in many cases, multiple modules allowing for multiple transfers from histidine to aspartate residues in what are known as phospho-relay schemes. In eukaryotic two-component systems, this kind of multiple phospho-relay is the norm rather than the exception and the phospho-relay systems

OCR for page 80
The Markey Scholars Conference: Proceedings often interface with more conventional eukaryotic signaling pathways. For instance, in the yeast osmoregulation system, the response regulator feeds into a Map kinase cascade. In Dictyostelium, one response regulator contains an effector domain with phosphodiesterase activity and the two-component pathway ties into a cyclic-neucleotide signaling cascade. The modularity of these proteins allows us to study them as individual components. The family of response regulators, with well over a thousand members to study, provides an opportunity to understand similarities and differences in the regulation of very large families. Response regulators have been the focus of research in our laboratory for a number of years, with one particular question in mind. How has nature designed a generic phosphorylation-activated switch that is capable of controlling the activity of a variety of different effector domains that are quite diverse with respect to both structure and function? Several different types of effector domains have been characterized biochemically and structurally. Examples of distinct effector domains that can be controlled by this common regulatory domain include different subfamilies of DNA-binding domains, as well as enzymes. We know a substantial amount about the functioning of the isolated regulatory domain from research that has accumulated in a number of laboratories over quite a number of years. The domain is a simple doubly-wound beta/alpha fold. There is a cluster of highly conserved aspartic acid residues that coordinates a Mg2+, creating the active site of the protein. It is this conserved response regulator domain that is actually responsible for the catalysis of phosphoryl transfer from the histidine kinase, specifically to an aspartic acid residue that is located at the C-terminal end of the third and central beta strand. There is also another set of conserved residues that are located on adjacent beta strands 3445, creating a diagonal path leading away from the active site. These residues are a hydroxyl containing residue, either serine or threonine, and an aromatic residue, either phenylalanine or tyrosine. Over the past couple of years the roles of these conserved residues have been illuminated by structures of several isolated regulatory domains in their active states, determined in a number of different laboratories. These additional conserved residues leading from the active site function as switches that help drive a conformational change in the domain. The conformations of these switch residues differ in inactive and active states of the domain. In the phosphorylated or activated state, the hydroxyl of the serine or threonine is positioned to form a hydrogen bond with the phosphate oxygen of the phosphorylated aspartate. The phenylalanine or tyrosine has flipped from an outward orientation to an inward orientation occupying the cavity that has been vacated by the movement

OCR for page 80
The Markey Scholars Conference: Proceedings of the serine or threonine. An animation demonstrates the reorientation of these residues in the active and inactive states of the regulatory domain. A minor conformational change in the backbone accompanies the flipping of these residues. The reorientation of the residues creates a conserved mechanism for the propagation of a conformational change that, in the study of four different proteins, has been found to affect a relatively large surface, involving approximately half of the molecule. The regions showing the most change involve the 3445 surface of the domain. The backbone deviations in the region that changes conformation as a result of phosphorylation range from one to several Å, at most. The changes are relatively subtle, but they create a sufficiently distinct surface that the different surfaces can be exploited for protein-protein interactions that are specific to one state or another. This allows the regulatory domain to function as a relatively versatile switch for regulation of events that can be controlled by differential protein-protein interactions. Our laboratory has been interested in trying to understand the details of how the altered surfaces of the regulatory domain are used for regulating different effector domain functions. A large number of structures of isolated regulatory and effector domains have been determined. There are dozens of structures of isolated regulatory domains and about a half dozen structures of isolated effector domains in the Protein Data Bank. But the structures of intact multidomain-response regulators have been relatively resistant to structural analysis. Thus the majority of what we know about the regulation of effector domain activity by regulatory domains has come from a variety of different biochemical rather than structural approaches. Again, this family with well over a thousand members allows us an opportunity to begin to ask: “within a conserved structural family, how similar or different are the mechanisms of regulation?” If we understand how one of these proteins is regulated, does it provide an understanding of how other proteins in the same family are regulated? This is a question that is becoming increasingly important in this era of structural genomics and proteomics, where we would like to be able to extrapolate common mechanisms of function from known structural relationships. I would like to discuss two brief stories about some of the studies that we have conducted with two representative response regulator proteins. One of them is the chemotaxis methylesterase CheB and the other, is a member of the OmpR/PhoB family of transcription factors. REGULATION OF CHEMOTAXIS METHYLESTERASE CHEB Before I get into the details of the methylesterase CheB, let me tell you about the system in which it functions. It belongs to the pathway of

OCR for page 80
The Markey Scholars Conference: Proceedings bacterial chemotaxis, which is a very extensively studied signal transduction system that is built on a two-component phosphotransfer scheme. Chemotaxis has long been advertised as a relatively simple sensory system for the study of signal transduction. However, it actually has some relatively complex regulatory features, one of which is the phenomenon of adaptation. The receptors in the chemotaxis system signal in response to changes in ligand occupancy, not in response to absolute ligand occupancy. That is, if we follow the behavior of cells upon administering an attractant stimulus, going from a low level of stimulus up to a higher level of stimulus, we see that the cells have a steady state behavior, followed by a transient physiological response, and then adapt back to their pre-stimulus behavior. To the extent that this is occurring at the level of the receptors, this implies that receptors at a low level of occupancy and at a high level of occupancy are capable of giving the same signaling output. In this system that signaling output is the regulation of the activity of the histidine kinase. We know that this phenomenon of adaptation is facilitated by the reversible covalent modification of receptors, specifically the methylation and demethylation of specific glutamate residues within the cytoplasmic domains of the receptors. In a very simple scheme, one can view the mechanism of adaptation in terms of a “balance model” for the receptor. In this scheme, the signaling activity of the receptor is reflective of both the ligand bound state of the receptor, and the modification state of the cytoplasmic domain. The pre-stimulus steady state is characterized by a balance between these two features. Upon addition of attractant, ligand occupancy outweighs methylation and the signaling state of the receptor is perturbed, leading to the response. Subsequently, methylation increases, counterbalancing the increased ligand and leading to a return to the pre-stimulus signaling state of the receptor. Thus adaptation is achieved, with a higher level of ligand occupancy compensated for by an increase in the level of receptor methylation. This, of course, requires that there is a very tight coupling between changes in chemo-effector level and the methylation system. So the methylation system is highly regulated. Reversible methylation involves two enzymes: a methyltransferase and a methylesterase. The majority of the regulation is contributed by the demethylating enzyme, the methylesterase, and involves a two-component pathway. In the chemotaxis system there are actually two-response regulators that obtain phosphoryl groups from the single histidine kinase. The chemoreceptors, through a coupling protein, control the activity of the histidine kinase CheA, which passes phosphoryl groups to CheY, which acts at the flagellar motor to induce the physiological response. At the same time, phosphoryl groups are passed to the methylesterase CheB, activating it to demethylate the receptors and attenuate the response.

OCR for page 80
The Markey Scholars Conference: Proceedings So how does phosphorylation work to activate the methylesterase? We know that the regulatory domain actually plays two roles in regulation of methylesterase activity. The intact unphosphorylated methylesterase has very low receptor demethylation activity. If the unphosphorylated regulatory domain is removed by genetic deletion or proteolysis, a 10-fold increase in demethylation activity is observed, indicating that the unphosphorylated regulatory domain serves as an inhibitor of the catalytic domain. The demethylation activity of the phosphorylated intact protein is ten-fold greater than that of the isolated catalytic domain. The phosphorylated domain enhances the activity of the catalytic domain. So there are actually two roles for the regulatory domain: an inhibitory activity of the unphosphorylated domain, and an activating function of the phosphorylated domain. The crystal structure of CheB, determined by postdoctoral fellows Senzana Djorjdevic and Ann West, has provided insight to the basis of the inhibitory interaction. A space-filling model of structure of the intact methylesterase illustrates a very tight packing of the regulatory domain against the catalytic domain, almost completely blocking access to the catalytic triad of the active site. In a modeling experiment, we attempted to dock the methylation region of the receptor, specifically a helix with glutamate residues, near the active site serine nucleophile of the catalytic domain. However, the glutamate residues cannot approach closer than ~7 Å without steric collision. The interface between the regulatory and catalytic domains is a tightly packed interface that has extensive contacts, hydrophobic in the center and more hydrophilic at the periphery. Notably, the core of the domain interface is formed by a pair of phenylalanine residues. The phenylalanine contributed by the regulatory domain is the conserved aromatic residue that in phosphorylated response regulators is part of the switch mechanism, flipping from an outward orientation in the inactive state to an inward orientation in the phosphorylated state. We anticipate that reorientation of this residue would disrupt the domain interface, exposing the active site for interaction with the receptor substrate. We wanted to probe this mechanism of activation. However, the short lifetime of the phosphorylated state of this protein, with a half-life of only one second, makes it relatively difficult to approach the structure of the phosphorylated state by conventional means. So Ganesh Anand, a graduate student in my laboratory, collaborated with Betsy Komives at UCSD to use a method of deuterium exchange analyzed by mass spectrometry to characterize solvent accessibility at the domain interface upon phosphorylation. In these deuterium exchange experiments, the protein is quenched at low pH, digested with pepsin, and then peptic fragments are isolated and

OCR for page 80
The Markey Scholars Conference: Proceedings analyzed by mass spectrometry. If the protein is briefly incubated in the presence of deuterium before this analysis, deuterons are exchanged for protons at solvent accessible backbone amides. The deuterium-substituted peptides, indicative of regions of increased solvent exposure, are shifted to higher masses relative to their hydrogen-containing counterparts. Various times of incubation in deuterium can be employed to generate rates of deuterium incorporation into different peptides. This analysis was performed on the methylesterase CheB in the absence of phosphorylation and under conditions of steady-state phosphorylation. As is typical of this method, there were a small number of peptides for which no quantitative data could be determined, but we obtained fairly complete coverage. The majority of the peptides showed no changes in deuterium-exchange rates upon phosphorylation. Two peptides showed increased solvent exchange upon phosphorylation. Both were located within the catalytic domain at the interdomain interface. When mapped onto the surface of the catalytic domain, it is apparent that these peptides form two edges of the domain interface. These regions become more solvent exposed upon phosphorylation, but not nearly as solvent exposed as what we observed in the isolated catalytic domain. These data indicate that the domains do readjust their orientation in the phosphorylated state. However, they do not separate completely. There is a substantial interface that remains between the two domains, perhaps providing a path for allosteric communication between the regulatory domain and the active site of the catalytic domain. Such contact may mediate the 10-fold activation contributed by the phosphorylated domain. THE OMPR/PHOB FAMILY OF TRANSCRIPTION FACTORS The chemotaxis system is a somewhat atypical two-component system. The two response regulators of the chemotaxis system are unusual in that they are not transcription factors. The vast majority of response regulators in two-component systems are responsible for regulation of gene expression and the effector domains of these response regulators are DNA-binding domains. The bacterial response regulators can be divided into subfamilies based on homology within their C-terminal DNA-binding domains. A survey of the E. coli genome conducted by Mizuno several years ago, revealed that out of the 32 response regulators in E. coli, 25 of them could be categorized as belonging to the previously identified three subfamilies of transcription factors based on homology within their DNA-binding domains. One of these three families, the OmpR/PhoB family, accounts for approximately 40 percent of all response regulators in bacte-

OCR for page 80
The Markey Scholars Conference: Proceedings rial genomes. This subfamily is characterized by a DNA-binding domain with a winged-helix fold. The OmpR/PhoB group is by far the largest subfamily. There are two other smaller subfamilies of transcription factors with different types of DNA-binding domains. A small number of proteins are unrelated to other response regulators and are grouped into a miscellaneous category. The two chemotaxis proteins discussed above are members of this group. We are interested in the large family of transcription factors represented by the OmpR/PhoB subfamily. The family is named after the OmpR protein that participates in osmoregulation in E. coli. The osmoregulation system involves a histidine kinase EnvZ that is a transmembrane kinase that senses changes in osmotic strength. EnvZ provides phosphoryl groups to the response regulator OmpR, which functions as a transcription factor. The C-terminal, or effector, domain of this response regulator is a DNA-binding domain, and OmpR, in its phosphorylated state, binds in a hierarchical fashion to binding sites that are located upstream of the genes that encode the major outer membrane porin proteins, OmpF and OmpC. These proteins are regulated not in a single on-off switch fashion, but in a much more subtle way. At low osmotic strength, there is a high level of expression of ompF. At high-osmotic strength, there is a high level of expression of ompC. Expression of these genes is differentially coordinated so that the total concentration of porin proteins remains constant while the composition varies. The OmpR protein serves as both an activator and repressor of transcription, depending on the binding sites that are occupied. Each of the binding sites preceding the ompF and ompC genes consists of two ten-base pair half-sites arranged as direct repeats. Structural analysis of the DNA-binding domain of OmpR by Erik Martínez-Hackert, a graduate student in my laboratory, defined the winged-helix fold for this family. Several experiments including site-specific cleavage analysis allowed us to predict a model for the interaction of this DNA-binding domain with its recognition sites in DNA. This model, though useful, is incomplete. In the cell, the phosphorylated regulatory domain is required for efficient DNA binding and transcriptional regulation. We would like to understand the specific role of the regulatory domain in allowing the DNA-binding domain to interact with its DNA sites. Despite the fact that this is a huge family of transcription factors with over 500 members, and a number of laboratories have been pursuing structures of different family members, these proteins have been quite resistant to crystallization. David Buckler, a postdoctoral fellow in my laboratory, finally made progress in this pursuit by using proteins from the thermophilic bacterium, Thermotoga maritima. When the Thermotoga genome was reported, it was possible to clone all of the response regula-

OCR for page 80
The Markey Scholars Conference: Proceedings tor members that belonged to the OmpR/PhoB family. There are four of these proteins. David expressed and purified all of them, and was lucky enough to be able to crystallize one. The structure of this protein, DrrD, revealed no particular surprises with regard to the folds of the individual domains. What was of interest was the orientation of the regulatory domain with respect to the DNA-binding domain. In this regard, the structure was surprising. In the crystal structure, the regulatory domain packs against the DNA-binding domain via an extremely small interface. There are only 250 Å2 of buried surface area. This value falls substantially below the area of any bona fide domain interface observed in any monomeric protein in the Protein Data Bank. The interface in DrrD is much smaller than the typical size of a domain interface, which would be ~1200 Å2. Of course, in a crystal structure, there needs to be an interaction between domains. Proteins have to be able to pack within the crystal lattice. We believe that crystallization has trapped these two domains in a fixed orientation that is not reflective of how they would exist in solution. The lack of a unique conformation would help explain why this particular OmpR/PhoB subfamily of response regulators has been so resistant to crystallization. Notably, within the very short linker that connects the two domains of DrrD together, there are two missing residues that lack electron density and the surrounding residues in the linker region have very high temperature factors. Again, these data suggest that this domain interface is not a stable interface, and that the linker that tethers the two domains to each other is flexible. The lack of a substantial domain interface makes DrrD unique among the three response regulators that have been crystallized as intact proteins to date. Comparing the buried surface areas at the domain interfaces of these three response regulators, two have substantial interfaces of ~1000 Å2. One of these is the chemotaxis methylesterase CheB, discussed above. The other is NarL, a transcription factor belonging to a completely different subfamily of response regulators. The structure of NarL, determined by Dick Dickerson’s group, reveals an extensive domain interface. These large interfaces in these two structures contrast with the very tiny patch of buried surface between the two domains of DrrD, implying that these proteins utilize different types of regulatory interactions. The two response regulators for which structures have been known for several years have served as a model for understanding interdomain regulation. In both cases, there are large domain interfaces and the regulatory domains pack against the C-terminal effector domains, essentially precluding access to the functional regions of the effector domains: the active site of the methylesterase CheB, and the recognition helix of the transcription factor NarL. In contrast, in DrrD, the transcription factor of

OCR for page 80
The Markey Scholars Conference: Proceedings the OmpR/PhoB subfamily, the recognition helix is completely unobstructed by the regulatory domain. The structure of DrrD suggests a model for regulation that is fundamentally different than that observed for other response regulators. Rather than a mechanism of regulation involving intra-molecular communication between the regulatory and effector domains within a monomer, the regulation within DrrD, and perhaps other members of the OmpR/PhoB subfamily, appears to occur through inter-molecular interactions. Specifically, activation is proposed to proceed primarily through dimerization of the phosphorylated regulatory domains with the effector domains participating as passive partners. Chimeric response regulators within the OmpR/PhoB subfamily, involving the regulatory domain of one member attached to the effector domain of another, have been constructed and function well, providing further support for this type of mechanism. These experiments have been performed with a limited number of OmpR/PhoB subfamily members at this point. One of the questions that remains at this time is whether within the subfamily, all members will function by a similar regulatory mechanism or whether within this large family, different members will have different mechanisms of regulation. This study has indicated that within large structural families, although sequence and structural similarity exist, there may be some very significant differences in the way proteins function and in particular, in the specific protein-protein interaction mechanisms that are being used for regulation in signal transduction proteins. ACKNOWLEDGMENTS The people associated with studies of methylesterase CheB were Ganesh Anand, a graduate student who performed the biochemical analyses; and Ann West and Snezana Djordjevic, postdoctoral fellows who determined the structure. The deuterium exchange study was done in collaboration with Betsy Komives at UCSD. The people associated with analyses of the OmpR family transcription factors were Erik Martínez-Hackert and Patricia Harrison-McMonagle, graduate students who determined the structure of the DNA-binding domain and performed the cleavage analysis that allowed us to model the interaction with DNA; and David Buckler, a postdoctoral fellow who determined the structure of the intact OmpR family member, DrrD. Financial support was provided by the NIH and HHMI.